git-commit-vandalism/unpack-trees.c

3032 lines
83 KiB
C
Raw Normal View History

#include "cache.h"
#include "strvec.h"
#include "repository.h"
#include "config.h"
#include "dir.h"
#include "tree.h"
#include "tree-walk.h"
#include "cache-tree.h"
#include "unpack-trees.h"
#include "progress.h"
#include "refs.h"
#include "attr.h"
#include "split-index.h"
#include "sparse-index.h"
#include "submodule.h"
#include "submodule-config.h"
#include "fsmonitor.h"
#include "object-store.h"
#include "promisor-remote.h"
#include "entry.h"
unpack-trees: add basic support for parallel checkout This new interface allows us to enqueue some of the entries being checked out to later uncompress them, apply in-process filters, and write out the files in parallel. For now, the parallel checkout machinery is enabled by default and there is no user configuration, but run_parallel_checkout() just writes the queued entries in sequence (without spawning additional workers). The next patch will actually implement the parallelism and, later, we will make it configurable. Note that, to avoid potential data races, not all entries are eligible for parallel checkout. Also, paths that collide on disk (e.g. case-sensitive paths in case-insensitive file systems), are detected by the parallel checkout code and skipped, so that they can be safely sequentially handled later. The collision detection works like the following: - If the collision was at basename (e.g. 'a/b' and 'a/B'), the framework detects it by looking for EEXIST and EISDIR errors after an open(O_CREAT | O_EXCL) failure. - If the collision was at dirname (e.g. 'a/b' and 'A'), it is detected at the has_dirs_only_path() check, which is done for the leading path of each item in the parallel checkout queue. Both verifications rely on the fact that, before enqueueing an entry for parallel checkout, checkout_entry() makes sure that there is no file at the entry's path and that its leading components are all real directories. So, any later change in these conditions indicates that there was a collision (either between two parallel-eligible entries or between an eligible and an ineligible one). After all parallel-eligible entries have been processed, the collided (and thus, skipped) entries are sequentially fed to checkout_entry() again. This is similar to the way the current code deals with collisions, overwriting the previously checked out entries with the subsequent ones. The only difference is that, since we no longer create the files in the same order that they appear on index, we are not able to determine which of the colliding entries will survive on disk (for the classic code, it is always the last entry). Co-authored-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Co-authored-by: Jeff Hostetler <jeffhost@microsoft.com> Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2021-04-19 02:14:53 +02:00
#include "parallel-checkout.h"
unpack-trees: allow Porcelain to give different error messages The plumbing output is sacred as it is an API. We _could_ change it if it is broken in such a way that it cannot convey necessary information fully, but we just do not _reword_ for the sake of rewording. If somebody does not like it, s/he is complaining too late. S/he should have been here in early May 2005 and make the language used by the API closer to what humans read. S/he wasn't here. Too bad, and it is too late. And people who complain should look at a bigger picture. Look at what was suggested by one of them and think for five seconds: $ git checkout mytopic -fatal: Entry 'frotz' not uptodate. Cannot merge. +fatal: Entry 'frotz' has local changes. Cannot merge. If you do not see something wrong with this output, your brain has already been rotten with use of git for too long a time. Nobody asked us to "merge" but why are we talking about "Cannot merge"? This patch introduces a mechanism to allow Porcelains to specify messages that are different from the ones that is given by the underlying plumbing implementation of read-tree, so that we can reword the message Porcelains give without disrupting the output from the plumbing. $ git-checkout pu error: You have local changes to 'Makefile'; cannot switch branches. There are other places that ask unpack_trees() to n-way merge, detect issues and let it issue error message on its own, but I did this as a demonstration and replaced only one message. Yes I know about C99 structure initializers. I'd love to use them but we try to be nice to compilers without it. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-05-17 21:03:49 +02:00
/*
* Error messages expected by scripts out of plumbing commands such as
* read-tree. Non-scripted Porcelain is not required to use these messages
* and in fact are encouraged to reword them to better suit their particular
* situation better. See how "git checkout" and "git merge" replaces
* them using setup_unpack_trees_porcelain(), for example.
unpack-trees: allow Porcelain to give different error messages The plumbing output is sacred as it is an API. We _could_ change it if it is broken in such a way that it cannot convey necessary information fully, but we just do not _reword_ for the sake of rewording. If somebody does not like it, s/he is complaining too late. S/he should have been here in early May 2005 and make the language used by the API closer to what humans read. S/he wasn't here. Too bad, and it is too late. And people who complain should look at a bigger picture. Look at what was suggested by one of them and think for five seconds: $ git checkout mytopic -fatal: Entry 'frotz' not uptodate. Cannot merge. +fatal: Entry 'frotz' has local changes. Cannot merge. If you do not see something wrong with this output, your brain has already been rotten with use of git for too long a time. Nobody asked us to "merge" but why are we talking about "Cannot merge"? This patch introduces a mechanism to allow Porcelains to specify messages that are different from the ones that is given by the underlying plumbing implementation of read-tree, so that we can reword the message Porcelains give without disrupting the output from the plumbing. $ git-checkout pu error: You have local changes to 'Makefile'; cannot switch branches. There are other places that ask unpack_trees() to n-way merge, detect issues and let it issue error message on its own, but I did this as a demonstration and replaced only one message. Yes I know about C99 structure initializers. I'd love to use them but we try to be nice to compilers without it. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-05-17 21:03:49 +02:00
*/
static const char *unpack_plumbing_errors[NB_UNPACK_TREES_WARNING_TYPES] = {
/* ERROR_WOULD_OVERWRITE */
unpack-trees: allow Porcelain to give different error messages The plumbing output is sacred as it is an API. We _could_ change it if it is broken in such a way that it cannot convey necessary information fully, but we just do not _reword_ for the sake of rewording. If somebody does not like it, s/he is complaining too late. S/he should have been here in early May 2005 and make the language used by the API closer to what humans read. S/he wasn't here. Too bad, and it is too late. And people who complain should look at a bigger picture. Look at what was suggested by one of them and think for five seconds: $ git checkout mytopic -fatal: Entry 'frotz' not uptodate. Cannot merge. +fatal: Entry 'frotz' has local changes. Cannot merge. If you do not see something wrong with this output, your brain has already been rotten with use of git for too long a time. Nobody asked us to "merge" but why are we talking about "Cannot merge"? This patch introduces a mechanism to allow Porcelains to specify messages that are different from the ones that is given by the underlying plumbing implementation of read-tree, so that we can reword the message Porcelains give without disrupting the output from the plumbing. $ git-checkout pu error: You have local changes to 'Makefile'; cannot switch branches. There are other places that ask unpack_trees() to n-way merge, detect issues and let it issue error message on its own, but I did this as a demonstration and replaced only one message. Yes I know about C99 structure initializers. I'd love to use them but we try to be nice to compilers without it. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-05-17 21:03:49 +02:00
"Entry '%s' would be overwritten by merge. Cannot merge.",
/* ERROR_NOT_UPTODATE_FILE */
unpack-trees: allow Porcelain to give different error messages The plumbing output is sacred as it is an API. We _could_ change it if it is broken in such a way that it cannot convey necessary information fully, but we just do not _reword_ for the sake of rewording. If somebody does not like it, s/he is complaining too late. S/he should have been here in early May 2005 and make the language used by the API closer to what humans read. S/he wasn't here. Too bad, and it is too late. And people who complain should look at a bigger picture. Look at what was suggested by one of them and think for five seconds: $ git checkout mytopic -fatal: Entry 'frotz' not uptodate. Cannot merge. +fatal: Entry 'frotz' has local changes. Cannot merge. If you do not see something wrong with this output, your brain has already been rotten with use of git for too long a time. Nobody asked us to "merge" but why are we talking about "Cannot merge"? This patch introduces a mechanism to allow Porcelains to specify messages that are different from the ones that is given by the underlying plumbing implementation of read-tree, so that we can reword the message Porcelains give without disrupting the output from the plumbing. $ git-checkout pu error: You have local changes to 'Makefile'; cannot switch branches. There are other places that ask unpack_trees() to n-way merge, detect issues and let it issue error message on its own, but I did this as a demonstration and replaced only one message. Yes I know about C99 structure initializers. I'd love to use them but we try to be nice to compilers without it. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-05-17 21:03:49 +02:00
"Entry '%s' not uptodate. Cannot merge.",
/* ERROR_NOT_UPTODATE_DIR */
unpack-trees: allow Porcelain to give different error messages The plumbing output is sacred as it is an API. We _could_ change it if it is broken in such a way that it cannot convey necessary information fully, but we just do not _reword_ for the sake of rewording. If somebody does not like it, s/he is complaining too late. S/he should have been here in early May 2005 and make the language used by the API closer to what humans read. S/he wasn't here. Too bad, and it is too late. And people who complain should look at a bigger picture. Look at what was suggested by one of them and think for five seconds: $ git checkout mytopic -fatal: Entry 'frotz' not uptodate. Cannot merge. +fatal: Entry 'frotz' has local changes. Cannot merge. If you do not see something wrong with this output, your brain has already been rotten with use of git for too long a time. Nobody asked us to "merge" but why are we talking about "Cannot merge"? This patch introduces a mechanism to allow Porcelains to specify messages that are different from the ones that is given by the underlying plumbing implementation of read-tree, so that we can reword the message Porcelains give without disrupting the output from the plumbing. $ git-checkout pu error: You have local changes to 'Makefile'; cannot switch branches. There are other places that ask unpack_trees() to n-way merge, detect issues and let it issue error message on its own, but I did this as a demonstration and replaced only one message. Yes I know about C99 structure initializers. I'd love to use them but we try to be nice to compilers without it. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-05-17 21:03:49 +02:00
"Updating '%s' would lose untracked files in it",
/* ERROR_CWD_IN_THE_WAY */
"Refusing to remove '%s' since it is the current working directory.",
/* ERROR_WOULD_LOSE_UNTRACKED_OVERWRITTEN */
"Untracked working tree file '%s' would be overwritten by merge.",
unpack-trees: allow Porcelain to give different error messages The plumbing output is sacred as it is an API. We _could_ change it if it is broken in such a way that it cannot convey necessary information fully, but we just do not _reword_ for the sake of rewording. If somebody does not like it, s/he is complaining too late. S/he should have been here in early May 2005 and make the language used by the API closer to what humans read. S/he wasn't here. Too bad, and it is too late. And people who complain should look at a bigger picture. Look at what was suggested by one of them and think for five seconds: $ git checkout mytopic -fatal: Entry 'frotz' not uptodate. Cannot merge. +fatal: Entry 'frotz' has local changes. Cannot merge. If you do not see something wrong with this output, your brain has already been rotten with use of git for too long a time. Nobody asked us to "merge" but why are we talking about "Cannot merge"? This patch introduces a mechanism to allow Porcelains to specify messages that are different from the ones that is given by the underlying plumbing implementation of read-tree, so that we can reword the message Porcelains give without disrupting the output from the plumbing. $ git-checkout pu error: You have local changes to 'Makefile'; cannot switch branches. There are other places that ask unpack_trees() to n-way merge, detect issues and let it issue error message on its own, but I did this as a demonstration and replaced only one message. Yes I know about C99 structure initializers. I'd love to use them but we try to be nice to compilers without it. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-05-17 21:03:49 +02:00
/* ERROR_WOULD_LOSE_UNTRACKED_REMOVED */
"Untracked working tree file '%s' would be removed by merge.",
unpack-trees: allow Porcelain to give different error messages The plumbing output is sacred as it is an API. We _could_ change it if it is broken in such a way that it cannot convey necessary information fully, but we just do not _reword_ for the sake of rewording. If somebody does not like it, s/he is complaining too late. S/he should have been here in early May 2005 and make the language used by the API closer to what humans read. S/he wasn't here. Too bad, and it is too late. And people who complain should look at a bigger picture. Look at what was suggested by one of them and think for five seconds: $ git checkout mytopic -fatal: Entry 'frotz' not uptodate. Cannot merge. +fatal: Entry 'frotz' has local changes. Cannot merge. If you do not see something wrong with this output, your brain has already been rotten with use of git for too long a time. Nobody asked us to "merge" but why are we talking about "Cannot merge"? This patch introduces a mechanism to allow Porcelains to specify messages that are different from the ones that is given by the underlying plumbing implementation of read-tree, so that we can reword the message Porcelains give without disrupting the output from the plumbing. $ git-checkout pu error: You have local changes to 'Makefile'; cannot switch branches. There are other places that ask unpack_trees() to n-way merge, detect issues and let it issue error message on its own, but I did this as a demonstration and replaced only one message. Yes I know about C99 structure initializers. I'd love to use them but we try to be nice to compilers without it. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-05-17 21:03:49 +02:00
/* ERROR_BIND_OVERLAP */
unpack-trees: allow Porcelain to give different error messages The plumbing output is sacred as it is an API. We _could_ change it if it is broken in such a way that it cannot convey necessary information fully, but we just do not _reword_ for the sake of rewording. If somebody does not like it, s/he is complaining too late. S/he should have been here in early May 2005 and make the language used by the API closer to what humans read. S/he wasn't here. Too bad, and it is too late. And people who complain should look at a bigger picture. Look at what was suggested by one of them and think for five seconds: $ git checkout mytopic -fatal: Entry 'frotz' not uptodate. Cannot merge. +fatal: Entry 'frotz' has local changes. Cannot merge. If you do not see something wrong with this output, your brain has already been rotten with use of git for too long a time. Nobody asked us to "merge" but why are we talking about "Cannot merge"? This patch introduces a mechanism to allow Porcelains to specify messages that are different from the ones that is given by the underlying plumbing implementation of read-tree, so that we can reword the message Porcelains give without disrupting the output from the plumbing. $ git-checkout pu error: You have local changes to 'Makefile'; cannot switch branches. There are other places that ask unpack_trees() to n-way merge, detect issues and let it issue error message on its own, but I did this as a demonstration and replaced only one message. Yes I know about C99 structure initializers. I'd love to use them but we try to be nice to compilers without it. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-05-17 21:03:49 +02:00
"Entry '%s' overlaps with '%s'. Cannot bind.",
/* ERROR_WOULD_LOSE_SUBMODULE */
"Submodule '%s' cannot checkout new HEAD.",
/* NB_UNPACK_TREES_ERROR_TYPES; just a meta value */
"",
/* WARNING_SPARSE_NOT_UPTODATE_FILE */
"Path '%s' not uptodate; will not remove from working tree.",
/* WARNING_SPARSE_UNMERGED_FILE */
"Path '%s' unmerged; will not remove from working tree.",
/* WARNING_SPARSE_ORPHANED_NOT_OVERWRITTEN */
"Path '%s' already present; will not overwrite with sparse update.",
unpack-trees: allow Porcelain to give different error messages The plumbing output is sacred as it is an API. We _could_ change it if it is broken in such a way that it cannot convey necessary information fully, but we just do not _reword_ for the sake of rewording. If somebody does not like it, s/he is complaining too late. S/he should have been here in early May 2005 and make the language used by the API closer to what humans read. S/he wasn't here. Too bad, and it is too late. And people who complain should look at a bigger picture. Look at what was suggested by one of them and think for five seconds: $ git checkout mytopic -fatal: Entry 'frotz' not uptodate. Cannot merge. +fatal: Entry 'frotz' has local changes. Cannot merge. If you do not see something wrong with this output, your brain has already been rotten with use of git for too long a time. Nobody asked us to "merge" but why are we talking about "Cannot merge"? This patch introduces a mechanism to allow Porcelains to specify messages that are different from the ones that is given by the underlying plumbing implementation of read-tree, so that we can reword the message Porcelains give without disrupting the output from the plumbing. $ git-checkout pu error: You have local changes to 'Makefile'; cannot switch branches. There are other places that ask unpack_trees() to n-way merge, detect issues and let it issue error message on its own, but I did this as a demonstration and replaced only one message. Yes I know about C99 structure initializers. I'd love to use them but we try to be nice to compilers without it. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-05-17 21:03:49 +02:00
};
#define ERRORMSG(o,type) \
( ((o) && (o)->msgs[(type)]) \
? ((o)->msgs[(type)]) \
: (unpack_plumbing_errors[(type)]) )
unpack-trees: allow Porcelain to give different error messages The plumbing output is sacred as it is an API. We _could_ change it if it is broken in such a way that it cannot convey necessary information fully, but we just do not _reword_ for the sake of rewording. If somebody does not like it, s/he is complaining too late. S/he should have been here in early May 2005 and make the language used by the API closer to what humans read. S/he wasn't here. Too bad, and it is too late. And people who complain should look at a bigger picture. Look at what was suggested by one of them and think for five seconds: $ git checkout mytopic -fatal: Entry 'frotz' not uptodate. Cannot merge. +fatal: Entry 'frotz' has local changes. Cannot merge. If you do not see something wrong with this output, your brain has already been rotten with use of git for too long a time. Nobody asked us to "merge" but why are we talking about "Cannot merge"? This patch introduces a mechanism to allow Porcelains to specify messages that are different from the ones that is given by the underlying plumbing implementation of read-tree, so that we can reword the message Porcelains give without disrupting the output from the plumbing. $ git-checkout pu error: You have local changes to 'Makefile'; cannot switch branches. There are other places that ask unpack_trees() to n-way merge, detect issues and let it issue error message on its own, but I did this as a demonstration and replaced only one message. Yes I know about C99 structure initializers. I'd love to use them but we try to be nice to compilers without it. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-05-17 21:03:49 +02:00
read-tree: add "--super-prefix" option, eliminate global The "--super-prefix" option to "git" was initially added in [1] for use with "ls-files"[2], and shortly thereafter "submodule--helper"[3] and "grep"[4]. It wasn't until [5] that "read-tree" made use of it. At the time [5] made sense, but since then we've made "ls-files" recurse in-process in [6], "grep" in [7], and finally "submodule--helper" in the preceding commits. Let's also remove it from "read-tree", which allows us to remove the option to "git" itself. We can do this because the only remaining user of it is the submodule API, which will now invoke "read-tree" with its new "--super-prefix" option. It will only do so when the "submodule_move_head()" function is called. That "submodule_move_head()" function was then only invoked by "read-tree" itself, but now rather than setting an environment variable to pass "--super-prefix" between cmd_read_tree() we: - Set a new "super_prefix" in "struct unpack_trees_options". The "super_prefixed()" function in "unpack-trees.c" added in [5] will now use this, rather than get_super_prefix() looking up the environment variable we set earlier in the same process. - Add the same field to the "struct checkout", which is only needed to ferry the "super_prefix" in the "struct unpack_trees_options" all the way down to the "entry.c" callers of "submodule_move_head()". Those calls which used the super prefix all originated in "cmd_read_tree()". The only other caller is the "unlink_entry()" caller in "builtin/checkout.c", which now passes a "NULL". 1. 74866d75793 (git: make super-prefix option, 2016-10-07) 2. e77aa336f11 (ls-files: optionally recurse into submodules, 2016-10-07) 3. 89c86265576 (submodule helper: support super prefix, 2016-12-08) 4. 0281e487fd9 (grep: optionally recurse into submodules, 2016-12-16) 5. 3d415425c7b (unpack-trees: support super-prefix option, 2017-01-17) 6. 188dce131fa (ls-files: use repository object, 2017-06-22) 7. f9ee2fcdfa0 (grep: recurse in-process using 'struct repository', 2017-08-02) Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2022-12-20 13:39:56 +01:00
static const char *super_prefixed(const char *path, const char *super_prefix)
unpack-trees: support super-prefix option In the future we want to support working tree operations within submodules, e.g. "git checkout --recurse-submodules", which will update the submodule to the commit as recorded in its superproject. In the submodule the unpack-tree operation is carried out as usual, but the reporting to the user needs to prefix any path with the superproject. The mechanism for this is the super-prefix. (see 74866d757, git: make super-prefix option) Add support for the super-prefix option for commands that unpack trees by wrapping any path output in unpacking trees in the newly introduced super_prefixed function. This new function prefixes any path with the super-prefix if there is one. Assuming the submodule case doesn't happen in the majority of the cases, we'd want to have a fast behavior for no super prefix, i.e. no reallocation/copying, but just returning path. Another aspect of introducing the `super_prefixed` function is to consider who owns the memory and if this is the right place where the path gets modified. As the super prefix ought to change the output behavior only and not the actual unpack tree part, it is fine to be that late in the line. As we get passed in 'const char *path', we cannot change the path itself, which means in case of a super prefix we have to copy over the path. We need two static buffers in that function as the error messages contain at most two paths. For testing purposes enable it in read-tree, which has no output of paths other than an unpack-trees.c. These are all converted in this patch. Signed-off-by: Stefan Beller <sbeller@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2017-01-18 02:05:20 +01:00
{
/*
* It is necessary and sufficient to have two static buffers
* here, as the return value of this function is fed to
* error() using the unpack_*_errors[] templates we see above.
*/
static struct strbuf buf[2] = {STRBUF_INIT, STRBUF_INIT};
static int super_prefix_len = -1;
static unsigned idx = ARRAY_SIZE(buf) - 1;
if (super_prefix_len < 0) {
if (!super_prefix) {
super_prefix_len = 0;
} else {
int i;
for (i = 0; i < ARRAY_SIZE(buf); i++)
strbuf_addstr(&buf[i], super_prefix);
super_prefix_len = buf[0].len;
}
}
if (!super_prefix_len)
return path;
if (++idx >= ARRAY_SIZE(buf))
idx = 0;
strbuf_setlen(&buf[idx], super_prefix_len);
strbuf_addstr(&buf[idx], path);
return buf[idx].buf;
}
void setup_unpack_trees_porcelain(struct unpack_trees_options *opts,
const char *cmd)
{
int i;
const char **msgs = opts->msgs;
const char *msg;
strvec_init(&opts->msgs_to_free);
if (!strcmp(cmd, "checkout"))
msg = advice_enabled(ADVICE_COMMIT_BEFORE_MERGE)
? _("Your local changes to the following files would be overwritten by checkout:\n%%s"
"Please commit your changes or stash them before you switch branches.")
: _("Your local changes to the following files would be overwritten by checkout:\n%%s");
else if (!strcmp(cmd, "merge"))
msg = advice_enabled(ADVICE_COMMIT_BEFORE_MERGE)
? _("Your local changes to the following files would be overwritten by merge:\n%%s"
"Please commit your changes or stash them before you merge.")
: _("Your local changes to the following files would be overwritten by merge:\n%%s");
else
msg = advice_enabled(ADVICE_COMMIT_BEFORE_MERGE)
? _("Your local changes to the following files would be overwritten by %s:\n%%s"
"Please commit your changes or stash them before you %s.")
: _("Your local changes to the following files would be overwritten by %s:\n%%s");
msgs[ERROR_WOULD_OVERWRITE] = msgs[ERROR_NOT_UPTODATE_FILE] =
strvec_pushf(&opts->msgs_to_free, msg, cmd, cmd);
msgs[ERROR_NOT_UPTODATE_DIR] =
_("Updating the following directories would lose untracked files in them:\n%s");
msgs[ERROR_CWD_IN_THE_WAY] =
_("Refusing to remove the current working directory:\n%s");
if (!strcmp(cmd, "checkout"))
msg = advice_enabled(ADVICE_COMMIT_BEFORE_MERGE)
? _("The following untracked working tree files would be removed by checkout:\n%%s"
"Please move or remove them before you switch branches.")
: _("The following untracked working tree files would be removed by checkout:\n%%s");
else if (!strcmp(cmd, "merge"))
msg = advice_enabled(ADVICE_COMMIT_BEFORE_MERGE)
? _("The following untracked working tree files would be removed by merge:\n%%s"
"Please move or remove them before you merge.")
: _("The following untracked working tree files would be removed by merge:\n%%s");
else
msg = advice_enabled(ADVICE_COMMIT_BEFORE_MERGE)
? _("The following untracked working tree files would be removed by %s:\n%%s"
"Please move or remove them before you %s.")
: _("The following untracked working tree files would be removed by %s:\n%%s");
msgs[ERROR_WOULD_LOSE_UNTRACKED_REMOVED] =
strvec_pushf(&opts->msgs_to_free, msg, cmd, cmd);
if (!strcmp(cmd, "checkout"))
msg = advice_enabled(ADVICE_COMMIT_BEFORE_MERGE)
? _("The following untracked working tree files would be overwritten by checkout:\n%%s"
"Please move or remove them before you switch branches.")
: _("The following untracked working tree files would be overwritten by checkout:\n%%s");
else if (!strcmp(cmd, "merge"))
msg = advice_enabled(ADVICE_COMMIT_BEFORE_MERGE)
? _("The following untracked working tree files would be overwritten by merge:\n%%s"
"Please move or remove them before you merge.")
: _("The following untracked working tree files would be overwritten by merge:\n%%s");
else
msg = advice_enabled(ADVICE_COMMIT_BEFORE_MERGE)
? _("The following untracked working tree files would be overwritten by %s:\n%%s"
"Please move or remove them before you %s.")
: _("The following untracked working tree files would be overwritten by %s:\n%%s");
msgs[ERROR_WOULD_LOSE_UNTRACKED_OVERWRITTEN] =
strvec_pushf(&opts->msgs_to_free, msg, cmd, cmd);
/*
* Special case: ERROR_BIND_OVERLAP refers to a pair of paths, we
* cannot easily display it as a list.
*/
msgs[ERROR_BIND_OVERLAP] = _("Entry '%s' overlaps with '%s'. Cannot bind.");
msgs[ERROR_WOULD_LOSE_SUBMODULE] =
_("Cannot update submodule:\n%s");
msgs[WARNING_SPARSE_NOT_UPTODATE_FILE] =
_("The following paths are not up to date and were left despite sparse patterns:\n%s");
msgs[WARNING_SPARSE_UNMERGED_FILE] =
_("The following paths are unmerged and were left despite sparse patterns:\n%s");
msgs[WARNING_SPARSE_ORPHANED_NOT_OVERWRITTEN] =
_("The following paths were already present and thus not updated despite sparse patterns:\n%s");
opts->show_all_errors = 1;
/* rejected paths may not have a static buffer */
for (i = 0; i < ARRAY_SIZE(opts->unpack_rejects); i++)
opts->unpack_rejects[i].strdup_strings = 1;
}
void clear_unpack_trees_porcelain(struct unpack_trees_options *opts)
{
strvec_clear(&opts->msgs_to_free);
memset(opts->msgs, 0, sizeof(opts->msgs));
}
static int do_add_entry(struct unpack_trees_options *o, struct cache_entry *ce,
unsigned int set, unsigned int clear)
{
clear |= CE_HASHED;
if (set & CE_REMOVE)
set |= CE_WT_REMOVE;
ce->ce_flags = (ce->ce_flags & ~clear) | set;
return add_index_entry(&o->result, ce,
ADD_CACHE_OK_TO_ADD | ADD_CACHE_OK_TO_REPLACE);
}
static void add_entry(struct unpack_trees_options *o,
const struct cache_entry *ce,
unsigned int set, unsigned int clear)
{
block alloc: allocate cache entries from mem_pool When reading large indexes from disk, a portion of the time is dominated in malloc() calls. This can be mitigated by allocating a large block of memory and manage it ourselves via memory pools. This change moves the cache entry allocation to be on top of memory pools. Design: The index_state struct will gain a notion of an associated memory_pool from which cache_entries will be allocated from. When reading in the index from disk, we have information on the number of entries and their size, which can guide us in deciding how large our initial memory allocation should be. When an index is discarded, the associated memory_pool will be discarded as well - so the lifetime of a cache_entry is tied to the lifetime of the index_state that it was allocated for. In the case of a Split Index, the following rules are followed. 1st, some terminology is defined: Terminology: - 'the_index': represents the logical view of the index - 'split_index': represents the "base" cache entries. Read from the split index file. 'the_index' can reference a single split_index, as well as cache_entries from the split_index. `the_index` will be discarded before the `split_index` is. This means that when we are allocating cache_entries in the presence of a split index, we need to allocate the entries from the `split_index`'s memory pool. This allows us to follow the pattern that `the_index` can reference cache_entries from the `split_index`, and that the cache_entries will not be freed while they are still being referenced. Managing transient cache_entry structs: Cache entries are usually allocated for an index, but this is not always the case. Cache entries are sometimes allocated because this is the type that the existing checkout_entry function works with. Because of this, the existing code needs to handle cache entries associated with an index / memory pool, and those that only exist transiently. Several strategies were contemplated around how to handle this: Chosen approach: An extra field was added to the cache_entry type to track whether the cache_entry was allocated from a memory pool or not. This is currently an int field, as there are no more available bits in the existing ce_flags bit field. If / when more bits are needed, this new field can be turned into a proper bit field. Alternatives: 1) Do not include any information about how the cache_entry was allocated. Calling code would be responsible for tracking whether the cache_entry needed to be freed or not. Pro: No extra memory overhead to track this state Con: Extra complexity in callers to handle this correctly. The extra complexity and burden to not regress this behavior in the future was more than we wanted. 2) cache_entry would gain knowledge about which mem_pool allocated it Pro: Could (potentially) do extra logic to know when a mem_pool no longer had references to any cache_entry Con: cache_entry would grow heavier by a pointer, instead of int We didn't see a tangible benefit to this approach 3) Do not add any extra information to a cache_entry, but when freeing a cache entry, check if the memory exists in a region managed by existing mem_pools. Pro: No extra memory overhead to track state Con: Extra computation is performed when freeing cache entries We decided tracking and iterating over known memory pool regions was less desirable than adding an extra field to track this stae. Signed-off-by: Jameson Miller <jamill@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-07-02 21:49:37 +02:00
do_add_entry(o, dup_cache_entry(ce, &o->result), set, clear);
}
/*
* add error messages on path <path>
* corresponding to the type <e> with the message <msg>
* indicating if it should be display in porcelain or not
*/
static int add_rejected_path(struct unpack_trees_options *o,
enum unpack_trees_error_types e,
const char *path)
{
if (o->quiet)
return -1;
if (!o->show_all_errors)
read-tree: add "--super-prefix" option, eliminate global The "--super-prefix" option to "git" was initially added in [1] for use with "ls-files"[2], and shortly thereafter "submodule--helper"[3] and "grep"[4]. It wasn't until [5] that "read-tree" made use of it. At the time [5] made sense, but since then we've made "ls-files" recurse in-process in [6], "grep" in [7], and finally "submodule--helper" in the preceding commits. Let's also remove it from "read-tree", which allows us to remove the option to "git" itself. We can do this because the only remaining user of it is the submodule API, which will now invoke "read-tree" with its new "--super-prefix" option. It will only do so when the "submodule_move_head()" function is called. That "submodule_move_head()" function was then only invoked by "read-tree" itself, but now rather than setting an environment variable to pass "--super-prefix" between cmd_read_tree() we: - Set a new "super_prefix" in "struct unpack_trees_options". The "super_prefixed()" function in "unpack-trees.c" added in [5] will now use this, rather than get_super_prefix() looking up the environment variable we set earlier in the same process. - Add the same field to the "struct checkout", which is only needed to ferry the "super_prefix" in the "struct unpack_trees_options" all the way down to the "entry.c" callers of "submodule_move_head()". Those calls which used the super prefix all originated in "cmd_read_tree()". The only other caller is the "unlink_entry()" caller in "builtin/checkout.c", which now passes a "NULL". 1. 74866d75793 (git: make super-prefix option, 2016-10-07) 2. e77aa336f11 (ls-files: optionally recurse into submodules, 2016-10-07) 3. 89c86265576 (submodule helper: support super prefix, 2016-12-08) 4. 0281e487fd9 (grep: optionally recurse into submodules, 2016-12-16) 5. 3d415425c7b (unpack-trees: support super-prefix option, 2017-01-17) 6. 188dce131fa (ls-files: use repository object, 2017-06-22) 7. f9ee2fcdfa0 (grep: recurse in-process using 'struct repository', 2017-08-02) Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2022-12-20 13:39:56 +01:00
return error(ERRORMSG(o, e), super_prefixed(path,
o->super_prefix));
/*
* Otherwise, insert in a list for future display by
* display_(error|warning)_msgs()
*/
string_list_append(&o->unpack_rejects[e], path);
return -1;
}
/*
* display all the error messages stored in a nice way
*/
static void display_error_msgs(struct unpack_trees_options *o)
{
int e;
unsigned error_displayed = 0;
for (e = 0; e < NB_UNPACK_TREES_ERROR_TYPES; e++) {
struct string_list *rejects = &o->unpack_rejects[e];
if (rejects->nr > 0) {
int i;
struct strbuf path = STRBUF_INIT;
error_displayed = 1;
for (i = 0; i < rejects->nr; i++)
strbuf_addf(&path, "\t%s\n", rejects->items[i].string);
read-tree: add "--super-prefix" option, eliminate global The "--super-prefix" option to "git" was initially added in [1] for use with "ls-files"[2], and shortly thereafter "submodule--helper"[3] and "grep"[4]. It wasn't until [5] that "read-tree" made use of it. At the time [5] made sense, but since then we've made "ls-files" recurse in-process in [6], "grep" in [7], and finally "submodule--helper" in the preceding commits. Let's also remove it from "read-tree", which allows us to remove the option to "git" itself. We can do this because the only remaining user of it is the submodule API, which will now invoke "read-tree" with its new "--super-prefix" option. It will only do so when the "submodule_move_head()" function is called. That "submodule_move_head()" function was then only invoked by "read-tree" itself, but now rather than setting an environment variable to pass "--super-prefix" between cmd_read_tree() we: - Set a new "super_prefix" in "struct unpack_trees_options". The "super_prefixed()" function in "unpack-trees.c" added in [5] will now use this, rather than get_super_prefix() looking up the environment variable we set earlier in the same process. - Add the same field to the "struct checkout", which is only needed to ferry the "super_prefix" in the "struct unpack_trees_options" all the way down to the "entry.c" callers of "submodule_move_head()". Those calls which used the super prefix all originated in "cmd_read_tree()". The only other caller is the "unlink_entry()" caller in "builtin/checkout.c", which now passes a "NULL". 1. 74866d75793 (git: make super-prefix option, 2016-10-07) 2. e77aa336f11 (ls-files: optionally recurse into submodules, 2016-10-07) 3. 89c86265576 (submodule helper: support super prefix, 2016-12-08) 4. 0281e487fd9 (grep: optionally recurse into submodules, 2016-12-16) 5. 3d415425c7b (unpack-trees: support super-prefix option, 2017-01-17) 6. 188dce131fa (ls-files: use repository object, 2017-06-22) 7. f9ee2fcdfa0 (grep: recurse in-process using 'struct repository', 2017-08-02) Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2022-12-20 13:39:56 +01:00
error(ERRORMSG(o, e), super_prefixed(path.buf,
o->super_prefix));
strbuf_release(&path);
}
string_list_clear(rejects, 0);
}
if (error_displayed)
fprintf(stderr, _("Aborting\n"));
}
/*
* display all the warning messages stored in a nice way
*/
static void display_warning_msgs(struct unpack_trees_options *o)
{
int e;
unsigned warning_displayed = 0;
for (e = NB_UNPACK_TREES_ERROR_TYPES + 1;
e < NB_UNPACK_TREES_WARNING_TYPES; e++) {
struct string_list *rejects = &o->unpack_rejects[e];
if (rejects->nr > 0) {
int i;
struct strbuf path = STRBUF_INIT;
warning_displayed = 1;
for (i = 0; i < rejects->nr; i++)
strbuf_addf(&path, "\t%s\n", rejects->items[i].string);
read-tree: add "--super-prefix" option, eliminate global The "--super-prefix" option to "git" was initially added in [1] for use with "ls-files"[2], and shortly thereafter "submodule--helper"[3] and "grep"[4]. It wasn't until [5] that "read-tree" made use of it. At the time [5] made sense, but since then we've made "ls-files" recurse in-process in [6], "grep" in [7], and finally "submodule--helper" in the preceding commits. Let's also remove it from "read-tree", which allows us to remove the option to "git" itself. We can do this because the only remaining user of it is the submodule API, which will now invoke "read-tree" with its new "--super-prefix" option. It will only do so when the "submodule_move_head()" function is called. That "submodule_move_head()" function was then only invoked by "read-tree" itself, but now rather than setting an environment variable to pass "--super-prefix" between cmd_read_tree() we: - Set a new "super_prefix" in "struct unpack_trees_options". The "super_prefixed()" function in "unpack-trees.c" added in [5] will now use this, rather than get_super_prefix() looking up the environment variable we set earlier in the same process. - Add the same field to the "struct checkout", which is only needed to ferry the "super_prefix" in the "struct unpack_trees_options" all the way down to the "entry.c" callers of "submodule_move_head()". Those calls which used the super prefix all originated in "cmd_read_tree()". The only other caller is the "unlink_entry()" caller in "builtin/checkout.c", which now passes a "NULL". 1. 74866d75793 (git: make super-prefix option, 2016-10-07) 2. e77aa336f11 (ls-files: optionally recurse into submodules, 2016-10-07) 3. 89c86265576 (submodule helper: support super prefix, 2016-12-08) 4. 0281e487fd9 (grep: optionally recurse into submodules, 2016-12-16) 5. 3d415425c7b (unpack-trees: support super-prefix option, 2017-01-17) 6. 188dce131fa (ls-files: use repository object, 2017-06-22) 7. f9ee2fcdfa0 (grep: recurse in-process using 'struct repository', 2017-08-02) Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2022-12-20 13:39:56 +01:00
warning(ERRORMSG(o, e), super_prefixed(path.buf,
o->super_prefix));
strbuf_release(&path);
}
string_list_clear(rejects, 0);
}
if (warning_displayed)
fprintf(stderr, _("After fixing the above paths, you may want to run `git sparse-checkout reapply`.\n"));
}
static int check_submodule_move_head(const struct cache_entry *ce,
const char *old_id,
const char *new_id,
struct unpack_trees_options *o)
{
unsigned flags = SUBMODULE_MOVE_HEAD_DRY_RUN;
const struct submodule *sub = submodule_from_ce(ce);
if (!sub)
return 0;
if (o->reset)
flags |= SUBMODULE_MOVE_HEAD_FORCE;
read-tree: add "--super-prefix" option, eliminate global The "--super-prefix" option to "git" was initially added in [1] for use with "ls-files"[2], and shortly thereafter "submodule--helper"[3] and "grep"[4]. It wasn't until [5] that "read-tree" made use of it. At the time [5] made sense, but since then we've made "ls-files" recurse in-process in [6], "grep" in [7], and finally "submodule--helper" in the preceding commits. Let's also remove it from "read-tree", which allows us to remove the option to "git" itself. We can do this because the only remaining user of it is the submodule API, which will now invoke "read-tree" with its new "--super-prefix" option. It will only do so when the "submodule_move_head()" function is called. That "submodule_move_head()" function was then only invoked by "read-tree" itself, but now rather than setting an environment variable to pass "--super-prefix" between cmd_read_tree() we: - Set a new "super_prefix" in "struct unpack_trees_options". The "super_prefixed()" function in "unpack-trees.c" added in [5] will now use this, rather than get_super_prefix() looking up the environment variable we set earlier in the same process. - Add the same field to the "struct checkout", which is only needed to ferry the "super_prefix" in the "struct unpack_trees_options" all the way down to the "entry.c" callers of "submodule_move_head()". Those calls which used the super prefix all originated in "cmd_read_tree()". The only other caller is the "unlink_entry()" caller in "builtin/checkout.c", which now passes a "NULL". 1. 74866d75793 (git: make super-prefix option, 2016-10-07) 2. e77aa336f11 (ls-files: optionally recurse into submodules, 2016-10-07) 3. 89c86265576 (submodule helper: support super prefix, 2016-12-08) 4. 0281e487fd9 (grep: optionally recurse into submodules, 2016-12-16) 5. 3d415425c7b (unpack-trees: support super-prefix option, 2017-01-17) 6. 188dce131fa (ls-files: use repository object, 2017-06-22) 7. f9ee2fcdfa0 (grep: recurse in-process using 'struct repository', 2017-08-02) Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2022-12-20 13:39:56 +01:00
if (submodule_move_head(ce->name, o->super_prefix, old_id, new_id,
flags))
return add_rejected_path(o, ERROR_WOULD_LOSE_SUBMODULE, ce->name);
return 0;
}
/*
* Perform the loading of the repository's gitmodules file. This function is
* used by 'check_update()' to perform loading of the gitmodules file in two
* different situations:
* (1) before removing entries from the working tree if the gitmodules file has
* been marked for removal. This situation is specified by 'state' == NULL.
* (2) before checking out entries to the working tree if the gitmodules file
* has been marked for update. This situation is specified by 'state' != NULL.
*/
static void load_gitmodules_file(struct index_state *index,
struct checkout *state)
{
int pos = index_name_pos(index, GITMODULES_FILE, strlen(GITMODULES_FILE));
if (pos >= 0) {
struct cache_entry *ce = index->cache[pos];
if (!state && ce->ce_flags & CE_WT_REMOVE) {
repo_read_gitmodules(the_repository, 0);
} else if (state && (ce->ce_flags & CE_UPDATE)) {
submodule_free(the_repository);
checkout_entry(ce, state, NULL, NULL);
repo_read_gitmodules(the_repository, 0);
}
}
}
static struct progress *get_progress(struct unpack_trees_options *o,
struct index_state *index)
{
unsigned cnt = 0, total = 0;
if (!o->update || !o->verbose_update)
return NULL;
for (; cnt < index->cache_nr; cnt++) {
const struct cache_entry *ce = index->cache[cnt];
if (ce->ce_flags & (CE_UPDATE | CE_WT_REMOVE))
total++;
}
return start_delayed_progress(_("Updating files"), total);
}
static void setup_collided_checkout_detection(struct checkout *state,
struct index_state *index)
{
int i;
state->clone = 1;
for (i = 0; i < index->cache_nr; i++)
index->cache[i]->ce_flags &= ~CE_MATCHED;
}
static void report_collided_checkout(struct index_state *index)
{
struct string_list list = STRING_LIST_INIT_NODUP;
int i;
for (i = 0; i < index->cache_nr; i++) {
struct cache_entry *ce = index->cache[i];
if (!(ce->ce_flags & CE_MATCHED))
continue;
string_list_append(&list, ce->name);
ce->ce_flags &= ~CE_MATCHED;
}
list.cmp = fspathcmp;
string_list_sort(&list);
if (list.nr) {
warning(_("the following paths have collided (e.g. case-sensitive paths\n"
"on a case-insensitive filesystem) and only one from the same\n"
"colliding group is in the working tree:\n"));
for (i = 0; i < list.nr; i++)
fprintf(stderr, " '%s'\n", list.items[i].string);
}
string_list_clear(&list, 0);
}
static int must_checkout(const struct cache_entry *ce)
{
return ce->ce_flags & CE_UPDATE;
}
static int check_updates(struct unpack_trees_options *o,
struct index_state *index)
{
unsigned cnt = 0;
int errs = 0;
struct progress *progress;
struct checkout state = CHECKOUT_INIT;
parallel-checkout: add configuration options Make parallel checkout configurable by introducing two new settings: checkout.workers and checkout.thresholdForParallelism. The first defines the number of workers (where one means sequential checkout), and the second defines the minimum number of entries to attempt parallel checkout. To decide the default value for checkout.workers, the parallel version was benchmarked during three operations in the linux repo, with cold cache: cloning v5.8, checking out v5.8 from v2.6.15 (checkout I) and checking out v5.8 from v5.7 (checkout II). The four tables below show the mean run times and standard deviations for 5 runs in: a local file system on SSD, a local file system on HDD, a Linux NFS server, and Amazon EFS (all on Linux). Each parallel checkout test was executed with the number of workers that brings the best overall results in that environment. Local SSD: Sequential 10 workers Speedup Clone 8.805 s ± 0.043 s 3.564 s ± 0.041 s 2.47 ± 0.03 Checkout I 9.678 s ± 0.057 s 4.486 s ± 0.050 s 2.16 ± 0.03 Checkout II 5.034 s ± 0.072 s 3.021 s ± 0.038 s 1.67 ± 0.03 Local HDD: Sequential 10 workers Speedup Clone 32.288 s ± 0.580 s 30.724 s ± 0.522 s 1.05 ± 0.03 Checkout I 54.172 s ± 7.119 s 54.429 s ± 6.738 s 1.00 ± 0.18 Checkout II 40.465 s ± 2.402 s 38.682 s ± 1.365 s 1.05 ± 0.07 Linux NFS server (v4.1, on EBS, single availability zone): Sequential 32 workers Speedup Clone 240.368 s ± 6.347 s 57.349 s ± 0.870 s 4.19 ± 0.13 Checkout I 242.862 s ± 2.215 s 58.700 s ± 0.904 s 4.14 ± 0.07 Checkout II 65.751 s ± 1.577 s 23.820 s ± 0.407 s 2.76 ± 0.08 EFS (v4.1, replicated over multiple availability zones): Sequential 32 workers Speedup Clone 922.321 s ± 2.274 s 210.453 s ± 3.412 s 4.38 ± 0.07 Checkout I 1011.300 s ± 7.346 s 297.828 s ± 0.964 s 3.40 ± 0.03 Checkout II 294.104 s ± 1.836 s 126.017 s ± 1.190 s 2.33 ± 0.03 The above benchmarks show that parallel checkout is most effective on repositories located on an SSD or over a distributed file system. For local file systems on spinning disks, and/or older machines, the parallelism does not always bring a good performance. For this reason, the default value for checkout.workers is one, a.k.a. sequential checkout. To decide the default value for checkout.thresholdForParallelism, another benchmark was executed in the "Local SSD" setup, where parallel checkout showed to be beneficial. This time, we compared the runtime of a `git checkout -f`, with and without parallelism, after randomly removing an increasing number of files from the Linux working tree. The "sequential fallback" column below corresponds to the executions where checkout.workers was 10 but checkout.thresholdForParallelism was equal to the number of to-be-updated files plus one (so that we end up writing sequentially). Each test case was sampled 15 times, and each sample had a randomly different set of files removed. Here are the results: sequential fallback 10 workers speedup 10 files 772.3 ms ± 12.6 ms 769.0 ms ± 13.6 ms 1.00 ± 0.02 20 files 780.5 ms ± 15.8 ms 775.2 ms ± 9.2 ms 1.01 ± 0.02 50 files 806.2 ms ± 13.8 ms 767.4 ms ± 8.5 ms 1.05 ± 0.02 100 files 833.7 ms ± 21.4 ms 750.5 ms ± 16.8 ms 1.11 ± 0.04 200 files 897.6 ms ± 30.9 ms 730.5 ms ± 14.7 ms 1.23 ± 0.05 500 files 1035.4 ms ± 48.0 ms 677.1 ms ± 22.3 ms 1.53 ± 0.09 1000 files 1244.6 ms ± 35.6 ms 654.0 ms ± 38.3 ms 1.90 ± 0.12 2000 files 1488.8 ms ± 53.4 ms 658.8 ms ± 23.8 ms 2.26 ± 0.12 From the above numbers, 100 files seems to be a reasonable default value for the threshold setting. Note: Up to 1000 files, we observe a drop in the execution time of the parallel code with an increase in the number of files. This is a rather odd behavior, but it was observed in multiple repetitions. Above 1000 files, the execution time increases according to the number of files, as one would expect. About the test environments: Local SSD tests were executed on an i7-7700HQ (4 cores with hyper-threading) running Manjaro Linux. Local HDD tests were executed on an Intel(R) Xeon(R) E3-1230 (also 4 cores with hyper-threading), HDD Seagate Barracuda 7200.14 SATA 3.1, running Debian. NFS and EFS tests were executed on an Amazon EC2 c5n.xlarge instance, with 4 vCPUs. The Linux NFS server was running on a m6g.large instance with 2 vCPUSs and a 1 TB EBS GP2 volume. Before each timing, the linux repository was removed (or checked out back to its previous state), and `sync && sysctl vm.drop_caches=3` was executed. Co-authored-by: Jeff Hostetler <jeffhost@microsoft.com> Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2021-04-19 02:14:55 +02:00
int i, pc_workers, pc_threshold;
trace_performance_enter();
read-tree: add "--super-prefix" option, eliminate global The "--super-prefix" option to "git" was initially added in [1] for use with "ls-files"[2], and shortly thereafter "submodule--helper"[3] and "grep"[4]. It wasn't until [5] that "read-tree" made use of it. At the time [5] made sense, but since then we've made "ls-files" recurse in-process in [6], "grep" in [7], and finally "submodule--helper" in the preceding commits. Let's also remove it from "read-tree", which allows us to remove the option to "git" itself. We can do this because the only remaining user of it is the submodule API, which will now invoke "read-tree" with its new "--super-prefix" option. It will only do so when the "submodule_move_head()" function is called. That "submodule_move_head()" function was then only invoked by "read-tree" itself, but now rather than setting an environment variable to pass "--super-prefix" between cmd_read_tree() we: - Set a new "super_prefix" in "struct unpack_trees_options". The "super_prefixed()" function in "unpack-trees.c" added in [5] will now use this, rather than get_super_prefix() looking up the environment variable we set earlier in the same process. - Add the same field to the "struct checkout", which is only needed to ferry the "super_prefix" in the "struct unpack_trees_options" all the way down to the "entry.c" callers of "submodule_move_head()". Those calls which used the super prefix all originated in "cmd_read_tree()". The only other caller is the "unlink_entry()" caller in "builtin/checkout.c", which now passes a "NULL". 1. 74866d75793 (git: make super-prefix option, 2016-10-07) 2. e77aa336f11 (ls-files: optionally recurse into submodules, 2016-10-07) 3. 89c86265576 (submodule helper: support super prefix, 2016-12-08) 4. 0281e487fd9 (grep: optionally recurse into submodules, 2016-12-16) 5. 3d415425c7b (unpack-trees: support super-prefix option, 2017-01-17) 6. 188dce131fa (ls-files: use repository object, 2017-06-22) 7. f9ee2fcdfa0 (grep: recurse in-process using 'struct repository', 2017-08-02) Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2022-12-20 13:39:56 +01:00
state.super_prefix = o->super_prefix;
state.force = 1;
state.quiet = 1;
state.refresh_cache = 1;
state.istate = index;
clone_checkout_metadata(&state.meta, &o->meta, NULL);
unpack-trees: exit check_updates() early if updates are not wanted check_updates() has a lot of code that repeatedly checks whether o->update or o->dry_run are set. (Note that o->dry_run is a near-synonym for !o->update, but not quite as per commit 2c9078d05bf2 ("unpack-trees: add the dry_run flag to unpack_trees_options", 2011-05-25).) In fact, this function almost turns into a no-op whenever the condition !o->update || o->dry_run is met. Simplify the code by checking this condition at the beginning of the function, and when it is true, do the few things that are relevant and return early. There are a few things that make the conversion not quite obvious: * The fact that check_updates() does not actually turn into a no-op when updates are not wanted may be slightly surprising. However, commit 33ecf7eb61 (Discard "deleted" cache entries after using them to update the working tree, 2008-02-07) put the discarding of unused cache entries in check_updates() so we still need to keep the call to remove_marked_cache_entries(). It's possible this call belongs in another function, but it is certainly needed as tests will fail if it is removed. * The original called remove_scheduled_dirs() unconditionally. Technically, commit 7847892716 (unlink_entry(): introduce schedule_dir_for_removal(), 2009-02-09) should have made that call conditional, but it didn't matter in practice because remove_scheduled_dirs() becomes a no-op when all the calls to unlink_entry() are skipped. As such, we do not need to call it. * When (o->dry_run && o->update), the original would have two calls to git_attr_set_direction() surrounding a bunch of skipped updates. These two calls to git_attr_set_direction() cancel each other out and thus can be omitted when o->dry_run is true just as they already are when !o->update. * The code would previously call setup_collided_checkout_detection() and report_collided_checkout() even when o->dry_run. However, this was just an expensive no-op because setup_collided_checkout_detection() merely cleared the CE_MATCHED flag for each cache entry, and report_collided_checkout() reported which ones had it set. Since a dry-run would skip all the checkout_entry() calls, CE_MATCHED would never get set and thus no collisions would be reported. Since we can't detect the collisions anyway without doing updates, skipping the collisions detection setup and reporting is an optimization. * The code previously would call get_progress() and display_progress() even when (!o->update || o->dry_run). This served to show how long it took to skip all the updates, which is somewhat useless. Since we are skipping the updates, we can skip showing how long it takes to skip them. Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-01-07 07:57:57 +01:00
if (!o->update || o->dry_run) {
remove_marked_cache_entries(index, 0);
trace_performance_leave("check_updates");
return 0;
}
if (o->clone)
setup_collided_checkout_detection(&state, index);
progress = get_progress(o, index);
/* Start with clean cache to avoid using any possibly outdated info. */
invalidate_lstat_cache();
unpack-trees: exit check_updates() early if updates are not wanted check_updates() has a lot of code that repeatedly checks whether o->update or o->dry_run are set. (Note that o->dry_run is a near-synonym for !o->update, but not quite as per commit 2c9078d05bf2 ("unpack-trees: add the dry_run flag to unpack_trees_options", 2011-05-25).) In fact, this function almost turns into a no-op whenever the condition !o->update || o->dry_run is met. Simplify the code by checking this condition at the beginning of the function, and when it is true, do the few things that are relevant and return early. There are a few things that make the conversion not quite obvious: * The fact that check_updates() does not actually turn into a no-op when updates are not wanted may be slightly surprising. However, commit 33ecf7eb61 (Discard "deleted" cache entries after using them to update the working tree, 2008-02-07) put the discarding of unused cache entries in check_updates() so we still need to keep the call to remove_marked_cache_entries(). It's possible this call belongs in another function, but it is certainly needed as tests will fail if it is removed. * The original called remove_scheduled_dirs() unconditionally. Technically, commit 7847892716 (unlink_entry(): introduce schedule_dir_for_removal(), 2009-02-09) should have made that call conditional, but it didn't matter in practice because remove_scheduled_dirs() becomes a no-op when all the calls to unlink_entry() are skipped. As such, we do not need to call it. * When (o->dry_run && o->update), the original would have two calls to git_attr_set_direction() surrounding a bunch of skipped updates. These two calls to git_attr_set_direction() cancel each other out and thus can be omitted when o->dry_run is true just as they already are when !o->update. * The code would previously call setup_collided_checkout_detection() and report_collided_checkout() even when o->dry_run. However, this was just an expensive no-op because setup_collided_checkout_detection() merely cleared the CE_MATCHED flag for each cache entry, and report_collided_checkout() reported which ones had it set. Since a dry-run would skip all the checkout_entry() calls, CE_MATCHED would never get set and thus no collisions would be reported. Since we can't detect the collisions anyway without doing updates, skipping the collisions detection setup and reporting is an optimization. * The code previously would call get_progress() and display_progress() even when (!o->update || o->dry_run). This served to show how long it took to skip all the updates, which is somewhat useless. Since we are skipping the updates, we can skip showing how long it takes to skip them. Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-01-07 07:57:57 +01:00
git_attr_set_direction(GIT_ATTR_CHECKOUT);
unpack-trees: exit check_updates() early if updates are not wanted check_updates() has a lot of code that repeatedly checks whether o->update or o->dry_run are set. (Note that o->dry_run is a near-synonym for !o->update, but not quite as per commit 2c9078d05bf2 ("unpack-trees: add the dry_run flag to unpack_trees_options", 2011-05-25).) In fact, this function almost turns into a no-op whenever the condition !o->update || o->dry_run is met. Simplify the code by checking this condition at the beginning of the function, and when it is true, do the few things that are relevant and return early. There are a few things that make the conversion not quite obvious: * The fact that check_updates() does not actually turn into a no-op when updates are not wanted may be slightly surprising. However, commit 33ecf7eb61 (Discard "deleted" cache entries after using them to update the working tree, 2008-02-07) put the discarding of unused cache entries in check_updates() so we still need to keep the call to remove_marked_cache_entries(). It's possible this call belongs in another function, but it is certainly needed as tests will fail if it is removed. * The original called remove_scheduled_dirs() unconditionally. Technically, commit 7847892716 (unlink_entry(): introduce schedule_dir_for_removal(), 2009-02-09) should have made that call conditional, but it didn't matter in practice because remove_scheduled_dirs() becomes a no-op when all the calls to unlink_entry() are skipped. As such, we do not need to call it. * When (o->dry_run && o->update), the original would have two calls to git_attr_set_direction() surrounding a bunch of skipped updates. These two calls to git_attr_set_direction() cancel each other out and thus can be omitted when o->dry_run is true just as they already are when !o->update. * The code would previously call setup_collided_checkout_detection() and report_collided_checkout() even when o->dry_run. However, this was just an expensive no-op because setup_collided_checkout_detection() merely cleared the CE_MATCHED flag for each cache entry, and report_collided_checkout() reported which ones had it set. Since a dry-run would skip all the checkout_entry() calls, CE_MATCHED would never get set and thus no collisions would be reported. Since we can't detect the collisions anyway without doing updates, skipping the collisions detection setup and reporting is an optimization. * The code previously would call get_progress() and display_progress() even when (!o->update || o->dry_run). This served to show how long it took to skip all the updates, which is somewhat useless. Since we are skipping the updates, we can skip showing how long it takes to skip them. Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-01-07 07:57:57 +01:00
if (should_update_submodules())
load_gitmodules_file(index, NULL);
for (i = 0; i < index->cache_nr; i++) {
Convert "struct cache_entry *" to "const ..." wherever possible I attempted to make index_state->cache[] a "const struct cache_entry **" to find out how existing entries in index are modified and where. The question I have is what do we do if we really need to keep track of on-disk changes in the index. The result is - diff-lib.c: setting CE_UPTODATE - name-hash.c: setting CE_HASHED - preload-index.c, read-cache.c, unpack-trees.c and builtin/update-index: obvious - entry.c: write_entry() may refresh the checked out entry via fill_stat_cache_info(). This causes "non-const struct cache_entry *" in builtin/apply.c, builtin/checkout-index.c and builtin/checkout.c - builtin/ls-files.c: --with-tree changes stagemask and may set CE_UPDATE Of these, write_entry() and its call sites are probably most interesting because it modifies on-disk info. But this is stat info and can be retrieved via refresh, at least for porcelain commands. Other just uses ce_flags for local purposes. So, keeping track of "dirty" entries is just a matter of setting a flag in index modification functions exposed by read-cache.c. Except unpack-trees, the rest of the code base does not do anything funny behind read-cache's back. The actual patch is less valueable than the summary above. But if anyone wants to re-identify the above sites. Applying this patch, then this: diff --git a/cache.h b/cache.h index 430d021..1692891 100644 --- a/cache.h +++ b/cache.h @@ -267,7 +267,7 @@ static inline unsigned int canon_mode(unsigned int mode) #define cache_entry_size(len) (offsetof(struct cache_entry,name) + (len) + 1) struct index_state { - struct cache_entry **cache; + const struct cache_entry **cache; unsigned int version; unsigned int cache_nr, cache_alloc, cache_changed; struct string_list *resolve_undo; will help quickly identify them without bogus warnings. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-07-09 17:29:00 +02:00
const struct cache_entry *ce = index->cache[i];
if (ce->ce_flags & CE_WT_REMOVE) {
display_progress(progress, ++cnt);
read-tree: add "--super-prefix" option, eliminate global The "--super-prefix" option to "git" was initially added in [1] for use with "ls-files"[2], and shortly thereafter "submodule--helper"[3] and "grep"[4]. It wasn't until [5] that "read-tree" made use of it. At the time [5] made sense, but since then we've made "ls-files" recurse in-process in [6], "grep" in [7], and finally "submodule--helper" in the preceding commits. Let's also remove it from "read-tree", which allows us to remove the option to "git" itself. We can do this because the only remaining user of it is the submodule API, which will now invoke "read-tree" with its new "--super-prefix" option. It will only do so when the "submodule_move_head()" function is called. That "submodule_move_head()" function was then only invoked by "read-tree" itself, but now rather than setting an environment variable to pass "--super-prefix" between cmd_read_tree() we: - Set a new "super_prefix" in "struct unpack_trees_options". The "super_prefixed()" function in "unpack-trees.c" added in [5] will now use this, rather than get_super_prefix() looking up the environment variable we set earlier in the same process. - Add the same field to the "struct checkout", which is only needed to ferry the "super_prefix" in the "struct unpack_trees_options" all the way down to the "entry.c" callers of "submodule_move_head()". Those calls which used the super prefix all originated in "cmd_read_tree()". The only other caller is the "unlink_entry()" caller in "builtin/checkout.c", which now passes a "NULL". 1. 74866d75793 (git: make super-prefix option, 2016-10-07) 2. e77aa336f11 (ls-files: optionally recurse into submodules, 2016-10-07) 3. 89c86265576 (submodule helper: support super prefix, 2016-12-08) 4. 0281e487fd9 (grep: optionally recurse into submodules, 2016-12-16) 5. 3d415425c7b (unpack-trees: support super-prefix option, 2017-01-17) 6. 188dce131fa (ls-files: use repository object, 2017-06-22) 7. f9ee2fcdfa0 (grep: recurse in-process using 'struct repository', 2017-08-02) Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2022-12-20 13:39:56 +01:00
unlink_entry(ce, o->super_prefix);
}
}
unpack-trees: exit check_updates() early if updates are not wanted check_updates() has a lot of code that repeatedly checks whether o->update or o->dry_run are set. (Note that o->dry_run is a near-synonym for !o->update, but not quite as per commit 2c9078d05bf2 ("unpack-trees: add the dry_run flag to unpack_trees_options", 2011-05-25).) In fact, this function almost turns into a no-op whenever the condition !o->update || o->dry_run is met. Simplify the code by checking this condition at the beginning of the function, and when it is true, do the few things that are relevant and return early. There are a few things that make the conversion not quite obvious: * The fact that check_updates() does not actually turn into a no-op when updates are not wanted may be slightly surprising. However, commit 33ecf7eb61 (Discard "deleted" cache entries after using them to update the working tree, 2008-02-07) put the discarding of unused cache entries in check_updates() so we still need to keep the call to remove_marked_cache_entries(). It's possible this call belongs in another function, but it is certainly needed as tests will fail if it is removed. * The original called remove_scheduled_dirs() unconditionally. Technically, commit 7847892716 (unlink_entry(): introduce schedule_dir_for_removal(), 2009-02-09) should have made that call conditional, but it didn't matter in practice because remove_scheduled_dirs() becomes a no-op when all the calls to unlink_entry() are skipped. As such, we do not need to call it. * When (o->dry_run && o->update), the original would have two calls to git_attr_set_direction() surrounding a bunch of skipped updates. These two calls to git_attr_set_direction() cancel each other out and thus can be omitted when o->dry_run is true just as they already are when !o->update. * The code would previously call setup_collided_checkout_detection() and report_collided_checkout() even when o->dry_run. However, this was just an expensive no-op because setup_collided_checkout_detection() merely cleared the CE_MATCHED flag for each cache entry, and report_collided_checkout() reported which ones had it set. Since a dry-run would skip all the checkout_entry() calls, CE_MATCHED would never get set and thus no collisions would be reported. Since we can't detect the collisions anyway without doing updates, skipping the collisions detection setup and reporting is an optimization. * The code previously would call get_progress() and display_progress() even when (!o->update || o->dry_run). This served to show how long it took to skip all the updates, which is somewhat useless. Since we are skipping the updates, we can skip showing how long it takes to skip them. Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-01-07 07:57:57 +01:00
remove_marked_cache_entries(index, 0);
remove_scheduled_dirs();
unpack-trees: exit check_updates() early if updates are not wanted check_updates() has a lot of code that repeatedly checks whether o->update or o->dry_run are set. (Note that o->dry_run is a near-synonym for !o->update, but not quite as per commit 2c9078d05bf2 ("unpack-trees: add the dry_run flag to unpack_trees_options", 2011-05-25).) In fact, this function almost turns into a no-op whenever the condition !o->update || o->dry_run is met. Simplify the code by checking this condition at the beginning of the function, and when it is true, do the few things that are relevant and return early. There are a few things that make the conversion not quite obvious: * The fact that check_updates() does not actually turn into a no-op when updates are not wanted may be slightly surprising. However, commit 33ecf7eb61 (Discard "deleted" cache entries after using them to update the working tree, 2008-02-07) put the discarding of unused cache entries in check_updates() so we still need to keep the call to remove_marked_cache_entries(). It's possible this call belongs in another function, but it is certainly needed as tests will fail if it is removed. * The original called remove_scheduled_dirs() unconditionally. Technically, commit 7847892716 (unlink_entry(): introduce schedule_dir_for_removal(), 2009-02-09) should have made that call conditional, but it didn't matter in practice because remove_scheduled_dirs() becomes a no-op when all the calls to unlink_entry() are skipped. As such, we do not need to call it. * When (o->dry_run && o->update), the original would have two calls to git_attr_set_direction() surrounding a bunch of skipped updates. These two calls to git_attr_set_direction() cancel each other out and thus can be omitted when o->dry_run is true just as they already are when !o->update. * The code would previously call setup_collided_checkout_detection() and report_collided_checkout() even when o->dry_run. However, this was just an expensive no-op because setup_collided_checkout_detection() merely cleared the CE_MATCHED flag for each cache entry, and report_collided_checkout() reported which ones had it set. Since a dry-run would skip all the checkout_entry() calls, CE_MATCHED would never get set and thus no collisions would be reported. Since we can't detect the collisions anyway without doing updates, skipping the collisions detection setup and reporting is an optimization. * The code previously would call get_progress() and display_progress() even when (!o->update || o->dry_run). This served to show how long it took to skip all the updates, which is somewhat useless. Since we are skipping the updates, we can skip showing how long it takes to skip them. Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-01-07 07:57:57 +01:00
if (should_update_submodules())
load_gitmodules_file(index, &state);
if (has_promisor_remote())
/*
* Prefetch the objects that are to be checked out in the loop
* below.
*/
prefetch_cache_entries(index, must_checkout);
unpack-trees: add basic support for parallel checkout This new interface allows us to enqueue some of the entries being checked out to later uncompress them, apply in-process filters, and write out the files in parallel. For now, the parallel checkout machinery is enabled by default and there is no user configuration, but run_parallel_checkout() just writes the queued entries in sequence (without spawning additional workers). The next patch will actually implement the parallelism and, later, we will make it configurable. Note that, to avoid potential data races, not all entries are eligible for parallel checkout. Also, paths that collide on disk (e.g. case-sensitive paths in case-insensitive file systems), are detected by the parallel checkout code and skipped, so that they can be safely sequentially handled later. The collision detection works like the following: - If the collision was at basename (e.g. 'a/b' and 'a/B'), the framework detects it by looking for EEXIST and EISDIR errors after an open(O_CREAT | O_EXCL) failure. - If the collision was at dirname (e.g. 'a/b' and 'A'), it is detected at the has_dirs_only_path() check, which is done for the leading path of each item in the parallel checkout queue. Both verifications rely on the fact that, before enqueueing an entry for parallel checkout, checkout_entry() makes sure that there is no file at the entry's path and that its leading components are all real directories. So, any later change in these conditions indicates that there was a collision (either between two parallel-eligible entries or between an eligible and an ineligible one). After all parallel-eligible entries have been processed, the collided (and thus, skipped) entries are sequentially fed to checkout_entry() again. This is similar to the way the current code deals with collisions, overwriting the previously checked out entries with the subsequent ones. The only difference is that, since we no longer create the files in the same order that they appear on index, we are not able to determine which of the colliding entries will survive on disk (for the classic code, it is always the last entry). Co-authored-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Co-authored-by: Jeff Hostetler <jeffhost@microsoft.com> Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2021-04-19 02:14:53 +02:00
parallel-checkout: add configuration options Make parallel checkout configurable by introducing two new settings: checkout.workers and checkout.thresholdForParallelism. The first defines the number of workers (where one means sequential checkout), and the second defines the minimum number of entries to attempt parallel checkout. To decide the default value for checkout.workers, the parallel version was benchmarked during three operations in the linux repo, with cold cache: cloning v5.8, checking out v5.8 from v2.6.15 (checkout I) and checking out v5.8 from v5.7 (checkout II). The four tables below show the mean run times and standard deviations for 5 runs in: a local file system on SSD, a local file system on HDD, a Linux NFS server, and Amazon EFS (all on Linux). Each parallel checkout test was executed with the number of workers that brings the best overall results in that environment. Local SSD: Sequential 10 workers Speedup Clone 8.805 s ± 0.043 s 3.564 s ± 0.041 s 2.47 ± 0.03 Checkout I 9.678 s ± 0.057 s 4.486 s ± 0.050 s 2.16 ± 0.03 Checkout II 5.034 s ± 0.072 s 3.021 s ± 0.038 s 1.67 ± 0.03 Local HDD: Sequential 10 workers Speedup Clone 32.288 s ± 0.580 s 30.724 s ± 0.522 s 1.05 ± 0.03 Checkout I 54.172 s ± 7.119 s 54.429 s ± 6.738 s 1.00 ± 0.18 Checkout II 40.465 s ± 2.402 s 38.682 s ± 1.365 s 1.05 ± 0.07 Linux NFS server (v4.1, on EBS, single availability zone): Sequential 32 workers Speedup Clone 240.368 s ± 6.347 s 57.349 s ± 0.870 s 4.19 ± 0.13 Checkout I 242.862 s ± 2.215 s 58.700 s ± 0.904 s 4.14 ± 0.07 Checkout II 65.751 s ± 1.577 s 23.820 s ± 0.407 s 2.76 ± 0.08 EFS (v4.1, replicated over multiple availability zones): Sequential 32 workers Speedup Clone 922.321 s ± 2.274 s 210.453 s ± 3.412 s 4.38 ± 0.07 Checkout I 1011.300 s ± 7.346 s 297.828 s ± 0.964 s 3.40 ± 0.03 Checkout II 294.104 s ± 1.836 s 126.017 s ± 1.190 s 2.33 ± 0.03 The above benchmarks show that parallel checkout is most effective on repositories located on an SSD or over a distributed file system. For local file systems on spinning disks, and/or older machines, the parallelism does not always bring a good performance. For this reason, the default value for checkout.workers is one, a.k.a. sequential checkout. To decide the default value for checkout.thresholdForParallelism, another benchmark was executed in the "Local SSD" setup, where parallel checkout showed to be beneficial. This time, we compared the runtime of a `git checkout -f`, with and without parallelism, after randomly removing an increasing number of files from the Linux working tree. The "sequential fallback" column below corresponds to the executions where checkout.workers was 10 but checkout.thresholdForParallelism was equal to the number of to-be-updated files plus one (so that we end up writing sequentially). Each test case was sampled 15 times, and each sample had a randomly different set of files removed. Here are the results: sequential fallback 10 workers speedup 10 files 772.3 ms ± 12.6 ms 769.0 ms ± 13.6 ms 1.00 ± 0.02 20 files 780.5 ms ± 15.8 ms 775.2 ms ± 9.2 ms 1.01 ± 0.02 50 files 806.2 ms ± 13.8 ms 767.4 ms ± 8.5 ms 1.05 ± 0.02 100 files 833.7 ms ± 21.4 ms 750.5 ms ± 16.8 ms 1.11 ± 0.04 200 files 897.6 ms ± 30.9 ms 730.5 ms ± 14.7 ms 1.23 ± 0.05 500 files 1035.4 ms ± 48.0 ms 677.1 ms ± 22.3 ms 1.53 ± 0.09 1000 files 1244.6 ms ± 35.6 ms 654.0 ms ± 38.3 ms 1.90 ± 0.12 2000 files 1488.8 ms ± 53.4 ms 658.8 ms ± 23.8 ms 2.26 ± 0.12 From the above numbers, 100 files seems to be a reasonable default value for the threshold setting. Note: Up to 1000 files, we observe a drop in the execution time of the parallel code with an increase in the number of files. This is a rather odd behavior, but it was observed in multiple repetitions. Above 1000 files, the execution time increases according to the number of files, as one would expect. About the test environments: Local SSD tests were executed on an i7-7700HQ (4 cores with hyper-threading) running Manjaro Linux. Local HDD tests were executed on an Intel(R) Xeon(R) E3-1230 (also 4 cores with hyper-threading), HDD Seagate Barracuda 7200.14 SATA 3.1, running Debian. NFS and EFS tests were executed on an Amazon EC2 c5n.xlarge instance, with 4 vCPUs. The Linux NFS server was running on a m6g.large instance with 2 vCPUSs and a 1 TB EBS GP2 volume. Before each timing, the linux repository was removed (or checked out back to its previous state), and `sync && sysctl vm.drop_caches=3` was executed. Co-authored-by: Jeff Hostetler <jeffhost@microsoft.com> Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2021-04-19 02:14:55 +02:00
get_parallel_checkout_configs(&pc_workers, &pc_threshold);
unpack-trees: add basic support for parallel checkout This new interface allows us to enqueue some of the entries being checked out to later uncompress them, apply in-process filters, and write out the files in parallel. For now, the parallel checkout machinery is enabled by default and there is no user configuration, but run_parallel_checkout() just writes the queued entries in sequence (without spawning additional workers). The next patch will actually implement the parallelism and, later, we will make it configurable. Note that, to avoid potential data races, not all entries are eligible for parallel checkout. Also, paths that collide on disk (e.g. case-sensitive paths in case-insensitive file systems), are detected by the parallel checkout code and skipped, so that they can be safely sequentially handled later. The collision detection works like the following: - If the collision was at basename (e.g. 'a/b' and 'a/B'), the framework detects it by looking for EEXIST and EISDIR errors after an open(O_CREAT | O_EXCL) failure. - If the collision was at dirname (e.g. 'a/b' and 'A'), it is detected at the has_dirs_only_path() check, which is done for the leading path of each item in the parallel checkout queue. Both verifications rely on the fact that, before enqueueing an entry for parallel checkout, checkout_entry() makes sure that there is no file at the entry's path and that its leading components are all real directories. So, any later change in these conditions indicates that there was a collision (either between two parallel-eligible entries or between an eligible and an ineligible one). After all parallel-eligible entries have been processed, the collided (and thus, skipped) entries are sequentially fed to checkout_entry() again. This is similar to the way the current code deals with collisions, overwriting the previously checked out entries with the subsequent ones. The only difference is that, since we no longer create the files in the same order that they appear on index, we are not able to determine which of the colliding entries will survive on disk (for the classic code, it is always the last entry). Co-authored-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Co-authored-by: Jeff Hostetler <jeffhost@microsoft.com> Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2021-04-19 02:14:53 +02:00
enable_delayed_checkout(&state);
parallel-checkout: add configuration options Make parallel checkout configurable by introducing two new settings: checkout.workers and checkout.thresholdForParallelism. The first defines the number of workers (where one means sequential checkout), and the second defines the minimum number of entries to attempt parallel checkout. To decide the default value for checkout.workers, the parallel version was benchmarked during three operations in the linux repo, with cold cache: cloning v5.8, checking out v5.8 from v2.6.15 (checkout I) and checking out v5.8 from v5.7 (checkout II). The four tables below show the mean run times and standard deviations for 5 runs in: a local file system on SSD, a local file system on HDD, a Linux NFS server, and Amazon EFS (all on Linux). Each parallel checkout test was executed with the number of workers that brings the best overall results in that environment. Local SSD: Sequential 10 workers Speedup Clone 8.805 s ± 0.043 s 3.564 s ± 0.041 s 2.47 ± 0.03 Checkout I 9.678 s ± 0.057 s 4.486 s ± 0.050 s 2.16 ± 0.03 Checkout II 5.034 s ± 0.072 s 3.021 s ± 0.038 s 1.67 ± 0.03 Local HDD: Sequential 10 workers Speedup Clone 32.288 s ± 0.580 s 30.724 s ± 0.522 s 1.05 ± 0.03 Checkout I 54.172 s ± 7.119 s 54.429 s ± 6.738 s 1.00 ± 0.18 Checkout II 40.465 s ± 2.402 s 38.682 s ± 1.365 s 1.05 ± 0.07 Linux NFS server (v4.1, on EBS, single availability zone): Sequential 32 workers Speedup Clone 240.368 s ± 6.347 s 57.349 s ± 0.870 s 4.19 ± 0.13 Checkout I 242.862 s ± 2.215 s 58.700 s ± 0.904 s 4.14 ± 0.07 Checkout II 65.751 s ± 1.577 s 23.820 s ± 0.407 s 2.76 ± 0.08 EFS (v4.1, replicated over multiple availability zones): Sequential 32 workers Speedup Clone 922.321 s ± 2.274 s 210.453 s ± 3.412 s 4.38 ± 0.07 Checkout I 1011.300 s ± 7.346 s 297.828 s ± 0.964 s 3.40 ± 0.03 Checkout II 294.104 s ± 1.836 s 126.017 s ± 1.190 s 2.33 ± 0.03 The above benchmarks show that parallel checkout is most effective on repositories located on an SSD or over a distributed file system. For local file systems on spinning disks, and/or older machines, the parallelism does not always bring a good performance. For this reason, the default value for checkout.workers is one, a.k.a. sequential checkout. To decide the default value for checkout.thresholdForParallelism, another benchmark was executed in the "Local SSD" setup, where parallel checkout showed to be beneficial. This time, we compared the runtime of a `git checkout -f`, with and without parallelism, after randomly removing an increasing number of files from the Linux working tree. The "sequential fallback" column below corresponds to the executions where checkout.workers was 10 but checkout.thresholdForParallelism was equal to the number of to-be-updated files plus one (so that we end up writing sequentially). Each test case was sampled 15 times, and each sample had a randomly different set of files removed. Here are the results: sequential fallback 10 workers speedup 10 files 772.3 ms ± 12.6 ms 769.0 ms ± 13.6 ms 1.00 ± 0.02 20 files 780.5 ms ± 15.8 ms 775.2 ms ± 9.2 ms 1.01 ± 0.02 50 files 806.2 ms ± 13.8 ms 767.4 ms ± 8.5 ms 1.05 ± 0.02 100 files 833.7 ms ± 21.4 ms 750.5 ms ± 16.8 ms 1.11 ± 0.04 200 files 897.6 ms ± 30.9 ms 730.5 ms ± 14.7 ms 1.23 ± 0.05 500 files 1035.4 ms ± 48.0 ms 677.1 ms ± 22.3 ms 1.53 ± 0.09 1000 files 1244.6 ms ± 35.6 ms 654.0 ms ± 38.3 ms 1.90 ± 0.12 2000 files 1488.8 ms ± 53.4 ms 658.8 ms ± 23.8 ms 2.26 ± 0.12 From the above numbers, 100 files seems to be a reasonable default value for the threshold setting. Note: Up to 1000 files, we observe a drop in the execution time of the parallel code with an increase in the number of files. This is a rather odd behavior, but it was observed in multiple repetitions. Above 1000 files, the execution time increases according to the number of files, as one would expect. About the test environments: Local SSD tests were executed on an i7-7700HQ (4 cores with hyper-threading) running Manjaro Linux. Local HDD tests were executed on an Intel(R) Xeon(R) E3-1230 (also 4 cores with hyper-threading), HDD Seagate Barracuda 7200.14 SATA 3.1, running Debian. NFS and EFS tests were executed on an Amazon EC2 c5n.xlarge instance, with 4 vCPUs. The Linux NFS server was running on a m6g.large instance with 2 vCPUSs and a 1 TB EBS GP2 volume. Before each timing, the linux repository was removed (or checked out back to its previous state), and `sync && sysctl vm.drop_caches=3` was executed. Co-authored-by: Jeff Hostetler <jeffhost@microsoft.com> Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2021-04-19 02:14:55 +02:00
if (pc_workers > 1)
init_parallel_checkout();
for (i = 0; i < index->cache_nr; i++) {
struct cache_entry *ce = index->cache[i];
if (must_checkout(ce)) {
size_t last_pc_queue_size = pc_queue_size();
if (ce->ce_flags & CE_WT_REMOVE)
BUG("both update and delete flags are set on %s",
ce->name);
ce->ce_flags &= ~CE_UPDATE;
unpack-trees: exit check_updates() early if updates are not wanted check_updates() has a lot of code that repeatedly checks whether o->update or o->dry_run are set. (Note that o->dry_run is a near-synonym for !o->update, but not quite as per commit 2c9078d05bf2 ("unpack-trees: add the dry_run flag to unpack_trees_options", 2011-05-25).) In fact, this function almost turns into a no-op whenever the condition !o->update || o->dry_run is met. Simplify the code by checking this condition at the beginning of the function, and when it is true, do the few things that are relevant and return early. There are a few things that make the conversion not quite obvious: * The fact that check_updates() does not actually turn into a no-op when updates are not wanted may be slightly surprising. However, commit 33ecf7eb61 (Discard "deleted" cache entries after using them to update the working tree, 2008-02-07) put the discarding of unused cache entries in check_updates() so we still need to keep the call to remove_marked_cache_entries(). It's possible this call belongs in another function, but it is certainly needed as tests will fail if it is removed. * The original called remove_scheduled_dirs() unconditionally. Technically, commit 7847892716 (unlink_entry(): introduce schedule_dir_for_removal(), 2009-02-09) should have made that call conditional, but it didn't matter in practice because remove_scheduled_dirs() becomes a no-op when all the calls to unlink_entry() are skipped. As such, we do not need to call it. * When (o->dry_run && o->update), the original would have two calls to git_attr_set_direction() surrounding a bunch of skipped updates. These two calls to git_attr_set_direction() cancel each other out and thus can be omitted when o->dry_run is true just as they already are when !o->update. * The code would previously call setup_collided_checkout_detection() and report_collided_checkout() even when o->dry_run. However, this was just an expensive no-op because setup_collided_checkout_detection() merely cleared the CE_MATCHED flag for each cache entry, and report_collided_checkout() reported which ones had it set. Since a dry-run would skip all the checkout_entry() calls, CE_MATCHED would never get set and thus no collisions would be reported. Since we can't detect the collisions anyway without doing updates, skipping the collisions detection setup and reporting is an optimization. * The code previously would call get_progress() and display_progress() even when (!o->update || o->dry_run). This served to show how long it took to skip all the updates, which is somewhat useless. Since we are skipping the updates, we can skip showing how long it takes to skip them. Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-01-07 07:57:57 +01:00
errs |= checkout_entry(ce, &state, NULL, NULL);
if (last_pc_queue_size == pc_queue_size())
display_progress(progress, ++cnt);
}
}
parallel-checkout: add configuration options Make parallel checkout configurable by introducing two new settings: checkout.workers and checkout.thresholdForParallelism. The first defines the number of workers (where one means sequential checkout), and the second defines the minimum number of entries to attempt parallel checkout. To decide the default value for checkout.workers, the parallel version was benchmarked during three operations in the linux repo, with cold cache: cloning v5.8, checking out v5.8 from v2.6.15 (checkout I) and checking out v5.8 from v5.7 (checkout II). The four tables below show the mean run times and standard deviations for 5 runs in: a local file system on SSD, a local file system on HDD, a Linux NFS server, and Amazon EFS (all on Linux). Each parallel checkout test was executed with the number of workers that brings the best overall results in that environment. Local SSD: Sequential 10 workers Speedup Clone 8.805 s ± 0.043 s 3.564 s ± 0.041 s 2.47 ± 0.03 Checkout I 9.678 s ± 0.057 s 4.486 s ± 0.050 s 2.16 ± 0.03 Checkout II 5.034 s ± 0.072 s 3.021 s ± 0.038 s 1.67 ± 0.03 Local HDD: Sequential 10 workers Speedup Clone 32.288 s ± 0.580 s 30.724 s ± 0.522 s 1.05 ± 0.03 Checkout I 54.172 s ± 7.119 s 54.429 s ± 6.738 s 1.00 ± 0.18 Checkout II 40.465 s ± 2.402 s 38.682 s ± 1.365 s 1.05 ± 0.07 Linux NFS server (v4.1, on EBS, single availability zone): Sequential 32 workers Speedup Clone 240.368 s ± 6.347 s 57.349 s ± 0.870 s 4.19 ± 0.13 Checkout I 242.862 s ± 2.215 s 58.700 s ± 0.904 s 4.14 ± 0.07 Checkout II 65.751 s ± 1.577 s 23.820 s ± 0.407 s 2.76 ± 0.08 EFS (v4.1, replicated over multiple availability zones): Sequential 32 workers Speedup Clone 922.321 s ± 2.274 s 210.453 s ± 3.412 s 4.38 ± 0.07 Checkout I 1011.300 s ± 7.346 s 297.828 s ± 0.964 s 3.40 ± 0.03 Checkout II 294.104 s ± 1.836 s 126.017 s ± 1.190 s 2.33 ± 0.03 The above benchmarks show that parallel checkout is most effective on repositories located on an SSD or over a distributed file system. For local file systems on spinning disks, and/or older machines, the parallelism does not always bring a good performance. For this reason, the default value for checkout.workers is one, a.k.a. sequential checkout. To decide the default value for checkout.thresholdForParallelism, another benchmark was executed in the "Local SSD" setup, where parallel checkout showed to be beneficial. This time, we compared the runtime of a `git checkout -f`, with and without parallelism, after randomly removing an increasing number of files from the Linux working tree. The "sequential fallback" column below corresponds to the executions where checkout.workers was 10 but checkout.thresholdForParallelism was equal to the number of to-be-updated files plus one (so that we end up writing sequentially). Each test case was sampled 15 times, and each sample had a randomly different set of files removed. Here are the results: sequential fallback 10 workers speedup 10 files 772.3 ms ± 12.6 ms 769.0 ms ± 13.6 ms 1.00 ± 0.02 20 files 780.5 ms ± 15.8 ms 775.2 ms ± 9.2 ms 1.01 ± 0.02 50 files 806.2 ms ± 13.8 ms 767.4 ms ± 8.5 ms 1.05 ± 0.02 100 files 833.7 ms ± 21.4 ms 750.5 ms ± 16.8 ms 1.11 ± 0.04 200 files 897.6 ms ± 30.9 ms 730.5 ms ± 14.7 ms 1.23 ± 0.05 500 files 1035.4 ms ± 48.0 ms 677.1 ms ± 22.3 ms 1.53 ± 0.09 1000 files 1244.6 ms ± 35.6 ms 654.0 ms ± 38.3 ms 1.90 ± 0.12 2000 files 1488.8 ms ± 53.4 ms 658.8 ms ± 23.8 ms 2.26 ± 0.12 From the above numbers, 100 files seems to be a reasonable default value for the threshold setting. Note: Up to 1000 files, we observe a drop in the execution time of the parallel code with an increase in the number of files. This is a rather odd behavior, but it was observed in multiple repetitions. Above 1000 files, the execution time increases according to the number of files, as one would expect. About the test environments: Local SSD tests were executed on an i7-7700HQ (4 cores with hyper-threading) running Manjaro Linux. Local HDD tests were executed on an Intel(R) Xeon(R) E3-1230 (also 4 cores with hyper-threading), HDD Seagate Barracuda 7200.14 SATA 3.1, running Debian. NFS and EFS tests were executed on an Amazon EC2 c5n.xlarge instance, with 4 vCPUs. The Linux NFS server was running on a m6g.large instance with 2 vCPUSs and a 1 TB EBS GP2 volume. Before each timing, the linux repository was removed (or checked out back to its previous state), and `sync && sysctl vm.drop_caches=3` was executed. Co-authored-by: Jeff Hostetler <jeffhost@microsoft.com> Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2021-04-19 02:14:55 +02:00
if (pc_workers > 1)
errs |= run_parallel_checkout(&state, pc_workers, pc_threshold,
progress, &cnt);
stop_progress(&progress);
errs |= finish_delayed_checkout(&state, o->verbose_update);
unpack-trees: exit check_updates() early if updates are not wanted check_updates() has a lot of code that repeatedly checks whether o->update or o->dry_run are set. (Note that o->dry_run is a near-synonym for !o->update, but not quite as per commit 2c9078d05bf2 ("unpack-trees: add the dry_run flag to unpack_trees_options", 2011-05-25).) In fact, this function almost turns into a no-op whenever the condition !o->update || o->dry_run is met. Simplify the code by checking this condition at the beginning of the function, and when it is true, do the few things that are relevant and return early. There are a few things that make the conversion not quite obvious: * The fact that check_updates() does not actually turn into a no-op when updates are not wanted may be slightly surprising. However, commit 33ecf7eb61 (Discard "deleted" cache entries after using them to update the working tree, 2008-02-07) put the discarding of unused cache entries in check_updates() so we still need to keep the call to remove_marked_cache_entries(). It's possible this call belongs in another function, but it is certainly needed as tests will fail if it is removed. * The original called remove_scheduled_dirs() unconditionally. Technically, commit 7847892716 (unlink_entry(): introduce schedule_dir_for_removal(), 2009-02-09) should have made that call conditional, but it didn't matter in practice because remove_scheduled_dirs() becomes a no-op when all the calls to unlink_entry() are skipped. As such, we do not need to call it. * When (o->dry_run && o->update), the original would have two calls to git_attr_set_direction() surrounding a bunch of skipped updates. These two calls to git_attr_set_direction() cancel each other out and thus can be omitted when o->dry_run is true just as they already are when !o->update. * The code would previously call setup_collided_checkout_detection() and report_collided_checkout() even when o->dry_run. However, this was just an expensive no-op because setup_collided_checkout_detection() merely cleared the CE_MATCHED flag for each cache entry, and report_collided_checkout() reported which ones had it set. Since a dry-run would skip all the checkout_entry() calls, CE_MATCHED would never get set and thus no collisions would be reported. Since we can't detect the collisions anyway without doing updates, skipping the collisions detection setup and reporting is an optimization. * The code previously would call get_progress() and display_progress() even when (!o->update || o->dry_run). This served to show how long it took to skip all the updates, which is somewhat useless. Since we are skipping the updates, we can skip showing how long it takes to skip them. Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-01-07 07:57:57 +01:00
git_attr_set_direction(GIT_ATTR_CHECKIN);
if (o->clone)
report_collided_checkout(index);
trace_performance_leave("check_updates");
return errs != 0;
}
static int verify_uptodate_sparse(const struct cache_entry *ce,
struct unpack_trees_options *o);
static int verify_absent_sparse(const struct cache_entry *ce,
enum unpack_trees_error_types,
struct unpack_trees_options *o);
static int apply_sparse_checkout(struct index_state *istate,
struct cache_entry *ce,
struct unpack_trees_options *o)
{
int was_skip_worktree = ce_skip_worktree(ce);
if (ce->ce_flags & CE_NEW_SKIP_WORKTREE)
ce->ce_flags |= CE_SKIP_WORKTREE;
else
ce->ce_flags &= ~CE_SKIP_WORKTREE;
if (was_skip_worktree != ce_skip_worktree(ce)) {
ce->ce_flags |= CE_UPDATE_IN_BASE;
mark_fsmonitor_invalid(istate, ce);
istate->cache_changed |= CE_ENTRY_CHANGED;
}
/*
* if (!was_skip_worktree && !ce_skip_worktree()) {
* This is perfectly normal. Move on;
* }
*/
/*
* Merge strategies may set CE_UPDATE|CE_REMOVE outside checkout
* area as a result of ce_skip_worktree() shortcuts in
* verify_absent() and verify_uptodate().
* Make sure they don't modify worktree if they are already
* outside checkout area
*/
if (was_skip_worktree && ce_skip_worktree(ce)) {
ce->ce_flags &= ~CE_UPDATE;
/*
* By default, when CE_REMOVE is on, CE_WT_REMOVE is also
* on to get that file removed from both index and worktree.
* If that file is already outside worktree area, don't
* bother remove it.
*/
if (ce->ce_flags & CE_REMOVE)
ce->ce_flags &= ~CE_WT_REMOVE;
}
if (!was_skip_worktree && ce_skip_worktree(ce)) {
/*
* If CE_UPDATE is set, verify_uptodate() must be called already
* also stat info may have lost after merged_entry() so calling
* verify_uptodate() again may fail
*/
if (!(ce->ce_flags & CE_UPDATE) &&
verify_uptodate_sparse(ce, o)) {
ce->ce_flags &= ~CE_SKIP_WORKTREE;
return -1;
}
ce->ce_flags |= CE_WT_REMOVE;
ce->ce_flags &= ~CE_UPDATE;
}
if (was_skip_worktree && !ce_skip_worktree(ce)) {
if (verify_absent_sparse(ce, WARNING_SPARSE_ORPHANED_NOT_OVERWRITTEN, o))
return -1;
ce->ce_flags |= CE_UPDATE;
}
return 0;
}
static int warn_conflicted_path(struct index_state *istate,
int i,
struct unpack_trees_options *o)
{
char *conflicting_path = istate->cache[i]->name;
int count = 0;
add_rejected_path(o, WARNING_SPARSE_UNMERGED_FILE, conflicting_path);
/* Find out how many higher stage entries are at same path */
while ((++count) + i < istate->cache_nr &&
!strcmp(conflicting_path, istate->cache[count + i]->name))
; /* do nothing */
return count;
}
static inline int call_unpack_fn(const struct cache_entry * const *src,
struct unpack_trees_options *o)
{
int ret = o->fn(src, o);
if (ret > 0)
ret = 0;
return ret;
}
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
static void mark_ce_used(struct cache_entry *ce, struct unpack_trees_options *o)
{
ce->ce_flags |= CE_UNPACKED;
if (o->cache_bottom < o->src_index->cache_nr &&
o->src_index->cache[o->cache_bottom] == ce) {
int bottom = o->cache_bottom;
while (bottom < o->src_index->cache_nr &&
o->src_index->cache[bottom]->ce_flags & CE_UNPACKED)
bottom++;
o->cache_bottom = bottom;
}
}
static void mark_all_ce_unused(struct index_state *index)
{
int i;
for (i = 0; i < index->cache_nr; i++)
index->cache[i]->ce_flags &= ~(CE_UNPACKED | CE_ADDED | CE_NEW_SKIP_WORKTREE);
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
}
static int locate_in_src_index(const struct cache_entry *ce,
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
struct unpack_trees_options *o)
{
struct index_state *index = o->src_index;
int len = ce_namelen(ce);
int pos = index_name_pos(index, ce->name, len);
if (pos < 0)
pos = -1 - pos;
return pos;
}
/*
* We call unpack_index_entry() with an unmerged cache entry
* only in diff-index, and it wants a single callback. Skip
* the other unmerged entry with the same name.
*/
static void mark_ce_used_same_name(struct cache_entry *ce,
struct unpack_trees_options *o)
{
struct index_state *index = o->src_index;
int len = ce_namelen(ce);
int pos;
for (pos = locate_in_src_index(ce, o); pos < index->cache_nr; pos++) {
struct cache_entry *next = index->cache[pos];
if (len != ce_namelen(next) ||
memcmp(ce->name, next->name, len))
break;
mark_ce_used(next, o);
}
}
static struct cache_entry *next_cache_entry(struct unpack_trees_options *o)
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
{
const struct index_state *index = o->src_index;
int pos = o->cache_bottom;
while (pos < index->cache_nr) {
struct cache_entry *ce = index->cache[pos];
if (!(ce->ce_flags & CE_UNPACKED))
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
return ce;
pos++;
}
return NULL;
}
Convert "struct cache_entry *" to "const ..." wherever possible I attempted to make index_state->cache[] a "const struct cache_entry **" to find out how existing entries in index are modified and where. The question I have is what do we do if we really need to keep track of on-disk changes in the index. The result is - diff-lib.c: setting CE_UPTODATE - name-hash.c: setting CE_HASHED - preload-index.c, read-cache.c, unpack-trees.c and builtin/update-index: obvious - entry.c: write_entry() may refresh the checked out entry via fill_stat_cache_info(). This causes "non-const struct cache_entry *" in builtin/apply.c, builtin/checkout-index.c and builtin/checkout.c - builtin/ls-files.c: --with-tree changes stagemask and may set CE_UPDATE Of these, write_entry() and its call sites are probably most interesting because it modifies on-disk info. But this is stat info and can be retrieved via refresh, at least for porcelain commands. Other just uses ce_flags for local purposes. So, keeping track of "dirty" entries is just a matter of setting a flag in index modification functions exposed by read-cache.c. Except unpack-trees, the rest of the code base does not do anything funny behind read-cache's back. The actual patch is less valueable than the summary above. But if anyone wants to re-identify the above sites. Applying this patch, then this: diff --git a/cache.h b/cache.h index 430d021..1692891 100644 --- a/cache.h +++ b/cache.h @@ -267,7 +267,7 @@ static inline unsigned int canon_mode(unsigned int mode) #define cache_entry_size(len) (offsetof(struct cache_entry,name) + (len) + 1) struct index_state { - struct cache_entry **cache; + const struct cache_entry **cache; unsigned int version; unsigned int cache_nr, cache_alloc, cache_changed; struct string_list *resolve_undo; will help quickly identify them without bogus warnings. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-07-09 17:29:00 +02:00
static void add_same_unmerged(const struct cache_entry *ce,
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
struct unpack_trees_options *o)
{
struct index_state *index = o->src_index;
int len = ce_namelen(ce);
int pos = index_name_pos(index, ce->name, len);
if (0 <= pos)
die("programming error in a caller of mark_ce_used_same_name");
for (pos = -pos - 1; pos < index->cache_nr; pos++) {
struct cache_entry *next = index->cache[pos];
if (len != ce_namelen(next) ||
memcmp(ce->name, next->name, len))
break;
add_entry(o, next, 0, 0);
mark_ce_used(next, o);
}
}
static int unpack_index_entry(struct cache_entry *ce,
struct unpack_trees_options *o)
{
const struct cache_entry *src[MAX_UNPACK_TREES + 1] = { NULL, };
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
int ret;
src[0] = ce;
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
mark_ce_used(ce, o);
if (ce_stage(ce)) {
if (o->skip_unmerged) {
add_entry(o, ce, 0, 0);
return 0;
}
}
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
ret = call_unpack_fn(src, o);
if (ce_stage(ce))
mark_ce_used_same_name(ce, o);
return ret;
}
static int find_cache_pos(struct traverse_info *, const char *p, size_t len);
unpack-trees.c: look ahead in the index This makes the traversal of index be in sync with the tree traversal. When unpack_callback() is fed a set of tree entries from trees, it inspects the name of the entry and checks if the an index entry with the same name could be hiding behind the current index entry, and (1) if the name appears in the index as a leaf node, it is also fed to the n_way_merge() callback function; (2) if the name is a directory in the index, i.e. there are entries in that are underneath it, then nothing is fed to the n_way_merge() callback function; (3) otherwise, if the name comes before the first eligible entry in the index, the index entry is first unpacked alone. When traverse_trees_recursive() descends into a subdirectory, the cache_bottom pointer is moved to walk index entries within that directory. All of these are omitted for diff-index, which does not even want to be fed an index entry and a tree entry with D/F conflicts. This fixes 3-way read-tree and exposes a bug in other parts of the system in t6035, test #5. The test prepares these three trees: O = HEAD^ 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x A = HEAD 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob 587be6b4c3f93f93c489c0111bba5596147a26cb a/x B = master 120000 blob a36b77384451ea1de7bd340ffca868249626bc52 a/b 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x With a clean index that matches HEAD, running git read-tree -m -u --aggressive $O $A $B now yields 120000 a36b77384451ea1de7bd340ffca868249626bc52 3 a/b 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 0 a/b-2/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 1 a/b/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 2 a/b/c/d 100644 587be6b4c3f93f93c489c0111bba5596147a26cb 0 a/x which is correct. "master" created "a/b" symlink that did not exist, and removed "a/b/c/d" while HEAD did not do touch either path. Before this series, read-tree did not notice the situation and resolved addition of "a/b" and removal of "a/b/c/d" independently. If A = HEAD had another path "a/b/c/e" added, this merge should conflict but instead it silently resolved "a/b" and then immediately overwrote it to add "a/b/c/e", which was quite bogus. Tests in t1012 start to work with this. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-09-20 09:03:39 +02:00
static void restore_cache_bottom(struct traverse_info *info, int bottom)
{
struct unpack_trees_options *o = info->data;
if (o->diff_index_cached)
return;
o->cache_bottom = bottom;
}
static int switch_cache_bottom(struct traverse_info *info)
{
struct unpack_trees_options *o = info->data;
int ret, pos;
if (o->diff_index_cached)
return 0;
ret = o->cache_bottom;
pos = find_cache_pos(info->prev, info->name, info->namelen);
unpack-trees.c: look ahead in the index This makes the traversal of index be in sync with the tree traversal. When unpack_callback() is fed a set of tree entries from trees, it inspects the name of the entry and checks if the an index entry with the same name could be hiding behind the current index entry, and (1) if the name appears in the index as a leaf node, it is also fed to the n_way_merge() callback function; (2) if the name is a directory in the index, i.e. there are entries in that are underneath it, then nothing is fed to the n_way_merge() callback function; (3) otherwise, if the name comes before the first eligible entry in the index, the index entry is first unpacked alone. When traverse_trees_recursive() descends into a subdirectory, the cache_bottom pointer is moved to walk index entries within that directory. All of these are omitted for diff-index, which does not even want to be fed an index entry and a tree entry with D/F conflicts. This fixes 3-way read-tree and exposes a bug in other parts of the system in t6035, test #5. The test prepares these three trees: O = HEAD^ 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x A = HEAD 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob 587be6b4c3f93f93c489c0111bba5596147a26cb a/x B = master 120000 blob a36b77384451ea1de7bd340ffca868249626bc52 a/b 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x With a clean index that matches HEAD, running git read-tree -m -u --aggressive $O $A $B now yields 120000 a36b77384451ea1de7bd340ffca868249626bc52 3 a/b 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 0 a/b-2/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 1 a/b/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 2 a/b/c/d 100644 587be6b4c3f93f93c489c0111bba5596147a26cb 0 a/x which is correct. "master" created "a/b" symlink that did not exist, and removed "a/b/c/d" while HEAD did not do touch either path. Before this series, read-tree did not notice the situation and resolved addition of "a/b" and removal of "a/b/c/d" independently. If A = HEAD had another path "a/b/c/e" added, this merge should conflict but instead it silently resolved "a/b" and then immediately overwrote it to add "a/b/c/e", which was quite bogus. Tests in t1012 start to work with this. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-09-20 09:03:39 +02:00
if (pos < -1)
o->cache_bottom = -2 - pos;
else if (pos < 0)
o->cache_bottom = o->src_index->cache_nr;
return ret;
}
static inline int are_same_oid(struct name_entry *name_j, struct name_entry *name_k)
{
return !is_null_oid(&name_j->oid) && !is_null_oid(&name_k->oid) && oideq(&name_j->oid, &name_k->oid);
}
unpack-trees: optimize walking same trees with cache-tree In order to merge one or many trees with the index, unpack-trees code walks multiple trees in parallel with the index and performs n-way merge. If we find out at start of a directory that all trees are the same (by comparing OID) and cache-tree happens to be available for that directory as well, we could avoid walking the trees because we already know what these trees contain: it's flattened in what's called "the index". The upside is of course a lot less I/O since we can potentially skip lots of trees (think subtrees). We also save CPU because we don't have to inflate and apply the deltas. The downside is of course more fragile code since the logic in some functions are now duplicated elsewhere. "checkout -" with this patch on webkit.git (275k files): baseline new -------------------------------------------------------------------- 0.056651714 0.080394752 s: read cache .git/index 0.183101080 0.216010838 s: preload index 0.008584433 0.008534301 s: refresh index 0.633767589 0.251992198 s: traverse_trees 0.340265448 0.377031383 s: check_updates 0.381884638 0.372768105 s: cache_tree_update 1.401562947 1.045887251 s: unpack_trees 0.338687914 0.314983512 s: write index, changed mask = 2e 0.411927922 0.062572653 s: traverse_trees 0.000023335 0.000022544 s: check_updates 0.423697246 0.073795585 s: unpack_trees 0.423708360 0.073807557 s: diff-index 2.559524127 1.938191592 s: git command: git checkout - Another measurement from Ben's running "git checkout" with over 500k trees (on the whole series): baseline new ---------------------------------------------------------------------- 0.535510167 0.556558733 s: read cache .git/index 0.3057373 0.3147105 s: initialize name hash 0.0184082 0.023558433 s: preload index 0.086910967 0.089085967 s: refresh index 7.889590767 2.191554433 s: unpack trees 0.120760833 0.131941267 s: update worktree after a merge 2.2583504 2.572663167 s: repair cache-tree 0.8916137 0.959495233 s: write index, changed mask = 28 3.405199233 0.2710663 s: unpack trees 0.000999667 0.0021554 s: update worktree after a merge 3.4063306 0.273318333 s: diff-index 16.9524923 9.462943133 s: git command: git.exe checkout This command calls unpack_trees() twice, the first time on 2way merge and the second 1way merge. In both times, "unpack trees" time is reduced to one third. Overall time reduction is not that impressive of course because index operations take a big chunk. And there's that repair cache-tree line. PS. A note about cache-tree invalidation and the use of it in this code. We do invalidate cache-tree in _source_ index when we add new entries to the (temporary) "result" index. But we also use the cache-tree from source index in this optimization. Does this mean we end up having no cache-tree in the source index to activate this optimization? The answer is twisted: the order of finding a good cache-tree and invalidating it matters. In this case we check for a good cache-tree first in all_trees_same_as_cache_tree(), then we start to merge things and potentially invalidate that same cache-tree in the process. Since cache-tree invalidation happens after the optimization kicks in, we're still good. But we may lose that cache-tree at the very first call_unpack_fn() call in traverse_by_cache_tree(). Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-08-18 16:41:24 +02:00
static int all_trees_same_as_cache_tree(int n, unsigned long dirmask,
struct name_entry *names,
struct traverse_info *info)
{
struct unpack_trees_options *o = info->data;
int i;
if (!o->merge || dirmask != ((1 << n) - 1))
return 0;
for (i = 1; i < n; i++)
if (!are_same_oid(names, names + i))
return 0;
return cache_tree_matches_traversal(o->src_index->cache_tree, names, info);
}
static int index_pos_by_traverse_info(struct name_entry *names,
struct traverse_info *info)
{
struct unpack_trees_options *o = info->data;
struct strbuf name = STRBUF_INIT;
unpack-trees: optimize walking same trees with cache-tree In order to merge one or many trees with the index, unpack-trees code walks multiple trees in parallel with the index and performs n-way merge. If we find out at start of a directory that all trees are the same (by comparing OID) and cache-tree happens to be available for that directory as well, we could avoid walking the trees because we already know what these trees contain: it's flattened in what's called "the index". The upside is of course a lot less I/O since we can potentially skip lots of trees (think subtrees). We also save CPU because we don't have to inflate and apply the deltas. The downside is of course more fragile code since the logic in some functions are now duplicated elsewhere. "checkout -" with this patch on webkit.git (275k files): baseline new -------------------------------------------------------------------- 0.056651714 0.080394752 s: read cache .git/index 0.183101080 0.216010838 s: preload index 0.008584433 0.008534301 s: refresh index 0.633767589 0.251992198 s: traverse_trees 0.340265448 0.377031383 s: check_updates 0.381884638 0.372768105 s: cache_tree_update 1.401562947 1.045887251 s: unpack_trees 0.338687914 0.314983512 s: write index, changed mask = 2e 0.411927922 0.062572653 s: traverse_trees 0.000023335 0.000022544 s: check_updates 0.423697246 0.073795585 s: unpack_trees 0.423708360 0.073807557 s: diff-index 2.559524127 1.938191592 s: git command: git checkout - Another measurement from Ben's running "git checkout" with over 500k trees (on the whole series): baseline new ---------------------------------------------------------------------- 0.535510167 0.556558733 s: read cache .git/index 0.3057373 0.3147105 s: initialize name hash 0.0184082 0.023558433 s: preload index 0.086910967 0.089085967 s: refresh index 7.889590767 2.191554433 s: unpack trees 0.120760833 0.131941267 s: update worktree after a merge 2.2583504 2.572663167 s: repair cache-tree 0.8916137 0.959495233 s: write index, changed mask = 28 3.405199233 0.2710663 s: unpack trees 0.000999667 0.0021554 s: update worktree after a merge 3.4063306 0.273318333 s: diff-index 16.9524923 9.462943133 s: git command: git.exe checkout This command calls unpack_trees() twice, the first time on 2way merge and the second 1way merge. In both times, "unpack trees" time is reduced to one third. Overall time reduction is not that impressive of course because index operations take a big chunk. And there's that repair cache-tree line. PS. A note about cache-tree invalidation and the use of it in this code. We do invalidate cache-tree in _source_ index when we add new entries to the (temporary) "result" index. But we also use the cache-tree from source index in this optimization. Does this mean we end up having no cache-tree in the source index to activate this optimization? The answer is twisted: the order of finding a good cache-tree and invalidating it matters. In this case we check for a good cache-tree first in all_trees_same_as_cache_tree(), then we start to merge things and potentially invalidate that same cache-tree in the process. Since cache-tree invalidation happens after the optimization kicks in, we're still good. But we may lose that cache-tree at the very first call_unpack_fn() call in traverse_by_cache_tree(). Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-08-18 16:41:24 +02:00
int pos;
strbuf_make_traverse_path(&name, info, names->path, names->pathlen);
strbuf_addch(&name, '/');
pos = index_name_pos(o->src_index, name.buf, name.len);
if (pos >= 0) {
if (!o->src_index->sparse_index ||
!(o->src_index->cache[pos]->ce_flags & CE_SKIP_WORKTREE))
BUG("This is a directory and should not exist in index");
} else {
pos = -pos - 1;
}
if (pos >= o->src_index->cache_nr ||
!starts_with(o->src_index->cache[pos]->name, name.buf) ||
(pos > 0 && starts_with(o->src_index->cache[pos-1]->name, name.buf)))
BUG("pos %d doesn't point to the first entry of %s in index",
pos, name.buf);
strbuf_release(&name);
unpack-trees: optimize walking same trees with cache-tree In order to merge one or many trees with the index, unpack-trees code walks multiple trees in parallel with the index and performs n-way merge. If we find out at start of a directory that all trees are the same (by comparing OID) and cache-tree happens to be available for that directory as well, we could avoid walking the trees because we already know what these trees contain: it's flattened in what's called "the index". The upside is of course a lot less I/O since we can potentially skip lots of trees (think subtrees). We also save CPU because we don't have to inflate and apply the deltas. The downside is of course more fragile code since the logic in some functions are now duplicated elsewhere. "checkout -" with this patch on webkit.git (275k files): baseline new -------------------------------------------------------------------- 0.056651714 0.080394752 s: read cache .git/index 0.183101080 0.216010838 s: preload index 0.008584433 0.008534301 s: refresh index 0.633767589 0.251992198 s: traverse_trees 0.340265448 0.377031383 s: check_updates 0.381884638 0.372768105 s: cache_tree_update 1.401562947 1.045887251 s: unpack_trees 0.338687914 0.314983512 s: write index, changed mask = 2e 0.411927922 0.062572653 s: traverse_trees 0.000023335 0.000022544 s: check_updates 0.423697246 0.073795585 s: unpack_trees 0.423708360 0.073807557 s: diff-index 2.559524127 1.938191592 s: git command: git checkout - Another measurement from Ben's running "git checkout" with over 500k trees (on the whole series): baseline new ---------------------------------------------------------------------- 0.535510167 0.556558733 s: read cache .git/index 0.3057373 0.3147105 s: initialize name hash 0.0184082 0.023558433 s: preload index 0.086910967 0.089085967 s: refresh index 7.889590767 2.191554433 s: unpack trees 0.120760833 0.131941267 s: update worktree after a merge 2.2583504 2.572663167 s: repair cache-tree 0.8916137 0.959495233 s: write index, changed mask = 28 3.405199233 0.2710663 s: unpack trees 0.000999667 0.0021554 s: update worktree after a merge 3.4063306 0.273318333 s: diff-index 16.9524923 9.462943133 s: git command: git.exe checkout This command calls unpack_trees() twice, the first time on 2way merge and the second 1way merge. In both times, "unpack trees" time is reduced to one third. Overall time reduction is not that impressive of course because index operations take a big chunk. And there's that repair cache-tree line. PS. A note about cache-tree invalidation and the use of it in this code. We do invalidate cache-tree in _source_ index when we add new entries to the (temporary) "result" index. But we also use the cache-tree from source index in this optimization. Does this mean we end up having no cache-tree in the source index to activate this optimization? The answer is twisted: the order of finding a good cache-tree and invalidating it matters. In this case we check for a good cache-tree first in all_trees_same_as_cache_tree(), then we start to merge things and potentially invalidate that same cache-tree in the process. Since cache-tree invalidation happens after the optimization kicks in, we're still good. But we may lose that cache-tree at the very first call_unpack_fn() call in traverse_by_cache_tree(). Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-08-18 16:41:24 +02:00
return pos;
}
/*
* Fast path if we detect that all trees are the same as cache-tree at this
* path. We'll walk these trees in an iterative loop using cache-tree/index
* instead of ODB since we already know what these trees contain.
unpack-trees: optimize walking same trees with cache-tree In order to merge one or many trees with the index, unpack-trees code walks multiple trees in parallel with the index and performs n-way merge. If we find out at start of a directory that all trees are the same (by comparing OID) and cache-tree happens to be available for that directory as well, we could avoid walking the trees because we already know what these trees contain: it's flattened in what's called "the index". The upside is of course a lot less I/O since we can potentially skip lots of trees (think subtrees). We also save CPU because we don't have to inflate and apply the deltas. The downside is of course more fragile code since the logic in some functions are now duplicated elsewhere. "checkout -" with this patch on webkit.git (275k files): baseline new -------------------------------------------------------------------- 0.056651714 0.080394752 s: read cache .git/index 0.183101080 0.216010838 s: preload index 0.008584433 0.008534301 s: refresh index 0.633767589 0.251992198 s: traverse_trees 0.340265448 0.377031383 s: check_updates 0.381884638 0.372768105 s: cache_tree_update 1.401562947 1.045887251 s: unpack_trees 0.338687914 0.314983512 s: write index, changed mask = 2e 0.411927922 0.062572653 s: traverse_trees 0.000023335 0.000022544 s: check_updates 0.423697246 0.073795585 s: unpack_trees 0.423708360 0.073807557 s: diff-index 2.559524127 1.938191592 s: git command: git checkout - Another measurement from Ben's running "git checkout" with over 500k trees (on the whole series): baseline new ---------------------------------------------------------------------- 0.535510167 0.556558733 s: read cache .git/index 0.3057373 0.3147105 s: initialize name hash 0.0184082 0.023558433 s: preload index 0.086910967 0.089085967 s: refresh index 7.889590767 2.191554433 s: unpack trees 0.120760833 0.131941267 s: update worktree after a merge 2.2583504 2.572663167 s: repair cache-tree 0.8916137 0.959495233 s: write index, changed mask = 28 3.405199233 0.2710663 s: unpack trees 0.000999667 0.0021554 s: update worktree after a merge 3.4063306 0.273318333 s: diff-index 16.9524923 9.462943133 s: git command: git.exe checkout This command calls unpack_trees() twice, the first time on 2way merge and the second 1way merge. In both times, "unpack trees" time is reduced to one third. Overall time reduction is not that impressive of course because index operations take a big chunk. And there's that repair cache-tree line. PS. A note about cache-tree invalidation and the use of it in this code. We do invalidate cache-tree in _source_ index when we add new entries to the (temporary) "result" index. But we also use the cache-tree from source index in this optimization. Does this mean we end up having no cache-tree in the source index to activate this optimization? The answer is twisted: the order of finding a good cache-tree and invalidating it matters. In this case we check for a good cache-tree first in all_trees_same_as_cache_tree(), then we start to merge things and potentially invalidate that same cache-tree in the process. Since cache-tree invalidation happens after the optimization kicks in, we're still good. But we may lose that cache-tree at the very first call_unpack_fn() call in traverse_by_cache_tree(). Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-08-18 16:41:24 +02:00
*/
static int traverse_by_cache_tree(int pos, int nr_entries, int nr_names,
struct traverse_info *info)
{
struct cache_entry *src[MAX_UNPACK_TREES + 1] = { NULL, };
struct unpack_trees_options *o = info->data;
struct cache_entry *tree_ce = NULL;
int ce_len = 0;
unpack-trees: optimize walking same trees with cache-tree In order to merge one or many trees with the index, unpack-trees code walks multiple trees in parallel with the index and performs n-way merge. If we find out at start of a directory that all trees are the same (by comparing OID) and cache-tree happens to be available for that directory as well, we could avoid walking the trees because we already know what these trees contain: it's flattened in what's called "the index". The upside is of course a lot less I/O since we can potentially skip lots of trees (think subtrees). We also save CPU because we don't have to inflate and apply the deltas. The downside is of course more fragile code since the logic in some functions are now duplicated elsewhere. "checkout -" with this patch on webkit.git (275k files): baseline new -------------------------------------------------------------------- 0.056651714 0.080394752 s: read cache .git/index 0.183101080 0.216010838 s: preload index 0.008584433 0.008534301 s: refresh index 0.633767589 0.251992198 s: traverse_trees 0.340265448 0.377031383 s: check_updates 0.381884638 0.372768105 s: cache_tree_update 1.401562947 1.045887251 s: unpack_trees 0.338687914 0.314983512 s: write index, changed mask = 2e 0.411927922 0.062572653 s: traverse_trees 0.000023335 0.000022544 s: check_updates 0.423697246 0.073795585 s: unpack_trees 0.423708360 0.073807557 s: diff-index 2.559524127 1.938191592 s: git command: git checkout - Another measurement from Ben's running "git checkout" with over 500k trees (on the whole series): baseline new ---------------------------------------------------------------------- 0.535510167 0.556558733 s: read cache .git/index 0.3057373 0.3147105 s: initialize name hash 0.0184082 0.023558433 s: preload index 0.086910967 0.089085967 s: refresh index 7.889590767 2.191554433 s: unpack trees 0.120760833 0.131941267 s: update worktree after a merge 2.2583504 2.572663167 s: repair cache-tree 0.8916137 0.959495233 s: write index, changed mask = 28 3.405199233 0.2710663 s: unpack trees 0.000999667 0.0021554 s: update worktree after a merge 3.4063306 0.273318333 s: diff-index 16.9524923 9.462943133 s: git command: git.exe checkout This command calls unpack_trees() twice, the first time on 2way merge and the second 1way merge. In both times, "unpack trees" time is reduced to one third. Overall time reduction is not that impressive of course because index operations take a big chunk. And there's that repair cache-tree line. PS. A note about cache-tree invalidation and the use of it in this code. We do invalidate cache-tree in _source_ index when we add new entries to the (temporary) "result" index. But we also use the cache-tree from source index in this optimization. Does this mean we end up having no cache-tree in the source index to activate this optimization? The answer is twisted: the order of finding a good cache-tree and invalidating it matters. In this case we check for a good cache-tree first in all_trees_same_as_cache_tree(), then we start to merge things and potentially invalidate that same cache-tree in the process. Since cache-tree invalidation happens after the optimization kicks in, we're still good. But we may lose that cache-tree at the very first call_unpack_fn() call in traverse_by_cache_tree(). Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-08-18 16:41:24 +02:00
int i, d;
if (!o->merge)
BUG("We need cache-tree to do this optimization");
/*
* Do what unpack_callback() and unpack_single_entry() normally
unpack-trees: optimize walking same trees with cache-tree In order to merge one or many trees with the index, unpack-trees code walks multiple trees in parallel with the index and performs n-way merge. If we find out at start of a directory that all trees are the same (by comparing OID) and cache-tree happens to be available for that directory as well, we could avoid walking the trees because we already know what these trees contain: it's flattened in what's called "the index". The upside is of course a lot less I/O since we can potentially skip lots of trees (think subtrees). We also save CPU because we don't have to inflate and apply the deltas. The downside is of course more fragile code since the logic in some functions are now duplicated elsewhere. "checkout -" with this patch on webkit.git (275k files): baseline new -------------------------------------------------------------------- 0.056651714 0.080394752 s: read cache .git/index 0.183101080 0.216010838 s: preload index 0.008584433 0.008534301 s: refresh index 0.633767589 0.251992198 s: traverse_trees 0.340265448 0.377031383 s: check_updates 0.381884638 0.372768105 s: cache_tree_update 1.401562947 1.045887251 s: unpack_trees 0.338687914 0.314983512 s: write index, changed mask = 2e 0.411927922 0.062572653 s: traverse_trees 0.000023335 0.000022544 s: check_updates 0.423697246 0.073795585 s: unpack_trees 0.423708360 0.073807557 s: diff-index 2.559524127 1.938191592 s: git command: git checkout - Another measurement from Ben's running "git checkout" with over 500k trees (on the whole series): baseline new ---------------------------------------------------------------------- 0.535510167 0.556558733 s: read cache .git/index 0.3057373 0.3147105 s: initialize name hash 0.0184082 0.023558433 s: preload index 0.086910967 0.089085967 s: refresh index 7.889590767 2.191554433 s: unpack trees 0.120760833 0.131941267 s: update worktree after a merge 2.2583504 2.572663167 s: repair cache-tree 0.8916137 0.959495233 s: write index, changed mask = 28 3.405199233 0.2710663 s: unpack trees 0.000999667 0.0021554 s: update worktree after a merge 3.4063306 0.273318333 s: diff-index 16.9524923 9.462943133 s: git command: git.exe checkout This command calls unpack_trees() twice, the first time on 2way merge and the second 1way merge. In both times, "unpack trees" time is reduced to one third. Overall time reduction is not that impressive of course because index operations take a big chunk. And there's that repair cache-tree line. PS. A note about cache-tree invalidation and the use of it in this code. We do invalidate cache-tree in _source_ index when we add new entries to the (temporary) "result" index. But we also use the cache-tree from source index in this optimization. Does this mean we end up having no cache-tree in the source index to activate this optimization? The answer is twisted: the order of finding a good cache-tree and invalidating it matters. In this case we check for a good cache-tree first in all_trees_same_as_cache_tree(), then we start to merge things and potentially invalidate that same cache-tree in the process. Since cache-tree invalidation happens after the optimization kicks in, we're still good. But we may lose that cache-tree at the very first call_unpack_fn() call in traverse_by_cache_tree(). Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-08-18 16:41:24 +02:00
* do. But we walk all paths in an iterative loop instead.
*
* D/F conflicts and higher stage entries are not a concern
* because cache-tree would be invalidated and we would never
* get here in the first place.
*/
for (i = 0; i < nr_entries; i++) {
int new_ce_len, len, rc;
unpack-trees: optimize walking same trees with cache-tree In order to merge one or many trees with the index, unpack-trees code walks multiple trees in parallel with the index and performs n-way merge. If we find out at start of a directory that all trees are the same (by comparing OID) and cache-tree happens to be available for that directory as well, we could avoid walking the trees because we already know what these trees contain: it's flattened in what's called "the index". The upside is of course a lot less I/O since we can potentially skip lots of trees (think subtrees). We also save CPU because we don't have to inflate and apply the deltas. The downside is of course more fragile code since the logic in some functions are now duplicated elsewhere. "checkout -" with this patch on webkit.git (275k files): baseline new -------------------------------------------------------------------- 0.056651714 0.080394752 s: read cache .git/index 0.183101080 0.216010838 s: preload index 0.008584433 0.008534301 s: refresh index 0.633767589 0.251992198 s: traverse_trees 0.340265448 0.377031383 s: check_updates 0.381884638 0.372768105 s: cache_tree_update 1.401562947 1.045887251 s: unpack_trees 0.338687914 0.314983512 s: write index, changed mask = 2e 0.411927922 0.062572653 s: traverse_trees 0.000023335 0.000022544 s: check_updates 0.423697246 0.073795585 s: unpack_trees 0.423708360 0.073807557 s: diff-index 2.559524127 1.938191592 s: git command: git checkout - Another measurement from Ben's running "git checkout" with over 500k trees (on the whole series): baseline new ---------------------------------------------------------------------- 0.535510167 0.556558733 s: read cache .git/index 0.3057373 0.3147105 s: initialize name hash 0.0184082 0.023558433 s: preload index 0.086910967 0.089085967 s: refresh index 7.889590767 2.191554433 s: unpack trees 0.120760833 0.131941267 s: update worktree after a merge 2.2583504 2.572663167 s: repair cache-tree 0.8916137 0.959495233 s: write index, changed mask = 28 3.405199233 0.2710663 s: unpack trees 0.000999667 0.0021554 s: update worktree after a merge 3.4063306 0.273318333 s: diff-index 16.9524923 9.462943133 s: git command: git.exe checkout This command calls unpack_trees() twice, the first time on 2way merge and the second 1way merge. In both times, "unpack trees" time is reduced to one third. Overall time reduction is not that impressive of course because index operations take a big chunk. And there's that repair cache-tree line. PS. A note about cache-tree invalidation and the use of it in this code. We do invalidate cache-tree in _source_ index when we add new entries to the (temporary) "result" index. But we also use the cache-tree from source index in this optimization. Does this mean we end up having no cache-tree in the source index to activate this optimization? The answer is twisted: the order of finding a good cache-tree and invalidating it matters. In this case we check for a good cache-tree first in all_trees_same_as_cache_tree(), then we start to merge things and potentially invalidate that same cache-tree in the process. Since cache-tree invalidation happens after the optimization kicks in, we're still good. But we may lose that cache-tree at the very first call_unpack_fn() call in traverse_by_cache_tree(). Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-08-18 16:41:24 +02:00
src[0] = o->src_index->cache[pos + i];
len = ce_namelen(src[0]);
new_ce_len = cache_entry_size(len);
if (new_ce_len > ce_len) {
new_ce_len <<= 1;
tree_ce = xrealloc(tree_ce, new_ce_len);
memset(tree_ce, 0, new_ce_len);
ce_len = new_ce_len;
tree_ce->ce_flags = create_ce_flags(0);
for (d = 1; d <= nr_names; d++)
src[d] = tree_ce;
}
unpack-trees: optimize walking same trees with cache-tree In order to merge one or many trees with the index, unpack-trees code walks multiple trees in parallel with the index and performs n-way merge. If we find out at start of a directory that all trees are the same (by comparing OID) and cache-tree happens to be available for that directory as well, we could avoid walking the trees because we already know what these trees contain: it's flattened in what's called "the index". The upside is of course a lot less I/O since we can potentially skip lots of trees (think subtrees). We also save CPU because we don't have to inflate and apply the deltas. The downside is of course more fragile code since the logic in some functions are now duplicated elsewhere. "checkout -" with this patch on webkit.git (275k files): baseline new -------------------------------------------------------------------- 0.056651714 0.080394752 s: read cache .git/index 0.183101080 0.216010838 s: preload index 0.008584433 0.008534301 s: refresh index 0.633767589 0.251992198 s: traverse_trees 0.340265448 0.377031383 s: check_updates 0.381884638 0.372768105 s: cache_tree_update 1.401562947 1.045887251 s: unpack_trees 0.338687914 0.314983512 s: write index, changed mask = 2e 0.411927922 0.062572653 s: traverse_trees 0.000023335 0.000022544 s: check_updates 0.423697246 0.073795585 s: unpack_trees 0.423708360 0.073807557 s: diff-index 2.559524127 1.938191592 s: git command: git checkout - Another measurement from Ben's running "git checkout" with over 500k trees (on the whole series): baseline new ---------------------------------------------------------------------- 0.535510167 0.556558733 s: read cache .git/index 0.3057373 0.3147105 s: initialize name hash 0.0184082 0.023558433 s: preload index 0.086910967 0.089085967 s: refresh index 7.889590767 2.191554433 s: unpack trees 0.120760833 0.131941267 s: update worktree after a merge 2.2583504 2.572663167 s: repair cache-tree 0.8916137 0.959495233 s: write index, changed mask = 28 3.405199233 0.2710663 s: unpack trees 0.000999667 0.0021554 s: update worktree after a merge 3.4063306 0.273318333 s: diff-index 16.9524923 9.462943133 s: git command: git.exe checkout This command calls unpack_trees() twice, the first time on 2way merge and the second 1way merge. In both times, "unpack trees" time is reduced to one third. Overall time reduction is not that impressive of course because index operations take a big chunk. And there's that repair cache-tree line. PS. A note about cache-tree invalidation and the use of it in this code. We do invalidate cache-tree in _source_ index when we add new entries to the (temporary) "result" index. But we also use the cache-tree from source index in this optimization. Does this mean we end up having no cache-tree in the source index to activate this optimization? The answer is twisted: the order of finding a good cache-tree and invalidating it matters. In this case we check for a good cache-tree first in all_trees_same_as_cache_tree(), then we start to merge things and potentially invalidate that same cache-tree in the process. Since cache-tree invalidation happens after the optimization kicks in, we're still good. But we may lose that cache-tree at the very first call_unpack_fn() call in traverse_by_cache_tree(). Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-08-18 16:41:24 +02:00
tree_ce->ce_mode = src[0]->ce_mode;
tree_ce->ce_namelen = len;
oidcpy(&tree_ce->oid, &src[0]->oid);
memcpy(tree_ce->name, src[0]->name, len + 1);
rc = call_unpack_fn((const struct cache_entry * const *)src, o);
if (rc < 0) {
free(tree_ce);
unpack-trees: optimize walking same trees with cache-tree In order to merge one or many trees with the index, unpack-trees code walks multiple trees in parallel with the index and performs n-way merge. If we find out at start of a directory that all trees are the same (by comparing OID) and cache-tree happens to be available for that directory as well, we could avoid walking the trees because we already know what these trees contain: it's flattened in what's called "the index". The upside is of course a lot less I/O since we can potentially skip lots of trees (think subtrees). We also save CPU because we don't have to inflate and apply the deltas. The downside is of course more fragile code since the logic in some functions are now duplicated elsewhere. "checkout -" with this patch on webkit.git (275k files): baseline new -------------------------------------------------------------------- 0.056651714 0.080394752 s: read cache .git/index 0.183101080 0.216010838 s: preload index 0.008584433 0.008534301 s: refresh index 0.633767589 0.251992198 s: traverse_trees 0.340265448 0.377031383 s: check_updates 0.381884638 0.372768105 s: cache_tree_update 1.401562947 1.045887251 s: unpack_trees 0.338687914 0.314983512 s: write index, changed mask = 2e 0.411927922 0.062572653 s: traverse_trees 0.000023335 0.000022544 s: check_updates 0.423697246 0.073795585 s: unpack_trees 0.423708360 0.073807557 s: diff-index 2.559524127 1.938191592 s: git command: git checkout - Another measurement from Ben's running "git checkout" with over 500k trees (on the whole series): baseline new ---------------------------------------------------------------------- 0.535510167 0.556558733 s: read cache .git/index 0.3057373 0.3147105 s: initialize name hash 0.0184082 0.023558433 s: preload index 0.086910967 0.089085967 s: refresh index 7.889590767 2.191554433 s: unpack trees 0.120760833 0.131941267 s: update worktree after a merge 2.2583504 2.572663167 s: repair cache-tree 0.8916137 0.959495233 s: write index, changed mask = 28 3.405199233 0.2710663 s: unpack trees 0.000999667 0.0021554 s: update worktree after a merge 3.4063306 0.273318333 s: diff-index 16.9524923 9.462943133 s: git command: git.exe checkout This command calls unpack_trees() twice, the first time on 2way merge and the second 1way merge. In both times, "unpack trees" time is reduced to one third. Overall time reduction is not that impressive of course because index operations take a big chunk. And there's that repair cache-tree line. PS. A note about cache-tree invalidation and the use of it in this code. We do invalidate cache-tree in _source_ index when we add new entries to the (temporary) "result" index. But we also use the cache-tree from source index in this optimization. Does this mean we end up having no cache-tree in the source index to activate this optimization? The answer is twisted: the order of finding a good cache-tree and invalidating it matters. In this case we check for a good cache-tree first in all_trees_same_as_cache_tree(), then we start to merge things and potentially invalidate that same cache-tree in the process. Since cache-tree invalidation happens after the optimization kicks in, we're still good. But we may lose that cache-tree at the very first call_unpack_fn() call in traverse_by_cache_tree(). Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-08-18 16:41:24 +02:00
return rc;
}
unpack-trees: optimize walking same trees with cache-tree In order to merge one or many trees with the index, unpack-trees code walks multiple trees in parallel with the index and performs n-way merge. If we find out at start of a directory that all trees are the same (by comparing OID) and cache-tree happens to be available for that directory as well, we could avoid walking the trees because we already know what these trees contain: it's flattened in what's called "the index". The upside is of course a lot less I/O since we can potentially skip lots of trees (think subtrees). We also save CPU because we don't have to inflate and apply the deltas. The downside is of course more fragile code since the logic in some functions are now duplicated elsewhere. "checkout -" with this patch on webkit.git (275k files): baseline new -------------------------------------------------------------------- 0.056651714 0.080394752 s: read cache .git/index 0.183101080 0.216010838 s: preload index 0.008584433 0.008534301 s: refresh index 0.633767589 0.251992198 s: traverse_trees 0.340265448 0.377031383 s: check_updates 0.381884638 0.372768105 s: cache_tree_update 1.401562947 1.045887251 s: unpack_trees 0.338687914 0.314983512 s: write index, changed mask = 2e 0.411927922 0.062572653 s: traverse_trees 0.000023335 0.000022544 s: check_updates 0.423697246 0.073795585 s: unpack_trees 0.423708360 0.073807557 s: diff-index 2.559524127 1.938191592 s: git command: git checkout - Another measurement from Ben's running "git checkout" with over 500k trees (on the whole series): baseline new ---------------------------------------------------------------------- 0.535510167 0.556558733 s: read cache .git/index 0.3057373 0.3147105 s: initialize name hash 0.0184082 0.023558433 s: preload index 0.086910967 0.089085967 s: refresh index 7.889590767 2.191554433 s: unpack trees 0.120760833 0.131941267 s: update worktree after a merge 2.2583504 2.572663167 s: repair cache-tree 0.8916137 0.959495233 s: write index, changed mask = 28 3.405199233 0.2710663 s: unpack trees 0.000999667 0.0021554 s: update worktree after a merge 3.4063306 0.273318333 s: diff-index 16.9524923 9.462943133 s: git command: git.exe checkout This command calls unpack_trees() twice, the first time on 2way merge and the second 1way merge. In both times, "unpack trees" time is reduced to one third. Overall time reduction is not that impressive of course because index operations take a big chunk. And there's that repair cache-tree line. PS. A note about cache-tree invalidation and the use of it in this code. We do invalidate cache-tree in _source_ index when we add new entries to the (temporary) "result" index. But we also use the cache-tree from source index in this optimization. Does this mean we end up having no cache-tree in the source index to activate this optimization? The answer is twisted: the order of finding a good cache-tree and invalidating it matters. In this case we check for a good cache-tree first in all_trees_same_as_cache_tree(), then we start to merge things and potentially invalidate that same cache-tree in the process. Since cache-tree invalidation happens after the optimization kicks in, we're still good. But we may lose that cache-tree at the very first call_unpack_fn() call in traverse_by_cache_tree(). Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-08-18 16:41:24 +02:00
mark_ce_used(src[0], o);
}
free(tree_ce);
unpack-trees: optimize walking same trees with cache-tree In order to merge one or many trees with the index, unpack-trees code walks multiple trees in parallel with the index and performs n-way merge. If we find out at start of a directory that all trees are the same (by comparing OID) and cache-tree happens to be available for that directory as well, we could avoid walking the trees because we already know what these trees contain: it's flattened in what's called "the index". The upside is of course a lot less I/O since we can potentially skip lots of trees (think subtrees). We also save CPU because we don't have to inflate and apply the deltas. The downside is of course more fragile code since the logic in some functions are now duplicated elsewhere. "checkout -" with this patch on webkit.git (275k files): baseline new -------------------------------------------------------------------- 0.056651714 0.080394752 s: read cache .git/index 0.183101080 0.216010838 s: preload index 0.008584433 0.008534301 s: refresh index 0.633767589 0.251992198 s: traverse_trees 0.340265448 0.377031383 s: check_updates 0.381884638 0.372768105 s: cache_tree_update 1.401562947 1.045887251 s: unpack_trees 0.338687914 0.314983512 s: write index, changed mask = 2e 0.411927922 0.062572653 s: traverse_trees 0.000023335 0.000022544 s: check_updates 0.423697246 0.073795585 s: unpack_trees 0.423708360 0.073807557 s: diff-index 2.559524127 1.938191592 s: git command: git checkout - Another measurement from Ben's running "git checkout" with over 500k trees (on the whole series): baseline new ---------------------------------------------------------------------- 0.535510167 0.556558733 s: read cache .git/index 0.3057373 0.3147105 s: initialize name hash 0.0184082 0.023558433 s: preload index 0.086910967 0.089085967 s: refresh index 7.889590767 2.191554433 s: unpack trees 0.120760833 0.131941267 s: update worktree after a merge 2.2583504 2.572663167 s: repair cache-tree 0.8916137 0.959495233 s: write index, changed mask = 28 3.405199233 0.2710663 s: unpack trees 0.000999667 0.0021554 s: update worktree after a merge 3.4063306 0.273318333 s: diff-index 16.9524923 9.462943133 s: git command: git.exe checkout This command calls unpack_trees() twice, the first time on 2way merge and the second 1way merge. In both times, "unpack trees" time is reduced to one third. Overall time reduction is not that impressive of course because index operations take a big chunk. And there's that repair cache-tree line. PS. A note about cache-tree invalidation and the use of it in this code. We do invalidate cache-tree in _source_ index when we add new entries to the (temporary) "result" index. But we also use the cache-tree from source index in this optimization. Does this mean we end up having no cache-tree in the source index to activate this optimization? The answer is twisted: the order of finding a good cache-tree and invalidating it matters. In this case we check for a good cache-tree first in all_trees_same_as_cache_tree(), then we start to merge things and potentially invalidate that same cache-tree in the process. Since cache-tree invalidation happens after the optimization kicks in, we're still good. But we may lose that cache-tree at the very first call_unpack_fn() call in traverse_by_cache_tree(). Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-08-18 16:41:24 +02:00
if (o->debug_unpack)
printf("Unpacked %d entries from %s to %s using cache-tree\n",
nr_entries,
o->src_index->cache[pos]->name,
o->src_index->cache[pos + nr_entries - 1]->name);
return 0;
}
static int traverse_trees_recursive(int n, unsigned long dirmask,
unsigned long df_conflicts,
struct name_entry *names,
struct traverse_info *info)
{
struct unpack_trees_options *o = info->data;
unpack-trees.c: look ahead in the index This makes the traversal of index be in sync with the tree traversal. When unpack_callback() is fed a set of tree entries from trees, it inspects the name of the entry and checks if the an index entry with the same name could be hiding behind the current index entry, and (1) if the name appears in the index as a leaf node, it is also fed to the n_way_merge() callback function; (2) if the name is a directory in the index, i.e. there are entries in that are underneath it, then nothing is fed to the n_way_merge() callback function; (3) otherwise, if the name comes before the first eligible entry in the index, the index entry is first unpacked alone. When traverse_trees_recursive() descends into a subdirectory, the cache_bottom pointer is moved to walk index entries within that directory. All of these are omitted for diff-index, which does not even want to be fed an index entry and a tree entry with D/F conflicts. This fixes 3-way read-tree and exposes a bug in other parts of the system in t6035, test #5. The test prepares these three trees: O = HEAD^ 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x A = HEAD 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob 587be6b4c3f93f93c489c0111bba5596147a26cb a/x B = master 120000 blob a36b77384451ea1de7bd340ffca868249626bc52 a/b 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x With a clean index that matches HEAD, running git read-tree -m -u --aggressive $O $A $B now yields 120000 a36b77384451ea1de7bd340ffca868249626bc52 3 a/b 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 0 a/b-2/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 1 a/b/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 2 a/b/c/d 100644 587be6b4c3f93f93c489c0111bba5596147a26cb 0 a/x which is correct. "master" created "a/b" symlink that did not exist, and removed "a/b/c/d" while HEAD did not do touch either path. Before this series, read-tree did not notice the situation and resolved addition of "a/b" and removal of "a/b/c/d" independently. If A = HEAD had another path "a/b/c/e" added, this merge should conflict but instead it silently resolved "a/b" and then immediately overwrote it to add "a/b/c/e", which was quite bogus. Tests in t1012 start to work with this. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-09-20 09:03:39 +02:00
int i, ret, bottom;
int nr_buf = 0;
struct tree_desc t[MAX_UNPACK_TREES];
void *buf[MAX_UNPACK_TREES];
struct traverse_info newinfo;
struct name_entry *p;
unpack-trees: optimize walking same trees with cache-tree In order to merge one or many trees with the index, unpack-trees code walks multiple trees in parallel with the index and performs n-way merge. If we find out at start of a directory that all trees are the same (by comparing OID) and cache-tree happens to be available for that directory as well, we could avoid walking the trees because we already know what these trees contain: it's flattened in what's called "the index". The upside is of course a lot less I/O since we can potentially skip lots of trees (think subtrees). We also save CPU because we don't have to inflate and apply the deltas. The downside is of course more fragile code since the logic in some functions are now duplicated elsewhere. "checkout -" with this patch on webkit.git (275k files): baseline new -------------------------------------------------------------------- 0.056651714 0.080394752 s: read cache .git/index 0.183101080 0.216010838 s: preload index 0.008584433 0.008534301 s: refresh index 0.633767589 0.251992198 s: traverse_trees 0.340265448 0.377031383 s: check_updates 0.381884638 0.372768105 s: cache_tree_update 1.401562947 1.045887251 s: unpack_trees 0.338687914 0.314983512 s: write index, changed mask = 2e 0.411927922 0.062572653 s: traverse_trees 0.000023335 0.000022544 s: check_updates 0.423697246 0.073795585 s: unpack_trees 0.423708360 0.073807557 s: diff-index 2.559524127 1.938191592 s: git command: git checkout - Another measurement from Ben's running "git checkout" with over 500k trees (on the whole series): baseline new ---------------------------------------------------------------------- 0.535510167 0.556558733 s: read cache .git/index 0.3057373 0.3147105 s: initialize name hash 0.0184082 0.023558433 s: preload index 0.086910967 0.089085967 s: refresh index 7.889590767 2.191554433 s: unpack trees 0.120760833 0.131941267 s: update worktree after a merge 2.2583504 2.572663167 s: repair cache-tree 0.8916137 0.959495233 s: write index, changed mask = 28 3.405199233 0.2710663 s: unpack trees 0.000999667 0.0021554 s: update worktree after a merge 3.4063306 0.273318333 s: diff-index 16.9524923 9.462943133 s: git command: git.exe checkout This command calls unpack_trees() twice, the first time on 2way merge and the second 1way merge. In both times, "unpack trees" time is reduced to one third. Overall time reduction is not that impressive of course because index operations take a big chunk. And there's that repair cache-tree line. PS. A note about cache-tree invalidation and the use of it in this code. We do invalidate cache-tree in _source_ index when we add new entries to the (temporary) "result" index. But we also use the cache-tree from source index in this optimization. Does this mean we end up having no cache-tree in the source index to activate this optimization? The answer is twisted: the order of finding a good cache-tree and invalidating it matters. In this case we check for a good cache-tree first in all_trees_same_as_cache_tree(), then we start to merge things and potentially invalidate that same cache-tree in the process. Since cache-tree invalidation happens after the optimization kicks in, we're still good. But we may lose that cache-tree at the very first call_unpack_fn() call in traverse_by_cache_tree(). Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-08-18 16:41:24 +02:00
int nr_entries;
nr_entries = all_trees_same_as_cache_tree(n, dirmask, names, info);
if (nr_entries > 0) {
int pos = index_pos_by_traverse_info(names, info);
if (!o->merge || df_conflicts)
BUG("Wrong condition to get here buddy");
/*
* All entries up to 'pos' must have been processed
* (i.e. marked CE_UNPACKED) at this point. But to be safe,
* save and restore cache_bottom anyway to not miss
* unprocessed entries before 'pos'.
*/
bottom = o->cache_bottom;
ret = traverse_by_cache_tree(pos, nr_entries, n, info);
unpack-trees: optimize walking same trees with cache-tree In order to merge one or many trees with the index, unpack-trees code walks multiple trees in parallel with the index and performs n-way merge. If we find out at start of a directory that all trees are the same (by comparing OID) and cache-tree happens to be available for that directory as well, we could avoid walking the trees because we already know what these trees contain: it's flattened in what's called "the index". The upside is of course a lot less I/O since we can potentially skip lots of trees (think subtrees). We also save CPU because we don't have to inflate and apply the deltas. The downside is of course more fragile code since the logic in some functions are now duplicated elsewhere. "checkout -" with this patch on webkit.git (275k files): baseline new -------------------------------------------------------------------- 0.056651714 0.080394752 s: read cache .git/index 0.183101080 0.216010838 s: preload index 0.008584433 0.008534301 s: refresh index 0.633767589 0.251992198 s: traverse_trees 0.340265448 0.377031383 s: check_updates 0.381884638 0.372768105 s: cache_tree_update 1.401562947 1.045887251 s: unpack_trees 0.338687914 0.314983512 s: write index, changed mask = 2e 0.411927922 0.062572653 s: traverse_trees 0.000023335 0.000022544 s: check_updates 0.423697246 0.073795585 s: unpack_trees 0.423708360 0.073807557 s: diff-index 2.559524127 1.938191592 s: git command: git checkout - Another measurement from Ben's running "git checkout" with over 500k trees (on the whole series): baseline new ---------------------------------------------------------------------- 0.535510167 0.556558733 s: read cache .git/index 0.3057373 0.3147105 s: initialize name hash 0.0184082 0.023558433 s: preload index 0.086910967 0.089085967 s: refresh index 7.889590767 2.191554433 s: unpack trees 0.120760833 0.131941267 s: update worktree after a merge 2.2583504 2.572663167 s: repair cache-tree 0.8916137 0.959495233 s: write index, changed mask = 28 3.405199233 0.2710663 s: unpack trees 0.000999667 0.0021554 s: update worktree after a merge 3.4063306 0.273318333 s: diff-index 16.9524923 9.462943133 s: git command: git.exe checkout This command calls unpack_trees() twice, the first time on 2way merge and the second 1way merge. In both times, "unpack trees" time is reduced to one third. Overall time reduction is not that impressive of course because index operations take a big chunk. And there's that repair cache-tree line. PS. A note about cache-tree invalidation and the use of it in this code. We do invalidate cache-tree in _source_ index when we add new entries to the (temporary) "result" index. But we also use the cache-tree from source index in this optimization. Does this mean we end up having no cache-tree in the source index to activate this optimization? The answer is twisted: the order of finding a good cache-tree and invalidating it matters. In this case we check for a good cache-tree first in all_trees_same_as_cache_tree(), then we start to merge things and potentially invalidate that same cache-tree in the process. Since cache-tree invalidation happens after the optimization kicks in, we're still good. But we may lose that cache-tree at the very first call_unpack_fn() call in traverse_by_cache_tree(). Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-08-18 16:41:24 +02:00
o->cache_bottom = bottom;
return ret;
}
p = names;
while (!p->mode)
p++;
newinfo = *info;
newinfo.prev = info;
newinfo.pathspec = info->pathspec;
newinfo.name = p->path;
newinfo.namelen = p->pathlen;
newinfo.mode = p->mode;
newinfo.pathlen = st_add3(newinfo.pathlen, tree_entry_len(p), 1);
newinfo.df_conflicts |= df_conflicts;
/*
* Fetch the tree from the ODB for each peer directory in the
* n commits.
*
* For 2- and 3-way traversals, we try to avoid hitting the
* ODB twice for the same OID. This should yield a nice speed
* up in checkouts and merges when the commits are similar.
*
* We don't bother doing the full O(n^2) search for larger n,
* because wider traversals don't happen that often and we
* avoid the search setup.
*
* When 2 peer OIDs are the same, we just copy the tree
* descriptor data. This implicitly borrows the buffer
* data from the earlier cell.
*/
for (i = 0; i < n; i++, dirmask >>= 1) {
if (i > 0 && are_same_oid(&names[i], &names[i - 1]))
t[i] = t[i - 1];
else if (i > 1 && are_same_oid(&names[i], &names[i - 2]))
t[i] = t[i - 2];
else {
const struct object_id *oid = NULL;
if (dirmask & 1)
oid = &names[i].oid;
buf[nr_buf++] = fill_tree_descriptor(the_repository, t + i, oid);
}
}
unpack-trees.c: look ahead in the index This makes the traversal of index be in sync with the tree traversal. When unpack_callback() is fed a set of tree entries from trees, it inspects the name of the entry and checks if the an index entry with the same name could be hiding behind the current index entry, and (1) if the name appears in the index as a leaf node, it is also fed to the n_way_merge() callback function; (2) if the name is a directory in the index, i.e. there are entries in that are underneath it, then nothing is fed to the n_way_merge() callback function; (3) otherwise, if the name comes before the first eligible entry in the index, the index entry is first unpacked alone. When traverse_trees_recursive() descends into a subdirectory, the cache_bottom pointer is moved to walk index entries within that directory. All of these are omitted for diff-index, which does not even want to be fed an index entry and a tree entry with D/F conflicts. This fixes 3-way read-tree and exposes a bug in other parts of the system in t6035, test #5. The test prepares these three trees: O = HEAD^ 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x A = HEAD 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob 587be6b4c3f93f93c489c0111bba5596147a26cb a/x B = master 120000 blob a36b77384451ea1de7bd340ffca868249626bc52 a/b 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x With a clean index that matches HEAD, running git read-tree -m -u --aggressive $O $A $B now yields 120000 a36b77384451ea1de7bd340ffca868249626bc52 3 a/b 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 0 a/b-2/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 1 a/b/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 2 a/b/c/d 100644 587be6b4c3f93f93c489c0111bba5596147a26cb 0 a/x which is correct. "master" created "a/b" symlink that did not exist, and removed "a/b/c/d" while HEAD did not do touch either path. Before this series, read-tree did not notice the situation and resolved addition of "a/b" and removal of "a/b/c/d" independently. If A = HEAD had another path "a/b/c/e" added, this merge should conflict but instead it silently resolved "a/b" and then immediately overwrote it to add "a/b/c/e", which was quite bogus. Tests in t1012 start to work with this. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-09-20 09:03:39 +02:00
bottom = switch_cache_bottom(&newinfo);
ret = traverse_trees(o->src_index, n, t, &newinfo);
unpack-trees.c: look ahead in the index This makes the traversal of index be in sync with the tree traversal. When unpack_callback() is fed a set of tree entries from trees, it inspects the name of the entry and checks if the an index entry with the same name could be hiding behind the current index entry, and (1) if the name appears in the index as a leaf node, it is also fed to the n_way_merge() callback function; (2) if the name is a directory in the index, i.e. there are entries in that are underneath it, then nothing is fed to the n_way_merge() callback function; (3) otherwise, if the name comes before the first eligible entry in the index, the index entry is first unpacked alone. When traverse_trees_recursive() descends into a subdirectory, the cache_bottom pointer is moved to walk index entries within that directory. All of these are omitted for diff-index, which does not even want to be fed an index entry and a tree entry with D/F conflicts. This fixes 3-way read-tree and exposes a bug in other parts of the system in t6035, test #5. The test prepares these three trees: O = HEAD^ 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x A = HEAD 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob 587be6b4c3f93f93c489c0111bba5596147a26cb a/x B = master 120000 blob a36b77384451ea1de7bd340ffca868249626bc52 a/b 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x With a clean index that matches HEAD, running git read-tree -m -u --aggressive $O $A $B now yields 120000 a36b77384451ea1de7bd340ffca868249626bc52 3 a/b 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 0 a/b-2/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 1 a/b/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 2 a/b/c/d 100644 587be6b4c3f93f93c489c0111bba5596147a26cb 0 a/x which is correct. "master" created "a/b" symlink that did not exist, and removed "a/b/c/d" while HEAD did not do touch either path. Before this series, read-tree did not notice the situation and resolved addition of "a/b" and removal of "a/b/c/d" independently. If A = HEAD had another path "a/b/c/e" added, this merge should conflict but instead it silently resolved "a/b" and then immediately overwrote it to add "a/b/c/e", which was quite bogus. Tests in t1012 start to work with this. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-09-20 09:03:39 +02:00
restore_cache_bottom(&newinfo, bottom);
for (i = 0; i < nr_buf; i++)
free(buf[i]);
unpack-trees.c: look ahead in the index This makes the traversal of index be in sync with the tree traversal. When unpack_callback() is fed a set of tree entries from trees, it inspects the name of the entry and checks if the an index entry with the same name could be hiding behind the current index entry, and (1) if the name appears in the index as a leaf node, it is also fed to the n_way_merge() callback function; (2) if the name is a directory in the index, i.e. there are entries in that are underneath it, then nothing is fed to the n_way_merge() callback function; (3) otherwise, if the name comes before the first eligible entry in the index, the index entry is first unpacked alone. When traverse_trees_recursive() descends into a subdirectory, the cache_bottom pointer is moved to walk index entries within that directory. All of these are omitted for diff-index, which does not even want to be fed an index entry and a tree entry with D/F conflicts. This fixes 3-way read-tree and exposes a bug in other parts of the system in t6035, test #5. The test prepares these three trees: O = HEAD^ 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x A = HEAD 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob 587be6b4c3f93f93c489c0111bba5596147a26cb a/x B = master 120000 blob a36b77384451ea1de7bd340ffca868249626bc52 a/b 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x With a clean index that matches HEAD, running git read-tree -m -u --aggressive $O $A $B now yields 120000 a36b77384451ea1de7bd340ffca868249626bc52 3 a/b 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 0 a/b-2/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 1 a/b/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 2 a/b/c/d 100644 587be6b4c3f93f93c489c0111bba5596147a26cb 0 a/x which is correct. "master" created "a/b" symlink that did not exist, and removed "a/b/c/d" while HEAD did not do touch either path. Before this series, read-tree did not notice the situation and resolved addition of "a/b" and removal of "a/b/c/d" independently. If A = HEAD had another path "a/b/c/e" added, this merge should conflict but instead it silently resolved "a/b" and then immediately overwrote it to add "a/b/c/e", which was quite bogus. Tests in t1012 start to work with this. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-09-20 09:03:39 +02:00
return ret;
}
/*
* Compare the traverse-path to the cache entry without actually
* having to generate the textual representation of the traverse
* path.
*
* NOTE! This *only* compares up to the size of the traverse path
* itself - the caller needs to do the final check for the cache
* entry having more data at the end!
*/
static int do_compare_entry_piecewise(const struct cache_entry *ce,
const struct traverse_info *info,
const char *name, size_t namelen,
unsigned mode)
{
int pathlen, ce_len;
const char *ce_name;
if (info->prev) {
int cmp = do_compare_entry_piecewise(ce, info->prev,
info->name, info->namelen,
info->mode);
if (cmp)
return cmp;
}
pathlen = info->pathlen;
ce_len = ce_namelen(ce);
/* If ce_len < pathlen then we must have previously hit "name == directory" entry */
if (ce_len < pathlen)
return -1;
ce_len -= pathlen;
ce_name = ce->name + pathlen;
return df_name_compare(ce_name, ce_len, S_IFREG, name, namelen, mode);
}
static int do_compare_entry(const struct cache_entry *ce,
const struct traverse_info *info,
const char *name, size_t namelen,
unsigned mode)
{
int pathlen, ce_len;
const char *ce_name;
int cmp;
unsigned ce_mode;
/*
* If we have not precomputed the traverse path, it is quicker
* to avoid doing so. But if we have precomputed it,
* it is quicker to use the precomputed version.
*/
if (!info->traverse_path)
return do_compare_entry_piecewise(ce, info, name, namelen, mode);
cmp = strncmp(ce->name, info->traverse_path, info->pathlen);
if (cmp)
return cmp;
pathlen = info->pathlen;
ce_len = ce_namelen(ce);
if (ce_len < pathlen)
return -1;
ce_len -= pathlen;
ce_name = ce->name + pathlen;
ce_mode = S_ISSPARSEDIR(ce->ce_mode) ? S_IFDIR : S_IFREG;
return df_name_compare(ce_name, ce_len, ce_mode, name, namelen, mode);
}
static int compare_entry(const struct cache_entry *ce, const struct traverse_info *info, const struct name_entry *n)
{
int cmp = do_compare_entry(ce, info, n->path, n->pathlen, n->mode);
if (cmp)
return cmp;
/*
* At this point, we know that we have a prefix match. If ce
* is a sparse directory, then allow an exact match. This only
* works when the input name is a directory, since ce->name
* ends in a directory separator.
*/
if (S_ISSPARSEDIR(ce->ce_mode) &&
ce->ce_namelen == traverse_path_len(info, tree_entry_len(n)) + 1)
return 0;
/*
* Even if the beginning compared identically, the ce should
* compare as bigger than a directory leading up to it!
*/
return ce_namelen(ce) > traverse_path_len(info, tree_entry_len(n));
}
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
static int ce_in_traverse_path(const struct cache_entry *ce,
const struct traverse_info *info)
{
if (!info->prev)
return 1;
if (do_compare_entry(ce, info->prev,
info->name, info->namelen, info->mode))
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
return 0;
/*
* If ce (blob) is the same name as the path (which is a tree
* we will be descending into), it won't be inside it.
*/
return (info->pathlen < ce_namelen(ce));
}
block alloc: add lifecycle APIs for cache_entry structs It has been observed that the time spent loading an index with a large number of entries is partly dominated by malloc() calls. This change is in preparation for using memory pools to reduce the number of malloc() calls made to allocate cahce entries when loading an index. Add an API to allocate and discard cache entries, abstracting the details of managing the memory backing the cache entries. This commit does actually change how memory is managed - this will be done in a later commit in the series. This change makes the distinction between cache entries that are associated with an index and cache entries that are not associated with an index. A main use of cache entries is with an index, and we can optimize the memory management around this. We still have other cases where a cache entry is not persisted with an index, and so we need to handle the "transient" use case as well. To keep the congnitive overhead of managing the cache entries, there will only be a single discard function. This means there must be enough information kept with the cache entry so that we know how to discard them. A summary of the main functions in the API is: make_cache_entry: create cache entry for use in an index. Uses specified parameters to populate cache_entry fields. make_empty_cache_entry: Create an empty cache entry for use in an index. Returns cache entry with empty fields. make_transient_cache_entry: create cache entry that is not used in an index. Uses specified parameters to populate cache_entry fields. make_empty_transient_cache_entry: create cache entry that is not used in an index. Returns cache entry with empty fields. discard_cache_entry: A single function that knows how to discard a cache entry regardless of how it was allocated. Signed-off-by: Jameson Miller <jamill@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-07-02 21:49:31 +02:00
static struct cache_entry *create_ce_entry(const struct traverse_info *info,
const struct name_entry *n,
int stage,
struct index_state *istate,
int is_transient,
int is_sparse_directory)
{
size_t len = traverse_path_len(info, tree_entry_len(n));
size_t alloc_len = is_sparse_directory ? len + 1 : len;
block alloc: add lifecycle APIs for cache_entry structs It has been observed that the time spent loading an index with a large number of entries is partly dominated by malloc() calls. This change is in preparation for using memory pools to reduce the number of malloc() calls made to allocate cahce entries when loading an index. Add an API to allocate and discard cache entries, abstracting the details of managing the memory backing the cache entries. This commit does actually change how memory is managed - this will be done in a later commit in the series. This change makes the distinction between cache entries that are associated with an index and cache entries that are not associated with an index. A main use of cache entries is with an index, and we can optimize the memory management around this. We still have other cases where a cache entry is not persisted with an index, and so we need to handle the "transient" use case as well. To keep the congnitive overhead of managing the cache entries, there will only be a single discard function. This means there must be enough information kept with the cache entry so that we know how to discard them. A summary of the main functions in the API is: make_cache_entry: create cache entry for use in an index. Uses specified parameters to populate cache_entry fields. make_empty_cache_entry: Create an empty cache entry for use in an index. Returns cache entry with empty fields. make_transient_cache_entry: create cache entry that is not used in an index. Uses specified parameters to populate cache_entry fields. make_empty_transient_cache_entry: create cache entry that is not used in an index. Returns cache entry with empty fields. discard_cache_entry: A single function that knows how to discard a cache entry regardless of how it was allocated. Signed-off-by: Jameson Miller <jamill@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-07-02 21:49:31 +02:00
struct cache_entry *ce =
is_transient ?
make_empty_transient_cache_entry(alloc_len, NULL) :
make_empty_cache_entry(istate, alloc_len);
ce->ce_mode = create_ce_mode(n->mode);
ce->ce_flags = create_ce_flags(stage);
ce->ce_namelen = len;
oidcpy(&ce->oid, &n->oid);
/* len+1 because the cache_entry allocates space for NUL */
make_traverse_path(ce->name, len + 1, info, n->path, n->pathlen);
if (is_sparse_directory) {
ce->name[len] = '/';
ce->name[len + 1] = '\0';
ce->ce_namelen++;
ce->ce_flags |= CE_SKIP_WORKTREE;
}
return ce;
}
/*
* Determine whether the path specified by 'p' should be unpacked as a new
* sparse directory in a sparse index. A new sparse directory 'A/':
* - must be outside the sparse cone.
* - must not already be in the index (i.e., no index entry with name 'A/'
* exists).
* - must not have any child entries in the index (i.e., no index entry
* 'A/<something>' exists).
* If 'p' meets the above requirements, return 1; otherwise, return 0.
*/
static int entry_is_new_sparse_dir(const struct traverse_info *info,
const struct name_entry *p)
{
int res, pos;
struct strbuf dirpath = STRBUF_INIT;
struct unpack_trees_options *o = info->data;
if (!S_ISDIR(p->mode))
return 0;
/*
* If the path is inside the sparse cone, it can't be a sparse directory.
*/
strbuf_add(&dirpath, info->traverse_path, info->pathlen);
strbuf_add(&dirpath, p->path, p->pathlen);
strbuf_addch(&dirpath, '/');
if (path_in_cone_mode_sparse_checkout(dirpath.buf, o->src_index)) {
res = 0;
goto cleanup;
}
pos = index_name_pos_sparse(o->src_index, dirpath.buf, dirpath.len);
if (pos >= 0) {
/* Path is already in the index, not a new sparse dir */
res = 0;
goto cleanup;
}
/* Where would this sparse dir be inserted into the index? */
pos = -pos - 1;
if (pos >= o->src_index->cache_nr) {
/*
* Sparse dir would be inserted at the end of the index, so we
* know it has no child entries.
*/
res = 1;
goto cleanup;
}
/*
* If the dir has child entries in the index, the first would be at the
* position the sparse directory would be inserted. If the entry at this
* position is inside the dir, not a new sparse dir.
*/
res = strncmp(o->src_index->cache[pos]->name, dirpath.buf, dirpath.len);
cleanup:
strbuf_release(&dirpath);
return res;
}
unpack-trees: optimize walking same trees with cache-tree In order to merge one or many trees with the index, unpack-trees code walks multiple trees in parallel with the index and performs n-way merge. If we find out at start of a directory that all trees are the same (by comparing OID) and cache-tree happens to be available for that directory as well, we could avoid walking the trees because we already know what these trees contain: it's flattened in what's called "the index". The upside is of course a lot less I/O since we can potentially skip lots of trees (think subtrees). We also save CPU because we don't have to inflate and apply the deltas. The downside is of course more fragile code since the logic in some functions are now duplicated elsewhere. "checkout -" with this patch on webkit.git (275k files): baseline new -------------------------------------------------------------------- 0.056651714 0.080394752 s: read cache .git/index 0.183101080 0.216010838 s: preload index 0.008584433 0.008534301 s: refresh index 0.633767589 0.251992198 s: traverse_trees 0.340265448 0.377031383 s: check_updates 0.381884638 0.372768105 s: cache_tree_update 1.401562947 1.045887251 s: unpack_trees 0.338687914 0.314983512 s: write index, changed mask = 2e 0.411927922 0.062572653 s: traverse_trees 0.000023335 0.000022544 s: check_updates 0.423697246 0.073795585 s: unpack_trees 0.423708360 0.073807557 s: diff-index 2.559524127 1.938191592 s: git command: git checkout - Another measurement from Ben's running "git checkout" with over 500k trees (on the whole series): baseline new ---------------------------------------------------------------------- 0.535510167 0.556558733 s: read cache .git/index 0.3057373 0.3147105 s: initialize name hash 0.0184082 0.023558433 s: preload index 0.086910967 0.089085967 s: refresh index 7.889590767 2.191554433 s: unpack trees 0.120760833 0.131941267 s: update worktree after a merge 2.2583504 2.572663167 s: repair cache-tree 0.8916137 0.959495233 s: write index, changed mask = 28 3.405199233 0.2710663 s: unpack trees 0.000999667 0.0021554 s: update worktree after a merge 3.4063306 0.273318333 s: diff-index 16.9524923 9.462943133 s: git command: git.exe checkout This command calls unpack_trees() twice, the first time on 2way merge and the second 1way merge. In both times, "unpack trees" time is reduced to one third. Overall time reduction is not that impressive of course because index operations take a big chunk. And there's that repair cache-tree line. PS. A note about cache-tree invalidation and the use of it in this code. We do invalidate cache-tree in _source_ index when we add new entries to the (temporary) "result" index. But we also use the cache-tree from source index in this optimization. Does this mean we end up having no cache-tree in the source index to activate this optimization? The answer is twisted: the order of finding a good cache-tree and invalidating it matters. In this case we check for a good cache-tree first in all_trees_same_as_cache_tree(), then we start to merge things and potentially invalidate that same cache-tree in the process. Since cache-tree invalidation happens after the optimization kicks in, we're still good. But we may lose that cache-tree at the very first call_unpack_fn() call in traverse_by_cache_tree(). Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-08-18 16:41:24 +02:00
/*
* Note that traverse_by_cache_tree() duplicates some logic in this function
* without actually calling it. If you change the logic here you may need to
* check and change there as well.
*/
static int unpack_single_entry(int n, unsigned long mask,
unsigned long dirmask,
struct cache_entry **src,
const struct name_entry *names,
const struct traverse_info *info,
int *is_new_sparse_dir)
{
int i;
struct unpack_trees_options *o = info->data;
unsigned long conflicts = info->df_conflicts | dirmask;
const struct name_entry *p = names;
*is_new_sparse_dir = 0;
if (mask == dirmask && !src[0]) {
/*
* If we're not in a sparse index, we can't unpack a directory
* without recursing into it, so we return.
*/
if (!o->src_index->sparse_index)
return 0;
/* Find first entry with a real name (we could use "mask" too) */
while (!p->mode)
p++;
/*
* If the directory is completely missing from the index but
* would otherwise be a sparse directory, we should unpack it.
* If not, we'll return and continue recursively traversing the
* tree.
*/
*is_new_sparse_dir = entry_is_new_sparse_dir(info, p);
if (!*is_new_sparse_dir)
return 0;
}
/*
* When we are unpacking a sparse directory, then this isn't necessarily
* a directory-file conflict.
*/
if (mask == dirmask &&
(*is_new_sparse_dir || (src[0] && S_ISSPARSEDIR(src[0]->ce_mode))))
conflicts = 0;
/*
* Ok, we've filled in up to any potential index entry in src[0],
* now do the rest.
*/
for (i = 0; i < n; i++) {
int stage;
unsigned int bit = 1ul << i;
if (conflicts & bit) {
src[i + o->merge] = o->df_conflict_entry;
continue;
}
if (!(mask & bit))
continue;
if (!o->merge)
stage = 0;
else if (i + 1 < o->head_idx)
stage = 1;
else if (i + 1 > o->head_idx)
stage = 3;
else
stage = 2;
block alloc: add lifecycle APIs for cache_entry structs It has been observed that the time spent loading an index with a large number of entries is partly dominated by malloc() calls. This change is in preparation for using memory pools to reduce the number of malloc() calls made to allocate cahce entries when loading an index. Add an API to allocate and discard cache entries, abstracting the details of managing the memory backing the cache entries. This commit does actually change how memory is managed - this will be done in a later commit in the series. This change makes the distinction between cache entries that are associated with an index and cache entries that are not associated with an index. A main use of cache entries is with an index, and we can optimize the memory management around this. We still have other cases where a cache entry is not persisted with an index, and so we need to handle the "transient" use case as well. To keep the congnitive overhead of managing the cache entries, there will only be a single discard function. This means there must be enough information kept with the cache entry so that we know how to discard them. A summary of the main functions in the API is: make_cache_entry: create cache entry for use in an index. Uses specified parameters to populate cache_entry fields. make_empty_cache_entry: Create an empty cache entry for use in an index. Returns cache entry with empty fields. make_transient_cache_entry: create cache entry that is not used in an index. Uses specified parameters to populate cache_entry fields. make_empty_transient_cache_entry: create cache entry that is not used in an index. Returns cache entry with empty fields. discard_cache_entry: A single function that knows how to discard a cache entry regardless of how it was allocated. Signed-off-by: Jameson Miller <jamill@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-07-02 21:49:31 +02:00
/*
* If the merge bit is set, then the cache entries are
* discarded in the following block. In this case,
* construct "transient" cache_entries, as they are
* not stored in the index. otherwise construct the
* cache entry from the index aware logic.
*/
src[i + o->merge] = create_ce_entry(info, names + i, stage,
&o->result, o->merge,
bit & dirmask);
}
if (o->merge) {
int rc = call_unpack_fn((const struct cache_entry * const *)src,
o);
for (i = 0; i < n; i++) {
struct cache_entry *ce = src[i + o->merge];
if (ce != o->df_conflict_entry)
block alloc: add lifecycle APIs for cache_entry structs It has been observed that the time spent loading an index with a large number of entries is partly dominated by malloc() calls. This change is in preparation for using memory pools to reduce the number of malloc() calls made to allocate cahce entries when loading an index. Add an API to allocate and discard cache entries, abstracting the details of managing the memory backing the cache entries. This commit does actually change how memory is managed - this will be done in a later commit in the series. This change makes the distinction between cache entries that are associated with an index and cache entries that are not associated with an index. A main use of cache entries is with an index, and we can optimize the memory management around this. We still have other cases where a cache entry is not persisted with an index, and so we need to handle the "transient" use case as well. To keep the congnitive overhead of managing the cache entries, there will only be a single discard function. This means there must be enough information kept with the cache entry so that we know how to discard them. A summary of the main functions in the API is: make_cache_entry: create cache entry for use in an index. Uses specified parameters to populate cache_entry fields. make_empty_cache_entry: Create an empty cache entry for use in an index. Returns cache entry with empty fields. make_transient_cache_entry: create cache entry that is not used in an index. Uses specified parameters to populate cache_entry fields. make_empty_transient_cache_entry: create cache entry that is not used in an index. Returns cache entry with empty fields. discard_cache_entry: A single function that knows how to discard a cache entry regardless of how it was allocated. Signed-off-by: Jameson Miller <jamill@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-07-02 21:49:31 +02:00
discard_cache_entry(ce);
}
return rc;
}
for (i = 0; i < n; i++)
if (src[i] && src[i] != o->df_conflict_entry)
if (do_add_entry(o, src[i], 0, 0))
return -1;
return 0;
}
static int unpack_failed(struct unpack_trees_options *o, const char *message)
{
discard_index(&o->result);
if (!o->quiet && !o->exiting_early) {
if (message)
return error("%s", message);
return -1;
}
return -1;
}
unpack-trees.c: look ahead in the index This makes the traversal of index be in sync with the tree traversal. When unpack_callback() is fed a set of tree entries from trees, it inspects the name of the entry and checks if the an index entry with the same name could be hiding behind the current index entry, and (1) if the name appears in the index as a leaf node, it is also fed to the n_way_merge() callback function; (2) if the name is a directory in the index, i.e. there are entries in that are underneath it, then nothing is fed to the n_way_merge() callback function; (3) otherwise, if the name comes before the first eligible entry in the index, the index entry is first unpacked alone. When traverse_trees_recursive() descends into a subdirectory, the cache_bottom pointer is moved to walk index entries within that directory. All of these are omitted for diff-index, which does not even want to be fed an index entry and a tree entry with D/F conflicts. This fixes 3-way read-tree and exposes a bug in other parts of the system in t6035, test #5. The test prepares these three trees: O = HEAD^ 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x A = HEAD 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob 587be6b4c3f93f93c489c0111bba5596147a26cb a/x B = master 120000 blob a36b77384451ea1de7bd340ffca868249626bc52 a/b 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x With a clean index that matches HEAD, running git read-tree -m -u --aggressive $O $A $B now yields 120000 a36b77384451ea1de7bd340ffca868249626bc52 3 a/b 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 0 a/b-2/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 1 a/b/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 2 a/b/c/d 100644 587be6b4c3f93f93c489c0111bba5596147a26cb 0 a/x which is correct. "master" created "a/b" symlink that did not exist, and removed "a/b/c/d" while HEAD did not do touch either path. Before this series, read-tree did not notice the situation and resolved addition of "a/b" and removal of "a/b/c/d" independently. If A = HEAD had another path "a/b/c/e" added, this merge should conflict but instead it silently resolved "a/b" and then immediately overwrote it to add "a/b/c/e", which was quite bogus. Tests in t1012 start to work with this. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-09-20 09:03:39 +02:00
/*
* The tree traversal is looking at name p. If we have a matching entry,
* return it. If name p is a directory in the index, do not return
* anything, as we will want to match it when the traversal descends into
* the directory.
*/
static int find_cache_pos(struct traverse_info *info,
const char *p, size_t p_len)
unpack-trees.c: look ahead in the index This makes the traversal of index be in sync with the tree traversal. When unpack_callback() is fed a set of tree entries from trees, it inspects the name of the entry and checks if the an index entry with the same name could be hiding behind the current index entry, and (1) if the name appears in the index as a leaf node, it is also fed to the n_way_merge() callback function; (2) if the name is a directory in the index, i.e. there are entries in that are underneath it, then nothing is fed to the n_way_merge() callback function; (3) otherwise, if the name comes before the first eligible entry in the index, the index entry is first unpacked alone. When traverse_trees_recursive() descends into a subdirectory, the cache_bottom pointer is moved to walk index entries within that directory. All of these are omitted for diff-index, which does not even want to be fed an index entry and a tree entry with D/F conflicts. This fixes 3-way read-tree and exposes a bug in other parts of the system in t6035, test #5. The test prepares these three trees: O = HEAD^ 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x A = HEAD 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob 587be6b4c3f93f93c489c0111bba5596147a26cb a/x B = master 120000 blob a36b77384451ea1de7bd340ffca868249626bc52 a/b 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x With a clean index that matches HEAD, running git read-tree -m -u --aggressive $O $A $B now yields 120000 a36b77384451ea1de7bd340ffca868249626bc52 3 a/b 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 0 a/b-2/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 1 a/b/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 2 a/b/c/d 100644 587be6b4c3f93f93c489c0111bba5596147a26cb 0 a/x which is correct. "master" created "a/b" symlink that did not exist, and removed "a/b/c/d" while HEAD did not do touch either path. Before this series, read-tree did not notice the situation and resolved addition of "a/b" and removal of "a/b/c/d" independently. If A = HEAD had another path "a/b/c/e" added, this merge should conflict but instead it silently resolved "a/b" and then immediately overwrote it to add "a/b/c/e", which was quite bogus. Tests in t1012 start to work with this. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-09-20 09:03:39 +02:00
{
int pos;
struct unpack_trees_options *o = info->data;
struct index_state *index = o->src_index;
int pfxlen = info->pathlen;
for (pos = o->cache_bottom; pos < index->cache_nr; pos++) {
Convert "struct cache_entry *" to "const ..." wherever possible I attempted to make index_state->cache[] a "const struct cache_entry **" to find out how existing entries in index are modified and where. The question I have is what do we do if we really need to keep track of on-disk changes in the index. The result is - diff-lib.c: setting CE_UPTODATE - name-hash.c: setting CE_HASHED - preload-index.c, read-cache.c, unpack-trees.c and builtin/update-index: obvious - entry.c: write_entry() may refresh the checked out entry via fill_stat_cache_info(). This causes "non-const struct cache_entry *" in builtin/apply.c, builtin/checkout-index.c and builtin/checkout.c - builtin/ls-files.c: --with-tree changes stagemask and may set CE_UPDATE Of these, write_entry() and its call sites are probably most interesting because it modifies on-disk info. But this is stat info and can be retrieved via refresh, at least for porcelain commands. Other just uses ce_flags for local purposes. So, keeping track of "dirty" entries is just a matter of setting a flag in index modification functions exposed by read-cache.c. Except unpack-trees, the rest of the code base does not do anything funny behind read-cache's back. The actual patch is less valueable than the summary above. But if anyone wants to re-identify the above sites. Applying this patch, then this: diff --git a/cache.h b/cache.h index 430d021..1692891 100644 --- a/cache.h +++ b/cache.h @@ -267,7 +267,7 @@ static inline unsigned int canon_mode(unsigned int mode) #define cache_entry_size(len) (offsetof(struct cache_entry,name) + (len) + 1) struct index_state { - struct cache_entry **cache; + const struct cache_entry **cache; unsigned int version; unsigned int cache_nr, cache_alloc, cache_changed; struct string_list *resolve_undo; will help quickly identify them without bogus warnings. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-07-09 17:29:00 +02:00
const struct cache_entry *ce = index->cache[pos];
unpack-trees.c: look ahead in the index This makes the traversal of index be in sync with the tree traversal. When unpack_callback() is fed a set of tree entries from trees, it inspects the name of the entry and checks if the an index entry with the same name could be hiding behind the current index entry, and (1) if the name appears in the index as a leaf node, it is also fed to the n_way_merge() callback function; (2) if the name is a directory in the index, i.e. there are entries in that are underneath it, then nothing is fed to the n_way_merge() callback function; (3) otherwise, if the name comes before the first eligible entry in the index, the index entry is first unpacked alone. When traverse_trees_recursive() descends into a subdirectory, the cache_bottom pointer is moved to walk index entries within that directory. All of these are omitted for diff-index, which does not even want to be fed an index entry and a tree entry with D/F conflicts. This fixes 3-way read-tree and exposes a bug in other parts of the system in t6035, test #5. The test prepares these three trees: O = HEAD^ 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x A = HEAD 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob 587be6b4c3f93f93c489c0111bba5596147a26cb a/x B = master 120000 blob a36b77384451ea1de7bd340ffca868249626bc52 a/b 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x With a clean index that matches HEAD, running git read-tree -m -u --aggressive $O $A $B now yields 120000 a36b77384451ea1de7bd340ffca868249626bc52 3 a/b 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 0 a/b-2/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 1 a/b/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 2 a/b/c/d 100644 587be6b4c3f93f93c489c0111bba5596147a26cb 0 a/x which is correct. "master" created "a/b" symlink that did not exist, and removed "a/b/c/d" while HEAD did not do touch either path. Before this series, read-tree did not notice the situation and resolved addition of "a/b" and removal of "a/b/c/d" independently. If A = HEAD had another path "a/b/c/e" added, this merge should conflict but instead it silently resolved "a/b" and then immediately overwrote it to add "a/b/c/e", which was quite bogus. Tests in t1012 start to work with this. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-09-20 09:03:39 +02:00
const char *ce_name, *ce_slash;
int cmp, ce_len;
unpack-trees: Make index lookahead less pessimal When traversing trees with an index, the current index pointer (o->cache_bottom) occasionally has to be temporarily advanced forwards to match the traversal order of the tree, which is not the same as the sort order of the index. The existing algorithm that did this (introduced in 730f72840cc50c523fe4cdd796ea2d2fc4571a28) would get "stuck" when the cache_bottom was popped and then repeatedly check the same index entries over and over. This represents a serious performance regression for large repositories compared to the old "broken" traversal order. This commit makes a simple change to mitigate this. Whenever find_cache_pos sees that the current pos is also the cache_bottom, and it has already been unpacked, it advances the cache_bottom as well as the current pos. This prevents the above "sticking" behavior without dramatically changing the algorithm. In addition, this commit moves the unpacked check above the ce_in_traverse_path() check. The simple bitmask check is cheaper, and in the case described above will be firing quite a bit to advance the cache_bottom after a tree pop. This yields considerable performance improvements for large trees. The following are the number of function calls for "git diff HEAD" on the Linux kernel tree, with 33,307 files: Symbol Calls Before Calls After ------------------- ------------ ----------- unpack_callback 35,332 35,332 find_cache_pos 37,357 37,357 ce_in_traverse_path 4,979,473 37,357 do_compare_entry 6,828,181 251,925 df_name_compare 6,828,181 251,925 And on a repository of 187,456 files: Symbol Calls Before Calls After ------------------- ------------ ----------- unpack_callback 197,958 197,958 find_cache_pos 208,460 208,460 ce_in_traverse_path 37,308,336 208,460 do_compare_entry 156,950,469 2,690,626 df_name_compare 156,950,469 2,690,626 On the latter repository, user time for "git diff HEAD" was reduced from 5.58 to 0.42 seconds. This is compared to 0.30 seconds before the traversal order fix was implemented. Signed-off-by: Brian Downing <bdowning@lavos.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-06-11 04:59:07 +02:00
if (ce->ce_flags & CE_UNPACKED) {
/*
* cache_bottom entry is already unpacked, so
* we can never match it; don't check it
* again.
*/
if (pos == o->cache_bottom)
++o->cache_bottom;
unpack-trees.c: look ahead in the index This makes the traversal of index be in sync with the tree traversal. When unpack_callback() is fed a set of tree entries from trees, it inspects the name of the entry and checks if the an index entry with the same name could be hiding behind the current index entry, and (1) if the name appears in the index as a leaf node, it is also fed to the n_way_merge() callback function; (2) if the name is a directory in the index, i.e. there are entries in that are underneath it, then nothing is fed to the n_way_merge() callback function; (3) otherwise, if the name comes before the first eligible entry in the index, the index entry is first unpacked alone. When traverse_trees_recursive() descends into a subdirectory, the cache_bottom pointer is moved to walk index entries within that directory. All of these are omitted for diff-index, which does not even want to be fed an index entry and a tree entry with D/F conflicts. This fixes 3-way read-tree and exposes a bug in other parts of the system in t6035, test #5. The test prepares these three trees: O = HEAD^ 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x A = HEAD 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob 587be6b4c3f93f93c489c0111bba5596147a26cb a/x B = master 120000 blob a36b77384451ea1de7bd340ffca868249626bc52 a/b 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x With a clean index that matches HEAD, running git read-tree -m -u --aggressive $O $A $B now yields 120000 a36b77384451ea1de7bd340ffca868249626bc52 3 a/b 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 0 a/b-2/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 1 a/b/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 2 a/b/c/d 100644 587be6b4c3f93f93c489c0111bba5596147a26cb 0 a/x which is correct. "master" created "a/b" symlink that did not exist, and removed "a/b/c/d" while HEAD did not do touch either path. Before this series, read-tree did not notice the situation and resolved addition of "a/b" and removal of "a/b/c/d" independently. If A = HEAD had another path "a/b/c/e" added, this merge should conflict but instead it silently resolved "a/b" and then immediately overwrote it to add "a/b/c/e", which was quite bogus. Tests in t1012 start to work with this. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-09-20 09:03:39 +02:00
continue;
unpack-trees: Make index lookahead less pessimal When traversing trees with an index, the current index pointer (o->cache_bottom) occasionally has to be temporarily advanced forwards to match the traversal order of the tree, which is not the same as the sort order of the index. The existing algorithm that did this (introduced in 730f72840cc50c523fe4cdd796ea2d2fc4571a28) would get "stuck" when the cache_bottom was popped and then repeatedly check the same index entries over and over. This represents a serious performance regression for large repositories compared to the old "broken" traversal order. This commit makes a simple change to mitigate this. Whenever find_cache_pos sees that the current pos is also the cache_bottom, and it has already been unpacked, it advances the cache_bottom as well as the current pos. This prevents the above "sticking" behavior without dramatically changing the algorithm. In addition, this commit moves the unpacked check above the ce_in_traverse_path() check. The simple bitmask check is cheaper, and in the case described above will be firing quite a bit to advance the cache_bottom after a tree pop. This yields considerable performance improvements for large trees. The following are the number of function calls for "git diff HEAD" on the Linux kernel tree, with 33,307 files: Symbol Calls Before Calls After ------------------- ------------ ----------- unpack_callback 35,332 35,332 find_cache_pos 37,357 37,357 ce_in_traverse_path 4,979,473 37,357 do_compare_entry 6,828,181 251,925 df_name_compare 6,828,181 251,925 And on a repository of 187,456 files: Symbol Calls Before Calls After ------------------- ------------ ----------- unpack_callback 197,958 197,958 find_cache_pos 208,460 208,460 ce_in_traverse_path 37,308,336 208,460 do_compare_entry 156,950,469 2,690,626 df_name_compare 156,950,469 2,690,626 On the latter repository, user time for "git diff HEAD" was reduced from 5.58 to 0.42 seconds. This is compared to 0.30 seconds before the traversal order fix was implemented. Signed-off-by: Brian Downing <bdowning@lavos.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-06-11 04:59:07 +02:00
}
if (!ce_in_traverse_path(ce, info)) {
/*
* Check if we can skip future cache checks
* (because we're already past all possible
* entries in the traverse path).
*/
if (info->traverse_path) {
if (strncmp(ce->name, info->traverse_path,
info->pathlen) > 0)
break;
}
unpack-trees.c: look ahead in the index This makes the traversal of index be in sync with the tree traversal. When unpack_callback() is fed a set of tree entries from trees, it inspects the name of the entry and checks if the an index entry with the same name could be hiding behind the current index entry, and (1) if the name appears in the index as a leaf node, it is also fed to the n_way_merge() callback function; (2) if the name is a directory in the index, i.e. there are entries in that are underneath it, then nothing is fed to the n_way_merge() callback function; (3) otherwise, if the name comes before the first eligible entry in the index, the index entry is first unpacked alone. When traverse_trees_recursive() descends into a subdirectory, the cache_bottom pointer is moved to walk index entries within that directory. All of these are omitted for diff-index, which does not even want to be fed an index entry and a tree entry with D/F conflicts. This fixes 3-way read-tree and exposes a bug in other parts of the system in t6035, test #5. The test prepares these three trees: O = HEAD^ 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x A = HEAD 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob 587be6b4c3f93f93c489c0111bba5596147a26cb a/x B = master 120000 blob a36b77384451ea1de7bd340ffca868249626bc52 a/b 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x With a clean index that matches HEAD, running git read-tree -m -u --aggressive $O $A $B now yields 120000 a36b77384451ea1de7bd340ffca868249626bc52 3 a/b 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 0 a/b-2/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 1 a/b/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 2 a/b/c/d 100644 587be6b4c3f93f93c489c0111bba5596147a26cb 0 a/x which is correct. "master" created "a/b" symlink that did not exist, and removed "a/b/c/d" while HEAD did not do touch either path. Before this series, read-tree did not notice the situation and resolved addition of "a/b" and removal of "a/b/c/d" independently. If A = HEAD had another path "a/b/c/e" added, this merge should conflict but instead it silently resolved "a/b" and then immediately overwrote it to add "a/b/c/e", which was quite bogus. Tests in t1012 start to work with this. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-09-20 09:03:39 +02:00
continue;
}
unpack-trees.c: look ahead in the index This makes the traversal of index be in sync with the tree traversal. When unpack_callback() is fed a set of tree entries from trees, it inspects the name of the entry and checks if the an index entry with the same name could be hiding behind the current index entry, and (1) if the name appears in the index as a leaf node, it is also fed to the n_way_merge() callback function; (2) if the name is a directory in the index, i.e. there are entries in that are underneath it, then nothing is fed to the n_way_merge() callback function; (3) otherwise, if the name comes before the first eligible entry in the index, the index entry is first unpacked alone. When traverse_trees_recursive() descends into a subdirectory, the cache_bottom pointer is moved to walk index entries within that directory. All of these are omitted for diff-index, which does not even want to be fed an index entry and a tree entry with D/F conflicts. This fixes 3-way read-tree and exposes a bug in other parts of the system in t6035, test #5. The test prepares these three trees: O = HEAD^ 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x A = HEAD 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob 587be6b4c3f93f93c489c0111bba5596147a26cb a/x B = master 120000 blob a36b77384451ea1de7bd340ffca868249626bc52 a/b 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x With a clean index that matches HEAD, running git read-tree -m -u --aggressive $O $A $B now yields 120000 a36b77384451ea1de7bd340ffca868249626bc52 3 a/b 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 0 a/b-2/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 1 a/b/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 2 a/b/c/d 100644 587be6b4c3f93f93c489c0111bba5596147a26cb 0 a/x which is correct. "master" created "a/b" symlink that did not exist, and removed "a/b/c/d" while HEAD did not do touch either path. Before this series, read-tree did not notice the situation and resolved addition of "a/b" and removal of "a/b/c/d" independently. If A = HEAD had another path "a/b/c/e" added, this merge should conflict but instead it silently resolved "a/b" and then immediately overwrote it to add "a/b/c/e", which was quite bogus. Tests in t1012 start to work with this. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-09-20 09:03:39 +02:00
ce_name = ce->name + pfxlen;
ce_slash = strchr(ce_name, '/');
if (ce_slash)
ce_len = ce_slash - ce_name;
else
ce_len = ce_namelen(ce) - pfxlen;
cmp = name_compare(p, p_len, ce_name, ce_len);
unpack-trees.c: look ahead in the index This makes the traversal of index be in sync with the tree traversal. When unpack_callback() is fed a set of tree entries from trees, it inspects the name of the entry and checks if the an index entry with the same name could be hiding behind the current index entry, and (1) if the name appears in the index as a leaf node, it is also fed to the n_way_merge() callback function; (2) if the name is a directory in the index, i.e. there are entries in that are underneath it, then nothing is fed to the n_way_merge() callback function; (3) otherwise, if the name comes before the first eligible entry in the index, the index entry is first unpacked alone. When traverse_trees_recursive() descends into a subdirectory, the cache_bottom pointer is moved to walk index entries within that directory. All of these are omitted for diff-index, which does not even want to be fed an index entry and a tree entry with D/F conflicts. This fixes 3-way read-tree and exposes a bug in other parts of the system in t6035, test #5. The test prepares these three trees: O = HEAD^ 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x A = HEAD 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob 587be6b4c3f93f93c489c0111bba5596147a26cb a/x B = master 120000 blob a36b77384451ea1de7bd340ffca868249626bc52 a/b 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x With a clean index that matches HEAD, running git read-tree -m -u --aggressive $O $A $B now yields 120000 a36b77384451ea1de7bd340ffca868249626bc52 3 a/b 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 0 a/b-2/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 1 a/b/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 2 a/b/c/d 100644 587be6b4c3f93f93c489c0111bba5596147a26cb 0 a/x which is correct. "master" created "a/b" symlink that did not exist, and removed "a/b/c/d" while HEAD did not do touch either path. Before this series, read-tree did not notice the situation and resolved addition of "a/b" and removal of "a/b/c/d" independently. If A = HEAD had another path "a/b/c/e" added, this merge should conflict but instead it silently resolved "a/b" and then immediately overwrote it to add "a/b/c/e", which was quite bogus. Tests in t1012 start to work with this. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-09-20 09:03:39 +02:00
/*
* Exact match; if we have a directory we need to
* delay returning it.
*/
if (!cmp)
return ce_slash ? -2 - pos : pos;
if (0 < cmp)
continue; /* keep looking */
/*
* ce_name sorts after p->path; could it be that we
* have files under p->path directory in the index?
* E.g. ce_name == "t-i", and p->path == "t"; we may
* have "t/a" in the index.
*/
if (p_len < ce_len && !memcmp(ce_name, p, p_len) &&
unpack-trees.c: look ahead in the index This makes the traversal of index be in sync with the tree traversal. When unpack_callback() is fed a set of tree entries from trees, it inspects the name of the entry and checks if the an index entry with the same name could be hiding behind the current index entry, and (1) if the name appears in the index as a leaf node, it is also fed to the n_way_merge() callback function; (2) if the name is a directory in the index, i.e. there are entries in that are underneath it, then nothing is fed to the n_way_merge() callback function; (3) otherwise, if the name comes before the first eligible entry in the index, the index entry is first unpacked alone. When traverse_trees_recursive() descends into a subdirectory, the cache_bottom pointer is moved to walk index entries within that directory. All of these are omitted for diff-index, which does not even want to be fed an index entry and a tree entry with D/F conflicts. This fixes 3-way read-tree and exposes a bug in other parts of the system in t6035, test #5. The test prepares these three trees: O = HEAD^ 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x A = HEAD 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob 587be6b4c3f93f93c489c0111bba5596147a26cb a/x B = master 120000 blob a36b77384451ea1de7bd340ffca868249626bc52 a/b 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x With a clean index that matches HEAD, running git read-tree -m -u --aggressive $O $A $B now yields 120000 a36b77384451ea1de7bd340ffca868249626bc52 3 a/b 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 0 a/b-2/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 1 a/b/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 2 a/b/c/d 100644 587be6b4c3f93f93c489c0111bba5596147a26cb 0 a/x which is correct. "master" created "a/b" symlink that did not exist, and removed "a/b/c/d" while HEAD did not do touch either path. Before this series, read-tree did not notice the situation and resolved addition of "a/b" and removal of "a/b/c/d" independently. If A = HEAD had another path "a/b/c/e" added, this merge should conflict but instead it silently resolved "a/b" and then immediately overwrote it to add "a/b/c/e", which was quite bogus. Tests in t1012 start to work with this. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-09-20 09:03:39 +02:00
ce_name[p_len] < '/')
continue; /* keep looking */
break;
}
return -1;
}
/*
* Given a sparse directory entry 'ce', compare ce->name to
unpack-trees: use traverse_path instead of name The sparse_dir_matches_path() method compares a cache entry that is a sparse directory entry against a 'struct traverse_info *info' and a 'struct name_entry *p' to see if the cache entry has exactly the right name for those other inputs. This method was introduced in 523506d (unpack-trees: unpack sparse directory entries, 2021-07-14), but included a significant mistake. The path comparisons used 'info->name' instead of 'info->traverse_path'. Since 'info->name' only stores a single tree entry name while 'info->traverse_path' stores the full path from root, this method does not work when 'info' is in a subdirectory of a directory. Replacing the right strings and their corresponding lengths make the method work properly. The previous change included a failing test that exposes this issue. That test now passes. The critical detail is that as we go deep into unpack_trees(), the logic for merging a sparse directory entry with a tree entry during 'git checkout' relies on this sparse_dir_matches_path() in order to avoid calling traverse_trees_recursive() during unpack_callback() in this hunk: if (!is_sparse_directory_entry(src[0], names, info) && traverse_trees_recursive(n, dirmask, mask & ~dirmask, names, info) < 0) { return -1; } For deep paths, the short-circuit never occurred and traverse_trees_recursive() was being called incorrectly and that was causing other strange issues. Specifically, the error message from the now-passing test previously included this: error: Your local changes to the following files would be overwritten by checkout: deep/deeper1/deepest2/a deep/deeper1/deepest3/a Please commit your changes or stash them before you switch branches. Aborting These messages occurred because the 'current' cache entry in twoway_merge() was showing as NULL because the index did not contain entries for the paths contained within the sparse directory entries. We instead had 'oldtree' given as the entry at HEAD and 'newtree' as the entry in the target tree. This led to reject_merge() listing these paths. Now that sparse_dir_matches_path() works the same for deep paths as it does for shallow depths, the rest of the logic kicks in to properly handle modifying the sparse directory entries as designed. Reported-by: Gustave Granroth <gus.gran@gmail.com> Reported-by: Mike Marcelais <michmarc@exchange.microsoft.com> Signed-off-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2021-12-06 15:10:37 +01:00
* info->traverse_path + p->path + '/' if info->traverse_path
* is non-empty.
*
* Compare ce->name to p->path + '/' otherwise. Note that
* ce->name must end in a trailing '/' because it is a sparse
* directory entry.
*/
static int sparse_dir_matches_path(const struct cache_entry *ce,
struct traverse_info *info,
const struct name_entry *p)
{
assert(S_ISSPARSEDIR(ce->ce_mode));
assert(ce->name[ce->ce_namelen - 1] == '/');
unpack-trees: use traverse_path instead of name The sparse_dir_matches_path() method compares a cache entry that is a sparse directory entry against a 'struct traverse_info *info' and a 'struct name_entry *p' to see if the cache entry has exactly the right name for those other inputs. This method was introduced in 523506d (unpack-trees: unpack sparse directory entries, 2021-07-14), but included a significant mistake. The path comparisons used 'info->name' instead of 'info->traverse_path'. Since 'info->name' only stores a single tree entry name while 'info->traverse_path' stores the full path from root, this method does not work when 'info' is in a subdirectory of a directory. Replacing the right strings and their corresponding lengths make the method work properly. The previous change included a failing test that exposes this issue. That test now passes. The critical detail is that as we go deep into unpack_trees(), the logic for merging a sparse directory entry with a tree entry during 'git checkout' relies on this sparse_dir_matches_path() in order to avoid calling traverse_trees_recursive() during unpack_callback() in this hunk: if (!is_sparse_directory_entry(src[0], names, info) && traverse_trees_recursive(n, dirmask, mask & ~dirmask, names, info) < 0) { return -1; } For deep paths, the short-circuit never occurred and traverse_trees_recursive() was being called incorrectly and that was causing other strange issues. Specifically, the error message from the now-passing test previously included this: error: Your local changes to the following files would be overwritten by checkout: deep/deeper1/deepest2/a deep/deeper1/deepest3/a Please commit your changes or stash them before you switch branches. Aborting These messages occurred because the 'current' cache entry in twoway_merge() was showing as NULL because the index did not contain entries for the paths contained within the sparse directory entries. We instead had 'oldtree' given as the entry at HEAD and 'newtree' as the entry in the target tree. This led to reject_merge() listing these paths. Now that sparse_dir_matches_path() works the same for deep paths as it does for shallow depths, the rest of the logic kicks in to properly handle modifying the sparse directory entries as designed. Reported-by: Gustave Granroth <gus.gran@gmail.com> Reported-by: Mike Marcelais <michmarc@exchange.microsoft.com> Signed-off-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2021-12-06 15:10:37 +01:00
if (info->pathlen)
return ce->ce_namelen == info->pathlen + p->pathlen + 1 &&
ce->name[info->pathlen - 1] == '/' &&
!strncmp(ce->name, info->traverse_path, info->pathlen) &&
!strncmp(ce->name + info->pathlen, p->path, p->pathlen);
return ce->ce_namelen == p->pathlen + 1 &&
!strncmp(ce->name, p->path, p->pathlen);
}
unpack-trees.c: look ahead in the index This makes the traversal of index be in sync with the tree traversal. When unpack_callback() is fed a set of tree entries from trees, it inspects the name of the entry and checks if the an index entry with the same name could be hiding behind the current index entry, and (1) if the name appears in the index as a leaf node, it is also fed to the n_way_merge() callback function; (2) if the name is a directory in the index, i.e. there are entries in that are underneath it, then nothing is fed to the n_way_merge() callback function; (3) otherwise, if the name comes before the first eligible entry in the index, the index entry is first unpacked alone. When traverse_trees_recursive() descends into a subdirectory, the cache_bottom pointer is moved to walk index entries within that directory. All of these are omitted for diff-index, which does not even want to be fed an index entry and a tree entry with D/F conflicts. This fixes 3-way read-tree and exposes a bug in other parts of the system in t6035, test #5. The test prepares these three trees: O = HEAD^ 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x A = HEAD 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob 587be6b4c3f93f93c489c0111bba5596147a26cb a/x B = master 120000 blob a36b77384451ea1de7bd340ffca868249626bc52 a/b 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x With a clean index that matches HEAD, running git read-tree -m -u --aggressive $O $A $B now yields 120000 a36b77384451ea1de7bd340ffca868249626bc52 3 a/b 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 0 a/b-2/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 1 a/b/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 2 a/b/c/d 100644 587be6b4c3f93f93c489c0111bba5596147a26cb 0 a/x which is correct. "master" created "a/b" symlink that did not exist, and removed "a/b/c/d" while HEAD did not do touch either path. Before this series, read-tree did not notice the situation and resolved addition of "a/b" and removal of "a/b/c/d" independently. If A = HEAD had another path "a/b/c/e" added, this merge should conflict but instead it silently resolved "a/b" and then immediately overwrote it to add "a/b/c/e", which was quite bogus. Tests in t1012 start to work with this. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-09-20 09:03:39 +02:00
static struct cache_entry *find_cache_entry(struct traverse_info *info,
const struct name_entry *p)
{
const char *path;
int pos = find_cache_pos(info, p->path, p->pathlen);
unpack-trees.c: look ahead in the index This makes the traversal of index be in sync with the tree traversal. When unpack_callback() is fed a set of tree entries from trees, it inspects the name of the entry and checks if the an index entry with the same name could be hiding behind the current index entry, and (1) if the name appears in the index as a leaf node, it is also fed to the n_way_merge() callback function; (2) if the name is a directory in the index, i.e. there are entries in that are underneath it, then nothing is fed to the n_way_merge() callback function; (3) otherwise, if the name comes before the first eligible entry in the index, the index entry is first unpacked alone. When traverse_trees_recursive() descends into a subdirectory, the cache_bottom pointer is moved to walk index entries within that directory. All of these are omitted for diff-index, which does not even want to be fed an index entry and a tree entry with D/F conflicts. This fixes 3-way read-tree and exposes a bug in other parts of the system in t6035, test #5. The test prepares these three trees: O = HEAD^ 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x A = HEAD 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob 587be6b4c3f93f93c489c0111bba5596147a26cb a/x B = master 120000 blob a36b77384451ea1de7bd340ffca868249626bc52 a/b 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x With a clean index that matches HEAD, running git read-tree -m -u --aggressive $O $A $B now yields 120000 a36b77384451ea1de7bd340ffca868249626bc52 3 a/b 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 0 a/b-2/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 1 a/b/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 2 a/b/c/d 100644 587be6b4c3f93f93c489c0111bba5596147a26cb 0 a/x which is correct. "master" created "a/b" symlink that did not exist, and removed "a/b/c/d" while HEAD did not do touch either path. Before this series, read-tree did not notice the situation and resolved addition of "a/b" and removal of "a/b/c/d" independently. If A = HEAD had another path "a/b/c/e" added, this merge should conflict but instead it silently resolved "a/b" and then immediately overwrote it to add "a/b/c/e", which was quite bogus. Tests in t1012 start to work with this. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-09-20 09:03:39 +02:00
struct unpack_trees_options *o = info->data;
if (0 <= pos)
return o->src_index->cache[pos];
/*
* Check for a sparse-directory entry named "path/".
* Due to the input p->path not having a trailing
* slash, the negative 'pos' value overshoots the
* expected position, hence "-2" instead of "-1".
*/
pos = -pos - 2;
if (pos < 0 || pos >= o->src_index->cache_nr)
unpack-trees.c: look ahead in the index This makes the traversal of index be in sync with the tree traversal. When unpack_callback() is fed a set of tree entries from trees, it inspects the name of the entry and checks if the an index entry with the same name could be hiding behind the current index entry, and (1) if the name appears in the index as a leaf node, it is also fed to the n_way_merge() callback function; (2) if the name is a directory in the index, i.e. there are entries in that are underneath it, then nothing is fed to the n_way_merge() callback function; (3) otherwise, if the name comes before the first eligible entry in the index, the index entry is first unpacked alone. When traverse_trees_recursive() descends into a subdirectory, the cache_bottom pointer is moved to walk index entries within that directory. All of these are omitted for diff-index, which does not even want to be fed an index entry and a tree entry with D/F conflicts. This fixes 3-way read-tree and exposes a bug in other parts of the system in t6035, test #5. The test prepares these three trees: O = HEAD^ 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x A = HEAD 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob 587be6b4c3f93f93c489c0111bba5596147a26cb a/x B = master 120000 blob a36b77384451ea1de7bd340ffca868249626bc52 a/b 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x With a clean index that matches HEAD, running git read-tree -m -u --aggressive $O $A $B now yields 120000 a36b77384451ea1de7bd340ffca868249626bc52 3 a/b 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 0 a/b-2/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 1 a/b/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 2 a/b/c/d 100644 587be6b4c3f93f93c489c0111bba5596147a26cb 0 a/x which is correct. "master" created "a/b" symlink that did not exist, and removed "a/b/c/d" while HEAD did not do touch either path. Before this series, read-tree did not notice the situation and resolved addition of "a/b" and removal of "a/b/c/d" independently. If A = HEAD had another path "a/b/c/e" added, this merge should conflict but instead it silently resolved "a/b" and then immediately overwrote it to add "a/b/c/e", which was quite bogus. Tests in t1012 start to work with this. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-09-20 09:03:39 +02:00
return NULL;
/*
* Due to lexicographic sorting and sparse directory
* entries ending with a trailing slash, our path as a
* sparse directory (e.g "subdir/") and our path as a
* file (e.g. "subdir") might be separated by other
* paths (e.g. "subdir-").
*/
while (pos >= 0) {
struct cache_entry *ce = o->src_index->cache[pos];
if (!skip_prefix(ce->name, info->traverse_path, &path) ||
strncmp(path, p->path, p->pathlen) ||
path[p->pathlen] != '/')
return NULL;
if (S_ISSPARSEDIR(ce->ce_mode) &&
sparse_dir_matches_path(ce, info, p))
return ce;
pos--;
}
return NULL;
unpack-trees.c: look ahead in the index This makes the traversal of index be in sync with the tree traversal. When unpack_callback() is fed a set of tree entries from trees, it inspects the name of the entry and checks if the an index entry with the same name could be hiding behind the current index entry, and (1) if the name appears in the index as a leaf node, it is also fed to the n_way_merge() callback function; (2) if the name is a directory in the index, i.e. there are entries in that are underneath it, then nothing is fed to the n_way_merge() callback function; (3) otherwise, if the name comes before the first eligible entry in the index, the index entry is first unpacked alone. When traverse_trees_recursive() descends into a subdirectory, the cache_bottom pointer is moved to walk index entries within that directory. All of these are omitted for diff-index, which does not even want to be fed an index entry and a tree entry with D/F conflicts. This fixes 3-way read-tree and exposes a bug in other parts of the system in t6035, test #5. The test prepares these three trees: O = HEAD^ 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x A = HEAD 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob 587be6b4c3f93f93c489c0111bba5596147a26cb a/x B = master 120000 blob a36b77384451ea1de7bd340ffca868249626bc52 a/b 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x With a clean index that matches HEAD, running git read-tree -m -u --aggressive $O $A $B now yields 120000 a36b77384451ea1de7bd340ffca868249626bc52 3 a/b 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 0 a/b-2/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 1 a/b/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 2 a/b/c/d 100644 587be6b4c3f93f93c489c0111bba5596147a26cb 0 a/x which is correct. "master" created "a/b" symlink that did not exist, and removed "a/b/c/d" while HEAD did not do touch either path. Before this series, read-tree did not notice the situation and resolved addition of "a/b" and removal of "a/b/c/d" independently. If A = HEAD had another path "a/b/c/e" added, this merge should conflict but instead it silently resolved "a/b" and then immediately overwrote it to add "a/b/c/e", which was quite bogus. Tests in t1012 start to work with this. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-09-20 09:03:39 +02:00
}
static void debug_path(struct traverse_info *info)
{
if (info->prev) {
debug_path(info->prev);
if (*info->prev->name)
putchar('/');
}
printf("%s", info->name);
}
static void debug_name_entry(int i, struct name_entry *n)
{
printf("ent#%d %06o %s\n", i,
n->path ? n->mode : 0,
n->path ? n->path : "(missing)");
}
static void debug_unpack_callback(int n,
unsigned long mask,
unsigned long dirmask,
struct name_entry *names,
struct traverse_info *info)
{
int i;
printf("* unpack mask %lu, dirmask %lu, cnt %d ",
mask, dirmask, n);
debug_path(info);
putchar('\n');
for (i = 0; i < n; i++)
debug_name_entry(i, names + i);
}
/*
* Returns true if and only if the given cache_entry is a
* sparse-directory entry that matches the given name_entry
* from the tree walk at the given traverse_info.
*/
static int is_sparse_directory_entry(struct cache_entry *ce,
const struct name_entry *name,
struct traverse_info *info)
{
if (!ce || !name || !S_ISSPARSEDIR(ce->ce_mode))
return 0;
return sparse_dir_matches_path(ce, info, name);
}
read-tree: make two-way merge sparse-aware Enable two-way merge with 'git read-tree' without expanding the sparse index. When in a sparse index, a two-way merge will trivially succeed as long as there are not changes to the same sparse directory in multiple trees (i.e., sparse directory-level "edit-edit" conflicts). If there are such conflicts, the merge will fail despite the possibility that individual files could merge cleanly. In order to resolve these "edit-edit" conflicts, "conflicted" sparse directories are - rather than rejected - merged by traversing their associated trees by OID. For each child of the sparse directory: 1. Files are merged as normal (see Documentation/git-read-tree.txt for details). 2. Subdirectories are treated as sparse directories and merged in 'twoway_merge'. If there are no conflicts, they are merged according to the rules in Documentation/git-read-tree.txt; otherwise, the subdirectory is recursively traversed and merged. This process allows sparse directories to be individually merged at the necessary depth *without* expanding a full index. The 't/t1092-sparse-checkout-compatibility.sh' test 'read-tree --merge with edit/edit conflicts in sparse directories' tests two-way merges with 1) changes inside sparse directories that do not conflict and 2) changes that do conflict (with the correct file(s) reported in the error message). Additionally, add two-way merge cases to 'sparse index is not expanded: read-tree' to confirm that the index is not expanded regardless of whether edit/edit conflicts are present in a sparse directory. Signed-off-by: Victoria Dye <vdye@github.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2022-03-01 21:24:30 +01:00
static int unpack_sparse_callback(int n, unsigned long mask, unsigned long dirmask, struct name_entry *names, struct traverse_info *info)
{
struct cache_entry *src[MAX_UNPACK_TREES + 1] = { NULL, };
struct unpack_trees_options *o = info->data;
int ret, is_new_sparse_dir;
read-tree: make two-way merge sparse-aware Enable two-way merge with 'git read-tree' without expanding the sparse index. When in a sparse index, a two-way merge will trivially succeed as long as there are not changes to the same sparse directory in multiple trees (i.e., sparse directory-level "edit-edit" conflicts). If there are such conflicts, the merge will fail despite the possibility that individual files could merge cleanly. In order to resolve these "edit-edit" conflicts, "conflicted" sparse directories are - rather than rejected - merged by traversing their associated trees by OID. For each child of the sparse directory: 1. Files are merged as normal (see Documentation/git-read-tree.txt for details). 2. Subdirectories are treated as sparse directories and merged in 'twoway_merge'. If there are no conflicts, they are merged according to the rules in Documentation/git-read-tree.txt; otherwise, the subdirectory is recursively traversed and merged. This process allows sparse directories to be individually merged at the necessary depth *without* expanding a full index. The 't/t1092-sparse-checkout-compatibility.sh' test 'read-tree --merge with edit/edit conflicts in sparse directories' tests two-way merges with 1) changes inside sparse directories that do not conflict and 2) changes that do conflict (with the correct file(s) reported in the error message). Additionally, add two-way merge cases to 'sparse index is not expanded: read-tree' to confirm that the index is not expanded regardless of whether edit/edit conflicts are present in a sparse directory. Signed-off-by: Victoria Dye <vdye@github.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2022-03-01 21:24:30 +01:00
assert(o->merge);
/*
* Unlike in 'unpack_callback', where src[0] is derived from the index when
* merging, src[0] is a transient cache entry derived from the first tree
* provided. Create the temporary entry as if it came from a non-sparse index.
*/
if (!is_null_oid(&names[0].oid)) {
src[0] = create_ce_entry(info, &names[0], 0,
&o->result, 1,
dirmask & (1ul << 0));
src[0]->ce_flags |= (CE_SKIP_WORKTREE | CE_NEW_SKIP_WORKTREE);
}
/*
* 'unpack_single_entry' assumes that src[0] is derived directly from
* the index, rather than from an entry in 'names'. This is *not* true when
* merging a sparse directory, in which case names[0] is the "index" source
* entry. To match the expectations of 'unpack_single_entry', shift past the
* "index" tree (i.e., names[0]) and adjust 'names', 'n', 'mask', and
* 'dirmask' accordingly.
*/
ret = unpack_single_entry(n - 1, mask >> 1, dirmask >> 1, src, names + 1, info, &is_new_sparse_dir);
read-tree: make two-way merge sparse-aware Enable two-way merge with 'git read-tree' without expanding the sparse index. When in a sparse index, a two-way merge will trivially succeed as long as there are not changes to the same sparse directory in multiple trees (i.e., sparse directory-level "edit-edit" conflicts). If there are such conflicts, the merge will fail despite the possibility that individual files could merge cleanly. In order to resolve these "edit-edit" conflicts, "conflicted" sparse directories are - rather than rejected - merged by traversing their associated trees by OID. For each child of the sparse directory: 1. Files are merged as normal (see Documentation/git-read-tree.txt for details). 2. Subdirectories are treated as sparse directories and merged in 'twoway_merge'. If there are no conflicts, they are merged according to the rules in Documentation/git-read-tree.txt; otherwise, the subdirectory is recursively traversed and merged. This process allows sparse directories to be individually merged at the necessary depth *without* expanding a full index. The 't/t1092-sparse-checkout-compatibility.sh' test 'read-tree --merge with edit/edit conflicts in sparse directories' tests two-way merges with 1) changes inside sparse directories that do not conflict and 2) changes that do conflict (with the correct file(s) reported in the error message). Additionally, add two-way merge cases to 'sparse index is not expanded: read-tree' to confirm that the index is not expanded regardless of whether edit/edit conflicts are present in a sparse directory. Signed-off-by: Victoria Dye <vdye@github.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2022-03-01 21:24:30 +01:00
if (src[0])
discard_cache_entry(src[0]);
return ret >= 0 ? mask : -1;
}
unpack-trees: optimize walking same trees with cache-tree In order to merge one or many trees with the index, unpack-trees code walks multiple trees in parallel with the index and performs n-way merge. If we find out at start of a directory that all trees are the same (by comparing OID) and cache-tree happens to be available for that directory as well, we could avoid walking the trees because we already know what these trees contain: it's flattened in what's called "the index". The upside is of course a lot less I/O since we can potentially skip lots of trees (think subtrees). We also save CPU because we don't have to inflate and apply the deltas. The downside is of course more fragile code since the logic in some functions are now duplicated elsewhere. "checkout -" with this patch on webkit.git (275k files): baseline new -------------------------------------------------------------------- 0.056651714 0.080394752 s: read cache .git/index 0.183101080 0.216010838 s: preload index 0.008584433 0.008534301 s: refresh index 0.633767589 0.251992198 s: traverse_trees 0.340265448 0.377031383 s: check_updates 0.381884638 0.372768105 s: cache_tree_update 1.401562947 1.045887251 s: unpack_trees 0.338687914 0.314983512 s: write index, changed mask = 2e 0.411927922 0.062572653 s: traverse_trees 0.000023335 0.000022544 s: check_updates 0.423697246 0.073795585 s: unpack_trees 0.423708360 0.073807557 s: diff-index 2.559524127 1.938191592 s: git command: git checkout - Another measurement from Ben's running "git checkout" with over 500k trees (on the whole series): baseline new ---------------------------------------------------------------------- 0.535510167 0.556558733 s: read cache .git/index 0.3057373 0.3147105 s: initialize name hash 0.0184082 0.023558433 s: preload index 0.086910967 0.089085967 s: refresh index 7.889590767 2.191554433 s: unpack trees 0.120760833 0.131941267 s: update worktree after a merge 2.2583504 2.572663167 s: repair cache-tree 0.8916137 0.959495233 s: write index, changed mask = 28 3.405199233 0.2710663 s: unpack trees 0.000999667 0.0021554 s: update worktree after a merge 3.4063306 0.273318333 s: diff-index 16.9524923 9.462943133 s: git command: git.exe checkout This command calls unpack_trees() twice, the first time on 2way merge and the second 1way merge. In both times, "unpack trees" time is reduced to one third. Overall time reduction is not that impressive of course because index operations take a big chunk. And there's that repair cache-tree line. PS. A note about cache-tree invalidation and the use of it in this code. We do invalidate cache-tree in _source_ index when we add new entries to the (temporary) "result" index. But we also use the cache-tree from source index in this optimization. Does this mean we end up having no cache-tree in the source index to activate this optimization? The answer is twisted: the order of finding a good cache-tree and invalidating it matters. In this case we check for a good cache-tree first in all_trees_same_as_cache_tree(), then we start to merge things and potentially invalidate that same cache-tree in the process. Since cache-tree invalidation happens after the optimization kicks in, we're still good. But we may lose that cache-tree at the very first call_unpack_fn() call in traverse_by_cache_tree(). Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-08-18 16:41:24 +02:00
/*
* Note that traverse_by_cache_tree() duplicates some logic in this function
* without actually calling it. If you change the logic here you may need to
* check and change there as well.
*/
static int unpack_callback(int n, unsigned long mask, unsigned long dirmask, struct name_entry *names, struct traverse_info *info)
{
struct cache_entry *src[MAX_UNPACK_TREES + 1] = { NULL, };
struct unpack_trees_options *o = info->data;
const struct name_entry *p = names;
int is_new_sparse_dir;
/* Find first entry with a real name (we could use "mask" too) */
while (!p->mode)
p++;
if (o->debug_unpack)
debug_unpack_callback(n, mask, dirmask, names, info);
/* Are we supposed to look at the index too? */
if (o->merge) {
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
while (1) {
int cmp;
unpack-trees.c: look ahead in the index This makes the traversal of index be in sync with the tree traversal. When unpack_callback() is fed a set of tree entries from trees, it inspects the name of the entry and checks if the an index entry with the same name could be hiding behind the current index entry, and (1) if the name appears in the index as a leaf node, it is also fed to the n_way_merge() callback function; (2) if the name is a directory in the index, i.e. there are entries in that are underneath it, then nothing is fed to the n_way_merge() callback function; (3) otherwise, if the name comes before the first eligible entry in the index, the index entry is first unpacked alone. When traverse_trees_recursive() descends into a subdirectory, the cache_bottom pointer is moved to walk index entries within that directory. All of these are omitted for diff-index, which does not even want to be fed an index entry and a tree entry with D/F conflicts. This fixes 3-way read-tree and exposes a bug in other parts of the system in t6035, test #5. The test prepares these three trees: O = HEAD^ 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x A = HEAD 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob 587be6b4c3f93f93c489c0111bba5596147a26cb a/x B = master 120000 blob a36b77384451ea1de7bd340ffca868249626bc52 a/b 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x With a clean index that matches HEAD, running git read-tree -m -u --aggressive $O $A $B now yields 120000 a36b77384451ea1de7bd340ffca868249626bc52 3 a/b 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 0 a/b-2/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 1 a/b/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 2 a/b/c/d 100644 587be6b4c3f93f93c489c0111bba5596147a26cb 0 a/x which is correct. "master" created "a/b" symlink that did not exist, and removed "a/b/c/d" while HEAD did not do touch either path. Before this series, read-tree did not notice the situation and resolved addition of "a/b" and removal of "a/b/c/d" independently. If A = HEAD had another path "a/b/c/e" added, this merge should conflict but instead it silently resolved "a/b" and then immediately overwrote it to add "a/b/c/e", which was quite bogus. Tests in t1012 start to work with this. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-09-20 09:03:39 +02:00
struct cache_entry *ce;
if (o->diff_index_cached)
ce = next_cache_entry(o);
unpack-trees.c: look ahead in the index This makes the traversal of index be in sync with the tree traversal. When unpack_callback() is fed a set of tree entries from trees, it inspects the name of the entry and checks if the an index entry with the same name could be hiding behind the current index entry, and (1) if the name appears in the index as a leaf node, it is also fed to the n_way_merge() callback function; (2) if the name is a directory in the index, i.e. there are entries in that are underneath it, then nothing is fed to the n_way_merge() callback function; (3) otherwise, if the name comes before the first eligible entry in the index, the index entry is first unpacked alone. When traverse_trees_recursive() descends into a subdirectory, the cache_bottom pointer is moved to walk index entries within that directory. All of these are omitted for diff-index, which does not even want to be fed an index entry and a tree entry with D/F conflicts. This fixes 3-way read-tree and exposes a bug in other parts of the system in t6035, test #5. The test prepares these three trees: O = HEAD^ 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x A = HEAD 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob 587be6b4c3f93f93c489c0111bba5596147a26cb a/x B = master 120000 blob a36b77384451ea1de7bd340ffca868249626bc52 a/b 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x With a clean index that matches HEAD, running git read-tree -m -u --aggressive $O $A $B now yields 120000 a36b77384451ea1de7bd340ffca868249626bc52 3 a/b 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 0 a/b-2/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 1 a/b/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 2 a/b/c/d 100644 587be6b4c3f93f93c489c0111bba5596147a26cb 0 a/x which is correct. "master" created "a/b" symlink that did not exist, and removed "a/b/c/d" while HEAD did not do touch either path. Before this series, read-tree did not notice the situation and resolved addition of "a/b" and removal of "a/b/c/d" independently. If A = HEAD had another path "a/b/c/e" added, this merge should conflict but instead it silently resolved "a/b" and then immediately overwrote it to add "a/b/c/e", which was quite bogus. Tests in t1012 start to work with this. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-09-20 09:03:39 +02:00
else
ce = find_cache_entry(info, p);
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
if (!ce)
break;
cmp = compare_entry(ce, info, p);
if (cmp < 0) {
if (unpack_index_entry(ce, o) < 0)
return unpack_failed(o, NULL);
continue;
}
if (!cmp) {
if (ce_stage(ce)) {
/*
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
* If we skip unmerged index
* entries, we'll skip this
* entry *and* the tree
* entries associated with it!
*/
if (o->skip_unmerged) {
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
add_same_unmerged(ce, o);
return mask;
}
}
src[0] = ce;
}
break;
}
}
if (unpack_single_entry(n, mask, dirmask, src, names, info, &is_new_sparse_dir))
return -1;
if (o->merge && src[0]) {
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
if (ce_stage(src[0]))
mark_ce_used_same_name(src[0], o);
else
mark_ce_used(src[0], o);
}
/* Now handle any directories.. */
if (dirmask) {
Optimize "diff-index --cached" using cache-tree When running "diff-index --cached" after making a change to only a small portion of the index, there is no point unpacking unchanged subtrees into the index recursively, only to find that all entries match anyway. Tweak unpack_trees() logic that is used to read in the tree object to catch the case where the tree entry we are looking at matches the index as a whole by looking at the cache-tree. As an exercise, after modifying a few paths in the kernel tree, here are a few numbers on my Athlon 64X2 3800+: (without patch, hot cache) $ /usr/bin/time git diff --cached --raw :100644 100644 b57e1f5... e69de29... M Makefile :100644 000000 8c86b72... 0000000... D arch/x86/Makefile :000000 100644 0000000... e69de29... A arche 0.07user 0.02system 0:00.09elapsed 102%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+0outputs (0major+9407minor)pagefaults 0swaps (with patch, hot cache) $ /usr/bin/time ../git.git/git-diff --cached --raw :100644 100644 b57e1f5... e69de29... M Makefile :100644 000000 8c86b72... 0000000... D arch/x86/Makefile :000000 100644 0000000... e69de29... A arche 0.02user 0.00system 0:00.02elapsed 103%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+0outputs (0major+2446minor)pagefaults 0swaps Cold cache numbers are very impressive, but it does not matter very much in practice: (without patch, cold cache) $ su root sh -c 'echo 3 >/proc/sys/vm/drop_caches' $ /usr/bin/time git diff --cached --raw :100644 100644 b57e1f5... e69de29... M Makefile :100644 000000 8c86b72... 0000000... D arch/x86/Makefile :000000 100644 0000000... e69de29... A arche 0.06user 0.17system 0:10.26elapsed 2%CPU (0avgtext+0avgdata 0maxresident)k 247032inputs+0outputs (1172major+8237minor)pagefaults 0swaps (with patch, cold cache) $ su root sh -c 'echo 3 >/proc/sys/vm/drop_caches' $ /usr/bin/time ../git.git/git-diff --cached --raw :100644 100644 b57e1f5... e69de29... M Makefile :100644 000000 8c86b72... 0000000... D arch/x86/Makefile :000000 100644 0000000... e69de29... A arche 0.02user 0.01system 0:01.01elapsed 3%CPU (0avgtext+0avgdata 0maxresident)k 18440inputs+0outputs (79major+2369minor)pagefaults 0swaps This of course helps "git status" as well. (without patch, hot cache) $ /usr/bin/time ../git.git/git-status >/dev/null 0.17user 0.18system 0:00.35elapsed 100%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+5336outputs (0major+10970minor)pagefaults 0swaps (with patch, hot cache) $ /usr/bin/time ../git.git/git-status >/dev/null 0.10user 0.16system 0:00.27elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+5336outputs (0major+3921minor)pagefaults 0swaps Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-05-21 00:57:22 +02:00
/* special case: "diff-index --cached" looking at a tree */
if (o->diff_index_cached &&
n == 1 && dirmask == 1 && S_ISDIR(names->mode)) {
int matches;
matches = cache_tree_matches_traversal(o->src_index->cache_tree,
names, info);
/*
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
* Everything under the name matches; skip the
* entire hierarchy. diff_index_cached codepath
* special cases D/F conflicts in such a way that
* it does not do any look-ahead, so this is safe.
Optimize "diff-index --cached" using cache-tree When running "diff-index --cached" after making a change to only a small portion of the index, there is no point unpacking unchanged subtrees into the index recursively, only to find that all entries match anyway. Tweak unpack_trees() logic that is used to read in the tree object to catch the case where the tree entry we are looking at matches the index as a whole by looking at the cache-tree. As an exercise, after modifying a few paths in the kernel tree, here are a few numbers on my Athlon 64X2 3800+: (without patch, hot cache) $ /usr/bin/time git diff --cached --raw :100644 100644 b57e1f5... e69de29... M Makefile :100644 000000 8c86b72... 0000000... D arch/x86/Makefile :000000 100644 0000000... e69de29... A arche 0.07user 0.02system 0:00.09elapsed 102%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+0outputs (0major+9407minor)pagefaults 0swaps (with patch, hot cache) $ /usr/bin/time ../git.git/git-diff --cached --raw :100644 100644 b57e1f5... e69de29... M Makefile :100644 000000 8c86b72... 0000000... D arch/x86/Makefile :000000 100644 0000000... e69de29... A arche 0.02user 0.00system 0:00.02elapsed 103%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+0outputs (0major+2446minor)pagefaults 0swaps Cold cache numbers are very impressive, but it does not matter very much in practice: (without patch, cold cache) $ su root sh -c 'echo 3 >/proc/sys/vm/drop_caches' $ /usr/bin/time git diff --cached --raw :100644 100644 b57e1f5... e69de29... M Makefile :100644 000000 8c86b72... 0000000... D arch/x86/Makefile :000000 100644 0000000... e69de29... A arche 0.06user 0.17system 0:10.26elapsed 2%CPU (0avgtext+0avgdata 0maxresident)k 247032inputs+0outputs (1172major+8237minor)pagefaults 0swaps (with patch, cold cache) $ su root sh -c 'echo 3 >/proc/sys/vm/drop_caches' $ /usr/bin/time ../git.git/git-diff --cached --raw :100644 100644 b57e1f5... e69de29... M Makefile :100644 000000 8c86b72... 0000000... D arch/x86/Makefile :000000 100644 0000000... e69de29... A arche 0.02user 0.01system 0:01.01elapsed 3%CPU (0avgtext+0avgdata 0maxresident)k 18440inputs+0outputs (79major+2369minor)pagefaults 0swaps This of course helps "git status" as well. (without patch, hot cache) $ /usr/bin/time ../git.git/git-status >/dev/null 0.17user 0.18system 0:00.35elapsed 100%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+5336outputs (0major+10970minor)pagefaults 0swaps (with patch, hot cache) $ /usr/bin/time ../git.git/git-status >/dev/null 0.10user 0.16system 0:00.27elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+5336outputs (0major+3921minor)pagefaults 0swaps Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-05-21 00:57:22 +02:00
*/
if (matches) {
unpack-trees: increment cache_bottom for sparse directories Correct tracking of the 'cache_bottom' for cases where sparse directories are present in the index. BACKGROUND ---------- The 'unpack_trees_options.cache_bottom' is a variable that tracks the in-progress "bottom" of the cache as 'unpack_trees()' iterates through the contents of the index. Most importantly, this value informs the sequential return values of 'next_cache_entry()' which, in the "diff cache" usage of 'unpack_callback()', are either unpacked as-is or are passed into the diff machinery. The 'cache_bottom' is intended to track the position of the first entry in the index that has not yet been diffed or unpacked. It is advanced in two main ways: either it is incremented when an index entry is marked as "used" (in 'mark_ce_used()'), indicating that it was unpacked or diffed, or when a directory is unpacked, in which case it is increased by an amount equaling the number of index entries inside that tree. In 17a1bb570b (unpack-trees: preserve cache_bottom, 2021-07-14), it was identified that sparse directories posed a problem to the above 'cache_bottom' advancement logic - because a sparse directory was both an index entry that could be "used" and a directory that can be unpacked, the 'cache_bottom' would be incremented too many times. To solve this problem, the 'mark_ce_used()' advancement of 'cache_bottom' was skipped for sparse directories. INCORRECT CACHE_BOTTOM TRACKING ------------------------------- Skipping the 'cache_bottom' advancement for sparse directories in 'mark_ce_used()' breaks down in two cases: 1. When the 'unpack_trees()' operation is *not* a "cache diff" (because the directory contents-based incrementing of 'cache_bottom' does not happen). 2. When a cache diff is performed with a pathspec (because 'unpack_index_entry()' will unpack a sparse directory not matched by the pathspec without performing the directory contents-based increment). The former luckily does not appear to affect 'git' behavior, likely because 'cache_bottom' is largely unused (non-"cache diff" 'unpack_trees()' uses 'find_index_entry()' - rather than 'next_cache_entry()' - to find the index entries to unpack). The latter, however, causes 'cache_bottom' to "lag behind" its intended position by an amount equal to the number of sparse directories unpacked so far with 'unpack_index_entry()'. If a repository is structured such that any sparse directories are ordered lexicographically *after* any pathspec-matching directories, though, this issue won't present any adverse behavior. This was the case with the 't1092-sparse-checkout-compatibility.sh' tests before the addition of the 'before/' sparse directory (ordered *before* the in-cone 'deep/' directory), therefore sidestepping the issue. Once the 'before/' directory was added, though, 'cache_bottom' began to lag behind its intended position, causing 'next_cache_entry()' to return index entries it had already processed and, ultimately, an incorrect diff. CORRECTING CACHE_BOTTOM ----------------------- The problems observed in 't1092' come from 'cache_bottom' lagging behind in cases where the cache tree-based advancement doesn't occur. To solve this, then, the fix in 17a1bb570b is "reversed"; rather than skipping 'cache_bottom' advancement in 'mark_ce_used()', we skip the directory contents-based advancement for sparse directories. Now, every index entry can be accounted for in 'cache_bottom': * if you're working with a single index entry, 'cache_bottom' is incremented in 'mark_ce_used()' * if you're working with a directory that contains index entries (but is not one itself), 'cache_bottom' is incremented by the number of entries in that directory. Finally, change the 'test_expect_failure' tests in 't1092' failing due to this bug back to 'test_expect_success'. Signed-off-by: Victoria Dye <vdye@github.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2022-03-17 16:55:35 +01:00
/*
* Only increment the cache_bottom if the
* directory isn't a sparse directory index
* entry (if it is, it was already incremented)
* in 'mark_ce_used()'
*/
if (!src[0] || !S_ISSPARSEDIR(src[0]->ce_mode))
o->cache_bottom += matches;
Optimize "diff-index --cached" using cache-tree When running "diff-index --cached" after making a change to only a small portion of the index, there is no point unpacking unchanged subtrees into the index recursively, only to find that all entries match anyway. Tweak unpack_trees() logic that is used to read in the tree object to catch the case where the tree entry we are looking at matches the index as a whole by looking at the cache-tree. As an exercise, after modifying a few paths in the kernel tree, here are a few numbers on my Athlon 64X2 3800+: (without patch, hot cache) $ /usr/bin/time git diff --cached --raw :100644 100644 b57e1f5... e69de29... M Makefile :100644 000000 8c86b72... 0000000... D arch/x86/Makefile :000000 100644 0000000... e69de29... A arche 0.07user 0.02system 0:00.09elapsed 102%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+0outputs (0major+9407minor)pagefaults 0swaps (with patch, hot cache) $ /usr/bin/time ../git.git/git-diff --cached --raw :100644 100644 b57e1f5... e69de29... M Makefile :100644 000000 8c86b72... 0000000... D arch/x86/Makefile :000000 100644 0000000... e69de29... A arche 0.02user 0.00system 0:00.02elapsed 103%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+0outputs (0major+2446minor)pagefaults 0swaps Cold cache numbers are very impressive, but it does not matter very much in practice: (without patch, cold cache) $ su root sh -c 'echo 3 >/proc/sys/vm/drop_caches' $ /usr/bin/time git diff --cached --raw :100644 100644 b57e1f5... e69de29... M Makefile :100644 000000 8c86b72... 0000000... D arch/x86/Makefile :000000 100644 0000000... e69de29... A arche 0.06user 0.17system 0:10.26elapsed 2%CPU (0avgtext+0avgdata 0maxresident)k 247032inputs+0outputs (1172major+8237minor)pagefaults 0swaps (with patch, cold cache) $ su root sh -c 'echo 3 >/proc/sys/vm/drop_caches' $ /usr/bin/time ../git.git/git-diff --cached --raw :100644 100644 b57e1f5... e69de29... M Makefile :100644 000000 8c86b72... 0000000... D arch/x86/Makefile :000000 100644 0000000... e69de29... A arche 0.02user 0.01system 0:01.01elapsed 3%CPU (0avgtext+0avgdata 0maxresident)k 18440inputs+0outputs (79major+2369minor)pagefaults 0swaps This of course helps "git status" as well. (without patch, hot cache) $ /usr/bin/time ../git.git/git-status >/dev/null 0.17user 0.18system 0:00.35elapsed 100%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+5336outputs (0major+10970minor)pagefaults 0swaps (with patch, hot cache) $ /usr/bin/time ../git.git/git-status >/dev/null 0.10user 0.16system 0:00.27elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+5336outputs (0major+3921minor)pagefaults 0swaps Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-05-21 00:57:22 +02:00
return mask;
}
}
if (!is_sparse_directory_entry(src[0], p, info) &&
!is_new_sparse_dir &&
traverse_trees_recursive(n, dirmask, mask & ~dirmask,
names, info) < 0) {
return -1;
}
return mask;
}
return mask;
}
static int clear_ce_flags_1(struct index_state *istate,
struct cache_entry **cache, int nr,
struct strbuf *prefix,
int select_mask, int clear_mask,
struct pattern_list *pl,
enum pattern_match_result default_match,
int progress_nr);
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
/* Whole directory matching */
static int clear_ce_flags_dir(struct index_state *istate,
struct cache_entry **cache, int nr,
struct strbuf *prefix,
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
char *basename,
int select_mask, int clear_mask,
struct pattern_list *pl,
enum pattern_match_result default_match,
int progress_nr)
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
{
struct cache_entry **cache_end;
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
int dtype = DT_DIR;
int rc;
enum pattern_match_result ret, orig_ret;
orig_ret = path_matches_pattern_list(prefix->buf, prefix->len,
basename, &dtype, pl, istate);
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
strbuf_addch(prefix, '/');
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
/* If undecided, use matching result of parent dir in defval */
if (orig_ret == UNDECIDED)
ret = default_match;
else
ret = orig_ret;
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
for (cache_end = cache; cache_end != cache + nr; cache_end++) {
struct cache_entry *ce = *cache_end;
if (strncmp(ce->name, prefix->buf, prefix->len))
break;
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
}
if (pl->use_cone_patterns && orig_ret == MATCHED_RECURSIVE) {
struct cache_entry **ce = cache;
unpack-trees: correctly compute result count The clear_ce_flags_dir() method processes the cache entries within a common directory. The returned int is the number of cache entries processed by that directory. When using the sparse-checkout feature in cone mode, we can skip the pattern matching for entries in the directories that are entirely included or entirely excluded. eb42feca (unpack-trees: hash less in cone mode, 2019-11-21) introduced this performance feature. The old mechanism relied on the counts returned by calling clear_ce_flags_1(), but the new mechanism calculated the number of rows by subtracting "cache_end" from "cache" to find the size of the range. However, the equation is wrong because it divides by sizeof(struct cache_entry *). This is not how pointer arithmetic works! A coverity build of Git for Windows in preparation for the 2.25.0 release found this issue with the warning, "Pointer differences, such as cache_end - cache, are automatically scaled down by the size (8 bytes) of the pointed-to type (struct cache_entry *). Most likely, the division by sizeof(struct cache_entry *) is extraneous and should be eliminated." This warning is correct. This leaves us with the question "how did this even work?" The problem that occurs with this incorrect pointer arithmetic is a performance-only bug, and a very slight one at that. Since the entry count returned by clear_ce_flags_dir() is reduced by a factor of 8, the loop in clear_ce_flags_1() will re-process entries from those directories. By inserting global counters into unpack-tree.c and tracing them with trace2_data_intmax() (in a private change, for testing), I was able to see count how many times the loop inside clear_ce_flags_1() processed an entry and how many times clear_ce_flags_dir() was called. Each of these are reduced by at least a factor of 8 with the current change. A factor larger than 8 happens when multiple levels of directories are repeated. Specifically, in the Linux kernel repo, the command git sparse-checkout set LICENSES restricts the working directory to only the files at root and in the LICENSES directory. Here are the measured counts: clear_ce_flags_1 loop blocks: Before: 11,520 After: 1,621 clear_ce_flags_dir calls: Before: 7,048 After: 606 While these are dramatic counts, the time spent in clear_ce_flags_1() is under one millisecond in each case, so the improvement is not measurable as an end-to-end time. Reported-by: Johannes Schindelin <Johannes.Schindelin@gmx.de> Signed-off-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-01-10 02:59:30 +01:00
rc = cache_end - cache;
while (ce < cache_end) {
(*ce)->ce_flags &= ~clear_mask;
ce++;
}
} else if (pl->use_cone_patterns && orig_ret == NOT_MATCHED) {
unpack-trees: correctly compute result count The clear_ce_flags_dir() method processes the cache entries within a common directory. The returned int is the number of cache entries processed by that directory. When using the sparse-checkout feature in cone mode, we can skip the pattern matching for entries in the directories that are entirely included or entirely excluded. eb42feca (unpack-trees: hash less in cone mode, 2019-11-21) introduced this performance feature. The old mechanism relied on the counts returned by calling clear_ce_flags_1(), but the new mechanism calculated the number of rows by subtracting "cache_end" from "cache" to find the size of the range. However, the equation is wrong because it divides by sizeof(struct cache_entry *). This is not how pointer arithmetic works! A coverity build of Git for Windows in preparation for the 2.25.0 release found this issue with the warning, "Pointer differences, such as cache_end - cache, are automatically scaled down by the size (8 bytes) of the pointed-to type (struct cache_entry *). Most likely, the division by sizeof(struct cache_entry *) is extraneous and should be eliminated." This warning is correct. This leaves us with the question "how did this even work?" The problem that occurs with this incorrect pointer arithmetic is a performance-only bug, and a very slight one at that. Since the entry count returned by clear_ce_flags_dir() is reduced by a factor of 8, the loop in clear_ce_flags_1() will re-process entries from those directories. By inserting global counters into unpack-tree.c and tracing them with trace2_data_intmax() (in a private change, for testing), I was able to see count how many times the loop inside clear_ce_flags_1() processed an entry and how many times clear_ce_flags_dir() was called. Each of these are reduced by at least a factor of 8 with the current change. A factor larger than 8 happens when multiple levels of directories are repeated. Specifically, in the Linux kernel repo, the command git sparse-checkout set LICENSES restricts the working directory to only the files at root and in the LICENSES directory. Here are the measured counts: clear_ce_flags_1 loop blocks: Before: 11,520 After: 1,621 clear_ce_flags_dir calls: Before: 7,048 After: 606 While these are dramatic counts, the time spent in clear_ce_flags_1() is under one millisecond in each case, so the improvement is not measurable as an end-to-end time. Reported-by: Johannes Schindelin <Johannes.Schindelin@gmx.de> Signed-off-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-01-10 02:59:30 +01:00
rc = cache_end - cache;
} else {
rc = clear_ce_flags_1(istate, cache, cache_end - cache,
prefix,
select_mask, clear_mask,
pl, ret,
progress_nr);
}
strbuf_setlen(prefix, prefix->len - 1);
return rc;
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
}
/*
* Traverse the index, find every entry that matches according to
* o->pl. Do "ce_flags &= ~clear_mask" on those entries. Return the
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
* number of traversed entries.
*
* If select_mask is non-zero, only entries whose ce_flags has on of
* those bits enabled are traversed.
*
* cache : pointer to an index entry
* prefix_len : an offset to its path
*
* The current path ("prefix") including the trailing '/' is
* cache[0]->name[0..(prefix_len-1)]
* Top level path has prefix_len zero.
*/
static int clear_ce_flags_1(struct index_state *istate,
struct cache_entry **cache, int nr,
struct strbuf *prefix,
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
int select_mask, int clear_mask,
struct pattern_list *pl,
enum pattern_match_result default_match,
int progress_nr)
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
{
avoid computing zero offsets from NULL pointer The Undefined Behavior Sanitizer in clang-11 seems to have learned a new trick: it complains about computing offsets from a NULL pointer, even if that offset is 0. This causes numerous test failures. For example, from t1090: unpack-trees.c:1355:41: runtime error: applying zero offset to null pointer ... not ok 6 - in partial clone, sparse checkout only fetches needed blobs The code in question looks like this: struct cache_entry **cache_end = cache + nr; ... while (cache != cache_end) and we sometimes pass in a NULL and 0 for "cache" and "nr". This is conceptually fine, as "cache_end" would be equal to "cache" in this case, and we wouldn't enter the loop at all. But computing even a zero offset violates the C standard. And given the fact that UBSan is noticing this behavior, this might be a potential problem spot if the compiler starts making unexpected assumptions based on undefined behavior. So let's just avoid it, which is pretty easy. In some cases we can just switch to iterating with a numeric index (as we do in sequencer.c here). In other cases (like the cache_end one) the use of an end pointer is more natural; we can keep that by just explicitly checking for the NULL/0 case when assigning the end pointer. Note that there are two ways you can write this latter case, checking for the pointer: cache_end = cache ? cache + nr : cache; or the size: cache_end = nr ? cache + nr : cache; For the case of a NULL/0 ptr/len combo, they are equivalent. But writing it the second way (as this patch does) has the property that if somebody were to incorrectly pass a NULL pointer with a non-zero length, we'd continue to notice and segfault, rather than silently pretending the length was zero. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-01-29 06:46:47 +01:00
struct cache_entry **cache_end = nr ? cache + nr : cache;
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
/*
* Process all entries that have the given prefix and meet
* select_mask condition
*/
while(cache != cache_end) {
struct cache_entry *ce = *cache;
const char *name, *slash;
int len, dtype;
enum pattern_match_result ret;
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
display_progress(istate->progress, progress_nr);
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
if (select_mask && !(ce->ce_flags & select_mask)) {
cache++;
progress_nr++;
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
continue;
}
if (prefix->len && strncmp(ce->name, prefix->buf, prefix->len))
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
break;
name = ce->name + prefix->len;
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
slash = strchr(name, '/');
/* If it's a directory, try whole directory match first */
if (slash) {
int processed;
len = slash - name;
strbuf_add(prefix, name, len);
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
processed = clear_ce_flags_dir(istate, cache, cache_end - cache,
prefix,
prefix->buf + prefix->len - len,
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
select_mask, clear_mask,
pl, default_match,
progress_nr);
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
/* clear_c_f_dir eats a whole dir already? */
if (processed) {
cache += processed;
progress_nr += processed;
strbuf_setlen(prefix, prefix->len - len);
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
continue;
}
strbuf_addch(prefix, '/');
processed = clear_ce_flags_1(istate, cache, cache_end - cache,
prefix,
select_mask, clear_mask, pl,
default_match, progress_nr);
cache += processed;
progress_nr += processed;
strbuf_setlen(prefix, prefix->len - len - 1);
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
continue;
}
/* Non-directory */
dtype = ce_to_dtype(ce);
ret = path_matches_pattern_list(ce->name,
ce_namelen(ce),
name, &dtype, pl, istate);
if (ret == UNDECIDED)
ret = default_match;
if (ret == MATCHED || ret == MATCHED_RECURSIVE)
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
ce->ce_flags &= ~clear_mask;
cache++;
progress_nr++;
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
}
display_progress(istate->progress, progress_nr);
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
return nr - (cache_end - cache);
}
static int clear_ce_flags(struct index_state *istate,
int select_mask, int clear_mask,
struct pattern_list *pl,
int show_progress)
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
{
static struct strbuf prefix = STRBUF_INIT;
char label[100];
int rval;
strbuf_reset(&prefix);
if (show_progress)
istate->progress = start_delayed_progress(
_("Updating index flags"),
istate->cache_nr);
xsnprintf(label, sizeof(label), "clear_ce_flags(0x%08lx,0x%08lx)",
(unsigned long)select_mask, (unsigned long)clear_mask);
trace2_region_enter("unpack_trees", label, the_repository);
rval = clear_ce_flags_1(istate,
istate->cache,
istate->cache_nr,
&prefix,
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
select_mask, clear_mask,
pl, 0, 0);
trace2_region_leave("unpack_trees", label, the_repository);
stop_progress(&istate->progress);
return rval;
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
}
/*
* Set/Clear CE_NEW_SKIP_WORKTREE according to $GIT_DIR/info/sparse-checkout
*/
static void mark_new_skip_worktree(struct pattern_list *pl,
struct index_state *istate,
int select_flag, int skip_wt_flag,
int show_progress)
{
int i;
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
/*
* 1. Pretend the narrowest worktree: only unmerged entries
* are checked out
*/
for (i = 0; i < istate->cache_nr; i++) {
struct cache_entry *ce = istate->cache[i];
if (select_flag && !(ce->ce_flags & select_flag))
continue;
if (!ce_stage(ce) && !(ce->ce_flags & CE_CONFLICTED))
ce->ce_flags |= skip_wt_flag;
else
ce->ce_flags &= ~skip_wt_flag;
}
unpack-trees: fix sparse checkout's "unable to match directories" Matching index entries against an excludes file currently has two problems. First, there's no function to do it. Code paths (like sparse checkout) that wanted to try it would iterate over index entries and for each index entry pass that path to excluded_from_list(). But that is not how excluded_from_list() works; one is supposed to feed in each ancester of a path before a given path to find out if it was excluded because of some parent or grandparent matching a bigsubdirectory/ pattern despite the path not matching any .gitignore pattern directly. Second, it's inefficient. The excludes mechanism is supposed to let us block off vast swaths of the filesystem as uninteresting; separately checking every index entry doesn't fit that model. Introduce a new function to take care of both these problems. This traverses the index in depth-first order (well, that's what order the index is in) to mark un-excluded entries. Maybe some day the in-core index format will be restructured to make this sort of operation easier. Or maybe we will want to try some binary search based thing. The interface is simple enough to allow all those things. Example: clear_ce_flags(the_index.cache, the_index.cache_nr, CE_CANDIDATE, CE_CLEARME, exclude_list); would clear the CE_CLEARME flag on all index entries with CE_CANDIDATE flag and not matched by exclude_list. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-11-26 19:17:46 +01:00
/*
* 2. Widen worktree according to sparse-checkout file.
* Matched entries will have skip_wt_flag cleared (i.e. "in")
*/
clear_ce_flags(istate, select_flag, skip_wt_flag, pl, show_progress);
}
static void populate_from_existing_patterns(struct unpack_trees_options *o,
struct pattern_list *pl)
{
if (get_sparse_checkout_patterns(pl) < 0)
o->skip_sparse_checkout = 1;
else
o->pl = pl;
}
static void update_sparsity_for_prefix(const char *prefix,
struct index_state *istate)
{
int prefix_len = strlen(prefix);
struct strbuf ce_prefix = STRBUF_INIT;
if (!istate->sparse_index)
return;
while (prefix_len > 0 && prefix[prefix_len - 1] == '/')
prefix_len--;
if (prefix_len <= 0)
BUG("Invalid prefix passed to update_sparsity_for_prefix");
strbuf_grow(&ce_prefix, prefix_len + 1);
strbuf_add(&ce_prefix, prefix, prefix_len);
strbuf_addch(&ce_prefix, '/');
/*
* If the prefix points to a sparse directory or a path inside a sparse
* directory, the index should be expanded. This is accomplished in one
* of two ways:
* - if the prefix is inside a sparse directory, it will be expanded by
* the 'ensure_full_index(...)' call in 'index_name_pos(...)'.
* - if the prefix matches an existing sparse directory entry,
* 'index_name_pos(...)' will return its index position, triggering
* the 'ensure_full_index(...)' below.
*/
if (!path_in_cone_mode_sparse_checkout(ce_prefix.buf, istate) &&
index_name_pos(istate, ce_prefix.buf, ce_prefix.len) >= 0)
ensure_full_index(istate);
strbuf_release(&ce_prefix);
}
static int verify_absent(const struct cache_entry *,
enum unpack_trees_error_types,
struct unpack_trees_options *);
/*
* N-way merge "len" trees. Returns 0 on success, -1 on failure to manipulate the
* resulting index, -2 on failure to reflect the changes to the work tree.
*
* CE_ADDED, CE_UNPACKED and CE_NEW_SKIP_WORKTREE are used internally
*/
int unpack_trees(unsigned len, struct tree_desc *t, struct unpack_trees_options *o)
{
struct repository *repo = the_repository;
int i, ret;
static struct cache_entry *dfc;
struct pattern_list pl;
int free_pattern_list = 0;
struct dir_struct dir = DIR_INIT;
2021-09-27 18:33:44 +02:00
if (o->reset == UNPACK_RESET_INVALID)
BUG("o->reset had a value of 1; should be UNPACK_TREES_*_UNTRACKED");
if (len > MAX_UNPACK_TREES)
die("unpack_trees takes at most %d trees", MAX_UNPACK_TREES);
if (o->dir)
BUG("o->dir is for internal use only");
trace_performance_enter();
trace2_region_enter("unpack_trees", "unpack_trees", the_repository);
prepare_repo_settings(repo);
if (repo->settings.command_requires_full_index) {
ensure_full_index(o->src_index);
if (o->dst_index)
ensure_full_index(o->dst_index);
}
2021-09-27 18:33:44 +02:00
if (o->reset == UNPACK_RESET_OVERWRITE_UNTRACKED &&
o->preserve_ignored)
BUG("UNPACK_RESET_OVERWRITE_UNTRACKED incompatible with preserved ignored files");
if (!o->preserve_ignored) {
o->dir = &dir;
o->dir->flags |= DIR_SHOW_IGNORED;
setup_standard_excludes(o->dir);
}
if (o->prefix)
update_sparsity_for_prefix(o->prefix, o->src_index);
if (!core_apply_sparse_checkout || !o->update)
o->skip_sparse_checkout = 1;
if (!o->skip_sparse_checkout && !o->pl) {
memset(&pl, 0, sizeof(pl));
free_pattern_list = 1;
populate_from_existing_patterns(o, &pl);
}
treewide: always have a valid "index_state.repo" member When the "repo" member was added to "the_index" in [1] the repo_read_index() was made to populate it, but the unpopulated "the_index" variable didn't get the same treatment. Let's do that in initialize_the_repository() when we set it up, and likewise for all of the current callers initialized an empty "struct index_state". This simplifies code that needs to deal with "the_index" or a custom "struct index_state", we no longer need to second-guess this part of the "index_state" deep in the stack. A recent example of such second-guessing is the "istate->repo ? istate->repo : the_repository" code in [2]. We can now simply use "istate->repo". We're doing this by making use of the INDEX_STATE_INIT() macro (and corresponding function) added in [3], which now have mandatory "repo" arguments. Because we now call index_state_init() in repository.c's initialize_the_repository() we don't need to handle the case where we have a "repo->index" whose "repo" member doesn't match the "repo" we're setting up, i.e. the "Complete the double-reference" code in repo_read_index() being altered here. That logic was originally added in [1], and was working around the lack of what we now have in initialize_the_repository(). For "fsmonitor-settings.c" we can remove the initialization of a NULL "r" argument to "the_repository". This was added back in [4], and was needed at the time for callers that would pass us the "r" from an "istate->repo". Before this change such a change to "fsmonitor-settings.c" would segfault all over the test suite (e.g. in t0002-gitfile.sh). This change has wider eventual implications for "fsmonitor-settings.c". The reason the other lazy loading behavior in it is required (starting with "if (!r->settings.fsmonitor) ..." is because of the previously passed "r" being "NULL". I have other local changes on top of this which move its configuration reading to "prepare_repo_settings()" in "repo-settings.c", as we could now start to rely on it being called for our "r". But let's leave all of that for now, and narrowly remove this particular part of the lazy-loading. 1. 1fd9ae517c4 (repository: add repo reference to index_state, 2021-01-23) 2. ee1f0c242ef (read-cache: add index.skipHash config option, 2023-01-06) 3. 2f6b1eb794e (cache API: add a "INDEX_STATE_INIT" macro/function, add release_index(), 2023-01-12) 4. 1e0ea5c4316 (fsmonitor: config settings are repository-specific, 2022-03-25) Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Acked-by: Derrick Stolee <derrickstolee@github.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2023-01-17 14:57:00 +01:00
index_state_init(&o->result, o->src_index->repo);
unpack_trees(): protect the handcrafted in-core index from read_cache() unpack_trees() rebuilds the in-core index from scratch by allocating a new structure and finishing it off by copying the built one to the final index. The resulting in-core index is Ok for most use, but read_cache() does not recognize it as such. The function is meant to be no-op if you already have loaded the index, until you call discard_cache(). This change the way read_cache() detects an already initialized in-core index, by introducing an extra bit, and marks the handcrafted in-core index as initialized, to avoid this problem. A better fix in the longer term would be to change the read_cache() API so that it will always discard and re-read from the on-disk index to avoid confusion. But there are higher level API that have relied on the current semantics, and they and their users all need to get converted, which is outside the scope of 'maint' track. An example of such a higher level API is write_cache_as_tree(), which is used by git-write-tree as well as later Porcelains like git-merge, revert and cherry-pick. In the longer term, we should remove read_cache() from there and add one to cmd_write_tree(); other callers expect that the in-core index they prepared is what gets written as a tree so no other change is necessary for this particular codepath. The original version of this patch marked the index by pointing an otherwise wasted malloc'ed memory with o->result.alloc, but this version uses Linus's idea to use a new "initialized" bit, which is conceptually much cleaner. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-08-23 21:57:30 +02:00
o->result.initialized = 1;
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
o->result.timestamp.sec = o->src_index->timestamp.sec;
o->result.timestamp.nsec = o->src_index->timestamp.nsec;
o->result.version = o->src_index->version;
if (!o->src_index->split_index) {
o->result.split_index = NULL;
} else if (o->src_index == o->dst_index) {
/*
* o->dst_index (and thus o->src_index) will be discarded
* and overwritten with o->result at the end of this function,
* so just use src_index's split_index to avoid having to
* create a new one.
*/
o->result.split_index = o->src_index->split_index;
o->result.split_index->refcount++;
} else {
o->result.split_index = init_split_index(&o->result);
}
oidcpy(&o->result.oid, &o->src_index->oid);
o->merge_size = len;
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
mark_all_ce_unused(o->src_index);
o->result.fsmonitor_last_update =
xstrdup_or_null(o->src_index->fsmonitor_last_update);
o->result.fsmonitor_has_run_once = o->src_index->fsmonitor_has_run_once;
if (!o->src_index->initialized &&
!repo->settings.command_requires_full_index &&
is_sparse_index_allowed(&o->result, 0))
o->result.sparse_index = 1;
/*
* Sparse checkout loop #1: set NEW_SKIP_WORKTREE on existing entries
*/
if (!o->skip_sparse_checkout)
mark_new_skip_worktree(o->pl, o->src_index, 0,
CE_NEW_SKIP_WORKTREE, o->verbose_update);
if (!dfc)
dfc = xcalloc(1, cache_entry_size(0));
o->df_conflict_entry = dfc;
if (len) {
const char *prefix = o->prefix ? o->prefix : "";
struct traverse_info info;
setup_traverse_info(&info, prefix);
info.fn = unpack_callback;
info.data = o;
info.show_all_errors = o->show_all_errors;
info.pathspec = o->pathspec;
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
if (o->prefix) {
/*
* Unpack existing index entries that sort before the
* prefix the tree is spliced into. Note that o->merge
* is always true in this case.
*/
while (1) {
struct cache_entry *ce = next_cache_entry(o);
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
if (!ce)
break;
if (ce_in_traverse_path(ce, &info))
break;
if (unpack_index_entry(ce, o) < 0)
goto return_failed;
}
}
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
trace_performance_enter();
trace2_region_enter("unpack_trees", "traverse_trees", the_repository);
ret = traverse_trees(o->src_index, len, t, &info);
trace2_region_leave("unpack_trees", "traverse_trees", the_repository);
trace_performance_leave("traverse_trees");
if (ret < 0)
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
goto return_failed;
}
/* Any left-over entries in the index? */
if (o->merge) {
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
while (1) {
struct cache_entry *ce = next_cache_entry(o);
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
if (!ce)
break;
if (unpack_index_entry(ce, o) < 0)
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
goto return_failed;
}
}
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
mark_all_ce_unused(o->src_index);
if (o->trivial_merges_only && o->nontrivial_merge) {
ret = unpack_failed(o, "Merge requires file-level merging");
goto done;
}
if (!o->skip_sparse_checkout) {
/*
* Sparse checkout loop #2: set NEW_SKIP_WORKTREE on entries not in loop #1
* If they will have NEW_SKIP_WORKTREE, also set CE_SKIP_WORKTREE
* so apply_sparse_checkout() won't attempt to remove it from worktree
*/
mark_new_skip_worktree(o->pl, &o->result,
CE_ADDED, CE_SKIP_WORKTREE | CE_NEW_SKIP_WORKTREE,
o->verbose_update);
ret = 0;
for (i = 0; i < o->result.cache_nr; i++) {
struct cache_entry *ce = o->result.cache[i];
/*
* Entries marked with CE_ADDED in merged_entry() do not have
* verify_absent() check (the check is effectively disabled
* because CE_NEW_SKIP_WORKTREE is set unconditionally).
*
* Do the real check now because we have had
* correct CE_NEW_SKIP_WORKTREE
*/
if (ce->ce_flags & CE_ADDED &&
unpack-trees: failure to set SKIP_WORKTREE bits always just a warning Setting and clearing of the SKIP_WORKTREE bit is not only done when users run 'sparse-checkout'; other commands such as 'checkout' also run through unpack_trees() which has logic for handling this special bit. As such, we need to consider how they handle special cases. A couple comparison points should help explain the rationale for changing how unpack_trees() handles these bits: Ignoring sparse checkouts for a moment, if you are switching branches and have dirty changes, it is only considered an error that will prevent the branch switching from being successful if the dirty file happens to be one of the paths with different contents. SKIP_WORKTREE has always been considered advisory; for example, if rebase or merge need or even want to materialize a path as part of their work, they have always been allowed to do so regardless of the SKIP_WORKTREE setting. This has been used for unmerged paths, but it was often used for paths it wasn't needed just because it made the code simpler. It was a best-effort consideration, and when it materialized paths contrary to the SKIP_WORKTREE setting, it was never required to even print a warning message. In the past if you trying to run e.g. 'git checkout' and: 1) you had a path that was materialized and had some dirty changes 2) the path was listed in $GITDIR/info/sparse-checkout 3) this path did not different between the current and target branches then despite the comparison points above, the inability to set SKIP_WORKTREE was treated as a *hard* error that would abort the checkout operation. This is completely inconsistent with how SKIP_WORKTREE is handled elsewhere, and rather annoying for users as leaving the paths materialized in the working copy (with a simple warning) should present no problem at all. Downgrade any errors from inability to toggle the SKIP_WORKTREE bit to a warning and allow the operations to continue. Reviewed-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-03-27 01:49:00 +01:00
verify_absent(ce, WARNING_SPARSE_ORPHANED_NOT_OVERWRITTEN, o))
ret = 1;
if (apply_sparse_checkout(&o->result, ce, o))
ret = 1;
}
unpack-trees: failure to set SKIP_WORKTREE bits always just a warning Setting and clearing of the SKIP_WORKTREE bit is not only done when users run 'sparse-checkout'; other commands such as 'checkout' also run through unpack_trees() which has logic for handling this special bit. As such, we need to consider how they handle special cases. A couple comparison points should help explain the rationale for changing how unpack_trees() handles these bits: Ignoring sparse checkouts for a moment, if you are switching branches and have dirty changes, it is only considered an error that will prevent the branch switching from being successful if the dirty file happens to be one of the paths with different contents. SKIP_WORKTREE has always been considered advisory; for example, if rebase or merge need or even want to materialize a path as part of their work, they have always been allowed to do so regardless of the SKIP_WORKTREE setting. This has been used for unmerged paths, but it was often used for paths it wasn't needed just because it made the code simpler. It was a best-effort consideration, and when it materialized paths contrary to the SKIP_WORKTREE setting, it was never required to even print a warning message. In the past if you trying to run e.g. 'git checkout' and: 1) you had a path that was materialized and had some dirty changes 2) the path was listed in $GITDIR/info/sparse-checkout 3) this path did not different between the current and target branches then despite the comparison points above, the inability to set SKIP_WORKTREE was treated as a *hard* error that would abort the checkout operation. This is completely inconsistent with how SKIP_WORKTREE is handled elsewhere, and rather annoying for users as leaving the paths materialized in the working copy (with a simple warning) should present no problem at all. Downgrade any errors from inability to toggle the SKIP_WORKTREE bit to a warning and allow the operations to continue. Reviewed-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-03-27 01:49:00 +01:00
if (ret == 1) {
/*
* Inability to sparsify or de-sparsify individual
* paths is not an error, but just a warning.
*/
if (o->show_all_errors)
display_warning_msgs(o);
ret = 0;
}
}
ret = check_updates(o, &o->result) ? (-2) : 0;
if (o->dst_index) {
unpack-trees: reuse (still valid) cache-tree from src_index We do n-way merge by walking the source index and n trees at the same time and add merge results to a new temporary index called o->result. The merge result for any given path could be either - keep_entry(): same old index entry in o->src_index is reused - merged_entry(): either a new entry is added, or an existing one updated - deleted_entry(): one entry from o->src_index is removed For some reason [1] we keep making sure that the source index's cache-tree is still valid if used by o->result: for all those merged/deleted entries, we invalidate the same path in o->src_index, so only cache-trees covering the "keep_entry" parts remain good. Because of this, the cache-tree from o->src_index can be perfectly reused in o->result. And in fact we already rely on this logic to reuse untracked cache in edf3b90553 (unpack-trees: preserve index extensions - 2017-05-08). Move the cache-tree to o->result before doing cache_tree_update() to reduce hashing cost. Since cache_tree_update() has risen up as one of the most expensive parts in unpack_trees() after the last few patches. This does help reduce unpack_trees() time significantly (on webkit.git): before after -------------------------------------------------------------------- 0.080394752 0.051258167 s: read cache .git/index 0.216010838 0.212106298 s: preload index 0.008534301 0.280521764 s: refresh index 0.251992198 0.218160442 s: traverse_trees 0.377031383 0.374948191 s: check_updates 0.372768105 0.037040114 s: cache_tree_update 1.045887251 0.672031609 s: unpack_trees 0.314983512 0.317456290 s: write index, changed mask = 2e 0.062572653 0.038382654 s: traverse_trees 0.000022544 0.000042731 s: check_updates 0.073795585 0.050930053 s: unpack_trees 0.073807557 0.051099735 s: diff-index 1.938191592 1.614241153 s: git command: git checkout - [1] I'm pretty sure the reason is an oversight in 34110cd4e3 (Make 'unpack_trees()' have a separate source and destination index - 2008-03-06). That patch aims to _not_ update the source index at all. The invalidation should have been done on o->result in that patch. But then there was no cache-tree on o->result even then so it's pointless to do so. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-08-18 16:41:26 +02:00
move_index_extensions(&o->result, o->src_index);
if (!ret) {
if (git_env_bool("GIT_TEST_CHECK_CACHE_TREE", 0))
cache_tree_verify(the_repository, &o->result);
if (!o->skip_cache_tree_update &&
!cache_tree_fully_valid(o->result.cache_tree))
cache_tree_update(&o->result,
WRITE_TREE_SILENT |
WRITE_TREE_REPAIR);
}
o->result.updated_workdir = 1;
discard_index(o->dst_index);
*o->dst_index = o->result;
} else {
discard_index(&o->result);
}
o->src_index = NULL;
done:
if (free_pattern_list)
clear_pattern_list(&pl);
if (o->dir) {
dir_clear(o->dir);
o->dir = NULL;
}
trace2_region_leave("unpack_trees", "unpack_trees", the_repository);
trace_performance_leave("unpack_trees");
return ret;
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
return_failed:
if (o->show_all_errors)
display_error_msgs(o);
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
mark_all_ce_unused(o->src_index);
ret = unpack_failed(o, NULL);
if (o->exiting_early)
ret = 0;
goto done;
}
unpack-trees: add a new update_sparsity() function Previously, the only way to update the SKIP_WORKTREE bits for various paths was invoking `git read-tree -mu HEAD` or calling the same code that this codepath invoked. This however had a number of problems if the index or working directory were not clean. First, let's consider the case: Flipping SKIP_WORKTREE -> !SKIP_WORKTREE (materializing files) If the working tree was clean this was fine, but if there were files or directories or symlinks or whatever already present at the given path then the operation would abort with an error. Let's label this case for later discussion: A) There is an untracked path in the way Now let's consider the opposite case: Flipping !SKIP_WORKTREE -> SKIP_WORKTREE (removing files) If the index and working tree was clean this was fine, but if there were any unclean paths we would run into problems. There are three different cases to consider: B) The path is unmerged C) The path has unstaged changes D) The path has staged changes (differs from HEAD) If any path fell into case B or C, then the whole operation would be aborted with an error. With sparse-checkout, the whole operation would be aborted for case D as well, but for its predecessor of using `git read-tree -mu HEAD` directly, any paths that fell into case D would be removed from the working copy and the index entry for that path would be reset to match HEAD -- which looks and feels like data loss to users (only a few are even aware to ask whether it can be recovered, and even then it requires walking through loose objects trying to match up the right ones). Refusing to remove files that have unsaved user changes is good, but refusing to work on any other paths is very problematic for users. If the user is in the middle of a rebase or has made modifications to files that bring in more dependencies, then for their build to work they need to update the sparse paths. This logic has been preventing them from doing so. Sometimes in response, the user will stage the files and re-try, to no avail with sparse-checkout or to the horror of losing their changes if they are using its predecessor of `git read-tree -mu HEAD`. Add a new update_sparsity() function which will not error out in any of these cases but behaves as follows for the special cases: A) Leave the file in the working copy alone, clear the SKIP_WORKTREE bit, and print a warning (thus leaving the path in a state where status will report the file as modified, which seems logical). B) Do NOT mark this path as SKIP_WORKTREE, and leave it as unmerged. C) Do NOT mark this path as SKIP_WORKTREE and print a warning about the dirty path. D) Mark the path as SKIP_WORKTREE, but do not revert the version stored in the index to match HEAD; leave the contents alone. I tried a different behavior for A (leave the SKIP_WORKTREE bit set), but found it very surprising and counter-intuitive (e.g. the user sees it is present along with all the other files in that directory, tries to stage it, but git add ignores it since the SKIP_WORKTREE bit is set). A & C seem like optimal behavior to me. B may be as well, though I wonder if printing a warning would be an improvement. Some might be slightly surprised by D at first, but given that it does the right thing with `git commit` and even `git commit -a` (`git add` ignores entries that are marked SKIP_WORKTREE and thus doesn't delete them, and `commit -a` is similar), it seems logical to me. Reviewed-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-03-27 01:48:52 +01:00
/*
* Update SKIP_WORKTREE bits according to sparsity patterns, and update
* working directory to match.
*
* CE_NEW_SKIP_WORKTREE is used internally.
*/
enum update_sparsity_result update_sparsity(struct unpack_trees_options *o)
{
enum update_sparsity_result ret = UPDATE_SPARSITY_SUCCESS;
struct pattern_list pl;
int i;
unpack-trees: add a new update_sparsity() function Previously, the only way to update the SKIP_WORKTREE bits for various paths was invoking `git read-tree -mu HEAD` or calling the same code that this codepath invoked. This however had a number of problems if the index or working directory were not clean. First, let's consider the case: Flipping SKIP_WORKTREE -> !SKIP_WORKTREE (materializing files) If the working tree was clean this was fine, but if there were files or directories or symlinks or whatever already present at the given path then the operation would abort with an error. Let's label this case for later discussion: A) There is an untracked path in the way Now let's consider the opposite case: Flipping !SKIP_WORKTREE -> SKIP_WORKTREE (removing files) If the index and working tree was clean this was fine, but if there were any unclean paths we would run into problems. There are three different cases to consider: B) The path is unmerged C) The path has unstaged changes D) The path has staged changes (differs from HEAD) If any path fell into case B or C, then the whole operation would be aborted with an error. With sparse-checkout, the whole operation would be aborted for case D as well, but for its predecessor of using `git read-tree -mu HEAD` directly, any paths that fell into case D would be removed from the working copy and the index entry for that path would be reset to match HEAD -- which looks and feels like data loss to users (only a few are even aware to ask whether it can be recovered, and even then it requires walking through loose objects trying to match up the right ones). Refusing to remove files that have unsaved user changes is good, but refusing to work on any other paths is very problematic for users. If the user is in the middle of a rebase or has made modifications to files that bring in more dependencies, then for their build to work they need to update the sparse paths. This logic has been preventing them from doing so. Sometimes in response, the user will stage the files and re-try, to no avail with sparse-checkout or to the horror of losing their changes if they are using its predecessor of `git read-tree -mu HEAD`. Add a new update_sparsity() function which will not error out in any of these cases but behaves as follows for the special cases: A) Leave the file in the working copy alone, clear the SKIP_WORKTREE bit, and print a warning (thus leaving the path in a state where status will report the file as modified, which seems logical). B) Do NOT mark this path as SKIP_WORKTREE, and leave it as unmerged. C) Do NOT mark this path as SKIP_WORKTREE and print a warning about the dirty path. D) Mark the path as SKIP_WORKTREE, but do not revert the version stored in the index to match HEAD; leave the contents alone. I tried a different behavior for A (leave the SKIP_WORKTREE bit set), but found it very surprising and counter-intuitive (e.g. the user sees it is present along with all the other files in that directory, tries to stage it, but git add ignores it since the SKIP_WORKTREE bit is set). A & C seem like optimal behavior to me. B may be as well, though I wonder if printing a warning would be an improvement. Some might be slightly surprised by D at first, but given that it does the right thing with `git commit` and even `git commit -a` (`git add` ignores entries that are marked SKIP_WORKTREE and thus doesn't delete them, and `commit -a` is similar), it seems logical to me. Reviewed-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-03-27 01:48:52 +01:00
unsigned old_show_all_errors;
int free_pattern_list = 0;
old_show_all_errors = o->show_all_errors;
o->show_all_errors = 1;
/* Sanity checks */
if (!o->update || o->index_only || o->skip_sparse_checkout)
BUG("update_sparsity() is for reflecting sparsity patterns in working directory");
if (o->src_index != o->dst_index || o->fn)
BUG("update_sparsity() called wrong");
trace_performance_enter();
/* If we weren't given patterns, use the recorded ones */
if (!o->pl) {
memset(&pl, 0, sizeof(pl));
free_pattern_list = 1;
populate_from_existing_patterns(o, &pl);
if (o->skip_sparse_checkout)
goto skip_sparse_checkout;
}
sparse-checkout: integrate with sparse index When modifying the sparse-checkout definition, the sparse-checkout builtin calls update_sparsity() to modify the SKIP_WORKTREE bits of all cache entries in the index. Before, we needed the index to be fully expanded in order to ensure we had the full list of files necessary that match the new patterns. Insert a call to reset_sparse_directories() that expands sparse directories that are within the new pattern list, but only far enough that every necessary file path now exists as a cache entry. The remaining logic within update_sparsity() will modify the SKIP_WORKTREE bits appropriately. This allows us to disable command_requires_full_index within the sparse-checkout builtin. Add tests that demonstrate that we are not expanding to a full index unnecessarily. We can see the improved performance in the p2000 test script: Test HEAD~1 HEAD ------------------------------------------------------------------------ 2000.24: git ... (sparse-v3) 2.14(1.55+0.58) 1.57(1.03+0.53) -26.6% 2000.25: git ... (sparse-v4) 2.20(1.62+0.57) 1.58(0.98+0.59) -28.2% These reductions of 26-28% are small compared to most examples, but the time is dominated by writing a new copy of the base repository to the worktree and then deleting it again. The fact that the previous index expansion was such a large portion of the time is telling how important it is to complete this sparse index integration. Signed-off-by: Derrick Stolee <derrickstolee@github.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2022-05-23 15:48:46 +02:00
/* Expand sparse directories as needed */
expand_index(o->src_index, o->pl);
unpack-trees: add a new update_sparsity() function Previously, the only way to update the SKIP_WORKTREE bits for various paths was invoking `git read-tree -mu HEAD` or calling the same code that this codepath invoked. This however had a number of problems if the index or working directory were not clean. First, let's consider the case: Flipping SKIP_WORKTREE -> !SKIP_WORKTREE (materializing files) If the working tree was clean this was fine, but if there were files or directories or symlinks or whatever already present at the given path then the operation would abort with an error. Let's label this case for later discussion: A) There is an untracked path in the way Now let's consider the opposite case: Flipping !SKIP_WORKTREE -> SKIP_WORKTREE (removing files) If the index and working tree was clean this was fine, but if there were any unclean paths we would run into problems. There are three different cases to consider: B) The path is unmerged C) The path has unstaged changes D) The path has staged changes (differs from HEAD) If any path fell into case B or C, then the whole operation would be aborted with an error. With sparse-checkout, the whole operation would be aborted for case D as well, but for its predecessor of using `git read-tree -mu HEAD` directly, any paths that fell into case D would be removed from the working copy and the index entry for that path would be reset to match HEAD -- which looks and feels like data loss to users (only a few are even aware to ask whether it can be recovered, and even then it requires walking through loose objects trying to match up the right ones). Refusing to remove files that have unsaved user changes is good, but refusing to work on any other paths is very problematic for users. If the user is in the middle of a rebase or has made modifications to files that bring in more dependencies, then for their build to work they need to update the sparse paths. This logic has been preventing them from doing so. Sometimes in response, the user will stage the files and re-try, to no avail with sparse-checkout or to the horror of losing their changes if they are using its predecessor of `git read-tree -mu HEAD`. Add a new update_sparsity() function which will not error out in any of these cases but behaves as follows for the special cases: A) Leave the file in the working copy alone, clear the SKIP_WORKTREE bit, and print a warning (thus leaving the path in a state where status will report the file as modified, which seems logical). B) Do NOT mark this path as SKIP_WORKTREE, and leave it as unmerged. C) Do NOT mark this path as SKIP_WORKTREE and print a warning about the dirty path. D) Mark the path as SKIP_WORKTREE, but do not revert the version stored in the index to match HEAD; leave the contents alone. I tried a different behavior for A (leave the SKIP_WORKTREE bit set), but found it very surprising and counter-intuitive (e.g. the user sees it is present along with all the other files in that directory, tries to stage it, but git add ignores it since the SKIP_WORKTREE bit is set). A & C seem like optimal behavior to me. B may be as well, though I wonder if printing a warning would be an improvement. Some might be slightly surprised by D at first, but given that it does the right thing with `git commit` and even `git commit -a` (`git add` ignores entries that are marked SKIP_WORKTREE and thus doesn't delete them, and `commit -a` is similar), it seems logical to me. Reviewed-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-03-27 01:48:52 +01:00
/* Set NEW_SKIP_WORKTREE on existing entries. */
mark_all_ce_unused(o->src_index);
mark_new_skip_worktree(o->pl, o->src_index, 0,
CE_NEW_SKIP_WORKTREE, o->verbose_update);
/* Then loop over entries and update/remove as needed */
ret = UPDATE_SPARSITY_SUCCESS;
for (i = 0; i < o->src_index->cache_nr; i++) {
struct cache_entry *ce = o->src_index->cache[i];
if (ce_stage(ce)) {
/* -1 because for loop will increment by 1 */
i += warn_conflicted_path(o->src_index, i, o) - 1;
ret = UPDATE_SPARSITY_WARNINGS;
continue;
}
unpack-trees: add a new update_sparsity() function Previously, the only way to update the SKIP_WORKTREE bits for various paths was invoking `git read-tree -mu HEAD` or calling the same code that this codepath invoked. This however had a number of problems if the index or working directory were not clean. First, let's consider the case: Flipping SKIP_WORKTREE -> !SKIP_WORKTREE (materializing files) If the working tree was clean this was fine, but if there were files or directories or symlinks or whatever already present at the given path then the operation would abort with an error. Let's label this case for later discussion: A) There is an untracked path in the way Now let's consider the opposite case: Flipping !SKIP_WORKTREE -> SKIP_WORKTREE (removing files) If the index and working tree was clean this was fine, but if there were any unclean paths we would run into problems. There are three different cases to consider: B) The path is unmerged C) The path has unstaged changes D) The path has staged changes (differs from HEAD) If any path fell into case B or C, then the whole operation would be aborted with an error. With sparse-checkout, the whole operation would be aborted for case D as well, but for its predecessor of using `git read-tree -mu HEAD` directly, any paths that fell into case D would be removed from the working copy and the index entry for that path would be reset to match HEAD -- which looks and feels like data loss to users (only a few are even aware to ask whether it can be recovered, and even then it requires walking through loose objects trying to match up the right ones). Refusing to remove files that have unsaved user changes is good, but refusing to work on any other paths is very problematic for users. If the user is in the middle of a rebase or has made modifications to files that bring in more dependencies, then for their build to work they need to update the sparse paths. This logic has been preventing them from doing so. Sometimes in response, the user will stage the files and re-try, to no avail with sparse-checkout or to the horror of losing their changes if they are using its predecessor of `git read-tree -mu HEAD`. Add a new update_sparsity() function which will not error out in any of these cases but behaves as follows for the special cases: A) Leave the file in the working copy alone, clear the SKIP_WORKTREE bit, and print a warning (thus leaving the path in a state where status will report the file as modified, which seems logical). B) Do NOT mark this path as SKIP_WORKTREE, and leave it as unmerged. C) Do NOT mark this path as SKIP_WORKTREE and print a warning about the dirty path. D) Mark the path as SKIP_WORKTREE, but do not revert the version stored in the index to match HEAD; leave the contents alone. I tried a different behavior for A (leave the SKIP_WORKTREE bit set), but found it very surprising and counter-intuitive (e.g. the user sees it is present along with all the other files in that directory, tries to stage it, but git add ignores it since the SKIP_WORKTREE bit is set). A & C seem like optimal behavior to me. B may be as well, though I wonder if printing a warning would be an improvement. Some might be slightly surprised by D at first, but given that it does the right thing with `git commit` and even `git commit -a` (`git add` ignores entries that are marked SKIP_WORKTREE and thus doesn't delete them, and `commit -a` is similar), it seems logical to me. Reviewed-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-03-27 01:48:52 +01:00
if (apply_sparse_checkout(o->src_index, ce, o))
ret = UPDATE_SPARSITY_WARNINGS;
}
skip_sparse_checkout:
if (check_updates(o, o->src_index))
ret = UPDATE_SPARSITY_WORKTREE_UPDATE_FAILURES;
display_warning_msgs(o);
unpack-trees: add a new update_sparsity() function Previously, the only way to update the SKIP_WORKTREE bits for various paths was invoking `git read-tree -mu HEAD` or calling the same code that this codepath invoked. This however had a number of problems if the index or working directory were not clean. First, let's consider the case: Flipping SKIP_WORKTREE -> !SKIP_WORKTREE (materializing files) If the working tree was clean this was fine, but if there were files or directories or symlinks or whatever already present at the given path then the operation would abort with an error. Let's label this case for later discussion: A) There is an untracked path in the way Now let's consider the opposite case: Flipping !SKIP_WORKTREE -> SKIP_WORKTREE (removing files) If the index and working tree was clean this was fine, but if there were any unclean paths we would run into problems. There are three different cases to consider: B) The path is unmerged C) The path has unstaged changes D) The path has staged changes (differs from HEAD) If any path fell into case B or C, then the whole operation would be aborted with an error. With sparse-checkout, the whole operation would be aborted for case D as well, but for its predecessor of using `git read-tree -mu HEAD` directly, any paths that fell into case D would be removed from the working copy and the index entry for that path would be reset to match HEAD -- which looks and feels like data loss to users (only a few are even aware to ask whether it can be recovered, and even then it requires walking through loose objects trying to match up the right ones). Refusing to remove files that have unsaved user changes is good, but refusing to work on any other paths is very problematic for users. If the user is in the middle of a rebase or has made modifications to files that bring in more dependencies, then for their build to work they need to update the sparse paths. This logic has been preventing them from doing so. Sometimes in response, the user will stage the files and re-try, to no avail with sparse-checkout or to the horror of losing their changes if they are using its predecessor of `git read-tree -mu HEAD`. Add a new update_sparsity() function which will not error out in any of these cases but behaves as follows for the special cases: A) Leave the file in the working copy alone, clear the SKIP_WORKTREE bit, and print a warning (thus leaving the path in a state where status will report the file as modified, which seems logical). B) Do NOT mark this path as SKIP_WORKTREE, and leave it as unmerged. C) Do NOT mark this path as SKIP_WORKTREE and print a warning about the dirty path. D) Mark the path as SKIP_WORKTREE, but do not revert the version stored in the index to match HEAD; leave the contents alone. I tried a different behavior for A (leave the SKIP_WORKTREE bit set), but found it very surprising and counter-intuitive (e.g. the user sees it is present along with all the other files in that directory, tries to stage it, but git add ignores it since the SKIP_WORKTREE bit is set). A & C seem like optimal behavior to me. B may be as well, though I wonder if printing a warning would be an improvement. Some might be slightly surprised by D at first, but given that it does the right thing with `git commit` and even `git commit -a` (`git add` ignores entries that are marked SKIP_WORKTREE and thus doesn't delete them, and `commit -a` is similar), it seems logical to me. Reviewed-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-03-27 01:48:52 +01:00
o->show_all_errors = old_show_all_errors;
if (free_pattern_list)
clear_pattern_list(&pl);
trace_performance_leave("update_sparsity");
return ret;
}
/* Here come the merge functions */
static int reject_merge(const struct cache_entry *ce,
struct unpack_trees_options *o)
{
return add_rejected_path(o, ERROR_WOULD_OVERWRITE, ce->name);
}
static int same(const struct cache_entry *a, const struct cache_entry *b)
{
if (!!a != !!b)
return 0;
if (!a && !b)
return 1;
"reset --merge": fix unmerged case Commit 9e8ecea (Add 'merge' mode to 'git reset', 2008-12-01) disallowed "git reset --merge" when there was unmerged entries. But it wished if unmerged entries were reset as if --hard (instead of --merge) has been used. This makes sense because all "mergy" operations makes sure that any path involved in the merge does not have local modifications before starting, so resetting such a path away won't lose any information. The previous commit changed the behavior of --merge to accept resetting unmerged entries if they are reset to a different state than HEAD, but it did not reset the changes in the work tree, leaving the conflict markers in the resulting file in the work tree. Fix it by doing three things: - Update the documentation to match the wish of original "reset --merge" better, namely, "An unmerged entry is a sign that the path didn't have any local modification and can be safely resetted to whatever the new HEAD records"; - Update read_index_unmerged(), which reads the index file into the cache while dropping any higher-stage entries down to stage #0, not to copy the object name from the higher stage entry. The code used to take the object name from the a stage entry ("base" if you happened to have stage #1, or "ours" if both sides added, etc.), which essentially meant that you are getting random results depending on what the merge did. The _only_ reason we want to keep a previously unmerged entry in the index at stage #0 is so that we don't forget the fact that we have corresponding file in the work tree in order to be able to remove it when the tree we are resetting to does not have the path. In order to differentiate such an entry from ordinary cache entry, the cache entry added by read_index_unmerged() is marked as CE_CONFLICTED. - Update merged_entry() and deleted_entry() so that they pay attention to cache entries marked as CE_CONFLICTED. They are previously unmerged entries, and the files in the work tree that correspond to them are resetted away by oneway_merge() to the version from the tree we are resetting to. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-01 08:04:04 +01:00
if ((a->ce_flags | b->ce_flags) & CE_CONFLICTED)
return 0;
return a->ce_mode == b->ce_mode &&
oideq(&a->oid, &b->oid);
}
/*
* When a CE gets turned into an unmerged entry, we
* want it to be up-to-date
*/
static int verify_uptodate_1(const struct cache_entry *ce,
struct unpack_trees_options *o,
enum unpack_trees_error_types error_type)
{
struct stat st;
if (o->index_only)
return 0;
/*
* CE_VALID and CE_SKIP_WORKTREE cheat, we better check again
* if this entry is truly up-to-date because this file may be
* overwritten.
*/
if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
; /* keep checking */
else if (o->reset || ce_uptodate(ce))
return 0;
if (!lstat(ce->name, &st)) {
int flags = CE_MATCH_IGNORE_VALID|CE_MATCH_IGNORE_SKIP_WORKTREE;
unsigned changed = ie_match_stat(o->src_index, ce, &st, flags);
if (submodule_from_ce(ce)) {
int r = check_submodule_move_head(ce,
"HEAD", oid_to_hex(&ce->oid), o);
if (r)
return add_rejected_path(o, error_type, ce->name);
return 0;
}
if (!changed)
return 0;
unpack-trees.c: assume submodules are clean during check-out Sven originally raised this issue: If you have a submodule checked out and you go back (or forward) to a revision of the supermodule that contains a different revision of the submodule and then switch to another revision, it will complain that the submodule is not uptodate, because git simply didn't update the submodule in the first move. The current policy is to consider it is perfectly normal that checked-out submodule is out-of-sync wrt the supermodule index. At least until we introduce a superproject repository configuration option that says "in this repository, I do care about this submodule and at any time I move around in the superproject, recursively check out the submodule to match", it is a reasonable policy, as we currently do not recursively checkout the submodules at all. The most extreme case of this policy is that the superproject index knows about the submodule but the subdirectory does not even have to be checked out. The function verify_uptodate(), called during the two-way merge aka branch switching, is about "make sure the filesystem entity that corresponds to this cache entry is up to date, lest we lose the local modifications". As we explicitly allow submodule checkout to drift from the supermodule index entry, the check should say "Ok, for submodules, not matching is the norm" for now. Later when we have the ability to mark "I care about this submodule to be always in sync with the superproject" (thereby implementing automatic recursive checkout and perhaps diff, among other things), we should check if the submodule in question is marked as such and perform the current test. Acked-by: Lars Hjemli <hjemli@gmail.com> Acked-by: Sven Verdoolaege <skimo@kotnet.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2007-08-04 07:13:09 +02:00
/*
* Historic default policy was to allow submodule to be out
* of sync wrt the superproject index. If the submodule was
* not considered interesting above, we don't care here.
unpack-trees.c: assume submodules are clean during check-out Sven originally raised this issue: If you have a submodule checked out and you go back (or forward) to a revision of the supermodule that contains a different revision of the submodule and then switch to another revision, it will complain that the submodule is not uptodate, because git simply didn't update the submodule in the first move. The current policy is to consider it is perfectly normal that checked-out submodule is out-of-sync wrt the supermodule index. At least until we introduce a superproject repository configuration option that says "in this repository, I do care about this submodule and at any time I move around in the superproject, recursively check out the submodule to match", it is a reasonable policy, as we currently do not recursively checkout the submodules at all. The most extreme case of this policy is that the superproject index knows about the submodule but the subdirectory does not even have to be checked out. The function verify_uptodate(), called during the two-way merge aka branch switching, is about "make sure the filesystem entity that corresponds to this cache entry is up to date, lest we lose the local modifications". As we explicitly allow submodule checkout to drift from the supermodule index entry, the check should say "Ok, for submodules, not matching is the norm" for now. Later when we have the ability to mark "I care about this submodule to be always in sync with the superproject" (thereby implementing automatic recursive checkout and perhaps diff, among other things), we should check if the submodule in question is marked as such and perform the current test. Acked-by: Lars Hjemli <hjemli@gmail.com> Acked-by: Sven Verdoolaege <skimo@kotnet.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2007-08-04 07:13:09 +02:00
*/
if (S_ISGITLINK(ce->ce_mode))
return 0;
errno = 0;
}
if (errno == ENOENT)
return 0;
return add_rejected_path(o, error_type, ce->name);
}
int verify_uptodate(const struct cache_entry *ce,
struct unpack_trees_options *o)
{
unpack-trees: fix accidental loss of user changes For sparse-checkouts, we don't want unpack-trees to error out on files that are missing from the worktree, so there has traditionally been logic to make it skip the verify_uptodate() check for these. Unfortunately, it was skipping the verify_uptodate() check for files that were expected to *become* SKIP_WORKTREE. For files that were not already SKIP_WORKTREE, that can cause us to later delete the file in apply_sparse_checkout(). Only skip the check for files that were already SKIP_WORKTREE as well to avoid lightly discarding important changes users may have made to files. Note 1: unpack-trees.c is already a bit complex, and the logic around CE_SKIP_WORKTREE and CE_NEW_SKIP_WORKTREE in that file are no exception. I also tried just replacing CE_NEW_SKIP_WORKTREE with CE_SKIP_WORKTREE in the verify_uptodate() check instead of checking for both flags, and found that it also fixed this bug and passed all the tests. I also attempted to devise a few testcases that might trip either variant of my fix and was unable to find any problems. It may be that just checking CE_SKIP_WORKTREE is a better fix, but I'm not sure. I thought it was a bit safer to strictly reduce the number of cases where we skip the up-to-date check rather than just toggling which kind of cases skip it, and thus went with the current variant of the fix. Note 2: I also wondered if verify_absent() might have a similar bug, but despite my attempts to try to devise a testcase that would trigger such a thing, I couldn't find any problematic testcases. Thus, this patch makes no attempt to apply similar changes to verify_absent() and verify_absent_if_directory(). Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2022-01-14 16:59:40 +01:00
if (!o->skip_sparse_checkout &&
(ce->ce_flags & CE_SKIP_WORKTREE) &&
(ce->ce_flags & CE_NEW_SKIP_WORKTREE))
return 0;
return verify_uptodate_1(ce, o, ERROR_NOT_UPTODATE_FILE);
}
static int verify_uptodate_sparse(const struct cache_entry *ce,
struct unpack_trees_options *o)
{
return verify_uptodate_1(ce, o, WARNING_SPARSE_NOT_UPTODATE_FILE);
}
/*
* TODO: We should actually invalidate o->result, not src_index [1].
* But since cache tree and untracked cache both are not copied to
* o->result until unpacking is complete, we invalidate them on
* src_index instead with the assumption that they will be copied to
* dst_index at the end.
*
* [1] src_index->cache_tree is also used in unpack_callback() so if
* we invalidate o->result, we need to update it to use
* o->result.cache_tree as well.
*/
static void invalidate_ce_path(const struct cache_entry *ce,
struct unpack_trees_options *o)
{
if (!ce)
return;
cache_tree_invalidate_path(o->src_index, ce->name);
untracked_cache_invalidate_path(o->src_index, ce->name, 1);
}
/*
* Check that checking out ce->sha1 in subdir ce->name is not
* going to overwrite any working files.
*/
static int verify_clean_submodule(const char *old_sha1,
const struct cache_entry *ce,
struct unpack_trees_options *o)
{
if (!submodule_from_ce(ce))
return 0;
return check_submodule_move_head(ce, old_sha1,
oid_to_hex(&ce->oid), o);
}
static int verify_clean_subdirectory(const struct cache_entry *ce,
struct unpack_trees_options *o)
{
/*
* we are about to extract "ce->name"; we would not want to lose
* anything in the existing directory there.
*/
int namelen;
int i;
struct dir_struct d;
char *pathbuf;
int cnt = 0;
if (S_ISGITLINK(ce->ce_mode)) {
struct object_id oid;
int sub_head = resolve_gitlink_ref(ce->name, "HEAD", &oid);
/*
* If we are not going to update the submodule, then
* we don't care.
*/
if (!sub_head && oideq(&oid, &ce->oid))
return 0;
return verify_clean_submodule(sub_head ? NULL : oid_to_hex(&oid),
ce, o);
}
/*
* First let's make sure we do not have a local modification
* in that directory.
*/
namelen = ce_namelen(ce);
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
for (i = locate_in_src_index(ce, o);
i < o->src_index->cache_nr;
i++) {
struct cache_entry *ce2 = o->src_index->cache[i];
int len = ce_namelen(ce2);
if (len < namelen ||
strncmp(ce->name, ce2->name, namelen) ||
ce2->name[namelen] != '/')
break;
/*
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
* ce2->name is an entry in the subdirectory to be
* removed.
*/
if (!ce_stage(ce2)) {
if (verify_uptodate(ce2, o))
return -1;
add_entry(o, ce2, CE_REMOVE, 0);
unpack-trees: add missing cache invalidation Any changes to the output index should be (confusingly) marked in the source index with invalidate_ce_path(). This is used to make sure we still have valid untracked cache and cache-tree extensions in the end. We do a pretty good job of invalidating except in two places. verify_clean_subdirectory() is part of verify_absent() and verify_absent_sparse(). The former is usually called by merged_entry() or directly in threeway_merge(). The latter is obviously used by sparse checkout. In these three call sites, only merged_entry() follows up with invalidate_ce_path(). The other two don't, but they should not trigger this ce removal because this is about D/F conflicts [1]. But let's be safe and invalidate_ce_path() here as well. The second place is keep_entry() which is also used by threeway_merge() to keep higher stage entries. In order to reuse cache-tree we need to invalidate these paths as well. It's not a problem in the past because whenever a higher stage entry is present, cache-tree will not be created [2]. Now we salvage cache-tree even when higher stage entries are present, we need more invalidation. [1] c81935348b (Fix switching to a branch with D/F when current branch has file D. - 2007-03-15) [2] This is probably too strict. We should be able to create and save cache-tree for the directories that do not have conflict entries in cache_tree_update(). And this becomes more important when cache-tree plays bigger role in terms of performance. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-08-18 16:41:27 +02:00
invalidate_ce_path(ce, o);
unpack-trees.c: prepare for looking ahead in the index This prepares but does not yet implement a look-ahead in the index entries when traverse-trees.c decides to give us tree entries in an order that does not match what is in the index. A case where a look-ahead in the index is necessary happens when merging branch B into branch A while the index matches the current branch A, using a tree O as their common ancestor, and these three trees looks like this: O A B t t t-i t-i t-i t-j t-j t/1 t/2 The traverse_trees() function gets "t", "t-i" and "t" from trees O, A and B first, and notices that A may have a matching "t" behind "t-i" and "t-j" (indeed it does), and tells A to give that entry instead. After unpacking blob "t" from tree B (as it hasn't changed since O in B and A removed it, it will result in its removal), it descends into directory "t/". The side that walked index in parallel to the tree traversal used to be implemented with one pointer, o->pos, that points at the next index entry to be processed. When this happens, the pointer o->pos still points at "t-i" that is the first entry. We should be able to skip "t-i" and "t-j" and locate "t/1" from the index while the recursive invocation of traverse_trees() walks and match entries found there, and later come back to process "t-i". While that look-ahead is not implemented yet, this adds a flag bit, CE_UNPACKED, to mark the entries in the index that has already been processed. o->pos pointer has been renamed to o->cache_bottom and it points at the first entry that may still need to be processed. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-07 23:59:54 +01:00
mark_ce_used(ce2, o);
}
cnt++;
}
/* Do not lose a locally present file that is not ignored. */
pathbuf = xstrfmt("%.*s/", namelen, ce->name);
memset(&d, 0, sizeof(d));
if (o->dir)
d.exclude_per_dir = o->dir->exclude_per_dir;
i = read_directory(&d, o->src_index, pathbuf, namelen+1, NULL);
2021-10-07 11:46:09 +02:00
dir_clear(&d);
free(pathbuf);
if (i)
return add_rejected_path(o, ERROR_NOT_UPTODATE_DIR, ce->name);
/* Do not lose startup_info->original_cwd */
if (startup_info->original_cwd &&
!strcmp(startup_info->original_cwd, ce->name))
return add_rejected_path(o, ERROR_CWD_IN_THE_WAY, ce->name);
return cnt;
}
/*
* This gets called when there was no index entry for the tree entry 'dst',
* but we found a file in the working tree that 'lstat()' said was fine,
* and we're on a case-insensitive filesystem.
*
* See if we can find a case-insensitive match in the index that also
* matches the stat information, and assume it's that other file!
*/
static int icase_exists(struct unpack_trees_options *o, const char *name, int len, struct stat *st)
{
Convert "struct cache_entry *" to "const ..." wherever possible I attempted to make index_state->cache[] a "const struct cache_entry **" to find out how existing entries in index are modified and where. The question I have is what do we do if we really need to keep track of on-disk changes in the index. The result is - diff-lib.c: setting CE_UPTODATE - name-hash.c: setting CE_HASHED - preload-index.c, read-cache.c, unpack-trees.c and builtin/update-index: obvious - entry.c: write_entry() may refresh the checked out entry via fill_stat_cache_info(). This causes "non-const struct cache_entry *" in builtin/apply.c, builtin/checkout-index.c and builtin/checkout.c - builtin/ls-files.c: --with-tree changes stagemask and may set CE_UPDATE Of these, write_entry() and its call sites are probably most interesting because it modifies on-disk info. But this is stat info and can be retrieved via refresh, at least for porcelain commands. Other just uses ce_flags for local purposes. So, keeping track of "dirty" entries is just a matter of setting a flag in index modification functions exposed by read-cache.c. Except unpack-trees, the rest of the code base does not do anything funny behind read-cache's back. The actual patch is less valueable than the summary above. But if anyone wants to re-identify the above sites. Applying this patch, then this: diff --git a/cache.h b/cache.h index 430d021..1692891 100644 --- a/cache.h +++ b/cache.h @@ -267,7 +267,7 @@ static inline unsigned int canon_mode(unsigned int mode) #define cache_entry_size(len) (offsetof(struct cache_entry,name) + (len) + 1) struct index_state { - struct cache_entry **cache; + const struct cache_entry **cache; unsigned int version; unsigned int cache_nr, cache_alloc, cache_changed; struct string_list *resolve_undo; will help quickly identify them without bogus warnings. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-07-09 17:29:00 +02:00
const struct cache_entry *src;
src = index_file_exists(o->src_index, name, len, 1);
return src && !ie_match_stat(o->src_index, src, st, CE_MATCH_IGNORE_VALID|CE_MATCH_IGNORE_SKIP_WORKTREE);
}
enum absent_checking_type {
COMPLETELY_ABSENT,
ABSENT_ANY_DIRECTORY
};
static int check_ok_to_remove(const char *name, int len, int dtype,
const struct cache_entry *ce, struct stat *st,
enum unpack_trees_error_types error_type,
enum absent_checking_type absent_type,
struct unpack_trees_options *o)
{
Convert "struct cache_entry *" to "const ..." wherever possible I attempted to make index_state->cache[] a "const struct cache_entry **" to find out how existing entries in index are modified and where. The question I have is what do we do if we really need to keep track of on-disk changes in the index. The result is - diff-lib.c: setting CE_UPTODATE - name-hash.c: setting CE_HASHED - preload-index.c, read-cache.c, unpack-trees.c and builtin/update-index: obvious - entry.c: write_entry() may refresh the checked out entry via fill_stat_cache_info(). This causes "non-const struct cache_entry *" in builtin/apply.c, builtin/checkout-index.c and builtin/checkout.c - builtin/ls-files.c: --with-tree changes stagemask and may set CE_UPDATE Of these, write_entry() and its call sites are probably most interesting because it modifies on-disk info. But this is stat info and can be retrieved via refresh, at least for porcelain commands. Other just uses ce_flags for local purposes. So, keeping track of "dirty" entries is just a matter of setting a flag in index modification functions exposed by read-cache.c. Except unpack-trees, the rest of the code base does not do anything funny behind read-cache's back. The actual patch is less valueable than the summary above. But if anyone wants to re-identify the above sites. Applying this patch, then this: diff --git a/cache.h b/cache.h index 430d021..1692891 100644 --- a/cache.h +++ b/cache.h @@ -267,7 +267,7 @@ static inline unsigned int canon_mode(unsigned int mode) #define cache_entry_size(len) (offsetof(struct cache_entry,name) + (len) + 1) struct index_state { - struct cache_entry **cache; + const struct cache_entry **cache; unsigned int version; unsigned int cache_nr, cache_alloc, cache_changed; struct string_list *resolve_undo; will help quickly identify them without bogus warnings. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-07-09 17:29:00 +02:00
const struct cache_entry *result;
/*
* It may be that the 'lstat()' succeeded even though
* target 'ce' was absent, because there is an old
* entry that is different only in case..
*
* Ignore that lstat() if it matches.
*/
if (ignore_case && icase_exists(o, name, len, st))
return 0;
if (o->dir &&
is_excluded(o->dir, o->src_index, name, &dtype))
/*
* ce->name is explicitly excluded, so it is Ok to
* overwrite it.
*/
return 0;
if (S_ISDIR(st->st_mode)) {
/*
* We are checking out path "foo" and
* found "foo/." in the working tree.
* This is tricky -- if we have modified
* files that are in "foo/" we would lose
* them.
*/
if (verify_clean_subdirectory(ce, o) < 0)
return -1;
return 0;
}
/* If we only care about directories, then we can remove */
if (absent_type == ABSENT_ANY_DIRECTORY)
return 0;
/*
* The previous round may already have decided to
* delete this path, which is in a subdirectory that
* is being replaced with a blob.
*/
result = index_file_exists(&o->result, name, len, 0);
if (result) {
if (result->ce_flags & CE_REMOVE)
return 0;
}
return add_rejected_path(o, error_type, name);
}
/*
* We do not want to remove or overwrite a working tree file that
* is not tracked, unless it is ignored.
*/
static int verify_absent_1(const struct cache_entry *ce,
enum unpack_trees_error_types error_type,
enum absent_checking_type absent_type,
struct unpack_trees_options *o)
{
int len;
struct stat st;
if (o->index_only || !o->update)
return 0;
if (o->reset == UNPACK_RESET_OVERWRITE_UNTRACKED) {
/* Avoid nuking startup_info->original_cwd... */
if (startup_info->original_cwd &&
!strcmp(startup_info->original_cwd, ce->name))
return add_rejected_path(o, ERROR_CWD_IN_THE_WAY,
ce->name);
/* ...but nuke anything else. */
return 0;
}
checkout: don't follow symlinks when removing entries At 1d718a5108 ("do not overwrite untracked symlinks", 2011-02-20), symlink.c:check_leading_path() started returning different codes for FL_ENOENT and FL_SYMLINK. But one of its callers, unlink_entry(), was not adjusted for this change, so it started to follow symlinks on the leading path of to-be-removed entries. Fix that and add a regression test. Note that since 1d718a5108 check_leading_path() no longer differentiates the case where it found a symlink in the path's leading components from the cases where it found a regular file or failed to lstat() the component. So, a side effect of this current patch is that unlink_entry() now returns early in all of these three cases. And because we no longer try to unlink such paths, we also don't get the warning from remove_or_warn(). For the regular file and symlink cases, it's questionable whether the warning was useful in the first place: unlink_entry() removes tracked paths that should no longer be present in the state we are checking out to. If the path had its leading dir replaced by another file, it means that the basename already doesn't exist, so there is no need for a warning. Sure, we are leaving a regular file or symlink behind at the path's dirname, but this file is either untracked now (so again, no need to warn), or it will be replaced by a tracked file during the next phase of this checkout operation. As for failing to lstat() one of the leading components, the basename might still exist only we cannot unlink it (e.g. due to the lack of the required permissions). Since the user expect it to be removed (especially with checkout's --no-overlay option), add back the warning in this more relevant case. Signed-off-by: Matheus Tavares <matheus.bernardino@usp.br> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2021-03-18 19:43:47 +01:00
len = check_leading_path(ce->name, ce_namelen(ce), 0);
if (!len)
return 0;
else if (len > 0) {
verify_absent: allow filenames longer than PATH_MAX When unpack-trees wants to know whether a path will overwrite anything in the working tree, we use lstat() to see if there is anything there. But if we are going to write "foo/bar", we can't just lstat("foo/bar"); we need to look for leading prefixes (e.g., "foo"). So we use the lstat cache to find the length of the leading prefix, and copy the filename up to that length into a temporary buffer (since the original name is const, we cannot just stick a NUL in it). The copy we make goes into a PATH_MAX-sized buffer, which will overflow if the prefix is longer than PATH_MAX. How this happens is a little tricky, since in theory PATH_MAX is the biggest path we will have read from the filesystem. But this can happen if: - the compiled-in PATH_MAX does not accurately reflect what the filesystem is capable of - the leading prefix is not _quite_ what is on disk; it contains the next element from the name we are checking. So if we want to write "aaa/bbb/ccc/ddd" and "aaa/bbb" exists, the prefix of interest is "aaa/bbb/ccc". If "aaa/bbb" approaches PATH_MAX, then "ccc" can overflow it. So this can be triggered, but it's hard to do. In particular, you cannot just "git clone" a bogus repo. The verify_absent checks happen before unpack-trees writes anything to the filesystem, so there are never any leading prefixes during the initial checkout, and the bug doesn't trigger. And by definition, these files are larger than PATH_MAX, so writing them will fail, and clone will complain (though it may write a partial path, which will cause a subsequent "git checkout" to hit the bug). We can fix it by creating the temporary path on the heap. The extra malloc overhead is not important, as we are already making at least one stat() call (and probably more for the prefix discovery). Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-08-19 20:12:37 +02:00
char *path;
int ret;
path = xmemdupz(ce->name, len);
if (lstat(path, &st))
ret = error_errno("cannot stat '%s'", path);
else {
if (submodule_from_ce(ce))
ret = check_submodule_move_head(ce,
oid_to_hex(&ce->oid),
NULL, o);
else
ret = check_ok_to_remove(path, len, DT_UNKNOWN, NULL,
&st, error_type,
absent_type, o);
}
verify_absent: allow filenames longer than PATH_MAX When unpack-trees wants to know whether a path will overwrite anything in the working tree, we use lstat() to see if there is anything there. But if we are going to write "foo/bar", we can't just lstat("foo/bar"); we need to look for leading prefixes (e.g., "foo"). So we use the lstat cache to find the length of the leading prefix, and copy the filename up to that length into a temporary buffer (since the original name is const, we cannot just stick a NUL in it). The copy we make goes into a PATH_MAX-sized buffer, which will overflow if the prefix is longer than PATH_MAX. How this happens is a little tricky, since in theory PATH_MAX is the biggest path we will have read from the filesystem. But this can happen if: - the compiled-in PATH_MAX does not accurately reflect what the filesystem is capable of - the leading prefix is not _quite_ what is on disk; it contains the next element from the name we are checking. So if we want to write "aaa/bbb/ccc/ddd" and "aaa/bbb" exists, the prefix of interest is "aaa/bbb/ccc". If "aaa/bbb" approaches PATH_MAX, then "ccc" can overflow it. So this can be triggered, but it's hard to do. In particular, you cannot just "git clone" a bogus repo. The verify_absent checks happen before unpack-trees writes anything to the filesystem, so there are never any leading prefixes during the initial checkout, and the bug doesn't trigger. And by definition, these files are larger than PATH_MAX, so writing them will fail, and clone will complain (though it may write a partial path, which will cause a subsequent "git checkout" to hit the bug). We can fix it by creating the temporary path on the heap. The extra malloc overhead is not important, as we are already making at least one stat() call (and probably more for the prefix discovery). Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-08-19 20:12:37 +02:00
free(path);
return ret;
} else if (lstat(ce->name, &st)) {
if (errno != ENOENT)
return error_errno("cannot stat '%s'", ce->name);
return 0;
} else {
if (submodule_from_ce(ce))
return check_submodule_move_head(ce, oid_to_hex(&ce->oid),
NULL, o);
return check_ok_to_remove(ce->name, ce_namelen(ce),
ce_to_dtype(ce), ce, &st,
error_type, absent_type, o);
}
}
static int verify_absent(const struct cache_entry *ce,
enum unpack_trees_error_types error_type,
struct unpack_trees_options *o)
{
if (!o->skip_sparse_checkout && (ce->ce_flags & CE_NEW_SKIP_WORKTREE))
return 0;
return verify_absent_1(ce, error_type, COMPLETELY_ABSENT, o);
}
static int verify_absent_if_directory(const struct cache_entry *ce,
enum unpack_trees_error_types error_type,
struct unpack_trees_options *o)
{
if (!o->skip_sparse_checkout && (ce->ce_flags & CE_NEW_SKIP_WORKTREE))
return 0;
return verify_absent_1(ce, error_type, ABSENT_ANY_DIRECTORY, o);
}
static int verify_absent_sparse(const struct cache_entry *ce,
enum unpack_trees_error_types error_type,
struct unpack_trees_options *o)
{
return verify_absent_1(ce, error_type, COMPLETELY_ABSENT, o);
}
static int merged_entry(const struct cache_entry *ce,
const struct cache_entry *old,
struct unpack_trees_options *o)
{
int update = CE_UPDATE;
block alloc: allocate cache entries from mem_pool When reading large indexes from disk, a portion of the time is dominated in malloc() calls. This can be mitigated by allocating a large block of memory and manage it ourselves via memory pools. This change moves the cache entry allocation to be on top of memory pools. Design: The index_state struct will gain a notion of an associated memory_pool from which cache_entries will be allocated from. When reading in the index from disk, we have information on the number of entries and their size, which can guide us in deciding how large our initial memory allocation should be. When an index is discarded, the associated memory_pool will be discarded as well - so the lifetime of a cache_entry is tied to the lifetime of the index_state that it was allocated for. In the case of a Split Index, the following rules are followed. 1st, some terminology is defined: Terminology: - 'the_index': represents the logical view of the index - 'split_index': represents the "base" cache entries. Read from the split index file. 'the_index' can reference a single split_index, as well as cache_entries from the split_index. `the_index` will be discarded before the `split_index` is. This means that when we are allocating cache_entries in the presence of a split index, we need to allocate the entries from the `split_index`'s memory pool. This allows us to follow the pattern that `the_index` can reference cache_entries from the `split_index`, and that the cache_entries will not be freed while they are still being referenced. Managing transient cache_entry structs: Cache entries are usually allocated for an index, but this is not always the case. Cache entries are sometimes allocated because this is the type that the existing checkout_entry function works with. Because of this, the existing code needs to handle cache entries associated with an index / memory pool, and those that only exist transiently. Several strategies were contemplated around how to handle this: Chosen approach: An extra field was added to the cache_entry type to track whether the cache_entry was allocated from a memory pool or not. This is currently an int field, as there are no more available bits in the existing ce_flags bit field. If / when more bits are needed, this new field can be turned into a proper bit field. Alternatives: 1) Do not include any information about how the cache_entry was allocated. Calling code would be responsible for tracking whether the cache_entry needed to be freed or not. Pro: No extra memory overhead to track this state Con: Extra complexity in callers to handle this correctly. The extra complexity and burden to not regress this behavior in the future was more than we wanted. 2) cache_entry would gain knowledge about which mem_pool allocated it Pro: Could (potentially) do extra logic to know when a mem_pool no longer had references to any cache_entry Con: cache_entry would grow heavier by a pointer, instead of int We didn't see a tangible benefit to this approach 3) Do not add any extra information to a cache_entry, but when freeing a cache entry, check if the memory exists in a region managed by existing mem_pools. Pro: No extra memory overhead to track state Con: Extra computation is performed when freeing cache entries We decided tracking and iterating over known memory pool regions was less desirable than adding an extra field to track this stae. Signed-off-by: Jameson Miller <jamill@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-07-02 21:49:37 +02:00
struct cache_entry *merge = dup_cache_entry(ce, &o->result);
"reset --merge": fix unmerged case Commit 9e8ecea (Add 'merge' mode to 'git reset', 2008-12-01) disallowed "git reset --merge" when there was unmerged entries. But it wished if unmerged entries were reset as if --hard (instead of --merge) has been used. This makes sense because all "mergy" operations makes sure that any path involved in the merge does not have local modifications before starting, so resetting such a path away won't lose any information. The previous commit changed the behavior of --merge to accept resetting unmerged entries if they are reset to a different state than HEAD, but it did not reset the changes in the work tree, leaving the conflict markers in the resulting file in the work tree. Fix it by doing three things: - Update the documentation to match the wish of original "reset --merge" better, namely, "An unmerged entry is a sign that the path didn't have any local modification and can be safely resetted to whatever the new HEAD records"; - Update read_index_unmerged(), which reads the index file into the cache while dropping any higher-stage entries down to stage #0, not to copy the object name from the higher stage entry. The code used to take the object name from the a stage entry ("base" if you happened to have stage #1, or "ours" if both sides added, etc.), which essentially meant that you are getting random results depending on what the merge did. The _only_ reason we want to keep a previously unmerged entry in the index at stage #0 is so that we don't forget the fact that we have corresponding file in the work tree in order to be able to remove it when the tree we are resetting to does not have the path. In order to differentiate such an entry from ordinary cache entry, the cache entry added by read_index_unmerged() is marked as CE_CONFLICTED. - Update merged_entry() and deleted_entry() so that they pay attention to cache entries marked as CE_CONFLICTED. They are previously unmerged entries, and the files in the work tree that correspond to them are resetted away by oneway_merge() to the version from the tree we are resetting to. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-01 08:04:04 +01:00
if (!old) {
/*
* New index entries. In sparse checkout, the following
* verify_absent() will be delayed until after
* traverse_trees() finishes in unpack_trees(), then:
*
* - CE_NEW_SKIP_WORKTREE will be computed correctly
* - verify_absent() be called again, this time with
* correct CE_NEW_SKIP_WORKTREE
*
* verify_absent() call here does nothing in sparse
* checkout (i.e. o->skip_sparse_checkout == 0)
*/
update |= CE_ADDED;
merge->ce_flags |= CE_NEW_SKIP_WORKTREE;
if (verify_absent(merge,
ERROR_WOULD_LOSE_UNTRACKED_OVERWRITTEN, o)) {
block alloc: add lifecycle APIs for cache_entry structs It has been observed that the time spent loading an index with a large number of entries is partly dominated by malloc() calls. This change is in preparation for using memory pools to reduce the number of malloc() calls made to allocate cahce entries when loading an index. Add an API to allocate and discard cache entries, abstracting the details of managing the memory backing the cache entries. This commit does actually change how memory is managed - this will be done in a later commit in the series. This change makes the distinction between cache entries that are associated with an index and cache entries that are not associated with an index. A main use of cache entries is with an index, and we can optimize the memory management around this. We still have other cases where a cache entry is not persisted with an index, and so we need to handle the "transient" use case as well. To keep the congnitive overhead of managing the cache entries, there will only be a single discard function. This means there must be enough information kept with the cache entry so that we know how to discard them. A summary of the main functions in the API is: make_cache_entry: create cache entry for use in an index. Uses specified parameters to populate cache_entry fields. make_empty_cache_entry: Create an empty cache entry for use in an index. Returns cache entry with empty fields. make_transient_cache_entry: create cache entry that is not used in an index. Uses specified parameters to populate cache_entry fields. make_empty_transient_cache_entry: create cache entry that is not used in an index. Returns cache entry with empty fields. discard_cache_entry: A single function that knows how to discard a cache entry regardless of how it was allocated. Signed-off-by: Jameson Miller <jamill@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-07-02 21:49:31 +02:00
discard_cache_entry(merge);
"reset --merge": fix unmerged case Commit 9e8ecea (Add 'merge' mode to 'git reset', 2008-12-01) disallowed "git reset --merge" when there was unmerged entries. But it wished if unmerged entries were reset as if --hard (instead of --merge) has been used. This makes sense because all "mergy" operations makes sure that any path involved in the merge does not have local modifications before starting, so resetting such a path away won't lose any information. The previous commit changed the behavior of --merge to accept resetting unmerged entries if they are reset to a different state than HEAD, but it did not reset the changes in the work tree, leaving the conflict markers in the resulting file in the work tree. Fix it by doing three things: - Update the documentation to match the wish of original "reset --merge" better, namely, "An unmerged entry is a sign that the path didn't have any local modification and can be safely resetted to whatever the new HEAD records"; - Update read_index_unmerged(), which reads the index file into the cache while dropping any higher-stage entries down to stage #0, not to copy the object name from the higher stage entry. The code used to take the object name from the a stage entry ("base" if you happened to have stage #1, or "ours" if both sides added, etc.), which essentially meant that you are getting random results depending on what the merge did. The _only_ reason we want to keep a previously unmerged entry in the index at stage #0 is so that we don't forget the fact that we have corresponding file in the work tree in order to be able to remove it when the tree we are resetting to does not have the path. In order to differentiate such an entry from ordinary cache entry, the cache entry added by read_index_unmerged() is marked as CE_CONFLICTED. - Update merged_entry() and deleted_entry() so that they pay attention to cache entries marked as CE_CONFLICTED. They are previously unmerged entries, and the files in the work tree that correspond to them are resetted away by oneway_merge() to the version from the tree we are resetting to. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-01 08:04:04 +01:00
return -1;
}
"reset --merge": fix unmerged case Commit 9e8ecea (Add 'merge' mode to 'git reset', 2008-12-01) disallowed "git reset --merge" when there was unmerged entries. But it wished if unmerged entries were reset as if --hard (instead of --merge) has been used. This makes sense because all "mergy" operations makes sure that any path involved in the merge does not have local modifications before starting, so resetting such a path away won't lose any information. The previous commit changed the behavior of --merge to accept resetting unmerged entries if they are reset to a different state than HEAD, but it did not reset the changes in the work tree, leaving the conflict markers in the resulting file in the work tree. Fix it by doing three things: - Update the documentation to match the wish of original "reset --merge" better, namely, "An unmerged entry is a sign that the path didn't have any local modification and can be safely resetted to whatever the new HEAD records"; - Update read_index_unmerged(), which reads the index file into the cache while dropping any higher-stage entries down to stage #0, not to copy the object name from the higher stage entry. The code used to take the object name from the a stage entry ("base" if you happened to have stage #1, or "ours" if both sides added, etc.), which essentially meant that you are getting random results depending on what the merge did. The _only_ reason we want to keep a previously unmerged entry in the index at stage #0 is so that we don't forget the fact that we have corresponding file in the work tree in order to be able to remove it when the tree we are resetting to does not have the path. In order to differentiate such an entry from ordinary cache entry, the cache entry added by read_index_unmerged() is marked as CE_CONFLICTED. - Update merged_entry() and deleted_entry() so that they pay attention to cache entries marked as CE_CONFLICTED. They are previously unmerged entries, and the files in the work tree that correspond to them are resetted away by oneway_merge() to the version from the tree we are resetting to. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-01 08:04:04 +01:00
invalidate_ce_path(merge, o);
unpack-trees: check for missing submodule directory in merged_entry Using `git checkout --recurse-submodules` to switch between a branch with no submodules and a branch with initialized nested submodules currently causes a fatal error: $ git checkout --recurse-submodules branch-with-nested-submodules fatal: exec '--super-prefix=submodule/nested/': cd to 'nested' failed: No such file or directory error: Submodule 'nested' could not be updated. error: Submodule 'submodule/nested' cannot checkout new HEAD. error: Submodule 'submodule' could not be updated. M submodule Switched to branch 'branch-with-nested-submodules' The checkout succeeds but the worktree and index of the first level submodule are left empty: $ cd submodule $ git -c status.submoduleSummary=1 status HEAD detached at b3ce885 Changes to be committed: (use "git restore --staged <file>..." to unstage) deleted: .gitmodules deleted: first.t deleted: nested fatal: not a git repository: 'nested/.git' Submodule changes to be committed: * nested 1e96f59...0000000: $ git ls-files -s $ # empty $ ls -A .git The reason for the fatal error during the checkout is that a child git process tries to cd into the yet unexisting nested submodule directory. The sequence is the following: 1. The main git process (the one running in the superproject) eventually reaches write_entry() in entry.c, which creates the first level submodule directory and then calls submodule_move_head() in submodule.c, which spawns `git read-tree` in the submodule directory. 2. The first child git process (the one in the submodule of the superproject) eventually calls check_submodule_move_head() at unpack_trees.c:2021, which calls submodule_move_head in dry-run mode, which spawns `git read-tree` in the nested submodule directory. 3. The second child git process tries to chdir() in the yet unexisting nested submodule directory in start_command() at run-command.c:829 and dies before exec'ing. The reason why check_submodule_move_head() is reached in the first child and not in the main process is that it is inside an if(submodule_from_ce()) construct, and submodule_from_ce() returns a valid struct submodule pointer, whereas it returns a null pointer in the main git process. The reason why submodule_from_ce() returns a null pointer in the main git process is because the call to cache_lookup_path() in config_from() (called from submodule_from_path() in submodule_from_ce()) returns a null pointer since the hashmap "for_path" in the submodule_cache of the_repository is not yet populated. It is not populated because both repo_get_oid(repo, GITMODULES_INDEX, &oid) and repo_get_oid(repo, GITMODULES_HEAD, &oid) in config_from_gitmodules() at submodule-config.c:639-640 return -1, as at this stage of the operation, neither the HEAD of the superproject nor its index contain any .gitmodules file. In contrast, in the first child the hashmap is populated because repo_get_oid(repo, GITMODULES_HEAD, &oid) returns 0 as the HEAD of the first level submodule, i.e. .git/modules/submodule/HEAD, points to a commit where .gitmodules is present and records 'nested' as a submodule. Fix this bug by checking that the submodule directory exists before calling check_submodule_move_head() in merged_entry() in the `if(!old)` branch, i.e. if going from a commit with no submodule to a commit with a submodule present. Also protect the other call to check_submodule_move_head() in merged_entry() the same way as it is safer, even though the `else if (!(old->ce_flags & CE_CONFLICTED))` branch of the code is not at play in the present bug. The other calls to check_submodule_move_head() in other functions in unpack_trees.c are all already protected by calls to lstat() somewhere in the program flow so we don't need additional protection for them. All commands in the unpack_trees machinery are affected, i.e. checkout, reset and read-tree when called with the --recurse-submodules flag. This bug was first reported in [1]. [1] https://lore.kernel.org/git/7437BB59-4605-48EC-B05E-E2BDB2D9DABC@gmail.com/ Reported-by: Philippe Blain <levraiphilippeblain@gmail.com> Reported-by: Damien Robert <damien.olivier.robert@gmail.com> Signed-off-by: Philippe Blain <levraiphilippeblain@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-02-17 05:53:05 +01:00
if (submodule_from_ce(ce) && file_exists(ce->name)) {
int ret = check_submodule_move_head(ce, NULL,
oid_to_hex(&ce->oid),
o);
if (ret)
return ret;
}
"reset --merge": fix unmerged case Commit 9e8ecea (Add 'merge' mode to 'git reset', 2008-12-01) disallowed "git reset --merge" when there was unmerged entries. But it wished if unmerged entries were reset as if --hard (instead of --merge) has been used. This makes sense because all "mergy" operations makes sure that any path involved in the merge does not have local modifications before starting, so resetting such a path away won't lose any information. The previous commit changed the behavior of --merge to accept resetting unmerged entries if they are reset to a different state than HEAD, but it did not reset the changes in the work tree, leaving the conflict markers in the resulting file in the work tree. Fix it by doing three things: - Update the documentation to match the wish of original "reset --merge" better, namely, "An unmerged entry is a sign that the path didn't have any local modification and can be safely resetted to whatever the new HEAD records"; - Update read_index_unmerged(), which reads the index file into the cache while dropping any higher-stage entries down to stage #0, not to copy the object name from the higher stage entry. The code used to take the object name from the a stage entry ("base" if you happened to have stage #1, or "ours" if both sides added, etc.), which essentially meant that you are getting random results depending on what the merge did. The _only_ reason we want to keep a previously unmerged entry in the index at stage #0 is so that we don't forget the fact that we have corresponding file in the work tree in order to be able to remove it when the tree we are resetting to does not have the path. In order to differentiate such an entry from ordinary cache entry, the cache entry added by read_index_unmerged() is marked as CE_CONFLICTED. - Update merged_entry() and deleted_entry() so that they pay attention to cache entries marked as CE_CONFLICTED. They are previously unmerged entries, and the files in the work tree that correspond to them are resetted away by oneway_merge() to the version from the tree we are resetting to. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-01 08:04:04 +01:00
} else if (!(old->ce_flags & CE_CONFLICTED)) {
/*
* See if we can re-use the old CE directly?
* That way we get the uptodate stat info.
*
* This also removes the UPDATE flag on a match; otherwise
* we will end up overwriting local changes in the work tree.
*/
if (same(old, merge)) {
copy_cache_entry(merge, old);
update = 0;
} else {
if (verify_uptodate(old, o)) {
block alloc: add lifecycle APIs for cache_entry structs It has been observed that the time spent loading an index with a large number of entries is partly dominated by malloc() calls. This change is in preparation for using memory pools to reduce the number of malloc() calls made to allocate cahce entries when loading an index. Add an API to allocate and discard cache entries, abstracting the details of managing the memory backing the cache entries. This commit does actually change how memory is managed - this will be done in a later commit in the series. This change makes the distinction between cache entries that are associated with an index and cache entries that are not associated with an index. A main use of cache entries is with an index, and we can optimize the memory management around this. We still have other cases where a cache entry is not persisted with an index, and so we need to handle the "transient" use case as well. To keep the congnitive overhead of managing the cache entries, there will only be a single discard function. This means there must be enough information kept with the cache entry so that we know how to discard them. A summary of the main functions in the API is: make_cache_entry: create cache entry for use in an index. Uses specified parameters to populate cache_entry fields. make_empty_cache_entry: Create an empty cache entry for use in an index. Returns cache entry with empty fields. make_transient_cache_entry: create cache entry that is not used in an index. Uses specified parameters to populate cache_entry fields. make_empty_transient_cache_entry: create cache entry that is not used in an index. Returns cache entry with empty fields. discard_cache_entry: A single function that knows how to discard a cache entry regardless of how it was allocated. Signed-off-by: Jameson Miller <jamill@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-07-02 21:49:31 +02:00
discard_cache_entry(merge);
return -1;
}
/* Migrate old flags over */
update |= old->ce_flags & (CE_SKIP_WORKTREE | CE_NEW_SKIP_WORKTREE);
invalidate_ce_path(old, o);
}
unpack-trees: check for missing submodule directory in merged_entry Using `git checkout --recurse-submodules` to switch between a branch with no submodules and a branch with initialized nested submodules currently causes a fatal error: $ git checkout --recurse-submodules branch-with-nested-submodules fatal: exec '--super-prefix=submodule/nested/': cd to 'nested' failed: No such file or directory error: Submodule 'nested' could not be updated. error: Submodule 'submodule/nested' cannot checkout new HEAD. error: Submodule 'submodule' could not be updated. M submodule Switched to branch 'branch-with-nested-submodules' The checkout succeeds but the worktree and index of the first level submodule are left empty: $ cd submodule $ git -c status.submoduleSummary=1 status HEAD detached at b3ce885 Changes to be committed: (use "git restore --staged <file>..." to unstage) deleted: .gitmodules deleted: first.t deleted: nested fatal: not a git repository: 'nested/.git' Submodule changes to be committed: * nested 1e96f59...0000000: $ git ls-files -s $ # empty $ ls -A .git The reason for the fatal error during the checkout is that a child git process tries to cd into the yet unexisting nested submodule directory. The sequence is the following: 1. The main git process (the one running in the superproject) eventually reaches write_entry() in entry.c, which creates the first level submodule directory and then calls submodule_move_head() in submodule.c, which spawns `git read-tree` in the submodule directory. 2. The first child git process (the one in the submodule of the superproject) eventually calls check_submodule_move_head() at unpack_trees.c:2021, which calls submodule_move_head in dry-run mode, which spawns `git read-tree` in the nested submodule directory. 3. The second child git process tries to chdir() in the yet unexisting nested submodule directory in start_command() at run-command.c:829 and dies before exec'ing. The reason why check_submodule_move_head() is reached in the first child and not in the main process is that it is inside an if(submodule_from_ce()) construct, and submodule_from_ce() returns a valid struct submodule pointer, whereas it returns a null pointer in the main git process. The reason why submodule_from_ce() returns a null pointer in the main git process is because the call to cache_lookup_path() in config_from() (called from submodule_from_path() in submodule_from_ce()) returns a null pointer since the hashmap "for_path" in the submodule_cache of the_repository is not yet populated. It is not populated because both repo_get_oid(repo, GITMODULES_INDEX, &oid) and repo_get_oid(repo, GITMODULES_HEAD, &oid) in config_from_gitmodules() at submodule-config.c:639-640 return -1, as at this stage of the operation, neither the HEAD of the superproject nor its index contain any .gitmodules file. In contrast, in the first child the hashmap is populated because repo_get_oid(repo, GITMODULES_HEAD, &oid) returns 0 as the HEAD of the first level submodule, i.e. .git/modules/submodule/HEAD, points to a commit where .gitmodules is present and records 'nested' as a submodule. Fix this bug by checking that the submodule directory exists before calling check_submodule_move_head() in merged_entry() in the `if(!old)` branch, i.e. if going from a commit with no submodule to a commit with a submodule present. Also protect the other call to check_submodule_move_head() in merged_entry() the same way as it is safer, even though the `else if (!(old->ce_flags & CE_CONFLICTED))` branch of the code is not at play in the present bug. The other calls to check_submodule_move_head() in other functions in unpack_trees.c are all already protected by calls to lstat() somewhere in the program flow so we don't need additional protection for them. All commands in the unpack_trees machinery are affected, i.e. checkout, reset and read-tree when called with the --recurse-submodules flag. This bug was first reported in [1]. [1] https://lore.kernel.org/git/7437BB59-4605-48EC-B05E-E2BDB2D9DABC@gmail.com/ Reported-by: Philippe Blain <levraiphilippeblain@gmail.com> Reported-by: Damien Robert <damien.olivier.robert@gmail.com> Signed-off-by: Philippe Blain <levraiphilippeblain@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2020-02-17 05:53:05 +01:00
if (submodule_from_ce(ce) && file_exists(ce->name)) {
int ret = check_submodule_move_head(ce, oid_to_hex(&old->oid),
oid_to_hex(&ce->oid),
o);
if (ret)
return ret;
}
"reset --merge": fix unmerged case Commit 9e8ecea (Add 'merge' mode to 'git reset', 2008-12-01) disallowed "git reset --merge" when there was unmerged entries. But it wished if unmerged entries were reset as if --hard (instead of --merge) has been used. This makes sense because all "mergy" operations makes sure that any path involved in the merge does not have local modifications before starting, so resetting such a path away won't lose any information. The previous commit changed the behavior of --merge to accept resetting unmerged entries if they are reset to a different state than HEAD, but it did not reset the changes in the work tree, leaving the conflict markers in the resulting file in the work tree. Fix it by doing three things: - Update the documentation to match the wish of original "reset --merge" better, namely, "An unmerged entry is a sign that the path didn't have any local modification and can be safely resetted to whatever the new HEAD records"; - Update read_index_unmerged(), which reads the index file into the cache while dropping any higher-stage entries down to stage #0, not to copy the object name from the higher stage entry. The code used to take the object name from the a stage entry ("base" if you happened to have stage #1, or "ours" if both sides added, etc.), which essentially meant that you are getting random results depending on what the merge did. The _only_ reason we want to keep a previously unmerged entry in the index at stage #0 is so that we don't forget the fact that we have corresponding file in the work tree in order to be able to remove it when the tree we are resetting to does not have the path. In order to differentiate such an entry from ordinary cache entry, the cache entry added by read_index_unmerged() is marked as CE_CONFLICTED. - Update merged_entry() and deleted_entry() so that they pay attention to cache entries marked as CE_CONFLICTED. They are previously unmerged entries, and the files in the work tree that correspond to them are resetted away by oneway_merge() to the version from the tree we are resetting to. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-01 08:04:04 +01:00
} else {
/*
* Previously unmerged entry left as an existence
* marker by read_index_unmerged();
*/
if (verify_absent_if_directory(merge,
ERROR_WOULD_LOSE_UNTRACKED_OVERWRITTEN, o)) {
discard_cache_entry(merge);
return -1;
}
"reset --merge": fix unmerged case Commit 9e8ecea (Add 'merge' mode to 'git reset', 2008-12-01) disallowed "git reset --merge" when there was unmerged entries. But it wished if unmerged entries were reset as if --hard (instead of --merge) has been used. This makes sense because all "mergy" operations makes sure that any path involved in the merge does not have local modifications before starting, so resetting such a path away won't lose any information. The previous commit changed the behavior of --merge to accept resetting unmerged entries if they are reset to a different state than HEAD, but it did not reset the changes in the work tree, leaving the conflict markers in the resulting file in the work tree. Fix it by doing three things: - Update the documentation to match the wish of original "reset --merge" better, namely, "An unmerged entry is a sign that the path didn't have any local modification and can be safely resetted to whatever the new HEAD records"; - Update read_index_unmerged(), which reads the index file into the cache while dropping any higher-stage entries down to stage #0, not to copy the object name from the higher stage entry. The code used to take the object name from the a stage entry ("base" if you happened to have stage #1, or "ours" if both sides added, etc.), which essentially meant that you are getting random results depending on what the merge did. The _only_ reason we want to keep a previously unmerged entry in the index at stage #0 is so that we don't forget the fact that we have corresponding file in the work tree in order to be able to remove it when the tree we are resetting to does not have the path. In order to differentiate such an entry from ordinary cache entry, the cache entry added by read_index_unmerged() is marked as CE_CONFLICTED. - Update merged_entry() and deleted_entry() so that they pay attention to cache entries marked as CE_CONFLICTED. They are previously unmerged entries, and the files in the work tree that correspond to them are resetted away by oneway_merge() to the version from the tree we are resetting to. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-01 08:04:04 +01:00
invalidate_ce_path(old, o);
}
if (do_add_entry(o, merge, update, CE_STAGEMASK) < 0)
return -1;
return 1;
}
read-tree: make two-way merge sparse-aware Enable two-way merge with 'git read-tree' without expanding the sparse index. When in a sparse index, a two-way merge will trivially succeed as long as there are not changes to the same sparse directory in multiple trees (i.e., sparse directory-level "edit-edit" conflicts). If there are such conflicts, the merge will fail despite the possibility that individual files could merge cleanly. In order to resolve these "edit-edit" conflicts, "conflicted" sparse directories are - rather than rejected - merged by traversing their associated trees by OID. For each child of the sparse directory: 1. Files are merged as normal (see Documentation/git-read-tree.txt for details). 2. Subdirectories are treated as sparse directories and merged in 'twoway_merge'. If there are no conflicts, they are merged according to the rules in Documentation/git-read-tree.txt; otherwise, the subdirectory is recursively traversed and merged. This process allows sparse directories to be individually merged at the necessary depth *without* expanding a full index. The 't/t1092-sparse-checkout-compatibility.sh' test 'read-tree --merge with edit/edit conflicts in sparse directories' tests two-way merges with 1) changes inside sparse directories that do not conflict and 2) changes that do conflict (with the correct file(s) reported in the error message). Additionally, add two-way merge cases to 'sparse index is not expanded: read-tree' to confirm that the index is not expanded regardless of whether edit/edit conflicts are present in a sparse directory. Signed-off-by: Victoria Dye <vdye@github.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2022-03-01 21:24:30 +01:00
static int merged_sparse_dir(const struct cache_entry * const *src, int n,
struct unpack_trees_options *o)
{
struct tree_desc t[MAX_UNPACK_TREES + 1];
void * tree_bufs[MAX_UNPACK_TREES + 1];
struct traverse_info info;
int i, ret;
/*
* Create the tree traversal information for traversing into *only* the
* sparse directory.
*/
setup_traverse_info(&info, src[0]->name);
info.fn = unpack_sparse_callback;
info.data = o;
info.show_all_errors = o->show_all_errors;
info.pathspec = o->pathspec;
/* Get the tree descriptors of the sparse directory in each of the merging trees */
for (i = 0; i < n; i++)
tree_bufs[i] = fill_tree_descriptor(o->src_index->repo, &t[i],
src[i] && !is_null_oid(&src[i]->oid) ? &src[i]->oid : NULL);
ret = traverse_trees(o->src_index, n, t, &info);
for (i = 0; i < n; i++)
free(tree_bufs[i]);
return ret;
}
static int deleted_entry(const struct cache_entry *ce,
const struct cache_entry *old,
struct unpack_trees_options *o)
{
/* Did it exist in the index? */
if (!old) {
if (verify_absent(ce, ERROR_WOULD_LOSE_UNTRACKED_REMOVED, o))
return -1;
return 0;
} else if (verify_absent_if_directory(ce, ERROR_WOULD_LOSE_UNTRACKED_REMOVED, o)) {
return -1;
}
"reset --merge": fix unmerged case Commit 9e8ecea (Add 'merge' mode to 'git reset', 2008-12-01) disallowed "git reset --merge" when there was unmerged entries. But it wished if unmerged entries were reset as if --hard (instead of --merge) has been used. This makes sense because all "mergy" operations makes sure that any path involved in the merge does not have local modifications before starting, so resetting such a path away won't lose any information. The previous commit changed the behavior of --merge to accept resetting unmerged entries if they are reset to a different state than HEAD, but it did not reset the changes in the work tree, leaving the conflict markers in the resulting file in the work tree. Fix it by doing three things: - Update the documentation to match the wish of original "reset --merge" better, namely, "An unmerged entry is a sign that the path didn't have any local modification and can be safely resetted to whatever the new HEAD records"; - Update read_index_unmerged(), which reads the index file into the cache while dropping any higher-stage entries down to stage #0, not to copy the object name from the higher stage entry. The code used to take the object name from the a stage entry ("base" if you happened to have stage #1, or "ours" if both sides added, etc.), which essentially meant that you are getting random results depending on what the merge did. The _only_ reason we want to keep a previously unmerged entry in the index at stage #0 is so that we don't forget the fact that we have corresponding file in the work tree in order to be able to remove it when the tree we are resetting to does not have the path. In order to differentiate such an entry from ordinary cache entry, the cache entry added by read_index_unmerged() is marked as CE_CONFLICTED. - Update merged_entry() and deleted_entry() so that they pay attention to cache entries marked as CE_CONFLICTED. They are previously unmerged entries, and the files in the work tree that correspond to them are resetted away by oneway_merge() to the version from the tree we are resetting to. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-01 08:04:04 +01:00
if (!(old->ce_flags & CE_CONFLICTED) && verify_uptodate(old, o))
return -1;
add_entry(o, ce, CE_REMOVE, 0);
invalidate_ce_path(ce, o);
return 1;
}
static int keep_entry(const struct cache_entry *ce,
struct unpack_trees_options *o)
{
add_entry(o, ce, 0, 0);
unpack-trees: add missing cache invalidation Any changes to the output index should be (confusingly) marked in the source index with invalidate_ce_path(). This is used to make sure we still have valid untracked cache and cache-tree extensions in the end. We do a pretty good job of invalidating except in two places. verify_clean_subdirectory() is part of verify_absent() and verify_absent_sparse(). The former is usually called by merged_entry() or directly in threeway_merge(). The latter is obviously used by sparse checkout. In these three call sites, only merged_entry() follows up with invalidate_ce_path(). The other two don't, but they should not trigger this ce removal because this is about D/F conflicts [1]. But let's be safe and invalidate_ce_path() here as well. The second place is keep_entry() which is also used by threeway_merge() to keep higher stage entries. In order to reuse cache-tree we need to invalidate these paths as well. It's not a problem in the past because whenever a higher stage entry is present, cache-tree will not be created [2]. Now we salvage cache-tree even when higher stage entries are present, we need more invalidation. [1] c81935348b (Fix switching to a branch with D/F when current branch has file D. - 2007-03-15) [2] This is probably too strict. We should be able to create and save cache-tree for the directories that do not have conflict entries in cache_tree_update(). And this becomes more important when cache-tree plays bigger role in terms of performance. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2018-08-18 16:41:27 +02:00
if (ce_stage(ce))
invalidate_ce_path(ce, o);
return 1;
}
#if DBRT_DEBUG
static void show_stage_entry(FILE *o,
const char *label, const struct cache_entry *ce)
{
if (!ce)
fprintf(o, "%s (missing)\n", label);
else
fprintf(o, "%s%06o %s %d\t%s\n",
label,
ce->ce_mode,
oid_to_hex(&ce->oid),
ce_stage(ce),
ce->name);
}
#endif
int threeway_merge(const struct cache_entry * const *stages,
struct unpack_trees_options *o)
{
const struct cache_entry *index;
const struct cache_entry *head;
const struct cache_entry *remote = stages[o->head_idx + 1];
int count;
int head_match = 0;
int remote_match = 0;
int df_conflict_head = 0;
int df_conflict_remote = 0;
int any_anc_missing = 0;
int no_anc_exists = 1;
int i;
for (i = 1; i < o->head_idx; i++) {
if (!stages[i] || stages[i] == o->df_conflict_entry)
any_anc_missing = 1;
else
no_anc_exists = 0;
}
index = stages[0];
head = stages[o->head_idx];
if (head == o->df_conflict_entry) {
df_conflict_head = 1;
head = NULL;
}
if (remote == o->df_conflict_entry) {
df_conflict_remote = 1;
remote = NULL;
}
/*
* First, if there's a #16 situation, note that to prevent #13
* and #14.
*/
if (!same(remote, head)) {
for (i = 1; i < o->head_idx; i++) {
if (same(stages[i], head)) {
head_match = i;
}
if (same(stages[i], remote)) {
remote_match = i;
}
}
}
/*
* We start with cases where the index is allowed to match
* something other than the head: #14(ALT) and #2ALT, where it
* is permitted to match the result instead.
*/
/* #14, #14ALT, #2ALT */
if (remote && !df_conflict_head && head_match && !remote_match) {
if (index && !same(index, remote) && !same(index, head)) {
if (S_ISSPARSEDIR(index->ce_mode))
return merged_sparse_dir(stages, 4, o);
else
return reject_merge(index, o);
}
return merged_entry(remote, index, o);
}
/*
* If we have an entry in the index cache, then we want to
* make sure that it matches head.
*/
if (index && !same(index, head)) {
if (S_ISSPARSEDIR(index->ce_mode))
return merged_sparse_dir(stages, 4, o);
else
return reject_merge(index, o);
}
if (head) {
/* #5ALT, #15 */
if (same(head, remote))
return merged_entry(head, index, o);
/* #13, #3ALT */
if (!df_conflict_remote && remote_match && !head_match)
return merged_entry(head, index, o);
}
/* #1 */
if (!head && !remote && any_anc_missing)
return 0;
/*
* Under the "aggressive" rule, we resolve mostly trivial
* cases that we historically had git-merge-one-file resolve.
*/
if (o->aggressive) {
int head_deleted = !head;
int remote_deleted = !remote;
const struct cache_entry *ce = NULL;
if (index)
ce = index;
else if (head)
ce = head;
else if (remote)
ce = remote;
else {
for (i = 1; i < o->head_idx; i++) {
if (stages[i] && stages[i] != o->df_conflict_entry) {
ce = stages[i];
break;
}
}
}
/*
* Deleted in both.
* Deleted in one and unchanged in the other.
*/
if ((head_deleted && remote_deleted) ||
(head_deleted && remote && remote_match) ||
(remote_deleted && head && head_match)) {
if (index)
return deleted_entry(index, index, o);
if (ce && !head_deleted) {
if (verify_absent(ce, ERROR_WOULD_LOSE_UNTRACKED_REMOVED, o))
return -1;
}
return 0;
}
/*
* Added in both, identically.
*/
if (no_anc_exists && head && remote && same(head, remote))
return merged_entry(head, index, o);
}
/* Handle "no merge" cases (see t/t1000-read-tree-m-3way.sh) */
if (index) {
/*
* If we've reached the "no merge" cases and we're merging
* a sparse directory, we may have an "edit/edit" conflict that
* can be resolved by individually merging directory contents.
*/
if (S_ISSPARSEDIR(index->ce_mode))
return merged_sparse_dir(stages, 4, o);
/*
* If we're not merging a sparse directory, ensure the index is
* up-to-date to avoid files getting overwritten with conflict
* resolution files
*/
if (verify_uptodate(index, o))
return -1;
}
o->nontrivial_merge = 1;
/* #2, #3, #4, #6, #7, #9, #10, #11. */
count = 0;
if (!head_match || !remote_match) {
for (i = 1; i < o->head_idx; i++) {
if (stages[i] && stages[i] != o->df_conflict_entry) {
keep_entry(stages[i], o);
count++;
break;
}
}
}
#if DBRT_DEBUG
else {
fprintf(stderr, "read-tree: warning #16 detected\n");
show_stage_entry(stderr, "head ", stages[head_match]);
show_stage_entry(stderr, "remote ", stages[remote_match]);
}
#endif
if (head) { count += keep_entry(head, o); }
if (remote) { count += keep_entry(remote, o); }
return count;
}
/*
* Two-way merge.
*
* The rule is to "carry forward" what is in the index without losing
* information across a "fast-forward", favoring a successful merge
* over a merge failure when it makes sense. For details of the
* "carry forward" rule, please see <Documentation/git-read-tree.txt>.
*
*/
int twoway_merge(const struct cache_entry * const *src,
struct unpack_trees_options *o)
{
const struct cache_entry *current = src[0];
const struct cache_entry *oldtree = src[1];
const struct cache_entry *newtree = src[2];
if (o->merge_size != 2)
return error("Cannot do a twoway merge of %d trees",
o->merge_size);
if (oldtree == o->df_conflict_entry)
oldtree = NULL;
if (newtree == o->df_conflict_entry)
newtree = NULL;
if (current) {
unpack-trees: fix "read-tree -u --reset A B" with conflicted index When we call "read-tree --reset -u HEAD ORIG_HEAD", the first thing we do with the index is to call read_cache_unmerged. Originally that would read the index, leaving aside any unmerged entries. However, as of d1a43f2 (reset --hard/read-tree --reset -u: remove unmerged new paths, 2008-10-15), it actually creates a new cache entry to serve as a placeholder, so that we later know to update the working tree. However, we later noticed that the sha1 of that unmerged entry was just copied from some higher stage, leaving you with random content in the index. That was fixed by e11d7b5 ("reset --merge": fix unmerged case, 2009-12-31), which instead puts the null sha1 into the newly created entry, and sets a CE_CONFLICTED flag. At the same time, it teaches the unpack-trees machinery to pay attention to this flag, so that oneway_merge throws away the current value. However, it did not update the code paths for twoway_merge, which is where we end up in the two-way read-tree with --reset. We notice that the HEAD and ORIG_HEAD versions are the same, and say "oh, we can just reuse the current version". But that's not true. The current version is bogus. Notice this case and make sure we do not keep the bogus entry; either we do not have that path in the tree we are moving to (i.e. remove it), or we want to have the cache entry we created for the tree we are moving to (i.e. resolve by explicitly saying the "newtree" version is what we want). [jc: this is from the almost year-old $gmane/212316] Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-12-29 21:51:54 +01:00
if (current->ce_flags & CE_CONFLICTED) {
if (same(oldtree, newtree) || o->reset) {
if (!newtree)
return deleted_entry(current, current, o);
else
return merged_entry(newtree, current, o);
}
return reject_merge(current, o);
} else if ((!oldtree && !newtree) || /* 4 and 5 */
unpack-trees: fix "read-tree -u --reset A B" with conflicted index When we call "read-tree --reset -u HEAD ORIG_HEAD", the first thing we do with the index is to call read_cache_unmerged. Originally that would read the index, leaving aside any unmerged entries. However, as of d1a43f2 (reset --hard/read-tree --reset -u: remove unmerged new paths, 2008-10-15), it actually creates a new cache entry to serve as a placeholder, so that we later know to update the working tree. However, we later noticed that the sha1 of that unmerged entry was just copied from some higher stage, leaving you with random content in the index. That was fixed by e11d7b5 ("reset --merge": fix unmerged case, 2009-12-31), which instead puts the null sha1 into the newly created entry, and sets a CE_CONFLICTED flag. At the same time, it teaches the unpack-trees machinery to pay attention to this flag, so that oneway_merge throws away the current value. However, it did not update the code paths for twoway_merge, which is where we end up in the two-way read-tree with --reset. We notice that the HEAD and ORIG_HEAD versions are the same, and say "oh, we can just reuse the current version". But that's not true. The current version is bogus. Notice this case and make sure we do not keep the bogus entry; either we do not have that path in the tree we are moving to (i.e. remove it), or we want to have the cache entry we created for the tree we are moving to (i.e. resolve by explicitly saying the "newtree" version is what we want). [jc: this is from the almost year-old $gmane/212316] Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-12-29 21:51:54 +01:00
(!oldtree && newtree &&
same(current, newtree)) || /* 6 and 7 */
(oldtree && newtree &&
same(oldtree, newtree)) || /* 14 and 15 */
(oldtree && newtree &&
!same(oldtree, newtree) && /* 18 and 19 */
same(current, newtree))) {
return keep_entry(current, o);
} else if (oldtree && !newtree && same(current, oldtree)) {
/* 10 or 11 */
return deleted_entry(oldtree, current, o);
} else if (oldtree && newtree &&
same(current, oldtree) && !same(current, newtree)) {
/* 20 or 21 */
return merged_entry(newtree, current, o);
unpack-trees: resolve sparse-directory/file conflicts When running unpack_trees() with a sparse index, we attempt to operate on the index without expanding the sparse directory entries. Thus, we operate by manipulating entire directories and passing them to the unpack function. In the case of the 'git checkout' command, this is the twoway_merge() function. There are several cases in twoway_merge() that handle different situations. One new one to add is the case of a directory/file conflict where the directory is sparse. Before the sparse index, such a conflict would appear as a list of file additions and deletions. Now, twoway_merge() initializes 'current', 'oldtree', and 'newtree' from src[0], src[1], and src[2], then sets 'oldtree' to NULL because it is equal to the df_conflict_entry. The way to determine that we have a directory/file conflict is to test that 'current' and 'newtree' disagree on being sparse directory entries. When we are in this case, we want to resolve the situation by calling merged_entry(). This allows replacing the 'current' entry with the 'newtree' entry. This is important for cases where we want to run 'git checkout' across the conflict and have the new HEAD represent the new file type at that path. The first NEEDSWORK comment dropped in t1092 demonstrates this necessary behavior. However, we still are in a confusing state when 'current' corresponds to a staged change within a sparse directory that is not present at HEAD. This should be atypical, because it requires adding a change outside of the sparse-checkout cone, but it is possible. Since we are unable to determine that this is a staged change within twoway_merge(), we cannot add a case to reject the merge at this point. I believe this is due to the use of df_conflict_entry in the place of 'oldtree' instead of using the valud at HEAD, which would provide some perspective to this decision. Any change that would allow this differentiation for staged entries would need to involve information further up in unpack_trees(). That work should be done, sometime, because we are further confusing the behavior of a directory/file conflict when staging a change in the directory. The two cases 'checkout behaves oddly with df-conflict-?' in t1092 demonstrate that even without a sparse-checkout, Git is not consistent in its behavior. Neither of the two options seems correct, either. This change makes the sparse-index behave differently than the typcial sparse-checkout case, but it does match the full checkout behavior in the df-conflict-2 case. Signed-off-by: Derrick Stolee <dstolee@microsoft.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2021-07-20 22:14:41 +02:00
} else if (current && !oldtree && newtree &&
S_ISSPARSEDIR(current->ce_mode) != S_ISSPARSEDIR(newtree->ce_mode) &&
ce_stage(current) == 0) {
/*
* This case is a directory/file conflict across the sparse-index
* boundary. When we are changing from one path to another via
* 'git checkout', then we want to replace one entry with another
* via merged_entry(). If there are staged changes, then we should
* reject the merge instead.
*/
return merged_entry(newtree, current, o);
read-tree: make two-way merge sparse-aware Enable two-way merge with 'git read-tree' without expanding the sparse index. When in a sparse index, a two-way merge will trivially succeed as long as there are not changes to the same sparse directory in multiple trees (i.e., sparse directory-level "edit-edit" conflicts). If there are such conflicts, the merge will fail despite the possibility that individual files could merge cleanly. In order to resolve these "edit-edit" conflicts, "conflicted" sparse directories are - rather than rejected - merged by traversing their associated trees by OID. For each child of the sparse directory: 1. Files are merged as normal (see Documentation/git-read-tree.txt for details). 2. Subdirectories are treated as sparse directories and merged in 'twoway_merge'. If there are no conflicts, they are merged according to the rules in Documentation/git-read-tree.txt; otherwise, the subdirectory is recursively traversed and merged. This process allows sparse directories to be individually merged at the necessary depth *without* expanding a full index. The 't/t1092-sparse-checkout-compatibility.sh' test 'read-tree --merge with edit/edit conflicts in sparse directories' tests two-way merges with 1) changes inside sparse directories that do not conflict and 2) changes that do conflict (with the correct file(s) reported in the error message). Additionally, add two-way merge cases to 'sparse index is not expanded: read-tree' to confirm that the index is not expanded regardless of whether edit/edit conflicts are present in a sparse directory. Signed-off-by: Victoria Dye <vdye@github.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2022-03-01 21:24:30 +01:00
} else if (S_ISSPARSEDIR(current->ce_mode)) {
/*
* The sparse directories differ, but we don't know whether that's
* because of two different files in the directory being modified
* (can be trivially merged) or if there is a real file conflict.
* Merge the sparse directory by OID to compare file-by-file.
*/
return merged_sparse_dir(src, 3, o);
} else
return reject_merge(current, o);
}
else if (newtree) {
if (oldtree && !o->initial_checkout) {
/*
* deletion of the path was staged;
*/
if (same(oldtree, newtree))
return 1;
return reject_merge(oldtree, o);
}
return merged_entry(newtree, current, o);
}
return deleted_entry(oldtree, current, o);
}
/*
* Bind merge.
*
* Keep the index entries at stage0, collapse stage1 but make sure
* stage0 does not have anything there.
*/
int bind_merge(const struct cache_entry * const *src,
struct unpack_trees_options *o)
{
const struct cache_entry *old = src[0];
const struct cache_entry *a = src[1];
if (o->merge_size != 1)
return error("Cannot do a bind merge of %d trees",
o->merge_size);
if (a && old)
return o->quiet ? -1 :
unpack-trees: support super-prefix option In the future we want to support working tree operations within submodules, e.g. "git checkout --recurse-submodules", which will update the submodule to the commit as recorded in its superproject. In the submodule the unpack-tree operation is carried out as usual, but the reporting to the user needs to prefix any path with the superproject. The mechanism for this is the super-prefix. (see 74866d757, git: make super-prefix option) Add support for the super-prefix option for commands that unpack trees by wrapping any path output in unpacking trees in the newly introduced super_prefixed function. This new function prefixes any path with the super-prefix if there is one. Assuming the submodule case doesn't happen in the majority of the cases, we'd want to have a fast behavior for no super prefix, i.e. no reallocation/copying, but just returning path. Another aspect of introducing the `super_prefixed` function is to consider who owns the memory and if this is the right place where the path gets modified. As the super prefix ought to change the output behavior only and not the actual unpack tree part, it is fine to be that late in the line. As we get passed in 'const char *path', we cannot change the path itself, which means in case of a super prefix we have to copy over the path. We need two static buffers in that function as the error messages contain at most two paths. For testing purposes enable it in read-tree, which has no output of paths other than an unpack-trees.c. These are all converted in this patch. Signed-off-by: Stefan Beller <sbeller@google.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2017-01-18 02:05:20 +01:00
error(ERRORMSG(o, ERROR_BIND_OVERLAP),
read-tree: add "--super-prefix" option, eliminate global The "--super-prefix" option to "git" was initially added in [1] for use with "ls-files"[2], and shortly thereafter "submodule--helper"[3] and "grep"[4]. It wasn't until [5] that "read-tree" made use of it. At the time [5] made sense, but since then we've made "ls-files" recurse in-process in [6], "grep" in [7], and finally "submodule--helper" in the preceding commits. Let's also remove it from "read-tree", which allows us to remove the option to "git" itself. We can do this because the only remaining user of it is the submodule API, which will now invoke "read-tree" with its new "--super-prefix" option. It will only do so when the "submodule_move_head()" function is called. That "submodule_move_head()" function was then only invoked by "read-tree" itself, but now rather than setting an environment variable to pass "--super-prefix" between cmd_read_tree() we: - Set a new "super_prefix" in "struct unpack_trees_options". The "super_prefixed()" function in "unpack-trees.c" added in [5] will now use this, rather than get_super_prefix() looking up the environment variable we set earlier in the same process. - Add the same field to the "struct checkout", which is only needed to ferry the "super_prefix" in the "struct unpack_trees_options" all the way down to the "entry.c" callers of "submodule_move_head()". Those calls which used the super prefix all originated in "cmd_read_tree()". The only other caller is the "unlink_entry()" caller in "builtin/checkout.c", which now passes a "NULL". 1. 74866d75793 (git: make super-prefix option, 2016-10-07) 2. e77aa336f11 (ls-files: optionally recurse into submodules, 2016-10-07) 3. 89c86265576 (submodule helper: support super prefix, 2016-12-08) 4. 0281e487fd9 (grep: optionally recurse into submodules, 2016-12-16) 5. 3d415425c7b (unpack-trees: support super-prefix option, 2017-01-17) 6. 188dce131fa (ls-files: use repository object, 2017-06-22) 7. f9ee2fcdfa0 (grep: recurse in-process using 'struct repository', 2017-08-02) Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2022-12-20 13:39:56 +01:00
super_prefixed(a->name, o->super_prefix),
super_prefixed(old->name, o->super_prefix));
if (!a)
return keep_entry(old, o);
else
return merged_entry(a, NULL, o);
}
/*
* One-way merge.
*
* The rule is:
* - take the stat information from stage0, take the data from stage1
*/
int oneway_merge(const struct cache_entry * const *src,
struct unpack_trees_options *o)
{
const struct cache_entry *old = src[0];
const struct cache_entry *a = src[1];
if (o->merge_size != 1)
return error("Cannot do a oneway merge of %d trees",
o->merge_size);
if (!a || a == o->df_conflict_entry)
return deleted_entry(old, old, o);
if (old && same(old, a)) {
int update = 0;
if (o->reset && o->update && !ce_uptodate(old) && !ce_skip_worktree(old) &&
!(old->ce_flags & CE_FSMONITOR_VALID)) {
struct stat st;
if (lstat(old->name, &st) ||
ie_match_stat(o->src_index, old, &st, CE_MATCH_IGNORE_VALID|CE_MATCH_IGNORE_SKIP_WORKTREE))
update |= CE_UPDATE;
}
if (o->update && S_ISGITLINK(old->ce_mode) &&
should_update_submodules() && !verify_uptodate(old, o))
update |= CE_UPDATE;
add_entry(o, old, update, CE_STAGEMASK);
return 0;
}
return merged_entry(a, old, o);
}
/*
* Merge worktree and untracked entries in a stash entry.
*
* Ignore all index entries. Collapse remaining trees but make sure that they
* don't have any conflicting files.
*/
int stash_worktree_untracked_merge(const struct cache_entry * const *src,
struct unpack_trees_options *o)
{
const struct cache_entry *worktree = src[1];
const struct cache_entry *untracked = src[2];
if (o->merge_size != 2)
BUG("invalid merge_size: %d", o->merge_size);
if (worktree && untracked)
return error(_("worktree and untracked commit have duplicate entries: %s"),
read-tree: add "--super-prefix" option, eliminate global The "--super-prefix" option to "git" was initially added in [1] for use with "ls-files"[2], and shortly thereafter "submodule--helper"[3] and "grep"[4]. It wasn't until [5] that "read-tree" made use of it. At the time [5] made sense, but since then we've made "ls-files" recurse in-process in [6], "grep" in [7], and finally "submodule--helper" in the preceding commits. Let's also remove it from "read-tree", which allows us to remove the option to "git" itself. We can do this because the only remaining user of it is the submodule API, which will now invoke "read-tree" with its new "--super-prefix" option. It will only do so when the "submodule_move_head()" function is called. That "submodule_move_head()" function was then only invoked by "read-tree" itself, but now rather than setting an environment variable to pass "--super-prefix" between cmd_read_tree() we: - Set a new "super_prefix" in "struct unpack_trees_options". The "super_prefixed()" function in "unpack-trees.c" added in [5] will now use this, rather than get_super_prefix() looking up the environment variable we set earlier in the same process. - Add the same field to the "struct checkout", which is only needed to ferry the "super_prefix" in the "struct unpack_trees_options" all the way down to the "entry.c" callers of "submodule_move_head()". Those calls which used the super prefix all originated in "cmd_read_tree()". The only other caller is the "unlink_entry()" caller in "builtin/checkout.c", which now passes a "NULL". 1. 74866d75793 (git: make super-prefix option, 2016-10-07) 2. e77aa336f11 (ls-files: optionally recurse into submodules, 2016-10-07) 3. 89c86265576 (submodule helper: support super prefix, 2016-12-08) 4. 0281e487fd9 (grep: optionally recurse into submodules, 2016-12-16) 5. 3d415425c7b (unpack-trees: support super-prefix option, 2017-01-17) 6. 188dce131fa (ls-files: use repository object, 2017-06-22) 7. f9ee2fcdfa0 (grep: recurse in-process using 'struct repository', 2017-08-02) Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2022-12-20 13:39:56 +01:00
super_prefixed(worktree->name, o->super_prefix));
return merged_entry(worktree ? worktree : untracked, NULL, o);
}