Earlier we made "rev-list --object-edge" more aggressively list the
objects at the edge commits, in order to reduce number of objects
fetched into a shallow repository, but the change affected cases
other than "fetching into a shallow repository" and made it
unusably slow (e.g. fetching into a normal repository should not
have to suffer the overhead from extra processing). Limit it to a
more specific case by introducing --objects-edge-aggressive, a new
option to rev-list.
* bc/fetch-thin-less-aggressive-in-normal-repository:
pack-objects: use --objects-edge-aggressive for shallow repos
rev-list: add an option to mark fewer edges as uninteresting
Documentation: add missing article in rev-list-options.txt
"git checkout-index --temp=$target $path" did not work correctly
for paths outside the current subdirectory in the project.
* es/checkout-index-temp:
checkout-index: fix --temp relative path mangling
t2004: demonstrate broken relative path printing
t2004: standardize file naming in symlink test
t2004: drop unnecessary write-tree/read-tree
t2004: modernize style
The get_merge_bases*() API was easy to misuse by careless
copy&paste coders, leaving object flags tainted in the commits that
needed to be traversed.
* jc/merge-bases:
get_merge_bases(): always clean-up object flags
bisect: clean flags after checking merge bases
checkout-index --temp only properly prints relative paths which are
descendants of the current directory. Paths in ancestor or sibling
directories (or their children) are often printed in mangled form. For
example:
mkdir a bbb &&
>file &&
>bbb/file &&
git update-index --add file bbb/file &&
cd a &&
git checkout-index --temp ../file ../bbb/file
prints:
.merge_file_ooblek le
.merge_file_igloo0 b/file
rather than the correct:
.merge_file_ooblek ../file
.merge_file_igloo0 ../bbb/file
Internally, given the above example, checkout-index prefixes each input
argument with the name of the current directory ("a/", in this case),
and then assumes that it can simply skip forward by strlen("a/") bytes
to recover the original name. This works for files in the current
directory or its descendants, but fails for files in ancestors or
siblings (or their children) due to path normalization.
For instance, given "../file", "a/" is prepended, giving "a/../file".
Path normalization folds out "a/../", resulting in "file". Attempting
to recover the original name by skipping strlen("a/") bytes gives the
incorrect "le" rather than the desired "../file".
Fix this by taking advantage of write_name_quoted_relative() to recover
the original name properly, rather than assuming that it can be
recovered by skipping strlen(prefix) bytes.
As a bonus, this also fixes a bug in which checkout-index --temp
accessed and printed memory beyond the end-of-string. For instance,
within a subdirectory named "subdirectory", and given argument
"../file", prefixing would give "subdirectory/../file", which would
become "file" after normalization. checkout-index would then attempt to
recover the original name by skipping strlen("subdirectory/") bytes of
"file", which placed it well beyond end-of-string. Despite this error,
it often appeared to give the correct result, but only due to an
accident of implementation which left an apparently correct copy of the
path in memory following the normalized value. In particular, handed
"subdirectory/../file", in-place processing by normalize_path_copy_len()
resulted in "file\0rectory/../file". When checkout-index skipped
strlen("subdirectory/") bytes, it ended up back at "../file" and thus
appeared to give the correct answer, despite being past end-of-string.
Reported-by: Russ Cox <rsc@golang.org>
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When fetching into or pushing from a shallow repository, we want to
aggressively mark edges as uninteresting, since this decreases the pack
size. However, aggressively marking edges can negatively affect
performance on large non-shallow repositories with lots of refs.
Teach pack-objects a --shallow option to indicate that we're pushing
from or fetching into a shallow repository. Use
--objects-edge-aggressive only for shallow repositories and otherwise
use --objects-edge, which performs better in the general case. Update
the callers to pass the --shallow option when they are dealing with a
shallow repository.
Signed-off-by: brian m. carlson <sandals@crustytoothpaste.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In commit fbd4a70 (list-objects: mark more commits as edges in
mark_edges_uninteresting - 2013-08-16), we marked an increasing number
of edges uninteresting. This change, and the subsequent change to make
this conditional on --objects-edge, are used by --thin to make much
smaller packs for shallow clones.
Unfortunately, they cause a significant performance regression when
pushing non-shallow clones with lots of refs (23.322 seconds vs.
4.785 seconds with 22400 refs). Add an option to git rev-list,
--objects-edge-aggressive, that preserves this more aggressive behavior,
while leaving --objects-edge to provide more performant behavior.
Preserve the current behavior for the moment by using the aggressive
option.
Signed-off-by: brian m. carlson <sandals@crustytoothpaste.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git update-ref --stdin"'s verify command did not work well when
<oldvalue>, which is documented as optional, was missing.
* mh/update-ref-verify:
update-ref: fix "verify" command with missing <oldvalue>
t1400: add some more tests of "update-ref --stdin"'s verify command
New tag object format validation added in 2.2 showed garbage
after a tagname it reported in its error message.
* js/fsck-tag-validation:
index-pack: terminate object buffers with NUL
fsck: properly bound "invalid tag name" error message
"git branch -d" (delete) and "git branch -m" (move) learned to
honor "-f" (force) flag; unlike many other subcommands, the way to
force these have been with separate "-D/-M" options, which was
inconsistent.
* mg/branch-d-m-f:
branch: allow -f with -m and -d
t3200-branch: test -M
"git ls-tree" does not support path selection based on negative
pathspecs, but did not error out when negative pathspecs are given.
* nd/ls-tree-pathspec:
t3102: style modernization
t3102: document that ls-tree does not yet support negated pathspec
ls-tree: disable negative pathspec because it's not supported
ls-tree: remove path filtering logic in show_tree
tree.c: update read_tree_recursive callback to pass strbuf as base
"git push" into a repository with a working tree normally refuses
to modify the branch that is checked out. The command learned to
optionally do an equivalent of "git reset --hard" only when there
is no change to the working tree and the index instead, which would
be useful to "deploy" by pushing into a repository.
* js/push-to-deploy:
t5516: more tests for receive.denyCurrentBranch=updateInstead
receive-pack: add another option for receive.denyCurrentBranch
The function sometimes returned a non-freeable memory and some
other times returned a piece of memory that must be freed.
* jc/exec-cmd-system-path-leak-fix:
system_path(): always return free'able memory to the caller
"git am" learned "--message-id" option to copy the message ID of
the incoming e-mail to the log message of resulting commit.
* pb/am-message-id-footer:
git-am: add --message-id/--no-message-id
git-mailinfo: add --message-id
"git remote update --prune" to drop many refs has been optimized.
* mh/simplify-repack-without-refs:
sort_string_list(): rename to string_list_sort()
prune_remote(): iterate using for_each_string_list_item()
prune_remote(): rename local variable
repack_without_refs(): make the refnames argument a string_list
prune_remote(): sort delete_refs_list references en masse
prune_remote(): initialize both delete_refs lists in a single loop
prune_remote(): exit early if there are no stale references
Some filesystems assign filemodes in a strange way, fooling then
automatic "filemode trustability" check done during a new
repository creation.
* tb/config-core-filemode-check-on-broken-fs:
init-db: improve the filemode trustability check
"git interpret-trailers" learned to properly handle the
"Conflicts:" block at the end.
* cc/interpret-trailers-more:
trailer: add test with an old style conflict block
trailer: reuse ignore_non_trailer() to ignore conflict lines
commit: make ignore_non_trailer() non static
merge & sequencer: turn "Conflicts:" hint into a comment
builtin/commit.c: extract ignore_non_trailer() helper function
merge & sequencer: unify codepaths that write "Conflicts:" hint
builtin/merge.c: drop a parameter that is never used
Git 2.0 was supposed to make the "simple" mode for the default of
"git push", but it didn't.
* jk/push-simple:
push: truly use "simple" as default, not "upstream"
"git init" (hence "git clone") initialized the per-repository
configuration file .git/config with x-bit by mistake.
* mh/config-flip-xbit-back-after-checking:
create_default_files(): don't set u+x bit on $GIT_DIR/config
"git config --get-color" did not parse its command line arguments
carefully.
* jk/colors-fix:
t4026: test "normal" color
config: fix parsing of "git config --get-color some.key -1"
docs: describe ANSI 256-color mode
"git checkout $treeish $path", when $path in the index and the
working tree already matched what is in $treeish at the $path,
still overwrote the $path unnecessarily.
* jk/checkout-from-tree:
checkout $tree: do not throw away unchanged index entries
Git 2.0 was supposed to make the "simple" mode for the default of
"git push", but it didn't.
* jk/push-simple:
push: truly use "simple" as default, not "upstream"
Call strbuf_complete_line() instead of open-coding it. Also remove
surrounding comments indicating the intent to complete a line since
this information is already included in the function name.
Signed-off-by: Rene Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
To figure out the author ident for a commit, we call
determine_author_info(). This function collects information
from the environment, other commits (in the case of
"--amend" or "-c/-C"), and the "--author" option. It then
uses fmt_ident to generate the final ident string that goes
into the commit object. fmt_ident is therefore responsible
for any quality or validation checks on what is allowed to
go into a commit.
Before returning, though, we call split_ident_line on the
result, and feed the individual components to hooks via the
GIT_AUTHOR_* variables. Furthermore, we do extra validation
by feeding the split to sane_ident_split(), which is pickier
than fmt_ident (in particular, it will complain about an empty
email field). If this parsing or validation fails, we skip
updating the environment variables.
This is bad, because it means that hooks may silently see a
different ident than what we are putting into the commit. We
should drop the extra sane_ident_split checks entirely, and
take whatever fmt_ident has fed us (and what will go into
the commit object).
If parsing fails, we should actually abort here rather than
continuing (and feeding the hooks bogus data). However,
split_ident_line should never fail here. The ident was just
generated by fmt_ident, so we know that it's sane. We can
use assert_split_ident to double-check this.
Note that we also teach that assertion to check that we
found a date (it always should, but until now, no caller
cared whether we found a date or not). Checking the return
value of sane_ident_split is enough to ensure we have the
name/email pointers set, and checking date_begin is enough
to know that all of the date/tz variables are set.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When we generate the commit-message template, we try to
report an author or committer ident that will be of interest
to the user: an author that does not match the committer, or
a committer that was auto-configured.
When doing so, if we encounter what we consider to be a
bogus ident, we immediately die. This is a bad idea, because
our use of the idents here is purely informational. Any
ident rules should be enforced elsewhere, because commits
that do not invoke the editor will not even hit this code
path (e.g., "git commit -mfoo" would work, but "git commit"
would not). So at best, we are redundant with other checks,
and at worse, we actively prevent commits that should
otherwise be allowed.
We should therefore do the minimal parsing we can to get a
value and not do any validation (i.e., drop the call to
sane_ident_split()).
In theory we could notice when even our minimal parsing
fails to work, and do the sane thing for each check (e.g.,
if we have an author but can't parse the committer, assume
they are different and print the author). But we can
actually simplify this even further.
We know that the author and committer strings we are parsing
have been generated by us earlier in the program, and
therefore they must be parseable. We could just call
split_ident_line without even checking its return value,
knowing that it will put _something_ in the name/mail
fields. Of course, to protect ourselves against future
changes to the code, it makes sense to turn this into an
assert, so we are not surprised if our assumption fails.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
If "git update-ref --stdin" was given a "verify" command with no
"<newvalue>" at all (not even zeros), the code was mistakenly setting
have_old=0 (and leaving old_sha1 uninitialized). But this is
incorrect: this command is supposed to verify that the reference
doesn't exist. So in this case we really need old_sha1 to be set to
null_sha1 and have_old to be set to 1.
Moreover, since have_old was being set to zero, *no* check of the old
value was being done, so the new value of the reference was being set
unconditionally to the value in new_sha1. new_sha1, in turn, was set
to null_sha1 in the expectation that that was the old value and it
shouldn't be changed. But because the precondition was not being
checked, the result was that the reference was being deleted
unconditionally.
So, if <oldvalue> is missing, set have_old unconditionally and set
old_sha1 to null_sha1.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Acked-by: Brad King <brad.king@kitware.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
-f/--force is the standard way to force an action, and is used by branch
for the recreation of existing branches, but not for deleting unmerged
branches nor for renaming to an existing branch.
Make "-m -f" equivalent to "-M" and "-d -f" equivalent to" -D", i.e.
allow -f/--force to be used with -m/-d also.
For the list modes, "-f" is simply ignored.
Signed-off-by: Michael J Gruber <git@drmicha.warpmail.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We have some tricky checks in fsck that rely on a side effect of
require_end_of_header(), and would otherwise easily run outside
non-NUL-terminated buffers. This is a bit brittle, so let's make sure
that only NUL-terminated buffers are passed around to begin with.
Jeff "Peff" King contributed the detailed analysis which call paths are
involved and pointed out that we also have to patch the get_data()
function in unpack-objects.c, which is what Johannes "Dscho" Schindelin
implemented.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Analyzed-by: Jeff King <peff@peff.net>
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A request to store an empty note via "git notes" meant to remove
note from the object but with --allow-empty we will store a (surprise!)
note that is empty. In the longer run, we might want to deprecate
the somewhat unintuitive "emptying means deletion" behaviour.
* jh/empty-notes:
t3301: modernize style
notes: empty notes should be shown by 'git log'
builtin/notes: add --allow-empty, to allow storing empty notes
builtin/notes: split create_note() to clarify add vs. remove logic
builtin/notes: simplify early exit code in add()
builtin/notes: refactor note file path into struct note_data
builtin/notes: improve naming
t3301: verify that 'git notes' removes empty notes by default
builtin/notes: fix premature failure when trying to add the empty blob
"git checkout $treeish $path", when $path in the index and the
working tree already matched what is in $treeish at the $path,
still overwrote the $path unnecessarily.
* jk/checkout-from-tree:
checkout $tree: do not throw away unchanged index entries
ls-tree uses read_tree_recursive() which already does path filtering
using pathspec. No need to filter one more time based on prefix
only. "ls-tree ../somewhere" does not work because of
this. write_name_quotedpfx() can now be retired because nobody else
uses it.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This allows the callback to use 'base' as a temporary buffer to
quickly assemble full path "without" extra allocation. The callback
has to restore it afterwards of course.
Helped-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
