This test had multiple issues causing it to fail for the wrong
reason(s):
* rename/rename(1to2) conflicts have always left the original source
path present in the working directory and index (at stage 1). Thus,
the triple rename/rename(1to2) should result in 9 unstaged files,
not 6.
* It messed up the three-way content merge for checking the results of
merging for one of the renames, accidentally turning it into a
two-way merge.
* It got the contents of the base files it was using to compare
against wrong, due to an off-by-one error, and overwrite-redirection
('>') instead of append-redirection ('>>').
* It used slightly too-long conflict markers
* It didn't include filenames in the conflict marker hunks (granted,
that was a shortcoming of the merge-recursive backend for rename/add
and rename/rename(2to1) conflicts, but since it's
test_expect_failure anyway we might as well make it expect our
preferred behavior rather than some compromise that we can't yet
reach anyway).
Fix these issues so that a merge backend which correctly handles these
kinds of nested conflicts will pass the test.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Commit da1e295e00 ("t604[236]: do not run setup in separate tests",
2019-10-22) removed approximately half the tests (which were setup-only
tests) in t6043 by turning them into functions that the subsequent test
would call as their first step. This ensured that any test from this
file could be run entirely independently of all the other tests in the
file. Unfortunately, the call to the new setup function was missed in
two of the test_expect_failure cases. Add them in.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Apparently I don't know how to count untracked files, and since the
tests in question were marked as test_expect_failure, no one ever
noticed it until now. Correct the count, as these tests clearly create
three untracked files ('out', 'err', and 'file_count').
(I believe this problem arose because earlier incarnations counted lines
via a pipe to 'wc -l'. Reviewers asked that it be replaced by writing
the output to a file and using test_line_count, but when the temporary
output was added to a separate file, the count of untracked files should
have increased.)
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The testcase only required that the merge complete without conflict,
without specifying what the correct resolution was. Since normalization
changed this from a modify/delete to a not-modified/delete, the correct
resolution is to have the file be removed at the end. Add a check for
this resolution.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The tests for the merge machinery are spread over several places.
Collect them into t64xx for simplicity. Some notes:
t60[234]*.sh:
Merge tests started in t602*, overgrew bisect and remote tracking
tests in t6030, t6040, and t6041, and nearly overtook replace tests
in t6050. This made picking out relevant tests that I wanted to run
in a tighter loop slightly more annoying for years.
t303*.sh:
These started out as tests for the 'merge-recursive' toplevel command,
but did not restrict to that and had lots of overlap with the
underlying merge machinery.
t7405, t7613:
submodule-specific merge logic started out in submodule.c but was
moved to merge-recursive.c in commit 18cfc08866 ("submodule.c: move
submodule merging to merge-recursive.c", 2018-05-15). Since these
tests are about the logic found in the merge machinery, moving these
tests to be with the merge tests makes sense.
t7607, t7609:
Having tests spread all over the place makes it more likely that
additional tests related to a certain piece of logic grow in all those
other places. Much like t303*.sh, these two tests were about the
underlying merge machinery rather than outer levels.
Tests that were NOT moved:
t76[01]*.sh:
Other than the four tests mentioned above, the remaining tests in
t76[01]*.sh are related to non-recursive merge strategies, parameter
parsing, and other stuff associated with the highlevel builtin/merge.c
rather than the recursive merge machinery.
t3[45]*.sh:
The rebase testcases in t34*.sh also test the merge logic pretty
heavily; sometimes changes I make only trigger failures in the rebase
tests. The rebase tests are already nicely coupled together, though,
and I didn't want to mess that up. Similar comments apply for the
cherry-pick tests in t35*.sh.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Update "git help guides" documentation organization.
* pb/guide-docs:
git.txt: add list of guides
Documentation: don't hardcode command categories twice
help: drop usage of 'common' and 'useful' for guides
command-list.txt: add missing 'gitcredentials' and 'gitremote-helpers'
All "mergy" operations that internally use the merge-recursive
machinery should honor the merge.renormalize configuration, but
many of them didn't.
* en/eol-attrs-gotchas:
checkout: support renormalization with checkout -m <paths>
merge: make merge.renormalize work for all uses of merge machinery
t6038: remove problematic test
t6038: make tests fail for the right reason
Small fixes and workarounds.
* jk/compiler-fixes-and-workarounds:
revision: avoid leak when preparing bloom filter for "/"
revision: avoid out-of-bounds read/write on empty pathspec
config: work around gcc-10 -Wstringop-overflow warning
Adjust tests in contrib/ to the recent change to fmt-merge-msg.
* es/adjust-subtree-test-for-merge-msg-update:
Revert "contrib: subtree: adjust test to change in fmt-merge-msg"
Code cleanup around "worktree" API implementation.
* es/worktree-cleanup:
worktree: retire special-case normalization of main worktree path
worktree: drop bogus and unnecessary path munging
worktree: drop unused code from get_linked_worktree()
worktree: drop pointless strbuf_release()
The argv_array API is useful for not just managing argv but any
"vector" (NULL-terminated array) of strings, and has seen adoption
to a certain degree. It has been renamed to "strvec" to reduce the
barrier to adoption.
* jk/strvec:
strvec: rename struct fields
strvec: drop argv_array compatibility layer
strvec: update documention to avoid argv_array
strvec: fix indentation in renamed calls
strvec: convert remaining callers away from argv_array name
strvec: convert more callers away from argv_array name
strvec: convert builtin/ callers away from argv_array name
quote: rename sq_dequote_to_argv_array to mention strvec
strvec: rename files from argv-array to strvec
argv-array: rename to strvec
argv-array: use size_t for count and alloc
Not all man5/man7 guides are mentioned in the 'git(1)' documentation,
which makes the missing ones somewhat hard to find.
Add a list of the guides to git(1) by leveraging the existing
`Documentation/cmd-list.perl` script to generate a file `cmds-guide.txt`
which gets included in git.txt.
Also, do not hard-code the manual section '1'. Instead, use a regex so
that the manual section is discovered from the first line of each
`git*.txt` file.
This addition was hinted at in 1b81d8cb19 (help: use command-list.txt
for the source of guides, 2018-05-20).
Helped-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Philippe Blain <levraiphilippeblain@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Instead of hard-coding the list of command categories in both
`Documentation/Makefile` and `Documentation/cmd-list.perl`, make the
Makefile the authoritative source and tweak `cmd-list.perl` so that it
receives the list of command categories as argument.
Signed-off-by: Philippe Blain <levraiphilippeblain@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since 1b81d8cb19 (help: use command-list.txt for the source of guides,
2018-05-20), all man5/man7 guides listed in command-list.txt appear in
the output of 'git help -g'.
However, 'git help -g' still prefixes this list with "The common Git
guides are:", which makes one wonder if there are others!
In the same spirit, the man page for 'git help' describes the '--guides'
option as listing 'useful' guides, which is not false per se but can
also be taken to mean that there are other guides that exist but are not
useful.
Instead of 'common' and 'useful', use 'Git concept guides' in both
places. To keep the code in line with this change, rename
help.c::list_common_guides_help to list_guides_help.
Signed-off-by: Philippe Blain <levraiphilippeblain@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The guides 'gitcredentials' and 'gitremote-helpers' do not currently
appear in command-list.txt.
'gitcredentials' was forgotten back when guides were added to
command-list.txt in 1b81d8cb19 (help: use command-list.txt for the
source of guides, 2018-05-20).
'gitremote-helpers' was moved to section 7 in 439cc74632 (docs: move
gitremote-helpers into section 7, 2019-03-25), but command-list.txt was
not updated at the time.
Add these two guides to the list of guides in 'command-list.txt', so
that they appear in the output of 'git help --guides', and capitalize
the first word of the description of 'gitcredentials', as was done in
1b81d8c (help: use command-list.txt for the source of guides,
2018-05-20) for the other guides.
While at it, add a comment in Documentation/Makefile to remind developers
to update command-list.txt if they add a new guide.
Signed-off-by: Philippe Blain <levraiphilippeblain@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Get rid of the trailing dot and mark for translation.
Signed-off-by: Sergey Organov <sorganov@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The pretend-object mechanism checks if the given object already
exists in the object store before deciding to keep the data
in-core, but the check would have triggered lazy fetching of such
an object from a promissor remote.
* jt/pretend-object-never-come-from-elsewhere:
sha1-file: make pretend_object_file() not prefetch
While packing many objects in a repository with a promissor remote,
lazily fetching missing objects from the promissor remote one by
one may be inefficient---the code now attempts to fetch all the
missing objects in batch (obviously this won't work for a lazy
clone that lazily fetches tree objects as you cannot even enumerate
what blobs are missing until you learn which trees are missing).
* jt/pack-objects-prefetch-in-batch:
pack-objects: prefetch objects to be packed
pack-objects: refactor to oid_object_info_extended
If we're given an empty pathspec, we refuse to set up bloom filters, as
described in f3c2a36810 (revision: empty pathspecs should not use Bloom
filters, 2020-07-01).
But before the empty string check, we drop any trailing slash by
allocating a new string without it. So a pathspec consisting only of "/"
will allocate that string, but then still cause us to bail, leaking the
new string. Let's make sure to free it.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Running t4216 with ASan results in it complaining of an out-of-bounds
read in prepare_to_use_bloom_filter(). The issue is this code to strip a
trailing slash:
last_index = pi->len - 1;
if (pi->match[last_index] == '/') {
because we have no guarantee that pi->len isn't zero. This can happen if
the pathspec is ".", as we translate that to an empty string. And if
that read of random memory does trigger the conditional, we'd then do an
out-of-bounds write:
path_alloc = xstrdup(pi->match);
path_alloc[last_index] = '\0';
Let's make sure to check the length before subtracting. Note that for an
empty pathspec, we'd end up bailing from the function a few lines later,
which makes it tempting to just:
if (!pi->len)
return;
early here. But our code here is stripping a trailing slash, and we need
to check for emptiness after stripping that slash, too. So we'd have two
blocks, which would require repeating some cleanup code.
Instead, just skip the trailing-slash for an empty string. Setting
last_index at all in the case is awkward since it will have a nonsense
value (and it uses an "int", which is a too-small type for a string
anyway). So while we're here, let's:
- drop last_index entirely; it's only used in two spots right next to
each other and writing out "pi->len - 1" in both is actually easier
to follow
- use xmemdupz() to duplicate the string. This is slightly more
efficient, but more importantly makes the intent more clear by
allocating the correct-sized substring in the first place. It also
eliminates any question of whether path_alloc is as long as
pi->match (which it would not be if pi->match has any embedded NULs,
though in practice this is probably impossible).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Compiling with gcc-10, -O2, and -fsanitize=undefined results in a
compiler warning:
config.c: In function ‘git_config_copy_or_rename_section_in_file’:
config.c:3170:17: error: writing 1 byte into a region of size 0 [-Werror=stringop-overflow=]
3170 | output[0] = '\t';
| ~~~~~~~~~~^~~~~~
config.c:3076:7: note: at offset -1 to object ‘buf’ with size 1024 declared here
3076 | char buf[1024];
| ^~~
This is a false positive. The interesting lines of code are:
int i;
char *output = buf;
...
for (i = 0; buf[i] && isspace(buf[i]); i++)
; /* do nothing */
...
int offset;
offset = section_name_match(&buf[i], old_name);
if (offset > 0) {
...
output += offset + i;
if (strlen(output) > 0) {
/*
* More content means there's
* a declaration to put on the
* next line; indent with a
* tab
*/
output -= 1;
output[0] = '\t';
}
}
So we do assign output to buf initially. Later we increment it based on
"offset" and "i" and then subtract "1" from it. That latter step is what
the compiler is complaining about; it could lead to going off the left
side of the array if "output == buf" at the moment of the subtraction.
For that to be the case, then "offset + i" would have to be 0. But that
can't happen:
- we know that "offset" is at least 1, since we're in a conditional
block that checks that
- we know that "i" is not negative, since it started at 0 and only
incremented over whitespace
So the sum must be at least 1, and therefore it's OK to subtract one
from "output".
But that's not quite the whole story. Since "i" is an int, it could in
theory be possible to overflow to negative (when counting whitespace on
a very large string). But we know that's impossible because we're
counting the 1024-byte buffer we just fed to fgets(), so it can never be
larger than that.
Switching the type of "i" to "unsigned" makes the warning go away, so
let's do that.
Arguably size_t is an even better type (for this and for the other
length fields), but switching to it produces a similar but distinct
warning:
config.c: In function ‘git_config_copy_or_rename_section_in_file’:
config.c:3170:13: error: array subscript -1 is outside array bounds of ‘char[1024]’ [-Werror=array-bounds]
3170 | output[0] = '\t';
| ~~~~~~^~~
config.c:3076:7: note: while referencing ‘buf’
3076 | char buf[1024];
| ^~~
If we were to ever switch off of fgets() to strbuf_getline() or similar,
we'd probably need to use size_t to avoid other overflow problems. But
for now we know we're safe because of the small fixed size of our
buffer.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This reverts commit 508fd8e8ba.
In 6e6029a8 (fmt-merge-msg: allow merge destination to be omitted again)
we get back the behavior where merges against 'master', by default, do
not include "into 'master'" at the end of the merge message. This test
fix is no longer needed.
Signed-off-by: Emily Shaffer <emilyshaffer@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Make it clear that the filename has only the rest of the object ID,
not the entirety of it.
Signed-off-by: Noam Yorav-Raphael <noamraph@gmail.com>
Acked-by: Taylor Blau <me@ttaylorr.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The 'merge' command is not the only one that does merges; other commands
like checkout -m or rebase do as well. Unfortunately, the only area of
the code that checked for the "merge.renormalize" config setting was in
builtin/merge.c, meaning it could only affect merges performed by the
"merge" command. Move the handling of this config setting to
merge_recursive_config() so that other commands can benefit from it as
well. Fixes a few tests in t6038.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
t6038.11, 'cherry-pick patch from after text=auto' was a test of
undefined behavior. To make matters worse, while there are a couple
possible correct answers, this test was coded to only check for an
obviously incorrect answer. And the final cherry on top is that the
test is marked test_expect_failure, meaning it can't provide much value,
other than possibly confusing future folks who come along and try to
work on attributes and look at existing tests. Because of all these
problems, just remove the test.
But for any future code spelunkers, here's my understanding of the two
possible correct answers:
This test was set up so that on a branch with no .gitattributes file,
you cherry-picked a patch from a branch that had a .gitattributes file
(containing '* text=auto'). Further, the two branches had a file which
differed only in line endings. In this situation, correct behavior is
not well defined: should the .gitattributes file affect the merge or
not?
If the .gitattributes file on the other branch should not affect the
merge, then we would have a content conflict with all three stages
different (the merge base didn't match either side).
If the .gitattributes file from the other branch should affect the
merge, then we would expect the line endings to be normalized to LF for
the version to be recorded in the repository. This would mean that when
doing a three-way content merge on the file that differed in line
endings, that the three-way content merge would see that the versions on
both sides matched and so the cherry-pick has no conflicts and can
succeed. The line endings in the file as recorded in the repository
will change from CRLF to LF. The version checked out in the working
copy will depend on the platform (since there's no eol attribute defined
for the file).
Also, as a final side note, this test expected an error message that was
built assuming cherry-pick was the old scripted version, because
cherry-pick no longer uses the error message that was encoded in this
test. So it was wrong for yet another reason.
Given that the handling of .gitattributes is not well defined and this
test was obviously broken and could do nothing but confuse future
readers, just remove it.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
t6038 had a pair of tests that were expected to fail, but weren't
failing for the expected reason. Both were meant to do a merge that
could be done cleanly after renormalization, but were supposed to fail
for lack of renormalization. Unfortunately, both tests had staged
changes, and checkout -m would abort due to the presence of those staged
changes before even attempting a merge.
Fix this first issue by utilizing git-restore instead of git-checkout,
so that the index is left alone and just the working directory gets the
changes we want.
However, there is a second issue with these tests. Technically, they
just wanted to verify that after renormalization, no conflicts would be
present. This could have been checked for by grepping for a lack of
conflict markers, but the test instead tried to compare the working
directory files to an expected result. Unfortunately, the setting of
"text=auto" without setting core.eol to any value meant that the content
of the file (in particular, the line endings) would be
platform-dependent and the tests could only pass on some platforms.
Replace the existing comparison with a call to 'git diff --no-index
--ignore-cr-at-eol' to verify that the contents, other than possible
carriage returns in the file, match the expected results and in
particular that the file has no conflicts from the checkout -m
operation.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Write the hexadecimal object ID directly into the destination buffer
using oid_to_hex_r() instead of writing it into a static buffer first
using oid_to_hex() and then copying it from there using memcpy().
This is shorter, simpler and a bit more efficient.
Reviewed-by: brian m. carlson <sandals@crustytoothpaste.net>
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Commits 7c0a6c8e47 ("merge-recursive: move some definitions around to
clean up the header", 2019-08-17), and b4db8a2b76 ("merge-recursive:
remove useless parameter in merge_trees()", 2019-08-17) added some
useful documentation to the functions, but had a few places where the
new comments were unclear or even misleading. Fix those comments.
Signed-off-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We use
printf '\0'
to generate a NUL byte which we then `dd` into the packfile to ensure
that we modify the first byte of the first object, thereby
(probabilistically) invalidating the checksum. Except the single quotes
we're using are interpreted to match with the ones we enclose the whole
test in. So we actually execute
printf \0
and end up injecting the ASCII code for "0", 0x30, instead.
The comment right above this `printf` invocation says that "at least one
of [the type bits] is not zero, so setting the first byte to 0 is
sufficient". Substituting "0x30" for "0" in that comment won't do: we'd
need to reason about which bits go where and just what the packfile
looks like that we're modifying in this test.
Let's avoid all of that by actually executing
printf "\0"
to generate a NUL byte, as intended.
Signed-off-by: Martin Ågren <martin.agren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git merge" learned to selectively omit " into <branch>" at the end
of the title of default merge message with merge.suppressDest
configuration.
* jc/fmt-merge-msg-suppress-destination:
fmt-merge-msg: allow merge destination to be omitted again
Revert "fmt-merge-msg: stop treating `master` specially"
In order for "git-worktree list" to present consistent results,
get_main_worktree() performs manual normalization on the repository
path (returned by get_common_dir()) after passing it through
strbuf_add_absolute_path(). In particular, it cleans up the path for
three distinct cases when the current working directory is (1) the main
worktree, (2) the .git/ subdirectory, or (3) a bare repository.
The need for such special-cases is a direct consequence of employing
strbuf_add_absolute_path() which, for the sake of efficiency, doesn't
bother normalizing the path (such as folding out redundant path
components) after making it absolute. Lack of normalization is not
typically a problem since redundant path elements make no difference
when working with paths at the filesystem level. However, when preparing
paths for presentation, possible redundant path components make it
difficult to ensure consistency.
Eliminate the need for these special cases by instead making the path
absolute via strbuf_add_real_path() which normalizes the path for us.
Once normalized, the only case we need to handle manually is converting
it to the path of the main worktree by stripping the "/.git" suffix.
This stripping of the "/.git" suffix is a regular idiom in
worktree-related code; for instance, it is employed by
get_linked_worktree(), as well.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The content of .git/worktrees/<id>/gitdir must be a path of the form
"/path/to/worktree/.git". Any other content would be indicative of a
corrupt "gitdir" file. To determine the path of the worktree itself one
merely strips the "/.git" suffix, and this is indeed how the worktree
path was determined from inception.
However, 5193490442 (worktree: add a function to get worktree details,
2015-10-08) extended the path manipulation in a mysterious way. If it is
unable to strip "/.git" from the path, then it instead reports the
current working directory as the linked worktree's path:
if (!strbuf_strip_suffix(&worktree_path, "/.git")) {
strbuf_reset(&worktree_path);
strbuf_add_absolute_path(&worktree_path, ".");
strbuf_strip_suffix(&worktree_path, "/.");
}
This logic is clearly bogus; it can never be generally correct behavior.
It materialized out of thin air in 5193490442 with neither explanation
nor tests to illustrate a case in which it would be desirable.
It's possible that this logic was introduced to somehow deal with a
corrupt "gitdir" file, so that it returns _some_ sort of meaningful
value, but returning the current working directory is not helpful. In
fact, it is quite misleading (except in the one specific case when the
current directory is the worktree whose "gitdir" entry is corrupt).
Moreover, reporting the corrupt value to the user, rather than fibbing
about it and hiding it outright, is more helpful since it may aid in
diagnosing the problem.
Therefore, drop this bogus path munging and restore the logic to the
original behavior of merely stripping "/.git".
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This code has been unused since fa099d2322 (worktree.c: kill parse_ref()
in favor of refs_resolve_ref_unsafe(), 2017-04-24), so drop it.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The content of this strbuf is unconditionally detached several lines
before the strbuf_release() and the strbuf is never touched again after
that point.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
b6839fda68 (ref-filter: add support for %(contents:size), 2020-07-16)
added a new format for ref-filter, and added a function to generate
tests for this new feature in t6300. Unfortunately, it tries to run
`test_expect_sucess' instead of `test_expect_success', and writes
$expect to `expected', but tries to read `expect'. Those two issues
were probably unnoticed because the script only printed errors, but did
not crash. This fixes these issues.
Signed-off-by: Alban Gruin <alban.gruin@gmail.com>
Reviewed-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
523fa69c (reflog: cleanse messages in the refs.c layer, 2020-07-10)
centralized reflog normalizaton. However, the normalizaton added a
leading "\t" to the message. This is an artifact of the reflog
storage format in the files backend, so it should be added there.
Routines that parse back the reflog (such as grab_nth_branch_switch)
expect the "\t" to not be in the message, so without this fix, git
with reftable cannot process the "@{-1}" syntax.
Signed-off-by: Han-Wen Nienhuys <hanwen@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>