When creating a pack using objects that reside in existing packs, we try
to avoid recomputing futile delta between an object (trg) and a candidate
for its base object (src) if they are stored in the same packfile, and trg
is not recorded as a delta already. This heuristics makes sense because it
is likely that we tried to express trg as a delta based on src but it did
not produce a good delta when we created the existing pack.
As the pack heuristics prefer producing delta to remove data, and Linus's
law dictates that the size of a file grows over time, we tend to record
the newest version of the file as inflated, and older ones as delta
against it.
When creating a thin-pack to transfer recent history, it is likely that we
will try to send an object that is recorded in full, as it is newer. But
the heuristics to avoid recomputing futile delta effectively forbids us
from attempting to express such an object as a delta based on another
object. Sending an object in full is often more expensive than sending a
suboptimal delta based on other objects, and it is even more so if we
could use an object we know the receiving end already has (i.e. preferred
base object) as the delta base.
Tweak the recomputation avoidance logic, so that we do not punt on
computing delta against a preferred base object.
The effect of this change can be seen on two simulated upload-pack
workloads. The first is based on 44 reflog entries from my git.git
origin/master reflog, and represents the packs that kernel.org sent me git
updates for the past month or two. The second workload represents much
larger fetches, going from git's v1.0.0 tag to v1.1.0, then v1.1.0 to
v1.2.0, and so on.
The table below shows the average generated pack size and the average CPU
time consumed for each dataset, both before and after the patch:
dataset
| reflog | tags
---------------------------------
before | 53358 | 2750977
size after | 32398 | 2668479
change | -39% | -3%
---------------------------------
before | 0.18 | 1.12
CPU after | 0.18 | 1.15
change | +0% | +3%
This patch makes a much bigger difference for packs with a shorter slice
of history (since its effect is seen at the boundaries of the pack) though
it has some benefit even for larger packs.
Signed-off-by: Jeff King <peff@peff.net>
Acked-by: Nicolas Pitre <nico@fluxnic.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
It is to give an alternate <name> instead of "origin" to the remote
we are cloning from.
Signed-off-by: Carlos Martín Nieto <cmn@elego.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"abbrev" and "commit_format" in struct rev_info get initialized in
init_revisions - no need to reinit in cmd_log_init_defaults.
Signed-off-by: Michael Schubert <mschub@elegosoft.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
On Solaris systems we'd warn about an implicit cast of mode_t when we
printed things out with the %d format. We'd get this warning under GCC
4.6.0 with Solaris headers:
builtin/init-db.c: In function ‘separate_git_dir’:
builtin/init-db.c:354:4: warning: format ‘%d’ expects argument of type ‘int’, but argument 2 has type ‘mode_t’ [-Wformat]
We've been doing this ever since v1.7.4.1-296-gb57fb80. Just work
around this by adding an explicit cast.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
gcc 4.6.2 (there may be others) does not realize that the variable "mode"
can never be used uninitialized in this function and issues a false warning
under -Wuninitialized option.
Squelch it with an unnecessary initialization; it is not like a single
assignment matters to the performance in this codepath that writes out
to the filesystem with checkout_entry() anyway.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When showing the raw timestamp, we format the numeric
seconds-since-epoch into a buffer, followed by the timezone
string. This string has come straight from the commit
object. A well-formed object should have a timezone string
of only a few bytes, but we could be operating on data
pushed by a malicious user.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When we are determining the list of refs to fetch via
fetch-pack, we have two sets of refs to compare: those on
the remote side, and a "match" list of things we want to
fetch. We iterate through the remote refs alphabetically,
seeing if each one is wanted by the "match" list.
Since def88e9 (Commit first cut at "git-fetch-pack",
2005-07-04), we have used the "path_match" function to do a
suffix match, where a remote ref is considered wanted if
any of the "match" elements is a suffix of the remote
refname.
This enables callers of fetch-pack to specify unqualified
refs and have them matched up with remote refs (e.g., ask
for "A" and get remote's "refs/heads/A"). However, if you
provide a fully qualified ref, then there are corner cases
where we provide the wrong answer. For example, given a
remote with two refs:
refs/foo/refs/heads/master
refs/heads/master
asking for "refs/heads/master" will first match
"refs/foo/refs/heads/master" by the suffix rule, and we will
erroneously fetch it instead of refs/heads/master.
As it turns out, all callers of fetch_pack do provide
fully-qualified refs for the match list. There are two ways
fetch_pack can get match lists:
1. Through the transport code (i.e., via git-fetch)
2. On the command-line of git-fetch-pack
In the first case, we will always be providing the names of
fully-qualified refs from "struct ref" objects. We will have
pre-matched those ref objects already (since we have to
handle more advanced matching, like wildcard refspecs), and
are just providing a list of the refs whose objects we need.
In the second case, users could in theory be providing
non-qualified refs on the command-line. However, the
fetch-pack documentation claims that refs should be fully
qualified (and has always done so since it was written in
2005).
Let's change this path_match call to simply check for string
equality, matching what the callers of fetch_pack are
expecting.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The get_remote_heads function reads the list of remote refs
during git protocol session. It dates all the way back to
def88e9 (Commit first cut at "git-fetch-pack", 2005-07-04).
At that time, the idea was to come up with a list of refs we
were interested in, and then filter the list as we got it
from the remote side.
Later, 1baaae5 (Make maximal use of the remote refs,
2005-10-28) stopped filtering at the get_remote_heads layer,
letting us use the non-matching refs to find common history.
As a result, all callers now simply pass an empty match
list (and any future callers will want to do the same). So
let's drop these now-useless parameters.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Tell the user what this command is intended for, and expand the
description of what it does.
Signed-off-by: Conrad Irwin <conrad.irwin@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When a user asks us to force a mv and overwrite the
destination, we print a warning. However, since a typical
use would be:
$ git mv one two
fatal: destination exists, source=one, destination=two
$ git mv -f one two
warning: overwriting 'two'
this warning is just noise. We already know we're
overwriting; that's why we gave -f!
This patch silences the warning unless "--verbose" is given.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When we try to "git mv" over an existing file, the error
message is fairly informative:
$ git mv one two
fatal: destination exists, source=one, destination=two
When the user forces the overwrite, we give a warning:
$ git mv -f one two
warning: destination exists; will overwrite!
This is less informative, but still sufficient in the simple
rename case, as there is only one rename happening.
But when moving files from one directory to another, it
becomes useless:
$ mkdir three
$ touch one two three/one
$ git add .
$ git mv one two three
fatal: destination exists, source=one, destination=three/one
$ git mv -f one two three
warning: destination exists; will overwrite!
The first message is helpful, but the second one gives us no
clue about what was overwritten. Let's mention the name of
the destination file:
$ git mv -f one two three
warning: overwriting 'three/one'
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Now that "git reset" no longer implicitly removes .git/sequencer that
the operator may or may not have wanted to keep, the logic to write a
backup copy of .git/sequencer and remove it when stale is not needed
any more. Simplify the sequencer API and repository layout by
dropping it.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
As v1.7.8-rc0~141^2~4 (2011-08-04) explains, git cherry-pick removes
the sequencer state just before applying the final patch. In the
single-pick case, that was a good thing, since --abort and --continue
work fine without access to such state and removing it provides a
signal that git should not complain about the need to clobber it ("a
cherry-pick or revert is already in progress") in sequences like the
following:
git cherry-pick foo
git read-tree -m -u HEAD; # forget that; let's try a different one
git cherry-pick bar
After the recent patch "allow single-pick in the middle of cherry-pick
sequence" we don't need that hack any more. In the new regime, a
traditional "git cherry-pick <commit>" command never looks at
.git/sequencer, so we do not need to cripple "git cherry-pick
<commit>..<commit>" for it any more.
So now you can run "git cherry-pick --abort" near the end of a
multi-pick sequence and it will abort the entire sequence, instead of
misbehaving and aborting just the final commit.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
After messing up a difficult conflict resolution in the middle of a
cherry-pick sequence, it can be useful to be able to
git checkout HEAD . && git cherry-pick that-one-commit
to restart the conflict resolution. The current code however errors out
saying that another cherry-pick is already in progress.
Suggested-by: Johannes Sixt <j6t@kdbg.org>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since 7e2bfd3f (revert: allow cherry-picking more than one commit,
2010-07-02), the pick/revert machinery has kept track of the set of
commits to be cherry-picked or reverted using commit_argc and
commit_argv variables, storing the corresponding command-line
parameters.
Future callers as other commands are built in (am, rebase, sequencer)
may find it easier to pass rev-list options to this machinery in
already-parsed form. Teach cmd_cherry_pick and cmd_revert to parse
the rev-list arguments in advance and pass the commit set to
pick_revisions() as a rev_info structure.
Original patch by Jonathan, tweaks and test from Ram.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Improved-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When "git cherry-pick ..bar" encounters conflicts, permit the operator
to use cherry-pick --continue after resolving them as a shortcut for
"git commit && git cherry-pick --continue" to record the resolution
and carry on with the rest of the sequence.
This improves the analogy with "git rebase" (in olden days --continue
was the way to preserve authorship when a rebase encountered
conflicts) and fits well with a general UI goal of making "git cmd
--continue" save humans the trouble of deciding what to do next.
Example: after encountering a conflict from running "git cherry-pick
foo bar baz":
CONFLICT (content): Merge conflict in main.c
error: could not apply f78a8d98c... bar!
hint: after resolving the conflicts, mark the corrected paths
hint: with 'git add <paths>' or 'git rm <paths>'
hint: and commit the result with 'git commit'
We edit main.c to resolve the conflict, mark it acceptable with "git
add main.c", and can run "cherry-pick --continue" to resume the
sequence.
$ git cherry-pick --continue
[editor opens to confirm commit message]
[master 78c8a8c98] bar!
1 files changed, 1 insertions(+), 1 deletions(-)
[master 87ca8798c] baz!
1 files changed, 1 insertions(+), 1 deletions(-)
This is done for both codepaths to pick multiple commits and a single
commit.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This makes pick_revisions() a little shorter and easier to read
straight through.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
If you try to "git mv" multiple files onto another
non-directory file, you confusingly get the "usage" message:
$ touch one two three
$ git add .
$ git mv one two three
usage: git mv [options] <source>... <destination>
[...]
From the user's perspective, that makes no sense. They just
gave parameters that exactly match that usage!
This behavior dates back to the original C version of "git
mv", which had a usage message like:
usage: git mv (<source> <destination> | <source>... <destination>)
This was slightly less confusing, because it at least
mentions that there are two ways to invoke (but it still
isn't clear why what the user provided doesn't work).
Instead, let's show an error message like:
$ git mv one two three
fatal: destination 'three' is not a directory
We could leave the usage message in place, too, but it
doesn't actually help here. It contains no hints that there
are two forms, nor that multi-file form requires that the
endpoint be a directory. So it just becomes useless noise
that distracts from the real error.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The code for a verbose flag has been here since "git mv" was
converted to C many years ago, but actually getting the "-v"
flag from the command line was accidentally lost in the
transition.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When we fetch from a remote, we print a status table like:
From url
* [new branch] foo -> origin/foo
We create this table in a static buffer using sprintf. If
the remote refnames are long, they can overflow this buffer
and smash the stack.
Instead, let's use a strbuf to build the string.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The content level merge machinery ll_merge() is prepared to merge
correctly in "both sides added differently" case by using an empty blob as
if it were the common ancestor. "checkout -m" could do the same, but didn't
bother supporting it and instead insisted on having all three stages.
Reported-by: Pete Harlan
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The comment on top of stripspace() claims that the buffer
will no longer be NUL-terminated. However, this has not been
the case at least since the move to using strbuf in 2007.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When git apply is passed something that is not a patch, it does not produce
an error message or exit with a non-zero status if it was not actually
"applying" the patch i.e. --check or --numstat etc were supplied on the
command line.
Fix this by producing an error when apply fails to find any hunks whatsoever
while parsing the patch.
This will cause some of the output formats (--numstat, --diffstat, etc) to
produce an error when they formerly would have reported zero changes and
exited successfully. That seems like the correct behavior though. Failure
to recognize the input as a patch should be an error.
Plus, add a test.
Reported-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: Brandon Casey <drafnel@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When on master, "git checkout -B master <commit>" is a more natural way to
say "git reset --keep <commit>", which was originally invented for the
exact purpose of moving to the named commit while keeping the local changes
around.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Overwriting the current branch with a different commit is forbidden, as it
will make the status recorded in the index and the working tree out of
sync with respect to the HEAD. There however is no reason to forbid it if
the current branch is renamed to itself, which admittedly is something
only an insane user would do, but is handy for scripts.
Test script is by Conrad Irwin.
Reported-by: Soeren Sonnenburg <sonne@debian.org>
Reported-by: Josh Chia (谢任中)
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Conrad Irwin <conrad.irwin@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Back in 1127148 (Loosen "working file will be lost" check in
Porcelain-ish - 2006-12-04), git-checkout.sh learned to quietly
overwrite ignored files. Howver the code only took .gitignore files
into account.
Standard ignored files include all specified in .gitignore files in
working directory _and_ $GIT_DIR/info/exclude. This patch makes sure
ignored files in info/exclude can also be overwritten automatically in
the spirit of the original patch.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The option --force should not put us in 'create branch' mode. The
fact that this option is only valid in 'create branch' mode is
already caught by the the next 'if' in which we assure that we
are in the correct mode.
Without this patch, "git branch -f" without any other argument ends
up calling create_branch without any branch name.
Signed-off-by: Vincent van Ravesteijn <vfr@lyx.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The "git cherry-pick --abort" command currently renames the
.git/sequencer directory to .git/sequencer-old instead of removing it
on success due to an accident. cherry-pick --abort is designed to
work in three steps:
1) find which commit to roll back to
2) call "git reset --merge <commit>" to move to that commit
3) remove the .git/sequencer directory
But the careless author forgot step 3 entirely. The only reason the
command worked anyway is that "git reset --merge <commit>" renames the
.git/sequencer directory as a secondary effect --- after moving to
<commit>, or so the logic goes, it is unlikely but possible that the
caller of git reset wants to continue the series of cherry-picks that
was in progress, so git renames the sequencer state to
.git/sequencer-old to be helpful while allowing the cherry-pick to be
resumed if the caller did not want to end the sequence after all.
By running "git cherry-pick --abort", the operator has clearly
indicated that she is not planning to continue cherry-picking. Remove
the (renamed) .git/sequencer directory as intended all along.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
On Windows, it is not possible to rename or remove a directory that has
open files. 'revert --abort' renamed .git/sequencer when it still had
.git/sequencer/head open. Close the file as early as possible to allow
the rename operation on Windows.
Signed-off-by: Johannes Sixt <j6t@kdbg.org>
Acked-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This fixes the following warning.
CC builtin/revert.o
builtin/revert.c: In function ‘write_cherry_pick_head’:
builtin/revert.c:311: warning: format not a string literal and no format arguments
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Remove the "git cherry-pick --reset" option, which has a different
preferred spelling nowadays ("--quit"). Luckily the old --reset name
was not around long enough for anyone to get used to it.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
After running some ill-advised command like "git cherry-pick
HEAD..linux-next", the bewildered novice may want to return to more
familiar territory. Introduce a "git cherry-pick --abort" command
that rolls back the entire cherry-pick sequence and places the
repository back on solid ground.
Just like "git merge --abort", this internally uses "git reset
--merge", so local changes not involved in the conflict resolution are
preserved.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When conflicts are encountered while reverting a commit, it can be
handy to have the name of that commit easily available. For example,
to produce a copy of the patch to refer to while resolving conflicts:
$ git revert 2eceb2a8
error: could not revert 2eceb2a8... awesome, buggy feature
$ git show -R REVERT_HEAD >the-patch
$ edit $(git diff --name-only)
Set a REVERT_HEAD pseudoref when "git revert" does not make a commit,
for cases like this. This also makes it possible for scripts to
distinguish between a revert that encountered conflicts and other
sources of an unmerged index.
After successfully committing, resetting with "git reset", or moving
to another commit with "git checkout" or "git reset", the pseudoref is
no longer useful, so remove it.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In the spirit of v1.6.3.3~3^2 (refuse to merge during a merge,
2009-07-01), "git cherry-pick" refuses to start a new cherry-pick when
in the middle of an existing conflicted cherry-pick in the following
sequence:
1. git cherry-pick HEAD..origin
2. resolve conflicts
3. git cherry-pick HEAD..origin (instead of "git cherry-pick
--continue", by mistake)
Good. However, the error message on attempting step 3 is more
convoluted than necessary:
$ git cherry-pick HEAD..origin
error: .git/sequencer already exists.
error: A cherry-pick or revert is in progress.
hint: Use --continue to continue the operation
hint: or --quit to forget about it
fatal: cherry-pick failed
Clarify by removing the redundant first "error:" message, simplifying
the advice, and using lower-case and no full stops to be consistent
with other commands that prefix their messages with "error:", so it
becomes
error: a cherry-pick or revert is already in progress
hint: try "git cherry-pick (--continue | --quit)"
fatal: cherry-pick failed
The "fatal: cherry-pick failed" line seems unnecessary, too, but
that can be fixed some other day.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Deal completely with "cherry-pick --quit" and --continue at the
beginning of pick_revisions(), leaving the rest of the function for
the more interesting "git cherry-pick <commits>" case.
No functional change intended. The impact is just to unindent the
code a little.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The option to "git cherry-pick" and "git revert" to discard the
sequencer state introduced by v1.7.8-rc0~141^2~6 (revert: Introduce
--reset to remove sequencer state, 2011-08-04) has a confusing name.
Change it now, while we still have the time.
The new name for "cherry-pick, please get out of my way, since I've
long forgotten about the sequence of commits I was cherry-picking when
you wrote that old .git/sequencer directory" is --quit. Mnemonic:
this is analagous to quiting a program the user is no longer using ---
we just want to get out of the multiple-command cherry-pick procedure
and not to reset HEAD or rewind any other old state.
The "--reset" option is kept as a synonym to minimize the impact. We
might consider dropping it for simplicity in a separate patch, though.
Adjust documentation and tests to use the newly preferred name (--quit)
instead of --reset. While at it, let's clarify the short descriptions
of these operations in "-h" output.
Before:
--reset forget the current operation
--continue continue the current operation
After:
--quit end revert or cherry-pick sequence
--continue resume revert or cherry-pick sequence
Noticed-by: Phil Hord <phil.hord@gmail.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Because git_path() calls vsnprintf(), code like
fd = open(git_path("SQUASH_MSG"), O_WRONLY | O_CREAT, 0666);
die_errno(_("Could not write to '%s'"), git_path("SQUASH_MSG"));
can end up printing an error indicator from vsnprintf() instead of
open() by mistake. Store the path we are trying to write to in a
temporary variable and pass _that_ to die_errno(), so the messages
written by git cherry-pick/revert and git merge can avoid this source
of confusion.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When --reuse-delta is in effect (which is the default), and an existing
pack in the repository has the same object registered twice (e.g. one copy
in a non-delta format and the other copy in a delta against some other
object), an attempt to repack the repository can result in a cyclic delta
dependency, causing write_one() function to infinitely recurse into
itself.
Detect such a case and break the loopy dependency by writing out an object
that is involved in such a loop in the non-delta format.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When receive-pack & fetch-pack are run and store the pack obtained over
the wire to a local repository, they internally run the index-pack command
with the --strict option. Make sure that we reject incoming packfile that
records objects twice to avoid spreading such a damage.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Some of the fatal messages printed by revert and cherry-pick look ugly
like the following:
fatal: Could not open .git/sequencer/todo.: No such file or directory
The culprit here is that these callers of the die_errno() function did not
take it into account that the message string they give to it is followed
by ": <strerror>", hence the message typically should not end with the
full-stop.
Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2564aa4 started to initialize buf.alloc, but that should actually be one
more byte than the string length due to the trailing \0. Also, do not
modify buf.alloc out of the strbuf code. Use the existing strbuf_attach
instead.
Signed-off-by: Sebastian Schuberth <sschuberth@gmail.com>
Reviewed-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Bad copy-paste.
Otherwise "git remote set-branches" without necessary argument
will result in an error message and help for set-url subcommand.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Both of these free() calls are freeing a "const unsigned char (*)[20]"
type while free() expects a "void *". This results in the following
warning under clang 2.9:
builtin/diff.c:185:7: warning: passing 'const unsigned char (*)[20]' to parameter of type 'void *' discards qualifiers
free(parent);
^~~~~~
submodule.c:394:7: warning: passing 'const unsigned char (*)[20]' to parameter of type 'void *' discards qualifiers
free(parents);
^~~~~~~
This free()-ing without a cast was added by Jim Meyering to
builtin/diff.c in v1.7.6-rc3~4 and later by Fredrik Gustafsson in
submodule.c in v1.7.7-rc1~25^2.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
According to the C standard size_t is always unsigned, therefore the
comparison "n1 < 0 || n2 < 0" when n1 and n2 are size_t will always be
false.
This was raised by clang 2.9 which throws this warning when compiling
apply.c:
builtin/apply.c:253:9: warning: comparison of unsigned expression < 0 is always false [-Wtautological-compare]
if (n1 < 0 || n2 < 0)
~~ ^ ~
builtin/apply.c:253:19: warning: comparison of unsigned expression < 0 is always false [-Wtautological-compare]
if (n1 < 0 || n2 < 0)
~~ ^ ~
This check was originally added in v1.6.5-rc0~53^2 by Giuseppe Bilotta
while adding an option to git-apply to ignore whitespace differences.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When pushing to delete a ref, it uses 0{40} as an object name to signal
that the request is a deletion. We shouldn't trigger "deletion of a
corrupt ref" warning in such a case, which was designed to notice that a
ref points at an object that is truly missing from the repository.
Reported-by: Stefan Näwe
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Modify the option parsing heuristic to handle all -m (rename) cases,
including the no-arg case.
Previously, this "fell through" to the (argc <= 2) case and caused
segfault.
Reported-by: Stefan Näwe <stefan.naewe@atlas-elektronik.com>
Signed-off-by: Tay Ray Chuan <rctay89@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The POSIX-function fork is not supported on Windows. Use our
start_command API instead.
As this is the last call-site that depends on the fork-stub in
compat/mingw.h, remove that as well.
Add an undocumented flag to git-archive that tells it that the
action originated from a remote, so features can be disabled.
Thanks to Jeff King for work on this part.
Remove the NOT_MINGW-prereq for t5000, as git-archive --remote
now works.
Signed-off-by: Erik Faye-Lund <kusmabite@gmail.com>
Helped-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
For a plain string where only the length is known, strbuf.alloc needs to
be initialized to the length. Otherwise strbuf.alloc is 0 and a later
call to strbuf_setlen() will fail.
This bug surfaced when calling git blame under Windows on a *.doc file.
The *.doc file is converted to plain text by antiword via the textconv
mechanism. However, the plain text returned by antiword contains DOS line
endings instead of Unix line endings which triggered the strbuf_setlen()
which previous to this patch failed.
Signed-off-by: Sebastian Schuberth <sschuberth@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Jeff King