The libiconv on Darwin uses the old iconv() interface (2nd argument is a
const char **, instead of a char **). Add OLD_ICONV to the Darwin
variable definitions to handle this.
Signed-off-by: Arjen Laarhoven <arjen@yaph.org> Acked-by: Brian Gernhardt <benji@silverinsanity.com> Signed-off-by: Junio C Hamano <junkio@cox.net>
t5300-pack-object.sh: portability issue using /usr/bin/stat
In the test 'compare delta flavors', /usr/bin/stat is used to get file size.
This isn't portable. There already is a dependency on Perl, use its '-s'
operator to get the file size.
Signed-off-by: Arjen Laarhoven <arjen@yaph.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
git-bisect: allow bisecting with only one bad commit.
This allows you to say:
git bisect start
git bisect bad $bad
git bisect next
to start bisection without knowing a good commit. This would
have you try a commit that is half-way since the beginning of
the history, which is rather wasteful if you already know a good
commit, but if you don't (or your history is short enough that
you do not care), there is no reason not to allow this.
This moves the knowledge about .git/config usage out of refs.c and into
builtin-branch.c instead, which allows git-branch to update HEAD to point
at the moved branch before attempting to update the config file. It also
allows git-branch to exit with an error code if updating the config file
should fail.
Signed-off-by: Lars Hjemli <hjemli@gmail.com> Signed-off-by: Junio C Hamano <junkio@cox.net>
rename_ref(): only print a warning when config-file update fails
If git_config_rename_section() fails, rename_ref() used to return 1, which
left HEAD pointing to an absent refs/heads file (since the actual renaming
had already occurred).
Signed-off-by: Lars Hjemli <hjemli@gmail.com> Signed-off-by: Junio C Hamano <junkio@cox.net>
The number of characters in a line MUST be no more than 998 characters,
and SHOULD be no more than 78 characters (RFC2822).
It is much safer to fold the header by ourselves.
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
cvsimport: Reorder options in documentation for better understanding
The current order the options are documented in makes no sense
at all to me. Reorder them so that similar options are grouped
together and also order them somehwhat by importance.
Signed-off-by: Frank Lichtenheld <frank@lichtenheld.de> Signed-off-by: Junio C Hamano <junkio@cox.net>
Sync both the usage lines in the code and the asciidoc
documentation with the real list of options. While
all options seems to be documented in the asciidoc
document, not all of them were listed in the usage line.
Signed-off-by: Frank Lichtenheld <frank@lichtenheld.de> Signed-off-by: Junio C Hamano <junkio@cox.net>
* maint:
Fix lseek(2) calls with args 2 and 3 swapped
Honor -p<n> when applying git diffs
Fix dependency of common-cmds.h
Fix renaming branch without config file
DESTDIR support for git/contrib/emacs
gitweb: Fix bug in "blobdiff" view for split (e.g. file to symlink) patches
Document --left-right option to rev-list.
Revert "builtin-archive: use RUN_SETUP"
rename contrib/hooks/post-receieve-email to contrib/hooks/post-receive-email.
rerere: make sorting really stable.
Fix t4200-rerere for white-space from "wc -l"
Bisect: teach "bisect start" to optionally use one bad and many good revs.
One bad commit is fundamentally needed for bisect to run,
and if we beforehand know more good commits, we can narrow
the bisect space down without doing the whole tree checkout
every time we give good commits.
git bisect start
git bisect bad $bad
git bisect good $good1 $good2...
On the other hand, there may be some confusion between revs
(<bad> and <good>...) and <pathspec>... if -- is not used
and if an invalid rev or a pathspec that looks like a rev is
given.
Signed-off-by: Christian Couder <chriscool@tuxfamily.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
This function was not called "add_file_to_cache()" only because
an ancient program, update-cache, used that name as an internal
function name that does something slightly different. Now that
is gone, we can take over the better name.
The plan is to name all functions that operate on the default
index xxx_cache(). Later patches create a variant of them that
take an explicit parameter xxx_index(), and then turn
xxx_cache() functions into macros that use "the_index".
Rename internal function "add_file_to_cache" in builtin-update-index.c
I'd like to consistently name all index-layer functions that
operate on the default index xxx_cache(), and this application
specific function interferes with the plan.
Propagate cache error internal to refresh_cache() via parameter.
The function refresh_cache() is the only user of cache_errno
that switches its behaviour based on what internal function
refresh_cache_entry() finds; pass the error status back in a
parameter passed down to it, to get rid of the global variable
cache_errno.
Fix bogus error message from merge-recursive error path
This error message should not usually trigger, but the function
make_cache_entry() called by add_cacheinfo() can return early
without calling into refresh_cache_entry() that sets cache_errno.
Also the error message had a wrong function name reported, and
it did not say anything about which path failed either.
Previously, a binary file in the diffstat would show as:
some-binary-file.bin | Bin
The space after the "Bin" was never used. This patch changes binary
lines in the diffstat to be:
some-binary-file.bin | Bin 12345 -> 123456 bytes
The very nice "->" notation was suggested by Johannes Schindelin, and
shows the before and after sizes more clearly than "+" and "-" would.
If a size is 0 it's not shown (although it would probably be better to
treat no-file differently from zero-byte-file).
The user can see what changed in the binary file, and how big the new
file is. This is in keeping with the information in the rest of the
diffstat.
The diffstat_t members "added" and "deleted" were unused when the file
was binary, so this patch loads them with the file sizes in
builtin_diffstat(). These figures are then read in show_stats() when
the file is marked binary.
Signed-off-by: Andy Parkins <andyparkins@gmail.com> Signed-off-by: Junio C Hamano <junkio@cox.net>
If the user is trying to apply a Git generated diff file and they
have specified a -p<n> option, where <n> is not 1, the user probably
has a good reason for doing this. Such as they are me, trying to
apply a patch generated in git.git for the git-gui subdirectory to
the git-gui.git repository, where there is no git-gui subdirectory
present.
Users shouldn't supply -p2 unless they mean it. But if they are
supplying it, they probably have thought about how to make this
patch apply to their working directory, and want to risk whatever
results may come from that.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
Fix lost-found to show commits only referenced by reflogs
Prior to 1.5.0 the git-lost-found utility was useful to locate
commits that were not referenced by any ref. These were often
amends, or resets, or tips of branches that had been deleted.
Being able to locate a 'lost' commit and recover it by creating a
new branch was a useful feature in those days.
Unfortunately 1.5.0 added the reflogs to the reachability analysis
performed by git-fsck, which means that most commits users would
consider to be lost are still reachable through a reflog. So most
(or all!) commits are reachable, and nothing gets output from
git-lost-found.
Now git-fsck can be told to ignore reflogs during its reachability
analysis, making git-lost-found useful again to locate commits
that are no longer referenced by a ref itself, but may still be
referenced by a reflog.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
clean up and optimize nth_packed_object_sha1() usage
Let's avoid the open coded pack index reference in pack-object and use
nth_packed_object_sha1() instead. This will help encapsulating index
format differences in one place.
And while at it there is no reason to copy SHA1's over and over while a
direct pointer to it in the index will do just fine.
Signed-off-by: Nicolas Pitre <nico@cam.org> Acked-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
Make git_config_rename_section return success if no config file
exists. Otherwise, renaming a branch would abort, leaving the
repository in an inconsistent state.
[jc: test]
Signed-off-by: Geert Bosch <bosch@gnat.com> Signed-off-by: Junio C Hamano <junkio@cox.net>
gitweb: Fix bug in "blobdiff" view for split (e.g. file to symlink) patches
git_patchset_body needs patch generated with --full-index option to
detect split patches, meaning two patches which corresponds to single
difftree (raw diff) entry. An example of such situation is changing
type (mode) of a file, e.g. from plain file to symbolic link.
Add, in git_blobdiff, --full-index option to patch generating git diff
invocation, for the 'html' format output ("blobdiff" view).
"blobdiff_plain" still uses shortened sha1 in the extended git diff
header "index <hash>..<hash>[ <mode>]" line.
Noticed-by: Martin Koegler <mkoegler@auto.tuwien.ac.at> Signed-off-by: Jakub Narebski <jnareb@gmail.com> Signed-off-by: Junio C Hamano <junkio@cox.net>
Commit 64edf4b2 cleaned up the initialization of git-archive,
at the cost of 'git-archive --list' now requiring a git repo.
This patch reverts the cleanup and documents the requirement
for this particular dirtyness in a test.
Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx> Signed-off-by: Junio C Hamano <junkio@cox.net>
The earlier code does not swap hunks when the beginning of the
first side is identical to the whole of the second side. In
such a case, the first one should sort later.
On OS X, wc outputs 6 spaces before the number of lines, so the test
expecting the string "10" failed. Do not quote $cmd to strip away
the problematic whitespace as other tests do.
Also fix the grammar of the test name while making changes to it.
There's only one preimage, so it's "has", not "have".
git-gui: Brown paper bag fix division by 0 in blame
If we generate a blame status string before we have obtained
any annotation data at all from the input file, or if the input
file is empty, our total_lines will be 0. This causes a division
by 0 error when we blindly divide by the 0 to compute the total
percentage of lines loaded. Instead we should report 0% done.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
If a dialog/window has a default button registered not every
platform associates the return key with that button, but all
users do. We have to register the binding of the return key
ourselves to make sure the user's expectations of pressing
return will activate the default button are met.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Do not break git-gui messages into multiple lines.
Many git-gui messages were broken into a multiple lines to make
good paragraph width. Unfortunately in reality it breaks the paragraph
width completely, because the dialog window width does not coincide
with the paragraph width created by the current font.
Tcl/Tk's standard dialog boxes are breaking the long lines
automatically, so it is better to make long lines and let the
interpreter do the job.
Signed-off-by: Eygene Ryabinkin <rea-git@codelabs.ru> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Made the default buttons on the dialog active and focused upon the
dialog appearence.
Bound 'Escape' and 'Return' keys to the dialog dismissal where it
was appropriate: mainly for dialogs with only one button and no
editable fields, but on console output dialogs as well.
Signed-off-by: Eygene Ryabinkin <rea-git@codelabs.ru> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
git-svn: bail out on incorrect command-line options
"git svn log" is the only command that needs the pass-through
option in Getopt::Long; otherwise we will bail out and let the
user know something is wrong.
Also, avoid printing out unaccepted mixed-case options (that
are reserved for the command-line) such as --useSvmProps
in the usage() function.
Signed-off-by: Eric Wong <normalperson@yhbt.net> Signed-off-by: Junio C Hamano <junkio@cox.net>
Fix switching to a branch with D/F when current branch has file D.
This loosens the over-eager verify_absent() check that gets
upset to find directory D in the current working tree when
switching to a branch that has a file there. The check needs to
make sure that we do not lose precious working tree files as a
result of removing directory D and replacing it with the file
from the other branch, which is a tad expensive but this is a
less common case.
Fix twoway_merge that passed d/f conflict marker to merged_entry().
When switching from one tree to another, we should not send a
marker that says "this file does not exist in the new tree -- I
am a placeholder to tell you that, and not a real blob" down to
merged_entry() as the result of the merge.
This variable keeps track of which entry in the original index
the traversal is looking at, and belongs to the unpack_trees_options
structure along with other traversal status information.
unpack_trees.c: pass unpack_trees_options structure to keep_entry() as well.
Other decision functions, deleted_entry() and merged_entry() take one as
their parameter, and this function should. I'll be introducing a separate
index to build the result in, and am planning to pass it as the part of the
structure.
* jc/bisect:
make the previous optimization work also on path-limited rev-list --bisect
rev-list --bisect: Fix "halfway" optimization.
t6004: add a bit more path optimization test.
git-rev-list --bisect: optimization
git-rev-list: add --bisect-vars option.
t6002: minor spelling fix.
* fl/doc:
Documentation: unbreak user-manual.
Documentation: Add version information to man pages
Documentation: Replace @@GIT_VERSION@@ in documentation
* post1.5.1/p4:
Added correct Python path to the RPM specfile.
Remove unused WITH_OWN_SUBPROCESS_PY from RPM spec
Added git-p4 package to the list of git RPMs.
Add the WITH_P4IMPORT knob to the Makefile.
This corrects the interface mistake of the previous one, and
gives a command line parameter to the only plumbing command that
currently needs it: "git-read-tree".
We can add the calls to set_alternate_index_output() to other
plumbing commands that update the index if/when needed.
_GIT_INDEX_OUTPUT: allow plumbing to output to an alternative index file.
When defined, this allows plumbing commands that update the
index (add, apply, checkout-index, merge-recursive, mv,
read-tree, rm, update-index, and write-tree) to write their
resulting index to an alternative index file while holding a
lock to the original index file. With this, git-commit that
jumps the index does not have to make an extra copy of the index
file, and more importantly, it can do the update while holding
the lock on the index.
However, I think the interface to let an environment variable
specify the output is a mistake, as shown in the documentation.
If a curious user has the environment variable set to something
other than the file GIT_INDEX_FILE points at, almost everything
will break. This should instead be a command line parameter to
tell these plumbing commands to write the result in the named
file, to prevent stupid mistakes.
Not that this release really matters, as we will be doing
1.5.1 tomorrow. This commit is to tie the loose ends and
merge all of "maint" branch into "master" in preparation.
Rewrite Tcl/Tk interpreter path for the GUI tools.
--with-tcltk=/path/to/wish sets the TCLTK_PATH variable that is
used to substitute the location of the wish interpreter in the
Tcl/Tk programs.
New tracking file, GIT-GUI-VARS, was introduced: it tracks the
location of the Tcl/Tk interpreter and activates the GUI tools
rebuild if the interpreter path was changed. The separate tracker
is better than the GIT-CFLAGS: there is no need to rebuild the whole
git if the interpreter path was changed.
Remove unused WITH_OWN_SUBPROCESS_PY from RPM spec
We don't have a copy of subprocess.py anymore, so we removed that
option from the Makefile. Let's not leave that cruft around the RPM
spec file either.
git-blame.el: pick a set of random colors for each git-blame turn
I thought it would be cool to have different set of colors for each
git-blame-mode. Function `git-blame-new-commit' does this for us
picking when possible, a random colors based on the set we build on
startup. When it fails, `git-blame-ancient-color' will be used. We
also take care not to use the same color more than once (thank you
David Kågedal, really).
* Prevent (future possible) namespace clash by renaming `color-scale'
into `git-blame-color-scale'. Definition has been changed to be more
in the "lisp" way (thanks for help to #emacs). Also added a small
description of what it does.
* Added docstrings at some point and instructed defvar when a variable
was candidate to customisation by users.
* Added missing defvar to silent byte-compilers (git-blame-file,
git-blame-current)
* Do not require 'cl at startup
* Added more informations on compatibility
Signed-off-by: Xavier Maillard <zedek@gnu.org> Acked-by: David Kågedal <davidk@lysator.liu.se> Signed-off-by: Junio C Hamano <junkio@cox.net>
git-blame.el: separate git-blame-mode to ease maintenance
git-blame-mode has been splitted into git-blame-mode-on and
git-blame-mode-off; it now conditionnaly calls one of them depending
of how we call it. Code is now easier to maintain and to understand.
Fixed `git-reblame' function: interactive form was at the wrong
place.
String displayed on the mode line is now configurable through
`git-blame-mode-line-string` (default to " blame").
Signed-off-by: Xavier Maillard <zedek@gnu.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
The way things are set up, you can now pass a "pathspec" to the
"read_directory()" function. If you pass NULL, it acts exactly
like it used to do (read everything). If you pass a non-NULL
pointer, it will simplify it into a "these are the prefixes
without any special characters", and stop any readdir() early if
the path in question doesn't match any of the prefixes.
NOTE! This does *not* obviate the need for the caller to do the *exact*
pathspec match later. It's a first-level filter on "read_directory()", but
it does not do the full pathspec thing. Maybe it should. But in the
meantime, builtin-add.c really does need to do first
read_directory(dir, .., pathspec);
if (pathspec)
prune_directory(dir, pathspec, baselen);
ie the "prune_directory()" part will do the *exact* pathspec pruning,
while the "read_directory()" will use the pathspec just to do some quick
high-level pruning of the directories it will recurse into.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
cvsserver: Don't lie about binary mode in asciidoc documentation
The git-cvsserver documentation claims that the server will set
-k modes if appropriate which is not really the case. On the other
hand the available gitcvs.allbinary variable is not documented at
all. Fix both these issues by rewording the related paragraph.
Signed-off-by: Frank Lichtenheld <frank@lichtenheld.de> Signed-off-by: Junio C Hamano <junkio@cox.net>
Keep rename/rename conflicts of intermediate merges while doing recursive merge
This patch leaves the base name in the resulting intermediate tree, to
propagate the conflict from intermediate merges up to the top-level merge.
Signed-off-by: Alex Riesen <raa.lkml@gmail.com> Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
contrib/workdir: add a simple script to create a working directory
Add a simple script to create a working directory that uses symlinks
to point at an exisiting repository. This allows having different
branches in different working directories but all from the same
repository.
Based on a description from Junio of how he creates multiple working
directories[1]. With the following caveat:
"This risks confusion for an uninitiated if you update a ref that
is checked out in another working tree, but modulo that caveat
it works reasonably well."
Reimplement emailing part of hooks--update in contrib/hooks/post-receive-email
The update hook is no longer the correct place to generate emails; there
is now the hooks/post-receive script which is run automatically after a
ref has been updated.
This patch is to make use of that new location, and to address some
faults in the old update hook.
The primary problem in the conversion was that in the update hook, the
ref has not actually been changed, but is about to be. In the
post-receive hook the ref has already been updated. That meant that
where we previously had lines like:
git rev-list --not --all
would now give the wrong list because "--all" in the post-receive hook
includes the ref that we are making the email for. This made it more
difficult to show only the new revisions added by this update.
The solution is not pretty; however it does work and doesn't need any
changes to git-rev-list itself. It also fixes (more accurately: reduces
the likelihood of) a nasty race when another update occurs while this
script is running. The solution, in short, looks like this (see the
source code for a longer explanation)
This uses git-rev-parse followed by grep to filter out the revision of
the ref in question before it gets to rev-list and inhibits the output
of itself. By using $(git rev-parse $revname) rather than $newrev as the
filter, it also takes care of the situation where another update to the
same ref has been made since $refname was $newrev.
The second problem that is addressed is that of tags inhibiting the
correct output of an update email. Consider this, with somebranch and
sometag pointing at the same revision:
That would work fine; the push of the branch would generate an email
containing all the new commits introduced by the update, then the push
of the tag would generate the shortlog formatted tag email. Now
consider:
When some branch comes to run its "--not --all" line, it will find
sometag, and filter those commits from the email - leaving nothing.
That meant that those commits would not show (in full) on any email.
The fix is to not use "--all", and instead use "--branches" in the
git-rev-parse command.
Other changes
* Lose the monstrous one-giant-script layout and put things in easy to
digest functions. This makes it much easier to find the place you
need to change if you wanted to customise the output. I've also
tried to write more verbose comments for the same reason. The hook
script is big, mainly because of all the different cases that it has
to handle, so being easy to navigate is important.
* All uses of "git-command" changed to "git command", to cope better
if a user decided not to install all the hard links to git;
* Cleaned up some of the English in the email
* The fact that the receive hook makes the ref available also allows me
to use Shawn Pearce's fantastic suggestion that an annotated tag can
be parsed with git-for-each-ref. This removes the potentially
non-portable use of "<<<" heredocs and the nasty messing around with
"date" to convert numbers of seconds UTC to a real date
* Deletions are now caught and notified (briefly)
* To help with debugging, I've retained the command line mode from the
update hook; but made it so that the output is not emailed, it's just
printed to the screen. This could then be redirected if the user
wanted
* Removed the "Hello" from the beginning of the email - it's just
noise, and no one seriously has their day made happier by "friendly"
programs
* The fact that it doesn't rely on repository state as an indicator any
more means that it's far more stable in its output; hopefully the
same arguments will always generate the same email - even if the
repository changes in the future. This means you can easily recreate
an email should you want to.
* Included Jim Meyering's envelope sender option for the sendmail call
* The hook is now so big that it was inappropriate to copy it
to every repository by keeping it in the templates directory.
Instead, I've put a comment saying to look in contrib/hooks, and
given an example of calling the script from that template hook. The
advantage of calling the script residing at some fixed location is
that if a future package of git included a bug fixed version of the
script, that would be picked up automatically, and the user would not
have to notice and manually copy the new hook to every repository
that uses it.
Signed-off-by: Andy Parkins <andyparkins@gmail.com> Signed-off-by: Junio C Hamano <junkio@cox.net>
git-svn: avoid respewing similar error messages for missing paths
We ignore errors if the path we're tracking did not exist for
a particular revision range, but we still print out warnings
telling the user about that.
As pointed out by Seth Falcon, this amounts to a lot of warnings
that could confuse and worry users. I'm not entirely comfortable
completely silencing the warnings, but showing one warning per
path that we track should be reasonable.
Signed-off-by: Eric Wong <normalperson@yhbt.net> Signed-off-by: Junio C Hamano <junkio@cox.net>
Rename warn() to warning() to fix symbol conflicts on BSD and Mac OS
This fixes a problem reported by Randal Schwartz:
>I finally tracked down all the (albeit inconsequential) errors I was getting
>on both OpenBSD and OSX. It's the warn() function in usage.c. There's
>warn(3) in BSD-style distros. It'd take a "great rename" to change it, but if
>someone with better C skills than I have could do that, my linker and I would
>appreciate it.
It was annoying to me, too, when I was doing some mergetool testing on
Mac OS X, so here's a fix.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Cc: "Randal L. Schwartz" <merlyn@stonehenge.com> Signed-off-by: Junio C Hamano <junkio@cox.net>
Don't translate the patch to UTF-8, instead preserve the data as
is. This also reverts a test case that was included in the
original patch series.
Also allow overwriting the authorship and title information we
gather from RFC2822 mail headers with additional in-body
headers, which was pointed out by Linus.
Signed-off-by: Don Zickus <dzickus@redhat.com> Signed-off-by: Junio C Hamano <junkio@cox.net>
gitweb: Support comparing blobs (files) with different names
Fix the bug that caused "blobdiff" view called with new style URI
for a rename with change diff to be show as new (added) file diff.
New style URI for "blobdiff" for rename means with $hash_base ('hb') and
$hash_parent_base ('hpb') paramaters denoting tree-ish (usually commit)
of a blobs being compared, together with both $file_name ('f') and
$file_parent ('fp') parameters.
It is done by adding $file_parent ('fp') to the path limiter, meaning
that diff command becomes:
git diff-tree [options] hpb hb -- fp f
Other option would be finding hash of a blob using git_get_hash_by_path
subroutine and comparing blobs using git-diff, or using extended SHA-1
syntax and compare blobs using git-diff:
git diff [options] hpb:fp hp:f
Signed-off-by: Jakub Narebski <jnareb@gmail.com> Signed-off-by: Junio C Hamano <junkio@cox.net>
* maint:
git-upload-pack: make sure we close unused pipe ends
Documentation/git-rev-parse.txt: fix example in SPECIFYING RANGES.
Documentation/git-svnimport.txt: fix typo.
Most bourne-ish shells I have here accept
x=$((echo x)|cat)
but all bourne-ish shells I have here accept
x=$( (echo x)|cat)
because $(( might mean arithmetic expansion.
Signed-off-by: Francis Daly <francis@daoine.org> Signed-off-by: Junio C Hamano <junkio@cox.net>