This applies to 'maint' to fix a rather serious data corruption
issue. When "git add -u" affects a subdirectory in such a way
that the only changes to its contents are path removals, the
next tree object written out of that index was bogus, as the
remove codepath forgot to invalidate the cache-tree entry.
Reported by Salikh Zakirov.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We attempt to remove directory that becomes empty after removal
of a file. We should do the same when we rename an existing
file away.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
If you have a working tree _file_ "foo", attempt to refer to a
branch "foo/bar" without -- to disambiguate, like this:
$ git log foo/bar
tried to make sure that foo/bar cannot be naming a working tree
file "foo/bar" (in which case we would say "which one do you
want? A rev or a working tree file? clarify with -- please").
We run lstat("foo/bar") to check that. If it does not succeed,
there is no ambiguity.
That is good. But we also checked the error status for the
lstat() and expected it to fail with ENOENT. In this particular
case, however, it fails with ENOTDIR. That should be treated as
"expected error" as well.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
For the html output we can use a stylesheet to make sure that the
listingblocks are presented in a monospaced font. For the manpages do
it manually by inserting a ".ft C" before and ".ft" after the block in
question.
In order for these roff commands to get through to the manpage they
have to be element encoded to prevent quoting.
Signed-off-by: Julian Phillips <julian@quantumfyre.co.uk>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When "git checkout-index" checks out path A/B/C, it makes sure A
and A/B are truly directories; if there is a regular file or
symlink at A, we prefer to remove it.
We used to do this by catching an error return from mkdir(2),
and on EEXIST did unlink(2), and when it succeeded, tried
another mkdir(2).
Thomas Glanzmann found out the above does not work on Solaris
for a root user, as unlink(2) was so old fashioned there that it
allowed to unlink a directory.
As pointed out, this still doesn't guarantee that git won't call
"unlink()" on a directory (race conditions etc), but that's
fundamentally true (there is no "funlink()" like there is
"fstat()"), and besides, that is in no way git-specific (ie it's
true of any application that gets run as root).
Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Earlier in 16a4c61, we taught "read-tree -m -u" not to be
confused when switching from a branch that has a path frotz/filfre
to another branch that has a symlink frotz that points at xyzzy/
directory. The fix was incomplete in that it was still confused
when coming back (i.e. switching from a branch with frotz -> xyzzy/
to another branch with frotz/filfre).
This fix is rather expensive in that for a path that is created
we would need to see if any of the leading component of that
path exists as a symbolic link in the filesystem (in which case,
we know that path itself does not exist, and the fact we already
decided to check it out tells us that in the index we already
know that symbolic link is going away as there is no D/F
conflict).
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* 'maint' of git://repo.or.cz/git-gui:
git-gui: Work around bad interaction between Tcl and cmd.exe on ^{tree}
git-gui: Don't linewrap within console windows
git-gui: Correct ls-tree buffering problem in browser
git-gui: Skip nicknames when selecting author initials
git-gui: Ensure windows shortcuts always have .bat extension
git-gui: Include a Push action on the left toolbar
git-gui: Bind M1-P to push action
git-gui: Don't bind F5/M1-R in all windows
git-gui: Unlock the index when cancelling merge dialog
git-gui: properly popup error if gitk should be started but is not installed
From Johannes Sixt <J.Sixt@eudaptics.com>:
> It seems that MSYS's wish does some quoting for Bourne shells,
> in particular, escape the first '{' of the "^{tree}" suffix, but
> then it uses cmd.exe to run "git rev-parse". However, cmd.exe does
> not remove the backslash, so that the resulting rev expression
> ends up in git's guts as unrecognizable garbage: rev-parse fails,
> and git-gui hickups in a way that it must be restarted.
Johannes originally submitted a patch to this section of commit.tcl
to use `git rev-parse $PARENT:`, but not all versions of Git will
accept that format. So I'm just taking the really simple approach
here of scanning the first line of the commit to grab its tree.
About the same cost, but works everywhere.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
The 'D' subcommand within a commit can also delete a directory
recursively. This wasn't clear in the prior version of the
documentation, leading to a question on the mailing list.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
If we get more than 80 characters of text in a single line odds
are it is output from git-fetch or git-push and its showing a
lot of detail off to the right edge that is not so important to
the average user. We still want to make sure we show everything
we need, but we can get away with that information being off to
the side with a horizontal scrollbar.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Our file browser was showing bad output as it did not properly buffer
a partial record when read from `ls-tree -z`. This did not show up on
my Mac OS X system as most trees are small, the pipe buffers generally
big and `ls-tree -z` was generally fast enough that all data was ready
before Tcl started to read. However on my Cygwin system one of my
production repositories had a large enough tree and packfile that it
took a couple of pipe buffers for `ls-tree -z` to complete its dump.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Our blame viewer only grabbed the first initial of the git.git
author string "Simon 'corecode' Schubert". Here the problem was we
looked at Simon, pulled the S into the author initials, then saw
the single quote as the start of the next name and did not like
this character as it was not an uppercase letter.
We now skip over single quoted nicknames placed within the author
name field and grab the initials following it. So the above name
will get the initials SS, rather than just S.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Recently a user on the mailing list complained that they'd read the
manual but couldn't figure out how to keep a couple private repositories
in sync. They'd tried using push, and were surprised by the effect.
Add a little text in an attempt to make it clear that:
- Pushing to a branch that is checked out will have odd results.
- It's OK to synchronize just using pull if that's simpler.
Signed-off-by: "J. Bruce Fields" <bfields@citi.umich.edu>
- "method of" is vulgar, "method for" is nicer
- "recovery" becomes "recovering" from Steve Hoelzer's original version
of this patch
- "if you want" is nicer as "if you wish"
- "you may" should be "you can"; "you may" is "you have permission to"
rather than "you can"'s "it is possible to"
Signed-off-by: Andy Parkins <andyparkins@gmail.com>
Signed-off-by: "J. Bruce Fields" <bfields@citi.umich.edu>
"git apply" used to take check the whitespace in the wrong
direction.
Noticed by Daniel Barkalow.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Apparently under some setups on Windows Tk is hiding our file
extension recommendation of ".bat" from the user and that is
allowing the user to create a shortcut file which has no file
extension. Double clicking on such a file in Windows Explorer
brings up the associate file dialog, as Windows does not know
what application to launch.
We now append the file extension ".bat" to the filename of the
shortcut file if it has no extension or if it has one but it is
not ".bat".
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Pushing changes to a remote system is a very common action for
many users of git-gui, so much so that in some workflows a user
is supposed to push immediately after they make a local commit
so that their change(s) are immediately available for their
teammates to view and build on top of.
Including the push button right below the commit button on the
left toolbar indicates that users should probably perform this
action after they have performed the commit action.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Users often need to be able to push the current branch so that they
can publish their recent changes to anyone they are collaborating
with on the project. Associating a keyboard action with this will
make it easier for keyboard-oriented users to quickly activate the
push features.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
We actually only want our F5/M1-R keystroke bound in the main window.
Within a browser/blame/console window pressing these keys should not
execute the rescan action.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Pressing the escape key while in the merge dialog cancels the merge
and correctly unlocks the index. Unfortunately this is not true of
the Cancel button, using it closes the dialog but does not release
the index lock, rendering git-gui frozen until you restart it. We
now properly release the index lock when the Cancel button is used.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
The -<n> option was not mentioned in git-format-patch's manpage till
now. Fix this.
Signed-off-by: Miklos Vajna <vmiklos@frugalware.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Without this patch, an added file would be reported as /dev/null.
Noticed by David Kastrup.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There is no restriction on the length of the name returned by
get_object_directory, other than the fact that it must be a stat'able
git object directory. That means its name may have length up to
PATH_MAX-1 (i.e., often 4095) not counting the trailing NUL.
Combine that with the assumption that the concatenation of that name and
suffixes like "/info/alternates" and "/pack/---long-name---.idx" will fit
in a buffer of length PATH_MAX, and you see the problem. Here's a fix:
sha1_file.c (prepare_packed_git_one): Lengthen "path" buffer
so we are guaranteed to be able to append "/pack/" without checking.
Skip any directory entry that is too long to be appended.
(read_info_alternates): Protect against a similar buffer overrun.
Before this change, using the following admittedly contrived environment
setting would cause many git commands to clobber their stack and segfault
on a system with PATH_MAX == 4096:
t=$(perl -e '$s=".git/objects";$n=(4096-6-length($s))/2;print "./"x$n . $s')
export GIT_ALTERNATE_OBJECT_DIRECTORIES=$t
touch g
./git-update-index --add g
If you run the above commands, you'll soon notice that many
git commands now segfault, so you'll want to do this:
unset GIT_ALTERNATE_OBJECT_DIRECTORIES
Signed-off-by: Jim Meyering <jim@meyering.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
I ran git-prune on a repository and got this:
$ git-prune
error: Object 228f8065b930120e35fc0c154c237487ab02d64a is a blob, not a commit
Segmentation fault (core dumped)
This repository was a strange one in that it was being used to provide
its own submodule. That is, the repository was cloned into a
subdirectory, an independent branch checked out in that subdirectory,
and then it was marked as a submodule. git-prune then failed in the
above manner.
The problem was that git-prune was not submodule aware in two areas.
Linus said:
> So what happens is that something traverses a tree object, looks at each
> entry, sees that it's not a tree, and tries to look it up as a blob. But
> subprojects are commits, not blobs, and then when you look at the object
> more closely, you get the above kind of object type confusion.
and included a patch to add an S_ISGITLINK() test to reachable.c's
process_tree() function. That fixed the first git-prune error, and
stopped it from trying to process the gitlink entries in trees as if
they were pointers to other trees (and of course failing, because
gitlinks _aren't_ trees). That part of this patch is his.
The second area is add_cache_refs(). This is called before starting the
reachability analysis, and was calling lookup_blob() on every object
hash found in the index. However, it is no longer true that every hash
in the index is a pointer to a blob, some of them are gitlinks, and are
not backed by any object at all, they are commits in another repository.
Normally this bug was not causing any problems, but in the case of the
self-referencing repository described above, it meant that the gitlink
hash was being marked as being of type OBJ_BLOB by add_cache_refs() call
to lookup_blob(). Then later, because that hash was also pointed to by
a ref, add_one_ref() would treat it as a commit; lookup_commit() would
return a NULL because that object was already noted as being an
OBJ_BLOB, not an OBJ_COMMIT; and parse_commit_buffer() would SEGFAULT on
that NULL pointer.
The fix made by this patch is to not blindly call lookup_blob() in
reachable.c's add_cache_refs(), and instead skip any index entries that
are S_ISGITLINK().
Signed-off-by: Andy Parkins <andyparkins@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The 'show' and 'prune' commands accept an option '-n'; document what
it does.
Signed-off-by: Sam Vilain <sam.vilain@catalyst.net.nz>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Some minor enhancements to the git-repack manual page.
Signed-off-by: Sam Vilain <sam.vilain@catalyst.net.nz>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
On 'Visualize ...', a gitk process is started. Since it is run in the
background, catching a possible startup error doesn't work, and the error
output goes to the console git-gui is started from. The most probable
startup error is that gitk is not installed; so before trying to start,
check for the existence of the gitk program, and popup an error message
unless it's found.
This was noticed and reported by Paul Wise through
http://bugs.debian.org/429810
Signed-off-by: Gerrit Pape <pape@smarden.org>
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
* 'maint' of git://repo.or.cz/git-gui:
git-gui: Don't require a .pvcsrc to create Tools/Migrate menu hack
git-gui: Don't nice git blame on MSYS as nice is not supported
git-gui: Don't require $DISPLAY just to get --version
git-gui: Bind Tab/Shift-Tab to cycle between panes in blame
git-gui: Correctly install to /usr/bin on Cygwin
The Tools/Migrate menu option is a hack just for me. Yes, that's
right, git-gui has a hidden feature that really only works for me,
and the users that I support within my day-job's great firewall.
The menu option is not supported outside of that environment.
In the past we only enabled Tools/Migrate if our special local
script 'gui-miga' existed in the proper location, and if there
was a special '.pvcsrc' in the top level of the working directory.
This latter test for the '.pvcsrc' file is now failing, as the file
was removed from all Git repositories due to changes made to other
tooling within the great firewall's realm.
I have changed the test to only work on Cygwin, and only if the
special 'gui-miga' is present. This works around the configuration
changes made recently within the great firewall's realm, but really
this entire Tools/Migrate thing should be abstracted out into some
sort of plugin system so other users can extend git-gui as they need.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Johannes Sixt reported that MinGW/MSYS does not have a nice.exe to
drop the priority of a child process when it gets spawned. So we
have to avoid trying to start `git blame` through nice when we are
on Windows and do not have Cygwin available to us.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Print no space after the name of a key without value.
Otherwise keys without values are printed exactly the
same as keys with empty values.
Signed-off-by: Frank Lichtenheld <frank@lichtenheld.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The asciidoc documentation of the --get-regexp option was
incomplete. Add some missing pieces:
- List the option in SYNOPSIS
- Mention that key names are printed
Signed-off-by: Frank Lichtenheld <frank@lichtenheld.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add a picture, and keep the setup and the tests together.
Signed-off-by: Sam Vilain <sam.vilain@catalyst.net.nz>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A pack-file can get created without any objects in it (to transfer "no
data" - which can happen if you use a reference git repo, for example,
or just otherwise just end up transferring only branch head information
and already have all the objects themselves).
And while we probably should never create an index for such a pack, if we
do (and we do), the index file size sanity checking was incorrect.
This fixes it.
Reported-and-tested-by: Jocke Tjernlund <tjernlund@tjernlund.se>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Without this patch, the code would look for the submodule
commits in the superproject and (needlessly) fail when it
couldn't find them.
Signed-off-by: Sven Verdoolaege <skimo@liacs.nl>
Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Junio asked that we don't force the user to have a valid X11 server
configured in $DISPLAY just to obtain the output of `git gui version`.
This makes sense, the user may be an automated tool that is running
without an X server available to it, such as a build script or other
sort of package management system. Or it might just be a user working
in a non-GUI environment and wondering "what version of git-gui do I
have installed?".
Tcl has a lot of warts, but one of its better ones is that a comment
can be continued to the next line by escaping the LF that would have
ended the comment using a backslash-LF sequence. In the past we have
used this trick to escape away the 'exec wish' that is actually a Bourne
shell script and keep Tcl from executing it.
I'm using that feature here to comment out the Bourne shell script and
hide it from the Tcl engine. Except now our Bourne shell script is a
few lines long and checks to see if it should print the version, or not.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
The blame viewer is composed of two different areas, the file
area on top and the commit area on the bottom. If users are
trying to shift the focus it is probably because they want to
shift from one area to the other, so we just setup Tab and
Shift-Tab to jump from the one half to the other in a cycle.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Mark Levedahl <mlevedahl@gmail.com> noted that installation on Cygwin
to /usr/bin can cause problems with the automatic guessing of our
library location. The problem is that installation to /usr/bin
means we actually have:
/usr/bin = c:\cygwin\bin
/usr/share = c:\cygwin\usr\share
So git-gui guesses that its library should be found within the
c:\cygwin\share directory, as that is where it should be relative
to the script itself in c:\cygwin\bin.
In my first version of this patch I tried to use `cygpath` to resolve
/usr/bin and /usr/share to test that they were in the same relative
locations, but that didn't work out correctly as we were actually
testing /usr/share against itself, so it always was equal, and we
always used relative paths. So my original solution was quite wrong.
Mark suggested we just always disable relative behavior on Cygwin,
because of the complexity of the mount mapping problem, so that's
all I'm doing.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
It turns out that the attribute definition we have had for a
long time to hide "^" character from AsciiDoc 7 was not honored
by AsciiDoc 8 even under "-a asciidoc7compatible" mode. There is
a similar breakage with the "compatible" mode with + characters.
The double colon at the end of definition list term needs
to be attached to the term, without a whitespace. After this
minimum fixups, AsciiDoc 8 (I used 8.2.1 on Debian) with
compatibility mode seems to produce reasonably good results.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Johannes and Marco discovered that "git log -z" spent cycles in diff even
though there is no need to actually compute diffs.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The original code did not take hierarchical branch names into account at all.
[jc: cherry-picked 11f68d9 from 'master']
Tested-by: Gerrit Pape <pape@smarden.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Junio C Hamano