gitk: Fix some bugs with path limiting in the diff display
First, we weren't putting "--" between the ids and the paths in the
git diff-tree/diff-index/diff-files command, so if there was a tag
and a file with the same name, we could get an ambiguity in the
command. This puts the "--" in to make it clear that the paths are
paths.
Secondly, this implements the path limiting for merge diffs as well
as the normal 2-way diffs.
gitk: Integrate the reset progress bar in the main frame
This makes the reset function use a progress bar in the same location
as the progress bars for reading in commits and for finding commits,
instead of a progress bar in a separate detached window. The progress
bar for resetting is red.
This also puts "Resetting" in the status window while the reset is in
progress. The setting of the status window is done through an
extension of the interface used for setting the watch cursor.
gitk: Ensure tabstop setting gets restored by Cancel button
We weren't restoring the tabstop setting if the user pressed the
Cancel button in the Edit/Preferences window. Also improved the
label for the checkbox (made it "Tab spacing" rather than the laconic
"tabstop") and moved it above the "Display nearby tags" checkbox.
gitk: Limit diff display to listed paths by default
When the user has specified a list of paths, either on the command line
or when creating a view, gitk currently displays the diffs for all files
that a commit has modified, not just the ones that match the path list.
This is different from other git commands such as git log. This change
makes gitk behave the same as these other git commands by default, that
is, gitk only displays the diffs for files that match the path list.
There is now a checkbox labelled "Limit diffs to listed paths" in the
Edit/Preferences pane. If that is unchecked, gitk will display the
diffs for all files as before.
When gitk is run with the --merge flag, it will get the list of unmerged
files at startup, intersect that with the paths listed on the command line
(if any), and use that as the list of paths.
Use PRIuMAX instead of 'unsigned long long' in show-index
Elsewhere in Git we already use PRIuMAX and cast to uintmax_t when
we need to display a value that is 'very big' and we're not exactly
sure what the largest display size is for this platform.
This particular fix is needed so we can do the incredibly crazy
temporary hack of:
allowing us to more easily look for locations where we are passing
a pointer to an 8 byte value to a function that expects a 4 byte
value. This can occur on some platforms where sizeof(long) == 8
and sizeof(size_t) == 4.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
* maint:
Describe more 1.5.3.5 fixes in release notes
Fix diffcore-break total breakage
Fix directory scanner to correctly ignore files without d_type
Improve receive-pack error message about funny ref creation
fast-import: Fix argument order to die in file_change_m
git-gui: Don't display CR within console windows
git-gui: Handle progress bars from newer gits
git-gui: Correctly report failures from git-write-tree
gitk.txt: Fix markup.
send-pack: respect '+' on wildcard refspecs
git-gui: accept versions containing text annotations, like 1.5.3.mingw.1
git-gui: Don't crash when starting gitk from a browser session
git-gui: Allow gitk to be started on Cygwin with native Tcl/Tk
git-gui: Ensure .git/info/exclude is honored in Cygwin workdirs
git-gui: Handle starting on mapped shares under Cygwin
git-gui: Display message box when we cannot find git in $PATH
git-gui: Avoid using bold text in entire gui for some fonts
Ok, so on the kernel list, some people noticed that "git log --follow"
doesn't work too well with some files in the x86 merge, because a lot of
files got renamed in very special ways.
In particular, there was a pattern of doing single commits with renames
that looked basically like
- rename "filename.h" -> "filename_64.h"
- create new "filename.c" that includes "filename_32.h" or
"filename_64.h" depending on whether we're 32-bit or 64-bit.
which was preparatory for smushing the two trees together.
Now, there's two issues here:
- "filename.c" *remained*. Yes, it was a rename, but there was a new file
created with the old name in the same commit. This was important,
because we wanted each commit to compile properly, so that it was
bisectable, so splitting the rename into one commit and the "create
helper file" into another was *not* an option.
So we need to break associations where the contents change too much.
Fine. We have the -B flag for that. When we break things up, then the
rename detection will be able to figure out whether there are better
alternatives.
- "git log --follow" didn't with with -B.
Now, the second case was really simple: we use a different "diffopt"
structure for the rename detection than the basic one (which we use for
showing the diffs). So that second case is trivially fixed by a trivial
one-liner that just copies the break_opt values from the "real" diffopts
to the one used for rename following. So now "git log -B --follow" works
fine:
however, the end result does *not* work. Because our diffcore-break.c
logic is totally bogus!
In particular:
- it used to do
if (base_size < MINIMUM_BREAK_SIZE)
return 0; /* we do not break too small filepair */
which basically says "don't bother to break small files". But that
"base_size" is the *smaller* of the two sizes, which means that if some
large file was rewritten into one that just includes another file, we
would look at the (small) result, and decide that it's smaller than the
break size, so it cannot be worth it to break it up! Even if the other
side was ten times bigger and looked *nothing* like the samell file!
That's clearly bogus. I replaced "base_size" with "max_size", so that
we compare the *bigger* of the filepair with the break size.
- It calculated a "merge_score", which was the score needed to merge it
back together if nothing else wanted it. But even if it was *so*
different that we would never want to merge it back, we wouldn't
consider it a break! That makes no sense. So I added
if (*merge_score_p > break_score)
return 1;
to make it clear that if we wouldn't want to merge it at the end, it
was *definitely* a break.
- It compared the whole "extent of damage", counting all inserts and
deletes, but it based this score on the "base_size", and generated the
damage score with
but that makes no sense either, since quite often, this will result in
a number that is *bigger* than MAX_SCORE! Why? Because base_size is
(again) the smaller of the two files we compare, and when you start out
from a small file and add a lot (or start out from a large file and
remove a lot), the base_size is going to be much smaller than the
damage!
Again, the fix was to replace "base_size" with "max_size", at which
point the damage actually becomes a sane percentage of the whole.
With these changes in place, not only does "git log -B --follow" work for
the case that triggered this in the first place, ie now
actually gives reasonable results. But I also wanted to verify it in
general, by doing a full-history
git log --stat -B -C
on my kernel tree with the old code and the new code.
There's some tweaking to be done, but generally, the new code generates
much better results wrt breaking up files (and then finding better rename
candidates). Here's a few examples of the "--stat" output:
Now, in some other cases, it does actually turn a rename into a real
"delete+create" pair, and then the diff is usually bigger, so truth in
advertizing: it doesn't always generate a nicer diff. But for what -B was
meant for, I think this is a big improvement, and I suspect those cases
where it generates a bigger diff are tweakable.
So I think this diff fixes a real bug, but we might still want to tweak
the default values and perhaps the exact rules for when a break happens.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Fix directory scanner to correctly ignore files without d_type
On Fri, 19 Oct 2007, Todd T. Fries wrote:
> If DT_UNKNOWN exists, then we have to do a stat() of some form to
> find out the right type.
That happened in the case of a pathname that was ignored, and we did
not ask for "dir->show_ignored". That test used to be *together*
with the "DTYPE(de) != DT_DIR", but splitting the two tests up
means that we can do that (common) test before we even bother to
calculate the real dtype.
Of course, that optimization only matters for systems that don't
have, or don't fill in DTYPE properly.
I also clarified the real relationship between "exclude" and
"dir->show_ignored". It used to do
if (exclude != dir->show_ignored) {
..
which wasn't exactly obvious, because it triggers for two different
cases:
- the path is marked excluded, but we are not interested in ignored
files: ignore it
- the path is *not* excluded, but we *are* interested in ignored
files: ignore it unless it's a directory, in which case we might
have ignored files inside the directory and need to recurse
into it).
so this splits them into those two cases, since the first case
doesn't even care about the type.
I also made a the DT_UNKNOWN case a separate helper function,
and added some commentary to the cases.
Linus
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Avoid a dup2(2) in apply_filter() - start_command() can do it for us.
When apply_filter() runs the external (clean or smudge) filter program, it
needs to pass the writable end of a pipe as its stdout. For this purpose,
it used to dup2(2) the file descriptor explicitly to stdout. Now we use
the facilities of start_command() to do it for us.
Furthermore, the path argument of a subordinate function, filter_buffer(),
was not used, so here we replace it to pass the fd instead.
Signed-off-by: Johannes Sixt <johannes.sixt@telecom.at> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
t0021-conversion.sh: Test that the clean filter really cleans content.
This test uses a rot13 filter, which is its own inverse. It tested only
that the content was the same as the original after both the 'clean' and
the 'smudge' filter were applied. This way it would not detect whether
any filter was run at all. Hence, here we add another test that checks
that the repository contained content that was processed by the 'clean'
filter.
Signed-off-by: Johannes Sixt <johannes.sixt@telecom.at> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
upload-pack: Run rev-list in an asynchronous function.
This gets rid of an explicit fork().
Since upload-pack has to coordinate two processes (rev-list and
pack-objects), we cannot use the normal finish_async(), but have to monitor
the process explicitly. Hence, there are no changes at this front.
Signed-off-by: Johannes Sixt <johannes.sixt@telecom.at> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
upload-pack: Move the revision walker into a separate function.
This allows us later to use start_async() with this function, and at
the same time is a nice cleanup that makes a long function
(create_pack_file()) shorter.
Signed-off-by: Johannes Sixt <johannes.sixt@telecom.at> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Use the asyncronous function infrastructure in builtin-fetch-pack.c.
We run the sideband demultiplexer in an asynchronous function.
Note that earlier there was a check in the child process that closed
xd[1] only if it was different from xd[0]; this test is no longer needed
because git_connect() always returns two different file descriptors
(see ec587fde0a76780931c7ac32474c8c000aa45134).
Signed-off-by: Johannes Sixt <johannes.sixt@telecom.at> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Add infrastructure to run a function asynchronously.
This adds start_async() and finish_async(), which runs a function
asynchronously. Communication with the caller happens only via pipes.
For this reason, this implementation forks off a child process that runs
the function.
[sp: Style nit fixed by removing unnecessary block on if condition
inside of start_async()]
Signed-off-by: Johannes Sixt <johannes.sixt@telecom.at> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
upload-pack: Use start_command() to run pack-objects in create_pack_file().
This gets rid of an explicit fork/exec.
Since upload-pack has to coordinate two processes (rev-list and
pack-objects), we cannot use the normal finish_command(), but have to
monitor the processes explicitly. Hence, the waitpid() call remains.
Signed-off-by: Johannes Sixt <johannes.sixt@telecom.at> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Have start_command() create a pipe to read the stderr of the child.
This adds another stanza that allocates a pipe that is connected to the
child's stderr and that the caller can read from. In order to request this
pipe, the caller sets cmd->err to -1.
The implementation is not exactly modeled after the stdout case: For stdout
the caller can supply an existing file descriptor, but this facility is
nowhere needed in the stderr case. Additionally, the caller is required to
close cmd->err.
Signed-off-by: Johannes Sixt <johannes.sixt@telecom.at> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Change git_connect() to return a struct child_process instead of a pid_t.
This prepares the API of git_connect() and finish_connect() to operate on
a struct child_process. Currently, we just use that object as a placeholder
for the pid that we used to return. A follow-up patch will change the
implementation of git_connect() and finish_connect() to make full use
of the object.
Old code had early-return-on-error checks at the calling sites of
git_connect(), but since git_connect() dies on errors anyway, these checks
were removed.
[sp: Corrected style nit of "conn == NULL" to "!conn"]
Signed-off-by: Johannes Sixt <johannes.sixt@telecom.at> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
* git://git.kernel.org/pub/scm/gitk/gitk:
gitk: Fix "can't unset prevlines(...)" Tcl error
gitk: Avoid an error when cherry-picking if HEAD has moved on
gitk: Check that we are running on at least Tcl/Tk 8.4
gitk: Do not pick up file names of "copy from" lines
gitk: Add support for OS X mouse wheel
gitk: disable colours when calling git log
Merge branch 'maint' of git://repo.or.cz/git-gui into maint
* 'maint' of git://repo.or.cz/git-gui:
git-gui: Don't display CR within console windows
git-gui: Handle progress bars from newer gits
git-gui: Correctly report failures from git-write-tree
git-gui: accept versions containing text annotations, like 1.5.3.mingw.1
git-gui: Don't crash when starting gitk from a browser session
git-gui: Allow gitk to be started on Cygwin with native Tcl/Tk
git-gui: Ensure .git/info/exclude is honored in Cygwin workdirs
git-gui: Handle starting on mapped shares under Cygwin
git-gui: Display message box when we cannot find git in $PATH
git-gui: Avoid using bold text in entire gui for some fonts
This fixes the error reported by Michele Ballabio, where gitk will
throw a Tcl error "can't unset prevlines(...)" when displaying a
commit that has a parent commit listed more than once, and the commit
is the first child of that parent.
The problem was basically that we had two variables, prevlines and
lineends, and were relying on the invariant that prevlines($id) was
set iff $id was in the lineends($r) list for some $r. But having
a duplicate parent breaks that invariant since we end up with the
parent listed twice in lineends.
This fixes it by simplifying the logic to use only a single variable,
lineend. It also rearranges things a little so that we don't try to
draw the line for the duplicated parent twice.
Define compat version of mkdtemp for systems lacking it
Solaris 9 doesn't have mkdtemp() so we need to emulate it for the
rsync transport implementation. Since Solaris 9 is lacking this
function we can also reasonably assume it is not available on
Solaris 8 either. The new Makfile definition NO_MKDTEMP can be
set to enable the git compat version.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Deduce exec_path also from calls to git with a relative path
There is already logic in the git wrapper to deduce the exec_path from
argv[0], when the git wrapper was called with an absolute path. Extend
that logic to handle relative paths as well.
For example, when you call "../../hello/world/git", it will not turn
"../../hello/world" into an absolute path, and use that.
Initial implementation by Scott R Parish.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
fast-import: Fix argument order to die in file_change_m
The arguments to the "Not a blob" die call in file_change_m were
transposed, so that the command was printed as the type, and the type
as the command. Switch them around so that the error message comes
out correctly.
Signed-off-by: Julian Phillips <julian@quantumfyre.co.uk> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Git progress bars from tools like git-push and git-fetch use CR
to skip back to the start of the current line and redraw it with
an updated progress. We were doing this in our Tk widget but had
failed to skip the CR, which Tk doesn't draw well.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Post Git 1.5.3 a new style progress bar has been introduced that
uses only one line rather than two. The formatting of the completed
and total section is also slightly different so we must adjust our
regexp to match. Unfortunately both styles are in active use by
different versions of Git so we need to look for both.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
The current git-p4 implementation does support file renames. However, because
it does not use the "p4 integrate" command, the history for the renamed file is
not linked to the new file.
This changeset adds support for perforce renames with the integrate command.
Currently this feature is only enabled when calling git-p4 submit with the -M
option. This is intended to look and behave similar to the "detect renames"
feature of other git commands.
The following sequence is used for renamed files:
p4 integrate -Dt x x'
p4 edit x'
rm x'
git apply
p4 delete x
By default, perforce will not allow an integration with a target file that has
been deleted. That is, if x' in the example above is the name of a previously
deleted file then perforce will fail the integrate. The -Dt option tells
perforce to allow the target of integrate to be a previously deleted file.
Signed-off-by: Chris Pettitt <cpettitt@gmail.com> Signed-off-by: Simon Hausmann <simon@lst.de>
gitk: Avoid an error when cherry-picking if HEAD has moved on
This fixes an error reported by Adam Piątyszek: if the current HEAD
is not in the graph that gitk knows about when we do a cherry-pick
using gitk, then gitk hits an error when trying to update its
internal representation of the topology. This avoids the error by
not doing that update if the HEAD before the cherry-pick was a
commit that gitk doesn't know about.
gitk: Check that we are running on at least Tcl/Tk 8.4
This checks that we have Tcl/Tk 8.4 or later, and puts up an error
message in a window and quits if not.
This was prompted by a patch submitted by Steffen Prohaska, but is
done a bit differently (this uses package require rather than
looking at [info tclversion], and uses show_error to display the
error rather than printing it to stderr).
git-gui: Correctly report failures from git-write-tree
If git-write-tree fails (such as if the index file is currently
locked and it wants to write to it) we were not getting the error
message as $tree_id was always the empty string so we shortcut
through the catch and never got the output from stderr.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
gitk: Do not pick up file names of "copy from" lines
A file copy would be detected only if the original file was modified in the
same commit. This implies that there will be a patch listed under the
original file name, and we would expect that clicking the original file
name in the file list warps the patch window to that file's patch. (If the
original file was not modified, the copy would not be detected in the first
place, the copied file would be listed as "new file", and this whole matter
would not apply.)
However, if the name of the copy is sorted after the original file's patch,
then the logic introduced by commit d1cb298b0b (which picks up the link
information from the "copy from" line) would overwrite the link
information that is already present for the original file name, which was
parsed earlier. Hence, this patch reverts part of said commit.
Signed-off-by: Johannes Sixt <johannes.sixt@telecom.at> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
(Väinö Järvelä supplied this patch a while ago for 1.5.2. It no longer
applied cleanly, so I'm reposting it.)
MacBook doesn't seem to recognize MouseRelease-4 and -5 events, at all.
So i added a support for the MouseWheel event, which i limited to Tcl/tk
aqua, as i couldn't test it neither on Linux or Windows. Tcl/tk needs to
be updated from the version that is shipped with OS X 10.4 Tiger, for
this patch to work.
Signed-off-by: Jonathan del Strother <jon.delStrother@bestbefore.tv> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
When matching source and destination refs, we were failing
to pull the 'force' parameter from wildcard refspecs (but
not explicit ones) and attach it to the ref struct.
This adds a test for explicit and wildcard refspecs; the
latter fails without this patch.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
The logic in stopfinding assumes that findcurline will be set if
find_dirn is, but findnext and findprev can set find_dirn without
setting findcurline. This makes sure we only set find_dirn in those
places if findcurline is already set.
Avoid scary errors about tagged trees/blobs during git-fetch
This is the same bug as 42a32174b600f139b489341b1281fb1bfa14c252.
The warning "Object $X is a tree, not a commit" is bogus and is
not relevant here. If its not a commit we just need to make sure
we don't mark it for merge as we fill out FETCH_HEAD.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
In 89d750bf6fa025edeb31ad258cdd09a27a5c02fa I got a little too
aggressive with changing "git diff" to "git diff-tree". This is
shown to the user, who expects to see a full diff on their console,
and will want to see the output of their custom diff drivers (if
any) as the whole point of this call site is to show the diff to
the end-user.
Noticed by Johannes Sixt <j.sixt@viscovery.net>.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
* maint:
Further 1.5.3.5 fixes described in release notes
Avoid invoking diff drivers during git-stash
attr: fix segfault in gitattributes parsing code
Define NI_MAXSERV if not defined by operating system
Ensure we add directories in the correct order
Avoid scary errors about tagged trees/blobs during git-fetch
Documentation/git-gc: improve description of --auto
This patch tries to make the description of --auto a little
more clear for new users, especially those referred by the
"git-gc --auto" notification message.
It also cleans up some grammatical errors and typos in the
original description, as well as rewording for clarity.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Documentation/git-gc: explain --auto in description
Now that git-gc --auto tells the user to look at the man
page, it makes sense to mention the auto behavior near the
top (since this is likely to be most users' first exposure
to git-gc).
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
The previous message had too much of a "boy, you should
really turn off this annoying gc" flair to it. Instead,
let's make sure the user understands what is happening, that
they can run it themselves, and where to find more info.
Suggested by Brian Gernhardt.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
git-stash needs to restrict itself to plumbing when running automated
diffs as part of its operation as the user may have configured a
custom diff driver that opens an interactive UI for certain/all
files. Doing that during scripted actions is very unfriendly to
the end-user and may cause git-stash to fail to work.
Reported by Johannes Sixt
Acked-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
git may segfault if gitattributes contains an invalid
entry. A test is added to t0020 that triggers the segfault.
The parsing code is fixed to avoid the crash.
Signed-off-by: Steffen Prohaska <prohaska@zib.de> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Define NI_MAXSERV if not defined by operating system
I found I needed NI_MAXSERV as it is defined in netdb.h, which is
not included by daemon.c. Rather than including the whole header
we can define a reasonable fallback value.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Avoid scary errors about tagged trees/blobs during git-fetch
Ok, what is going on is:
- append_fetch_head() looks up the SHA1 for all heads (including tags):
if (get_sha1(head, sha1))
return error("Not a valid object name: %s", head);
- it then wants to check if it's a candidate for merging (because
fetching also does the whole "list which heads to merge" in case
it is going to be part of a "pull"):
commit = lookup_commit_reference(sha1);
if (!commit)
not_for_merge = 1;
- and that "lookup_commit_reference()" is just very vocal about the
case where it fails. It really shouldn't be, and it shouldn't
affect the actual end result, but that basically explains why
you get that scary warning.
In short, the warning is just bogus, and should be harmless, but
I agree that it's ugly. I think the appended patch should fix it.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
mergetool: use path to mergetool in config var mergetool.<tool>.path
This commit adds a mechanism to provide absolute paths to the
external programs called by 'git mergetool'. A path can be
specified in the configuation variable mergetool.<tool>.path.
The configuration variable is similar to how we name branches
and remotes. It is extensible if we need to specify more details
about a tool.
The mechanism is especially useful on Windows, where external
programs are unlikely to be in PATH.
[sp: Fixed a few minor issues prior to applying]
Signed-off-by: Steffen Prohaska <prohaska@zib.de> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
The first test (updating local refs) should succeed without the
prior commit, but the second one (not updating on error) used to
fail before the prior commit was written.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Previously, we updated the tracking refs (which match refs we
are pushing) while generating the list of refs to send.
However, at that point we don't know whether the refs were
accepted.
Instead, we now wait until we get a response code from the
server. If an error was indicated, we don't update any local
tracking refs. Technically some refs could have been updated
on the remote, but since the local ref update is just an
optimization to avoid an extra fetch, we are better off
erring on the side of correctness.
The user-visible message is now generated much later in the
program, and has been tweaked to make more sense.
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
* maint:
Yet more 1.5.3.5 fixes mentioned in release notes
cvsserver: Use exit 1 instead of die when req_Root fails.
git-blame shouldn't crash if run in an unmerged tree
git-config: print error message if the config file cannot be read
fixing output of non-fast-forward output of post-receive-email
git-blame shouldn't crash if run in an unmerged tree
If we are in the middle of resolving a merge conflict there may be
one or more files whose entries in the index represent an unmerged
state (index entries in the higher-order stages).
Attempting to run git-blame on any file in such a working directory
resulted in "fatal: internal error: ce_mode is 0" as we use the magic
marker for an unmerged entry is 0 (set up by things like diff-lib.c's
do_diff_cache() and builtin-read-tree.c's read_tree_unmerged())
and the ce_match_stat_basic() function gets upset about this.
I'm not entirely sure that the whole "ce_mode = 0" case is a good
idea to begin with, and maybe the right thing to do is to remove
that horrid freakish special case, but removing the internal error
seems to be the simplest fix for now.
Linus
[sp: Thanks to Björn Steinbrink for the test case]
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
fixing output of non-fast-forward output of post-receive-email
post-receive-email has one place where the variable fast_forward is not
spelled correctly. At the same place the logic was reversed. The
combination of both bugs made the script work correctly for fast-forward
commits but not for non-fast-forward ones. This change fixes this to
be correct in both cases.
Signed-off-by: Robert Schiele <rschiele@gmail.com> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
This allows one to easily retrieve a single SVN property from within
git-svn without requiring svn or remembering the URL of a repository
* git-svn.perl (%cmd): Add the new command `propget'.
($cmd_dir_prefix): New global.
(&get_svnprops): New helper.
(&cmd_propget): New. Use &get_svnprops.
* t/t9101-git-svn-props.sh: Add a test case for propget.
[ew: make sure the rev-parse --show-prefix call doesn't break
the `git-svn clone' command]
Signed-off-by: Benoit Sigoure <tsuna@lrde.epita.fr> Acked-by: Eric Wong <normalperson@yhbt.net> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
git svn create-ignore (to create one .gitignore per directory
from the svn:ignore properties. This has the disadvantage of
committing the .gitignore during the next dcommit, but when you
import a repo with tons of ignores (>1000), using git svn show-ignore
to build .git/info/exclude is *not* a good idea, because things like
git-status will end up doing >1000 fnmatch *per file* in the repo,
which leads to git-status taking more than 4s on my Core2Duo 2Ghz 2G
RAM)
* git-svn.perl (%cmd): Add the new command `create-ignore'.
(&cmd_create_ignore): New.
* t/t9101-git-svn-props.sh: Adjust the test-case for show-ignore and
add a test case for create-ignore.
[ew: added commit message from
<05CAB148-56ED-4FF1-8AAB-4BA2A0B70C2C@lrde.epita.fr> ]
Signed-off-by: Benoit Sigoure <tsuna@lrde.epita.fr> Acked-by: Eric Wong <normalperson@yhbt.net> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
* maint:
Document additional 1.5.3.5 fixes in release notes
Avoid 'expr index' on Mac OS X as it isn't supported
filter-branch: update current branch when rewritten
fix filter-branch documentation
helpful error message when send-pack finds no refs in common.
Fix setup_git_directory_gently() with relative GIT_DIR & GIT_WORK_TREE
Correct typos in release notes for 1.5.3.5
The man page for filter-branch still talked about writing the result
to the branch "newbranch". This is hopefully the last place where the
old behaviour was described.
Noticed by Bill Lear.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
git-svn: use "no warnings 'once'" to disable false-positives
Some variables coming from the Subversion's Perl bindings are used
in our code only once, so the interpreter warns us about it. These
warnings are false-positives, because the variables themselves are
initialized in the binding's guts, that are made by SWIG.
Credits to Sam Vilain for his note about "no warnings 'once'".
[ew: minor formatting change]
Signed-off-by: Eygene Ryabinkin <rea-git@codelabs.ru> Acked-by: Eric Wong <normalperson@yhbt.net> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Fix setup_git_directory_gently() with relative GIT_DIR & GIT_WORK_TREE
There are a few programs, such as config and diff, which allow running
without a git repository. Therefore, they have to call
setup_git_directory_gently().
However, when GIT_DIR and GIT_WORK_TREE were set, and the current
directory was a subdirectory of the work tree,
setup_git_directory_gently() would return a bogus NULL prefix.
This patch fixes that.
Noticed by REPLeffect on IRC.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
fetch: if not fetching from default remote, ignore default merge
When doing "git fetch <remote>" on a remote that does not have the
branch referenced in branch.<current-branch>.merge, git fetch failed.
It failed because it tried to add the "merge" ref to the refs to be
fetched.
Fix that. And add a test case.
Incidentally, this unconvered a bug in our own test suite, where
"git pull <some-path>" was expected to merge the ref given in the
defaults, even if not pulling from the default remote.
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Lars Hjemli <hjemli@gmail.com> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
git-rebase: document suppression of duplicate commits
git-rebase uses format-patch's --ignore-if-in-upstream
option, but we never document the user-visible behavior. The
example is placed near the top of the example list rather
than at the bottom because it is:
a. a simple example
b. a reasonably common scenario for many projects (mail
some patches which get accepted upstream, then rebase)
[sp: Corrected direction of 'HEAD..<upstream>' set comparsion]
Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
Documentation/git-tag.txt: Document how to backdate tags
Added a new section beneath "On Automatic following" called "On
Backdating Tags". This includes an explanation of when to use this
method, a brief explanation of the kind of date that can be used in
GIT_AUTHOR_DATE, and an example invocation of git-tag using a custom
setting of GIT_AUTHOR_DATE.
[sp: Corrected s/you/your/, noticed by Jeff King]
Signed-off-by: Michael W. Olson <mwolson@gnu.org> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>