On Windows, git-gui suggests running the garbage collector if it finds
1 or more files in .git/objects/42 (as opposed to 8 files on other
platforms). The probability of that happening if the repo contains
about 100 loose objects is 32%. The probability for the same to happen
when searching 4 directories is only 8%, which is bit more reasonable.

Also remove $objects_limit from the message, because we already know
that we are above (or close to) that limit. Telling the user about
that number does not really give him any useful information.

The following octave script shows the probability for at least m*q
objects to be found in q subdirectories of .git/objects if n is the
total number of objects.

q = 4;
m = [1 2 8];
n = 0:10:2000;

P = zeros(length(n), length(m));
for k = 1:length(n)
        P(k, :) = 1-binocdf(q*m-1, n(k), q/(256-q));
end
plot(n, P);

n \ q   1       4
50      18%     1%
100     32%     8%
200     54%     39%
500     86%     96%
Signed-off-by: NClemens Buchacher <drizzd@aon.at>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

We already do the same when locally deleting a branch.
Signed-off-by: NHeiko Voigt <hvoigt@hvoigt.net>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

When the main window is maximized, if the screen height is shorter (e.g.
Netbook screen 1024x600), both the partial commit pane and the status bar
are hidden. The diff pane is resizable, so that it can use less vertical
height, allowing the overall window to be shorter and still display both
the entire commit pane and status bar.
Signed-off-by: NVietor Liu <vietor@vxwo.org>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

This fixes a bug introduced by the "display summary when showing diff of a
submodule" patch. It lead to a "no such variable" error when opening the
diff context menu while no diff was shown.
Signed-off-by: NJens Lehmann <Jens.Lehmann@web.de>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

I often close git gui window when it is maximized, and when I reopen
it next time the it would usually become out of place (e.g. a huge
window with a top-left corner somewhere close to the center of the
screen). Fix it by storing and restoring wm state in config, as well
as setting wm state to normal before retrieving wm geometry info.
Signed-off-by: NAlexey Borzenkov <snaury@gmail.com>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

When quoting an arbitrary user string in Tcl, its better to use
[list ...] than to use {...}, in case the user string has spaces
or { embedded within it.
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

When a submodule commit had already been staged and another commit had
been checked out inside the submodule, the diff always displayed the
submodule commit log messages between the last supermodule commit and
the working tree, totally ignoring the commit in the index.
Signed-off-by: NJens Lehmann <Jens.Lehmann@web.de>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

Signed-off-by: NAlex Riesen <raa.lkml@gmail.com>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

When there is a large number of new or modified files,
"display_all_files" takes a long time, and git-gui appears to hang.

This change limits the number of files that are displayed.  This
limit can be set as gui.maxfilesdisplayed, and is 5000 by default.

A warning is shown the first time the list of files is truncated
in this GUI session.  Subsequent truncations are not mentioned to
the user.
Signed-off-by: NDan Zwell <dzwell@zwell.net>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

If the user wants to delete a remote branch and selects the correct
"merged into" we should not warn that "Recovering deleted branches is
difficult". For local branches we do the same already.
Signed-off-by: NHeiko Voigt <hvoigt@hvoigt.net>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

As it is hard to say what changed in a submodule by looking at the hashes,
let's show the colored submodule summary instead.
Signed-off-by: NJens Lehmann <Jens.Lehmann@web.de>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

- detect more Tk.framework variants
- fix apple menu setup, use native preferences menu item
- don't set menu font
Signed-off-by: NDaniel A. Steffen <das@users.sourceforge.net>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

Also, the previous translations of the words 'tag' and 'merge' were
changed. Added translation of the 'Tool' submenu.

Thanks go to Alexander Gavrilov and Dmitry Potapov for proofreading
and suggestions.
Signed-off-by: NAlex Riesen <raa.lkml@gmail.com>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

git-gui is using "git-read-tree -u" when cloning which doesn't
invoke the post-checkout hook as a plain git-clone would.
So git-gui must call the hook itself.
Signed-off-by: NJens Lehmann <Jens.Lehmann@web.de>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

* maint:
  git-gui: Ensure consistent usage of mergetool.keepBackup
  git-gui: fix use of undeclared variable diff_empty_count

In several places merge.keepBackup is used i.s.o.
mergetool.keepBackup. This patch makes it all
consistent.
Signed-off-by: NFerry Huberts <ferry.huberts@pelagic.nl>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

Commit 584fa9cc introduced the global variable diff_empty_count, which
is used in diff.tcl. This variable wasn't declared anywhere which
resulted in an ugly error message box instead of the intended
informative message.
Signed-off-by: NJoerg Bornemann <joerg.bornemann@web.de>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

This works around git-gui's error message

    Cannot use funny .git directory: .

when started from the .git/ directory, which is useful in repositories
without any directories for the right click.

Now git-gui can be started via Windows Explorer shell extension (Git GUI
Here) from the .git/ directory.
Signed-off-by: NMarkus Heidelberg <markus.heidelberg@web.de>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

Starting the Explorer from the git-gui menu "Explore Working Copy"
didn't work, when git-gui was started via Windows Explorer shell
extension (Git GUI Here) from a directory within the project.
The Explorer raised an error message like this:

    Path "C:/somedir/worktree" is not available or not a directory

It worked when started from the project directory itself, because then
the path argument for the Explorer was just '.' (current directory)
without any problematic forward slashes.

To make it work, convert the path given as argument to explorer.exe to
its native format with backslashes.
Signed-off-by: NMarkus Heidelberg <markus.heidelberg@web.de>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

git-gui is using "git-read-tree -u" for checkout which doesn't
invoke the post-checkout hook as a plain git-checkout would.
So git-gui must call the hook itself.
Signed-off-by: NJens Lehmann <Jens.Lehmann@web.de>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

Before calling the post-commit hook, the variable "pc_err" is cleared
while later only "pch_error" is used. "pch_error$cmt_id" only appeared in
"upvar"-Statements (which were changed to "global") and was removed.
Signed-off-by: NJens Lehmann <Jens.Lehmann@web.de>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

Previously a hardcoded path $GIT_EXEC_PATH/../Documentation/ was used to
search for the documentation, when the user has asked for it via menu
"Help -> Online Documentation".
This didn't work for the default directory structure.

To find the path reliably, use the new git command line option, which
returns the correct path.

If the output of `git --html-path` is empty because git is not found or
the option is not yet supported in the installed git, the documentation
from kernel.org is launched. There is no additional guessing of the
right location of the installed docs.
Signed-off-by: NMarkus Heidelberg <markus.heidelberg@web.de>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

* maint:
  git-gui: fix deleting from the context menu with empty selection

An "Application Error" was raised when trying to delete text from the
commit message field when no text was selected.
Signed-off-by: NMarkus Heidelberg <markus.heidelberg@web.de>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

Properly spell "successful" and slightly rewrite a couple of strings
that actually say the same thing in order to reduce translation work.

Update .pot and .po files accordingly since no new translation is
required.
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

Mostly grammar, spelling and typography fixes, but also a few wording
enhancements here and there.
Signed-off-by: NSam Hocevar <sam@zoy.org>
Acked-by: NChristian Couder <chriscool@tuxfamily.org>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

When a merge conflict occurs in a file with spaces in the filename,
git-gui showed wrongly "LOCAL: deleted".
Signed-off-by: NJens Lehmann <Jens.Lehmann@web.de>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

* maint:
  git-gui: don't hide the Browse button when resizing the repo chooser

Rather shrink the input field for "Create New Repository" and "Open
Existing Repository" as it's already done for "Clone Existing
Repository".
Signed-off-by: NMarkus Heidelberg <markus.heidelberg@web.de>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

The git-gui menu item "Repository | Create Desktop Icon" creates a
shortcut (.lnk file) on the Windows desktop.  The purpose of the
created shortcut is to make it easy for a user to launch git-gui
for a particular repo in the future.

A Windows user would expect to see git gui launch when they click
the shortcut; they would not expect (nor want) to see a cmd window
open and remain open in the background.

msysGit avoids opening a command window altogether when it's Git GUI
shortcut is used.  Ideally, git on cygwin would also have shortcuts
that simply open the GUI, but as a first step, this change allows
the shell window to politely disappear after starting git gui as a
background process.
Signed-off-by: NPhil Lawrence <prlawrence@gmail.com>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

Recently the msysgit repository has got a '1.6.1-msysgit1'
tag, which, when used to build the git version, is not
handled gracefully by the git-gui version code.

This patch changes the regular expressions to fix it, and
removes the hardcoded 'rc' string. Now git-gui can accept
a version tail like '.foo123.GIT.bar.456.7.g89ab'
Signed-off-by: NAlexander Gavrilov <angavrilov@gmail.com>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

If the index update machinery and git diff happen to disagree
on whether a particular file is modified, it may cause git-gui
to enter an infinite index rescan loop, where an empty diff
starts a rescan, which finds the same set of files modified,
and tries to display the diff for the first one, which happens
to be the empty one. A current example of a possible disagreement
point is the autocrlf filter.

This patch breaks the loop by using a global counter to track
the auto-rescans. The variable is reset whenever a non-empty
diff is displayed.

Another suggested approach, which is based on giving the
--exit-code argument to git diff, cannot be used, because
diff-files seems to trust the timestamps in the index, and
returns a non-zero code even if the file is actually
unchanged, which essentially defeats the purpose of the
auto-rescan logic.
Signed-off-by: NAlexander Gavrilov <angavrilov@gmail.com>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

As pointed out in msysgit bug #181, when a non-locale encoding is
used for commits, post-commit status messages display the subject
incorrectly.  It happens because the file handle is not properly
configured before the subject is read back.

This patch fixes it by factoring out the code that is used to setup
the output handle into a separate function, and calling it from
the reading code.
Signed-off-by: NAlexander Gavrilov <angavrilov@gmail.com>
Acked-by: NRobin Rosenberg <robin.rosenberg@gmail.com>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

In dc871831(Only use GIT_CONFIG in "git config", not other programs),
GIT_CONFIG_LOCAL was rested in peace, in favor of not reading
/etc/gitconfig and $HOME/.gitconfig at all when GIT_CONFIG is set.
Signed-off-by: NJohannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

Signed-off-by: NMiklos Vajna <vmiklos@frugalware.org>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

Signed-off-by: NPeter Krefting <peter@softwolves.pp.se>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>

Signed-off-by: NPeter Krefting <peter@softwolves.pp.se>
Signed-off-by: NShawn O. Pearce <spearce@spearce.org>