This commit fixes two problems with commit encodings:

1) git-log actually uses i18n.logoutputencoding to generate
   its output, and falls back to i18n.commitencoding only
   when that option is not set.  Thus, gitk should use its
   value to read the results, if available.

2) The readcommit function did not process encodings at all.
   This led to randomly appearing misconverted commits if
   the commit encoding differed from the current locale.

Now commit messages should be displayed correctly, except
when logoutputencoding is set to an encoding that cannot
represent charecters in the message.  For example, it is
impossible to convert Japanese characters from Shift-JIS
to CP-1251 (although the reverse conversion works).

The reason for using git log to read the commit and then getting
Tcl to convert its output is that is essentially what happens in
the normal path through getcommitlines, hence there is less chance
for unintended differences in how commits are processed in
getcommitlines and do_readcommit.
Signed-off-by: NAlexander Gavrilov <angavrilov@gmail.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

Transient windows cause problems on these platforms:

- On Win32 the windows appear in the top left corner
  of the screen. In order to fix it, this patch causes
  them to be explicitly centered on their parents by
  an idle handler.

- On MacOS with Tk 8.4 they appear without a title bar.
  Since it is clearly unacceptable, this patch disables
  transient on that platform.
Signed-off-by: NAlexander Gavrilov <angavrilov@gmail.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

This commit documents keyboard accelerators used for menu
commands in the menu, as it is usually done, and adds some
more, e.g. F4 to invoke Edit View (or New View if the current
view is the un-editable "All files" view).

The changes include a workaround for handling Shift-F4 on
systems where XKB binds special XF86_Switch_VT_* symbols
to Ctrl-Alt-F* combinations.  Tk often receives these codes
when Shift-F* is pressed, so it is necessary to bind the
relevant actions to them as well.
Signed-off-by: NAlexander Gavrilov <angavrilov@gmail.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

Originally gitk required the user to specify all limiting
options manually in the same field with the list of commits.
It is rather unfriendly for new users, who may not know
which options can be used, or, indeed, that it is possible
to specify them at all.

This commit modifies the dialog to present the most useful
options as individual fields.  Note that options that may
be useful to an extent, but produce a severely broken view,
are deliberately not included.

It is still possible to specify options in the commit list
field, but when the dialog is reopened, they will be extracted
into their own separate fields.
Signed-off-by: NAlexander Gavrilov <angavrilov@gmail.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

This adds to the regexps that are used to work out what sort of error
we encountered in trying to do a cherry-pick so that it recognizes
some additional common error messages.  It adds a confirmation dialog
when the error is a merge conflict so the user can choose whether or
not to run git citool.  Finally, it arranges to update the display
after a cherry-pick failed so that any local changes made by the
cherry-pick become visible.
Signed-off-by: NPaul Mackerras <paulus@samba.org>

Now that git-gui has facilities to help users resolve
conflicts, it makes sense to launch it from other GUI
tools when they happen.
Signed-off-by: NAlexander Gavrilov <angavrilov@gmail.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

Transient windows are always kept above their parent, and don't occupy
any space in the taskbar, which is useful for dialogs.  Also, when
transient is used, it is important to bind windows to the correct
parent.

This commit adds transient annotations to all dialogs, ensures usage
of the correct parent for error and confirmation popups, and, as a
side job, makes gitk preserve the create tag dialog window in case of
errors.
Signed-off-by: NAlexander Gavrilov <angavrilov@gmail.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

It is often more convenient to dismiss or accept a dialog using the
keyboard, than by clicking buttons on the screen.  This commit adds
key binding to make it possible with gitk's dialogs.
Signed-off-by: NAlexander Gavrilov <angavrilov@gmail.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

This modifies gettreediffline so that it when we get both a "U" line
and an "M" line for the same file in the output from git diff-files
or git diff-index --cached (used when the user clicks on a fake commit)
we don't add the same filename to the treediff list twice.

This also makes getblobdiffline recognize the "* Unmerged path ..."
lines we get when we ask for the actual diffs, and makes a tiny
optimization in makediffhdr.
Signed-off-by: NPaul Mackerras <paulus@samba.org>

This makes the "Show origin of this line" menu item work correctly
on the fake commits that gitk shows for local uncommitted changes.
With the fake commit for changes that aren't checked in to the index,
we can actually get a 3-way diff shown, which means we might have to
blame either the parent or the commit being merged in (which we get
from .git/MERGE_HEAD).

If the parent is the fake commit which shows the changes that have
been checked in to the index, then we need to get the SHA1 of the blob
for the version of the file that is in the index, then use git cat-file
blob to get the contents of the blob, and give that to git blame with --contents - so that git blame will do the blame on the index version
of the file.  In that case, we might get the all-zeroes SHA1 back from
git blame, meaning that the line is new in the index version of the
file, so then we have to use $nullid2 (the pseudo-SHA1 of the fake
commit for the checked-in changes).
Signed-off-by: NPaul Mackerras <paulus@samba.org>

This adds code to getblobdiffline to make it able to recognize and
display merge diffs (i.e. N-way diffs for N >= 3) as well as normal
two-way diffs.  This means that it can also correctly display the
3-way diff of the local changes when the git repository is in the
middle of a merge with conflicts.

This also removes getmergediffline and changes mergediff to invoke
getblobdiffline rather than getmergediffline.
Signed-off-by: NPaul Mackerras <paulus@samba.org>

This makes it possible for the user to configure the background color
of lines that are "marked".  At the moment only the "show the origin
of this line" function marks lines.  This also makes the user's choice
persistent by saving it in ~/.gitk.
Signed-off-by: NPaul Mackerras <paulus@samba.org>

This adds a menu item to the pop-up menu for the diff display window
which makes gitk find which commit added the line (via git blame)
and show that commit, with the line highlighted with a light-blue
background.
Signed-off-by: NPaul Mackerras <paulus@samba.org>

When walking back from the line where a right-click happened to the
previous hunk separator line to calculate the line number to work on,
we were counting every line including the one clicked on.  That isn't
right; if the user clicked on the line immediately after the hunk
separator then the correct line number would be the one from the
hunk separator.  Therefore this looks at the clicked-on line to work
out which parent to blame (or whether to blame the current commit),
and then looks only at the preceding lines to work out the offset from
the line number in the hunk separator.

This also fixes an off-by-one error when we are showing files rather
than diffs.  In this case diff_menu_filebase is the line number of
the banner showing the file name, so the first line of the file is
at line $diff_menu_filebase + 1.

This also simplifies the code in find_hunk_blamespec a bit and arranges
that we don't pop up the context menu if the user clicks on a file
separator line or a hunk separator line.
Signed-off-by: NPaul Mackerras <paulus@samba.org>

This adds a context menu item to the diff viewer pane that calls git
gui blame, focusing it on the clicked line.  In case of combined
diffs, it also automatically deduces which parent is to be blamed.
Lines added by the diff are blamed on the current commit itself.

The context menu itself is added by this patch.  It would be possible
to add the commands from the flist menu to it.
Signed-off-by: NAlexander Gavrilov <angavrilov@gmail.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

Currently it displays an ugly error box, because the treediffs array
is not filled for such commits.  This fixes it by making
getmergediffline add the filenames it sees to the treediffs array
like gettreediffline does.
Signed-off-by: NAlexander Gavrilov <angavrilov@gmail.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

If gitk knows that the branch the user tries to create exists,
it should ask whether it should overwrite it.  This way the user
can either decide to choose a new name, or move the head while
preserving the reflog.
Signed-off-by: NAlexander Gavrilov <angavrilov@gmail.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

This changes the link detection logic to accept strings of between 6
and 40 hex characters as a possible SHA1 ID of another commit, rather
than insisting on seeing the full 40 hex characters.

To make the logic that turns a possible link into an actual link work
with abbreviated IDs, this changes the way the commitinterest array is
used, and puts the code that deals with it in a pair of new functions.
The commitinterest array is now indexed by just the first 4 characters
of the interesting SHA1 ID, and each element is a list of id + command
pairs.  This also pulls out the logic for expanding an abbreviated
SHA1 to the list of matching full IDs into its own function (the way
it is done is still the same slow way it was done before, which should
be improved some day).

This also fixes the bug where clicking on a link would take you to the
wrong commit if the line number of the target had changed since the
link was made.

This is based on a patch by Linus Torvalds, but totally rewritten by me.
Signed-off-by: NPaul Mackerras <paulus@samba.org>

This is the result of running make update-po and removing or fixing
the strings that were fuzzily matched.  The ones that were fixed were
the ones where the only change was "git rev-list" to "git log", and
the "about gitk" message where the copyright year got updated.

To get xgettext to see the menu labels as needing translation, it
was necessary for arrange for them to be preceded by "mc".  This
therefore changes makemenu to ignore the first element in each
menu item so that it can be "mc" in the makemenu call.
Signed-off-by: NPaul Mackerras <paulus@samba.org>

This is inspired by patches from Robin Rosenberg but takes a different
approach.  This adds a "makemenu" procedure for constructing menus
that allows the menu layout to be specified in a clear fashion, and
provides one place where the alt+letter accelerators can be detected
and handled.

The alt+letter accelerator is specified by putting an ampersand (&)
before the letter for the accelerator in the menu item name.  (Two
ampersands in succession produce one ampersand in the menu item as
it appears on screen.)  This is handled in makemenu.

We also add an mca procedure which is like mc but also does the
ampersand translation, for use when we want to refer to a menu item
by name.  The mca name and the locations where we use it were
shamelessly stolen from Robin Rosenberg's patch.

This doesn't actually add any alt+letter accelerators yet.
Signed-off-by: NPaul Mackerras <paulus@samba.org>

The Return key can now be used as well as pressing the Create button
from the dialog box that is shown when selecting "Create new branch".
Signed-off-by: NRichard Quirk <richard.quirk@gmail.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

This adds a break so that gitk doesn't go and execute the global
binding for <Return> (i.e. find next) when the user presses the
return key in the sha1 entry field to indicate that gitk should
jump to the commit identified by what they just put into the
sha1 field.
Signed-off-by: NPaul Mackerras <paulus@samba.org>

This adds an option allowing the user to select whether gitk should
look up per-file encoding settings using git check-attr or not.  If
not, gitk uses the global encoding set in the git config (as reported
by git config --get gui.encoding) for all files, or if that is not
set, then the system encoding.

The option is controlled by a checkbox in the Edit->Preferences
window, and defaults to off for now because git check-attr is so
slow.  When the user turns it on we discard any cached diff file
lists in treediffs, because we may not have encodings cached for
the files listed in those lists, meaning that getblobdiffline will
do it for each file, which will be really really slow.

This adjusts the limit of how many paths cache_gitattr passes to each
instance of git check-attr depending on whether we're running under
windows or not.  Passing only 30 doesn't effectively amortize the
startup costs of git check-attr, but it's all we can do under windows
because of the 32k limit on arguments to a command.  Under other OSes
we pass up to 1000.

Similarly we adjust how many lines gettreediffline processes depending
on whether we are doing per-file encodings so that we don't run for
too long.  When we are, 500 seems to be a reasonable limit, leading
to gettreediffline taking about 60-70ms under Linux (almost all of
which is in cache_gitattr, unfortunately).  This means that we can
take out the update call in cache_gitattr.

This adds a simple cache on [tclencoding].  Now that we get repeated
calls to translate the same encoding, this is useful.

This reindents the new code added in the last couple of commits to
conform to the gitk 4-space indent and makes various other improvements:
use regexp in gitattr and cache_gitattr instead of split + join + regsub,
make gui_encoding be the value from [tclencoding] to avoid having to
do [tcl_encoding $gui_encoding] in each call to get_path_encoding,
and print a warning message at startup if $gui_encoding isn't
supported by Tcl.
Signed-off-by: NPaul Mackerras <paulus@samba.org>

When the diff contains thousands of files, calling git-check-attr once
per file is very slow.  With this patch gitk does attribute lookup in
batches of 30 files while reading the diff file list, which leads to a
very noticeable speedup.

It may be possible to reimplement this even more efficiently, if
git-check-attr is modified to support a --stdin-paths option.
Additionally, it should quote the ':' character in file paths, or
provide a more robust way of column separation.
Signed-off-by: NAlexander Gavrilov <angavrilov@gmail.com>
Tested-by: NJohannes Sixt <johannes.sixt@telecom.at>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

This allows the encoding to be specified for file contents and used
when displaying files and diffs in the bottom-left pane.  When
displaying diffs, the encoding for each diff hunk is that for the file
that the diff hunk is from, so it can change through the course of the
diff.

The encoding for file contents is determined as follows:

- File encoding defaults to the system encoding.
- It can be overridden by setting the gui.encoding option.
- Finally, the 'encoding' attribute is checked on
  per-file basis; it has the last word.

Note: Since git-check-attr does not provide support for reading
attributes from trees, attribute lookup is done using files from the
working directory.

This also extends the range of supported encoding names, adding
ShiftJIS and Shift-JIS as aliases for Shift_JIS, and allowing
cp-*, cp_*, ibm-*, ibm_*, jis-* and jis_* as aliases for cp*,
ibm* and jis* respectively.

This also fixes some bugs in handling of non-ASCII filenames.  Core
git apparently supports only locale-encoded filenames, so processing
is done using the system encoding.
Signed-off-by: NAlexander Gavrilov <angavrilov@gmail.com>
Tested-by: NJohannes Sixt <johannes.sixt@telecom.at>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

Signed-off-by: NChristian Stimming <stimming@tuhh.de>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

To reproduce: expand a tree like this, then collapse A:

+A
 +B
   C
   D

The result is:

-A
   C
   D

I.e. sub-nodes expanded from the last sub-node of the item
being collapsed are not removed.
Signed-off-by: NAlexander Gavrilov <angavrilov@gmail.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

An OSX user has reported that gitk's context menus are not usable
under OSX because it doesn't provide a way to generate <Button-3>
events.  Users can generate <Button-2> events with command+click,
so use that for the context menus instead on OSX.
Signed-off-by: NPaul Mackerras <paulus@samba.org>

This adds a new item to the file list popup menu, that calls git gui
blame for the selected file, starting with the first parent of the
current commit.
Signed-off-by: NAlexander Gavrilov <angavrilov@gmail.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

Other GUI tools may need to start gitk and make it automatically
select a certain commit.  This adds a new command-line option
--select-commit=id to make that possible.
Signed-off-by: NAlexander Gavrilov <angavrilov@gmail.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

Originally dorunq assumed that the queue entry remained first
in the queue after the script eval, and blindly removed it.
However, if the handler calls nukefile, it may not be the
case anymore, and a random queue entry gets dropped instead.

This makes dorunq remove the entry before calling the
script, and adds a global variable to allow other functions
to determine if they are called from within a dorunq handler.
Signed-off-by: NAlexander Gavrilov <angavrilov@gmail.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

Signed-off-by: NPaul Mackerras <paulus@samba.org>

This includes suggestions by Stephan Beyer.
Signed-off-by: NChristian Stimming <stimming@tuhh.de>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

Try selecting the head, if the previously selected commit
is not available in the new view.
Signed-off-by: NAlexander Gavrilov <angavrilov@gmail.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

- Switching views now actually preserves the selected commit.
- Reloading (also Edit View) preserves the currently selected commit.
- Initial selection does not produce weird scrolling.
Signed-off-by: NAlexander Gavrilov <angavrilov@gmail.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

If the tree diff command failed to start for some
random reason, treepending remained set, and thus
no more diffs were shown after that.
Signed-off-by: NAlexander Gavrilov <angavrilov@gmail.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

MSysGit compiles git binaries as native Windows executables,
so they cannot be killed unless a special flag is specified.

This flag is implemented by the Cygwin version of kill,
which is also included in MSysGit.
Signed-off-by: NAlexander Gavrilov <angavrilov@gmail.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

Local change analysis can take a noticeable amount of time on large
file sets, and produce no output if there are no changes.  Register
the back-ends in commfd, so that they get properly killed on window
close.
Signed-off-by: NAlexander Gavrilov <angavrilov@gmail.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

When collecting commits for a rarely changed, or recently
created file or directory, rev-list may work for a noticeable
period of time without producing any output.  Such processes
don't receive SIGPIPE for a while after gitk is closed, thus
becoming runaway CPU hogs.
Signed-off-by: NAlexander Gavrilov <angavrilov@gmail.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>