Stuff that ended up in the result is shown in bold with a "+" at the
beginning of the line; stuff that didn't is in the normal font with
a "-" at the beginning of the line.  The color shows which parent
the stuff was in; red for the first parent, blue for the second, then
green, purple, brown, and the rest are grey.  If the result is different
from all of the parents it is shown in black (and bold).

If there is a GCA for the parents of the merge, then a file is
interesting if some parent has a version that is different from both
the child and the GCA.  If there is no GCA (e.g. for a merge that
pulls in an external project) then a file is interesting if the child's
version is different from all of the parents.

Next step is to actually show the differences for the interesting
files...

The old style is still available as an option (we still need a
preferences window so we can set/clear it though).

Also simplified the parsing of the git-diff-tree -p output and got
rid of the unused 'seenfile' variable.

We display the files where the child differs from the first parent
first in black in the file list window, followed by the second parent
in blue, and so on using different colors for each parent (provided
you don't have more than 6 parents; if you do we cycle around).

We now kick off a single git-diff-tree -r --stdin and feed it all
the commit pairs we want to know about, instead of doing a separate
git-diff-tree invocation for each.

"Files" matches the find string against each of the files modified
by each commit, and can do exact, case-ignoring or regexp matching.

"Pickaxe" uses git-diff-tree -S'string' and can only do exact
matching.  I called it "pickaxe" rather than "find within patch"
since it only finds commits where the string is present in the child
but not the parents or vice versa, and "pickaxe" is what the author
of that feature calls it.

This just invokes git-diff-tree.  Also changed the readonly entry
widgets in the "write patch" and "generate tag" windows to have
flat relief, so it doesn't look like you should be able to modify
what's in them.

If we have a very long commit message, and we end up getting a
bufferfull of data from git-rev-list that all belongs to one commit,
we ended up throwing away the data from a previous read that should
have been included.  The result was a error message about not being
able to parse the output of git-rev-list.

Also, if the git-rev-list output that we can't parse is long, only put
the first 80 chars in the error message.  Otherwise we end up with an
enormous error window.

I had code in there to put "No commits selected" on the canvas
but it needed some globals.

Check that $GIT_DIR (or .git, if GIT_DIR is not set) is a directory.
This means we can give a more informative error message if the user
runs gitk somewhere that isn't a git repository.

If the user pastes in the selection (with the middle mouse button) and
it already has 40 characters in it, clear it before pasting.

Right-click on a context row now brings up a menu allowing the user to
generate a diff between that row and the selected row.  Left-click on
a graph line shows the parent and children connected by the line in
the details pane.  Left-click on a circle in the graph selects that
commit.  Left-click elsewhere in the graph does nothing.

When displaying a diff, the bottom-right file list box behaves
slightly differently now; instead of eliding all other files' diffs,
it now just scrolls the details pane so that the selected file's diff
starts at the top of the pane.

Since the diffs can be rather large, arrange for an update to be done
every 100ms while reading diffs.

Also removed the CVS revision keywords and bumped the version number
to 1.2.

In particular try hard to give different colors to lines that cross
at a corner in one of the lines.

This involves adding indentation when we read a commit with
git-cat-file and trimming the whitespace from the headline.

With --header, git-rev-list gives us the contents of the commit
in-line, so we don't need to exec a git-cat-file to get it, and we
don't need the readobj command either.
Also fixed a residual problem with handling the commit that
has a parent listed twice.

This allows the user to specify ranges more flexibly; for instance
the user can now do "gitk v2.6.12.." and see all the changes since
2.6.12.

This happens in the linux-2.6 tree.  We draw the graph line
double-thick to show that this happened.

Also fix a bug where we got a bogus "No commit information available"
line at the end on simple repositories like this one.

Added context menu on lines
Added headline display when the mouse hovers over a line
Removed some debug messages

Display a watch cursor when reading the commits initially

Some speedups from not doing update so often

Fix bug in changing font size in entry widgets.
Fix bug with B1 click before anything has been drawn.
Use "units" and "pages" instead of "u" and "p" for tk8.5.

with a label.
Allow SHA1 ids or tags to be entered in the SHA1 ID field.

Add 'f' key for moving to next file

Put (deleted) or (created, mode xxx) in header lines
Fix scrolling to bring lines on screen

Make up/down/pgup/pgdn work again
Return in find string entry does find
Scale circles and lines with font size
Fix scrolling to make entire selected line visible
Use white circle for commits not listed but put in to terminate lines
Fix diff parsing for created and deleted files

Stop . bindings firing on find string entry keypresses
Fix geometry saving/restoring a bit
Show the terminal commits
Highlight comment matches in the comment window

to keep the proportionality of the pane widths as the overall
window is resized.

Added control+/- to increase/decrease font sizes
Rearranged code a little.

Add a find facility to search within the commits
Cope with multiple starting points.

Fix crash benh saw with currentid undefined
Add menu with file/quit and help/about items
Add ^Q for quit