Dirkjan Ochtman <dirkjan@ochtman.nl> [Mon, 03 Nov 2008 16:48:23 +0100] rev 502
patch: turn patch.diff() into a generator
This should even be a little faster than passing in an fp argument.
[ original upstream description ]
Christian Ebert <blacktrash@gmx.net> [Thu, 23 Oct 2008 10:25:24 +0200] rev 501
Merge with stable
Matt Mackall <mpm@selenic.com> [Wed, 22 Oct 2008 17:34:52 -0500] rev 500
extensions: use new wrapper functions
[ original upstream description]
Christian Ebert <blacktrash@gmx.net> [Sat, 18 Oct 2008 20:51:12 +0200] rev 499
Merge with stable
Dirkjan Ochtman <dirkjan@ochtman.nl> [Sat, 18 Oct 2008 16:56:39 +0200] rev 498
help: better documentation intro for a few extensions
[ original upstream description ]
Dirkjan Ochtman <dirkjan@ochtman.nl> [Sat, 18 Oct 2008 16:50:03 +0200] rev 497
show enabled extensions in hg help
[ original upstream description ]
Peter Gervai <grin@grin.hu> [Wed, 15 Oct 2008 15:52:34 +0200] rev 496
untest
Peter Gervai <grin@grin.hu> [Wed, 15 Oct 2008 15:16:15 +0200] rev 495
cross them fingahs
Christian Ebert <blacktrash@gmx.net> [Tue, 02 Sep 2008 08:48:48 -0400] rev 494
Merge with stable
Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Tue, 02 Sep 2008 01:19:12 +0200] rev 493
fix test output broken by a04d8cadb6af
Christian Ebert <blacktrash@gmx.net> [Sun, 24 Aug 2008 23:44:52 +0200] rev 492
Merge with stable
Gilles Moris <gilles.moris@free.fr> [Fri, 22 Aug 2008 22:32:53 +0200] rev 491
Have verbose and debug flag print the changeset rev and hash when committing.
Besides updating the tests output, I had to push back at the end the test for
issue1193 in test-commit, solved by cset 4386a7706828, because it was using the
current date for commit and was randomizing the subsequent hash.
[ original upstream description ]
Patrick Mezard <pmezard@gmail.com> [Sun, 24 Aug 2008 22:13:39 +0200] rev 490
Backed out changeset d557749c627a
Replace this change with a better one.
[ original upstream description ]
Christian Ebert <blacktrash@gmx.net> [Sun, 24 Aug 2008 20:13:45 +0200] rev 489
(0.9.2compat) remove new verbose commit message for backwards compatibility
Christian Ebert <blacktrash@gmx.net> [Sun, 24 Aug 2008 20:02:02 +0200] rev 488
(stable) merge with default
Gilles Moris <gilles.moris@free.fr> [Fri, 22 Aug 2008 13:25:35 +0200] rev 487
Have verbose flag print the changeset hash when committing
[ original upstream message ]
Christian Ebert <blacktrash@gmx.net> [Fri, 22 Aug 2008 11:53:30 +0200] rev 486
(0.9.2compat) add merge, resolve to restricted
This is not really backwards compatible (test only as comment)
and doesn't prevent extra conflicts with older Hg versions.
On the other hand it doesn't do any damage either and provides
the 3-way merge ease if possible.
Christian Ebert <blacktrash@gmx.net> [Fri, 22 Aug 2008 08:26:18 +0200] rev 485
Avoid additional conflicts during merge/resolve
Make merge and resolve trigger kwtemplater.restricted to compare
data without keyword expansion.
The keyword stays outside the conflict:
$Keyword$
<<<<<<< local
bar
=======
foo
>>>>>>> other
and will again be expanded on commit.
Demonstrate in test case.
Christian Ebert <blacktrash@gmx.net> [Thu, 21 Aug 2008 21:32:06 +0200] rev 484
(stable) merge with default
Christian Ebert <blacktrash@gmx.net> [Thu, 21 Aug 2008 19:07:33 +0200] rev 483
(0.9.2compat) fix test compatibility
Comment out non-backwards compatible hgweb test (at default position).
Make kwexpand nonexistent work on Windows (originally by pmezard).
Christian Ebert <blacktrash@gmx.net> [Thu, 21 Aug 2008 10:20:44 +0200] rev 482
Thorough hgweb testing
- test hgweb while keyword extension is _enabled_
- check expansion with hgweb.file
- besides log and diff, also check annotate for not expanding
Christian Ebert <blacktrash@gmx.net> [Thu, 07 Aug 2008 11:20:34 +0200] rev 481
(stable) merge with default
Christian Ebert <blacktrash@gmx.net> [Wed, 30 Jul 2008 20:05:14 +0200] rev 480
(0.9.2compat) add verify to nokwcommands
Christian Ebert <blacktrash@gmx.net> [Wed, 30 Jul 2008 16:26:38 +0200] rev 479
Add verify to nokwcommands after refactor in e79a8f36c2a5
Add test as well.
Thanks to Guy Brand for spotting this.
Christian Ebert <blacktrash@gmx.net> [Mon, 30 Jun 2008 12:42:48 +0200] rev 478
(0.9.2compat) context, status refactor keeping backwards compatible fallback
Christian Ebert <blacktrash@gmx.net> [Mon, 30 Jun 2008 12:33:36 +0200] rev 477
(0.9.2compat) ddb8a222249b is backwards compatible
Christian Ebert <blacktrash@gmx.net> [Sat, 28 Jun 2008 15:49:54 +0200] rev 476
Merge with stable
Matt Mackall <mpm@selenic.com> [Fri, 27 Jun 2008 18:28:45 -0500] rev 475
util: add sort helper
[ original upstream description]
Matt Mackall <mpm@selenic.com> [Fri, 27 Jun 2008 13:43:29 -0500] rev 474
status: clean up all users for unknown files
[ original upstream description]
Matt Mackall <mpm@selenic.com> [Thu, 26 Jun 2008 14:35:50 -0500] rev 473
repo.status: eliminate list_
[ original upstream description]
Matt Mackall <mpm@selenic.com> [Thu, 26 Jun 2008 14:35:50 -0500] rev 472
add __len__ and __iter__ methods to repo and revlog
[ original upstream description]
Matt Mackall <mpm@selenic.com> [Thu, 26 Jun 2008 14:35:50 -0500] rev 471
manifest: remove execf/linkf methods
[ original upstream description]
Matt Mackall <mpm@selenic.com> [Thu, 26 Jun 2008 14:35:46 -0500] rev 470
use repo[changeid] to get a changectx
[ original upstream description]
Matt Mackall <mpm@selenic.com> [Thu, 26 Jun 2008 13:46:34 -0500] rev 469
simplify flag handling
add _checklink var to dirstate
introduce dirstate.flagfunc
switch users of util.execfunc/linkfunc to flagfunc
change manifestdict.set to take a flags string
change ctx.fileflags to ctx.flags
change gitmode func to a dict
remove util.execfunc/linkfunc
[ original upstream description]
Matt Mackall <mpm@selenic.com> [Thu, 26 Jun 2008 13:46:29 -0500] rev 468
use repo.changectx(None) to get a workingctx
[ original upstream description]
Matt Mackall <mpm@selenic.com> [Wed, 25 Jun 2008 17:35:20 -0500] rev 467
context: avoid using None for working parent
[ original upstream description ]
Matt Mackall <mpm@selenic.com> [Wed, 25 Jun 2008 17:34:28 -0500] rev 466
use repo.changectx rather than context.changectx
[ original upstream description ]
Christian Ebert <blacktrash@gmx.net> [Thu, 12 Jun 2008 02:27:06 +0200] rev 465
(stable) mimic cmdutil.bail_if_changed even more
At the check before overwriting with kwexpand, kwshrink:
- abort when in middle of merge
- give terser abort message
Christian Ebert <blacktrash@gmx.net> [Fri, 13 Jun 2008 17:41:26 +0200] rev 464
(0.9.2compat) secure keyword arguments in backwards compatible wrappers
Explicitly assign _all_ keyword args when TypeErrors must be detected.
Christian Ebert <blacktrash@gmx.net> [Thu, 12 Jun 2008 11:10:04 +0200] rev 463
keyword: disable expansion for annotate
Keyword expansion should not be annotated.
Annotate revision when unexpanded keyword was inserted.
Add test.
Christian Ebert <blacktrash@gmx.net> [Thu, 12 Jun 2008 16:32:16 +0200] rev 462
(0.9.2compat) disable expansion for annotate
Christian Ebert <blacktrash@gmx.net> [Thu, 12 Jun 2008 11:10:04 +0200] rev 461
keyword: disable expansion for annotate
Keyword expansion should not be annotated.
Annotate revision when unexpanded keyword was inserted.
Add test.
Christian Ebert <blacktrash@gmx.net> [Thu, 12 Jun 2008 02:35:38 +0200] rev 460
Merge after backout
Revert to dispatch._parse for grabbing current command name.
Christian Ebert <blacktrash@gmx.net> [Thu, 12 Jun 2008 02:33:41 +0200] rev 459
Backed out changeset 10ebb9f44b99
Tests with "hg annotate" reveal that dispatch._runcommand is
unreliable for our purpose.
Christian Ebert <blacktrash@gmx.net> [Fri, 06 Jun 2008 17:21:00 +0200] rev 458
(0.9.2compat) get some changes from default branch
We stay with old-fashioned command _parsing because a _*runcommand
method was introduced after 0.9.2
Christian Ebert <blacktrash@gmx.net> [Fri, 06 Jun 2008 17:08:56 +0200] rev 457
keyword: mimic cmdutil.bail_if_changed even more
At the check before overwriting with kwexpand, kwshrink:
- abort when in middle of merge
- give terser abort message
Christian Ebert <blacktrash@gmx.net> [Sun, 01 Jun 2008 13:35:12 +0200] rev 456
keyword: use dirstate.parents to obtain current first parent
This should be quicker than changectx().node()
Christian Ebert <blacktrash@gmx.net> [Sun, 01 Jun 2008 23:47:32 +0200] rev 455
keyword: monkeypatch dispatch._runcommand instead of _parse
We want the current command, and hooking into this method
makes it more transparent what we are doing.