Brodie Rao <brodie@bitheap.org> [Sun, 12 Sep 2010 18:05:53 -0500] rev 792
merge: suggest 'hg up -C .' for discarding changes, not 'hg up -C'
Without specifying the parent revision of the working copy, users will
update to tip, which is most likely the other head they were trying to
merge, not the revision they were at before the merge.
[ original upstream message ]
Christian Ebert <blacktrash@gmx.net> [Fri, 10 Sep 2010 21:22:11 +0200] rev 791
(stable) merge
Christian Ebert <blacktrash@gmx.net> [Fri, 10 Sep 2010 16:58:13 +0200] rev 790
Explain file-wise expansion in help
Christian Ebert <blacktrash@gmx.net> [Fri, 03 Sep 2010 15:59:08 +0100] rev 789
Merge with stable
Martin Geisler <mg@lazybytes.net> [Fri, 03 Sep 2010 12:58:51 +0200] rev 788
diff: recurse into subrepositories with --subrepos/-S flag
[ original upstream message ]
Christian Ebert <blacktrash@gmx.net> [Tue, 31 Aug 2010 18:18:28 +0100] rev 787
Merge with stable
Nicolas Dumazet <nicdumz.commits@gmail.com> [Mon, 30 Aug 2010 13:15:30 +0900] rev 786
test-keyword: use regular expressions instead of grepping
[ original upstream message]
Christian Ebert <blacktrash@gmx.net> [Mon, 16 Aug 2010 01:01:41 +0100] rev 785
Merge with stable
Matt Mackall <mpm@selenic.com> [Sun, 15 Aug 2010 13:26:12 -0500] rev 784
tests: drop big sed from test-keyword.t
[ original upstream message ]
Christian Ebert <blacktrash@gmx.net> [Mon, 16 Aug 2010 01:00:29 +0100] rev 783
(stable) merge
Christian Ebert <blacktrash@gmx.net> [Sat, 14 Aug 2010 09:47:57 +0100] rev 782
tests: unify test-keyword
Christian Ebert <blacktrash@gmx.net> [Sat, 14 Aug 2010 09:47:41 +0100] rev 781
Merge with stable
Matt Mackall <mpm@selenic.com> [Fri, 14 May 2010 10:01:09 -0500] rev 780
rollback: fix up tests
[ original upstream message ]
Christian Ebert <blacktrash@gmx.net> [Sat, 14 Aug 2010 09:42:14 +0100] rev 779
Merge after backout
Christian Ebert <blacktrash@gmx.net> [Sat, 14 Aug 2010 09:30:42 +0100] rev 778
Backed out changeset bceb04857ce1 (not all upstream hunks applied)
Nicolas Dumazet <nicdumz.commits@gmail.com> [Thu, 12 Aug 2010 14:53:34 +0900] rev 777
tests: catch re.error if test line is not a valid regular expression
[ original upstream message ]
Matt Mackall <mpm@selenic.com> [Mon, 02 Aug 2010 23:27:22 -0500] rev 776
tests: basic support for unified tests
[ original upstream message ]
Matt Mackall <mpm@selenic.com> [Mon, 02 Aug 2010 15:44:54 -0500] rev 775
tests: move script execution in runner helpers
[ original upstream message ]
Christian Ebert <blacktrash@gmx.net> [Mon, 26 Jul 2010 12:34:33 +0200] rev 774
(stable) merge
Christian Ebert <blacktrash@gmx.net> [Mon, 26 Jul 2010 12:32:45 +0200] rev 773
Move collecting of [keyword] patterns to reposetup (issue2303)
When cloning, prevent [keyword] filename patterns configured locally
in the source directory to persist during the update in the destination.
a) move [keyword] retrieval (back) to reposetup
b) remove the corresponding global kwtools attributes
Add test cases.
Christian Ebert <blacktrash@gmx.net> [Mon, 19 Jul 2010 08:13:48 +0100] rev 772
(stable) merge
Christian Ebert <blacktrash@gmx.net> [Thu, 15 Jul 2010 10:24:31 +0200] rev 771
Add extra datefilters in a single update call
Christian Ebert <blacktrash@gmx.net> [Wed, 16 Jun 2010 11:08:43 +0200] rev 770
(stable) merge
Christian Ebert <blacktrash@gmx.net> [Tue, 15 Jun 2010 21:58:53 +0200] rev 769
Postpone manifest calculation in kwtemplater.overwrite
We can check for file existence in the working directory (needed
in case of recording) by simply using the given context and
calculate the manifest only when there are in fact candidates
for expansion/shrinking.
Christian Ebert <blacktrash@gmx.net> [Thu, 10 Jun 2010 11:33:01 +0100] rev 768
Merge with stable
FUJIWARA Katsunori <foozy@lares.dti.ne.jp> [Sun, 06 Jun 2010 17:25:00 +0900] rev 767
help: show value requirement and multiple occurrence of options
this helps users to know what kind of option is:
- no value is required(flag option)
- value is required
- value is required, and multiple occurrences are allowed
each kinds are shown as below:
-f --force force push
-e --ssh CMD specify ssh command to use
-b --branch BRANCH [+] a specific branch you would like to push
if one or more 3rd type options are shown, explanation for '[+]' mark
is also shown as footnote.
[ original upstream message ]
Christian Ebert <blacktrash@gmx.net> [Wed, 09 Jun 2010 22:50:30 +0100] rev 766
(stable) merge
Christian Ebert <blacktrash@gmx.net> [Tue, 01 Jun 2010 01:05:45 +0200] rev 765
Force dirstate normal when all changes in a file are recorded
Before this bugfix a file whose changes were entirely recorded was still
considered modified by "hg status".
Note: the test must use hg record -l/--logfile, because this is not
reproducible with hg record -m/--message.
Christian Ebert <blacktrash@gmx.net> [Wed, 09 Jun 2010 17:21:12 +0100] rev 764
Pass context to kwtemplater.overwrite
Now that we have retrieved the context in every calling function
except commit, pass it as argument to kwtemplater.overwrite to
avoid looking it up twice.
Reorder arguments to kwtemplater.overwrite to reflect their
importance.
Turn node argument into a simple boolean and rename it to iswctx.
Christian Ebert <blacktrash@gmx.net> [Tue, 01 Jun 2010 03:45:10 +0200] rev 763
Retrieve added and modified files from commitctx
Christian Ebert <blacktrash@gmx.net> [Tue, 08 Jun 2010 09:55:08 +0100] rev 762
Merge with stable
Dirkjan Ochtman <dirkjan@ochtman.nl> [Mon, 07 Jun 2010 20:03:32 +0200] rev 761
move working dir/dirstate methods from localrepo to workingctx
[ original upstream message ]
Christian Ebert <blacktrash@gmx.net> [Sun, 06 Jun 2010 20:19:24 +0100] rev 760
(stable) merge
Christian Ebert <blacktrash@gmx.net> [Thu, 03 Jun 2010 10:37:31 +0100] rev 759
Use context to detect uncommitted merge
Christian Ebert <blacktrash@gmx.net> [Sat, 22 May 2010 14:10:49 +0200] rev 758
(stable) merge
Christian Ebert <blacktrash@gmx.net> [Fri, 21 May 2010 22:12:39 +0200] rev 757
Offer svn-like default keywordmaps
svn-like default keywords can be set in a new configuration section
called [keywordset] -- thanks to timeless for the name.
Move setup of default keywordmaps into dedicated function used by
kwtemplater.__init__ and demo.
HeadURL/URL is not supported (by default).
Christian Ebert <blacktrash@gmx.net> [Fri, 21 May 2010 22:12:39 +0200] rev 756
Add 2 svn-like date filters
svnisodate yields the format the date in svn's Id keyword expands to.
svnutcdate yields the format svn's Date/LastChangedDate expands to.
http://svnbook.red-bean.com/en/1.5/svn.advanced.props.special.keywords.html
Christian Ebert <blacktrash@gmx.net> [Tue, 18 May 2010 22:22:59 +0200] rev 755
Merge with stable
Martin Geisler <mg@aragost.com> [Tue, 18 May 2010 16:31:10 +0200] rev 754
Use our custom hg reStructuredText role some more
I missed these occurrences on my first scan through the source.
[ original upstream message ]
Christian Ebert <blacktrash@gmx.net> [Fri, 14 May 2010 17:15:38 +0200] rev 753
Merge with stable
Matt Mackall <mpm@selenic.com> [Fri, 14 May 2010 10:01:09 -0500] rev 752
rollback: fix up tests
[ original upstream message ]
Christian Ebert <blacktrash@gmx.net> [Fri, 14 May 2010 10:56:47 +0200] rev 751
(stable) merge
Matt Mackall <mpm@selenic.com> [Thu, 13 May 2010 17:24:21 -0500] rev 750
commit: note new branch heads rather than topological heads
Move to using contexts while we're at it.
[ original upstream message ]
Christian Ebert <blacktrash@gmx.net> [Sun, 09 May 2010 11:14:43 +0200] rev 749
Support extensions using dorecord, e.g. crecord
Provide extendable keyword.recordextensions variable, so other
extensions beside hgext.record which provide the dorecord function
can cooperate with hgext.keyword like so (example from crecord):
def extsetup():
try:
keyword = extensions.find('keyword')
keyword.restricted += ' crecord qcrecord'
try:
# use record support in keyword.py if present
keyword.recordcommands += ' crecord qcrecord'
keyword.recordextensions += ' crecord'
except AttributeError:
pass
except KeyError:
pass
Christian Ebert <blacktrash@gmx.net> [Wed, 05 May 2010 17:16:00 +0200] rev 748
(stable) merge
Christian Ebert <blacktrash@gmx.net> [Wed, 05 May 2010 14:02:45 +0200] rev 747
Omit setting extra variable for record context
Since dc2f37864348 the context is always retrieved in
kwtemplater.overwrite().
Christian Ebert <blacktrash@gmx.net> [Tue, 04 May 2010 01:15:34 +0200] rev 746
(stable) merge
Christian Ebert <blacktrash@gmx.net> [Sun, 02 May 2010 22:39:43 +0200] rev 745
Simplify record switch in kwtemplater.overwrite
1) use kwtemplater.record attribute for clarity
2) drop optional context argument; consider the speed loss by
duplicating the dictionary lookup repo['.'] as negligible
Christian Ebert <blacktrash@gmx.net> [Sat, 01 May 2010 23:18:42 +0200] rev 744
(stable) merge
Christian Ebert <blacktrash@gmx.net> [Sat, 01 May 2010 20:49:40 +0200] rev 743
Cleanup test and make it portable
- replace sed call with python command
- no need to back up hgrc before record
Christian Ebert <blacktrash@gmx.net> [Fri, 30 Apr 2010 16:30:09 +0200] rev 742
Merge with stable
Matt Mackall <mpm@selenic.com> [Thu, 29 Apr 2010 22:04:05 -0500] rev 741
run-tests: add --view switch to use external diff viewer
[original upstream message ]
Matt Mackall <mpm@selenic.com> [Thu, 29 Apr 2010 18:25:45 -0500] rev 740
run-tests: sort options
[original upstream message ]
Matt Mackall <mpm@selenic.com> [Thu, 29 Apr 2010 18:25:45 -0500] rev 739
run-tests: add -l short option for --local
[original upstream message ]
Christian Ebert <blacktrash@gmx.net> [Fri, 30 Apr 2010 15:21:59 +0200] rev 738
(stable) merge
Christian Ebert <blacktrash@gmx.net> [Mon, 26 Apr 2010 04:01:13 +0200] rev 737
Test recording
Christian Ebert <blacktrash@gmx.net> [Mon, 26 Apr 2010 04:01:07 +0200] rev 736
Support (q)record
Monkeypatch hgext.dorecord to trigger keyword expansion.
Read data from working directory, not from filelog.
Prevent keyword expansion from within record's commitfunc,
thereby fixing a bug/inconsistency where files which are clean
after recording were overwritten twice.
Christian Ebert <blacktrash@gmx.net> [Mon, 26 Apr 2010 03:54:18 +0200] rev 735
Remove mq commands from restricted list
Monkeypatching patch.diff takes care of this since 911f5be5d159.
Test mq more thoroughly by loosening [keywordmaps] and comparing
the output of hg cat with keyword expansion enabled and disabled.
Christian Ebert <blacktrash@gmx.net> [Fri, 16 Apr 2010 14:30:13 +0200] rev 734
Replace deprecated mq commands in test
Christian Ebert <blacktrash@gmx.net> [Fri, 23 Apr 2010 10:37:08 +0200] rev 733
Merge with stable
Martin Geisler <mg@aragost.com> [Thu, 22 Apr 2010 10:24:49 +0200] rev 732
Use hg role in help strings
[ original upstream message ]
Christian Ebert <blacktrash@gmx.net> [Tue, 20 Apr 2010 02:37:20 +0200] rev 731
Merge with stable
Augie Fackler <durin42@gmail.com> [Sat, 17 Apr 2010 13:38:42 -0500] rev 730
keyword: monkeypatch patch so that optional args can be passed as kwargs
[ original upstream message ]
Christian Ebert <blacktrash@gmx.net> [Thu, 15 Apr 2010 22:16:14 +0200] rev 729
Merge with stable
Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Thu, 15 Apr 2010 20:25:07 +0200] rev 728
run-tests.py: can't remove from os.environ on solaris
[ original upstream message ]
Christian Ebert <blacktrash@gmx.net> [Wed, 14 Apr 2010 15:23:22 +0200] rev 727
Merge with stable
Nicolas Dumazet <nicdumz.commits@gmail.com> [Wed, 14 Apr 2010 17:58:10 +0900] rev 726
pylint, pyflakes: remove unused or duplicate imports
[ original upstream message ]
Ry4an Brase <ry4an-hg@ry4an.org> [Wed, 14 Apr 2010 00:24:47 -0500] rev 725
Fix --blacklist when --jobs > 1 in run_tests.py.
The options parsing logic replaces the options.blacklist initial value,
a list of filenames, with a dict of test names to filenames. When
runchildren rebuilds the command line to launch child processes the dict
is flattened and provided as a malformed argument when a filename is
expected.
Fix is to remove the blacklist option from child invocations since the
filtering is handled in the parent anyway.
[ original upstream message ]
Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Wed, 14 Apr 2010 09:08:47 +0200] rev 724
run-tests.py: reset env variables set by hooks
[ original upstream message ]
Christian Ebert <blacktrash@gmx.net> [Tue, 13 Apr 2010 10:28:48 +0200] rev 723
Merge with stable
Ronny Pfannschmidt <Ronny.Pfannschmidt@gmx.de> [Tue, 23 Mar 2010 15:23:04 +0100] rev 722
run-tests: force to test pure on pypy as well
[ original upstream message ]
Ronny Pfannschmidt <Ronny.Pfannschmidt@gmx.de> [Tue, 23 Mar 2010 15:22:17 +0100] rev 721
Fix run-tests.py -jX after 2ed667a9dfcb
[ original upstream message ]
Ronny Pfannschmidt <Ronny.Pfannschmidt@gmx.de> [Tue, 23 Mar 2010 11:37:31 +0100] rev 720
tests: adapt the test runner to work with jython
[ original upstream message ]
Brodie Rao <brodie@bitheap.org> [Mon, 22 Mar 2010 01:36:12 -0400] rev 719
run-tests: make sure GREP_OPTIONS isn't set
This can interfere with test output, especially for options like --color.
[ original upstream message ]
Dirkjan Ochtman <djc.ochtman@kentyde.com> [Thu, 11 Mar 2010 15:32:24 +0100] rev 718
tests: use external coverage, mandate newer version
This removes the option for including the stdlib in coverage reports.
[ original upstream message ]
Dirkjan Ochtman <dirkjan@ochtman.nl> [Thu, 05 Nov 2009 15:19:54 +0100] rev 717
templater: readable dates older than 24 months revert to ISO8601 (issue1006)
[ original upstream message ]
Christian Ebert <blacktrash@gmx.net> [Tue, 13 Apr 2010 10:13:41 +0200] rev 716
(stable) merge
Christian Ebert <blacktrash@gmx.net> [Tue, 13 Apr 2010 10:13:06 +0200] rev 715
Add forgotten comment
Christian Ebert <blacktrash@gmx.net> [Mon, 12 Apr 2010 17:34:12 +0200] rev 714
Make the templater a local variable
Christian Ebert <blacktrash@gmx.net> [Tue, 13 Apr 2010 10:05:51 +0200] rev 713
Merge with stable
Matt Mackall <mpm@selenic.com> [Mon, 12 Apr 2010 17:21:30 -0700] rev 712
rollback: improve message
[ original upstream message ]
Steve Borho <steve@borho.org> [Fri, 09 Apr 2010 17:23:37 -0500] rev 711
rollback: add dry-run argument, emit transaction description
[ original upstream message ]
Christian Ebert <blacktrash@gmx.net> [Tue, 16 Mar 2010 22:29:16 +0100] rev 710
Do not bother about detecting extension path in demo
Detecting and showing the path to a keyword extension in a
non-standard place only made sense while keyword.py was not
shipped with Mercurial.
The test output has changed because we do not have a spurious
space at eol anymore.
Christian Ebert <blacktrash@gmx.net> [Mon, 12 Apr 2010 14:12:56 +0200] rev 709
Merge with stable
Matt Mackall <mpm@selenic.com> [Mon, 05 Apr 2010 15:25:08 -0500] rev 708
keywords: build a new templater to work around caching interaction
Christian Ebert <blacktrash@gmx.net> [Tue, 30 Mar 2010 11:12:09 +0100] rev 707
(0.9.2compat) improve quoting of environment variable in test
Christian Ebert <blacktrash@gmx.net> [Tue, 30 Mar 2010 11:08:13 +0100] rev 706
Merge with stable
Mads Kiilerich <mads@kiilerich.com> [Wed, 24 Mar 2010 01:43:24 +0100] rev 705
Tests with spaces in paths
This allows most tests to succeed with
./run-tests.py --tmpdir='/tmp/hg test'
and introduces other tests for spaces and shell quoting
[ original upstream message ]
Christian Ebert <blacktrash@gmx.net> [Wed, 17 Mar 2010 23:16:44 +0100] rev 704
(0.9.2compat) get applicable changes from default
Christian Ebert <blacktrash@gmx.net> [Wed, 17 Mar 2010 15:04:07 +0100] rev 703
(stable) merge
Even though this is a fake standalone branch, we do not include
the patch that removed extension path detection in kwdemo.
Christian Ebert <blacktrash@gmx.net> [Tue, 16 Mar 2010 22:28:58 +0100] rev 702
Make kwdemo less verbose
1) Set the branchname always silently with
dirstate.setbranch().
We create a branch so that testing the {branches} template
does not come up empty. But kwdemo is hardly the place to
inform the user by inference why {branches} is empty on the
default branch.
"demobranch" is ascii and cannot be changed, so using the
internal command instead of commands.branch() is safe.
2) Do not show full path to temporary directory
(distracting long lines on Mac OS X).
3) No special debug output. Output only related to keyword,
no internals like unsetting of commit hooks etc.
Christian Ebert <blacktrash@gmx.net> [Tue, 16 Mar 2010 22:28:23 +0100] rev 701
1 variable name for overwriting candidates
Christian Ebert <blacktrash@gmx.net> [Tue, 16 Mar 2010 22:27:22 +0100] rev 700
Do not customize ui.note() when overwriting
Behaviour change:
hg commit --verbose now prints out the names of the files
which are overwritten.
Christian Ebert <blacktrash@gmx.net> [Thu, 18 Feb 2010 05:55:05 +0100] rev 699
Mark improved demo commit message for translation
s/config/configuration/
Inside the quoted command put the message in single quotes in case
the translation contains apostrophes.
Christian Ebert <blacktrash@gmx.net> [Thu, 11 Mar 2010 19:56:41 +0100] rev 698
(0.9.2compat) update copyright _and_ licence
Christian Ebert <blacktrash@gmx.net> [Thu, 11 Mar 2010 19:48:38 +0100] rev 697
(stable) merge
Christian Ebert <blacktrash@gmx.net> [Thu, 11 Mar 2010 18:49:41 +0100] rev 696
Update copyright
Christian Ebert <blacktrash@gmx.net> [Thu, 11 Mar 2010 18:49:14 +0100] rev 695
Remove deprecated options
Christian Ebert <blacktrash@gmx.net> [Sun, 07 Mar 2010 22:57:55 +0100] rev 694
Merge with stable
Mads Kiilerich <mads@kiilerich.com> [Tue, 09 Feb 2010 01:12:29 +0100] rev 693
run-tests.py: skipped tests shouldn't change working directory
[ original upstream message ]
Wagner Bruna <wbruna@softwareexpress.com.br> [Tue, 29 Dec 2009 19:02:26 -0200] rev 692
run-tests.py: clears http_proxy for all tests
[ original upstream message ]
Christian Ebert <blacktrash@gmx.net> [Sun, 07 Mar 2010 22:48:01 +0100] rev 691
(stable) merge
Christian Ebert <blacktrash@gmx.net> [Sun, 07 Mar 2010 21:50:36 +0100] rev 690
(0.9.2compat) upgrade kwfiles behaviour to match default wrt --unknown
Switch long option name from --untracked to --unknown.
Christian Ebert <blacktrash@gmx.net> [Sun, 07 Mar 2010 21:09:04 +0100] rev 689
(0.9.2compat) use status' names for readability
Christian Ebert <blacktrash@gmx.net> [Sun, 07 Mar 2010 20:59:43 +0100] rev 688
Remove spurious locks, improve handling of wlock
- kwcommitctx is inside the wlock of repo.commit: no lock
- _kwfwrite only needs wlock
wlock outside try block, so we don't need to import lock.release
_kwfwrite should even be safer that way, as we moved the status
call inside the try-except block.
Thanks to Benoit Boissinot for help.
Christian Ebert <blacktrash@gmx.net> [Sun, 07 Mar 2010 19:56:37 +0100] rev 687
Use status' names instead of index
Christian Ebert <blacktrash@gmx.net> [Wed, 17 Feb 2010 21:57:52 +0100] rev 686
(stable) merge
Christian Ebert <blacktrash@gmx.net> [Wed, 17 Feb 2010 21:56:15 +0100] rev 685
Do not postpone commit hooks
With 0e64d814d7d0 in place, this hack is not needed any more.
Christian Ebert <blacktrash@gmx.net> [Wed, 17 Feb 2010 11:15:18 +0100] rev 684
Merge with stable
Matt Mackall <mpm@selenic.com> [Tue, 09 Feb 2010 14:12:22 -0600] rev 683
fix up a bunch of check-code warnings
[ original upstream message ]
Brendan Cully <brendan@kublai.com> [Sat, 06 Feb 2010 15:18:36 +0100] rev 682
run-tests: kill daemons on ^C with -j.
This is not perfect, but the problem is fundamentally racy.
[ original upstream message ]
Nicolas Dumazet <nicdumz.commits@gmail.com> [Thu, 10 Dec 2009 17:21:31 +0900] rev 681
run-tests: split tests/blacklist in tests/blacklists/*
Following discussions with Gilles Morris [1], it seems that it is preferable to
use several blacklist files in a blacklists/ directory. It is easier to add an
unversioned file for experiments than modifying a tracked file.
Also fall back to a simpler syntax, giving up ConfigParser, now that section
names are not needed anymore.
And allow --blacklist parameter to be a complete path, instead of only one
of the filenames contained in tests/blacklists/
[1] http://www.selenic.com/pipermail/mercurial-devel/2009-December/017317.html
[ original upstream message ]
Matt Mackall <mpm@selenic.com> [Mon, 25 Jan 2010 00:05:27 -0600] rev 680
many, many trivial check-code fixups
[ original upstream message ]
Ronny Pfannschmidt <Ronny.Pfannschmidt@gmx.de> [Tue, 16 Feb 2010 21:04:04 +0100] rev 679
filelog: sort meta entries, ensure deterministic order
[ original upstream message ]
Christian Ebert <blacktrash@gmx.net> [Fri, 05 Feb 2010 18:08:07 +0000] rev 678
(0.9.2compat) fix utcdate filter
Christian Ebert <blacktrash@gmx.net> [Fri, 05 Feb 2010 17:57:36 +0000] rev 677
Merge with stable
Christian Ebert <blacktrash@gmx.net> [Fri, 05 Feb 2010 16:50:31 +0000] rev 676
Fix utcdate filter
So far the utcdate filter only omitted the timezone,
now it actually converts to utc.
Thanks to Kevin Grover for pointing out the issue.
Christian Ebert <blacktrash@gmx.net> [Wed, 20 Jan 2010 10:14:36 +0100] rev 675
Merge with stable
Matt Mackall <mpm@selenic.com> [Tue, 19 Jan 2010 22:20:08 -0600] rev 674
Update license to GPLv2+
[ original upstream message ]
Christian Ebert <blacktrash@gmx.net> [Thu, 24 Dec 2009 11:02:48 +0000] rev 673
(0.9.2compat) load extensions with "ext ="