Peter Ingebretson <pingebre@yahoo.com> [Wed, 23 Sep 2009 15:36:51 +0200] rev 9474
convert: fix marshalling in P4 convert extension to use a binary stream
The Perforce convert extension uses the p4 -G command line option and
marshal.load to obtain information about the history of a Perforce depot.
The method marshal.load must be provided with a binary stream (mode='rb') in
order to function reliably. Certain changelist descriptions or other elements
in a depot history could trigger a premature EOFError, causing the conversion
process to fail.
This may resolve the issue discussed in the following thread:
http://www.selenic.com/pipermail/mercurial/2009-July/026813.html
Dirkjan Ochtman <dirkjan@ochtman.nl> [Wed, 23 Sep 2009 15:19:26 +0200] rev 9473
merge with mpm
Patrick Mezard <pmezard@gmail.com> [Tue, 22 Sep 2009 13:30:32 +0200] rev 9472
convert/cvs: update debugcvsps documentation
Martin Geisler <mg@lazybytes.net> [Sat, 19 Sep 2009 13:55:18 +0200] rev 9471
Merge with crew-stable
Martin Geisler <mg@lazybytes.net> [Sat, 19 Sep 2009 13:52:54 +0200] rev 9470
dispatch: catch ConfigError while constructing ui
Nicolas Dumazet <nicdumz.commits@gmail.com> [Sat, 12 Sep 2009 11:17:07 +0200] rev 9469
config: abort on indented non-continuation lines (issue1829)
Previously, as soon as a continuation would be met, "cont" would stay
forever set to True, but "item" was set back to "None".
This caused the continuation code bits to run every time, until the next
"self.get(section, item) + '\n'" which would crash.
Stuart W Marks <smarks@smarks.org> [Thu, 17 Sep 2009 22:39:38 -0700] rev 9468
hg: say 'hg update' instead of 'hg up' in merge status message
Includes the change to the message itself, along with several test
output files that also had to change.
Martin Geisler <mg@lazybytes.net> [Sat, 19 Sep 2009 01:15:38 +0200] rev 9467
do not attempt to translate ui.debug output
Martin Geisler <mg@lazybytes.net> [Fri, 18 Sep 2009 17:37:51 +0200] rev 9466
Merge with crew-stable (again)
I had pulled 661bc51f09b7 into my local clone and accidentially
rebased it after pulling a new changeset from crew.
Shun-ichi GOTO <shunichi.goto@gmail.com> [Fri, 18 Sep 2009 21:25:47 +0900] rev 9465
win32mbcs: util.normpath should be wrapped.
It is because util.normpath() finally calls windows.pconvert()
which is not wrapped (by changeset:25c41ddb3978).