Matt Mackall <mpm@selenic.com> [Fri, 13 Apr 2012 22:55:46 -0500] rev 16424
revlog: drop some unneeded rev.node calls in revdiff
Bryan O'Sullivan <bryano@fb.com> [Thu, 12 Apr 2012 20:26:33 -0700] rev 16423
revlog: avoid an expensive string copy
This showed up in a statprof profile of "hg svn rebuildmeta", which
is read-intensive on the changelog. This two-line patch improved
the performance of that command by 10%.
Matt Mackall <mpm@selenic.com> [Fri, 13 Apr 2012 22:55:46 -0500] rev 16422
changegroupsubset: avoid setdefault in inner loop
Matt Mackall <mpm@selenic.com> [Fri, 13 Apr 2012 22:55:40 -0500] rev 16421
changegroupsubset: drop repeated len() calls
Matt Mackall <mpm@selenic.com> [Fri, 13 Apr 2012 22:27:29 -0500] rev 16420
changegroup: reduce progress overhead
Matt Mackall <mpm@selenic.com> [Fri, 13 Apr 2012 22:23:45 -0500] rev 16419
changegroupsubset: renest inner loop
The mdata list should generally be shorter than changedfiles
Matt Mackall <mpm@selenic.com> [Fri, 13 Apr 2012 21:35:48 -0500] rev 16418
revlog: increase readahead size
Matt Mackall <mpm@selenic.com> [Fri, 13 Apr 2012 15:32:49 -0500] rev 16417
revset: avoid demandimport bug
Apparently the "import x as xy" doesn't manage to update xy in the
current scope's dictionary after load, which causes nodemod.nullrev to do a huge amount of demandload magic in the inner loop.
Patrick Mezard <patrick@mezard.eu> [Fri, 13 Apr 2012 15:07:13 +0200] rev 16416
mq: replace hasattr() with util.safehasattr(), update check-code.py
Matt Mackall <mpm@selenic.com> [Thu, 12 Apr 2012 20:52:39 -0500] rev 16415
merge with stable