Siddharth Agarwal <sid0@fb.com> [Tue, 04 Dec 2012 10:29:18 -0800] rev 18032
dirstate: test normalize is truthy instead of using a no-op lambda
hg perfstatus -u on a working directory with 170,000 files, without this
change:
! wall 1.869404 comb 1.850000 user 1.170000 sys 0.680000 (best of 6)
With this change:
! wall 1.839561 comb 1.830000 user 1.120000 sys 0.710000 (best of 6)
Kevin Bullock <kbullock@ringworld.org> [Tue, 04 Dec 2012 11:19:32 -0600] rev 18031
merge with stable
Matt Mackall <mpm@selenic.com> [Thu, 06 Dec 2012 13:21:27 -0600] rev 18030
hgweb: avoid generator exhaustion with branches
Matt Mackall <mpm@selenic.com> [Wed, 05 Dec 2012 15:38:18 -0600] rev 18029
hgweb: fix iterator reuse in atom feed generation
Julien Cristau <julien.cristau@logilab.fr> [Tue, 04 Dec 2012 14:35:02 +0100] rev 18028
tests: don't hardcode errno==2 for ENOENT
Hurd seems to set ENOENT to 2 + 2**30, unlike everyone else.
Bryan O'Sullivan <bryano@fb.com> [Mon, 03 Dec 2012 13:17:01 -0800] rev 18027
osutil: tab damage, how i hate thee
Bryan O'Sullivan <bryano@fb.com> [Mon, 03 Dec 2012 12:40:24 -0800] rev 18026
osutil: write a C implementation of statfiles for unix
This makes a big difference to performance.
In a clean working directory containing 170,000 files, performance of
"hg --time diff" improves from 2.38 seconds to 1.69.