Regression(r154816): clang produces chromium binaries that are 0.875 MB (1.4%) larger than clang 154815

[Quuxplusone]

Bugzilla Link	PR12645
Status	NEW
Importance	P enhancement
Reported by	Nico Weber (nicolasweber@gmx.de)
Reported on	2012-04-24 12:44:13 -0700
Last modified on	2012-04-24 20:17:08 -0700
Version	unspecified
Hardware	PC All
CC	chandlerc@gmail.com, llvm-bugs@lists.llvm.org, rafael@espindo.la
Fixed by commit(s)
Attachments
Blocks
Blocked by
See also

See http://build.chromium.org/f/chromium/perf/mac-release/sizes/report.html?history=150&rev=133696&graph=ChromiumFramework

[Quuxplusone]

Bisect steps (clang revision: chromium framework size in bytes):

155352: 63413872
155000: 63413872
154750: 62591628
154875: 63413872
154810: 62591628
154840: 63414924
154825: 63414924
154818: 63414924
154815: 62591628

154817 and 18 are documentation changes, so it's 154816 "Flip the new block-
placement pass to be on by default.".

Since that's a somewhat involved change, and it's "just" 870kB, I'd be willing
to live with this in chromium if the test of the clang community thinks it's
acceptable.

If 154816 causing size regressions isn't new information, feel free to WONTFIX.
Else, please cc relevant folks and discuss :-)