Do not compile pyautolib with -faddress-sanitizer on Mac.

Do not compile pyautolib with -faddress-sanitizer on Mac.
Building asan_stub.c with -faddress-sanitizer causes Clang to crash,
see http://code.google.com/p/address-sanitizer/issues/detail?id=67

BUG=address-sanitizer:67
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=133451

[Alexander Potapenko, 2012-04-20 14:21:40]

Nico,

what happens to test/ui/ui_test.cc and other source files that may be shared
between pyautolib and other targets? Is -faddress-sanitizer passed to clang to
build it?

[Alexander Potapenko, 2012-04-23 10:49:57]

Ping there

[Nico, 2012-04-23 15:17:23]

LGTM

On 2012/04/20 14:21:40, Alexander Potapenko wrote:
> Nico,
> 
> what happens to test/ui/ui_test.cc and other source files that may be shared
> between pyautolib and other targets? Is -faddress-sanitizer passed to clang to
> build it?

ui_test.cc will be built with -faddress-sanitizer when part of test_support_ui,
and without it when part of pyautolib. I hope no target depends on both
test_support_ui and pyautolib (that would be an ODR violation without this
change too, but without this change it might be harmless).

> Ping there

Sorry, I was on a conference last wed/thu

[Nico, 2012-04-23 15:17:28]

http://codereview.chromium.org/10134009/diff/1/chrome/chrome_tests.gypi
File chrome/chrome_tests.gypi (right):

http://codereview.chromium.org/10134009/diff/1/chrome/chrome_tests.gypi#newco...
chrome/chrome_tests.gypi:4521: },
Huh, I asked if we need this on mac here:
http://codereview.chromium.org/10037007/diff/1/chrome/chrome_tests.gypi :-/

[Alexander Potapenko, 2012-04-23 15:22:31]

> Huh, I asked if we need this on mac here:
> http://codereview.chromium.org/10037007/diff/1/chrome/chrome_tests.gypi :-/

We didn't build pyautolib for Mac before.

[Alexander Potapenko, 2012-04-23 15:27:34]

On 2012/04/23 15:17:23, Nico wrote:
> LGTM
Thank you!