Reland https://codereview.chromium.org/581983003/: Enable ASan default options on Mac.

Reland https://codereview.chromium.org/581983003/: Enable ASan default options on Mac.

This CL links libsanitizer_options into every executable built with ASan on OSX.
The existing implementation of __asan_default_options for Chromium.app is merged
with that in sanitizer_options.cc

Also now use_sanitizer_options is only set when building with sanitizers so that
there isn't an unconditional dependency on an empty object file in every
executable in non-sanitizer builds.

In addition to that, the dependency on sanitizers.gyp:sanitizer_options is removed
from ui/base/ui_base_tests.gyp:ui_base_tests_bundle which is not a real shared lib,
despite marked as such.

BUG=302040
TBR=earthdok@chromium.org,cpu@chromium.org
R=avi@chromium.org

Committed: https://crrev.com/7dbed1574ed37e1ec4d8417e6d4c4205fcfdf382
Cr-Commit-Position: refs/heads/master@{#296200}

[Alexander Potapenko, 2014-09-23 13:11:28]

Carlos, Sergey: TBR

Avi, can you please take a look at the ui/ bits?
This CL has been reviewed at https://codereview.chromium.org/581983003/, but
upon landing it broke ui_unittests which were unsuccessfully trying to load
"@executable_path/../Versions/1.0.0.0/ui_unittests
Framework.framework/ui_unittests Framework"
The trick is that "ui_unittests Framework" is not a real framework, despite it's
marked as a shared library. Adding a dependency on sanitizer_options to that
framework produces an ugly fake shared library with an empty text section, which
ends up being linked into ui_unittests (without ASan the ui_unittests executable
doesn't depend on that framework at all), so I've decided to remove that
dependency.

A side question that puzzles me is whether "ui_unittests Framework" must be
marked as a shared_library in GYP.

[Avi (use Gerrit), 2014-09-23 14:07:03]

I'm fumbling my way around the magical build requirements of bundling the ui
tests as you are. This looks reasonable to me, and the comment is quite
explanatory. LGTM

[Avi (use Gerrit), 2014-09-23 14:07:30]

On 2014/09/23 14:07:03, Avi wrote:
> I'm fumbling my way around the magical build requirements of bundling the ui
> tests as you are. This looks reasonable to me, and the comment is quite
> explanatory. LGTM

BTW, that was a review of ui/base. I didn't look at the rest.

[Alexander Potapenko, 2014-09-23 14:08:52]

On 2014/09/23 14:07:30, Avi wrote:
> On 2014/09/23 14:07:03, Avi wrote:
> > I'm fumbling my way around the magical build requirements of bundling the ui
> > tests as you are. This looks reasonable to me, and the comment is quite
> > explanatory. LGTM
> 
> BTW, that was a review of ui/base. I didn't look at the rest.

The rest has been copied verbatim from
https://codereview.chromium.org/581983003/, I've CCed the reviewers of that CL.
Thank you!

[earthdok, 2014-09-23 14:40:22]

rslgtm

[Alexander Potapenko, 2014-09-23 15:16:40]

The CQ bit was checked by glider@chromium.org

[commit-bot: I haz the power, 2014-09-23 15:17:17]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/594843002/1

[commit-bot: I haz the power, 2014-09-23 16:38:30]

Committed patchset #1 (id:1) as d00ec406565210eb847e9d7f641f48ee4d1166f8

[commit-bot: I haz the power, 2014-09-23 16:39:32]

Patchset 1 (id:??) landed as
https://crrev.com/7dbed1574ed37e1ec4d8417e6d4c4205fcfdf382
Cr-Commit-Position: refs/heads/master@{#296200}

[Nico, 2015-09-17 04:12:01]

thakis@chromium.org changed reviewers:
+ thakis@chromium.org

[Nico, 2015-09-17 04:12:02]

https://codereview.chromium.org/594843002/diff/1/ui/base/ui_base_tests.gyp
File ui/base/ui_base_tests.gyp (right):

https://codereview.chromium.org/594843002/diff/1/ui/base/ui_base_tests.gyp#ne...
ui/base/ui_base_tests.gyp:248: # ui_base_tests_bundle doesn't actually contain a
shared library and
…wouldn't it be a better fix to make this target not be a shared_library if it
isn't one? :-/