Note we compile on 10.12 while testing on 10.13. Maybe that can be a problem?

The following revision refers to this bug:
  https://chromium.googlesource.com/infra/infra/+/dffd7edc0ae20cf1fe40ac1e148581cae29d5cc8

commit dffd7edc0ae20cf1fe40ac1e148581cae29d5cc8
Author: Michael Achenbach <machenbach@chromium.org>
Date: Mon Mar 12 08:10:18 2018

V8: Temporarily remove mac asan bot from lkgr

TBR=sergiyb@chromium.org

Bug: 820416
Change-Id: Ieb581af2282f3357b5823853e8b0d350afda7dc7
Reviewed-on: https://chromium-review.googlesource.com/958841
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Commit-Queue: Michael Achenbach <machenbach@chromium.org>

[modify] https://crrev.com/dffd7edc0ae20cf1fe40ac1e148581cae29d5cc8/infra/services/lkgr_finder/config/v8_cfg.pyl

Logged on a Mac VM while it was hanging on the tests. The only running v8 exe seemed to be a unittest:

chrome-bot@vm1151-m4:(Mac 10.13.3):~$ ps -ax | grep unittests
18322 ??         0:00.00 /b/s/w/ir/out/Release/unittests --random-seed=-976563914 --invoke-weak-callbacks --omit-quit --nohard-abort --gtest_filter=RandomNumberGenerator.NextSampleInvalidParam --gtest_random_seed=-976563914 --gtest_print_time=0

I tried to send a sigterm to the main python process, but nothing happened.

The hanging test uses GTEST_HAS_DEATH_TEST which reminds me on the hanging test on mac in issue  https://crbug.com/v8/6927 

I'm not too familiar with these RNG tests (they're not related to ASLR which I did change). cc'ing majeski who added the hanging test.

Another one that hung:
27990 ??         0:00.00 /b/s/w/ir/out/Release/unittests --random-seed=-976563914 --invoke-weak-callbacks --omit-quit --nohard-abort --gtest_filter=LoggingDeathTest.V8_DcheckCanBeOverridden --gtest_random_seed=-976563914 --gtest_print_time=0

After killing it on command line, the test driver reported it as PASS :(

Another one:
65108 ??         0:00.00 /b/s/w/ir/out/Release/unittests --random-seed=-976563914 --invoke-weak-callbacks --omit-quit --nohard-abort --gtest_filter=RandomNumberGenerator.NextSampleSlowInvalidParam1 --gtest_random_seed=-976563914 --gtest_print_time=0

This seems to affect all tests using gtest's EXPECT_DEATH_IF_SUPPORTED. I'll work on skipping just those for now. Maybe we shouldn't run mac asan anyways on those low-level unittests. I'd expect to find mostly bugs in tests.

The following revision refers to this bug:
  https://chromium.googlesource.com/v8/v8.git/+/188cca4dacbb76025a824d29263b67951b56d7e2

commit 188cca4dacbb76025a824d29263b67951b56d7e2
Author: Michael Achenbach <machenbach@chromium.org>
Date: Mon Mar 19 14:13:59 2018

[test] Skip death-tests for mac64 asan

NOTRY=true

Bug: chromium:820416
Change-Id: If995357345c2916ed5a3c8d1ad94f783cf6d633a
Reviewed-on: https://chromium-review.googlesource.com/968249
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Michael Achenbach <machenbach@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52031}
[modify] https://crrev.com/188cca4dacbb76025a824d29263b67951b56d7e2/test/unittests/unittests.status

Another hanging death-test on mac:
https://chromium-swarm.appspot.com/task?id=3dab8ae26342c310&refresh=10&show_raw=1
32088 ??         0:00.00 /b/s/w/ir/out/Release/unittests --random-seed=-976563914 --invoke-weak-callbacks --omit-quit --nohard-abort --gtest_filter=LoggingTest.CompareWithDifferentSignedness --gtest_random_seed=-976563914 --gtest_print_time=0

Will skip some more.

The following revision refers to this bug:
  https://chromium.googlesource.com/v8/v8.git/+/cffe6247ad06d90d2fb69e3710358983eaed39d5

commit cffe6247ad06d90d2fb69e3710358983eaed39d5
Author: Michael Achenbach <machenbach@chromium.org>
Date: Thu May 24 16:02:51 2018

[test] Skip more hanging death tests on mac asan

TBR=sigurds@chromium.org
NOTRY=true

Bug: chromium:820416
Change-Id: Ibd6fec04a9ea610c83c9347b4971f5ffea1fa80d
Reviewed-on: https://chromium-review.googlesource.com/1071521
Commit-Queue: Michael Achenbach <machenbach@chromium.org>
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53344}
[modify] https://crrev.com/cffe6247ad06d90d2fb69e3710358983eaed39d5/test/unittests/unittests.status

Please note that a similar issue started manifesting on V8 Linux64 - verify csa, see:
https://ci.chromium.org/p/v8/builders/luci.v8.ci/V8%20Linux64%20-%20verify%20csa/9740

It seems to happen pretty often (at least 1/10 runs) and Michael Achenbach was able to pinpoint mjsunit/big-object-literal as one of the faulty tests manually.

I filed a separate issue for comment 11: https://crbug.com/v8/8292