install_static_unittests Chromium/InstallStaticUtilTest.GetSandboxSidPrefix failing on CrWinAsan |
||||
Issue descriptionStarted here: https://ci.chromium.org/buildbot/chromium.clang/CrWinAsan/1071 [ RUN ] Chromium/InstallStaticUtilTest.GetSandboxSidPrefix/1 ../../chrome/install_static/install_util_unittest.cc(665): error: Expected equality of these values: GetSandboxSidPrefix() Which is: 00007FF67BEDAD60 kSandBoxSids[std::get<0>(GetParam())] Which is: 00007FF67BE85660 Stack trace: Backtrace: testing::internal::UnitTestImpl::CurrentOsStackTraceExceptTop [0x00007FF67BA97FB4+214] (C:\b\c\b\CrWinAsan\src\third_party\googletest\src\googletest\src\gtest.cc:810) testing::internal::AssertHelper::operator= [0x00007FF67BA96F35+265] (C:\b\c\b\CrWinAsan\src\third_party\googletest\src\googletest\src\gtest.cc:386) install_static::InstallStaticUtilTest_GetSandboxSidPrefix_Test::TestBody [0x00007FF67BA3B779+559] (C:\b\c\b\CrWinAsan\src\chrome\install_static\install_util_unittest.cc:665) [ FAILED ] Chromium/InstallStaticUtilTest.GetSandboxSidPrefix/1, where GetParam() = (0, 00007FF67BE868A0 pointing to "system") (2 ms) https://chromium-review.googlesource.com/1157509 looks like it might possibly be vaguely related, but seems somewhat unlikely. Not sure what's up. Also happens on the new main waterfall win-asan bot (if I enable tests; they aren't yet): https://ci.chromium.org/p/chromium/builders/luci.chromium.try/win-asan/9
,
Aug 17
Hey, this code is less than one month old: https://chromium-review.googlesource.com/c/chromium/src/+/1094576 Maybe it's flaky or doesn't work on asan or something. wahahab, could you take a look?
,
Aug 18
I see what's wrong, I'll make a CL.
,
Aug 18
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/e57c623d6abcd190829747d789636097b8c4d7cb commit e57c623d6abcd190829747d789636097b8c4d7cb Author: Nico Weber <thakis@chromium.org> Date: Sat Aug 18 19:58:05 2018 win/asan: Fix GetSandboxSidPrefix test and run install_static_unittests on bot. The test was comparing two wchar_ts for pointer equality. In non-asan builds, equal constant strings in different TUs got merged, but in asan builds this seems ot not happen. Do a string comparison instead. Bug: 875225 Change-Id: I6a4031d1d6bca5697fb0a2dd6224323667711a81 Reviewed-on: https://chromium-review.googlesource.com/1180552 Reviewed-by: Reid Kleckner <rnk@chromium.org> Commit-Queue: Nico Weber <thakis@chromium.org> Cr-Commit-Position: refs/heads/master@{#584320} [modify] https://crrev.com/e57c623d6abcd190829747d789636097b8c4d7cb/chrome/install_static/install_util_unittest.cc [modify] https://crrev.com/e57c623d6abcd190829747d789636097b8c4d7cb/testing/buildbot/chromium.memory.json [modify] https://crrev.com/e57c623d6abcd190829747d789636097b8c4d7cb/testing/buildbot/test_suite_exceptions.pyl
,
Aug 18
https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/win-asan/1013 From the review: rnk says that after r334417 ASan should deduplicate identical string literals. I checked and that got reverted in r335684 and as far as I can tell didn't reland yet. The code is more correct after the fix in any case. It'd be nice if asan builds behaved like other builds though, so hopefully we can reland r334417 eventually.
,
Aug 19
My apologies for seeing this late. Thanks for helping me fix this! |
||||
►
Sign in to add a comment |
||||
Comment 1 by thakis@chromium.org
, Aug 17