[PATCH] D82188: [clang-tidy] Reworked enum options handling(again)

[Nico Weber via Phabricator via cfe-commits]

thakis added a comment.

In D82188#2119393 <https://reviews.llvm.org/D82188#2119393>, @njames93 wrote:

> @thakis Do those bots use `gn`, could that be the cause of the failures?


It's possible, but unlikely, given that your change doesn't touch any build files.

…and indeed, I checked out `8f73c4432b5fa8510c99a5053c07dc70a610e1fb^` (the commit before the revert) and ran `ninja -j200 check-clang-tools` in a cmake build and it repros there too:

  -- Build files have been written to: /Users/thakis/src/llvm-build-project
  [2185/2306] Linking CXX static library lib/libclangDriver.a
  /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/libtool: warning same member name (VE.cpp.o) in output file used for input files: tools/clang/lib/Driver/CMakeFiles/obj.clangDriver.dir/ToolChains/VE.cpp.o and: tools/clang/lib/Driver/CMakeFiles/obj.clangDriver.dir/ToolChains/Arch/VE.cpp.o (due to use of basename, truncation, blank padding or duplicate input files)
  [2305/2306] Running the Clang extra tools' regression tests
  FAIL: Extra Tools Unit Tests :: clang-tidy/./ClangTidyTests/ClangTidyDiagnosticConsumer.SortsErrors (799 of 924)
  ******************** TEST 'Extra Tools Unit Tests :: clang-tidy/./ClangTidyTests/ClangTidyDiagnosticConsumer.SortsErrors' FAILED ********************
  Note: Google Test filter = ClangTidyDiagnosticConsumer.SortsErrors
  [==========] Running 1 test from 1 test case.
  [----------] Global test environment set-up.
  [----------] 1 test from ClangTidyDiagnosticConsumer
  [ RUN      ] ClangTidyDiagnosticConsumer.SortsErrors
  /Users/thakis/src/llvm-project/clang-tools-extra/unittests/clang-tidy/ClangTidyDiagnosticConsumerTest.cpp:27: Failure
        Expected: 2ul
        Which is: 2
  To be equal to: Errors.size()
        Which is: 0
  0  ClangTidyTests           0x0000000106468c65 llvm::sys::PrintStackTrace(llvm::raw_ostream&) + 37
  1  ClangTidyTests           0x0000000106467b36 llvm::sys::RunSignalHandlers() + 198
  2  ClangTidyTests           0x0000000106469246 SignalHandler(int) + 262
  3  libsystem_platform.dylib 0x00007fff6b5055fd _sigtramp + 29
  4  libsystem_malloc.dylib   0x00007fff6b4c5a3d tiny_malloc_from_free_list + 555
  5  ClangTidyTests           0x000000010647186f testing::Test::Run() + 527
  6  ClangTidyTests           0x00000001064726e6 testing::TestInfo::Run() + 566
  7  ClangTidyTests           0x0000000106472f17 testing::TestCase::Run() + 247
  8  ClangTidyTests           0x000000010647b737 testing::internal::UnitTestImpl::RunAllTests() + 1287
  9  ClangTidyTests           0x000000010647b1bb testing::UnitTest::Run() + 171
  10 ClangTidyTests           0x000000010646a4c9 main + 121
  11 libdyld.dylib            0x00007fff6b30ccc9 start + 1
  
  ********************
  ********************
  Failed Tests (1):
    Extra Tools Unit Tests :: clang-tidy/./ClangTidyTests/ClangTidyDiagnosticConsumer.SortsErrors
  
  
  Testing Time: 21.55s
    Unsupported      :   3
    Passed           : 918
    Expectedly Failed:   2
    Failed           :   1

I then did the same again with the change reverted, and everything passes there.

> From what I can see `gn` isn't fully supported by llvm and could certainly be the cause of failing builds.
>  Do we have any build bots on mac that don't use `gn` but passed the test case?

It's correct that the `gn` build isn't supported (not just not fully: not at all) and if this would break things only there, then that wouldn't be a reason for revert. In practice, things very rarely are broken in the GN build only, and the GN bots cycle an order of magnitude faster, so they tend to show problems before the cmake bots do.

In this case, trunk was broken for > 12h, so I'd expect there's lots of evidence of this being broken on cmake bots too. However, the mac bots aren't on buildbot but on http://green.lab.llvm.org/green/ for whatever reason, and I find that page hard to read. http://green.lab.llvm.org/green/job/clang-stage1-RA/ has a bunch of red on the LHS … ah yeah http://green.lab.llvm.org/green/job/clang-stage1-RA/11966/console has "  Extra Tools Unit Tests :: clang-tidy/./ClangTidyTests/ClangTidyDiagnosticConsumer.SortsErrors" at the bottom.

So this showed up on the regular mac bots, yes.

---

The tree's been broken for over half a day to this. I'd suggest taking breakage supports a bit more serious going forward, instead of replying with "it can't possibly be this patch" and "it must be bot setup weirdness" going forward :)


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D82188/new/

https://reviews.llvm.org/D82188