[PATCH] D147059: [asan][test] Fix tests or mark XFAIL for MinGW target
Martin Storsjö via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Wed Mar 29 00:35:18 PDT 2023
mstorsjo added inline comments.
================
Comment at: compiler-rt/lib/asan/tests/asan_test.cpp:208
+ // https://google.github.io/googletest/advanced.html#regular-expression-syntax
+ // GoogleTest's regular expression engine on Windows does not support `[]`
+ // brackets.
----------------
pengfei wrote:
> alvinhochun wrote:
> > mstorsjo wrote:
> > > If this is specific to Windows, why is this different from the case with MSVC/clang-cl?
> > MSVC target shouldn't reach here anyway because `if (sizeof(long double) == sizeof(double))` is true. I guess I can change it to `_WIN32` though.
> Clang has option to override it, see https://godbolt.org/z/1x9Pd9Tfx
Thanks - although we're not interested in making the MSVC target run this code - I was just wondering why this issue only appeared now.
@alvinhochun Changing it to _WIN32 probably is a good idea for clarity, and then mention the long double issue in the commit message regarding that particular change.
================
Comment at: compiler-rt/test/asan/TestCases/Windows/report_after_syminitialize.cpp:6
+
+// The first build command is intended for MSVC target, which fails on MinGW.
+// The second build command contains the flags required to get PDB debug info
----------------
alvinhochun wrote:
> mstorsjo wrote:
> > In which way does it fail - what does `%clangxx_asan` expand to in that case?
> It fails because it's missing `-ldbghelp`, which MSVC target doesn't need thanks to the `#pragma` lib comment.
Ok, I see. This is kinda tricky and somewhat brittle, but I guess it can be ok.
================
Comment at: compiler-rt/test/asan/TestCases/atoll_strict.c:14
// FIXME: Needs Windows interceptor.
-// XFAIL: target={{.*windows-msvc.*}}
+// XFAIL: target={{.*windows-(msvc.*|gnu)}}
----------------
alvinhochun wrote:
> mstorsjo wrote:
> > Or just generalize into `{{.*windows-.*}}?
> I was thinking of making it more explicit to point out that, yes this test is XFAIL for both msvc and gnu targets.
Ok, fair enough.
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D147059/new/
https://reviews.llvm.org/D147059
More information about the llvm-commits
mailing list