labath requested changes to this revision. labath added a comment. This revision now requires changes to proceed.
I don't think this should be necessary. Based on my experiments, the fact whether clang-cl defines `char32_t` depends on the version of visual studio it is emulating (`-fms-compatibility-version`). For instance I can do this: $ cat /tmp/a.cc char32_t XX; $ bin/clang-cl /tmp/a.cc /GS- /GR- /c -fms-compatibility-version=18 /tmp/a.cc(1,1) : error: unknown type name 'char32_t' char32_t XX; ^ 1 error generated. $ bin/clang-cl /tmp/a.cc /GS- /GR- /c -fms-compatibility-version=19 $ bin/clang-cl /tmp/a.cc /GS- /GR- /c So, what I suspect is happening is that clang uses %VCINSTALLDIR% to detect the MSVC version, and then sets the compatibility based on that. If the detected version is high enough, `char32_t` will be automatically defined. The interesting part is that, for me, the default version seems to be high enough that even without any special setting `char32_t` "just works". However, I'm on linux, so the differences could be due to that. So I think this is good news. I hope all that's needed here is to get `build.py` to add the -fms-compatibility thingy and things should work. Until we actually need to vary this version, I think we can just unconditionally add that flag to all clang-cl invocations in build.py. Repository: rLLDB LLDB CHANGES SINCE LAST ACTION https://reviews.llvm.org/D57809/new/ https://reviews.llvm.org/D57809 _______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
