Superseded by D68172.

@aaronpuchert thanks, that looks like a better fix than mine for PR43430, although I'm not sure D68172 obsoletes this patch.
@Szelethus could you confirm that thus examples shouldn't be built when CLANG_BUILD_EXAMPLES is OFF, and that this patch is still valid?

Here you go: https://bugs.llvm.org/show_bug.cgi?id=43430

I don't have commit rights, I'd be glad if someone could commit this if the review pass, thanks.

Forgot to apply clang-format-diff.

Add tests, added a note that LLVMDWARFMacinfoRecordType values come from the DWARF spec.

@Szelethus take your time, I'm surprised to be the first with the problem (couldn't find reference on google or llvm bug tracker), I wouldn't be surprised if I fucked up somewhere and that this patch isn't really necessary.

Update to add -U999999, although dunno why I don't get more context on clang/lib/Analysis/plugins/CMakeLists.txt.

Mybad, llvm::dwarf::MacinfoRecordType (in llvm/BinaryFormat/Dwarf.h) isn't generated using Dwarf.def, so I just used the same values for another enum in llvm-c/DebugInfo.h.
The values in llvm::dwarf::MacinfoRecordType won't move as they are defined by the DWARFv4 standard (Figure 39. Macinfo Type Encodings, at page 194/325), so it is safe to static_cast it from the new LLVMDWARFMacinfoRecordType to MacinfoRecordType.

It seem to also miss macro related bindings, I'll probably make another differential today.