tejohnson added a comment. In https://reviews.llvm.org/D31508#716087, @davide wrote:
> clang -cc1 crashes if an invalid reloc model is passed (via > `-mreloc-model=`), see https://bugs.llvm.org/show_bug.cgi?id=32490 > I'm not sure if the bug was already there or this refactoring exposed it. > Either way, I'm going to submit a patch to get this fixed. > Ideally, we shouldn't keep two codepath in sync (as we do already for > `getCodeModel()`, but that's the status quo). I may decide to refactor the > whole handling to avoid this sync issue, but that won't be today. Thanks for fixing. I don't think this refactoring would have affected this though, the non-ThinLTO backend changes were NFC, it is the same switch statement that was there before. Repository: rL LLVM https://reviews.llvm.org/D31508 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits