mstorsjo added a comment. FWIW, I'm also interested in this. I wasn't aware of the other ways of achieving this used by chromium, but although neat, I'd prefer something that doesn't require that kind of setup.
As for performance - although I haven't studied how the code works - wouldn't it be possible to try the normal straightforward case sensitive open first? Then if that fails, we'd fall back on trying doing a case insensitive search (in either direction). That way, the performance hit only affects compilations that would otherwise fail, so it shouldn't be prohibitive for anything else than the specific headers that are problematic. Or perhaps that's already been taken into consideration? Repository: rC Clang https://reviews.llvm.org/D48626 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
