mstorsjo wrote: > Also I didn't consider non-ASCII at all. I don't have any strong opinions on > the discussion in #161400 but I do note that this keeps coming up
+1 - I didn't join in that discussion, but as long as exact matches still work as they should (for e.g. unicode directory names further up in the path), I don't exactly see what it would break - as long as we only try case insensitive matches when we fail to find things. (Practically, for the Windows SDK, we only need to do case insensitive checks for the filename in the exact path as such - we don't need to try the higher level directory names insensitively. But doing that might be good anyway though; I haven't checked what this implementation does.) https://github.com/llvm/llvm-project/pull/178585 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
