sammccall added a comment. In D141478#4043134 <https://reviews.llvm.org/D141478#4043134>, @hokein wrote:
> Thanks for tracking it down. The solution looks good to me. Since this is a > fragile and subtle issue, is it possible to have a unittest for it? If it is > not too hard, it would be nice to have a test. I think it's hard to write a test that's worth anything :-( It can't be a test that poking FileManager doesn't affect spelling, because I didn't fix that. I can test specifically that getExports() doesn't affect the getName() of relevant fileentries, but if we regress this symptom that's very unlikely to be the mechanism. An end-to-end test for the symptom needs to be set up in a way that tickles the bad side-effect at exactly the wrong time, and it's hard to predict how to do that in general. I'm not happy with this patch, but I'm not sure a bad test would help with that :-\ Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D141478/new/ https://reviews.llvm.org/D141478 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
