benlangmuir added a subscriber: JDevlieghere. benlangmuir added a comment. > I think it should be fine to allow dropping the > A.framework/Frameworks/B.framework directory from the reproducer VFS
I think technically this is wrong, since if you're missing the symlink, then A might not build -- e.g. it could be doing a relative include that needs the symlink. But am I understanding correctly that the reproducer was already broken in this case? If so I'm fine with this. The right thing to do would be to capture both the framework and the symlink. I'm not sure how practical that is with the current architecture. @JDevlieghere any thoughts? Longer term, our CAS work could ultimately end up solving this in a better way by fully capturing all the inputs. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D151855/new/ https://reviews.llvm.org/D151855 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
