bruno added a comment. > It might make more sense to have the module loaded from the AST file shadow > the module from the module map, especially for an explicit module build, now > that we have that functionality.)
+1, seems a much more consistent model. ================ Comment at: lib/Lex/ModuleMap.cpp:2574-2575 + llvm::SaveAndRestore<bool> OldExplicit(CurrentModuleMapIsExplicitlyProvided); + CurrentModuleMapIsExplicitlyProvided |= IsExplicitlyProvided; + ---------------- rsmith wrote: > It would seem cleaner to make this a member of `ModuleMapParser` (and > explicitly pass down the flag when parsing an `extern module` declaration). > Is there a reason to use (essentially) global state for this? I don't believe there's any reason for using a global state here (and Ben doesn't recall any specific reason either). I changed the patch to pass down the flag and it works fine. https://reviews.llvm.org/D31269 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
