delesley added a comment. I agree with Aaron here. Those thread safety errors are supposed to fire; simply disabling the unit tests because they no longer fire is not acceptable. I also don't understand how this could be a bug with the thread safety analysis, since these particular errors are issued when the attributes are parsed, not when the analysis is run. I haven't looked at the relevant code in a long time; have there been significant changes to the attribute-parsing mechanism? There's no member initialization going on here, so why would changes to member initialization have this side effect? I'm confused.
DeLesley https://reviews.llvm.org/D29685 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
