This patch could've used some more (any) test coverage, btw. Oh, I see now that the test was added in r172648, but missed at least the case below
On Tue, Jan 15, 2013 at 5:22 PM, Eric Christopher <[email protected]> wrote: > Author: echristo > Date: Tue Jan 15 19:22:32 2013 > New Revision: 172591 > > URL: http://llvm.org/viewvc/llvm-project?rev=172591&view=rev > Log: > Collect both normal and static data members of a class in source > order. Describe static data members to metadata using new interfaces. > > Part of PR14471. > > Patch by Paul Robinson! [snip] > +/// CollectRecordStaticField - Helper for CollectRecordFields. > +void CGDebugInfo:: > +CollectRecordStaticField(const VarDecl *Var, > + SmallVectorImpl<llvm::Value *> &elements, > + llvm::DIType RecordTy) { > + // Create the descriptor for the static variable, with or without > + // constant initializers. > + llvm::DIFile VUnit = getOrCreateFile(Var->getLocation()); > + llvm::DIType VTy = getOrCreateType(Var->getType(), VUnit); > + > + // Do not describe enums as static members. > + if (VTy.getTag() == llvm::dwarf::DW_TAG_enumeration_type) > + return; I changed this to an assert in r188494/r188496, since no test case actually touched this codepath (I'm sort of surprised none of the buildbots fired - maybe we really don't have any static member variables of enum type in all of Clang and LLVM?), which resulted in PR16927, which I fixed in r188612. _______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
