uecker wrote: > It seems to me that we could simplify the implementation by being a bit more > literal in the interpretation of the requirements: look only at the type, > alignment, name, and bitfield-width of members. Just don't bother checking > other stuff > > I'd hope it can look more like a simple loop over the fields of the two > record types (similar to mergeFunctionTypes's checks of the functions' param > types), rather than starting with the very-complex > ASTStructuralEquivalence.cpp, and adding even more complexity to validate > attributes.
What I do in GCC is mostly look at those things required in the standard and I make sure that all field offsets are the same (e.g. if some attribute would change how fields are layed out, this would catch it). This is a simple loop. GCC already had a generic mechanism to compare attributes on types which decides whether an attribute makes a type incompatible (.e.g. affects ABI such as "packed"), or requires a warning, or whether it is just fine. https://github.com/llvm/llvm-project/pull/132939 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
