Xazax-hun wrote: > To fix it, PR moves part of the modeling related to acquiring and releasing > to evalCall.
`evalCall` is a really big cannon and we should use it sparingly. The problem is, evalCall needs to modify all aspects of the function. Moreover, there is only one check that could do evalCall for a single function. I'd love to better understand what is the root cause for these weird diagnostics when the functions are inlined just in case there is a less intrusive solution. If there is nothing else we could do, I am not opposed to evalCall, but I think it should be a last resort. https://github.com/llvm/llvm-project/pull/111588 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
