cor3ntin wrote: > Regarding issues associated with > https://github.com/llvm/llvm-project/pull/95660, I realized they are actually > part of a bigger problem in constant evaluation. As discussed in > https://github.com/llvm/llvm-project/issues/59966, it might be beneficial to > implement on-demand template instantiation for constant evaluation. That way, > we don't have to touch the logic of deferral instantiation in > MarkFunctionReferenced(), (which probably can be removed then) and we could > instantiate those local lambdas before evaluation.
I haven't had time to look in the issue in details, but this might be the best solution. We will need to do that for reflection too @Endilll is planning to wire Sema in constant evaluation in the next few months, that will be a good opportunity to fix all of these issues https://github.com/llvm/llvm-project/pull/98758 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
