[clang] [Sema] Fix computations of "unexpanded packs" in substituted lambdas (PR #99882)

[via cfe-commits]

cor3ntin wrote:

I think we should first rebase on top of the default argument PR, remove the attribute change (definitively the wrong way to do that), and reassess where we go from there.

My main concern is this line which seems very artificial  https://github.com/llvm/llvm-project/pull/99882/files#diff-10901a3bb08d903a57820f09c2eb5f40cb4cd47c54ca3cad472106814c4bf15dR359 

If we both agree that you would rather centralize everything during transform, I think it's fine but during parsing we already have the information available so we should try to use it - unless I am missing something?


https://github.com/llvm/llvm-project/pull/99882