[Lldb-commits] [lldb] [lldb][Fortran] Added DWARFASTParser for Fortran (PR #218024)

[Michael Buch via lldb-commits]

https://github.com/Michael137 commented:

Might be good to get @Teemperor/@adrian-prantl to chime in given the proposal in https://discourse.llvm.org/t/rfc-a-faster-more-reliable-type-system-for-c-languages/91459

We will eventually introduce a layer between DWARFASTParser and the AST representation.

With that in mind, question i'd have is: is the proposed intermediate layer a per-language representation or per-debug-info? I.e., should we use a single DWARF->IR parser and have language plugins implement just the IR->AST translation. Or, is each language plugin responsible for its own intermediate.

And what if a language plugin refuses to implement the intermediate IR step? Would that prevent/restrict the RFC in any way?

https://github.com/llvm/llvm-project/pull/218024