[Lldb-commits] [lldb] [lldb][SymbolFileDWARF] CompleteType: Lookup type in the declaration … (PR #120569)

[Michael Buch via lldb-commits]

Michael137 wrote:

The opens a couple cans of worms (though I think this fix still makes sense in isolation):
1. This one might be by design and hard to change at this point but do we want to silently ignore failures to resolve a type in `ResolveCompilerType`?: https://github.com/llvm/llvm-project/blob/eb812d28f542bf0de54c157a7391e446739570cc/lldb/source/Symbol/Type.cpp#L647
2. The order of the entries in the debug-map for the test-case matter! Which is worrying because that is probably not a deterministic property we can rely on? If `lib.cpp` and `main.cpp` were swapped we actually run into an entirely different assertion (which I'm looking into separately)! If anyone has ideas on how to make this more deterministic let me know. Perhaps a unit-test for `ParseTypeFromDWARF`/`CompleteTypeFromDWARF` is the best way forward for that?

https://github.com/llvm/llvm-project/pull/120569