Michael137 wrote: > I am not sure that there is a reliable approach here so it may be better to > have two ways to get data type. The current solution works both for gcc and > clang on my laptop, but it might depend on the compiler version.
That's a fair point. But just for testability, I think it's best if we use the same method of retrieval here. If Clang decides to omit those typedefs at some point then we can add another branch. But don't feel strongly about it > I found an old [review](https://reviews.llvm.org/D14358) where mentioned that > omitting certain fields can be an intentional debug info optimization. That is true, but I've only seen this apply to types or functions. First time I'm seeing this for other kinds of tags like `DW_TAG_template_type_parameter` (@dwblaikie @adrian-prantl is this a size optimisation you are aware of?) > This [bug](https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66116) was already > submitted to GCC, but there has been no feedback. Good find. Interesting that this applies to `std::allocator` again. https://github.com/llvm/llvm-project/pull/164251 _______________________________________________ lldb-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
