Closed jameysharp closed 6 months ago
No problem; done. Almost everything I know about DWARF today is stuff I learned this week, so you could easily convince me either way. I'm always happy to take advice.
Do you have plans for more gimli work? If not I'll do a release soon (along with object and addr2line).
I don't have anything more planned right now. I kind of wanted to improve the string formatting of the Error
type, both to include the data associated with the various Unknown
and Unsupported
variants, and also ideally to attach a ReaderOffsetId
to every error, not just UnexpectedEof
. But I don't think I actually have the energy for either of those projects at the moment.
I intend to dig into the support for DWARF that's in Wasmtime and Cranelift, so it's possible I'll find more things I'd like from gimli along the way. But for now, I think a new release would be great. Thanks for your help getting these changes merged so promptly!
Keeping this type distinct from the
SectionId
enum ensures that the compiler will report an error if someone adds incomplete support for new separate-debug sections.As discussed in: https://github.com/gimli-rs/gimli/pull/714#discussion_r1611164319
The existing
SectionId::dwo_name
method did not match the list of sections actually supported elsewhere in gimli. I've added the one section ID that was missing (.debug_macinfo.dwo
), but I have not removed the section IDs which are only present there:.debug_str.dwo
,.debug_cu_index
, and.debug_tu_index
.