https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83935
--- Comment #4 from Pierre-Marie de Rodat <derodat at adacore dot com> --- (In reply to Tom Tromey from comment #3) > TBH this did not make sense to me, either, which is partly why I originally > wrote my patch the "more natural" way -- then this got caught in review, > see https://reviews.llvm.org/D42082 Thanks for the pointer. So first, thanks for your efforts in adding support for these tags in GDB! During the review, Paul said: > Adding a TAG_variant_part that had no children could also work, although it > seems a little odd and might trip up an unwary non-Rust-aware debugger. Well, for now, we do generate child-less variant parts for Ada types such as: type Integer_Option (B : Boolean) is record case B is when False => null; when True => Value : Integer; end record; So I hope non-Rust-aware debuggers can cope with this. ;-) > It's worth a shot, though I think it was tried before, see > http://lists.dwarfstd.org/pipermail/dwarf-discuss-dwarfstd.org/2006-August/ > 001710.html Ah, I see. Looks like I have the same reasoning that Ron and Todd with respect to variant parts nesting had 12 years ago! :-D Everybody seems to agree that having a top-level member to serve as a variant part’s discriminant is a good thing to have, and it’s even suggested to post a proposal to reword the troublesome paragraph. So I suggest we indeed post a submission on dwarf-discuss@ and leave GCC’s DWARF back-end as it is today. What do you think? It’s a bit late for me, so I’ll give it a try tomorrow.
