Use exact equality for category language matching, for all languages, except those specifically mentioned.
Details
Diff Detail
- Repository
- rL LLVM
Event Timeline
Though... wouldn't "category_lang == valobj_lang" be a sane default behavior for any languages not mentioned specifically below?
On the surface, that looks correct to me. However I don't work on LLDB anymore and have been out of the loop enough that I don't feel comfortable approving this change. I believe Jim Ingham is Data Formatter Overlord these days.
It does make sense to me that any category language that has a name should by default to be an exact match for the valobj language. If we're going to choose that as a rule, then we should eliminate all the explicitly listed languages that just require an exact match. That will make the code clearer.
It looks like Go has been using the default "all valobj's match" rule. I don't know it that was intentional or not, if somebody who uses Go knows, they should chime in.
Rust and Go are not the only missing languages. Looks like all of the newly-added DWARF5 language codes are not covered here.
Yes, the only reason I mentioned Go is that there is actually an lldb Language plugin for Go, but not for any of the other languages, so there's a little more support there to potentially break.