[clang] [llvm] [DebugInfo] Place local ODR-uniqued types in decl DISubprograms (PR #119001)

[Jeremy Morse via llvm-commits]

jmorse wrote:

> Ah, fair enough, because the subprogram will contian a list of these local types, then the local type's scope will be a (potentially series of) lexical scope - so the lexical scopes are preserved?

That's certainly the outcome that the original patch series is aiming for -- I'm not completely certain that we can still achieve it. Part of the patch series clones types contained within distinct DISubprograms `retainedNodes` when they're inlined, which suggests that there's a connection between the types and the duplicated/distinct scopes that must be maintained, something that this patch would damage.

 I suppose I should hold off landing this patch until I fully understand that mechanism, and check we can still reach the desired outcome.

> > Given that the ODR-uniquing issue was only found experimentally in-the-field, I think we have to accept the incremental approach and discover problems by experience.
>
> Not sure I follow this.

I mean: at some point we just have to commit code and see what breaks, to discover obscure requirements.


> > Stepping back slightly: I think the ODR-uniquing of these types is directly opposed to scope-precision: reducing the number of type definitions down to one necessarily means discarding the extra locality information about where it's scoped in different function instances.
>
> Not sure I follow this either - I think that if each concrete lexical scope references the abstract ones, we could emit correct abstract origins to get the local static variables and types to be scoped appropriately in the resulting DWARF/for the consumer.

Indeed, if that connection is present then we should be fine. I'm worried that some of the extra plumbing (i.e. the cloning-of-types-during-inlining) indicates that this property requires maintenance during optimisation, and that that maintenance could be impossible in the presence of the ODR-uniquing feature.

https://github.com/llvm/llvm-project/pull/119001