rsmith added inline comments.

================
Comment at: clang/lib/Sema/SemaTemplate.cpp:2174
+      Args.addOuterTemplateArguments(SubstArgs);
+      Args.addOuterRetainedLevel();
+      NamedDecl *NewParam = transformTemplateParameter(Param, Args);
----------------
aeubanks wrote:
> rsmith wrote:
> > This outer retained level would correspond to the parameters you're 
> > building -- we shouldn't add this. Instead, we should have two different 
> > `SubstArgs` lists, one for the outer parameters and one for the inner 
> > parameters; we should add just the outer parameters here, and add the inner 
> > and outer parameters below. (And at the end of each iteration of this loop 
> > we should accumulate arguments onto the outer arguments list like we 
> > accumulate arguments onto the inner arguments list in the next loop.)
> I think this is what you're referring to? I still don't understand exactly 
> what `Args.addOuterRetainedLevel()` does, and the current draft doesn't fix 
> the modules crash.
I found a few more places that need to be adjusted. Here's where I got to: 
https://github.com/zygoloid/llvm-project/commit/fffb4b688eb5a91bc2aca673773bc2da018f3dba
 (this is on top of the earlier version of this patch, though).

There are still some test failures there, so I think some more updates are 
still needed. We also need to do the same cloning process in 
`buildSimpleDeductionGuide`, which I think is why the modules unit test is 
still failing.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D116983/new/

https://reviews.llvm.org/D116983

_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
  • [PATCH] D116983: ... Arthur Eubanks via Phabricator via cfe-commits
    • [PATCH] D116... Arthur Eubanks via Phabricator via cfe-commits
    • [PATCH] D116... Arthur Eubanks via Phabricator via cfe-commits
    • [PATCH] D116... Richard Smith - zygoloid via Phabricator via cfe-commits

Reply via email to