zwuis wrote:

> > * the added test case is **invalid**, which doesn't represent the original 
> > reproducer, and
> 
> I think i verified this added case will crash before this patch, and the 
> original reproducer doesn't crash now too(https://godbolt.org/z/GW8YhxGbs), 
> any more details?

I see. Thanks for the clarification.

> > * the added comment is wrong.
> 
> You are saying this is a FP?

No. The root cause is `getAsCXXRecordDecl` returns `nullptr` for dependent 
types, regardless of whether it is a target base class. If we change the test 
to `class X: public C<T>`, it still crashed.

https://github.com/llvm/llvm-project/pull/192786
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

Reply via email to