================
@@ -11451,7 +11451,11 @@ bool Sema::CheckDeductionGuideDeclarator(Declarator
&D, QualType &R,
bool MightInstantiateToSpecialization = false;
if (auto RetTST =
TSI->getTypeLoc().getAsAdjusted<TemplateSpecializationTypeLoc>()) {
- TemplateName SpecifiedName = RetTST.getTypePtr()->getTemplateName();
+ const TemplateSpecializationType *TST = RetTST.getTypePtr();
+ while (TST && TST->isTypeAlias())
+ TST = TST->getAliasedType()->getAs<TemplateSpecializationType>();
+
+ TemplateName SpecifiedName = TST->getTemplateName();
----------------
antangelo wrote:
If there is some test case that would trip the assert (i.e. some input for
which the alias type is not a template specialization type), then we should
handle it more gracefully. If it (conceivably) can't happen, then an assert is
fine.
I thought about it for a bit and came up with this (somewhat contrived) invalid
code that might fall into this path:
```c++
template<class T> struct A {};
struct B {};
template<typename T>
using X = B;
template<typename T>
A() -> X<T>;
```
https://godbolt.org/z/shzK9eccz
https://github.com/llvm/llvm-project/pull/117450
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
