Richard, thanks for the review.
================
Comment at: include/clang/Basic/DiagnosticSemaKinds.td:3654-3657
@@ -3653,1 +3653,6 @@
" expected type %3">;
+def err_mismatched_exception_spec_explicit_instantiation : Error<
+ "exception specification in explicit instantiation does not match deduced
one">;
+def ext_mismatched_exception_spec_explicit_instantiation : ExtWarn<
+ "exception specification in explicit instantiation does not match deduced
one">,
+ InGroup<Microsoft>;
----------------
rsmith wrote:
> "deduced" is not correct here; "instantiated" would be better.
Ok, fixed
================
Comment at: lib/Sema/SemaTemplate.cpp:7637-7639
@@ +7636,5 @@
+ // other declarations of that function.
+ if (D.isFunctionDeclarator() &&
+ D.getFunctionTypeInfo().getExceptionSpecType() != EST_None) {
+ auto TSI = GetTypeForDeclarator(D, S);
+ unsigned DiagID =
----------------
rsmith wrote:
> This is duplicating work the function has already done. (See variables `T`
> and `R` declared earlier.) Instead of checking whether the declarator has an
> exception specification, you can instead check whether `R` is a
> `FunctionProtoType` with an exception specification.
Agree, missed these variables somehow.
http://reviews.llvm.org/D5822
_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
