emilio added inline comments.
================ Comment at: clang/test/Index/print-type.cpp:118 // CHECK: TemplateRef=Foo:4:8 [type=] [typekind=Invalid] [isPOD=0] -// CHECK: FieldDecl=qux:29:38 (Definition) [type=Qux<int, char *, Foo<int>, outer::inner::Bar::FooType>] [typekind=Unexposed] [templateargs/4= [type=int] [typekind=Int] [type=char *] [typekind=Pointer] [type=Foo<int>] [typekind=Unexposed] [type=outer::inner::Bar::FooType] [typekind=Typedef]] [canonicaltype=outer::Qux<int, char *, outer::Foo<int>, int>] [canonicaltypekind=Record] [canonicaltemplateargs/4= [type=int] [typekind=Int] [type=char *] [typekind=Pointer] [type=outer::Foo<int>] [typekind=Record] [type=int] [typekind=Int]] [isPOD=1] // CHECK: TemplateRef=Qux:12:8 [type=] [typekind=Invalid] [isPOD=0] ---------------- arphaman wrote: > Hmm, that's an interesting change. I assume the test passed prior to this > patch, right? Did this output change purely because of the change in > `GetTemplateArguments`? Yup, that's right. This makes the type and the canonical type match in this case. In this case we lose a tiny bit of information (mainly, that the template argument was used through a typedef), but the alternative is to lose information about default template parameters, which seems worse to me (and is a regression, while template argument packs haven't been inspectionable until libclang 4.0). https://reviews.llvm.org/D32348 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
