dblaikie added inline comments.
================
Comment at: clang/lib/AST/DeclPrinter.cpp:1101
Out << ", ";
- Args[I].print(Policy, Out);
+ if (TemplOverloaded || !Params)
+ Args[I].print(Policy, Out, /*IncludeType*/ true);
----------------
Looks like this (& the `TemplOverloaded` in the other function, below) is
undertested.
Hardcoding:
true in this function results in no failures
false in this function results in 1 failure in
`SemaTemplate/temp_arg_enum_printing.cpp`
- this failure, at least to me, seems to not have much to do with overloading
- at least syntactically foo<2> (where the parameter is an enum type) doesn't
compile, not due to any overloading ambiguity, but due to a lack of implicit
conversion from int to the enum type, I think? and the example doesn't look
like it involves any overloading...
true in the other overload results in no failures
false in the other overload results in no failures
I came across this because I was confused by how this feature works - when the
suffixes are used and when they are not to better understand the implications
this might have for debug info. (though I'm still generally leaning towards
just always putting the suffixes on for debug info)
@rsmith - could you give an example of what you meant by passing in a parameter
when the template is overloaded & that should use the suffix? My simple
examples, like this:
```
template<unsigned V>
void f1() { }
template<long L>
void f1() { }
int main() {
f1<3U>();
}
```
That's just ambiguous, apparently, and doesn't compile despite the type suffix
on the literal. If I put a parameter on one of the functions it doesn't seem to
trigger the "TemplOverloaded" case - both instantiations still get un-suffixed
names "f1<3>"...
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D77598/new/
https://reviews.llvm.org/D77598
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
