[cfe-commits] r119583 - in /cfe/trunk: include/clang/Basic/DiagnosticGroups.td include/clang/Basic/DiagnosticSemaKinds.td include/clang/Basic/LangOptions.h include/clang/Driver/CC1Options.td include/clang/Driver/Options.td include/clang/Sema/Sema.h lib/Driver/Tools.cpp lib/Frontend/CompilerInvocation.cpp lib/Sema/SemaDecl.cpp test/SemaCXX/warn-argument-larger-than.cpp

[Chris Lattner]

On Nov 17, 2010, at 5:54 PM, Argyrios Kyrtzidis wrote:

> On Nov 17, 2010, at 3:51 PM, Argyrios Kyrtzidis wrote:
> 
>> 
>>> 
>>>> +  if (ReturnTy->isPODType() &&
>>>> +      Diags.getDiagnosticLevel(diag::warn_return_value_size) !=
>>>> +          Diagnostic::Ignored) {
>>> 
>>> Is this worth doing a 'is disabled' check for it?  It doesn't seem that expensive.  Does it cause a lot of PCH deserialization or something?
>> 
>> This is to allow disabling the warning through #pragma diagnostic.
> 
> Oops, disregard the above comment, I don't know what I was thinking.
> 
> ASTContext::getTypeInfo will calculate a ASTRecordLayout if one is not available.
> It's debatable whether we will avoid it but better be on the safe side ?

When does this warning trigger?  IMO, it would make sense to only do it on function definitions, not prototypes (because if you just have a proto, you probably don't have the option to change it, and that way you don't get the warning in every translation unit).

If it just triggers on definitions, the layout will have to be computed at codegen time, so it shouldn't add any cost to -O0 compiles.

-Chris