r215365 - Coverage mapping: emit mapping for cxx constructors that use microsoft's ABI

[Alex L]

I haven't gotten a chance to resubmit this yet, will do soon though.


2014-08-22 10:15 GMT-07:00 Reid Kleckner <rnk at google.com>:

> Has this been resolved? I'm trying to grind through my unread commit mail.
>
>
> On Mon, Aug 11, 2014 at 11:19 AM, Justin Bogner <mail at justinbogner.com>
> wrote:
>
>> Alex Lorenz <arphaman at gmail.com> writes:
>> > Author: arphaman
>> > Date: Mon Aug 11 12:45:49 2014
>> > New Revision: 215365
>> >
>> > URL: http://llvm.org/viewvc/llvm-project?rev=215365&view=rev
>> > Log:
>> > Coverage mapping: emit mapping for cxx constructors that use
>> microsoft's ABI
>> >
>> > Modified:
>> >     cfe/trunk/lib/CodeGen/CodeGenPGO.cpp
>> >
>> > Modified: cfe/trunk/lib/CodeGen/CodeGenPGO.cpp
>> > URL:
>> >
>> http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/CodeGen/CodeGenPGO.cpp?rev=215365&r1=215364&r2=215365&view=diff==============================================================================
>> > --- cfe/trunk/lib/CodeGen/CodeGenPGO.cpp (original)
>> > +++ cfe/trunk/lib/CodeGen/CodeGenPGO.cpp Mon Aug 11 12:45:49 2014
>> > @@ -847,8 +847,13 @@ void CodeGenPGO::checkGlobalDecl(GlobalD
>> >    // a class. Every function is instrumented, but we only want to
>> provide
>> >    // coverage for one of them. Because of that we only emit the
>> coverage mapping
>> >    // for the base constructor/destructor.
>> > +  // For Microsoft's C++ ABI Clang emits only the complete constructor,
>> > +  // therefore we have to emit the coverage mapping for it instead of
>> the base
>> > +  // one.
>> > +  const CXXCtorType AcceptedCtor =
>> > +      CGM.getTarget().getCXXABI().isMicrosoft()? Ctor_Complete :
>> Ctor_Base;
>>
>> There should be spaces around the ? here.
>>
>> >    if ((isa<CXXConstructorDecl>(GD.getDecl()) &&
>> > -       GD.getCtorType() != Ctor_Base) ||
>> > +       GD.getCtorType() != AcceptedCtor) ||
>> >        (isa<CXXDestructorDecl>(GD.getDecl()) &&
>> >         GD.getDtorType() != Dtor_Base)) {
>> >      SkipCoverageMapping = true;
>>
>> You should really commit the tests this is fixing along with the
>> change. It makes life much easier for people digging through history
>> later if code changes and tests are in the same commit.
>> _______________________________________________
>> cfe-commits mailing list
>> cfe-commits at cs.uiuc.edu
>> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20140822/cc150743/attachment.html>