vsk added a comment.
In https://reviews.llvm.org/D25040#556209, @arphaman wrote:
> In https://reviews.llvm.org/D25040#556165, @vsk wrote:
>
> > In https://reviews.llvm.org/D25040#556109, @arphaman wrote:
> >
> > > Thanks for working on this, I have a couple of comments/questions:
> > >
> > > - Th
arphaman added a comment.
In https://reviews.llvm.org/D25040#556165, @vsk wrote:
> In https://reviews.llvm.org/D25040#556109, @arphaman wrote:
>
> > Thanks for working on this, I have a couple of comments/questions:
> >
> > - The current logic doesn't work correctly for a test case that's shown
vsk added a comment.
In https://reviews.llvm.org/D25040#556109, @arphaman wrote:
> Thanks for working on this, I have a couple of comments/questions:
>
> - The current logic doesn't work correctly for a test case that's shown below:
>
> ``` class C2 { void __attribute__((unused)) m4() { } void
arphaman added a comment.
Thanks for working on this, I have a couple of comments/questions:
- The current logic doesn't work correctly for a test case that's shown below:
class C2 {
void __attribute__((unused)) m4() { }
void __attribute__((unused)) m5();
};
The method 'm4' actually
vsk created this revision.
vsk added a reviewer: arphaman.
vsk added a subscriber: cfe-commits.
This lets us suppress coverage reporting for ~30 functions across llvm and
clang.
I passed the test case through llvm-cov and verified that 'unused' functions
are skipped, e.g:
```
3| |voi
