This is the mail archive of the
mailing list for the GCC project.
Re: RFC [ABI]: C++ PATCH for c++/44540 - avoid mangling attribute noreturn
Richard Guenther wrote:
>> I guess I think this ought to require an -fabi-version=N option. Your
>> change is clearly correct, but to the extent that this matters to
>> anyone, it is going to break existing working code.
> Wouldn't it be better to steal some other bits for this purpose and
> "fix" it on the middle-end side?
I think that at some point the mangling has to change. It's valid to
declare these functions with and without an attribute. So, we have to
I agree that an orthogonal improvement would be to stop abusing
TYPE_CONST and TYPE_VOLATILE for these attributes and record them on
TYPE_ATTRIBUTES. But, that would be much bigger than Jason's patch.
(650) 331-3385 x713