llvm dev - May 2016 - [RFC] Fail early on unimplementable append linkage uses

On 13 May 2016 at 08:51, Renato Golin <renato.golin at linaro.org>
wrote:
> On 13 May 2016 at 13:43, Rafael Espíndola <llvm-dev at
lists.llvm.org> wrote:
>> That means that appending linkage can only be used with llvm special
>> symbols that never hit a real .o.
>
> So they should have been resolved by emission time, or it's a back-end
error.
Not resolved, translated. For example, llvm.ctors gets expanded to
whatever form the object file uses for processing static constructors.
> Do you have an idea of what could be affected if we start emitting an
> error, here? Test-Suite, other programs?
No. Clang only produces the well known ones. Anything producing a
appending linkage global that actually got written to the output would
hit a link error if it tried combining any two of those files (and can
get the same behaviour with an external linkage).

Cheers,
Rafale

On 13 May 2016 at 13:56, Rafael Espíndola <rafael.espindola at gmail.com>
wrote:
> No. Clang only produces the well known ones. Anything producing a
> appending linkage global that actually got written to the output would
> hit a link error if it tried combining any two of those files (and can
> get the same behaviour with an external linkage).
I wonder what was the motivation to add that piece of code, then. It
sounds to me that it was wrong to do that in the first place.

--renato

On 13 May 2016 at 10:11, Renato Golin <renato.golin at linaro.org>
wrote:
> On 13 May 2016 at 13:56, Rafael Espíndola <rafael.espindola at
gmail.com> wrote:
>> No. Clang only produces the well known ones. Anything producing a
>> appending linkage global that actually got written to the output would
>> hit a link error if it tried combining any two of those files (and can
>> get the same behaviour with an external linkage).
>
> I wonder what was the motivation to add that piece of code, then. It
> sounds to me that it was wrong to do that in the first place.
I agree. The revisions in question seem to be r41048 and before that
r8417. From reading them I think it was just a mistake about how
object files work.

Cheers,
Rafael