This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] PR c++/27574


> Dodji Seketeli wrote:
> >Jason Merrill a écrit :
> >>Seems like the right fix would be to include clones in the inlined_to 
> >>list of the abstract function.
> >
> >After talking with Honza on IRC, it appeared that it wouldn't work 
> >because nodes that are added to inline_to must be actually inlined 
> >otherwise, later sanity check that would break.
> 
> Can't the later sanity check check for inlining or cloning?

We definitly can remove the sanity check, but the inline clones are
designed for representing inline plan (i.e. producing virtual clones
that do not have actual tree representation), so using them for linking
functions that was originaly produced as clones by frontend but appear
as real functions in backend does not make that much sense.

If we want to have functions related together, we definitly can add new
lists for that, same way as we do for nested functions and such.

Honza
> 
> Jason


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]