> On 06/17/2013 09:07 AM, Jan Hubicka wrote: > >>Yes. Also for template instantiations and inline functions > >>(basically, decls with TREE_COMDAT set). That isn't very > > > >Can't those be just merged based on assembler name? > > Yes, though they can have local classes that are also subject to the ODR.
I see, I guess my ODR predicate should just work for those. I will write some code looking for unmerged ODR cases on the top of Richard's patch and we can decide what to do. To get meaningful warnings, we need to know what decls/types are subject to ODR. Do you think you can make C++ FE to drop a flag so middle-end know? We can LTO ODR and non-ODR languages together. Thanks, Honza > > >I think this is something Richard can handle (semi-easilly) with his > >new tree merging patch. > > Great. > > >Shall I understand this as approval with those fixes and testing? > > Yes. > > Jason
