More issues

[Christophe de Dinechin]

Jason Merrill wrote:
> 
> >>>>> Christophe de Dinechin <ddd at cup.hp.com> writes:
> 
>  > We need it if we have an entry point calling operator new.
> 
> But we decided not to have such an entry point.
> 

Did we? Well, you seem right: C-6 is indicated as closed in October. I certainly
did not notice :-(

Is there any conclusive argument against having a COMDAT-generated allocation
entry point whenever you have a new expression?  I guess that you can do that
without impacting the ABI... The writeup on C-6 is definitely very short.


Christophe