named warnings & individual warning control

[Ian Lance Taylor]

DJ Delorie <dj@redhat.com> writes:

> > And the strings would be fixed for a given instantiation of the
> > compiler,
> 
> Assuming you're able to scan all the possible target backends and
> language frontends, yes.  Relying on that automatically precludes
> third parties adding their own targets or languages without figuring
> out how to reliably regenerate the hash.

I suppose one could use different hash tables for each frontend, for
the backend, and for the target files.

> I suspect a more general purpose hash would be easier to manage.

Well, probably.

Ian