Re: Which HPPA targets do we still support?

[Andrew Cagney <ac131313 at redhat dot com>]

> Actually, I believe that no one is using any of these targets except for
> HP/UX.  hppa*-*-pro* was kept around because it is (was?) standalone;
> you could build an hppa-proelf cross debugger to make sure you didn't
> break compilation for the PA.
Even ``was'' is being generous here for the hppa*-pro*.  I don't think 
it built on anything other than HP/UX :-(

> I've seen bug reports for HP/UX, but never for any of the others; and
> we know the HPPA target has broken periodically, so that's a good hint
> that no one's tracking them.  Of course if it's not much trouble, we
> can keep them for now and deal with it later - but I'm not sure that we
> need to hang on to all of them.
>
> Looking at GCC's supported targets, I wouldn't be surprised if
> hppa-openbsd is in use, but that won't match the existing pattern
> anyway... ditto hppa-rtems.
>
>
> > Is the list above the correct list to look at to get the list of new
> > OSABI enums? I would like to suggest the addition of 
> >     GDB_OSABI_HPPA
> >     GDB_OSABI_HPPA_64
> >     GDB_OSABI_HPPA_BSD
> >     GDB_OSABI_HPPA_HPUX
> >     GDB_OSABI_HPPA_OSF
> >     GDB_OSABI_HPPA_PRO
> >
> > Does this look ok?
>
> I don't think they're necessary, by analogy with the existing code...
> certainly not GDB_OSABI_HPPA or GDB_OSABI_HPPA_64.  We probably need
> GDB_OSABI_HPUX.  We've already got OSF1, which is presumably the right
> OSF target.
Personally, I'd just, initally worry about:
	HPPA
	HPPA_64
(and perhaphs HPPA-elf if that is meaningingful?).  I'd expect all the 
others to be broken.

> If we're going to keep the anonymous "hppa-bsd" target we may need
> GDB_OSABI_BSD.  I don't know if hppa-proelf has its own OSABI or not.
Andrew