Re: PING 3: [patch, testsuite] General ARM target triplet

["Ulrich Weigand" <uweigand@de.ibm.com>]

Richard Earnshaw wrote:
> On Wed, 2011-01-19 at 09:08 -0700, Yao Qi wrote:
> > On 01/13/2011 04:47 AM, Yao Qi wrote:
> > > On 01/06/2011 08:10 AM, Yao Qi wrote:
> > >> On 12/13/2010 11:39 PM, Pedro Alves wrote:
> > >>>>> OK, I combine them together in one single patch this time.
> > >>> Thanks. If/when everyone's happy with the triplet bits
> > >>> proper, this is okay.
> > >>
> > >> Ping. Are you happy with this triplet? :-)
> > >>
> > >> http://sourceware.org/ml/gdb-patches/2010-12/msg00187.html
> > >>
> > >
> > > Ping.
> > >
> >=20
> > Ping again with a small fix pointed out by Matthew.  Richard, are you OK=
> =20
> > with that?
> >=20
> > --=20
> > Yao Qi
> 
> +proc dwarf2_support {} {
> +    if {![istarget *-*-linux*]
> +       && ![istarget *-*-gnu*]
> +       && ![istarget *-*-elf*]
> +       && ![istarget *-*-openbsd*]
> +       && ![istarget arm*-*-eabi*]
> +       && ![istarget arm*-*-symbianelf*]
> +       && ![istarget powerpc-*-eabi*]} {
> +       return 0
> +    }
> +
> +    return 1
> +}
> 
> Why is this negated?  It would seem more sensible to me to have a list
> of supported targets that returns 1 and then default to returning 0 for
> everything else.  Why?  Because then if the list gets too long its
> easier to add a second cluster in an independent if-clause.

Agreed, it seems more maintainable to have this as a positive list.

> The ARM bits are all fine.  The dwarf2 changes look generally sensible,
> but should be reviewed by a full gdb maintainer, which I'm not.

I think the general bits were already approved by Pedro a while back.
Yao, please apply (with the change Richard suggested).

Thanks,
Ulrich

-- 
  Dr. Ulrich Weigand
  GNU Toolchain for Linux on System z and Cell BE
  Ulrich.Weigand@de.ibm.com