Clarify the warning displayed when you debug a core file for the wrong architecture.

Pushing one of the old patches from Daniel into FSF.

2002-08-30  Daniel Jacobowitz  <drowmvista.com>

        * corelow.c (get_core_register_section): Clarify
warning.

Index: gdb-6.3.90_20051109/gdb/corelow.c
============================================================
=======
--- gdb-6.3.90_20051109.orig/gdb/corelow.c      2005-07-04 
13:29:10.000000000 +0000
+++ gdb-6.3.90_20051109/gdb/corelow.c   2005-11-09
23:23:54.000000000 +0000
 -462,7
+462,10 
   if (! section)
     {
       if (required)
+      {
        warning (_("Couldn't find %s registers in core
file."), human_name);
+       warning (_("Core file may be for a different
architecture.n"));
+      }
       return;
     }

Nitin Gupta <nguptamvista.com> writes:
> Pushing one of the old patches from Daniel into FSF.
>
> 2002-08-30  Daniel Jacobowitz  <drowmvista.com>
>
>        * corelow.c (get_core_register_section): Clarify
warning.
>
> Index: gdb-6.3.90_20051109/gdb/corelow.c
>
============================================================
=======
> --- gdb-6.3.90_20051109.orig/gdb/corelow.c     
2005-07-04
> 13:29:10.000000000 +0000
> +++ gdb-6.3.90_20051109/gdb/corelow.c   2005-11-09
23:23:54.000000000 +0000
>  -462,7 +462,10 
>   if (! section)
>     {
>       if (required)
> +      {
>        warning (_("Couldn't find %s registers in
core file."), human_name);
> +       warning (_("Core file may be for a
different architecture.n"));
> +      }
>       return;
>     }

I don't object to this patch, but I wonder if there isn't a
better
place to do the check, where we could issue a more
informative
message.

Does this warning ever come about when the core file is
*not* of a
different architecture?  I'd assume not, since the caller
indicated
that the section in question is required.

(Sorry if this patch has been discussed before --- I
couldn't find the
thread in the archives.)

>
>  Nitin Gupta <nguptamvista.com> writes:
> > Pushing one of the old patches from Daniel into
FSF.
> >
> > 2002-08-30  Daniel Jacobowitz  <drowmvista.com>
> >
> >        * corelow.c (get_core_register_section):
Clarify warning.
> >
> > Index: gdb-6.3.90_20051109/gdb/corelow.c
> >
============================================================
=======
> > --- gdb-6.3.90_20051109.orig/gdb/corelow.c     
2005-07-04
> > 13:29:10.000000000 +0000
> > +++ gdb-6.3.90_20051109/gdb/corelow.c   2005-11-09
23:23:54.000000000
> > +0000
> >  -462,7 +462,10 
> >   if (! section)
> >     {
> >       if (required)
> > +      {
> >        warning (_("Couldn't find %s registers
in core file."),
> > human_name);
> > +       warning (_("Core file may be for a
different architecture.n"));
> > +      }
> >       return;
> >     }
>
>  I don't object to this patch, but I wonder if there
isn't a better
>  place to do the check, where we could issue a more
informative
>  message.
>
>  Does this warning ever come about when the core file
is *not* of a
>  different architecture?  I'd assume not, since the
caller indicated
>  that the section in question is required.
>
>  (Sorry if this patch has been discussed before --- I
couldn't find the
>  thread in the archives.)

I don't like the change.  It's replacing a factual statement
with a
specualtion.

"Mark Kettenis" <mark.kettenisxs4all.nl> writes:
> I don't like the change.  It's replacing a factual
statement with a
> specualtion.

It doesn't *replace* the factual statement.  It *augments*
it with a
speculation.