Re: [RFC] solib for darwin

[Tristan Gingold <gingold@adacore.com>]

On Dec 14, 2008, at 6:37 PM, Ulrich Weigand wrote:
> The following patch combines the ops->bfd_open callback I need for
> the Cell debugger with the OBJF_KEEPBFD, and also extracts the bulk
> of the solib_bfd_open functionality into subroutines that might be
> re-used by target-specific implementations.
>
> Can you verify that this would also allow a Darwin implementation?

Hi Ulrich,

yes this patch also allows a Darwin implementation.  I also think that  
this
approach is better than mine as it is more generic and is smaller.

I would be very happy if it were integrated.

Just a side comment on your patch:

> diff -urNp gdb-orig/gdb/objfiles.c gdb-head/gdb/objfiles.c
> --- gdb-orig/gdb/objfiles.c	2008-12-06 20:28:59.000000000 +0100
> +++ gdb-head/gdb/objfiles.c	2008-12-14 17:52:16.000000000 +0100
> @@ -424,7 +424,7 @@ free_objfile (struct objfile *objfile)
>
>   /* We always close the bfd. */
>
> -  if (objfile->obfd != NULL)
> +  if (objfile->obfd != NULL && !(objfile->flags & OBJF_KEEPBFD))

Humm, the above comment is not true anymore!

Thank you,
Tristan.