Re: [PATCH,gdb,Update4]: ensures that cie ptr of an fda is a cie

[Tom Tromey <tromey@redhat.com>]

>>>>> "Fawzi" == Fawzi Mohamed <fawzi.mohamed@nokia.com> writes:

Fawzi> 2011-07-07  Fawzi Mohamed  <fawzi.mohamed@nokia.com>
Fawzi> 	* dwarf2-frame.c (decode_frame_entry, decode_frame_entry_1): Ensure
Fawzi> 	that CIE pointer of an FDE really points to a CIE .

Looks pretty good.

Fawzi> +	error (_("Found a CIE when not expecting it, broken eh_info"));

Remove ", broken eh_info" here.

Fawzi> +	error (_("Found an FDE when not expecting it, broken eh_info."));

And here.

The reason is that the error that is printed mentions .eh_frame, and I
think that is sufficient.

Fawzi> +	      warning (_("skipping .eh_frame info of %s: %s"),objfile->name, e.message);

Line too long.  Also, missing a space after the first "," -- but that is
where you should split the line anyhow.

Fawzi> +		  xfree(fde_table.entries);

Space before the paren.

Fawzi> +	  warning (_("skipping .debug_frame info of %s: %s"),objfile->name, e.message);

Likewise.

Fawzi> +		  xfree(fde_table.entries);

Space.

Fawzi> +		{
Fawzi> +		    xrealloc (fde_table.entries,
Fawzi> +			      fde_table.num_entries * sizeof (fde_table.entries[0]));

Wrong indentation for the xrealloc.

I think it is weird not to use the result of the xrealloc call, even
though we know it is just shrinking the buffer.

Tom