Re: [PATCH] Bug 23438. Support for 128-bit long double types on PowerPC 64]

[Mark Kettenis <mark dot kettenis at xs4all dot nl>]

> Date: Tue, 11 Jul 2006 15:17:24 -0700
> From: Pete Carr <pecarr@linux.ibm.com>
> 
> Hi,

Hi Pete, Welcome to GDB.

> Please note that there are changes to floatformat.h and floatformat.c which
> are held under include and libiberty subdirectories.

Those bits need to be handled seperately; see the toplevel MAINTAINERS
file for details.

> This is my first patch submission so I'd appreciate guidance if I have
> not followed the rules.

You'll need to work a bit on your coding style ;-).  I noticed that
your lines tend to be too long (there's some debate what the maximum
length is, but they definitely should be under 79 characters), and you
comments are no formatted correctly.  Please take some time to read
the GNU coding standards; feel free to ask any questions that remain
after doing that.

As for the patch, the change to rs6000-tdep.c is almost certainly
wrong, since it also seems to affect 32-bit powerpc and POWER stuff.

Regarding the ppc-linux-tdep.c change: we generally don't #ifdef out
code, but just remove it, since CVS tracks the history for us.  But
again, I don't think this change is right, since the code is also for
32-bit powerpc support.  And even for powerpc64, there still are
systems out there that use the old ABI with 64-bit long doubles isn't
it?  How does this change affect those systems?

Mark