Re: Huge Apple gdb code dropping^H^H^H^H

[Jason Molenda <jmolenda at apple dot com>]

David Relston wrote:

> This is good news!  Objective C is one of the languages supported by
> gcc.  It is also the language used in the GNUstep project, which brings
> NeXTSTEP/OpenStep capability to Linux, BSD, Windoze, etc.  Of 
> importance to
> me, it allows a project I wrote for NeXTSTEP/OpenStep to run on Linux.

Yep, all the ObjC changes are in that mega patch.  As I wrote in the 
note, we
are working on a clean ObjC support patch to be submitted -- this is one 
of
the most useful big changes in our tree and we want to get it integrated 
into
the FSF sources.  I believe we'll have that patch out soon; Andrew is 
less
optimistic and wants an unfiltered patch in the hand instead of two in 
the
bush. :-)

> Objective C support has not been part of mainline gdb for several 
> years -
> at least since gdb 4.17.  There have been "unofficial" patches for 4.17,
> 4.18, and 5.0, but as gdb has evolved over the years it has become
> increasingly difficult to port the unofficial patches to each successive
> version of gdb.  Having the Objective C support code be part of the
> official, mainline of gdb is something I've wanted for the last couple 
> of
> years.

I'm not any kind of expert on the history of ObjC patches, but I believe 
these
patches are originally written by Michael Snyder when he was working at 
NeXT.
He's been trying to get a clear copyright assignment to the FSF for them 
for
years, but hasn't gotten that permission from the NeXT management.  
NeXT+Apple
has Seen The Light and is happy to assign changes to gdb back to the FSF.
(in theory, we could probably even get a clear statement about the old 
ObjC
support patches if someone was really interested in trying to bring them 
up to
date and merge them in.  But I'd obviously prefer that we go with the 
ObjC
patches in use at Apple. :-)

The changes on the net are probably not the same as the changes in the 
Apple
patch I submitted.  I gather there were multiple ObjC implementations 
done at
NeXT, and the one on the net was a big rewrite by Michael intended to 
fix a
number of bogosities in the other one.  We may have this 'other one' as 
our
main sources these days.  As I said, we've got no revision history from 
before
2000, and none of the people currently working on gdb have been at it 
for more
than a few years.

Jason