Re: New gdb 31 & 64 bit patches for S/390

[Andrew Cagney <ac131313 at cygnus dot com>]

> If anything the code quality may be deteriorating owing to being asked to
> move code around so much & possibly being tired of retesting it,
> I currently have 4 versions of the code to support the post 5.0 32 & 64 bit
> patches the 5.0 32 & 64 bit patches & a lot of other work to do
> besides & probably falling back on it, when enthuasism is low my
> productivity deteriorates badly, this crap is pissing me off.
>  I will remove the non multi arch definitions when I don't have to support
> gdb-5.0.

That is your decision.

As things stand, the FSF doesn't own this code and, as a consequence, 
people have strong reservations about working on it.  (IBM will only 
assign the copyright using a letter that cites a digitally signed patch. 
  I understand this process to be slow and complicated.).

This is unfortunate.  As you'll recall, I have, on several occasions, 
suggested that instead of trying to get the source code up-to-scratch, 
it instead be assigned as-is to the FSF.  That way, others (probably me) 
would have been able to more directly help with the task of merging the 
changes in.

Instead, as you note, we've ended up with a slow painful process where 
the code would be submited and then I would eventually (since no one 
else step forward :-) review it.

Could I again suggest that the patches be assigned, as-is, to the FSF.

> We did our first 31 bit gdb drop 20 months ago.
> Even redhat has a experimental 64 bit drop for the last 6 months & it still
> isn't gnu.org the repository .

Yes, isn't this ironic.  I won't go near (and Red Hat can't contribute) 
those changes either since they are also legally owned by IBM :-)

	Andrew