GCC 2.95 deprecation (Was Re: [Gnustep-cvs] GNUstep Testfarm Results)

[Gregory John Casamento]

Riccardo,

I'm not sure if I got that email, what "core" stuff are you referring to?   Please send it to the list again.

In the meantime, allow me to be clear on this: GCC 2.95.x is an ancient version of the compiler which doesn't support modern conventions in C99 as well such things as namespaces, certain types of foward declarations, -fconstant-string, and it also has problems with spin-lock (mutex locking) and etc.  It's silly for us to officially support it anywhere.

Also, any architecture which doesn't have GCC 3 support can and probably should be dropped as well.

The non-release critical policy, in this case, needs to cover all of GNUstep, not just "core pieces".

Later, GJC
--
Gregory John Casamento

----- Original Message ----
From: Riccardo <address@hidden>
To: Fred Kiefer <address@hidden>
Cc: Developer GNUstep <address@hidden>
Sent: Sunday, August 13, 2006 4:29:04 AM
Subject: Re: [Gnustep-cvs] GNUstep Testfarm Results

Hello,

On Friday, August 11, 2006, at 12:27 PM, Fred Kiefer wrote:

> I am not that sure about this. gcc 2.95 is a very old version, but it
> still has its own merits. It was the most popular release before the
> re-architecture and in many cases it is still faster than the later gcc
> releases. It also is in some cases the only version of gcc at hand. (I
> had to learn this when trying to do cross compilation of GNUstep)

I agree here. Sensing the hint of Gregory I would go as far as
considering 2.95 bugs non-critical, but trying not to actively break
2.95 and officially supporting it in the core stuff I mentioned in the
previous email.


Cheers,
   R



_______________________________________________
Gnustep-dev mailing list
address@hidden
http://lists.gnu.org/mailman/listinfo/gnustep-dev