From: Godmar Back (gback@cs.utah.edu)
Date: Tue Nov 03 1998 - 20:01:01 EST
>
> On Nov 1, 1998, Godmar Back <gback@cs.utah.edu> wrote:
>
> > Whatever happened to egcs/gcj? Cygnus's attitude isn't exactly what
> > I'd call fair, given how they started out.
>
> What do you mean?
>
I meant Per Bothner's attitude, as shown in the following email exchange
that occurred on the java-discuss list at Cygnus:
Archie> [...] GCJ, which is based around the 0.9.2 version of kaffe?
Per> GCJ is not based on Kaffe. Some Kaffe-isms (such as field names)
Per> may still exist, but consider them historical artifacts.
Tim> And the fact that the internal structures generated by GCJ are
Tim> almost identical to those used by Kaffe, especially the old 0.9.x
Tim> release, is a mere coincidence? Even the field ordering is the same,
Tim> the layout of the constant pool, the numbering of the states a class
Tim> goes through while initialising. the method structure ... would you
Tim> like me to go on ...? I even think it's bug compatible in places :)
Per> [quoting Tim up to ``coincidence'']
Per> No, they are historical artifacts.
Per>
Per> We have no plans for Kaffe compatibility. Our goal is to
Per> replace any Kaffe code we might use (except that written by
Per> Cygnus employees). This is one of the things we have to do
Per> before a release. We feel this is necessary for legal reasons,
Per> since we do not have any of the necessary legal paper-work
Per> from Transvirtual.
Per>
Per> This is unfortunate, but the GNU project requires copyright
Per> assignments/disclaimers in general; for Java, we need to be
Per> even more careful.
Now I don't mind their general attitude --- I can see them competing
and everything --- but to say that their current gcj is not based on
Kaffe, when it clearly is, seems pushing it.
- Godmar
This archive was generated by hypermail 2b29 : Sat Sep 23 2000 - 19:57:02 EDT