FC4 slimfast slimfest

[Elliot Lee]

On Mon, 21 Feb 2005, Phil Muldoon wrote:

> > Regardless of your opinions of the merits of the Java language, it's
> > just not heavily used by any of the things that desktop users bump into
> > every day.  (Except, *possibly*, web page applets, but I gather you're
> > talking about a much bigger scope than that.)
> 
> One of the products we have in RHEL is Red Hat Developer Suite, which is
> based on java. We do active development on this, and we'd like to bring
> it to Fedora. There is a huge community of Eclipse users out there.
> Also, there is a large sphere of java usage out there. I'm not going to
> argue usage metrics, but suffice to say there is enough usage that
> warrants interest im(h)o.

Java is critical because it's the only language and development
environment that's a serious alternative to .NET.

So why am I leaning towards pulling eclipse (and Java stuff) out of
FC4test1? Because it seems like that's part of the best short-term
solution to a short-term problem. All we need to do is have FC4 come out
on four CDs. If we can wind up removing enough "other" stuff from FC to
allow putting eclipse back for test2, that'll be awesome. These removals
aren't permanent by any means, but for now they seem the best way to go.

We already know what the long-term solution is - multi-repo support in
anaconda, and moving stuff to Extras (hopefully allowing us to slim Core
down even more).

Another problem is the distribution of ecj - having the main Java compiler
be part of the eclipse tarball is really an aweful situation. Fortunately,
gcjx will help solve this - hopefully we won't need ecj at all in the
future.

Cheers,
-- Elliot