Heya,
So I figured I'd stop being lame and send out an update of where we are
with the opensolaris.org release of JDS within Sun. Well, we're getting
there, but have a couple of hurdles still to figure out -
o We've got a whole heap of excellent people [1] working on
getting GNOME 2.10 to build properly on Solaris, and another
set of excellent people [2] kicking so much arse on fixing
some of the obvious bugs. So from a code and packaging
perspective, I think we're in pretty good shape despite
some functionality missing that we still need to think
about [3]
o From a process point of view, we're not quite there yet, and
have a number of stumbling blocks to figure out -
o Export Control - We've got to go through a full review
of the code that we're putting out onto opensolaris.org.
While we only ever intended to release the *deltas*,
it's still code and we need to play by the rules Sun has
in place. I think if the JDS community is going to
succeed, we'll need to have a mechanism of generating
binary packages as well - all the more need for review.
o SCM - There's the very obvious lack of a source code
management system. We use CVS internally for storing
our patches, in a module called 'spec-files' - kind of
ooooky name though. Right now there's lot of discussion
going on between Teamware and Subversion [4], but the
time lines for a decision don't really favour us. Right
now, if we can get past the export control stuff, we
could easily provide regular CVS snapshots. Ideally I'd
like an anonymous CVS mirror at the very least until we
sync up with the plans of everyone else.
o Website - I've started working on the website stuff, but
there's still a significant chunk of things to do -
http://www.opensolaris.org/os/community/desktop/communities/jds/building
I'd like to expand this, especially detailing some of
the functionality of pkgbuild, and some of the internal
scripts we use. If anyone feels like doing a tutorial of
pkgbuild, grab the tarball and start writing -
http://pkgbuild.sourceforge.net/
It's likely you'll need some of the spec files to help
you out - I can easily send you a tarball of the Solaris
10 branch if you're keen.
So I think that's about it - hopefully we'll make some progress over the
next week or two
Glynn
[1] Brian Cameron, Balamurali Viswanathan, Damien Carbery, Laszlo Peter,
and Matt Keenan
[2] Narayana Pattipati, Srirama Sharma, Archana Shah, Muktha Narayan,
and Arvind Samptur
[3] This includes things that haven't been properly thought about on
Solaris just yet, along with things that we've relied upon which
are now not available in the latest community code
[4] http://blogs.sun.com/roller/page/jag?entry=happily_subversive is
a good example of that discussion