JDK ... FYI

[mkes at ra.rockwell.com]

Zrovna ted prisel takovy mail o pokroku v implementaci JDK pro FreeBSD:

On Wed, Jun 26, 2002 at 12:24:13PM -0400, Dylan Carlson wrote:
> On Tuesday 25 June 2002 07:10pm, Bill Huey wrote:
> How lean is the FreeBSD Java team right now?   Does this compromise the
target
> of FreeBSD shipping a Sun-approved JVM at all?   I think everyone would
> probably love to know more about the people behind the project here.

The Porting team is probably pretty lean at the moment.  Bill is working on
HotSpot and I'm working on JCK compliance.  I'm not sure on the status of
anyone else (feel free to let me know :).  This is all on a volunteer basis
and, at least for me, has to fit around work and everything else going on.

> Also I would like to know whether the JVM patchset releases are pretty
much
> frozen because of the work around KSEs.  I was hoping to see a more
> aggressive patchset release schedule (I would have expected us to be at 8
or
> 9 by now) ... but it seems like we're just dealing with the status quo
until
> FreeBSD 5.0 ships.  (That's my rough perception of it)

Almost all of the fixes that are in the repository (I can think of one
minor exception) are being installed as patches as part of the jdk13 port.
Bill's recent HotSpot work is the only really missing piece and that will
be addressed with patchset 7.  Its fine to want a more aggressive release
schedule, but there is no point releasing a new patchset if there isn't
really anything new to show.

That said, we're not bug free and I'll attempt to keep the bug fixes
coming as reports come in and fixes are found.  Right now, for instance,
I'm working on a bug in exec'ing an external process and waiting for it
to complete which is causing some of the JCK tests to fail bogusly.  If
I had a test case I'd post it, but I haven't narrowed it down yet.  I'll
also attempt to construct test cases for those JCK tests which are
failing and post them too so people can pitch in.  I can't just post
the tests themselves for obvious reasons.

For those who are interested there are currently 19 compiler failures
and 183 runtime failures that stand in the way of a binary release.  As
I mentioned above, many of the runtime failures are bogus and are due
to an exec bug (the tests pass when they are run by hand).

If anyone is interested in contributing, here are a couple of tasks in
no particular order:

1. Get native threads fully functional.  You probably need some pthreads
   knowledge for this, although I have none and I've hacked on things ;).
2. Start working on a port of 1.4.  I've imported the code but haven't
   had time to work on it yet.
3. Pick a bug from those listed for the jdk13 port and work on it.

If your primary interest is NetBSD or OpenBSD then work on making sure
things work there as well (OpenBSD in particular isn't even close with
respect to 1.3.1).

In short, its not a matter of status quo, its just that there hasn't been
a real justification for another patchset before Bill's HotSpot stuff.
--
Greg Lewis                            Email : glewis at eyesbeyond.com
Eyes Beyond                           Web   : http://www.eyesbeyond.com
Information Technology