Re: Java and Sparc for 3.1

[Tom Tromey <tromey at redhat dot com>]

>>>>> "Brad" == Brad Lucier <lucier@math.purdue.edu> writes:

Brad> Unfortunately, I do not believe that Java works well enough on
Brad> sparc-sun-solaris2.7 for release.  I'm running Solaris 2.8, and
Brad> lately I've only been testing on sparcv9, but the last time I
Brad> tested libgjc, I saw 262 errors in the test suite for libjava
Brad> tests.

There have been other much more favorable reports:

    http://gcc.gnu.org/ml/gcc-testresults/2002-03/msg00316.html

This one shows many execution failures for -m64, see also PR 6092, but
it does fine on the other run.

Rainer suspects a test suite problem:

    http://gcc.gnu.org/ml/java/2002-03/msg00289.html

I haven't looked into it.  I don't even know if I have access to the
hardware.

Can you find out why your tests are failing?  Ordinarily when all the
libgcj execution tests fail it indicates some kind of systematic error.

Brad> libgjc didn't build properly the last time I tried it when there
Brad> already were older versions of the libraries sitting in the
Brad> install directory.

This is a known problem.  Gerald is going to submit a PR for it.  We
can work around this at configure time.  However my understanding is
that the problem is in Sun's linker and so we can't actually fix it.
There's an ongoing discussion of this on the java list.

Brad> But in my opinion it will take some significant work before java
Brad> is ready for the big time on sparc.

There is work, but I'd guess it actually isn't that significant.


More generally, I don't think we can necessarily expect libgcj to work
on all the platforms that gcc itself works on.  Porting the Java
runtime requires a lot of different things to all work together.  I'd
rather we agree to promise that release N+1 of gcj will work on at
least all the platforms that release N worked on.  By that measure I
think we're doing ok.  Anybody with contrary information, please tell
me about it.

Tom