[kaffe] [FW] [RFC] Debian Bugs information: Bug list for package kaffe
Dalibor Topic
robilad at kaffe.org
Sun Aug 15 10:50:36 PDT 2004
Arnaud Vandyck wrote:
> I think this one did not reach the mailing list :(
No idea why it didn't make it. But this one did ;)
> Subject:
> [RFC] Debian Bugs information: Bug list for package kaffe
> From:
> Arnaud Vandyck <avdyk at debian.org>
> Date:
> Fri, 13 Aug 2004 00:15:12 +0200
> To:
> Kaffe Mailing List <kaffe at kaffe.org>
>
> To:
> Kaffe Mailing List <kaffe at kaffe.org>
> CC:
> Debian Java Kaffe Maintainers <pkg-java-kaffe at lists.alioth.debian.org>,
> Ean Schuessler <ean at debian.org>, Adam Heath <doogie at debian.org>
>
>
> Hi Kaffe lovers,
>
> Here is a summary of bugs that has been reported to kaffe Debian
> package. I try to comment every bugs, but thanks to comment it also.
>
> Each bug is separated by a line. First, there is the bug number (you can
> find it if you point your browser to http://bugs.debian.org/XXX where
> XXX is the number without the '#'), title of the bug, number of days the
> bug has been reported. After that, I boxed a custom summary of different
> mails that have been exchanged about the bug; then a personnal note
> about the bug. There is a title to the 'box': `Debian' title means the
> bug has to be resolved by Debian (or anyone who want to help);
> `Kaffe' title means the bug is in the kaffe side.
>
> To be clear in your replies, cleanup all non relevant informations
> (including this text ;-))
>
> Debian Bug report logs: package kaffe
>
> Important bugs - outstanding (4 bugs)
>
> * #227741: kaffe FTBFS on buildd for ARM; 210 days old.
>
> ,----[ Debian ]
> | > jikes is broken on arm
> | > see http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=230601
> | >
> | > The problem will be solved when the arch and indep builds will be
> | > completely split because rt.jar does not have to be built on allarches.
> |
> | [1] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=257066&msg=6&att=0
> `----
>
> Ean, Doogie, the Debian package is really complicated and I don't have
> that time to give to a single package. We have to split the build
> process to arch-dep and indep. The split is done *after* the package has
> been built... we need to split the process *before*.
Yeah. Or we
a) wait for jikes to be fixed :) or
b) make sure that debian does not need to rebuild rt.jar anymore and can
use the one shipped with kaffe tarballs or
c) split the rt.jar away into its own package (kaffe-classpath?).
> * #259055: Kaffe: FTBFS on mips and mipsel; 30 days old.
>
> ,----[ Kaffe ]
> | >>[kaffe] Fixing mipsel-linux: differences between jit3 and jit1
> | >
> | > [...]
> | >
> | >>Done & checked in. Cross compiles for both i386 and mipsel, so I
> | >>guess it's o.k. :)
> | >>
> | >>cheers,
> | >>dalibor topic
> | > It does seem to be the LOUT stuff referred to in the thread, as
> | > that's what I see in the buildd logs.
> | >
> | > [1] http://www.kaffe.org/pipermail/kaffe/2004-July/098810.html
> |
> | Yep, I've also fixed some other build problems with jit3 for mips(el),
> | so the CVS head crosscompiles fine for mips. I'm working with arnaud
> | on making a 'dist' of CVS head build fine on all debian platforms,
> | natively.
> |
> | cheers,
> | dalibor topic
> `----
>
> 2004-07-11 Dalibor Topic <robilad at kaffe.org>
>
> * config/mips/jit3-mips.def: Changed LOUT to use
> jit3 syntax.
>
> Dalibor, Is #259055 closed for you (or maybe another Changelog entry?)?
Yeah, the LOUT problems have been fixed. I have no access to a
mips-linux box, but I'd assume that we don't have the problems anymore.
> * #264919: javax-servletapi2.3: FTBFS on amd64
>
> ,----[ Kaffe ]
> | > Hi, javax-servletapi2.3 fails to build on amd64, see attached buildd
> | > log.
> |
> | [...]
> |
> | > /usr/lib/kaffe/bin/java -classpath
> | > /usr/share/java/gnujaxp.jar:/usr/share/ant1.6/lib/ant.jar:\
> | > /usr/share/ant1.6/lib/ant-launcher.jar \
> | > org.apache.tools.ant.Main -Dbuild.compiler=jikes \
> | > -Dant.home=/usr/share/ant1.6 dist
> | >
> | > make: *** [build-stamp] Error 1
> | > ***********************************************************
> | > Build finished at 20040810-1218
> | > FAILED [dpkg-buildpackage died]
> | > -----------------------------------------------------------
> |
> | Hard to say something! Maybe kaffe is not doing the job on AMD64!
> |
> | I've just rebuild servletapi with pbuilder on powerpc and everything
> | was fine.
> `----
>
> Is Kaffe running well on amd64? Is Kaffe running ant1.6 on amd64 has
> already been reported successfully?
It used to. But then some of my warnings fixing effort broke it for
good. See http://www.kaffe.org/pipermail/kaffe/2004-July/098763.html and
followups for details.
I'll go back to amd64 once I'm happy enough with the build process and
have the kaffeh->cp-tools transition done.
> Normal bugs - outstanding (6 bugs)
>
> * #159162: sid Kaffe does not support LFS; 1 year and 345 days old.
>
> ,----[ Kaffe ]
> | Kaffe fails when extending files to more than 2Gb-1 (when running
> | freenet). This suggests it has been compiled without large file
> | support.
> `----
>
> This bug was reported for release 1.0.6 of Kaffe. I parsed all the
> change logs and did not find anything about that.
I'd guess that all that's needed is adding AC_SYS_LARGEFILE to
configure.ac and reconfiguring. And `AC_FUNC_FSEEKO' if we use fseek &
friends.
> _________________________________________________________________
>
> * #253755: kaffe: Javasound no longer works; 62 days old.
>
> ,----[ Debian ]
> | Reporter> Javasound has stopped working since the last release.
> |
> | Dalibor> Kaffe should work with both ALSA and ESD. Chances are that you
> | Dalibor> need to include the dependencies of these packages, too, I
> | Dalibor> don;t know if debian's packaging system resolves that kind of
> | Dalibor> things automatically.
> `----
>
> I already tried to build with sound but I think it did not build on some
> arches. Am I right? Does the sound work on every arches?
>
> If not, which architectures are supposed to work, not work with sound?
No idea ;) In theory, both alsa and esd should work on all linux
platforms, but none of my hacking boxes has sound or speakers so I can't
say if or how well it works.
> _________________________________________________________________
>
> * #261188: kaffe: Kaffe javah not registered in alternatives;
> 19 days old.
>
> ,----[ Debian ]
> | Kaffe does not install it's javah in the alternatives
> | system during installation.
> `----
>
> We (Debian) need to add a manpage and an entry to the postinst
Reminds me to pull in yout xml-ized man pages eventually.
> _________________________________________________________________
>
> Normal bugs - forwarded to upstream software authors (1 bug)
>
> * #255502: kaffe: Incorrect result due to computations in extended
> precision on x86; 52 days old.
>
> ,----[ Kaffe ]
> | Concerning the following Java source:
> |
> | ------------------------------------------------------------------
> | // $Id: test.java 3730 2004-06-21 13:02:51Z lefevre $
> |
> | public class test {
> | public static void main(String[] args) throws Exception {
> | test t = new test();
> | t.doTest();
> | }
> |
> | volatile double x, y, z;
> |
> | public void doTest() {
> | x = 9007199254740994.0; /* 2^53 + 2 */
> | y = 1.0 - 1/65536.0;
> | z = x + y;
> | System.out.println("z = " + z);
> | }
> | }
> | ------------------------------------------------------------------
> |
> | I've compiled it with "gcj -C test.java" (GCC 3.3.4).
> |
> | Sun's JVM gives the correct result:
> |
> | greux:~/wd/src/fp> /usr/local/j2re1.4.1/bin/java test
> | z = 9.007199254740994E15
> |
> | but not kaffe:
> |
> | greux:~/wd/src/fp> kaffe test
> | z = 9.007199254740996E15
> |
> | kaffe should switch the FPU of the x86 processor (Pentium III in my
> | case) to rounding in double precision to avoid the effect of the
> | "double rounding" (you may find some information about this effect
> | here: <http://www.srware.com/linux_numerics.txt>).
> `----
>
> Good luck for this one ;-)
Well, no idea how to do in the jit, but there are some nice fpcontrol
functions in C99, that might be able to serve for that purpose.
cheers,
dalibor topic
More information about the kaffe
mailing list