Hi Ito,
Thanks for making the test. Yes. Kaffe normally supports 1.5 class
files. Though the new information are not necessarily yet well
integrated with the VM core they should be exported to classpath.
The patch looks right to me. Go ahead !
Regards,
Guilhem.
Ito Kazumitsu wrote:
Hi,
I
Fu, Elva wrote:
Hi all,
I installed kaffe-1.1.6 on Fedora 4 , but it failed when I try to
execute awt demo. My install steps as following:
$ ./configure
$ make
$ make install
Everything seems ok. Then I add kaffe path into PATH:
$ export PATH=$PATH:/usr/local/kaffe/bin
Makefile.in files in the tree and updated then
with cvs -qz3 update ...
Each make call leads now to the make[3]: F: command not found error
as described above ...
Do you have an idea how I could manage it? What happens?
Thank you for any reply in advance!
Best Regards
Vladimir
Guilhem Lavaux wrote
Vladimir
Guilhem Lavaux wrote:
Hi Vlad,
If you use --disable-gconf-peer your problem should vanish. But if you
have ubuntu 6.06 then you should be able to install the development
package of gconf2 and gtk using
apt-get install libgconf2-dev libgtk2.0-dev
Regards,
Guilhem.
vlad
Hi Vlad,
If you use --disable-gconf-peer your problem should vanish. But if you
have ubuntu 6.06 then you should be able to install the development
package of gconf2 and gtk using
apt-get install libgconf2-dev libgtk2.0-dev
Regards,
Guilhem.
vlad wrote:
Hello,
I've fixed my previous
Hi Jim,
Thank you very much for the move ! You are supporting kaffe's hardware
for such a long time now. :-)
Cheers !
Guilhem.
Jim Pick wrote:
Jim Pick wrote:
In about half an hour, I'm planning to move pogo.kaffe.org from it's
current dedicated machine to a new Xen session on one of my
Tim Bevan wrote:
I can reduce my real problem to the attached class and native metbod.
We are simply creating a weak reference to the class then releasing it
once we have finished with it. Deleting the weak reference aborts the
program!
Hello, world from java first
Got weak reference
Tim Bevan wrote:
I can reduce my real problem to the attached class and native metbod.
We are simply creating a weak reference to the class then releasing it
once we have finished with it. Deleting the weak reference aborts the
program!
Hello, world from java first
Got weak reference
Hi Michael,
The author of JamVM is trying to implement this behaviour. It is
strongly in development and probably CPU intensive when the heap is
shrinked because you have to update all pointers in your memory. This is
a somewhat complicated procedure as you have to scan the heap for all
pointers
Hi pancake,
What version of kaffe are you using ? CVS ? I am getting a right OOM
error when trying to call System.out.println. This is normal because
the object o may not yet be relaxed even if you put 'o = null'. What
sort of assertion failure do you get ?
I would expect something thrown
Michael Koch wrote:
Hello list,
I tried current CVS and it failed when building fastjar. There were
several problems which I tried to fix locally but failed. There are even
different errors when building --with-system-zlib or with the in-tree
zlib. I used $builddir == $srcdir if that matters.
Dalibor Topic wrote:
On Wed, 2006-04-05 at 09:29 +0200, Riccardo wrote:
Hello,
I was trying to compile kaffe on AIX 4.2 with gcc on PowerPC. The setup
is similar to the one I used at home where I experienced also
endianness issues which I fixed at the time, I did not check if
classpath
Hi,
First, the GC is scanning all stacks including the native stacks looking
for references to GC objects which have been allocated using gcMalloc.
So the only point which poses problems is obj = gcMalloc at the moment
of the =. In the facts it is not a problem because the thread is
suspended in
Hi Ioannis,
This is good news that at least the fix seems to correct most of the
problems. Concerning the cpu usage with pthreads I expect it is correct
that it really eats all the CPU (actually your test is meant to no ? :-))
The problem with jthreads is a bit more worrying. We really need
Hi Fernando,
ExceptionOccured() returns a reference to the exception object. The
object is pushed on the local reference table (as do some other VM,
which is logical as it is a reference). You should use ExceptionCheck
instead: it returns a boolean which is true if an exception occured and
Hi Ioannis,
I have downloaded the file. However I am not sure I will be able to
reproduce your exact system state and this is a problem because core
dumps depends on the exact libc, the exact gcc, ... So probably _I_ will
not be able to exploit it. However _you_ can. ;-) I can guide you if
Ioannis Liverezas wrote:
Hi all,
Any progress made with the jthreads problem? Last post about this problem was
on 13/01/2006.
It's my first priority currently. I didn't manage to reproduce the
problem with jthreads atm however pthreads has a really painful bug. For
some reason the lists
+1,9 @@ +2006-01-14
Guilhem Lavaux [EMAIL PROTECTED] + + *
libraries/clib/sound/alsa/check_constants.h: Check + for the macro
definition because using the assert to check +the value (needed for
alsa-lib-1.0.11-rc2). + 2006-01-14 Dalibor Topic [EMAIL PROTECTED]
* README, configure.ac
Hi all,
I am pleased to announce that Kaffe.org has now a functional bugzilla
online. Now all kaffe bugs shoud go into this bugzilla except if they
are classpath's related bugs and in that case the bugs will go into
classpath's bug tracker. There is a detailed notice in the front page of
the
Hi,
I have forgotten the essential ! :)
To access bugzilla directly:
http://www.kaffe.org/bugzilla/
Guilhem.
Guilhem Lavaux wrote:
Hi all,
I am pleased to announce that Kaffe.org has now a functional bugzilla
online. Now all kaffe bugs shoud go into this bugzilla except
Hi Kiyo,
Paths are not right in Makefile.am (you can see -I../zlib/ but it should
be -I$(srcdir)/../zlib/) . I think I have fixed it now. Could you try
again ?
Cheers,
Guilhem.
Kiyo Inaba wrote:
Hi,
Thanks to improve using fastjar for kaffe. After this introduction,
yesterday is the
Hi Ioannis,
I am writing to you to say that I am really considering your problem.
Actually the VM does look into a strange state where it eats all the CPU
without going further anymore (though I am not sure yet at the moment I
have to analyze the dynamic of what it is doing). I am
Ito Kazumitsu wrote:
From: Ito Kazumitsu [EMAIL PROTECTED]
Date: Thu, 22 Dec 2005 11:08:15 +0900 (JST)
The runtime access control is much simpler than the compile-time
access control. And we may be able to omit the checking of nested
relations of classes at runtime.
Simply omitting
Hi Ito,
I have met a new IllegalAccessError in eclipse. I have attached two
files which reproduce the problem.
The protected bit is difficult to handle apparently. I have talked with
Tom Tromey and it says that normally we should completely ignore the
accessibility flag in inner classes.
Hi Ito,
Here is a better example. The previous could not be compiled directly if
you try to compile indepedently then you'll see the runtime accepts it.
Here I have an example which compiles and should run on a VM. However
kaffe throws an IllegalAccessError.
Regards,
Guilhem.
Ito Kazumitsu
Ito Kazumitsu wrote:
Hi Guilhem,
I have met a new IllegalAccessError in eclipse. I have attached two
files which reproduce the problem.
legal2.java did not compile.
Found 1 semantic error compiling legal2.java:
7. legal1.A a = legal1.getA();
Hi,
After checking a few things here and there it appears that we are too
strict. The VM must only rely on access flags of the class to check the
accessibility. The inner class flags are used by the other compiler and
the reflections. So we may remove everything related to inner/nested
Hi Ito,
Ito Kazumitsu wrote:
From: Ito Kazumitsu [EMAIL PROTECTED]
Date: Fri, 09 Dec 2005 13:31:58 +0900 (JST)
So the question is whether the assumption that returned test array is
consistently sorted is valid or not. If the assumption is valid,
then ecj may be wrong. Otherwise we will have
Hi Wolfgang,
I'll check in a patch for that tonight... I have also noticed that ecj
was causing some troubles to kaffe and I'll remedy to them a bit later.
Regards,
Guilhem.
Wolfgang Baer wrote:
Hi Ito,
Ito Kazumitsu wrote:
Using the following poor man's ecj,
$ cat bin/ecj
#!/bin/sh
Hi Wolfgang,
Wolfgang Baer wrote:
Hi Guilhem,
Kaffe CVS wrote:
PatchSet 6980 Date: 2005/12/03 12:54:16
Author: guilhem
Branch: HEAD
Tag: (none) Log:
Fix to be able to use ECJ instead of JIKES.
* configure.ac: Added a new option --with-ecj to be able
to compile with ecj. Reworked the
Ito Kazumitsu wrote:
From: Kaffe CVS [EMAIL PROTECTED]
Date: Sat, 03 Dec 2005 05:04:09 -0800
Fix to be able to use ECJ instead of JIKES.
Being an old-fashined programmer, I am not interested in
Eclipse, but would like to try ECJ because JIKES seems to
be inactive now. Unfortunately, I can
Hi Wolfgang,
It looks like ecj/ant discover the bootclasspath automatically
somewhere. However with the integration of GNU Classpath the runtime jar
is now called glibj.zip (maybe we should change this). Ant does not
discover this and produce a wrong bootclasspath/classpath. As a
consequence
Hi Daniel,
If you don't have enough memory then the VM should tell you explicitly.
In that case I expect it is more a bug in the trampolines of kaffe. 16MB
of ram should be sufficient for the minimal usage of kaffe (we had
reports that we need approximately 5MB at minimum to load GNU
Hi Jean-Frederic,
The problem should be fixed now in CVS. I have tested your program
locally with kaffe and it seems to work now. Please tell me if you are
still experiencing some troubles...
Regards,
Guilhem.
Jean-frederic Clere wrote:
Hi,
I have tried a small test program that runs
Gianluca Moro wrote:
hi all,
I'm running kagge with a graphical output
directly on X, without any window manager.
I see that all the applicatio I start are placed in (0,0)
(upper lefr corner of the screen) while starting
my java application under kaffe it is placed in a bit
on the left, and
Marko Merl wrote:
hey,
today i installed antivir-workstation-6.32.0.56-linux. it runs perfactly
in the console but if i start the gui (antivir-gui) i got this
error-message:
--- ERROR ---
kaffe-bin:
/build/buildd/kaffe-1.1.6/build-tree/kaffe-1.1.6/kaffe/kaffevm/support.c:351:
Hi,
sysdepCallMethod is not simply this. It ensures that the arguments are
passed correctly according to the ABI for some architecture/os/compiler
combination. Generally it must be written in assembly or using some
tricks of the compiler. sysdepCallMethod takes a pointer to some
function,
Ito Kazumitsu wrote:
Hi,
From: Guilhem Lavaux [EMAIL PROTECTED]
Date: Wed, 19 Oct 2005 22:55:35 +0200
The classpath tree has been merged. I know that it builds, runs
regression tests install peacefully here but I haven't yet tested
make dist. If there are problems they will be fixed
for GCC Bugzilla Bug 23008 and 22968.
+
2005-10-19 Guilhem Lavaux [EMAIL PROTECTED]
* developers/update-vm-list: New script adapted from update-class-list
Index:
kaffe/libraries/javalib/external/classpath/gnu/java/nio/charset/UTF_16Decoder.java
diff -u
kaffe/libraries/javalib/external
Ito Kazumitsu wrote:
Hi,
From: Guilhem Lavaux [EMAIL PROTECTED]
Date: Thu, 20 Oct 2005 20:59:38 +0200
Could you push your patch yourself on classpath-patches ml ? :) I could
do it but it is maybe better if you do it.
I reported the bugs and submitted the patches to Savannah last June
Dalibor Topic wrote:
Guilhem Lavaux wrote:
Hi,
Salut Guilhem,
As some people know I have nearly finished merging the classpath tree
into kaffe tree. There are still a couple of issues to solve with
installing classpath libraries into the right places and how to deal
with glibj.zip (build
Hi,
The classpath tree has been merged. I know that it builds, runs
regression tests install peacefully here but I haven't yet tested
make dist. If there are problems they will be fixed progressively.
Cheers,
Guilhem.
___
kaffe mailing list
Hi Marcus,
Thanks for the bug report. As we still don't have a bug database it is a
bit difficult to track them sometimes. So thanks again for reminding us
about that fact. I'll check if I can reproduce the bug (in that case
I'll fix it asap) if not I'll ask you to do a few things.
ATM, I
Antoine Reilles wrote:
Hi,
There is a bug report in the netbsd bug tracker about the use of
sigcontext in the pthreads code of kaffe.
http://www.netbsd.org/cgi-bin/query-pr-single.pl?number=30585
The bug report and the proposed patches are for kaffe-1.1.5, but they
are propably relevant
Hi Gwenole,
I have just fixed an important bug in the GC on AMD64. GC was freeing
used memory on it and now the stack is walked correctly. However, I have
noticed there is still SMP oddities sometimes.
Cheers,
Guilhem Lavaux.
Gwenole Beauchesne wrote:
Hi,
Dalibor told me you'd know more
Gwenole Beauchesne wrote:
Hi,
Dalibor told me you'd know more precisely the status of boehm-gc vs.
kaffe-gc. I am asking because on x86_64 (gcc 4.0-branch, glibc 2.3.5), I
sometimes get failures in the testsuite where all threads are waiting for
something. This is with the default kaffe-gc.
just in case.
Regards,
Guilhem.
2005-08-15, h keltezéssel 08.44-kor Kaffe CVS ezt írta:
PatchSet 6830
Date: 2005/08/15 15:36:08
Author: guilhem
Branch: HEAD
Tag: (none)
Log:
Fixlet for Security
2005-08-15 Guilhem Lavaux [EMAIL PROTECTED]
* libraries/javalib
Hi Ito,
Two things:
* Commons-Net is flawed.
* Kaffe is flawed.
:)
First kaffe: I know that from the transvirtual times we have inherited a
synchronized access to native socket operations. That way we are sure
two threads are not accessing the same socket at the same time. However
it is not
!
Regards,
Guilhem Lavaux.
On Mon, 2005-07-25 at 11:39 +0900, Ito Kazumitsu wrote:
From: Ito Kazumitsu [EMAIL PROTECTED]
Subject: Re: [kaffe] SecureRandomTest failed
Date: Sun, 24 Jul 2005 08:18:31 +0900 (JST)
java.io.IOException: Inappropriate ioctl for device
I am afraid that on FreeBSD
and loadProviders. I know that some exceptions are hidden there and
these functions are in charge of loading the providers.
Thanks !
Regards,
Guilhem Lavaux.
On Sat, 2005-07-23 at 07:13 +0900, Ito Kazumitsu wrote:
I am afraid something has changed so that Security.getProviders() returns
Hi,
The patch looks ok to me. I will check it in reality just in case and
commit it.
Many thanks !
Regards,
Guilhem.
On Tue, 2005-07-19 at 14:41 -0400, Kurt Miller wrote:
From: Kurt Miller [EMAIL PROTECTED]
When jar is passed a path as a parameter using -t or -x, it
doesn't recurse into
Hi,
Thank you very much for your patch ! We are sometimes a bit low to
answer and sorry for that. I will try it and if it looks ok it will go
into CVS.
Regards,
Guilhem Lavaux.
On Fri, 2005-07-15 at 17:07 -0400, Kurt Miller wrote:
From: Kurt Miller [EMAIL PROTECTED]
can kaffeh be used
Hi Kiyo,
I've updated CVS once again to use old sigcontext structure if no
sys/ucontext.h is found.
I have tested it on qemu/netbsd 1.6/x86 and it was working without a
failure.
Cheers,
Guilhem.
On Mon, 2005-07-11 at 20:35 +0900, Kiyo Inaba wrote:
Hi Guilhem,
Thanks. I've missed that in
Hi Kiyo,
Thanks. I've missed that in Riccardo's patch.
Could you try again CVS head now ?
Guilhem.
On Sun, 2005-07-10 at 09:10 +0900, Kiyo Inaba wrote:
Latest snap failed to be compiled on NetBSD 1.6 (ia32).
Error log said,
In file included from ../../config/md.h:1,
from
Hi,
We are only missing two functions in our JNI 1.4 implementation:
NewWeakGlobalRef and DeleteWeakGlobalRef. These are intended as
lightweight weak global reference builders in the VM. They create weak
object reference from any object reference. From the documentation, it
looks the weak
On Sat, 2005-06-25 at 11:36 +, Riccardo wrote:
Hello,
Hi Riccardo,
I was building on netbsd ppc an intrp+jthread version (since pthreads
are flawed in the kernel according to Guilhem)
In file included from ../../../../kaffe/libraries/clib/fdlibm/mprec.h:34,
from
Hi,
I am not a JVMPI expert but SetThreadLocalStorage is only there to
record that some pointer is affected to some thread. I am not sure this
must be walked by the GC as this may be a weak reference. The
documentation on sun's site is not clear on that point and would require
further testing to
Hi,
I'm proposing to move KaffeVM_callMethodX to its own file as they are
very hackish and I need a few of the defined macros to write another
machine dependent function to count the real number of arguments which
will be used by callMethod.
Anything against this ?
Regards,
Guilhem.
Hi Daniel,
I've just fixed the problem in CVS head.
Tell me if there's something else you see in the regression tests. :)
Cheers,
Guilhem.
On Thu, 2005-06-09 at 18:26 +0200, Daniel Bonniot wrote:
Helmer, thanks a lot for your fix!
With the same setup, here is another bug (I have been
Hi Ito,
Could you add (*env)-ExceptionDescribe(env) at the point where there is
a failed assertion. It should work now in any case. I guess that
FindClass returns NULL because it cannot find RawData* but we must
check.
Regards,
Guilhem.
On Sat, 2005-06-04 at 08:33 +0900, Ito Kazumitsu wrote:
On Thu, 2005-06-02 at 14:45 -0500, Adam Heath wrote:
I've done a small straw poll on irc.kaffe.org#kaffe, to see what people think
about switching kaffe away from cvs, and to svn. Everyone who's responded has
been for the idea.
So now, I'm posting publically. What do all the existing
Hi,
The standard detection of stack boundaries involves the production of a
segmentation violation by the program. It is caught by kaffe and handled
correctly. You can safely ignore the message as kaffe seems to work.
Regards,
Guilhem.
On Sun, 2005-05-01 at 17:36 +0800, huangtonglang wrote:
On Sun, 2005-04-17 at 18:53 -0400, Michael Franz wrote:
On 12/30/04, Michael Franz [EMAIL PROTECTED] wrote:
Here is the diff -u .
Any reason why this was not applied?
Michael
No special reason besides forgetting. :)
Checking it in now.
Guilhem.
On Sat, 2005-04-16 at 13:13 -0700, Eric Anholt wrote:
Seeing only one regression failure on FreeBSD/amd64 (GCTest, which fails
100% of the time), I went on another warnings cleanup spree. No
improvement in regressions, but fewer warnings as a result. It's mostly
type cleanups, but there's
Hmmm... it seems kdevelop has a good indentation algorithm. I haven't
noticed it has modified all these lines.
Guilhem.
On Sat, 2005-04-16 at 07:19 -0700, Kaffe CVS wrote:
PatchSet 5677
Date: 2005/04/16 14:15:27
Author: guilhem
Branch: HEAD
Tag: (none)
Log:
Fixlet for OS which do not
On Sat, 2005-04-16 at 14:53 -0500, Adam Heath wrote:
On Sat, 16 Apr 2005, Guilhem Lavaux wrote:
Hmmm... it seems kdevelop has a good indentation algorithm. I haven't
noticed it has modified all these lines.
That's why I always run diff before committing.
Actually this should
Hi,
This bug has already been fixed in CVS head.
Thanks for reporting it.
Regards,
Guilhem.
On Thu, 2005-04-14 at 17:19 +0200, Laurent Martelli wrote:
The following class raises a NullPointerException with Kaffe because
of the null passed to the WeakReference constructor. It should not.
kernel
Control: 107D
thank you for help.
regards.
On 4/13/05, Guilhem Lavaux [EMAIL PROTECTED] wrote:
On Wed, 2005-04-13 at 09:38 +, Hammam Mounir wrote:
hi all;
in file .../kaffevm/jar.c we need to map rt.jar file in memory
Hi,
Thanks for the bug report and the patch !
I'll commit it asap.
Regards,
Guilhem.
On Tue, 2005-04-12 at 19:41 -0700, Mike Castle wrote:
* Minor fix to include directories ordering
In my case, I build kaffe with --prefix=/usr/X11R6. Since I had 1.1.4
installed already, it picked up
On Wed, 2005-04-13 at 09:38 +, Hammam Mounir wrote:
hi all;
in file .../kaffevm/jar.c we need to map rt.jar file in memory
(function OpenJarFile(..)), but uClinux does not support file mapping
(CPU=ARM OS=uClinux).
Hi,
Normally openJarFile should handle systems which does not support
On Tue, 2005-04-12 at 17:16 +0300, Nektarios K. Papadopoulos wrote:
Hi all,
the attached patch fix a bug in jthreadedRecvfrom.
The case of NOTIMEOUT was not handled correctly, passing negative values to
waitForTimeout. The result was maximum cpu consumption.
Hi !
Thanks for the patch.
On Tue, 2005-04-12 at 12:45 -0400, Barry Hawkins wrote:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Michael Koch wrote:
[...]
I have a little feature request. Please output the default values for
stak size, maximum and initial heap size. These values differ from arch
to arch and tis
Hi,
Just to mention that I've finally have made kaffe passed all tests on
jakarta-commons/collections (0 failures). I will run some more tests on
the other packages.
Regards,
Guilhem.
___
kaffe mailing list
kaffe@kaffe.org
On Mon, 2005-04-11 at 23:22 +0900, Rei Odaira wrote:
Hi,
+ while (1)
+{
+ r = select (n, readfds, writefds, exceptfds, timeout ? delay :
NULL);
+
+ if (r 0 errno != EINTR)
+ return -errno;
+ else if (r = 0)
+ return r;
Could anyone replace this
Hi Rei,
Thanks ! I've checked the patch and it looks ok. There isn't any
regression either. So I've checked it in.
Regards,
Guilhem.
On Tue, 2005-04-12 at 01:10 +0900, Rei Odaira wrote:
Hi,
div_int_const_optimize() in jit3 cannot handle
a negative dividend correctly.
Hi Rei,
Thanks for the patch ! I haven't yet reread it thoroughly though. As
this is a GNU Classpath bug it should also be sent to:
http://savannah.gnu.org/bugs/?group=classpath
Regards,
Guilhem.
On Sun, 2005-04-10 at 01:27 +0900, Rei Odaira wrote:
Fernando Lozano wrote:
Having read many
Hi,
No need to file the bug. I have done it myself.
Regards,
Guilhem.
On Sun, 2005-04-10 at 01:27 +0900, Rei Odaira wrote:
Fernando Lozano wrote:
Having read many reports on the archieve of success running Tomcat under
Kaffe, I decided to try myself using Tomcat 5.0.28. Although many JSP
Hi Rei,
Thank you very much for the patch ! It looks good to me. The last two
files should also be sent to GNU Classpath though. As they are trivial
bugs I think there should not be any problem going into the CVS.
I am committing this now.
Regards,
Guilhem.
On Sun, 2005-04-03 at 17:04 +0900,
On Sun, 2005-04-03 at 12:28 +0200, Michael Koch wrote:
On Sun, Apr 03, 2005 at 05:04:32PM +0900, Rei Odaira wrote:
--- libraries/clib/nio/gnu_java_nio_VMSelector.c1 Apr 2005 01:02:56
- 1.5
+++ libraries/clib/nio/gnu_java_nio_VMSelector.c3 Apr 2005 07:24:27
-
@@
On Sun, 2005-04-03 at 20:08 +0200, Christian Thalinger wrote:
While testing and implementing CACAO's defineClass, if found this issue:
$ cacao MyClassLoader
Exception in thread main java.lang.ArrayIndexOutOfBoundsException
No stacktrace available
$ kaffe MyClassLoader
class
compliant (at least concerning the
invokation interface because some JNI functions are still missing).
Regards,
Guilhem Lavaux.
___
kaffe mailing list
kaffe@kaffe.org
http://kaffe.org/cgi-bin/mailman/listinfo/kaffe
Ito Kazumitsu wrote:
ThreadLocalTest fails recently. It does not finish until killed by the
killer process.
Thanks Ito !
I'm also experiencing other deadlocks on freebsd. I'll have a look at
where does they come from.
Guilhem.
bash-3.00$ uname -a
FreeBSD mq.maczuka.gcd.org 5.3-RELEASE FreeBSD
Riccardo wrote:
Hello,
I've applied the patch however I don't whether it has been activated
(I've introduced a #define somewhere in config-jthreads.h). You may
check this by putting a dprintf in setupSigAltStack. I've attached a
patch for this.
I get no output if I insert that print statement.
Riccardo wrote:
Hey,
I'll try pthreads later and possibly on the other irix box whith gcc 2.
95.
running pthreads instead of jthreads makes only 20 reg. fail:
BadFloatTest.fail IntfTest.fail CatchDeath.fail
NetworkInterfaceTest.fail ClassGC.fail Preempt.fail
.
[EMAIL PROTECTED] wrote:
Dear all,
I just build fresh Kaffe with the new lock API committed by
Guilhem Lavaux. However, it failed to get compiled, and then
I attempt to fix that with the following patch:
$ cvs diff -up kaffe/kaffevm/locks.c
Index: kaffe/kaffevm/locks.c
Helmer Krämer wrote:
Guilhem Lavaux [EMAIL PROTECTED] wrote:
Hi,
I've just finished a functional patch for the locking system. I don't
know yet how much slower it is. But I've already had some interesting
results with some private app which was not working previously. If noone
is against I
Guilhem Lavaux wrote:
Helmer Krämer wrote:
Guilhem Lavaux [EMAIL PROTECTED] wrote:
Hi,
I've just finished a functional patch for the locking system. I don't
know yet how much slower it is. But I've already had some interesting
results with some private app which was not working previously
Timothy Stack wrote:
On Mar 8, 2005, at 5:38 AM, Helmer Krämer wrote:
If *lkp does not point to a heavy lock, a new one is allocated on the
heap.
Hi Tim,
I just want to interject a little note here, which I'm not sure is even
relevant...
Basically, Pat Tullmann (i think) once observed that the
Hi,
I've been quite busy these latter days so I haven't yet a patch ready
for locks. Here is a summary of the problems/solutions:
* currently kaffe uses a fast lock mechanism which needs to work to get
a unique pointer for each thread. These pointer represents the holder.
The stack pointer is
something for unix-jthreads asap and commit it.
Cheers,
Guilhem.
Daniel Bonniot wrote:
Guilhem Lavaux wrote:
Hi Daniel,
I had a few minutes free and checked it in the CVS. Now whenever there
is an uncaught exception happened kaffe-bin will return with 1 instead
of 0.
Cheers,
Guilhem.
Guilhem
Michael Franz wrote:
Hi,
I have been going over kaffe.def and slots.h to figure out how the
interpreter works. I understand what the slots.h macros are doing
(stack, pop, push, ) but I don't understand what the logical
difference is amoung rstack, wstack and stack? I am thinking Read
stack,
Hi Max,
I have tried to reproduce it without success so far. I'll try again on a
debian/linux 2.4 this evening. Meanwhile you can enable the production
of core dumps (ulimit -c 10 e.g.) start kaffe and get the stack
trace of the problem using gdb.
gdb /home/max/sys/kaffe/jre/bin/kaffe-bin
Hi Daniel,
I had a few minutes free and checked it in the CVS. Now whenever there
is an uncaught exception happened kaffe-bin will return with 1 instead of 0.
Cheers,
Guilhem.
Daniel Bonniot a écrit:
Hi list,
When execution terminates because an exception was thrown and never
caught, kaffe
Riccardo wrote:
Hey,
linux/x86
1 of 147 tests failed
cat RefTest.fail
Reference cleared
Reference not cleared
finalizer called
Hi,
Yeah. I know this one. When I have some time, I try to remove the alloca
hack I've added for the finalizer but it's difficult as I cannot act on
the pthread_*
bug ?).
* Copyright notices aren't there.
* There are no differences between the different reference flavour.
Implementing this would require a modification to kaffe-gc which will be
able to chose between 'must free' and 'should free'. :)
Cheers,
Guilhem Lavaux.
P.S.: This diff has been generated
I was a bit lazy to do :( ).
Regards,
Guilhem Lavaux.
___
kaffe mailing list
kaffe@kaffe.org
http://kaffe.org/cgi-bin/mailman/listinfo/kaffe
Helmer Krämer wrote:
Guilhem Lavaux [EMAIL PROTECTED] wrote:
Helmer Krämer wrote:
Davanum Srinivas [EMAIL PROTECTED] wrote:
looks like we need a lock for locking access to utfLockRoot :(
I'd think that the whole utfLockRoot thing is broken (you'd
need one utfLockRoot per thread since it's
other idea ?
Regards,
Guilhem Lavaux.
Please check your CLASSPATH and your installation.
java/lang/Error: Unexpected problems with default encoder
at java.lang.VMThrowable.fillInStackTrace(VMThrowable.java:native)
at java.lang.VMThrowable.fillInStackTrace(VMThrowable.java:76
Michael Franz wrote:
Hi,
When running (trying to run) kaffe on darwin/x86 I get the same native
library load issue as on darwin/ppc (I think this is currently fixed),
however, for the PPC version the paths are complete and on x86 there
are relative. Is there code that is making them complete for
Michael Franz wrote:
Thanks to your question I have reread the JNI spec and it looks like we
should on the contrary enable the JNI refs.
They are not needed in most cases because we store objects on the stack
but if the native call stores object in the heap we may lose track of
used objects and
1 - 100 of 335 matches
Mail list logo