In <[EMAIL PROTECTED]> Dalibor Topic
wrote:
> looks like you've found a bug. ;)
>
> could you compile and run developers/sp_ofset.c and check if the value
> matches the one from CVS/kaffe/config/parisc/threads.h ?
torus 31: ./test
#define SP_OFFSET 1
this matches the one in the header
Hello,
continuing in my cross-platform benchmarking project, I tried to compile
kaffe (cvs version) on hp-ux 10.20 on a PA-RISC 1.1 (32bit)
during compile I get the following blocking error:
Making all in libraries/javalib gmake[1]: Entering directory `/usr/local/
kaffe-cvs/kaffe/libraries/jav
Hello,
as once Dalibtor requested, I tried to check out the 68k platform. Ince
I updated to the latest NetBSD (1.6.1) and during the update process
something went wrong, I reinstalled everything, this is why I had an
absecne.
Well I checked out CVS yesterday, waited faithfully all the hours t
In <[EMAIL PROTECTED]> Kiyo Inaba wrote:
> I am hesitate to reply for this, but if old story is still true, any
> c function arguments are converted to double from float. See, pp.137
> of "C a reference manual".
this can be no longer true using ANSI-C, instead of K&R
>
> # There should be better
In <[EMAIL PROTECTED]> Greg Wooledge wrote:
> I think it depends on the platform; not all platforms have the jit
> or jit3 yet. You can see which one you're using by running
> "java -version" (or "kaffe -version").
it reports on my solaris box using kaffe-cvs:
Engine: Just-in-time Version: 1.0
Hello,
as I told Dalibor, I'm mainly interested in kaffe because I am doing
some benchmarls about cross-platfrom java perfomance and I couldn't
resist to compare kaffe to JDK 1.1 that runs on my sparc.
I used kaffe-cvs configured with no options. If understand correctly,
this gives me "jit" as
Hello,
as Dalibor suggested, I tried to build kaffe-cvs on my NetBSD68k box to
see if 68k still works well on BSD. 68030 isn?t the fastest thing around
but well, for embedded stuff it is stilll used.
I checked out cvs and started configuring.
configure is S slow (about one hour, or maybe mo
In <[EMAIL PROTECTED]> Dalibor Topic
wrote:
>> I am running the checks now, the fail rate is not hight, but it is
>> there: FAIL: DoubleCvt.java FAIL: ThreadState.java FAIL: ProcessTest.
>> java FAIL: UDPTest.java ..
>
> could you post a gzipped tar file with test/regression/*.fail? the
> log
Hello,
well, me made a *significant* step forward: CVS builds and runs on
OpenBSD 3.3 for sparc!
I am running the checks now, the fail rate is not hight, but it is there:
FAIL: DoubleCvt.java
FAIL: ThreadState.java
FAIL: ProcessTest.java
FAIL: UDPTest.java
..
The DoubleCvt fails on my solaris
Hello
I know it is a horrid habit of asking again...
when attempt to build CVS on my OpenBSD box i get the following error
mkdir lib
/bin/sh ./rebuildLib @essential.files
Compiling classes from @essential.files using /home/multix/kaffe-cvs/
kaffe/kaffe/kaffe/kaffe-bin -verbosegc at.dms.kjc.M
In <[EMAIL PROTECTED]> Dalibor Topic
wrote:
> Since Helmer fixed the stack offset problem with your help, I'd say
> switch to CVS. You should get yourself the latest autotools as well,
> automake 1.7.5 and autoconf 2.57, since the configure script may ask
> for them. Make sure that you don't i
BUG !
SP_OFFEST refers to the stack pointer ?
In <[EMAIL PROTECTED]> Helmer Kr?mer wrote:
> would you be so kind and compile and run developers/sp_offset.c?
> If it prints out something other than "#define SP_OFFSET 1", we've
> already found the bug ;)
mallorn:developers {118} cc -o test sp_off
Hello,
as suggested I cleaned all of the old kaffe soruce dir and the kaffe
install dir. I rceonfigured with the interpreter engine and the debugger
information.
It still core dumps.
So I launched it with the debugging active, I get:
Starting program: /usr/local/kaffe/jre/bin/kaffe-bin Loops.j
In <[EMAIL PROTECTED]> Dalibor Topic
wrote:
> hm, openbsd works on i386, and sparc-solaris2 works as well. so it
> looks like something stupid happens with spacr-openbsd. let's try
> making sure that the intrp works first before we start debugging the
> jitter.
i reconfigured for the interpre
In <[EMAIL PROTECTED]> Dalibor Topic
wrote:
> If 1.1.0 doesn't really work either, please try to configure 1.1.0
> --with-engine=intrp and build it again. I'd be interested in the
> results of make check and any compiler warnings/errors.
for now I got 1.1.0, I removed the old /usr/local/kaffe
Hello
I tried to compile myself kafe 1.07 on OpenBSD 3.3 (sparc)
It configured and compiled and even installed without apparent problems
But running make check fails at the first test!
Even trying to run a small program causes a problem.
mallorn:multix {151} /usr/local/kaffe/bin/kaffe Loops
Int
16 matches
Mail list logo