Re: [kaffe] Does Kaffe's JIT support ARM?
Jari Korva wrote: Hi Dalibor, I think that all patches that I have suggested are in cvs. The problem now is that the reason to current problems can't be found just by debugging (at least not easily) and me (and others who have observed the same problem) do not know enough about Kaffe internals (and haven't had enough time to sit down and educate ourselves :( No worries. A simple way to get started is to look at the compiler warnings for your $arch and get those fixed :) And don't get me wrong, I'm not blaming Kaffe. Last year when I was starting my project I tested around ten free/open source virtual machines and contacted four commercial vendors and Kaffe was the only one that worked on my big-endian XScale! That's not so bad then. I hope it will get back to that. :) cheers, dalibor topic ___ kaffe mailing list kaffe@kaffe.org http://kaffe.org/cgi-bin/mailman/listinfo/kaffe
Re: [kaffe] Does Kaffe's JIT support ARM?
Kiyo Inaba wrote: That's the only one, afaict. Kiyo has been working on m68k jits as well, Just before to getting into airplane, and becomes short. The m68k port with jit works with 20 or 30 errors. But I hope 'main part' of it works not only for ia32. I plan to check alpha since I got one WS recently ;-) great! For a bit of status info on alpha-intrp: Debian gnu/linux buildlogs of 1.1.6-3 include make check output: http://buildd.debian.org/fetch.php?pkg=kaffever=2%3A1.1.6-3arch=alphastamp=1128345978file=logas=raw intrp jthreads: 2 of 148 tests failed intrp: pthreads: 4 of 148 tests failed cheers, dalibor topic ___ kaffe mailing list kaffe@kaffe.org http://kaffe.org/cgi-bin/mailman/listinfo/kaffe
Re: [kaffe] Does Kaffe's JIT support ARM?
Kiyo Inaba wrote: OK, now I came back to Tokyo, my home city. Because of the classpath merge, latest weekly regression does not work at all for m68k/netbsd1.6, and for the time being I can not invoke weekly regression for m68k/linux (I ordered IIci logic boards which are supposed to come this week). So the latest weekly regression test result I have for m68k ports is m68k/netbsd1.6 at 2005-09-29. And the results are jit3: 26 of 148 tests failed jit: 32 of 148 tests failed intrp: 22 of 148 tests failed Thanks for the update, Kiyo. Since NetBSD-1.6 uses gcc-2.95.3, and jikes-1.19 and later can not be compiled with this compiler, I first have to try compile newer gcc, or make some other machines with newer NetBSD :- Unfortunately, gcjx (from the gcjx-branch in gcc's subversion repository) from Tom Tromey also demands a pretty recent gcc. (gcc4, I believe). cheers, dalibor topic ___ kaffe mailing list kaffe@kaffe.org http://kaffe.org/cgi-bin/mailman/listinfo/kaffe
Re: [kaffe] Problem to locate libnative.a
[EMAIL PROTECTED] wrote: Hello people, Hallo Daniel, now loading.. /libnative.a user_search_path:(null) libnative.a LTDL_SEARCHPATH_VAR:(null) libnative.a sys_search_path:/lib:/usr/lib libnative.a xtryall_dlopen /libnative.a tryall_dlopen loader:366a2 /libnative.a sys-name:libnative.a Failed to locate native library libnative in path / Aborting. If you are trying to dlopen static libraries, make sure the libnative.la file is there as well. If that does not work, you can try stepping through the library loading code in a debugger, the code for loading native libraries in Kaffe is located in loadNativeLibrary in file kaffe/kaffevm/external.c the actual library loading is performed by libtool, the code for that is in the ltdl subdirectory of the source code. cheers, dalibor topic ___ kaffe mailing list kaffe@kaffe.org http://kaffe.org/cgi-bin/mailman/listinfo/kaffe
[kaffe] paint and repaint ...
sorry for the probably dumb question: I have a small example that draws a rect in a paint method of the canvas. based on an event I redraw the rect with a different color (with a call to fillRect). Then i call repaint with the including modified area to avoid flickering while updating. The update is correct with java from j2sdk1.4.2_06, while in kaffe 1.1.6 with Tiny-X on a MIPS machine, the windows is cleared and only the modified rectangle is drawn: seems that repaint clear all the window and repaint only the requested rectangle: should not it clear only the requested rectangle? thanks giammy -- Gianluca Moro http://groups.yahoo.com/group/embeddeditalia/ ISCRIVITI alla Mailing List Italiana su LINUX EMBEDDED [EMAIL PROTECTED] Visit http://ilpinguino.altervista.org/ ___ Yahoo! Messenger: chiamate gratuite in tutto il mondo http://it.messenger.yahoo.com ___ kaffe mailing list kaffe@kaffe.org http://kaffe.org/cgi-bin/mailman/listinfo/kaffe
[kaffe] Question about function 'sysdepCallMethod'
Hi all, Please advise if 'sysdepCallMethod' can be implemented by this way as bellow: method 1: in JVM: typedef int (*FUNCTYPE)(); //function's return type is int FUNCTYPE *nativefunc = (FUNCTYPE)dlsym(handle, foo); //four argument (*nativefunc)(1, 2.5); in share library: int foo(int, double); method 2: in JVM: typedef int (*FUNCTYPE)(...); //function's return type is int, argument list is '...' FUNCTYPE *nativefunc = (FUNCTYPE)dlsym(handle, foo); //four argument (*nativefunc)(1, 2.5); in share library: int foo(int, double); _ 与联机的朋友进行交流,请使用 MSN Messenger: http://messenger.msn.com/cn ___ kaffe mailing list kaffe@kaffe.org http://kaffe.org/cgi-bin/mailman/listinfo/kaffe