Re: SableVM on ARM Zaurus
Cortez The Killer <[email protected]>
| Newsgroups | gmane.comp.java.vm.sablevm.general |
|---|---|
| Message-ID | <[email protected]> |
> > > > I hope you can see something special, because I don't ;-) > > Kind of. I grepped for "libgtkpeer" and it looks like the library is > properly found, opened, read etc... Though I am not quite sure why > do I see a the end SableVM being "killed by SIGINT" ! After the stack trace is shown, the bottom lines says "...1 more". I have to admit that I don't know how the rest of the stack trace can be displayed, so I exit forcively by pressing Ctrl-C. I guess that's why you see the "killed by SIGINT" message. > Can you add --enable-debugging-features to 'sablevm's ./configure, > remove any --with-threading=... option, and recompile reinstall it? > (no need to touch sablevm-classpath or any other software) Using this configure parameter, make exits with following error: In file included from libsablevm.c:18: /usr/java/include/ffi.h:118:6: use of C99 long long integer constant make[5]: ***[labsablevm.lo] Error 1 ... Any ideas? > Do you get any additional information (ie. a failed assertion) when > ran normally? > > There's also other things I have not asked you about yet, i.e. > does the app you're trying to run on ARM run when you use SableVM > on an x86? Good question, I haven't tried that yet, mostly because the app is so simple. I'll come back to you about that. > I also went thru all the code involved in initiating gtkpeers > and then loading a library, up to sablevm native method > Java_java_lang_VMRuntime_nativeLoad in > src/libsablevm/java_lang_VMRuntime.c > > I do not know if you have gdb running reliably on your Taurus > (I remember there were problems with it sometimes), so if I were > you, I'd put in the above function a few 'printf(); fflush(NULL);' > statments that would show with what parameters its called, > what is the control flow inside and where exactly it fails! I'll try that too. > For now I suspect a libtool problem that we might need to work > around (though the strange thing is that the .so library actually > is found, opened and read). > > It might seem a vague advice, but you should not give up. It's > probable we're very near finding the exact problem. Just a little > more patience :-) Again, I'm glad you helped me out so far, so I'm patient, no problem. It would be great to have full Java (including Swing) on the Zaurus. There are a few VM's, but most of them use Qt/E, and none of them (Blackdown, Sun, IBM) support X11.