[kaffe] Odd behavior of jawt in cvs head

[Dalibor Topic]

jserv at linux2.cc.ntu.edu.tw wrote:
> Hi list,
> 
>   I just copy the JAWT example from GNU Classpath cvs repository, and
> I am about to have a test. After packaging[1], I find some odd behavior
> with cvs head of Kaffe.

AWT has been broken on Classpath head these days in interesting ways,
afaik, so things may not be as stable as they seem ;) Chances are it may
get better with the next checkins and resyncing round.

>   The library loading message will be output to jawt-test.log, and I
> found some entries like:
> 
> Error message is /home/jserv/jhacker/kaffe/test/awt/jawt/libDemoJAWT.so:
> undefined symbol: gnu_java_awt_peer_gtk_GdkFontPeer_initState.
> Error message is /home/jserv/jhacker/kaffe/test/awt/jawt/libDemoJAWT.so:
> undefined symbol: gnu_java_awt_peer_gtk_GdkFontPeer_initStaticState.
> Error message is /home/jserv/jhacker/kaffe/test/awt/jawt/libDemoJAWT.so:
> undefined symbol: gnu_java_awt_peer_gtk_GdkFontPeer_setFont.
> Error message is /home/jserv/jhacker/kaffe/test/awt/jawt/libDemoJAWT.so:
> undefined symbol: gnu_java_awt_peer_gtk_GdkGraphics_dispose.
> 
>   But it should not happen since libjawt.so was linked to libgtkpeer.so,
> which contains the above symbols.

Yeah, those entries are ok, if you looks through the logs, later the
Java_* symbols get resolved.
> 
>   Any idea?

Other than firing up gdb ... nope ;) i've looked at the output of
-vmdebug NATIVE, and I can get the library loaded and all that, but I
can't get jawt to do anything save to give me a frame.

cheers,
dalibor topic