[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Patches 1059 and 1094
From: |
Brian Jones |
Subject: |
Re: Patches 1059 and 1094 |
Date: |
14 Feb 2003 18:08:39 -0500 |
User-agent: |
Gnus/5.09 (Gnus v5.9.0) Emacs/21.2 |
Julian Dolby <address@hidden> writes:
> Those of us at IBM at planning to cut a new release of Jikes RVM fairly
> soon, and we would like to make some claims about what the combination of
> it and the GNU Classpath libraries can run. Is there any chance that my
> recent patches (1059 and 1094) will be merged soon? If they are merged we
> can make some nice statements about running both Eclipse and some AWT code.
Patch 1059 can probably be applied since it does nothing.
I saw Tom make a post/reply regarding 1094, did you respond to that?
Brian
--
Brian Jones <address@hidden>
>>>>> "Julian" == Julian Dolby <address@hidden> writes:
Julian> I found the gthread-jni files, and I basically filled in all
Julian> the functions in gthread-jni.c to use JNI to talk to
Julian> Java-level synchronization mechanisms. I also passed the
Julian> magic structure of those function pointers to g_thread_init in
Julian> GtkMainThread.gtkInit. This code should work for any VM;
Julian> nonethelessI put the changes to use this stuff under a
Julian> --enable-portable-native-sync option to configure, so that VMs
Julian> that do not support all the JNI I used will not be bitten by
Julian> it.
I think it would make sense for us to simply require that a JVM
implement these JNI methods correctly. Is there really a JVM that
Classpath can run on that doesn't? (I assume they are monitor
enter/exit calls...)
Tom