harmony-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Evgueni Brevnov (JIRA)" <j...@apache.org>
Subject [jira] Updated: (HARMONY-1582) [drlvm] Implementation of Invocation API for DRLVM
Date Mon, 02 Oct 2006 14:21:22 GMT
     [ http://issues.apache.org/jira/browse/HARMONY-1582?page=all ]

Evgueni Brevnov updated HARMONY-1582:

    Attachment: invocation_api.3.patch

I have fixed several places according to discussions on the list. Namely

1) move  "daemon" attribute from native to java layer.
2) return  jthread_create signature back
3) remove hythread_create_ex
4) fix unexpected bug in encoder
5) split patches for vm and build modules.

> [drlvm] Implementation of Invocation API for DRLVM
> --------------------------------------------------
>                 Key: HARMONY-1582
>                 URL: http://issues.apache.org/jira/browse/HARMONY-1582
>             Project: Harmony
>          Issue Type: Improvement
>          Components: DRLVM
>         Environment: Windows-ia32, Linux-ia32
>            Reporter: Evgueni Brevnov
>         Assigned To: Geir Magnusson Jr
>         Attachments: invocation_api.2.patch, invocation_api.3.patch, invocation_api.patch,
> Hi,
> I've finished implementation of Invocation API for DRLVM. Eclipse runs OK. Smoke and
kernel tests passes on Windows and Linux (with some exceptions of course :-)). I have some
expected failures and some new one in kernel test java.lang.Thread on Windows and java.lang.ThreadGroup
on Linux. Actually these tests worked fine before this Monday (09/25). I think the last patch
 for condition vars (HARMONY-1519?) introduced these failures. Also I had to switch cunit
tests off because of linking problems. Hopefully original authors of cunit tests can fix the
problem in the right way.
> One more workaround is disabling warnings on Linux. We need to find a way to get rid
of "warning: invalid access to non-static data member `JavaVM_Internal::link' of NULL object".
It happenes because of APR_OFFSETOF. I believe we can resolve this together.
> There are two restrictions in the current implementation:
> 1) Only on VM can be created in the process address space. Actualy it's not a problem
of Invocation API implementation rather than VM design itself. Additional work required to
have this feature.
> 2) DestroyJavaVM doesn't clean its internals properly. It kills the process instead.
It is not a trivial task to implement it properly. I suggest do it separetly (not in this
> So waiting for your comments......and quick integration :-)
> Thanks
> Evgueni

This message is automatically generated by JIRA.
If you think it was sent incorrectly contact one of the administrators: http://issues.apache.org/jira/secure/Administrators.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira


View raw message