This is the mail archive of the
java@gcc.gnu.org
mailing list for the Java project.
Re: OpenOffice and gcj at runtime
- From: Caolan McNamara <caolanm at redhat dot com>
- To: java at gcc dot gnu dot org
- Date: Wed, 09 Feb 2005 12:07:06 +0000
- Subject: Re: OpenOffice and gcj at runtime
- References: <1102435824.8012.23.camel@sheol.homelinux.org> <m3eki2cbii.fsf@localhost.localdomain> <1102586332.8525.18.camel@sheol.homelinux.org>
On Thu, 2004-12-09 at 09:58 +0000, Caolan McNamara wrote:
> Yeah, I've whipped up a better dlopening example which is a fairer
> reflection of what OOo wants to do
> http://people.redhat.com/caolanm/gcj.problems/runtime2.tar.gz
> and with it I can see that libgcj.so.4 is ok, while libgcj.so.5
> hangs/crashes on the second CallObjectMethodA, and libgcj.so.6 silently
> fails on the second CallObjectMethodA. Forcing the module which causes
> grief in OOo to use libgcj.so.4 at runtime works, so its apparently down
> to the same type of problem which had slightly different results between
> libgcj.so.5/libgcj.so.6.
This all works fine now with gcj (thanks). So openoffice.org works
relatively ok. The only thing is that I have a very similiar error under
ppc trying to register java components which work under x86, I have no
problems with the build-time java tools, just with calling java
components through the OOo uno c++ bridge. Now the original testcase
itself works fine, and I've been unable to date to extract a new
testcase that pinpoints the ppc issue so it might be something on my
side, but are there known difficulties with libgcj under ppc in this
arena which don't arise under ix86 ?
C.