Things are a little ugly in fc6 extras java land...
* Azureus makes my computer practically unusable because of libgcj's new SecureRandom implementation. I've filed a bug for a work around here: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=200653 Mark tells me this is the root cause of some crypto test timeout problems on the autobuilder as well. Casey says that this same code doesn't seem to have a problem on jamvm.
* RSSOwl no longer fails to start because of the zip problem (thanks tromey!), but now crashes in _Jv_ClassReader::checkExtends (super = 0x3). I'm rebuilding libgcj with -O0 -g3 to debug.
* I've updated jogl to jsr231 1.0 beta 5. The good news is that it builds sans patching! The bad news is that demos crash the X server (like, try resizing the gears demo). I think this is an X problem, so I'm going to push this out anyways because the latest WW2D (and "WW2D plus one", which looks awesome -- http://www.alpix.com/3d/worldwin/WW2d_Java.html) use this version of the jogl API.
* kawa is fine.
AG
On Sun, 2006-07-30 at 11:25 -0700, Anthony Green wrote:
- Azureus makes my computer practically unusable because of libgcj's new
SecureRandom implementation. I've filed a bug for a work around here: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=200653 Mark tells me this is the root cause of some crypto test timeout problems on the autobuilder as well. Casey says that this same code doesn't seem to have a problem on jamvm.
It does seem to be not very efficient on my version of jamvm. I just committed the following to classpath which works pretty nicely:
2006-07-30 Mark Wielaard mark@klomp.org
* resource/java/security/classpath.security: Add /dev/urandom as default securerandom.source.
This will be part of 0.92.
Cheers,
Mark
diff -u -r1.2.2.4 classpath.security --- resource/java/security/classpath.security 23 Jul 2006 21:28:36 -0000 1.2.2.4 +++ resource/java/security/classpath.security 30 Jul 2006 21:39:58 -0000 @@ -45,3 +45,7 @@ # The VM-wide default callback handler class name. MUST be a subclass of # javax.security.auth.callback.CallbackHandler auth.login.defaultCallbackHandler=gnu.javax.security.auth.callback.DefaultCallbackHandler + +# If this file isn't found we fall back to generating entropy through +# "battling threads". +securerandom.source=file:///dev/urandom
On Sun, 2006-07-30 at 23:44 +0200, Mark Wielaard wrote:
It does seem to be not very efficient on my version of jamvm. I just committed the following to classpath which works pretty nicely:
2006-07-30 Mark Wielaard mark@klomp.org
* resource/java/security/classpath.security: Add /dev/urandom as default securerandom.source.This will be part of 0.92.
rebuild-security-providers from jpackage-utils writes over this file. We need to hack this tool for fedora.
AG
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Anthony Green wrote:
rebuild-security-providers from jpackage-utils writes over this file. We need to hack this tool for fedora.
AG
Actually, it keeps the original file. So all that needs to be done is to update the initial classpath.security file to something more recent.
The one from 0.92 would be usable if the security providers are also getting backported for fc6, or in any case, some part of it is usable.
- -- Sincerely,
David Walluck david@zarb.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
David Walluck wrote:
The one from 0.92 would be usable if the security providers are also getting backported for fc6, or in any case, some part of it is usable.
Actually, since rebuild-security-providers replaces the providers only, it doesn't matter what providers are in the initial classpath.security, but the names have changed and number of providers have increased since the version of classpath in gcc 4.1.x.
- -- Sincerely,
David Walluck david@zarb.org
"Mark" == Mark Wielaard mark@klomp.org writes:
Mark> 2006-07-30 Mark Wielaard mark@klomp.org Mark> * resource/java/security/classpath.security: Add /dev/urandom as Mark> default securerandom.source. Mark> This will be part of 0.92.
This should probably go in GCC svn as soon as possible. Also we should pull it into the FC6 RPM.
Tom
On Sun, 2006-07-30 at 11:25 -0700, Anthony Green wrote:
- RSSOwl no longer fails to start because of the zip problem (thanks
tromey!), but now crashes in _Jv_ClassReader::checkExtends (super = 0x3). I'm rebuilding libgcj with -O0 -g3 to debug.
This has beaten me. It looks like a hairy class loading bug. Anybody interested in having a look? If so, I'll have to figure out the best way to give you the bits to test with.
AG
Anthony Green writes:
On Sun, 2006-07-30 at 11:25 -0700, Anthony Green wrote:
- RSSOwl no longer fails to start because of the zip problem (thanks
tromey!), but now crashes in _Jv_ClassReader::checkExtends (super = 0x3). I'm rebuilding libgcj with -O0 -g3 to debug.
This has beaten me. It looks like a hairy class loading bug. Anybody interested in having a look?
Yes, definitely. Hairy class loading bugs are the spice of my life.
Andrew.
On Mon, 2006-07-31 at 09:54 +0100, Andrew Haley wrote:
This has beaten me. It looks like a hairy class loading bug. Anybody interested in having a look?
Yes, definitely. Hairy class loading bugs are the spice of my life.
Thanks!
http://people.redhat.com/~green/FE/devel/rssowl-1.2.1-2.src.rpm
AG
Anthony Green writes:
On Mon, 2006-07-31 at 09:54 +0100, Andrew Haley wrote:
This has beaten me. It looks like a hairy class loading bug. Anybody interested in having a look?
Yes, definitely. Hairy class loading bugs are the spice of my life.
Thanks!
http://people.redhat.com/~green/FE/devel/rssowl-1.2.1-2.src.rpm
The real exception is
java.lang.ClassNotFoundException@2aaab4c55a50 { java.lang.Throwable ex = null, java.lang.VMThrowable vmState = 0x2aaab28a8530, [Ljava.lang.StackTraceElement; stackTrace = null, java.lang.Throwable cause = 0x2aaab4c55a50, java.lang.String detailMessage = "org.eclipse.core.commands.common.EventManager not found in gnu.gcj.runtime.SystemClassLoader{urls=[file:/usr/share/java/rssowl.jar,file:/usr/share/java/xerces-j2.jar,file:/usr/share/java/itext.jar,file:/usr/share/eclipse/plugins/org.eclipse.core.runtime_3.2.0.v20060603.jar,file:/usr/share/eclipse/plugins/org.eclipse.ui.forms_3.2.0.v20060602.jar,file:/usr/share/java/swt-gtk-3.2.jar,file:/usr/share/java/commons-logging.jar,file:/usr/share/eclipse/plugins/org.eclipse.jface_3.2.0.I20060605-1400.jar,file:/usr/share/java/jdom.jar,file:/usr/share/java/commons-httpclient.jar,file:/usr/share/java/commons-codec.jar,file:/usr/share/java/glib0.2.jar,file:/usr/share/java/gconf2.12.jar,file:/usr/share/java/gtk2.8.jar,file:/usr/share/java/], parent=gnu.gcj.runtime.ExtensionClassLoader{urls=[], parent=null}}", }
org.eclipse.core.commands.common.EventManager is in /usr/share/eclipse/plugins/org.eclipse.core.commands_3.2.0.I20060605-1400.jar
Yous RSSowl start line should look like
exec java -Djava.library.path=/usr/lib64/ -cp /usr/share/java/rssowl.jar:/usr/share/java/xerces-j2.jar:/usr/share/java/itext.jar:/usr/share/eclipse/plugins/org.eclipse.core.runtime_3.2.0.v20060603.jar:/usr/share/eclipse/plugins/org.eclipse.core.commands_3.2.0.I20060605-1400.jar:/usr/share/eclipse/plugins/org.eclipse.ui.forms_3.2.0.v20060602.jar:/usr/share/java/swt-gtk-3.2.jar:/usr/share/java/commons-logging.jar:/usr/share/eclipse/plugins/org.eclipse.jface_3.2.0.I20060605-1400.jar:/usr/share/eclipse/plugins/com.ibm.icu_3.4.4.1.jar:/usr/share/eclipse/plugins/org.eclipse.equinox.common_3.2.0.v20060603.jar:/usr/share/java/jdom.jar:/usr/share/java/commons-httpclient.jar:/usr/share/java/commons-codec.jar:/usr/share/java/glib0.2.jar:/usr/share/java/gconf2.12.jar:/usr/share/java/gtk2.8.jar:/usr/share/java/ net.sourceforge.rssowl.controller.RSSOwlLoader "$@"
Anthony Green writes:
On Mon, 2006-07-31 at 18:20 +0100, Andrew Haley wrote:
Yous RSSowl start line should look like
Awesome, thanks. Did you see a NullPointer exception, like I did?
Yes.
In any case, rssowl appears to be working well now (modulo old feed finding bug - probably a regexp issue).
Good.
Andrew.
Anthony> Things are a little ugly in fc6 extras java land... [...]
Anthony> * I've updated jogl to jsr231 1.0 beta 5. The good news is that it Anthony> builds sans patching! The bad news is that demos crash the X server Anthony> (like, try resizing the gears demo).
Is there a PR for this one?
Tom
java-devel@lists.fedoraproject.org
-
Andrew Haley -
Anthony Green -
David Walluck -
Mark Wielaard -
Tom Tromey