fc6 extras java status

[Anthony Green]

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

[Mark Wielaard]

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(a)klomp.org&gt;

    * 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

[Anthony Green]

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(a)klomp.org&gt;
>  
>      * 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

[David Walluck]

-----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(a)zarb.org&gt;
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.3 (GNU/Linux)
> Comment: Using GnuPG with Mandriva - http://enigmail.mozdev.org
>
> iD8DBQFEzVA8N5thZBYlTwkRAoI3AJsEwQx/bYjfajoGIQGm+NaasruHUwCgmZPk
> uDofsaUSn1afs8JhK21M8Qg=
> =Vy0g
> -----END PGP SIGNATURE-----

[David Walluck]

-----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(a)zarb.org&gt;
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.3 (GNU/Linux)
> Comment: Using GnuPG with Mandriva - http://enigmail.mozdev.org
>
> iD8DBQFEzVG6N5thZBYlTwkRAnplAJ9xG1Fl4eQHGIOkM1E8Kxt+wW1wCQCfUtS2
> d3OElQ6P3nfTuYo0znJkCfY=
> =Y30w
> -----END PGP SIGNATURE-----

[Tom Tromey]

> >>>> "Mark" == Mark Wielaard
> <mark(a)klomp.org&gt; writes: 
Mark> 2006-07-30  Mark Wielaard  <mark(a)klomp.org&gt;
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

[Anthony Green]

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

[Andrew Haley]

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.

[Anthony Green]

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

[Andrew Haley]

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]

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?

In any case, rssowl appears to be working well now (modulo old feed
finding bug - probably a regexp issue).

AG

[Andrew Haley]

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.

[Tom Tromey]

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