You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@geronimo.apache.org by Joe O'Pecko <op...@gmail.com> on 2006/10/26 15:47:13 UTC
Geronimo 1.1.1 not respecting WAR MANIFEST.MF for GBean
Hi,
I have an existing GBean packaged in an external jar file within an
EAR file, running on a Geronimo 1.0 server that refuses to load in
Geronimo 1.1.1. The simplified structure of the EAR is as follows:
example.ear
| - geronimo-application.xml (gbean defined herein)
| - gbean.jar
| - web-app.war
| - META-INF
| - MANIFEST.MF (Class-Path: gbean.jar)
This deploys in a 1.0 container and starts the GBean. However, in
1.1.1, I receive a DeploymentException caused by a
ClassNotFoundException. It appears to me that Geronimo 1.1.1 does not
respect the WAR Manifest CLASSPATH, WRT GBeans, in the same fashion as
1.0. Is this a regression in 1.1.1, or was this a "feature" of 1.0
that has since been corrected?
Thanks for any information you can provide.
Joe
Re: Geronimo 1.1.1 not respecting WAR MANIFEST.MF for GBean
Posted by Aaron Mulder <am...@alumni.princeton.edu>.
The problem seems to be that the GBean is defined at the EAR level but
the classes are provided at the WAR level. These are two different
class loaders. It should work, for example, if you put the GBean in
geronimo-web.xml in the WAR rather than geronimo-application.xml in
the EAR.
Thanks,
Aaron
On 10/26/06, Joe O'Pecko <op...@gmail.com> wrote:
> Hi,
>
> I have an existing GBean packaged in an external jar file within an
> EAR file, running on a Geronimo 1.0 server that refuses to load in
> Geronimo 1.1.1. The simplified structure of the EAR is as follows:
>
> example.ear
> | - geronimo-application.xml (gbean defined herein)
> | - gbean.jar
> | - web-app.war
> | - META-INF
> | - MANIFEST.MF (Class-Path: gbean.jar)
>
> This deploys in a 1.0 container and starts the GBean. However, in
> 1.1.1, I receive a DeploymentException caused by a
> ClassNotFoundException. It appears to me that Geronimo 1.1.1 does not
> respect the WAR Manifest CLASSPATH, WRT GBeans, in the same fashion as
> 1.0. Is this a regression in 1.1.1, or was this a "feature" of 1.0
> that has since been corrected?
>
> Thanks for any information you can provide.
>
> Joe
>