Well, I figured something out.  I was then working on a simple example
to show to Stuart and using pax runner and couldn't reproduce the
problem.  After trying all different combinations to make the 2 have
the exact same configuration, the last (desperate) thing I tried was
replacing the Felix OSGi compendium bundle (downloaded from maven
<http://repo2.maven.org/maven2/org/apache/felix/org.osgi.compendium/1.2.0/>)
with the regular OSGi compendium bundle (also from ops4j
<http://repository.ops4j.org/maven2/org/osgi/org.osgi.compendium/4.1.0/>)
it started working.

I have no idea why that makes a lick of difference, but I'm happy to
say everything is working well now.

Thanks,
Rich

Fri, May 1, 2009 at 11:00 AM, Stuart McCulloch <[email protected]> wrote:
> 2009/5/2 Richard Wallace <[email protected]>
>>
>> On Thu, Apr 30, 2009 at 11:32 PM, Niclas Hedhman <[email protected]>
>> wrote:
>> > On Fri, May 1, 2009 at 9:52 AM, Richard Wallace
>> > <[email protected]> wrote:
>> >> Hey all,
>> >>
>> >> I'm running into a problem where I've got the
>> >> osgi_R4_compendium-1.0.jar and the Pax Web bundle jar load in Felix.
>> >> I'm running into a problem where one bundle is using the HttpService
>> >> from one and another (Peaberry) is trying to use the one from the
>> >> other, causing a bunch of problems.  Any reason the HttpService can't
>> >> be taken out of the Pax Web bundles?
>> >
>> > Exactly what is the problem, because OSGi is supposed to handle
>> > exactly(!) this kind of situation and ONLY choose one of the exporters
>> > of the packages, and consolidate everything under a single
>> > classloader.
>> >
>> > Is it a version problem, or is it missing Import/Export-Package of
>> > something in Pax Web?
>> >
>>
>> I was mistaken, I think.  I was using Peaberry to do the service
>> imports and getting a "NoClassDefFound: javax/servlet/Servlet" when it
>> tried to import the HttpService.  But I switched to using straight
>> ServiceTrackers and everything seems to be working now.  I'll follow
>> up on the Peaberry list with Stuart.
>
> interesting, we use the same underlying OSGi APIs (getServiceReferences +
> addServiceListener)
> so it should behave the same wrt. assignable services - perhaps there's an
> additional check done
> somewhere in the tracker on the instance that's not done by the raw APIs...
>
> if you can boil this down to a simple testcase that would be great :)
>
>> Thanks,
>> Rich
>>
>> >
>> > Cheers
>> > --
>> > Niclas Hedhman, Software Developer
>> > http://www.qi4j.org - New Energy for Java
>> >
>> > I  live here; http://tinyurl.com/2qq9er
>> > I  work here; http://tinyurl.com/2ymelc
>> > I relax here; http://tinyurl.com/2cgsug
>> >
>> > _______________________________________________
>> > general mailing list
>> > [email protected]
>> > http://lists.ops4j.org/mailman/listinfo/general
>> >
>>
>> _______________________________________________
>> general mailing list
>> [email protected]
>> http://lists.ops4j.org/mailman/listinfo/general
>
>
>
> --
> Cheers, Stuart
>
> _______________________________________________
> general mailing list
> [email protected]
> http://lists.ops4j.org/mailman/listinfo/general
>
>

_______________________________________________
general mailing list
[email protected]
http://lists.ops4j.org/mailman/listinfo/general

Reply via email to