[
https://issues.apache.org/jira/browse/FELIX-624?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Carsten Ziegeler closed FELIX-624.
--
Resolution: Fixed
Fix Version/s: scr-1.0.4
> Wrong classloader is used to detect bind/unb
[
https://issues.apache.org/jira/browse/FELIX-620?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12611436#action_12611436
]
Sahoo commented on FELIX-620:
-
Thanks for fixing the bug so quickly. When is the next version o
You would have to do the following around all calls into the framework:
ClassLoader tccl = Thread.currentThread().getContextClassLoader();
try
{
Thread.currentThread().setContextClassLoader(null);
// call into the framework
}
finally
{
Thread.currentThread().setContextClassLoader(tccl);
}
Hi,
- I can't use Equinox because I discovered a Null Pointer Exception when
attempting to restart it (in the same "context" loosely speaking -- embedded);
Also, equinox as embedded is not really "instance" based out of the box; I
logged a bug on the NPE and it's probably fixed by now, but it'
On Tue, Jul 8, 2008 at 4:40 AM, Karl Pauls <[EMAIL PROTECTED]> wrote:
>> So, depending on how you are setting this up, you need to do
>> "something", like either replicate the ContextFinder approach, or
>> plainly set the TCL to null prior to loading up Felix.
>
> Is that something pax logging can
On Mon, Jul 7, 2008 at 10:27 PM, Niclas Hedhman <[EMAIL PROTECTED]> wrote:
> On Tue, Jul 8, 2008 at 3:26 AM, Niclas Hedhman <[EMAIL PROTECTED]> wrote:
>> On Tue, Jul 8, 2008 at 1:31 AM, Karl Pauls <[EMAIL PROTECTED]> wrote:
>>> Is it somehow possible that log4j is
>>> doing some classloader magic i
On Tue, Jul 8, 2008 at 3:26 AM, Niclas Hedhman <[EMAIL PROTECTED]> wrote:
> On Tue, Jul 8, 2008 at 1:31 AM, Karl Pauls <[EMAIL PROTECTED]> wrote:
>> Is it somehow possible that log4j is
>> doing some classloader magic itself and maybe is using the context
>> classloader? The context classloader wou
On Tue, Jul 8, 2008 at 1:31 AM, Karl Pauls <[EMAIL PROTECTED]> wrote:
> Is it somehow possible that log4j is
> doing some classloader magic itself and maybe is using the context
> classloader? The context classloader would be the WebAppClassloader I
> think ...
Good Suggestion. Will investigate th
On Mon, Jul 7, 2008 at 7:17 PM, Niclas Hedhman <[EMAIL PROTECTED]> wrote:
> On Tue, Jul 8, 2008 at 12:22 AM, Stuart McCulloch
> <[EMAIL PROTECTED]> wrote:
>
>> again, this wouldn't solve the Pax-Logging issue because
>> then the application will have all it's logging calls bound to
>> the original
On Tue, Jul 8, 2008 at 12:22 AM, Stuart McCulloch
<[EMAIL PROTECTED]> wrote:
> again, this wouldn't solve the Pax-Logging issue because
> then the application will have all it's logging calls bound to
> the original Log4J jar, which lacks the OSGi support that
> Pax-Logging provides (various other
2008/7/8 Sahoo <[EMAIL PROTECTED]>:
> Karl Pauls wrote:
>
>> Since Felix is started by the web app, Felix would use that web app's
>>> class
>>> loader for any delegation. By having this property, both OSGi bundles
>>> should
>>> see same log4j classes as the web app.
>>>
>>>
>>
>> Right. I'll try
Karl Pauls wrote:
Since Felix is started by the web app, Felix would use that web app's class
loader for any delegation. By having this property, both OSGi bundles should
see same log4j classes as the web app.
Right. I'll try to investigate this a bit more.
Since the parent class loader
On Mon, Jul 7, 2008 at 11:24 PM, Stuart McCulloch
<[EMAIL PROTECTED]> wrote:
> however, setting this would break Pax-Logging, because then the application
> would load the Log4J public API classes from the standard Log4J jar - rather
> than the custom classes from Pax-Logging that fix Log4J logging
2008/7/7 Sahoo <[EMAIL PROTECTED]>:
> Karl Pauls wrote:
>
>> On Mon, Jul 7, 2008 at 4:08 PM, Sahoo <[EMAIL PROTECTED]> wrote:
>>
>>
>>> log4j:ERROR A "org.apache.log4j.ConsoleAppender" object is not
assignable to a "
org.apache.log4j.Appender" variable.
log4j:ERROR The cla
On Mon, Jul 7, 2008 at 4:50 PM, Sahoo <[EMAIL PROTECTED]> wrote:
> Karl Pauls wrote:
>>
>> On Mon, Jul 7, 2008 at 4:08 PM, Sahoo <[EMAIL PROTECTED]> wrote:
>>
log4j:ERROR A "org.apache.log4j.ConsoleAppender" object is not
assignable to a "
org.apache.log4j.Appender" variabl
Stuart McCulloch wrote:
looking at the current code that trawls the super-classes, I believe the
problem could be solved by changing 'DependencyManager.java:734'
so a ClassNotFoundException will drop down to the exhaustive search
of declared methods (this code would need to be refactored out, so
[
https://issues.apache.org/jira/browse/FELIX-624?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Carsten Ziegeler updated FELIX-624:
---
Summary: Wrong classloader is used to detect bind/unbind methods (was:
Import statement incom
Karl Pauls wrote:
On Mon, Jul 7, 2008 at 4:08 PM, Sahoo <[EMAIL PROTECTED]> wrote:
log4j:ERROR A "org.apache.log4j.ConsoleAppender" object is not
assignable to a "
org.apache.log4j.Appender" variable.
log4j:ERROR The class "org.apache.log4j.Appender" was loaded by
log4j:ERROR [7.0] whereas
[
https://issues.apache.org/jira/browse/FELIX-625?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Clement Escoffier resolved FELIX-625.
-
Resolution: Fixed
Add the onTimeout support to the temporal dependencies (revision 674506)
Karl,
I can send you a zip in a few...
Re: strange... yeah... I want to believe that I should not have any
class loader issues in OSGi in that I simply rely on the manifests and
then "world is good"... Of course, I rely on the BND tool, which I
couldn't live without, incidentally (I avoid maven l
[
https://issues.apache.org/jira/browse/FELIX-625?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Clement Escoffier updated FELIX-625:
Summary: Temporal Dependency improvements (was: Temporal Dependency
improvement)
> Tempora
Temporal Dependency improvement
---
Key: FELIX-625
URL: https://issues.apache.org/jira/browse/FELIX-625
Project: Felix
Issue Type: Improvement
Components: iPOJO
Affects Versions: iPOJO-0.8.0
Hi,
This seemed to work, sort of:
> org.osgi.framework.bootdelegation=sun.*,com.sun.*, org.apache.log4j.*
There was a class cast exception, as follows:
> ERROR: EventDispatcher: Error during dispatch.
(java.lang.ClassCastException:
org.ops4j.pax.logging.internal.PaxLoggingServiceImpl$1ManagedPa
Richard S. Hall wrote:
It probably makes sense to change this in some fashion. We have the same
issue inside the framework when we are trying to figure out how if a
class is visible to a given bundle (in Util.loadClassUsingClass()). So,
it makes sense that SCR might have to do something simil
2008/7/7 Carsten Ziegeler <[EMAIL PROTECTED]>:
> Stuart McCulloch wrote:
>
>>
>> imho SCR should be using the classloader of the class that declared the
>> bind/unbind method, ie:
>>
>> method.getDeclaringClass().getClassLoader()
>>
>> when loading parameter classes - always using B's classloade
On Mon, Jul 7, 2008 at 4:08 PM, Sahoo <[EMAIL PROTECTED]> wrote:
>
>> log4j:ERROR A "org.apache.log4j.ConsoleAppender" object is not
>> assignable to a "
>>
>> org.apache.log4j.Appender" variable.
>>
>> log4j:ERROR The class "org.apache.log4j.Appender" was loaded by
>>
>> log4j:ERROR [7.0] whereas
Carsten Ziegeler wrote:
Stuart McCulloch wrote:
imho SCR should be using the classloader of the class that declared the
bind/unbind method, ie:
method.getDeclaringClass().getClassLoader()
when loading parameter classes - always using B's classloader doesn't
sound
right to me
Yes, I agre
Strange. Somehow a class is loaded from the outside and then used
inside a bundle...
Are you maybe bridging something from the outside to the inside (maybe
via the list parameter in the felix constructor)?
Can you make an example that fails available to me (either send it to
me directly or create
log4j:ERROR A "org.apache.log4j.ConsoleAppender" object is not
assignable to a "
org.apache.log4j.Appender" variable.
log4j:ERROR The class "org.apache.log4j.Appender" was loaded by
log4j:ERROR [7.0] whereas object of type
log4j:ERROR "org.apache.log4j.ConsoleAppender" was loaded by
[WebappC
Stuart McCulloch wrote:
imho SCR should be using the classloader of the class that declared the
bind/unbind method, ie:
method.getDeclaringClass().getClassLoader()
when loading parameter classes - always using B's classloader doesn't sound
right to me
Yes, I agree - now, the current implem
[
https://issues.apache.org/jira/browse/FELIX-620?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Stuart McCulloch resolved FELIX-620.
Resolution: Fixed
Fixed in trunk - manifest goal now follows export instructions from includ
[
https://issues.apache.org/jira/browse/FELIX-620?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Stuart McCulloch reassigned FELIX-620:
--
Assignee: Stuart McCulloch
> manifest goal of maven-bundle-plugin ignores _include instr
Clement Escoffier a écrit :
Hi,
Providing "automatic" OBR bundle descriptions is on the iPOJO roadmap (and
is currently under development). It should describe service requirements /
capabilities as well as provided and required handlers. The complex stuff
comes from the differentiation between f
2008/7/7 Carsten Ziegeler <[EMAIL PROTECTED]>:
> Hi,
>
> I came across a problem related to class loading in SCR and before trying
> to fix this I would get some feedback if this is really a bug (see
> FELX-624).
>
> Now, the problem is as follows - we have three bundles.
> Bundle C : Provides a s
Hi,
Providing "automatic" OBR bundle descriptions is on the iPOJO roadmap (and
is currently under development). It should describe service requirements /
capabilities as well as provided and required handlers. The complex stuff
comes from the differentiation between factories and instances (only
i
Hi,
I came across a problem related to class loading in SCR and before
trying to fix this I would get some feedback if this is really a bug
(see FELX-624).
Now, the problem is as follows - we have three bundles.
Bundle C : Provides a service for SERVICE_X which is registered through SCR.
Bun
[
https://issues.apache.org/jira/browse/FELIX-620?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12611146#action_12611146
]
Sahoo commented on FELIX-620:
-
Thanks for pointing out the typo in osgi.bundl2. Will the merge b
[
https://issues.apache.org/jira/browse/FELIX-624?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Carsten Ziegeler updated FELIX-624:
---
Attachment: Archiv.zip
Updated test case with three bundles
> Import statement incomplete whe
[
https://issues.apache.org/jira/browse/FELIX-624?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12611141#action_12611141
]
Carsten Ziegeler commented on FELIX-624:
Ok, i've updated my test case and it seems
[
https://issues.apache.org/jira/browse/FELIX-624?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Carsten Ziegeler updated FELIX-624:
---
Attachment: (was: test.zip)
> Import statement incomplete when inheriting from an abstract
Hi,
Hmmm... looks like the dev mailing list doesn't like attachments... OK,
I'll copy/paste the thing in... again, this would work if you run
felix.jar standalone... I'm not going to do all the jre entries, they
go on forever... Craig
org.osgi.framework.system.packages=org.osgi.framework; versi
Hi, Attached... Craig
PS - Note: this basic setup works fine if you run felix.jar standalone;
I think the key error log entry is:
whereas object of type "org.apache.log4j.ConsoleAppender" was loaded by
[WebappClassLoader
-Original Message-
From: Karl Pauls [mailto:[EMAIL PROTECTED]
Sent:
Could you maybe post your complete config.properties?
regards,
Karl
On Mon, Jul 7, 2008 at 12:59 PM, Craig Phillips
<[EMAIL PROTECTED]> wrote:
> Hi, I probably need to direct this to PAX, but I'll try here (felix dev)
> first... Sorry to trouble you, if anyone with felix class loader
> intimacy
[
https://issues.apache.org/jira/browse/FELIX-624?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Stuart McCulloch updated FELIX-624:
---
Component/s: (was: Maven Bundle Plugin)
Declarative Services (
Hi, I probably need to direct this to PAX, but I'll try here (felix dev)
first... Sorry to trouble you, if anyone with felix class loader
intimacy could take a look, I'd be appreciative... for a logger, this
isn't critical and I'll probably just write my own simple FileLogger
that implements LogSer
[
https://issues.apache.org/jira/browse/FELIX-624?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12611001#action_12611001
]
Stuart McCulloch commented on FELIX-624:
Yes, that sounds like an SCR issue - is it
[
https://issues.apache.org/jira/browse/FELIX-624?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12611000#action_12611000
]
Stuart McCulloch commented on FELIX-624:
I ran javap over the subclass and cannot fi
[
https://issues.apache.org/jira/browse/FELIX-624?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12610954#action_12610954
]
Carsten Ziegeler commented on FELIX-624:
No, the subclass is not directly using the
[
https://issues.apache.org/jira/browse/FELIX-624?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12610949#action_12610949
]
Stuart McCulloch commented on FELIX-624:
I'm not sure that the import is actually ne
[
https://issues.apache.org/jira/browse/FELIX-624?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Carsten Ziegeler updated FELIX-624:
---
Attachment: test.zip
Test case with two bundles.
> Import statement incomplete when inheritin
Import statement incomplete when inheriting from an abstract class from another
bundle
--
Key: FELIX-624
URL: https://issues.apache.org/jira/browse/FELIX-624
Project:
[
https://issues.apache.org/jira/browse/FELIX-620?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12610946#action_12610946
]
Stuart McCulloch commented on FELIX-620:
No - it does read the instructions from osg
[
https://issues.apache.org/jira/browse/FELIX-620?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12610939#action_12610939
]
Sahoo commented on FELIX-620:
-
That's good. But, as I mentioned in the README, the problem is no
[
https://issues.apache.org/jira/browse/FELIX-620?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12610934#action_12610934
]
Stuart McCulloch commented on FELIX-620:
Actually the include instruction is being p
Clement Escoffier a écrit :
Hi,
I go on my investigations about an OBR for Felix. I'm working on writing
descriptions for all released bundles. Indeed, Bindex generate correctly
package capabilities and requirement in term of package, but is not very
useful about services. The maven-bundle-plugi
55 matches
Mail list logo
