[
https://issues.apache.org/jira/browse/FELIX-2066?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12832923#action_12832923
]
Shrini Joshi commented on FELIX-2066:
-------------------------------------
I agree with Richard that this is not a straight forward fix.
"The only platform-neutral way is to modify File Install to deploy the fragment
directory before the host directory and perform a refresh on all hosts."
But I think above proposal assumes that all fragments are dpeloyed in same
directory which may not be a case.
I was thinking of following option:
After every scan, if one or more (non-fragment) bundles have been installed,
then call packageadmin.resolve(null). As per the javadoc call to this method
with "null" means that framework will resolve all the unresolved bundles. This
wil ensure that all unresolved fragments can get resolved.
This should be done before starting these new bundles to ensure that fragment
bundles are attached to their host.
Any comments?
> File Install fails to move fragment bundle into RESOLVED state
> --------------------------------------------------------------
>
> Key: FELIX-2066
> URL: https://issues.apache.org/jira/browse/FELIX-2066
> Project: Felix
> Issue Type: Bug
> Components: File Install
> Environment: Windows XP, Java 6, Equinox , File-Install 2.0.8
> Reporter: Shrini Joshi
>
> The file-install will install the fragment bundle but fails to move it to
> resolve. The fragment bundle remains in INSTALLED state.
> If we install it from equinox shell, then fragment bundle is installed and
> resolved correctly.
> Can someone please comment ?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
