Hello.

I'm using a BundleTracker to instantiate services based on the contents
of specially-crafted bundle manifest entries. This is working nicely,
but the one thing I don't really have is good error reporting: The code
that does the processing of bundles doesn't expose any services itself
and I don't really have any way to determine from elsewhere in the
program if processing a bundle has failed. Is there a preferred design
pattern for handling this?

One good option would appear to be the Event Admin, but the downside to
this is that (unless I'm wrong), I'd need to use the core OSGi APIs to
subscribe to events. That's not a huge downside, but it does mean that
code that otherwise has no interest in OSGi specifically has to depend
on the APIs (unless I write some other abstraction that relays
events to something like an rxjava Observable, and external code can
subscribe to this instead).

Perhaps the right thing to do in this case is to simply publish errors
to the Event Admin and just assume that some other code will subscribe
and take care of the errors?

Any opinions? I appreciate that the question is a bit vague, but I've
not actually seen error reporting discussed in any literature.

M

Attachment: pgpY8ZpgsCeyk.pgp
Description: OpenPGP digital signature

_______________________________________________
OSGi Developer Mail List
osgi-dev@mail.osgi.org
https://mail.osgi.org/mailman/listinfo/osgi-dev

Reply via email to