Oops, I forgot the links: [1] http://team.ops4j.org/browse/PAXSB-55 [2] http://team.ops4j.org/browse/PAXSB-41 [3] http://team.ops4j.org/browse/PAXWEB-254
Am 16.06.2012 15:33, schrieb Harald Wellmann:
I just ran into an issue with the BundleWatcher worker threads silently swallowing exceptions: http://team.ops4j.org/browse/PAXSB-55 While working on a fix, I realized this class is rather a crossbreed: It uses a SynchronousBundleListener, but then delegates STARTED events to an Executor whereas STOPPED events are handled in the same thread. Is there a good reason for that? As far as I can tell, the Executor was introduced to fix [2] and [3], and I wonder if the fact that it is not used for STOPPED events is just an omission. On the other hand, what's the point of making a synchronous listener asynchronous by means of an Executor, why not simply use a BundleListener which is asynchronous by default? Maybe we should really create a SynchronousBundleWatcher and an AsynchronousBundleWatcher, to handle _all_ events either synchronously or asynchronously? Best regards, Harald
_______________________________________________ general mailing list general@lists.ops4j.org http://lists.ops4j.org/mailman/listinfo/general
