I guess, I've never been really sure if it was actually necessary to have the filter. I assume that the filter would get by default only the messages for the current user, right?
Any stanza send to an client will always, as per the XMPP spec, have one of the following values as to attribute:
- the full JID of the client
- the bare JID of the client
- no to attribute at all
Therefore the ToContains filter is actually
- harmful, because there could be valid stanzas without a to value
- unnecessary, because every stanza send by a standard compliant XMPP server to the client, is addressed, implicitly or explicitly, to the client
Well the point of a library is that it will also provide helpful classes, which are not used in the library itself, but only by clients of it, isn't it?
That's right, but it's a harmful and in most cases unnecessary, trival piece of code that can easily be re-implemented if you really need it (but then I would to eager to hear the use case).
Basically, I just wondered that it was removed without notice.
Smack 4 has undergone the greatest refactoring in the history of Smack. Yes it should be documented, but I'm currently a one man show and therefore have not many resources available. I'm happy about every report or contribution, so that I can update the "Smack 4 Readme and Upgrade Guide", which I just did. Thanks for you report.
Makes total sense to have integration tests and then not use them
Again, one-man show, not much time/resources. Integration tests are not a top priority right now, especially due to their server implementation dependend nature (you want tests against the XMPP spec, not against a particular XMPP server). Good unit tests are more important (which we have, but there is also room for improvement). Contributions are always welcome. :-)