Message12468

Author jeff.allen
Recipients bmvanwyk, jeff.allen, tomluk, wfouche2, zyasoft
Date 2019-04-28.16:59:33
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <1556470773.35.0.446160402159.issue2445@roundup.psfhosted.org>
In-reply-to
Content
Fixed, I claim: https://hg.python.org/jython/rev/36db7416b3b7

@Werner I'm confident MQQueue will import, but would you see if it behaves too?

The historic work-arounds seem still to be necessary. The special-case for Map/Iterable still is and I had to add one for Enumeration/Iterator. I have not tried to remove the one based on assignableFrom() Jim added, nor make other inessential improvements that ocurred to me. Maybe in another change set.

If it's not obvious yet, I think that when it comes to treating Java collections as Python ones, or giving proxied Java objects whatever API, explicit would be better than implicit. E.g. m.map really is (only) a Python map, but wraps the Map that m is proxy for, while m has exactly the methods its type requires. (Choose name/syntax.)

Also, I don't think I'm able to express "Java rules" in MRO-speak. It may be possible (suspect not), but by far the most reliable course is surely to ask Java itself what is inherited and from where. Not for 2.7.2 I think.
History
Date User Action Args
2019-04-28 16:59:33jeff.allensetmessageid: <1556470773.35.0.446160402159.issue2445@roundup.psfhosted.org>
2019-04-28 16:59:33jeff.allensetrecipients: + jeff.allen, zyasoft, tomluk, bmvanwyk, wfouche2
2019-04-28 16:59:33jeff.allenlinkissue2445 messages
2019-04-28 16:59:33jeff.allencreate