Bugs item #3012516, was opened at 2010-06-07 04:02 Message generated for change (Comment added) made by bewareofdoug You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=603138&aid=3012516...

Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: interwiki Group: None Status: Open Resolution: Accepted Priority: 7 Private: No Submitted By: Kenrick Anggara (kenrick95) Assigned to: Nobody/Anonymous (nobody) Summary: cosmetic_changes.py : bot remove double interwiki

Initial Comment: My bot : http://id.wikipedia.org/w/index.php?title=Rangkaian_seri_dan_paralel&dif... http://id.wikipedia.org/w/index.php?title=Rangkaian_seri_dan_paralel&dif...

Other's bot but I'm sure it is the same version as mine : http://id.wikipedia.org/w/index.php?title=Piala_Thomas_dan_Uber_2010&dif... http://id.wikipedia.org/w/index.php?title=Piala_Thomas_dan_Uber_2010&dif...

My bot (Kenrick95Bot) version : Pywikipedia nightly:pywikipedia (r28, 2010/03/16, 11:21:32) Python 2.6.5 (r265:79096, Mar 19 2010, 21:48:26) [MSC v.1500 32 bit (Intel)] config-settings: use_api = True use_api_login = True

Please fix... Thanks...

----------------------------------------------------------------------

Comment By: Doug (bewareofdoug)

Date: 2011-12-30 04:31

Message: I agree completely, at least for now. This is the solution that has been adopted on wikisources and implemented in a script by Candalua. Essentially nobody else touches mainspace works which is where the multiple interwikis show up.

Possibly in the long run Wikisources need to abandon interlanguage linking by this method and switch to a different method of linking translations, but that's a problem for wikisource.org and the wikimedia bugzilla world.

I will attach Candalua's solution (after I confirm with him that's it's free licensed) for discussion/review.

----------------------------------------------------------------------

Comment By: xqt (xqt) Date: 2011-12-30 04:23

Message: The currently most widely accepted paradigm is a 1:1 relationship between pages of different sites. Multiple interwiki links means these pages does not match exact each other. Ok some people wants links which match pages they are quite similar. I know at wikisource this a most imported issue. As valhallash says implementation is not trivial because a central paradigma the bot is working with must be changed. I think there are several points: MediaWiki too cannot deal with multiple interwiki links by returning them via api pwb cannot deal with multiple iw links. Changing it concerns the data structures and most of the given methods dealing with iw links. My idea is ignoring all source and target pages with multiple interwiki links like it does with interwiki conflicts (if the methods are able to handle with ist)

----------------------------------------------------------------------

Comment By: Doug (bewareofdoug) Date: 2011-12-29 10:32

Message: The interlanguage extension exists and is stable: http://www.mediawiki.org/wiki/Extension:Interlanguage However, it does *not* solve this problem at all as far as I can see, it simply replaces the interwiki language link system currently in use. It won't help on any project it isn't adopted on and it won't likely be adopted on wikisources unless it adds a lot more than it takes away. Waiting on the across the board adoption of an extension to fix a bug that can be easily avoided (by making the bot skip such pages) seems like a very bad idea. Interwiki.py's behavior is wrong in this case as well, which is part of the issue.

----------------------------------------------------------------------

Comment By: Merlijn S. van Deen (valhallasw) Date: 2011-12-29 10:23

Message: It's not entirely trivial to fix this at the moment. Cosmetic_changes essentially does what interwiki.py does: it only uses the *last* mention of a certain language.

Changing this behaviour will break interwiki.py, so it has to be implemented with some care.

----------------------------------------------------------------------

Comment By: xqt (xqt) Date: 2011-12-29 09:33

Message: We are waiting for the interlanguage extension ;)

----------------------------------------------------------------------

Comment By: Doug (bewareofdoug) Date: 2011-12-29 07:18

Message: bug https://sourceforge.net/tracker/?func=detail&aid=3465680&group_id=93... has been identified as a duplicate of this bug, I don't think we should close the other bug report yet because, although the code involved may be the same, the actual issue is slightly different. cosmetic_changes.py should convert [[en:Steveville|Steveville]] to [[:en:Steveville|Steveville]] because it's in the body of the article rather than treating it as an interwiki language link.

----------------------------------------------------------------------

Comment By: Doug (bewareofdoug) Date: 2011-12-27 14:01

Message: It is open, what you see is Valhallasw changing the close_date to none when he essentially reversed xqt's change of the same day. I've increased the priority and removed assignment to xqt as his last action was to close it as "Wont Fix", so there's no reason to believe that he's actually working on this. Last I knew interwiki.py did the same thing. Is that the same issue? I mean does cosmetic_changes.py get the removal function from interwiki.py or is the same code at least. We have a work around on ws (where double and triple interwikis are common) though it really means skipping such pages.

----------------------------------------------------------------------

Comment By: Yevhen Movsesov () Date: 2011-12-26 23:04

Message: 1. Is this issue is "Open" ? Ask, because on this page I see "close_date 2010-06-09 00:36"

2. If it "Open" please increase priority, or provide workaround.

3. s this problem with "standardizePageFooter" module ? If yes, which line should I comment as workaround ?

----------------------------------------------------------------------

Comment By: Merlijn S. van Deen (valhallasw) Date: 2010-06-09 00:54

Message: In no case should a bot destroy data, especially not if the reasoning behind it is 'the software cannot cope with it'. A double interwiki might be wrong from a MediaWiki/pwb perspective, but it is *not* wrong from a content perspective. As such, they should simply be left alone.

----------------------------------------------------------------------

Comment By: xqt (xqt) Date: 2010-06-09 00:36

Message: MediaWiki also could not handle double interwiki to the same site. It only saves one link into its database and only one link is reachable via api or database request. Unless this is not changed by MediaWiki's behavior I do propose not to changed bot's behavior.

----------------------------------------------------------------------

Comment By: JAn (jandudik) Date: 2010-06-07 07:46

Message: Double interwiki in page is incorrect - interwiki bot cannot handle it and it makes problem for them.

----------------------------------------------------------------------

Comment By: Kenrick Anggara (kenrick95) Date: 2010-06-07 04:09

Message: Update : I have updated my Pywikipediabot into the latest version. I test it on that Thomas & Uber Cup page and the result is : http://id.wikipedia.org/w/index.php?title=Piala_Thomas_dan_Uber_2010&dif...

----------------------------------------------------------------------

You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=603138&aid=3012516...