Last modified: 2012-09-13 16:07:35 UTC
from #wikimedia irc chanel blockquote: 17:41:55 - Base: hi 17:44:34 - Base: http://uk.wikipedia.org/wiki/%D0%A4%D0%B0%D0%B9%D0%BB:%D0%AE%D0%9C%D0%97_%D0%95186_1.jpg has duplicates but query http://uk.wikipedia.org/w/api.php?action=query&prop=duplicatefiles&format=xml&dflimit=max&titles=File:ЮМЗ_Е186_1.jpg returns nothing - what is the reason and how fix it? 17:55:23 - Vulpix: Base: if you see the duplicate file, it's hosted on Commons 17:55:47 - Vulpix: that means that the file is hosted in commons and also on uk.wikipedia 17:56:17 - Base: I know 17:56:42 - Vulpix: since the file on Commons has the attribution already, you may safely delete the file on uk.wikipedia 17:56:50 - Base: Why API request dont return duplicate file name? 17:57:25 - Vulpix: because the api query only the local file repository, while the duplicate file is hosted on a remote repository 17:58:20 - Base: http://uk.wikipedia.org/w/api.php?action=query&prop=duplicatefiles&format=xml&dflimit=max&titles=File:%D8%EA%EE%EB%E02.jpg 17:58:32 - Base: this request work correctly 17:58:46 - Base: - it return name of duplicate 17:58:53 - Base: which is on commons 17:59:52 - Vulpix: but this second file has a different name 18:00:11 - Vulpix: while the first one you mentioned has the exact same name 18:01:08 - Vulpix: if probably could be considered a bug, but not sure how to fix it 18:03:04 - Base: i create list of files on ukwiki that has same names as files on commons. and i whant create list of files from this list that is duplicates of commons ones to delete them... 18:05:05 - Base: May be there is some other way to discover has file duplicates on commons or no? 18:06:45 - Vulpix: I'm sure this is a bug. I'm reading the code of ApiQueryDuplicateFiles and I think I've found the problem 18:07:17 - marktraceur: Base: The server does hash checks on files and should return a warning if there are two exactly similar files 18:08:16 - Vulpix: marktraceur: in this case, the file on commons was upload by a bot after it was uploaded to ukwikipedia 18:09:02 - Vulpix: He probably wants to bot-delete those kind of files from ukwikipedia so the commons version is used instead 18:10:03 - marktraceur: Base: So look at the contribs page for the bot? 18:10:37 - Base: marktraceur: which bot? 18:11:25 - Vulpix: This seems to be the problematic line: https://gerrit.wikimedia.org/r/gitweb?p=mediawiki/core.git;a=blob;f=includes/api/ApiQueryDuplicateFiles.php;hb=HEAD#l121 18:12:46 - Base: can you fix it? 18:13:34 - Vulpix: No, sorry, I have no mediawiki installation with a commons repository to check it 18:13:46 - Vulpix: File a bug on https://bugzilla.wikimedia.org/ end of blockquote
Sounds like the problem, which is fixed with Gerrit change #20519, which is part of 1.20wmf11 and that is not live right now.
(In reply to comment #1) > Sounds like the problem, which is fixed with Gerrit change #20519, which is part of > 1.20wmf11 and that is not live right now. May be. When WMF wikis will be upgraded to 1.20wmf11 ?
See [[mw:MediaWiki 1.20/Roadmap]] http://www.mediawiki.org/wiki/MediaWiki_1.20/Roadmap
Works now: http://uk.wikipedia.org/w/api.php?action=query&prop=duplicatefiles&format=xml&dflimit=max&titles=%D0%A4%D0%B0%D0%B9%D0%BB:%D0%AE%D0%9C%D0%97_%D0%95186_1.jpg