Page MenuHomePhabricator

STOP adding zh-yue captions on Commons. Migrate existing ones to yue.
Closed, ResolvedPublic

Description

On wikimedia commons, it is now possible to add captions in zh-yue, which is NOT a valid iso langcode. It should only be yue.

Please prevent users from adding zh-yue.

A short-term solution is for the Android app team, please stop users adding zh-yue. Make them add yue captions instead. Most of the zh-yue captions were added by your app's Suggestededit.

To clean up the mess, please migrate existing ones to yue as well.

(Added comments.)
I forgot to mention, the same should be done for the following pairs and possibly others:

invalid codecorrect code
zh-min-nannan
zh-classicallzh

Event Timeline

Restricted Application added a subscriber: Aklapper. · View Herald Transcript

I think this affects Wikidata as well – I just added separate and distinct yue and zh-yue labels to the sandbox item, which probably shouldn’t be possible. It looks like both MediaWiki (meta=languageinfo) and Wikibase (meta=wbcontentlanguages) consider yue and zh-yue to be different languages to some degree (albeit with the same name and autonym).

The same seems to be true for other deprecated language codes, e. g. als (⇒ gsw) or be-x-old (⇒ be-tarask).

Likely related to (possibly a duplicate of) T44396: duplicate/invalid language codes.

That task notes that the regular Wikidata user interface already doesn’t allow you to set terms using deprecated language codes – if you add e. g. ?uselang=zh-yue to the URL, MediaWiki normalizes it (in this case to yue) before Wikibase ever sees it. (The example edits I linked above were made via Special:SetLabel, which doesn’t use the user interface language and so still allows those language codes.) Maybe the Commons app should apply a similar mapping even if the underlying Wikibase problem isn’t fixed soon (so it would still be possible to set zh-yue via the API, but it wouldn’t happen so often through the Commons app).