The autonyms in Names.php and langdb for the two Arabic-script Kazakh entries, kk-arab and kk-cn, seem to be wrong.
kk-arab:
- In Names.php, the part in brackets is "تٴوتە" (teh, high hamsa, waw, teh, ae)
- In langdb, the part in brackets is "تٶتە" (teh, high hamza waw, teh, ae)
According to https://en.wikipedia.org/wiki/Kazakh_alphabets#Use_of_Hamza, the hamza can only come at the beginning of a word, which would imply these spellings are not possible. It also says it is not used in words containing "e". "e" is written with the character "ae", which would mean this word should not have a hamza.
kk.json in core has "توتە" (teh waw teh ae) without a hamza. That is also the only spelling I've found on other websites.
kk-cn:
- The part in brackets is "جۇنگو" (jeem, u, noon, gaf, waw)
A Google search for it finds almost zero results which aren't linked to MediaWiki. There are two from kazakh.people.com.cn but on those pages it's part of someone's name. DuckDuckGo only finds pages on Commons.
The word is a transliteration of 中国 (Zhōngguó) and the only spelling I've found elsewhere (including the Chinese Wiktionary) is "جۇڭگو" (jeem u ng gaf waw).
Based on a recent discussion with @Amire80 an ideal next step here would be to remove the two language codes. Adding steps below:
Draft plan, may change:
- Remove kk-arab and kk-cn from language-data or declare it as deprecated
- Remove kk-arab and kk-cn from jquery.uls
- Remove kk-arab and kk-cn from ULS extension
- Remove kk-arab and kk-cn from ULS search index
- Redirect kk-arab and kk-cn to kk and kk-cyrl in core for backwards compatibility
- Declare kk-arab and kk-cn as deprecated in core RELEASE_NOTES
- Deploy core MediaWiki with the new code on translatewiki
- Verify all the projects that have kk-arab and kk-cn translations:
- Remove all translations in kk-arab and kk-cn in all projects on translatewiki ( https://translatewiki.net/wiki/Renaming_language_codes )
- Verify correct functionality in translatewiki
- Verify all the projects that have kk-arab translations: https://translatewiki.net/wiki/Special:LanguageStats?language=kk=arab&suppressempty=1&x=D should be empty after purging
- Verify all the projects that have kk-cn translations: https://translatewiki.net/wiki/Special:LanguageStats?language=kk-cn&suppressempty=1&x=D should be empty after purging
- Export all the projects from translatewiki to Gerrit, etc.
- Prohibit translation into kk-arab and kk-cn on translatewiki in mw-config/TranslateSettings.php
- Deploy translatewiki configuration with prohibited kk-arab and kk-cn on translatewiki
- Set translatewiki language portal for kk-arab and kk-cn to "disabled".
- Set translatewiki language category for kk-arab and kk-cn to "disabled".
- Change kk-arab and kk-cn in translatewiki content wiki pages to relevant language codes (if needed)
- Test https://translatewiki.net/wiki/Special:ActiveLanguages
- Deploy updated codes on Wikimedia sites (probably in the train)
- Remove data tagged as kk-arab and kk-cn from Wikidata
- Maybe? (Under review!) Remove MessagesKk_arab.php, MessagesKk_cn.php and add kk-arab, kk-cn to DEPRECATED_LANGUAGE_CODE_MAPPING in core