Hard-code check for Multilingual Wikisource to avoid errors #470
Conversation
Rather than looking for an empty subdomain, check for the Wikidata ID of Multilingual Wikisource. Bug: T342520
| } | ||
| $out[$subdomain] = $datum['label']; | ||
| preg_match( '|https://([a-z-_]*)\.?wikisource\.org|', $datum['website'], $matches ); | ||
| $out[$matches[1]] = $datum['label']; |
There was a problem hiding this comment.
$matches[1] seems a bit fragile to me if Wikidata data are invalid for some reason.
There was a problem hiding this comment.
Good point. I did initially try it with Wikimedia language code (P424), but that doesn't get the correct value for some, like syl. Although maybe https://www.wikidata.org/wiki/Q120844812 should not be called a Wikisource language edition yet, while it's in incubator.
If someone puts an invalid value for Official website then you're right, this would break, but then whatever we pull from Wikidata could get broken at some point so I'm not sure it's too different.
There was a problem hiding this comment.
Sorry, I have made this disucssion slip. Indeed, that's a good point. I am fine with merging this PR.
Rather than looking for an empty subdomain, check for the Wikidata ID of Multilingual Wikisource.
Bug: T342520