Page MenuHomePhabricator

Cannot find term that is mixed with other characters
Closed, DuplicatePublicBUG REPORT

Description

I don't get any hits. I cannot search for links that I've added to MediaWiki pages.

Event Timeline

David_Hedlund renamed this task from Cannot find terms that is not separated by white space to Cannot find terms that is mixed with other characters.Mar 4 2023, 7:29 PM
David_Hedlund updated the task description. (Show Details)
David_Hedlund renamed this task from Cannot find terms that is mixed with other characters to Cannot find term that is mixed with other characters.Mar 4 2023, 7:34 PM

See https://www.mediawiki.org/wiki/Help:Searching - what happens if you try *wikimedia.org*? Which MediaWiki version is this? Please fill in the sections in the template instead of deleting them. Thanks!

@David_Hedlund, I'm having trouble following what's happening here. Your link in "Search: wikimedia.org" points to localhost so I can't really follow it.

I think I've figured out what you mean by "mixed with other characters". If you have phabricator.wikimedia.org in the document, and you search for wikimedia.org, you want them to match. Is that correct?

Right now, different language wikis handle this differently. English and Italian are configured to break on periods, but other languages are not. This configuration inconsistency goes back before my time (more than 8 years) and I don't know the historical reason for it. I want to enable it everywhere, but it currently has some negatives, too. In particular, acronyms (like N.A.S.A) get split up. In addition to not matching NASA without periods, this case is particularly bad because a is a very common word and all sorts of problems arise from it being split out.

My plan is to fix the acronym problem in T170625, test the effect across languages of the acronym fix and generally splitting on periods, underscores, etc., and enable both everywhere (or almost everywhere if there is a language where it causes problems).

So, if the issue is phabricator.wikimedia.org not matching wikimedia.org, then this ticket can be closed or merged into T170625, which will address the issue. (I'm hoping to work on it this year, but planning for the second half of the year is still ongoing.)