Page MenuHomePhabricator

Exact phrase intitle search incorrectly includes category match results
Closed, ResolvedPublic2 Estimated Story PointsBUG REPORT

Description

Steps to reproduce

  1. Search intitle:Massacres on enwiki. Link.
    • While first result "Massacre" does not fit query, it has a redirect "Massacres" that does (though the search results don't show the redirect).
    • Get to second page, and entries no long fit expect results. See "Chios massacre" which has no redirects that contain the plural form "Massacres".
  2. Search intitle:"Massacres" (with quotes) in the same manner. Link.
    • By first page, "Sand Creek massacre" is already among the results. No redirects.

What should have happened instead?:
These results should not have been included.

Software version (if not a Wikimedia wiki), browser information, screenshots, other information, etc:
My user agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/96.0.4664.45 Safari/537.36

Event Timeline

See https://www.mediawiki.org/wiki/Help:CirrusSearch : Stemming is automatic but can be turned off using an "exact phrase". Use intitle:"Massacres" instead.

See https://www.mediawiki.org/wiki/Help:CirrusSearch : Stemming is automatic but can be turned off using an "exact phrase". Use intitle:"Massacres" instead.

I did? "Sand Creek massacre" (not plural) still appears on the first page of the results.

The point is that intitle also seems to cover category entries, it seems?

Babi Yar (category Massacres in the Soviet Union)
Sand Creek massacre (category Massacres in 1864)
Manila massacre (category Massacres in the 1940s)

Aklapper renamed this task from intitle search leading to incorrect results to Exact phrase intitle search incorrectly includes category match results.Dec 1 2021, 7:49 AM

Change 757038 had a related patch set uploaded (by Ebernhardson; author: Ebernhardson):

[mediawiki/extensions/CirrusSearch@master] quoted intitle search must only query plain

https://gerrit.wikimedia.org/r/757038

Change 757038 merged by jenkins-bot:

[mediawiki/extensions/CirrusSearch@master] quoted intitle search must only query plain

https://gerrit.wikimedia.org/r/757038

This looks to be resolved, Collecting all 304 results no longer includes sand creek massacre.