1180144 - 'kumascript_macros' parameter doesn't match exact macro name

Status: RESOLVED FIXED

(developer.mozilla.org Graveyard :: Search, defect)

Description

[Sebastian Zartner [:sebo]]

What did you do?
================
Searched for occurrences of the 'csssyntax' macro in en-US locale:
https://developer.mozilla.org/en-US/search?locale=en-US&kumascript_macros=csssyntax

What happened?
==============
I also got matches for pages, which include the 'csssyntax2' macro.

What should have happened?
==========================
The parameter value of 'kumascript_macros' should be matched exactly, so searching for 'xyz' should match pages including the macro {{xyz}}, but not pages including the macros {{abcxyz}}, {{xyz123}} or {{abcxyz123}}.

Is there anything else we should know?
======================================
If substring matching for macros is needed, this could be implemented using wildcards. So kumascript_macros=*xyz* would match all three variants above.

Comment 1

[Florian Scholz (Open Web Docs)]

CC'ing Jannis our search expert.

I guess a start is this file:
https://github.com/mozilla/kuma/blob/master/kuma/wiki/search.py

Jannis, if this is a good first bug, any pointers are welcome.

Comment 2

[Eric Shepherd [:sheppy]]

(In reply to Sebastian Zartner [:sebo] from comment #0)

> What should have happened?
> ==========================
> The parameter value of 'kumascript_macros' should be matched exactly, so
> searching for 'xyz' should match pages including the macro {{xyz}}, but not
> pages including the macros {{abcxyz}}, {{xyz123}} or {{abcxyz123}}.

Agreed.

Comment 3

[Sebastian Zartner [:sebo]]

Note that this is needed to ensure there are no pages anymore using csssyntax, which will get broken once the CSS info pages are killed. See bug 1198949.
So it would be great if this could be fixed.

Sebastian

Comment 4

[Sebastian Zartner [:sebo]]

(In reply to Florian Scholz [:fscholz] (MDN) from comment #1)
> CC'ing Jannis our search expert.
> 
> I guess a start is this file:
> https://github.com/mozilla/kuma/blob/master/kuma/wiki/search.py
> 
> Jannis, if this is a good first bug, any pointers are welcome.

As Jannis just removed himself from CC, I redirect this question to John.

Sebastian

Comment 5

[John Whitlock [:jwhitlock]]

No, this is not a good first bug. All search bugs require 1) domain knowledge of ElasticSearch, 2) enough content for local development tests. The first is rare in first contributors, and the second requires staff access to anonymized production data. It should be prioritized in the staff bugs process.

Comment 6

[[github robot]]

Commit pushed to master at https://github.com/mozilla/kuma

https://github.com/mozilla/kuma/commit/d8ad405b2396c89cfaa43208464819c81c94f88a
bug 1180144: Use exact matches for keywords

For the keyword queries (kumascript_macros, css_classnames,
html_attributes) match on the exact value provided, rather than add in a
prefix match by default.

If a '*' or '?' is included, continue with an exact match without those
characters, and a wildcard match with the additional characters.

A keyword query of just '*' or '?' is ignored.

Comment 7

[John Whitlock [:jwhitlock]]

Deployed with the Elasticsearch 5.6 update.