[wp-trac] [WordPress Trac] #37009: When two different tags generate the same slug, the second tag is rejected

[WordPress Trac]

#37009: When two different tags generate the same slug, the second tag is rejected
------------------------------+---------------------------
 Reporter:  michael.costanza  |       Owner:  boonebgorges
     Type:  defect (bug)      |      Status:  assigned
 Priority:  normal            |   Milestone:  4.6
Component:  Taxonomy          |     Version:  4.5.2
 Severity:  normal            |  Resolution:
 Keywords:                    |     Focuses:
------------------------------+---------------------------

Comment (by boonebgorges):

 [attachment:37009.diff] starts to move to the more precise
 `get_term_by()`, but it's still broken.

 When you query for `slug` - in `term_exists()`, `get_term_by()`, or
 `get_terms()` - the value is passed through `sanitize_title()`. This
 sanitization is done on the way into the database to ensure that the slug
 is URL-safe. It's done while querying mostly to avoid SQL injection
 issues. This mixing of concerns makes it impossible to do a strict query
 for a slug match: `get_term_by( 'slug', '$foo' )` will return a term with
 the slug `foo`. This pretty clearly seems wrong.

 Ideally, we would stop doing so much heavy sanitization on values passed
 to `get_term_by()` and `get_terms()`, so that we would stop getting false-
 positive matches for things like `$foo`/`foo`. Simply removing the
 `sanitize_title()` sanitization would probably solve the immediate
 problem, but has compatibility ramifications for plugins that are
 expecting the current fuzzy matching. My inclination is to stop supporting
 fuzzy matches in the case of `get_term_by()`, because the nature of the
 function is such that (IMO) you should expect exact matches only. But a
 second opinion would be helpful here.

--
Ticket URL: <https://core.trac.wordpress.org/ticket/37009#comment:6>
WordPress Trac <https://core.trac.wordpress.org/>
WordPress publishing platform