[wp-trac] [WordPress Trac] #52998: Add a filter to paginate_links to allow bypassing merging query params

[WordPress Trac]

#52998: Add a filter to paginate_links to allow bypassing merging query params
----------------------------+-----------------------------
 Reporter:  jonoaldersonwp  |      Owner:  (none)
     Type:  enhancement     |     Status:  new
 Priority:  normal          |  Milestone:  Awaiting Review
Component:  General         |    Version:
 Severity:  normal          |   Keywords:  seo
  Focuses:  performance     |
----------------------------+-----------------------------
 The [https://core.trac.wordpress.org/browser/tags/5.7/src/wp-includes
 /general-template.php#L4154 paginate_links] function constructs the HTML
 markup used in archive pagination.

 As part of this, any parameters present in the requesting URL are merged
 into the pagination links
 ([https://core.trac.wordpress.org/browser/tags/5.7/src/wp-includes
 /general-template.php#L4197 ref]).

 This is a safeguard to ensure that any logic which relies on `$_GET`
 parameters in a paginated series still functions as expected as the user
 navigates through paginated states. E.g., a plugin may use query
 parameters to describe a date range in an archive; such as
 `/archive/page/2/?date_from=2001-12-20&date_to=2020-05-14`.

 I'd like for theme/plugin authors to be able to disable/bypass this
 'merging' process.

 **The problem(s)**

 The current behaviour makes it possible to request (paginated) archive
 templates with arbitrary, meaningless query parameters (e.g.,
 https://wordpress.org/support/topic-
 tag/iframe/page/5/?cats=yes&dogs=no&cake=please&utm_tracking=broken).
 Those parameters persist through all subsequent paginated requests.

 This bypasses caches, breaks digital analytics, and inflates crawling from
 third party systems (search engines, social media, etc), as it 'creates'
 many additional URLs.

 Furthermore, because WordPress doesn't natively output canonical URL tags
 on paginated archives, this can be ''extremely'' destructive from an SEO
 perspective; particularly on large sites, where the 'existence' of many
 such URLs compounds the problems above. And whilst the presence of a
 canonical tag would mitigate ''some'' of that damage, it wouldn't address
 those other problems.

 **A potential solution**

 I propose that we should introduce a new filter into
 [https://github.com/WordPress/WordPress/blob/master/wp-includes/general-
 template.php#L4154 paginate_links] which allows theme/plugin authors to
 bypass the code block which merges in query parameters.

 The filter should accept a simple boolean value, which acts as a switch to
 skip the merging logic.

 ** A note on the existing `paginate_links` filter**

 Note that [https://github.com/WordPress/WordPress/blob/master/wp-includes
 /general-template.php#L4302 a paginate_links filter already exists] at the
 end of the function, but this only allows for manipulating the `href`
 attribute of the link. Whilst this could ''technically'' be used to
 ''strip'' parameters via brute force, this is cumbersome, and requires
 knowledge of the correct canonical URL (which WordPress doesn't
 natively/reliably have).

-- 
Ticket URL: <https://core.trac.wordpress.org/ticket/52998>
WordPress Trac <https://core.trac.wordpress.org/>
WordPress publishing platform