Add filters to handle custom post statuses when generating sitemap

[tolnem]

Summary

This PR can be summarized in the following changelog entry:

• Add filters to handle custom post statuses when generating sitemap

Test instructions

This PR can be tested by following these steps:

1. Create a custom post type
2. Enable Edit Flow plugin or otherwise create custom post status for the custom post type
3. Create one or more posts with the custom post status - see that they do not appear in the sitemap.xml created by Yoast SEO
4. Add a filter for 'wpseo_sitemap_poststatus' and 'wpseo_sitemap_poststatus_lastmodified' to add the custom post status to the array of post status names.
5. See that the posts are now included in the sitemap.xml

UI changes

• This PR does not change the UI in the plugin.

Quality assurance

• I have tested this code to the best of my abilities

Fixes #12074

[abotteram]

Thank you for contributing to Yoast SEO!

I have contacted our SEO team to review this PR. In the meanwhile please address the provided code review feedback.

[abotteram]

Acceptance 👍

Thank you for your contribution. This feature will be included in our 10.2 release on April 9th.

[stodorovic]

@xyfi I think that we should little postpone the milestone. I've found couple issues:

• There is still query which isn't covered with the filter:
  

  wordpress-seo/inc/sitemaps/class-post-type-sitemap-provider.php

  Lines 167 to 175 in 69f652c

  	$sql = "
  	SELECT post_modified_gmt
  	FROM ( SELECT @rownum:=0 ) init
  	JOIN {$wpdb->posts} USE INDEX( type_status_date )
  	WHERE post_status IN ( 'publish', 'inherit' )
  	AND post_type = %s
  	AND ( @rownum:=@rownum+1 ) %% %d = 0
  	ORDER BY post_modified_gmt ASC
  	";

  
  I think that we should use unique filter for all queries which will be more clear for users (and prevent possible differences between the root sitemap, CPT sitemaps and tax sitemaps). I'm preparing PR which will fix it.

• There is hard-coded "publish" in post object.
  

  wordpress-seo/inc/sitemaps/class-post-type-sitemap-provider.php

  Lines 563 to 566 in 69f652c

  	$post->post_type = $post_type;
  	$post->post_status = 'publish';
  	$post->filter = 'sample';
  	$post->ID = (int) $post->ID;

It could confuse users which uses filter wpseo_sitemap_entry.

• Attachments still uses only 'publish' status for parent posts.

• We should test performance if there are 100K posts because we've changed sensitive SQL queries (I don't expect differences, but we should test it).. It's related to Performance fix for Yoast sitemaps #12161

[stodorovic]

I just found new possible issue:

wordpress-seo/inc/sitemaps/class-taxonomy-sitemap-provider.php

Lines 172 to 183 in 69f652c

	$sql = "
	SELECT MAX(p.post_modified_gmt) AS lastmod
	FROM $wpdb->posts AS p
	INNER JOIN $wpdb->term_relationships AS term_rel
	ON term_rel.object_id = p.ID
	INNER JOIN $wpdb->term_taxonomy AS term_tax
	ON term_tax.term_taxonomy_id = term_rel.term_taxonomy_id
	AND term_tax.taxonomy = %s
	AND term_tax.term_id = %d
	WHERE p.post_status IN ('publish','inherit')
	AND p.post_password = ''
	";

Also, I created PR #12511, but it needs some modifications (eg. taxonomy).

[abotteram]

@tolnem I'm sorry, but based on @stodorovic 's findings we decided to revert this PR. As it looks now the 10.2 release has been pushed back by a week (to April 16th). If you would like to help you could review and test #12511 so we can try to get this into 10.2.