WP should do sanity checks for paginated posts, pages and comments

[Denis-de-Bernardy]

Create a post with two pages using the <!--nextpage--> tag. Publish, and browse the post's *third* page.

WP should return a 404 here, rather than the last page of the post.

Along the same lines, it should return the correct canonical urls for paginated posts. Currently, rel=canonical will only ever return the post's first page.

[elky]

The rel=canonical issue really ought to be split out in to a separate bug, imho.

[dd32]

I'm attaching a patch I've got locally for this.

This will need a more thought out patch however.

[SergeyBiryukov]

Closed #21278 as a duplicate. Related: #11857 (for archive pages).

As noted in ticket:21278:9, paginated comments are affected too.

[mdgl]

See also #21579 where a similar issue occurs with paged comments. Invalid comment pages, for example ​http://example.com/content/comment-page-999/ do not return a 404 but equally do not find any comments! Is a new ticket needed or can we use this one to fix the 404 issue in both cases and #21579 to generate the proper rel="canonical" in both cases?

[SergeyBiryukov]

Replying to mdgl:

Is a new ticket needed or can we use this one to fix the 404 issue in both cases

No need for a new ticket. Updated the summary.

[here]

See also #18660

@elky was correct 3 years ago, this is two issues.

1) site/post/## - where ## is non-existant - should 404 or 301 not 200 duplicate content
2) site/post/## - where ## > 1 - rel=canonical should be site/post/## not duplicate first page site/post based on :

​http://support.google.com/webmasters/bin/answer.py?hl=en&answer=1663744

"rel="next" and rel="prev" are orthogonal concepts to rel="canonical". You can include both declarations."

[joostdevalk]

Canonical - Pagination Patch

[joostdevalk]

The patch I just added fixes point 1 in @here's comment above: ie. the fact that WP serves a 200 for a page that doesn't exist by, if the post is paginated, checking how much pages are in the post and acting accordingly, by either redirecting to the first page of the post or by providing a proper URL.

[joostdevalk]

Patch for rel=canonical on paginated posts

[joostdevalk]

This second patch adds the page query variable to the canonical. If it's not used in conjunction with the other patch, this should check whether this page can actually exist.

[joostdevalk]

Looking at how it's dealt with in post-template.php a bit more, this probably should actually be done a bit different. I think _wp_link_page should be split up and part of that should be in link-template.php, which should then be used by _wp_link_page and the functions above...

[joostdevalk]

rel-canonical patch which adds paginated post link API call

[joostdevalk]

The last ticket introduces get_paginated_post_link, for which the functionality is removed from _wp_link_page and now used in both _wp_link_page in post-template.php and rel_canonical in link-template.php. This looks a lot cleaner to me than what we had so far and also allows for future use of get_paginated_post_link within a patch for #18660.

[joostdevalk]

Patch for rel=canonical on paginated posts

[joostdevalk]

v3 of the patch changes get_paginated_post_link into the semantically correct name get_paginated_post_url.

[wonderboymusic]

these are all marked 3.7-early

[sirzooro]

Related: #21266

[SergeyBiryukov]

#26123 was marked as a duplicate.

[nacin]

There is some cool work from joostdevalk here, but it's too late for 3.8. Maybe we can tackle the messiness of pagination in 3.9.

[SergeyBiryukov]

Related: #28081

[SergeyBiryukov]

Related: #29655

[boonebgorges]

Removing from trunk as per #30284

[ocean90]

#32505 was marked as a duplicate.

[wonderboymusic]

I found an amazingly awesome bug in Rewrite - makes it where get_query_var( 'page' ) returns /4 .... no problem, let's just cast to (int) .... what's that you say? The cast results in 0?!?!?

See 11694.2.diff​

[wonderboymusic]

In 34492:

Canonical/Rewrite: sanity check posts that are paged with <!--nextpage-->. Page numbers past the max number of pages are returning the last page of content and causing infinite duplicate content.

Awesome rewrite bug: the page query var was being set to '/4' in $wp. When cast to int, it returns 0 (Bless you, PHP). WP_Query calls trim( $page, '/' ) when setting its own query var. The few places that were checking page before posts were queried now have sanity checks, so that these changes work without flushing rewrites.

Adds/updates unit tests.

Props wonderboymusic, dd32.
See #11694.

[wonderboymusic]

In 34494:

After [34492], no need to import the global instance when we are, in fact, currently, that instance.

See #11694.

[wonderboymusic]

In 34496:

Canonical/Rewrite: After [34492], fix rel="canonical" URLs for paginated posts.

Props wonderboymusic, joostdevalk.
See #11694.

[wonderboymusic]

The only remaining piece is redirecting out-of-bounds comment pages - which can be tackled once we finish ripping apart WP_Comment_Query in #8071

[wonderboymusic]

Closing as fixed because #11248 is the last piece.

[SergeyBiryukov]

#34194 was marked as a duplicate.

[danielbachhuber]

Changing rewrite regex means anyone filtering rewrite_rules will experience unexpected behavior with this change.

[wonderboymusic]

Replying to danielbachhuber:

Changing rewrite regex means anyone filtering rewrite_rules will experience unexpected behavior with this change.

what's your point? you want to keep the bugs?

[danielbachhuber]

Replying to wonderboymusic:

what's your point? you want to keep the bugs?

My comment was observational, not inflammatory. Just noting that r34492 breaks backwards compatibility in the manner I specified. I've had to ​update WP-CLI's test accordingly.

[dd32]

Just noting that #40773 is a follow up to this.

[34492] only applies when <!--nextpage--> exists in a post, therefor, /post-name/1234567/ will only redirect if there's pages defined, if no pages it just displays the duplicate content.

[SergeyBiryukov]

In 47727:

Canonical: Redirect paged requests for non-paginated posts to the post permalink.

This avoids displaying duplicate content of the same post under different URLs and ensures the canonical URL is correct.

Previously, requests for invalid page numbers were only redirected to the post permalink if the post was actually paginated using the <!--nextpage--> marker.

Follow-up to [34492].

Props jeremyfelt, prografika, sachit.tandukar, subrataemfluence, hronak, ekatherine, henry.wright, chesio, dd32, SergeyBiryukov.
Fixes #40773. See #45337, #28081, #11694.

[SergeyBiryukov]

In 47760:

Canonical: Redirect paged requests for a static page assigned as the "Posts page".

This avoids displaying duplicate content of the home page under different URLs with appended page numbers.

This change only affects the <!--nextpage--> pagination (page query variable) and not the regular multiple posts pagination (paged query variable).

The posts page does not support the <!--nextpage--> pagination, so requests for invalid page numbers should be redirected to the page permalink, applying the logic previously implemented for single posts or pages.

Follow-up to [34492], [47727].

Props jeremyfelt, sachit.tandukar, SergeyBiryukov.
Fixes #45337. See #40773, #28081, #11694.

[SergeyBiryukov]

In 47761:

Canonical: Only redirect non-existing page requests to the post permalink if the post is found.

Follow-up to [47760].

See #45337, #40773, #28081, #11694.