wp_publish_post() does not fire post_updated action

[Rarst]

do_action( 'post_updated', $post_ID, $post_after, $post_before);

Gives convenient access to before/after state of the post for actions that are conditional on exact changes performed.

wp_publish_post() inconsistently doesn't fire this hook, even though it does fire edit_post and save_* hooks that come before and after it. This makes it impossible to rely on post_updated hook since it will never fire for any update performed by wp_publish_post().

[SergeyBiryukov]

Related: #24957

[chandrapatel]

I have added post_updated action in wp_publish_post() which is fire after edit_post action.

[peterwilsoncc]

Within core wp_publish_post() is only called from check_and_publish_future_post(), it lacks two features of wp_update_post() and wp_insert_post():

• post_updated does not fire
• wp_post->date_modified does not change
• wp_post->date_modified_gmt does not change

My educated guess is it's by design and publishing a scheduled post is not considered an update per se.

A number of suitable actions are fired in wp_transition_post_status(), my inclination is to leave wp_publish_post() unchanged.

[Rarst]

My educated guess is it's by design and publishing a scheduled post is not considered an update per se.

It changes one of the most important post fields in a major way. If it's by design I would like to hear a reason why it shouldn't be handled consistently with every other post change.

A number of suitable actions are fired in wp_transition_post_status()

None of the hooks in status transition give access to full before/after state of the post.

[DrewAPicture]

Replying to Rarst:

My educated guess is it's by design and publishing a scheduled post is not considered an update per se.

It changes one of the most important post fields in a major way. If it's by design I would like to hear a reason why it shouldn't be handled consistently with every other post change.

I can get on board with the idea that publishing a scheduled post constitutes a post update.

On whether we should add the hook, I actually wonder if maybe we shouldn't look at deprecating wp_publish_post() and switch to using wp_insert_post() in check_and_publish_future_post() instead. You can attempt to maintain a parallel use-case for only so long before it becomes pointless.

Of course we pick up a hook or two in the process, but it also would streamline the API a bit. It would need a good bit of tests to make sure we aren't breaking something.

[afercia]

Replying to peterwilsoncc:

My educated guess is it's by design and publishing a scheduled post is not considered an update per se.

That's the point I guess :) Users can actually modify a scheduled post before publishing it and as far as I understand there's no way to detect this. By the way seems the history here is a bit complicated see [21942] and [23206] and the related tickets and see also 21963.

[voldemortensen]

This seems to need significant discussion which probably won't happen in time for 4.6, punting.