Make WordPress Core

Opened 6 years ago

Closed 5 years ago

#9065 closed defect (bug) (wontfix)

Can not store 0 (zero) in meta_value using update_usermeta() in user.php

Reported by: stevejohnson Owned by: jfarthing84
Milestone: Priority: normal
Severity: normal Version: 2.8.4
Component: General Keywords:
Focuses: Cc:


The following lines in wp-includes/user.php evaluate $meta_value passed to update_usermeta() (current line #s from trunk):

397 	    if (empty($meta_value)) {
398 	        return delete_usermeta($user_id, $meta_key);
399 	    }

A numerical or string zero evals as empty, resulting in the inability to store zero as a meta_value.

A better test for truly empty might be:

397 	    if ($meta_value == '')) {
398 	        return delete_usermeta($user_id, $meta_key);
399 	    }

Attachments (1)

user.php.diff (609 bytes) - added by jfarthing84 6 years ago.

Download all attachments as: .zip

Change History (18)

comment:1 @stevejohnson6 years ago

  • Priority changed from normal to lowest
  • Severity changed from normal to trivial

I've been told that since this code section has been the same for over 2 years, it's probably best to leave it alone as changing it might break something that relies on this behavior.

comment:2 @DD326 years ago

Another thing: Everything is stored as a string (except arrays and objects), So passing (int)1 in, would result in (string)1 out, (bool)true results in (string)1 coming out.

I've found (This was with options, but the same holds for meta's) that wraping everything in an array is best array( (int)1 ); results in array( (int)1 ); coming out.

comment:3 @jacobsantos6 years ago

  • Milestone changed from 2.8 to 2.9

Wish for the best of luck on creating a patch for this.

comment:4 follow-up: @stevejohnson6 years ago

Is it worth creating a patch for? If this is the way it's supposed to act, the ticket should be closed, yes?

comment:5 in reply to: ↑ 4 @jacobsantos6 years ago

Replying to stevejohnson:

Is it worth creating a patch for? If this is the way it's supposed to act, the ticket should be closed, yes?

I do not feel that is for me to decide. If you wish to close it, then please do so.

comment:6 @stevejohnson6 years ago

  • Resolution set to invalid
  • Status changed from new to closed

That was more of an "I'm new here, what should I do?" question.


comment:7 @jacobsantos6 years ago

  • Milestone 2.9 deleted

PS: please remember to reset the milestone as well.

comment:8 @jfarthing846 years ago

  • Priority changed from lowest to normal
  • Resolution invalid deleted
  • Severity changed from trivial to normal
  • Status changed from closed to reopened

I believe this should be re-opened because, although the original poster's method as storing the value as an array may work, it creates unnecessary storage amounts storing a serialized array for one single '0'.

@jfarthing846 years ago


comment:9 @jfarthing846 years ago

  • Keywords has_patch added

comment:10 @jfarthing846 years ago

  • Owner changed from anonymous to jfarthing84
  • Status changed from reopened to accepted

comment:11 @jfarthing846 years ago

  • Version changed from 2.7 to 2.8.4

comment:12 @jfarthing846 years ago

  • Cc jfarthing84 added

comment:13 @scribu6 years ago

Related: #10837.

comment:15 @jfarthing845 years ago

  • Cc jfarthing84 removed

comment:16 @jfarthing845 years ago

  • Cc jeff@… added

comment:17 @nacin5 years ago

  • Keywords has_patch removed
  • Resolution set to wontfix
  • Status changed from accepted to closed

The _usermeta API is now deprecated in favor of _user_meta, which uses our generic _metadata functions.

Thus, closing as wontfix for _usermeta. And we're now standardized across all meta APIs.

Note: See TracTickets for help on using tickets.