Page MenuHomePhabricator

Duplicate revert notification received
Closed, DeclinedPublic

Description

@Redrose64 received a duplicate revert notification for this edit.

Reported here.
(keywords for later searching: repeat repeated double twice)

Event Timeline

Restricted Application added subscribers: StudiesWorld, Aklapper. · View Herald Transcript

Strange. It looks like the event was inserted twice with identical data. I wonder if there was a timeout after which insertion was retried, or something like that.

The two events are timestamped 20:29:24 and 20:29:27. The revision timestamp is 20:29:24.

*************************** 1. row ***************************
                        event_id: NNNNNN38
                      event_type: reverted
                   event_variant: NULL
                  event_agent_id: 424023
                  event_agent_ip: NULL
            event_page_namespace: NULL
                event_page_title: NULL
                     event_extra: a:5:{s:5:"revid";i:707425870;s:16:"reverted-user-id";i:9612106;s:20:"reverted-revision-id";s:9:"706986494";s:6:"method";s:4:"undo";s:7:"summary";s:199:"Undid revision 706986494 by [[Special:Contributions/Redrose64|Redrose64]] ([[User talk:Redrose64|talk]]) See talk and the Hammersmith & City Line article. Service terminated at Whitechapel until 1941";}
                   event_page_id: 373502
              notification_event: NNNNNN38
               notification_user: 9612106
          notification_timestamp: 20160228202924
     notification_read_timestamp: 20160228NNNNNN
        notification_bundle_base: 1
        notification_bundle_hash: 
notification_bundle_display_hash: 
*************************** 5. row ***************************
                        event_id: NNNNNN39
                      event_type: reverted
                   event_variant: NULL
                  event_agent_id: 424023
                  event_agent_ip: NULL
            event_page_namespace: NULL
                event_page_title: NULL
                     event_extra: a:5:{s:5:"revid";i:707425870;s:16:"reverted-user-id";i:9612106;s:20:"reverted-revision-id";s:9:"706986494";s:6:"method";s:4:"undo";s:7:"summary";s:199:"Undid revision 706986494 by [[Special:Contributions/Redrose64|Redrose64]] ([[User talk:Redrose64|talk]]) See talk and the Hammersmith & City Line article. Service terminated at Whitechapel until 1941";}
                   event_page_id: 373502
              notification_event: NNNNNN39
               notification_user: 9612106
          notification_timestamp: 20160228202927
     notification_read_timestamp: 20160228NNNNNN
        notification_bundle_base: 1
        notification_bundle_hash: 
notification_bundle_display_hash:

I can't find any evidence of DB errors in the log (I read exception.log for the minute between 2016-02-28 20:29:00 and 20:30:00). Instead, I think this might be another occurrence of the null edit bug (T123843: Making a null edit triggers a new message notification although null edit is not recorded in page history): if the reverter pressed Save twice, the second save could have been processed as a null edit.

matmarex subscribed.

I don't think it's realistic to debug this 5 years later, and I haven't seen any more recent reports of this bug. If this was caused by T123843 as @Catrope guessed, then it should be fixed.