Page MenuHomePhabricator

page history shows unreviewed revisions, reverts in wrong order
Open, LowPublic

Description

The vandalism https://de.wikipedia.org/w/index.php?title=Myokardinfarkt&oldid=110726356
was reverted in https://de.wikipedia.org/w/index.php?title=Myokardinfarkt&oldid=110726365

The history contains these two versions in reverse order https://de.wikipedia.org/w/index.php?title=Myokardinfarkt&offset=20121127000000&action=history&limit=4

(Aktuell | Vorherige) 16:19, 20. Nov. 2012‎ 178.2.15.57 (Diskussion)‎ . . (61.254 Byte) (+7)‎ . . (→‎Epidemiologie) (rückgängig) [Markierung ausstehend]
(Aktuell | Vorherige) 16:19, 20. Nov. 2012‎ D (Diskussion | Beiträge)‎ K . . (61.247 Byte) (-7)‎ . . (Änderungen von 178.2.15.57 (Diskussion) wurden auf die letzte Version von Buchexperte zurückgesetzt) (rückgängig) [automatisch gesichtet]
(Aktuell | Vorherige) 12:45, 20. Nov. 2012‎ Buchexperte (Diskussion | Beiträge)‎ K . . (61.247 Byte) (-5)‎ . . (Fix Überschriftenstruktur) (rückgängig) [automatisch gesichtet]


See Also:
https://phabricator.wikimedia.org/T4930

Details

Reference
bz42296

Event Timeline

bzimport raised the priority of this task from to Low.Nov 22 2014, 12:48 AM
bzimport set Reference to bz42296.
bzimport added a subscriber: Unknown Object (MLST).

Looks like a FlaggedRevs problem, unreviewed revs are always shown at the top and I guess FR just didn't 178.2.15.57's revision as reverted for some reason.

Looking at the seconds:

  1. Nov. 2012, 16:19:34‎ 178.2.15.57
  2. Nov. 2012, 16:19:32‎ D

But that means the inserting of the revision from 178.2.15.57 was done 2 seconds after the revert, but without the revision, you cannot revert ??

Looking at the revids:

oldid=110726356 178.2.15.57
oldid=110726365 D

That means, the revision id of the ip was created earlier than the revert. Maybe User:D is using the IRC Feed, which was sent, but the revision was not ready inserted? Or the RecentChanges were faster than the revision table?

sumanah wrote:

Has this been reproduced again within the last few weeks (in the current deployment cycle)?

Code problem is very likely a dup of bug 2930, the instance where this happened could block bug 16660.

Can someone make a database query to count how often such a problem exist?

Aklapper renamed this task from page history shows unreviewed revisions, reversions in wrong order to page history shows unreviewed revisions, reverts in wrong order.Nov 7 2020, 5:18 PM