Page MenuHomePhabricator
Create Task
Maniphest T121338

Better handling of revision deletion on import
Closed, ResolvedPublic
Actions

Description

Currently our Import process doesn't handle revision deletion at all, even though the relevant data is exported into the XML format.

This has the following effects for revisions:

Before exportAfter import
Username hiddenThe revision is skipped completely
Edit summary hiddenBlank edit summary
Content hiddenBlank revision (no content)

To me, the bottom two behaviours seem acceptable. We shouldn't skip revisions for which the username is hidden, though. The suppressed username should be replaced by user ID 0 and a user_text such as "Unknown user".

For log entries, any kind of selective deletion seems to just blow up WikiImporter; see T34876.

Details

SubjectRepoBranchLines +/-
Import: Properly handle deleted usernames in XML dumpsmediawiki/coremaster+6 -2
Handle missing titles and usernames when importing log itemsmediawiki/coremaster+43 -10
Customize query in gerrit

Related Objects

Event Timeline

TTO created this task.Dec 13 2015, 4:41 AM
TTO raised the priority of this task from to Low.
TTO updated the task description. (Show Details)
TTO added projects: MediaWiki-Core-Snapshots, MediaWiki-Revision-deletion.
TTO subscribed.
Restricted Application added a subscriber: Aklapper. · View Herald TranscriptDec 13 2015, 4:41 AM
TTO assigned this task to Georggi1999.Dec 17 2015, 6:39 AM
gerritbot subscribed.Dec 18 2015, 9:15 AM

Change 259987 had a related patch set uploaded (by Georggi199):
Made possible to import dumps without usernames

https://gerrit.wikimedia.org/r/259987

gerritbot added a project: Patch-For-Review.Dec 18 2015, 9:15 AM
gerritbot added a comment.Dec 19 2015, 10:26 AM

Change 259987 merged by jenkins-bot:
Handle missing titles and usernames when importing log items

https://gerrit.wikimedia.org/r/259987

ReleaseTaggerBot added projects: MW-1.27-release-notes, MW-1.27-release (WMF-deploy-2016-01-12_(1.27.0-wmf.10)).Dec 19 2015, 11:00 AM
gerritbot added a comment.Dec 21 2015, 12:05 PM

Change 260349 had a related patch set uploaded (by Georggi199):
Fixed contributor and text handling in Import.php

https://gerrit.wikimedia.org/r/260349

gerritbot added a comment.Dec 21 2015, 1:00 PM

Change 260349 merged by jenkins-bot:
Import: Properly handle deleted usernames in XML dumps

https://gerrit.wikimedia.org/r/260349

TTO closed this task as Resolved.Dec 21 2015, 2:17 PM
TTO mentioned this in T26519: RevDel'd imported edits may be blanked.Dec 22 2015, 3:43 AM
matmarex mentioned this in T158474: User:Unknown_user editing in dewiki.Feb 18 2017, 5:38 PM
Alsee subscribed.Oct 21 2017, 6:02 PM

Note user:Unknown_user already existed on EnWiki (and only on EnWiki). It was created and briefly used for a few days in 2004. It has 126 edits. Any username should always be checked for existence before using it for any system purpose.

In 2017 this username was added to $wgReservedUsernames in T158474. The account exists, but it can no longer be logged into.

Content licensed under Creative Commons Attribution-ShareAlike (CC BY-SA) 4.0 unless otherwise noted; code licensed under GNU General Public License (GPL) 2.0 or later and other open source licenses. By using this site, you agree to the Terms of Use, Privacy Policy, and Code of Conduct. · Wikimedia Foundation · Privacy Policy · Code of Conduct · Terms of Use · Disclaimer · CC-BY-SA · GPL