Page MenuHomePhabricator

XML-Import of a page version prior to existing history spoils diff page?
Closed, InvalidPublic

Description

See http://bugzilla.wikimedia.org/show_bug.cgi?id=5438 for background info.

In order to prepare several 1000 xml files for being imported (name space prefixes in page titles, and contributor
user names need adjustment), I made few tests.

There was already a page http://ksh.wikipedia.org/wiki/User_talk:Sockpuppet with an edit history, all of today
(2006-Apr-22). I tried to import an older version of that page, which failed because there was an ASCII 0x1A (Ctrl/
Z) at the end of the file. Having it replaced by ASCII 0x0A (Ctrl/J, Line Feed), the import succeeded.

Next, I found the newly imported page listed twice in the article history.
Also, diff-ing against the imported page yielded unexpected results. Effectively, it refuses to show any
differences, even no traces of the referred page are to be seen, only data from the imported page shows.

Consequentially, we shall not import any page histories for the time being, where a page of that name already
exists. That is at most three dozen pages of thousands, i.e. no big deal.

Note that the kshwiki uses Mediawiki Version 1.7alpha (r13760) see http://ksh.wikipedia.org/wiki/Special:Version
, which is not currently selectable in the "Enter Bug" form.

If you need to run tests in the kshwiki, feel free to do so. Using pages from User:Sockpuppet's user- an talkpage-
namespace won't harm anyone, he has been created by http://ksh.wikipedia.org/wiki/User:Purodha for this knd of
testing.


Version: 1.7.x
Severity: normal
URL: http://ksh.wikipedia.org/w/index.php?title=User_talk:Sockpuppet&diff=prev&oldid=1834

Details

Reference
bz5689

Event Timeline

bzimport raised the priority of this task from to Low.Nov 21 2014, 9:12 PM
bzimport set Reference to bz5689.
bzimport added a subscriber: Unknown Object (MLST).

The imported xml data file.

See initial bug description.
True file format is not allowed to be entered manually:
Content-Type: text/xml; charset=utf-8

Attached:

It shows twice because you imported the same file twice.

Diff looks fine.