Last modified: 2012-11-05 17:58:59 UTC

Wikimedia Bugzilla is closed!

Wikimedia migrated from Bugzilla to Phabricator. Bug reports are handled in Wikimedia Phabricator.
This static website is read-only and for historical purposes. It is not possible to log in and except for displaying bug reports and their history, links might be broken. See T43336, the corresponding Phabricator task for complete and up-to-date bug report information.
Bug 41336 - Click tracking no associating valid rev_id with some edit_success events.
Click tracking no associating valid rev_id with some edit_success events.
Status: RESOLVED FIXED
Product: MediaWiki extensions
Classification: Unclassified
ArticleFeedbackv5 (Other open bugs)
unspecified
All All
: Highest normal (vote)
: ---
Assigned To: Nobody - You can work on this!
:
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2012-10-23 23:13 UTC by Aaron Halfaker
Modified: 2012-11-05 17:58 UTC (History)
3 users (show)

See Also:
Web browser: ---
Mobile Platform: ---
Assignee Huggle Beta Tester: ---


Attachments

Description Aaron Halfaker 2012-10-23 23:13:15 UTC
Relevant to #40775

I'm running into edit_success events that do not have a valid revision_id associated with them.  For example, at the end of clicktracking.log-20121022.gz.  

$ zcat clicktracking.log-20121022.gz | grep "cta_edit-edit_success" | tail -n1
enwiki ext.articleFeedbackv5@10-option6X-cta_edit-edit_success	20121022062446	0	sE35RgbXXhnXHM5rluqKSK4QOCe8ik3w	0	0	0	0	0	Dowry system in India|0

Notice the "Dowry system in India|0" at the end of the event.  This portion of the event should contain <title>|<new revision id>, but the <new revision id> == 0.  

This bug appears to affect 3731/69933 edit_success events that were captured in my sample.  This wouldn't be a big deal, but it affects a full 1/3rd of the cta_edt-edit_success events (668/1694),
Comment 1 Aaron Halfaker 2012-10-23 23:26:03 UTC
I did a little bit more digging to find that event I referenced above does not have a revision reflected in the history of the article "Dowry system in India".  

http://en.wikipedia.org/w/index.php?title=Dowry_system_in_India&action=history

Currently, the above history page shows that the last revision was saved at 2012-10-14 10:22.  In the log, the edit success event has the associated timestamp of 2012-10-22 06:24, nearly a week later.  

The enwiki database confirms this:
> select max(rev_timestamp) from revision where rev_page = 31999935;
+--------------------+
| max(rev_timestamp) |
+--------------------+
| 20121014102253     |
+--------------------+
1 row in set (0.11 sec)
Comment 2 Dario Taraborelli 2012-10-24 00:25:53 UTC
(copying from thread) this is a known issue (not necessarily a bug): if you try and save an empty edit, i.e. submit an edit with no modification, AFT still logs this as an edit_success but since MW doesn't generate a rev_id as a result of this transaction, a 0 value is stored to flag this as a "silent fail". Ideally we wouldn't even log these events as they are potentially inflating edit success counts, in the past we've just discarded them when processing the log data.
Comment 3 Andre Klapper 2012-10-24 00:32:45 UTC
Wondering how severe this is. Should this block the next deployment? Fabrice?
Comment 4 Matthias Mullie 2012-10-24 11:47:48 UTC
I've pushed a patchset that will no longer register such no-modification submissions as "edit_success", but will now register then as "edit_norevision".

The change is part of https://gerrit.wikimedia.org/r/#/c/29766/

Note You need to log in before you can comment on or make changes to this bug.


Navigation
Links