No reliable way to detect when an API edit is disallowed by the abuse filter
Open, NormalPublic


I set up an abuse filter on testwiki that warns the user with the MediaWiki message "snorkel-abuse" and also disallows the action. When triggering this filter using the API (action=edit) I get this:

<?xml version="1.0"?><api><edit code="snorkel-abuse" info="Hit AbuseFilter: TTO&#039;s test filter" warning="Don&#039;t create snorkel pages" result="Failure" /></api>

The only way to test for AbuseFilter being hit appears to be to check whether the "info" parameter starts with the string "Hit AbuseFilter". This seems a bit hackish.

The spam blacklist manages to do better than this, providing its own spamblacklist attribute on the <edit> element. AbuseFilter should be able to do better too.

Version: unspecified
Severity: normal


bzimport raised the priority of this task from to Normal.
bzimport set Reference to bz55460.
bzimport added a subscriber: Unknown Object (MLST).
TTO created this task.Oct 8 2013, 8:45 AM
werdna removed a subscriber: werdna.Dec 10 2014, 5:27 PM
TTO set Security to None.
TTO removed a subscriber: wikibugs-l-list.
Restricted Application added a subscriber: Aklapper. · View Herald TranscriptFeb 10 2016, 4:39 AM
Jay8g added a subscriber: Jay8g.Apr 4 2016, 5:15 AM

I solved a similar issue for TitleBlacklist (T115258) a while ago with 97b433286dc086ea5261b1b88736600b5454887f. The same approach should work here.

Krinkle added a subscriber: Krinkle.