Abuse filter management

From EncyclopAtys

Abuse Filter navigation (Home | Recent filter changes | Examine past edits | Abuse log)
Jump to: navigation, search

Back to filter editor | Back to filter history

Differences between versions

ItemVersion from 12:09, 16 April 2012 by SirCotareVersion from 12:32, 16 April 2012 by SirCotare
Basic information

Description:

(publicly viewable)
 
New user blanking articles
 
New user blanking articles
Flags:
 
Enabled
 
Enabled
Notes:
 
Should there be a note in the warning about what to do if you're a BLP subject attempting to blank your own article?
 
Should there be a note in the warning about what to do if you're a BLP subject attempting to blank your own article?
  
 
Also, yay, first comment! - east
 
Also, yay, first comment! - east
  
 
These notes are not private, by the way.
 
These notes are not private, by the way.
  
 
The false positives page has not been created yet, so the warning page is incomplete. I deactivated the action 'warn'.
 
The false positives page has not been created yet, so the warning page is incomplete. I deactivated the action 'warn'.
  
 
Re-enabled after creating it -- Andrew
 
Re-enabled after creating it -- Andrew
  
 
We really need a better way to talk. Do tildes work here? ~~~ (east)
 
We really need a better way to talk. Do tildes work here? ~~~ (east)
 
Guess not :'(
 
Guess not :'(
  
 
Ignore if editor is the only editor of the page (faked by checking last 10 editors)
 
Ignore if editor is the only editor of the page (faked by checking last 10 editors)
  
 
Redefine blanking as any size under 50, where the page started at >500 (min delta of 450)
 
Redefine blanking as any size under 50, where the page started at >500 (min delta of 450)
  
 
Logic fixes, some of the conditions were redundant - Hersfold
 
Logic fixes, some of the conditions were redundant - Hersfold
  
 
Note that filter applies only to new users. (eg non autoconfirmed but new user is shorter)
 
Note that filter applies only to new users. (eg non autoconfirmed but new user is shorter)
  
 
Optimizing - Hersfold
 
Optimizing - Hersfold
  
 
Add redirect avoidance, necessary if this is going to allow pages to not be fully blank.  - DF
 
Add redirect avoidance, necessary if this is going to allow pages to not be fully blank.  - DF
  
 
changed "user_name in article_recent_contributors" to "!(user_name in article_recent_contributors)" - Hersfold
 
changed "user_name in article_recent_contributors" to "!(user_name in article_recent_contributors)" - Hersfold
  
 
Adding & !(user_name == 69.226.103.13) to temporarily exempt that IP address from this filter, see my talk page for details. - Hersfold
 
Adding & !(user_name == 69.226.103.13) to temporarily exempt that IP address from this filter, see my talk page for details. - Hersfold
 
Scratch that, removing per Prodego's comments on filter 30. If it got removed there, I shouldn't be adding it here. - Hersfold
 
Scratch that, removing per Prodego's comments on filter 30. If it got removed there, I shouldn't be adding it here. - Hersfold
 
I added an exception for {{db-attack}} template. - Ruslik
 
I added an exception for {{db-attack}} template. - Ruslik
 
I added an exception for {{db-g10}} template. - Od Mishehu
 
I added an exception for {{db-g10}} template. - Od Mishehu
 
Optimization of the code. - Ruslik
 
Optimization of the code. - Ruslik
 
Moved new size check before namespace check - rationale very few edits result in size < 50, most are to namespace 0.  Moved old size before namespace check - rationale _given_ new size< 50 very few edits will have oldsize > 500 (especially by new users), however most will be to mainspace.  Previous to change average run time 6.52ms, 27 conditions. RF
 
Moved new size check before namespace check - rationale very few edits result in size < 50, most are to namespace 0.  Moved old size before namespace check - rationale _given_ new size< 50 very few edits will have oldsize > 500 (especially by new users), however most will be to mainspace.  Previous to change average run time 6.52ms, 27 conditions. RF
 
OK condition stats seem meaningless varying from 8 to 500 I'll review in 10,000 edits time.
 
OK condition stats seem meaningless varying from 8 to 500 I'll review in 10,000 edits time.
 
Run 5.97 conditions 14, but I don't trust them. RF.
 
Run 5.97 conditions 14, but I don't trust them. RF.
  
 
Now monitoring namespace 100 (portal) Shirik 12 Feb 2010
 
Now monitoring namespace 100 (portal) Shirik 12 Feb 2010
  
 
Optimize --Tim
 
Optimize --Tim
  
 
Adding Wiktionary template to prevent false positives. -v^_^v
 
Adding Wiktionary template to prevent false positives. -v^_^v
  
 
Optimize using regex. - KoH
 
Optimize using regex. - KoH
 
Portals are simply not worth the runtime. - KoH
 
Portals are simply not worth the runtime. - KoH
Filter conditions
Conditions:
 
!"autoconfirmed" in user_groups
 
!"autoconfirmed" in user_groups
 
& new_size < 50
 
& new_size < 50
 
& old_size > 500
 
& old_size > 500
 
& article_namespace == 0
 
& article_namespace == 0
 
& !user_name in article_recent_contributors
 
& !user_name in article_recent_contributors
 
& !(lcase(added_lines) rlike "#redirect|{{((db-(attack|g10))|wi)}}")
 
& !(lcase(added_lines) rlike "#redirect|{{((db-(attack|g10))|wi)}}")
Actions to take when matched
Actions to take when matched
 
Tag: blanking
 
Tag: blanking
Warn: abusefilter-warning-blanking
+
Warn: abusefilter-warning