Jump to content

Edit filter configuration

Differences between versions

ItemVersion from 18:56, 25 December 2012 by Sole SoulVersion from 17:09, 24 July 2015 by Rich Farmbrough
Basic information
Notes:
A long string of letters and/or numbers may be indicative of vandalism. - KoH
A long string of letters and/or numbers may be indicative of vandalism. - KoH


FP: Adding links. - KoH
FP: Adding links. - KoH


Made the string length allowed before tagging longer, as it's catching URLs..  -Versageek
Made the string length allowed before tagging longer, as it's catching URLs..  -Versageek


Hope this check for http is better... - KoH
Hope this check for http is better... - KoH


Also check https. - KoH
Also check https. - KoH


Virtually everything triggering this hits another filter, and this uses a lot of conditions, disabled, consider merging in to 135. -Prodego
Virtually everything triggering this hits another filter, and this uses a lot of conditions, disabled, consider merging in to 135. -Prodego
-ok -P
-ok -P
Turn to disallow. -Sole Soul
Turn to disallow. -Sole Soul


I can think of legitimate uses for 50+ character strings. From experience, in a warn-tag filter, a warning already blocks out over 75% of the vandalism. - KoH
I can think of legitimate uses for 50+ character strings. From experience, in a warn-tag filter, a warning already blocks out over 75% of the vandalism. - KoH


False positives are associated with the word "word". -Sole Soul
False positives are associated with the word "word". -Sole Soul
Optimize for conditions. RF 20150724
Filter conditions
Conditions:
(documentation)
!("autoconfirmed" in user_groups) &
!"autoconfirmed" in user_groups & (
(article_namespace == 0) &
  article_namespace == 0 & (
(added_lines rlike "[a-zA-Z0-9]{50,}") &
    added_lines rlike "[a-zA-Z0-9]{50,}" & (
!(removed_lines rlike "[a-zA-Z0-9]{50,}") &
      !removed_lines rlike "[a-zA-Z0-9]{50,}" & (
!(added_lines rlike "https?://[^ ]*?[a-zA-Z0-9]{50,}")
        !added_lines rlike "https?://[^ ]*?[a-zA-Z0-9]{50,}" &
& !(lcase(added_lines) rlike "word|thumb|file|sequence")
        !lcase(added_lines) rlike "word|thumb|file|sequence"
      )
    )
  )
)