[08:09:46] We are experiencing a high amount of spam vandalism at the Basque Wikipedia, with random insertion of youtube videos in articles. All of them follow the same pattern: [08:09:47] VIDEO URL [08:09:49] VIDEO URL [08:09:50] VIDEO URL [08:09:52] Is there a way to have a spam blacklist that detects the consecutive insertion of the same video in different lines? I have tried with some regex but it doesn't seem to work. [14:35:19] You can match exact copies of a specific pattern with regular expressions using a capture group reuse pattern. The pattern `\1` will automatically match the first capture group, `\2` will match the second, etc. You're probably look for something like `added_lines rlike "(https:\/\/.+?)\n\n\1\n\n\1"` (which will match the same URL three times, separated by [14:35:19] newlines). Sample filter [14:35:20] here (https://test.wikipedia.org/wiki/Special:AbuseFilter/299), sample edit here (https://test.wikipedia.org/wiki/Special:AbuseLog/116540). Feel free to narrow it down to, say, YouTube links only. (re @Galder: We are experiencing a high amount of spam vandalism at the Basque Wikipedia, with random insertion of youtube videos in articles...) [14:51:58] if you do that YouTube has multiple domain names... mdot, regular, youtu(.)be. (re @ChlodAlejandro: You can match exact copies of a specific pattern with regular expressions using a capture group reuse pattern. The pattern \1 wi...) [14:52:59] you could repeat the domain matching for each group occurrence and capture only the path. but even then query string could vary. [14:53:39] have they discovered Rick roll? [14:54:17] yeah but maybe simple prevention of inserting the same link 3 times in a row (even if it is something else) could help right away. (re @jeremy_b: you could repeat the domain matching for each group occurrence and capture only the path. but even then query string could vary.) [14:54:39] yeah, if you filter this person too hard they just evade, but preventing the same link many times in the same edit would be broadly useful [14:55:49] the only concern here (as always with AF which I avoid for that reason) is collateral damage, usually my imagination is too poor to image what ELSE could be broken by the rule [14:56:14] start with report only? (re @saperski: the only concern here (as always with AF which I avoid for that reason) is collateral damage, usually my imagination is too poor...) [14:56:41] I'd rather say what bothers me that I never monitor the rules after X months once they are in place. [14:57:09] but for that quick kind of prevention Abuse Filter is probably the best tool [18:12:08] thanks! This seems to work. Let's see how long for the spamer to change the pattern (re @ChlodAlejandro: You can match exact copies of a specific pattern with regular expressions using a capture group reuse pattern. The pattern \1 wi...) [19:26:55] All movies & series available in PopcornTV ⭐️🤩 [19:26:56] Netflix & Disney Plus and more [19:26:58] Join now: https://astrurl.io/popcrntv [19:26:59] ⭐️🤩⭐️🤩⭐️🤩 [23:11:23] @albertoleoncio: It looks like that spammer got past our gate because they spammed before the bot had a chance to restrict them. Not sure if they were just really fast or if the bot had a hiccup. [23:16:24] bd808: It seems to be working normally. The spammer was just faster at joining, posting, and then leaving before the bot could act.