- We've been ignoring that for too long .. the list is too long, takes too long to sort and hence time outs on different parts - main bot procedure thinks things are hanging and kills the process. --Dirk Beetstra T C (en: U, T) 12:52, 19 December 2017 (UTC)
Hi. mw: is being for some time getting targeted by spambots. If not already, could it be possible to monitor its local (black|whitelists) and IRC feed? Thanks! —MarcoAurelio (talk) 10:45, 6 February 2018 (UTC)
- @MarcoAurelio: not sure if it is, I will check when I have access to putty. --Dirk Beetstra T C (en: U, T) 12:11, 6 February 2018 (UTC)
COIBot significant backlog ...
- Hmm, there were only 2 running (should be 6). I've added 4, will take some time to catch up .. --Dirk Beetstra T C (en: U, T) 13:22, 15 February 2018 (UTC)
- Thanks. — billinghurst sDrewth 00:17, 18 March 2018 (UTC)
- @SDrewth: kicked them on the server. No clue what happened. --Dirk Beetstra T C (en: U, T) 06:48, 18 March 2018 (UTC)
User encoding problems on lisankom report
When creating User:COIBot/LinkReports/lisankom.com COIBot had some problems encoding a couple of user names (encoded twice as utf-8?). These are shown in the wiki page as “Ø³Ø¹ÙØ¯ Ø¹ÙØ³Ù Ø§ÙÙ” and “Ø¹Ø²ÙØ²Ù”, although they seem to have been by MediaWiki, from the recent changes, it seems to have been using “Ø³Ø¹Ù<U+008A>Ø¯_Ø¹Ù<U+008A>Ø³Ù<U+0089>_Ø§Ù<U+0084>Ù<U+0085>Ø¹Ø²Ù<U+0088>Ø²Ù<U+008A>“ (hexadecimal: c3 98 c2 b3 c3 98 c2 b9 c3 99 c2 8a c3 98 c2 af 5f c3 98 c2 b9 c3 99 c2 8a c3 98 c2 b3 c3 99 c2 89 5f c3 98 c2 a7 c3 99 c2 84 c3 99).
U+0084, U+0085, U+0088, U+0089, U+008A are control characters (formerly known as INDEX, NEXT LINE, CHARACTER TABULATION SET, CHARACTER TABULATION WITH JUSTIFICATION, LINE TABULATION SET) and, -not only are forbidden in MediaWiki usernames- but are also problematic to insert in a text.
In case it helps, attempting to decode the string from COIBot fails (illegal input sequence), but encoding in utf-8 the already-in-utf-8 bytes of the reported user (سعيد عيسى المعزوزي) leads to the following output similar to what COIBot provided:
- COIBot: c3 98 c2 b3 c3 98 c2 b9 c3 99 c2 8a c3 98 c2 af 5f c3 98 c2 b9 c3 99 c2 8a c3 98 c2 b3 c3 99 c2 89 5f c3 98 c2 a7 c3 99 c2 84 c3 99
- Output: c3 98 c2 b3 c3 98 c2 b9 c3 99 c5 a0 c3 98 c2 af 5f c3 98 c2 b9 c3 99 c5 a0 c3 98 c2 b3 c3 99 e2 80 b0 5f c3 98 c2 a7 c3 99 e2 80 9e c3 99 e2 80 a6 c3 98 c2 b9 c3 98 c2 b2 c3 99 cb 86 c3 98 c2 b2 c3 99 c5
- @Platonides: Yes, I know it seems to have problems there. I though I forgot to en/decode somewhere .. but maybe it is a double encoding. It does seem to have the same problem with the domains (report names) sometimes. I'll try and dig through the code once more, see what happens. Problem is that the console (putty) does not seem to understand the encoding either, and throws garbage at me, and perl has a nasty tendency to convert stuff when it shouldn't - which makes it difficult to check where it goes wrong. But it might very well be that I encode the name-string, and then finally encode the whole text again. Thanks for the analysis! --Dirk Beetstra T C (en: U, T) 07:54, 10 April 2018 (UTC)
Hi. COIBot is hanging around, though its mates the LiWas are missing. When you are next around and available, it would be great if you could kick whatever needs kicking. Thx. — billinghurst sDrewth 11:49, 18 May 2018 (UTC)