User talk:Beetstra
From Meta, a Wikimedia project coordination wiki
 |
Dirk is taking a short wikibreak and will be back on Meta soon. |
Archives:
[edit] COIBot not reporting in
Leaving you a note here since I'm probably not irc tomorrow. Coibot seems to hang regularly in saving linkreports, I did one restart this afternoon but now it seeems to be hanging again... I'll restart it now, but you'll probably want to check Special:Contributions/COIBot when (if?) you check in tomorrow. Regards, Finn Rindahl 00:39, 28 March 2009 (UTC)
- I have to write a bit better logging for the bot so I can finally figure out where it hangs. I as yet don't understand. --Dirk Beetstra T C (en: U, T) 12:06, 30 March 2009 (UTC)
[edit] User:COIBot/XWiki
Could you leave cells blank instead of "ND" where applicable? I think this will allow the table to be sorted as numbers instead of text, which would be useful, I think. — Mike.lifeguard | @en.wb 01:57, 1 April 2009 (UTC)
- OK, I'll try that. --Dirk Beetstra T C (en: U, T) 10:23, 1 April 2009 (UTC)
[edit] Looping/hanging of the LinkSaver
Seems it got itself into a bit of a loop regarding User:COIBot/XWiki/fuck this one... Finn Rindahl 23:02, 7 April 2009 (UTC)
- This is the same problem as the hanging of the LinkSaver. I built in a catch for cases where the LinkSaver can't save a link for some reason, where the link is put back into the database for another attempt (e.g. if there was a timeout or the wiki was down at that moment, as that might result in loss of reports). However, this did not catch the cases where the bot, for whichever reason, REALLY can't save the report for other reasons (blacklisting of a link, title-blacklist, impossible names for the page, etc.).
- I patched it, it now should do three attempts to save a link, if it fails, it deletes it (but restarting the loop of three until the link is not returned into the database for other reasons, like ongoing spamming, poking or IRC-report-requesting). Error messages regarding this go to #wikimedia-external-links. --Dirk Beetstra T C (en: U, T) 10:06, 8 April 2009 (UTC)
[edit] coibot report swmt domain.com
The SWMTBots got renamed... could we use CVNBot9 now, instead of SWMTBot1? — Mike.lifeguard | @en.wb 03:23, 11 April 2009 (UTC)
- It's now a setting in User:COIBot/Settings. Hence: {{solvified}}. --Dirk Beetstra T C (en: U, T) 10:36, 11 April 2009 (UTC)
[edit] broken regexes?
Broken regex \bbr\.geocities\.com\ (perl-corrected: \bbr\.geocities\.com\) on pt.wikiquote.org blacklist, error: Trailing \ in regex m/\bbr\.geocities\.com\/ at LinkSaver.pl line 2041, <LIST> line 853.
Something like this appears on a lot (all?) of reports. Is that normal? — Mike.lifeguard | @en.wb 23:09, 20 April 2009 (UTC)
- Well .. no, it is not normal. I have contacted a local admin to repair it (as the regex is broken!), but to no avail. Maybe I should poke somewhere someone again. --Dirk Beetstra T C (en: U, T) 11:29, 21 April 2009 (UTC)
- See q:pt:Usuário_Discussão:Chico#Question. --Dirk Beetstra T C (en: U, T) 11:31, 21 April 2009 (UTC)
[edit] COIBot rights command
Hello Beetstra, I've found a bug on the COIBot rights [...] , here is the source: [10:20] <MrDferg> rights Dferg
[10:20] <COIBot> Dferg is sysop on w:es, meta; bureaucrat on w:es; rollbacker on w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en, w:en I think the bug is only in the rollbacker group cause I checked Drini and: [10:21] <COIBot> Drini is sysop on w:en, w:es, w:nah, n:es, q:es, meta, commons; bureaucrat on w:es, w:nah, n:es, meta; steward on meta; checkuser on w:es, q:es, meta, commons; editor on w:als Best regards, —Dferg (talk) 08:20, 24 April 2009 (UTC)
- Yep .. it was one of the things that the LinkSaver was hanging on (this was a secondary task of that part of the bot, when it did not have to save linkreports). I have separated it from that bot now .. annoying, I'll have to write something special for this. Don't know though why it duplicated. Thanks for reporting. --Dirk Beetstra T C (en: U, T) 11:18, 24 April 2009 (UTC)
[edit] q:pt
--Chico 22:22, 24 April 2009 (UTC)
- Gracias! --Dirk Beetstra T C (en: U, T) 10:43, 25 April 2009 (UTC)
[edit] coibot has left the building
or #external-links at least, guess you're the only one who can start it again.... Finn Rindahl 15:54, 6 May 2009 (UTC)
- Heh, most bots left the building. Most reconnected after the, what I assume was, a netsplit, others have been restarted. It is one of the ways of getting the bots 'down' (though most still run, and still work in the background, we only lost the IRC tools .. ). --Dirk Beetstra T C (en: U, T) 15:18, 7 May 2009 (UTC)
[edit] COIBot stats
Hi!
COIBot does log (almost) all link additions, doesn't it? So are there somewhere statistics about the number of link additions for example in w:de?
related: de:WP:FZW#Externe_Links. -- seth 10:38, 10 May 2009 (UTC)
- I could write a command for that in the Commander. Yes, in principle the linkwatchers log everything, except for some hard-ignored links (xxx.wikipedia.org/xxx.wiktionary.org, say, the internal links which look like external links). --Dirk Beetstra T C (en: U, T) 11:46, 10 May 2009 (UTC)
-
- Ok, thanks, that would be a nice gimmick to have, although it won't really help us in anti spam fighting. -- seth 15:38, 10 May 2009 (UTC)
- Indeed, mostly a gimmick, but it may help making general statements in guidelines and policies. --Dirk Beetstra T C (en: U, T) 20:49, 10 May 2009 (UTC)
[edit] I had...
...to ban-forward (from cvn-wp-es and -external-links) your bot to #wikimedia-overflow because it is part/joining constantly the channels. Please unban it from #wikimedia-external-links when the connection of the bot is fixed. Regards and sorry, —Dferg (talk) 11:07, 28 May 2009 (UTC)
- Dei Beetstra, ter ynformaasje: itselde is niiskrekt dien op #cvn-sw troch Kylu, wêr't itselde oan 'e gong wie. Groetenis, Wutsje 12:11, 28 May 2009 (UTC)
- Nothing in the logs, nothing to see. No clue what happened, but ping me or Versageek, we both can kill it and restart it if it happens again. Thanks! --Dirk Beetstra T C (en: U, T) 10:03, 29 May 2009 (UTC)
[edit] COIBot reports
Hello Beetstra, please correct me if I'm wrong, but, aren't COIBot reports suposed not to be indexed by search engines? (see diff of Robots.txt). I saw that sometimes those reports appears in Google results. Would be possible for COIBot to add {{NOINDEX}} to every new report it creates? Thank you. — df| 15:30, 4 August 2009 (UTC)
- No, you are correct. Thing is that, already before that edit to the Robots.txt, COIBot is adding {{NOINDEX}} to all his reports (I added that because of the complaints, while we were still waiting for the 'global' noindex). I am however curious to see a couple of those reports which show up on Google, somewhere must be something wrong. --Dirk Beetstra T C (en: U, T) 19:04, 4 August 2009 (UTC)
- Hello Beetstra, thanks for your reply. I asked because I've found the reply on Talk:Spam_blacklist#www.travel2macedonia.com.mk strange. I quote: [...]The Google hit (ranked 12 in my search) links to the COIBot report[...].
- I checked this on my own and I have found that it is true, at least for User:COIBot/LinkReports/travel2macedonia.com.mk and I supose it is not the unique. Perhaps Google is ignoring the {{NOINDEX}}es tags? Thanks for your time. — df| 20:41, 4 August 2009 (UTC)
- Now I don't know. Maybe we should either first discuss it on the spam blacklist, or maybe immediately a bug-report? It would be really bad if Google is ignoring the noindex. --Dirk Beetstra T C (en: U, T) 06:07, 6 August 2009 (UTC)
- I'll post this discussion on the discussion section of the SBL so that other admins/experienced users can share their opinions. Thank you, — df| 10:58, 6 August 2009 (UTC) P.D.: Talk:Spam_blacklist#COIBot_reports_showing_up_in_Google_results — df| 11:06, 6 August 2009 (UTC)
[edit] Category:Open Local reports for es.wikipedia.org
Hello. Will it be possible for COIBot to close all the reports on that category that are older than 2 weeks? Thanks, —Dferg (disputatio) 15:52, 17 November 2009 (UTC)
- I am regenerating, that will clean the old ones. --Dirk Beetstra T C (en: U, T) 20:34, 21 November 2009 (UTC)
[edit] Huge resource consumption by your bot
Hi, did you know that your bot produced 217775 edits of avg. 95732 bytes for a total of 19 Gb raw data. Is meta the right place for this kind of data? Erik Zachte 03:04, 20 December 2009 (UTC) http://stats.wikimedia.org/EN/TableRankArticleHistoryByArchiveSize.html
- Eh. Wow. Well, yes, the problem is, where else? Meta is the place where the global spam blacklist is located, and that is also where the bot reports it's reports. I could consider localising a subset of it, but I think that it will still be by far the largest editor then anyway. I must say, we are working on another solution, but that will take time. --Dirk Beetstra T C (en: U, T) 15:05, 20 December 2009 (UTC)
- Are these operational data where older history is irrelevant ? (I have no idea). I don't assume you could autodelete obsolete info with a bot? Probably not, if bots could do that that would be very risky. Maybe you could start a new page each month, and file pages older than x months for deletion? But then I see you have many satelite pages. Maybe you could host a separate wiki? Or we could discuss a separate WMF wiki for operational data like this, where auto delete is allowed. Besides general concern my worry is that dump creation gets more and more difficult when more and more bots use this online storage scheme. (Ronald also does this for proxy data) Erik Zachte 18:14, 21 December 2009 (UTC)
