[00:03:44] Simetrical: around? [00:04:02] werdna, sure. [00:04:08] Simetrical: can I have your .git/config [00:04:13] werdna, why do you always ask if I'm around? [00:04:17] !ask | werdna [00:04:17] --mwbot-- werdna: Don't say "I have a question", or ask "Is anyone around?" or "Can anyone help?". Just ask the question, and someone will help you if they can. Also, please read < http://workaround.org/moin/GettingHelpOnIrc > for a good explanation of getting help on IRC. [00:04:18] :) [00:04:29] <_mary_kate_> any simetricals around? [00:04:32] I don't want to have to pull from your non-running git daemon every time I want to get some updates :) [00:04:39] and, for that matter, it'd be nice to be able to commit :p [00:04:50] are any simetricals around? I need a simetrical! [00:05:07] werdna, http://pastebin.ca/1254799 [00:05:11] !around | _mary_kate_ [00:05:11] --mwbot-- _mary_kate_: Are you around? [00:06:25] . . . [00:06:54] !around | Wegge [00:06:54] --mwbot-- Wegge: Are you around? [00:06:56] oops! [00:07:01] !around | werdna [00:07:01] --mwbot-- werdna: Are you around? [00:07:12] !around | mwbot [00:07:12] --mwbot-- mwbot: Are you around? [00:07:33] !stab AaronSchulz [00:07:33] --mwbot-- I don't know anything about "stab". You might try: !download !release [00:07:42] Zing! [00:07:44] !stab is ACTION stabs `e1 [00:07:44] --mwbot-- You don't have permission to do that. [00:07:48] ugh! [00:07:52] double zing! [00:08:02] !stab is ACTION stabs `e1 [00:08:02] --mwbot-- Successfully added keyword: stab [00:08:06] !stab AaronSchulz [00:08:06] --mwbot-- ACTION stabs AaronSchulz [00:08:09] dammit. [00:08:12] <_mary_kate_> werdna: you forgot the ^A [00:08:21] !stab is ^Astabs `e1 [00:08:21] --mwbot-- Keyword "stab" already exists. Use "!stab del" to remove it. [00:08:26] !stab del [00:08:26] --mwbot-- Successfully removed keyword: stab [00:08:28] !stab is ^Astabs `e1 [00:08:28] --mwbot-- Successfully added keyword: stab [00:08:30] !stab AaronSchulz [00:08:30] --mwbot-- ^Astabs AaronSchulz [00:08:31] <_mary_kate_> no, a literal control-A character [00:08:32] jesus fucking spam [00:08:39] <_mary_kate_> ^A^A [00:08:46] werdna: #mwbot Plz ^_^ [00:08:48] <_mary_kate_> where the ctcp content would be "ACTION whatever" [00:09:42] guess his client didn't like ^A [00:11:21] *Simetrical 's beeps [00:11:40] Splarka, is that like the game where you say in a multiplayer game, "Alt-F4 to open cheat menu", and then see how many people disappear? [00:12:04] <_mary_kate_> ^A shouldn't make anyone's client quit... unless it has really strange key bindings [00:12:36] Simetrical: that and the "to access your secret windows control panel... hold down alt, ctrl, and hit del twice real fast" [00:13:00] Splarka, doesn't work these days, I don't think. [00:13:09] *werdna grumbles about crashing irssi [00:13:11] Ctrl-Alt-Backspace works on Linux, though! [00:13:12] hey, that was Win3.11 days [00:13:17] Kind of. [00:13:30] Or, what is it, Alt-SysRq-B? [00:13:46] One of those does a hard shutdown of the computer. Without unmounting filesystems or anything. [00:13:57] ctrl-alt-backspace rox. [00:13:57] yeah, alt-sysrq-b [00:14:01] <_mary_kate_> Simetrical: but only if it's enabled [00:14:07] no, sorry [00:14:11] -b is reboot [00:14:11] _mary_kate_, yes, well, you can disable anything. [00:14:11] -o is shutdown [00:14:15] -u is remount-ro [00:14:23] <_mary_kate_> Simetrical: the kernel doesn't enable it by default.. or didn't last i looked [00:14:28] <_mary_kate_> (but maybe some distributions do) [00:14:37] _mary_kate_, desktop distros do, AFAIK. At least Ubuntu does, I know that. [00:14:42] Obviously server distros likely don't. [00:14:57] Splarka, in Win3.1, wasn't Ctrl-Alt-Delete instant reboot with no confirmation? [00:15:07] double ctrl-alt-delete was [00:15:09] I think it was 95 where they added the Task Manager, and you had to hit it again to reboot. [00:15:12] Simetrical: double [00:15:18] *Simetrical 's memory fails [00:15:19] oh no, Simetrical is right [00:15:20] <_mary_kate_> ctrl-amiga-amiga! [00:15:20] IIRC anyways [00:15:33] I think Simetrical is right in that it was double in 95, and single in 3.1 [00:15:36] werdna: I thought 3.11 added the check? [00:15:52] what's that really easy to remember nameserver which I have forgotten? [00:15:54] no one can remember back that far I guess [00:15:57] <_mary_kate_> 4.2.2.2 [00:16:03] <_mary_kate_> / 4.2.2.1 [00:17:15] *werdna installs a DNS server so he never has to deal with that again [00:17:26] <_mary_kate_> with what? [00:17:34] bind.. [00:17:41] *werdna shrugs, apt-get install bind9 [00:17:45] <_mary_kate_> your ISP doesn't provide nameservers? [00:18:06] I think they might, but I wasn't resolving anything, so I'd rather just run up a nameserver. [00:18:12] this was on my VPS, might have been a transient thing. [00:18:54] Wait, what are you using a DNS server for? [00:19:10] resolving names to IP addresses. [00:19:22] *Simetrical reads http://www.cato.org/pub_display.php?pub_id=9775 [00:19:23] <_mary_kate_> Simetrical: 'i googled it' is much more memorable than 'i 64.233.187.99d it' [00:19:29] . . . [00:19:40] That's a rather narrow interpretation of my question [00:19:41] . [00:19:54] <_mary_kate_> i could only assume the question wasn't serious.. [00:20:02] <_mary_kate_> since everyone needs a name server to use the internet ;) [00:20:06] What are you *installing* a DNS server for, as opposed to using someone else's? [00:20:09] I think he thought I was using it to serve my own DNS [00:20:16] Even if not. [00:20:33] Why would you even need to run a caching DNS server? [00:21:03] <_mary_kate_> apparently his ISP's weren't working [00:21:13] https://bugzilla.wikimedia.org/show_bug.cgi?id=2170 added 'random page in category' feature but removed it due to big cats (mew) [00:21:18] <_mary_kate_> (probably PEBKAC ;) [00:21:19] Ah. [00:21:34] darkcode has an idea regarding PAGESINCAT though [00:21:40] Splarka, we have no cl_random column. [00:21:43] *Splarka prods darkcode [00:21:46] _mary_kate_: no, I just logged into my VPS and it didn't work, so I figured I'd take the lazy way out. [00:21:53] And MySQL isn't clever enough to allow an efficient ORDER BY RAND() otherwise. [00:21:59] Simetrical: right, so you'd have to build it each query, so that'd be a pain, but [00:22:18] you have a number of category members (re PAGESINCAT) [00:22:29] we could pick a random number between 0 and PAGESINCAT and than retrieve the page name for that item in the category list [00:22:30] can it not just choose a random number from that, and offset the index? [00:23:01] Splarka, so you're planning on scanning half the category's rows average for this feature? [00:23:03] <_mary_kate_> darkcode: i don't think there's any 'category id', and even if there were, that would break if there were gaps in the sequence [00:23:46] Splarka, you could, sure, but it's not scalable. It might only take a few seconds right now on our largest categories, but "only a few seconds" is a phrase that will induce vigorous stabbings by domas in the context of database queries. [00:23:48] Simetrical: I am not planning anything, I'm wondering [00:24:01] domas? he can't hurt me anymore [00:24:07] _mary_kate_, cat_id. We have a category table now, you know. :) [00:24:09] *Splarka is fully vented [00:24:25] <_mary_kate_> really? how uncharicteristly useful [00:24:28] darkcode, see response to Splarka. You'd have to scan an average of half the rows in the category. [00:24:33] _mary_kate_, I added it! [00:24:35] 03(mod) New Special page: Random page in category - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=15824 (10mikelifeguard) [00:24:41] <_mary_kate_> is it populated? [00:24:44] Yes. [00:24:46] I'm not talking about using category ids, mary kate, just get page_title from row X [00:24:47] 03(mod) Random page in this category feature - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=2170 (10mikelifeguard) [00:24:54] That's what {{PAGESINCAT:}} and similar use now. [00:25:10] That's also why category pages now give the actual count and don't just pretend they can't have more than 200 members. [00:25:16] <_mary_kate_> darkcode: how will that work without loading the entire list of categories? [00:25:25] where X is some random number between 0 and the number of pages in the category [00:25:32] darkcode, yes, and that will scan hundreds of thousands of rows in categorylinks for Living people. [00:26:17] _mary_kate_, SELECT cl_from FROM categorylinks WHERE cl_to='Foo' ORDER BY cl_sortkey LIMIT 1 OFFSET 25321; [00:26:29] <_mary_kate_> Simetrical: how is that not loading all the pages? :) [00:26:39] yknow [00:26:48] I wasn't sure what "the entire list of categories" was supposed to mean. [00:26:50] It's a list of pages. [00:26:54] 03tparscal * r43429 10/trunk/extensions/ContributionReporting/ (3 files): Added Fundraiser Statistics. [00:26:54] it seems like much of the problems with categories are always citing "Living people" [00:27:02] as if to say "we'll fix it when you kill some of those people [00:27:07] <_mary_kate_> s/categories/pages.. wasn't sure if the problem was random category or random page [00:27:08] Anyway, it only loads 25321 of them, not all. ;) But that's what I said, you can't do that quickly enough. [00:27:19] _mary_kate_, "random page in category" [00:27:25] <_mary_kate_> yes, i see that now [00:28:20] We'd need a cl_random. [00:28:24] Because the categorylinks table isn't big enough yet. [00:28:36] so add it? [00:28:40] (on the other hand, if we changed it, we could convert cl_to to an id instead of a varchar) [00:28:44] he suggested schema change! [00:28:46] off with his head [00:28:53] darkcode, sysadmins tend to get grumpy when you talk about modifying the categorylinks table. [00:29:37] just sneak cl_random in when you change cl_to to an id ;p [00:29:49] hmm, it does have a timestamp index though, you could just choose randomly from that (though it'd be insanely heavily weighted to new pages I would guess) [00:31:28] Simetrical: uh oh, git svn fetch seems to be starting from r1 [00:31:52] werdnum, maybe git clone doesn't play nicely with git svn? [00:32:01] Or maybe there was some other file you needed to copy? [00:32:02] I did hear that somewhere. [00:32:10] .git/svn maybe? [00:32:37] *Simetrical sees such a directory [00:32:47] $ ls .git/svn/ | wc -l [00:32:47] 107 [00:32:56] what does it contain? [00:33:12] Branches and tags, maybe? [00:33:23] is it plaintext? [00:33:47] *Simetrical points to above file count [00:33:50] 107 directories in it. [00:34:02] Seem to be files called "unhandled.log" in at least a few of them. [00:34:38] If that's what you need, I could tar it up. [00:34:49] It's only 29M. [00:35:17] 03tparscal * r43430 10/trunk/extensions/ContributionReporting/ContributionHistory_body.php: Made paging links follow language. [00:36:59] might work [00:37:03] I'll background git svn fetch for now [00:39:39] *werdnum ponders the best way to get it to me [00:39:56] tar -zcv .git/svn | nc werdn.us 7892 [00:40:16] and on my end, nc -lp 7892 | tar -zxv [00:40:18] *werdnum giggles. [00:40:54] *Simetrical was thinking uploading it to his server for download by HTTP, but whatever ;) [00:41:23] HTTP, schmHTTP [00:41:53] *Simetrical has to go now for the night [00:42:30] did you upload it yet? [00:43:18] 03tparscal * r43431 10/trunk/extensions/ContributionReporting/FundraiserStatistics_body.php: Fixed scale issue [00:44:20] *Simetrical goes poof [00:44:27] :/ [00:44:27] night [00:50:52] I'm having a problem with all my wiki pages using � in the title not linking from other pages [00:51:02] If I change the � to ' then it works.... [00:51:12] is there a setting I can use to fix this [00:51:20] 03pinky * r43432 10/USERINFO/pinky: Adding USERINFO for myself [00:51:24] or do I have to manually change every instance of � to ' ? [00:52:10] I'm hoping its the former rather than the latter because i have hundreds of pages with it :( [00:54:00] roan's not around :( [00:59:39] "Parson's Function Question" [00:59:48] "I would like to discuss with someone who is an EXPERT on PARSON FUNCTIONS how to get them "activated" on a Wiki site." [01:00:07] *werdnum giggles. [01:00:31] <_mary_kate_> {{#baptise}} [01:02:14] here's a new bug (I'll file it): if an API edit triggers the spam filter, what you get back from the API is a generic 500 HTTP response [01:02:32] 03brion * r43433 10/trunk/phase3/includes/ (GlobalFunctions.php MessageCache.php): [01:02:32] Run parsemag stuff through the proper language when wfMsgExt() asks for a specific one. [01:02:32] Previously we only passed on if we were interface or content, but if we're rendering for some third alternative it didn't get used. [01:02:32] Note that parse and parseinline are probably still wrong.... some better model may be required here. [01:02:34] Also it may be wrong in the case of fallbacks ... if you get an English message, is it right to use the Czech magic word parsing? [01:02:37] Hmm. [01:02:39] Might back it out if there's a better way. :) [01:03:02] I liked how the CIA spamming threshhold stopped it just before 'hmm' for a second [01:03:05] heh [01:03:09] carl-m - bug 16120 [01:03:17] anomie: thanks! [01:03:24] I won't duplicate it then [01:05:47] hmm. fixed ten days but not scapped [01:06:48] will finish scappin' soon :) [01:08:00] brion: I know it takes a while, just beating my head against a wall here [01:08:12] *werdnum offers carl-m a pillow. [01:08:25] here's the deal: I can edit [[Talk:Icons of Evolution]] just fine [01:08:38] but if I try to make a new page that says {{Talk:Icons of Evolution}} then I trigger the spam filter [01:09:46] how can that be? [01:12:32] O.o [01:12:42] carl-m: what matches? [01:12:49] http://www.amazon.com/s/002-3951650-1744850?ie=UTF8&tag=mozilla-20&index=blended&link%5Fcode=qs&field-keywords=icons%20evolution%20dvd&sourceid=Mozilla-search [01:13:45] why doesn't the spam filter trigger when I edit the page itself? [01:14:02] *werdnum still tries to get his head around the Configure extension. [01:14:08] Where is ialex when you need him? :( [01:14:28] carl-m: it only triggers when you /add/ a blacklisted link [01:14:37] which was probably a terrible idea, but people whined a lot [01:14:42] *Mike||antispam grumbles [01:14:59] Mike||antispam: when I transclude a page I am not adding any new content [01:15:24] that's not what the software thinks [01:16:05] so here's the problem. my bot is saving a log that has a lot of entries of the form {{Talk:Foo}}. Now, if one of those pages violates the spam filter, my bot gets a spam notice when it edits the log page [01:16:09] man I love kate (the text editor) so much. [01:16:24] but my bot has no way to fix the problem, it just wants to make a log page [01:16:50] Yes [01:16:56] thoughts? [01:17:24] fallback to a link instead of a transclusion? [01:17:32] or better yet, remove the spamlink XD [01:17:41] I probably have no idea which page actually contains the spam link [01:19:25] the "you tripped the spam filter" message does not say which transcluded page is the problem [01:19:55] it has no way of knowing [01:20:53] so I don't know how to proceed here. Once the API bug is fixed, my bot can just skip that page and move to the next one. But really I have no idea what else I can do to work around the problem [01:22:48] carl-m: what api bug is that? [01:23:02] 16120 [01:23:49] which, by the way, is a fun thing to debug in your api code by hand :) [01:25:08] 03(WONTFIX) Problem with $wgSiteNotice after upgrading to MW 1.13 - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=15901 (10maziminke) [01:25:47] and... I don't understand how that fixes the problem you're having. It simply makes api return an error properly when an edit fails due to SBL [01:26:06] you still don't know which page it is that contains the blacklisted link, and it doesn't let you save the page [01:26:16] right, but at least I can skip to the next page, and log the error [01:26:54] why can't you do that now? [01:27:09] because the API returns a "squid error, try again" when the SBL is tripped [01:27:27] which is to say it returns a 500 response code [01:27:33] and no detailed error message [01:27:37] oh, I thought you were using the ui [01:27:43] o no [01:27:44] as a fallback for such cases [01:27:58] or something like that [01:28:03] no, usually a 500 error means you just need to wait and try again [01:28:42] anyway, I have to run now, I will need to think over this and discuss it here at some later point. thanks for your help [01:44:36] 03laner * r43434 10/trunk/extensions/LdapAuthentication/LdapAuthentication.php: Made debug code a little less stupid. [01:49:44] Simetrical: so to get around the lack of cl_random... [01:50:02] the category in question is on 10.5% of pages in the namespace (en.wikibooks) [01:50:19] &generator=random&grnnamespace=0&grnlimit=10&prop=categories [01:50:25] guess what I'm doing ^_^ [01:55:00] <_mary_kate_> cattelan: that's very annoying, please don't do it [01:55:45] sorry my client was throwing a fit [01:56:11] NICKLEN=16, a-zA-Z0-9 and `^-_[]{}\|, can't start with number [01:56:15] no need to experiment [02:02:31] 03(mod) Request: Different "MediaWiki:Spamprotectiontext" for local and global spam-blacklist - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=12034 (10mikelifeguard) [02:04:38] hiya. i have a few questions regarding mediawiki and the wikipedia software... my first question is how do they get the little menu at the top for pages like Template:Navbar/doc to say "< Template:Navbar" to help people go back to Template:Navbar? [02:04:44] that's cool and i want to implement that on my wiki [02:04:48] <_mary_kate_> RingtailedFox: enable subpages [02:04:53] <_mary_kate_> !wg NamespacesWithSubpages [02:04:53] --mwbot-- http://www.mediawiki.org/wiki/Manual:%24wgNamespacesWithSubpages [02:05:08] thanks [02:06:59] do i edit localsettings.php, or language.php? [02:07:24] <_mary_kate_> The normal way to enable subpages for a given namespace is to edit the LocalSettings.php and insert the following: [02:08:28] # Enable subpages in all namespaces [02:08:28] $wgNamespacesWithSubpages = array_fill(0, 200, true); ? [02:08:40] <_mary_kate_> what about it? [02:08:50] is that what i insert? [02:09:00] <_mary_kate_> do you want to enable subpages in all namespaces? [02:09:03] yesh [02:09:23] <_mary_kate_> that seems like a reasonable assumption then [02:11:48] that works great [02:11:58] my second issue is that SVG files do not show properly on my wiki [02:20:07] SVG images are grey and have text instad of the image [02:50:08] How can I move where the TOC are in an article [02:50:30] <^demon> __TOC__ [02:50:45] <^demon> or __FORCETOC__, if you want it to appear regardless of whether it should. [02:51:26] http://www.mediawiki.org/wiki/Help:Magic_words#Behaviour_switches [03:02:07] Whoa! Who redesigned SpecialSearch [03:03:52] http://svn.wikimedia.org/viewvc/mediawiki/trunk/phase3/includes/specials/SpecialSearch.php?view=log [03:19:33] Mike_lifeguard: still around? [03:19:44] yes [03:20:21] I had a late meeting that I had to attend, sorry for running so suddenly. I was trying to think about how to resolve my problem with the spam filter [03:20:58] one easy option, but maybe not practical, is to exempt bots from the spam filter, on the assumption they will be banned if they spam [03:22:02] but that may not be politically possible [03:23:20] That has been discussed before and rejected, IIRC. [03:23:25] werdnum: are you around? [03:23:33] mike: I'm not surprised [03:23:52] another option would be to just break the pages for everyone, following the principle of least surprise [03:24:07] what do you mean "break" the pages? [03:24:17] so you think this problem has come up before? [03:24:26] break = trip the spam filter on all edits [03:24:46] I wonder how the archiving bots deal with it [03:24:53] It doesn't [03:24:56] my guess is they don't [03:25:23] carl-m: yes, why? [03:25:34] werdnum: how does your archive bot deal with the possibility that it might trip the spam filter? [03:25:42] *werdnum shrugs. [03:25:48] haven't bothered thinking about it. [03:25:54] I forgot that thing still runs. [03:26:02] s/does/did/ [03:26:41] carl-m: so you mean undoing https://bugzilla.wikimedia.org/show_bug.cgi?id=1505 [03:27:07] How would I check to see where a template is used? [03:27:10] I would like that quite a lot - that was a bad idea. [03:27:17] chuck: the "what links here" link [03:27:19] Special:Whatlinkshere [03:27:33] and you can then exclude links and redirects to see just transclusions [03:27:35] Mike_lifeguard: yes that's what I mean by break the pages for everyone [03:28:02] I would support that. The question is whether anyone else (ie brion, being the Man who changed things in the first place) would agree [03:35:01] carl-m: FYI, the whole extension should really be rewritten to have a proper log, use a special page to manage things, settings for where to block expressions etc [03:35:17] so, perhaps if/when someone does that, it will be easier to get such things done [03:35:34] Mike_lifeguard: apparently it is not a common problem? [03:36:10] it's not? [03:36:14] one thing that would help is for the spam filter to automatically make a category, "Pages that triger the spam filter", that is populated when a page is saved that already had a filtered link [03:36:20] trigger [03:36:33] well, if it was common, someone else would have complained about it by now [03:36:40] no, they complain [03:36:44] there's just no real solution [03:36:44] I see [03:37:04] there have been requests a few times to make bots exempt [03:37:20] however that was deemed a Bad Idea [03:37:57] I'd prefer to break it for everyone, really. making bots exempt is a security hole [03:38:14] yes, indeed [03:38:30] well, often also is the call to make sysops exempt [03:38:37] that's still a bad idea [03:38:44] "we can't revert vandals who vandalize a page with a newly blacklisted external link" [03:38:50] yes, it is [03:38:53] the only case I can advocate is making /rollback/ excempt [03:38:54] but it doesn't prevent people asking [03:38:58] +spelling [03:39:11] which I've already requested [03:39:32] yes, it makes sense that rollback should be exempt from the spam filter, but not a regular edit [03:40:57] hmmm [03:41:21] well, that's dumb, you can bypass the spam blacklist with: [03:41:25] [{{#ifeq:{{CURRENTHOUR}}|03||http://www.detoxiy.com}} spam me later] [03:41:29] yes [03:41:31] and then purge an hour later [03:41:42] also, the message "you tripped the spam filter" should include a button that says "remove the spammy link and save my damn edit" [03:41:43] it is quite fragile with respect to parserfunctions [03:41:55] carl-m: kinda hard to do I think [03:42:01] but we can give better instructions [03:42:25] Mike_lifeguard: how's that? add an html form, put all the edit info into hidden variables, and add another input wpDeleteTheSpammyLinks [03:42:51] oh, well that sounds lovely then [03:43:01] we have a kinda buggy javascript to simulate that [03:44:21] another option would be to add a parameter to the edit api, to automatically just remove spam filtered links [03:44:36] which is what I am going to do with my bot, anyway, once the API reports to me which link triggered the problem [03:45:01] <_mary_kate_> since the blacklist applies against the parsed page, it's not quite that easy [03:45:17] _mary_kate_: s/$badlink//g. [03:45:36] <_mary_kate_> carl-m: that will not work in all cases, as the blacklist applies against the parsed page, not the source [03:45:40] http://buffalopundit.wnymedia.net/blogs/files/2008/09/vw-ohsnap.gif [03:45:42] of course that only affects the output [03:46:05] but really, why does the spam filter let the filtered links through to the output anyway, when we've told it they're spam?! [03:46:31] _mary_kate_: I think it will probably eventually apply at several levels (ie input and output) [03:46:53] <_mary_kate_> Mike_lifeguard: which still doesn't help if the blacklist only matches the parsed text [03:46:55] so you cannot do something like http://spam and .org in different templates then add them to the page [03:47:01] Mike_lifeguard: {{if:{{a}}||{{c}}|{{d}}}} [03:47:05] <_mary_kate_> you already can't do that... unless something changed recently [03:47:14] <_mary_kate_> that was the point of doing it this way [03:47:31] _mary_kate_: so here is the problem that led me to discuss this [03:47:47] I have a bot, say, that every day makes one edit to save a page that says {{X}}{{Y}}{{Z}}. [03:47:49] I'm not that familiar with how it applies the regex to find hits; see https://bugzilla.wikimedia.org/show_bug.cgi?id=15582 [03:48:01] one day, one of those pages trips the spam filter. now my bot's edit trips it. What do I do ? [03:48:23] <_mary_kate_> Mike_lifeguard: it applies it after parsing templates etc [03:48:34] ok [03:48:41] I thought the spam filter only tripped if you actually added the link, that it ignored links already on the page [03:48:42] well, there are other ways around it [03:48:46] MrZ-man: yes [03:48:55] MrZ-man: or if you transclude a page that already includes the link [03:48:56] <_mary_kate_> MrZ-man: in carl's example, he is inadvertantly adding the link [03:49:00] ah [03:50:26] _mary_kate_: because of the way the filter works, it cannot tell my bot which transclusion caused the problem. and so I don't know what I can do except punt [03:50:37] Misza: !!! hello !!! [03:50:43] I have a $10,000 question for you [03:51:33] 03(NEW) Blacklisted links should mean the page can't be saved - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=16325 normal; normal; MediaWiki extensions: Spam Blacklist; (mikelifeguard) [03:52:14] one idea would be for the spam filter to be more thoroughly integrated in the parser, [03:52:30] so that, whenever the parser finishes converting a page, it then applies the spamfilter [03:52:42] thus it would be possible to tell the first place in the parse tree where the bad link appears [03:52:55] yes, that'd be very nice [03:53:30] well, I just remember from when it only checked the input [03:53:41] 03(WONTFIX) CodeReview's diff shows whitespace changes while ViewVC's doesn 't - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=16161 (10JSchulz_4587) [03:53:59] people were spamming Wikia with http://on.nimp.ugh (url changed for the innocent) [03:54:00] but then... do you just hide the link? the whole page? [03:54:18] 03(mod) Blacklisted links should mean the page can't be saved - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=16325 +comment (10cbm) [03:54:32] I told brion in a PM (BEANS), he said "yah, known problem, let us know if it gets out of hand on WMF" [03:54:40] a few months later it did, and was fixed apparently [03:54:58] but even later, as soon as Tim created LinkSearch, I checked for nimp.ugh links... [03:55:28] and found one IP had added 65 fake {{welcome}} messages with all links pointing to that crashsite, using , after I'd mentioned it to brion, but before the bug was fixed [03:55:31] Mike_lifeguard: I think that just hiding the links, or replacing them with a link to a message page, would be fine [03:55:43] 03(mod) Blacklisted links should mean the page can't be saved - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=16325 (10mikelifeguard) [03:55:44] the message page would say, "this link has been flagged as spam, sorry" [03:56:16] That's what brion and I were thinking. [03:56:30] "added 65 fake {{welcome}}" <- on en.wikipedia, that is [03:56:37] well, actually more like "http://spam.org is blacklisted, but still included in this page. We've hidden it, but please help by removing it" [03:56:50] which is the function that not being able to save the page previously played [03:56:51] 65 people were {{welcome}}'d to nimp.ugh links, the IP was never banned, the links were never fixed, for months [03:56:59] *Mike_lifeguard lesigh [03:57:05] abuse filter! [03:57:06] at least edit summaries are included now [03:57:07] Mike_lifeguard: I was thinking of replacing it in the output HTML; your idea is also good [03:57:15] oh, well both would be best [03:57:21] banner at the top plus a thing in the body [03:57:22] the problem with deleting it from the output HTML is is makes the conversation broken [03:57:36] oh who cares - the link should get removed anyways [03:57:41] so it sure isn't fair to block users/bots/sysops who try to edit a page that was spammed with [{{#ifeq:{{CURRENTHOUR}}|03||http://www.detoxiy.com}} spam me later] .. [03:57:42] *Mike_lifeguard is cavalier! [03:57:50] but it sure isn't fair to allow that sort of thing either [03:57:51] if I say in a comment "Here's a link: [nimp.org]" [03:57:59] and you remove the link, then my comment doesn;t make any sense [03:58:07] right - so fixing /that/ problem is paramount [03:58:11] Heeey everybody! I'm looking at... [03:58:19] if you just replace nimp.org with Special:ThisIsASpamLink [03:58:23] then everything works [03:58:28] meh, sure [03:58:31] Krimpet: you are? [03:58:34] *werdnum offers. [03:58:36] *werdnum hides. [03:58:39] and then that page can automagically remove the link (or try to) [03:58:41] \o/ [03:58:51] well, why not just make it a passthrough? [03:59:08] Mike_lifeguard: I was thinking that page just says "sorry, that was a spam link and has been disabled" [03:59:17] that page = the page the link is relaced with [03:59:22] so the link is disarmed so to speak [03:59:36] carl-m: can you add a request for your automatic spam removal link when the link is blacklisted to bugzilla? [03:59:41] [http://nimp.ugh click here] [03:59:50] click here [03:59:56] but having that removal tool on that informational page makes a lot of sense [03:59:59] -> This is a bad link, you don't want to go here [04:00:05] (or somewhere, at least) [04:00:28] Splarka: yep [04:00:38] Mike_lifeguard: what component is the spam filter? [04:00:51] 03(mod) Allow optional replacement of user talk page - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=16305 (10JSchulz_4587) [04:00:52] extensions [04:00:55] then spam blacklist [04:00:59] got it [04:01:21] 03(mod) Allow optional replacement of user talk page - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=16305 +comment (10mikelifeguard) [04:01:34] thanks [04:01:38] *Splarka thinks: [04:01:57] try to add a blacklisted link that wasn't on the page before (raw wikitext): edit fails [04:02:23] if the link existed before or was added via transclusion/nowiki/parserfuction: success, but redir to Special:ThisIsASpamLink [04:02:36] plaintext? [04:02:49] *Mike_lifeguard *hates* plaintext spammers >:| [04:02:56] most of my SVGs are either mis-coloured, or they show up as grey boxes with text..like this one: http://wiki.mobianlegends.com/wiki/Image:Template-info.svg [04:02:58] why is that? [04:03:02] luckily we have global blocking now, but still it is a problem [04:03:03] can't stop them now anyways, except with $wgSpamRegex [04:03:11] *Mike_lifeguard knows :'( [04:03:13] or abusefilter... *pokes Werd* [04:03:23] 03(NEW) spam blacklist should replace blacklisted links with a safe special page - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=16326 15enhancement; normal; MediaWiki extensions: Spam Blacklist; (cbm) [04:04:40] *Mike_lifeguard hearts werdna [04:04:50] I have $wgGroupPermissions['*']['read'] = false to disable general reads. I want to enable anonymous reads on ONLY the main page, which explains who to email to get an account made. How can I do this? [04:05:30] the main page is usually whitelistread [04:05:35] Oh I suppose I could edit the content of the loginrequired page. [04:05:43] Where is the login required page? [04:05:55] rberckma: http://www.mediawiki.org/wiki/Manual:$wgWhitelistRead [04:05:58] 03(mod) spam blacklist should replace blacklisted links with a safe special page - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=16326 +comment (10mikelifeguard) [04:06:13] Thanks Splarka! [04:06:19] $wgWhitelistRead = array("Main Page", "Special:Userlogin"); [04:06:30] np [04:07:06] wave [04:08:14] how do i fix my SVG issues? [04:08:37] 03aaron * r43435 10/trunk/extensions/ContributionReporting/ContributionHistory_body.php: Remove unused var [04:11:10] What on Earth does $wgInterwikiFallbackSite do? [04:11:14] Used by maintenance scripts to dump interwiki table [04:11:18] (from MediaWiki.org) [04:20:47] woohoo! [04:20:57] I've written a description for every configuration setting appropriate for Special:Configure [04:29:00] is there a way to add additional info to the category display (i.e. adding a little text beside the title in the list)? [04:30:21] I moved my wiki from /var/www/wiki to /var/www/somewiki and now all of the styling seems to be messed up. It just dumps the page. I don't think I changed anything else. [04:30:38] (I tried moving it back to /var/www/wiki, too) [04:31:10] rberckma: what files did you edit when you moved it? [04:32:10] None when I moved it. I modified LocalSettings a bit, including accidentally modifying wgScriptPath, but I have reversed those changes. [04:32:23] so it's back to the way it was? [04:32:33] A diff between this wiki and a working wiki's LocalSettings yields changes in only db privs and the script path [04:32:34] Yes [04:32:50] *MrZ-man gives werdnum a cookie [04:33:24] rberckma: ok, that's reassuring [04:33:32] [6~~ [04:53:48] 03(NEW) Search possibilities cover up search buttons - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=16327 normal; normal; Wikimedia: wikibugs; (probephage) [04:56:55] surely that's a dupe... [04:56:57] hey tomaszf [04:57:15] hey there [04:57:27] dinner time ... [04:59:52] I always scare him off [05:07:18] ta [05:08:20] how do you run the db update script without MySQL admin privs? [05:09:59] oh you just feed the normal db user to it [05:11:55] give it an account which has CREATE, DROP privileges [05:25:29] *werdnum waves tomaszf, predicts a run-away. [05:25:40] :) [06:52:29] 03siebrand * r43436 10/trunk/extensions/ContributionReporting/ContributionReporting.i18n.php: Update capitalisation [06:54:41] http://en.wikipedia.org/w/index.php?title=List_of_fictional_locations_in_The_Simpsons&action=history [06:54:53] *AaronSchulz loves assholes [06:55:25] [/sarcasm] [06:56:27] siebrand: lol, there seems to be a capitalization war going on [06:57:07] He's not on IRC, I believe. [06:57:34] hmm, usually he is, ah well... [07:10:07] Can anyone help me setup a dynamic link in the skin to the talk page? [07:11:30] I have a dynamic link setup to point to the edit and history page like this but I can't get the variables correct for the talk page...any body available to help? [07:12:51] what do you want it to do if you're on a talk page? (or rather, in an odd-numbered namespace) [07:13:17] what I was looking to do is just create a link in the footer of each page to link to the corresponding talk page [07:16:43] wb splarka [07:18:03] any chance you have a solution? [07:18:58] $this->mTitle->getTalkPage(); ? [07:19:58] I was trying to have it php coded into the skin but i'm not much of a php expert [07:20:09] not even sure how to code it but I'll give it a try in a sec [07:21:21] who is a php expert? or rather, who wants to be [07:23:52] you probably want to run a check like if( $this->iscontent && !$this->mTitle->isTalkPage() ) (is that valid php?) [07:24:05] shrugs [07:24:10] I really dont understand much of this [07:26:28] what i'm trying to do is something like this [07:26:29] $talkPageName = $talkPageTitle->getFullUrl('action=edit'); [07:26:29] echo($talkPageName); [07:26:39] $talkPageName = $talkPageTitle->getFullUrl('action=edit'); echo($talkPageName); [07:27:32] [07:27:43] I think some variables in there isn't proper... [07:27:53] not sure if you're understanding what I'm trying to do... [07:29:25] on each content page at the top... is a link to the subject page and the talk page, (one of which is the page you are on), so if you're on [[Foo]] it links to "Foo" and "Talk:Foo", if you're on [[Talk:Foo]] it links to the same [07:29:40] you want to duplicate the action of the second tab in the footer? [07:29:48] yes [07:29:53] pretty much [07:30:10] just the second tab, the discussion page [07:30:25] then you probably want to check if you're on a content page, I think via if( $this->iscontent ) [07:30:47] how would the whole code look like? [07:30:53] this would be added into the skin [07:31:28] I dunno php, so can't help you with the "Whole code" ^_^ [07:31:31] $this->mTitle->getTalkPage(); [07:31:36] dope :( [07:31:48] excuse me? [07:32:07] never heard of the phrase dope? [07:32:11] like darnit [07:32:40] not in that context, usually it is a subject-directed prejorative [07:32:56] its nothing offensive :) [07:33:19] its more of like "too bad for me" [07:33:31] anyway, check what includes/SkinTemplate.php does [07:33:43] function buildContentActionUrls() [07:33:46] yah I looked through it [07:34:00] if( $this->iscontent ) { [07:34:00] $talkpage = $this->mTitle->getTalkPage(); [07:34:02] but I don't know how to piece it together into a php link [07:34:12] so that I can just place it directly on the skin [07:35:19] anyone else around to lend a helping hand? [07:37:26] hello, i have problem with installing gadget, ie i created MediaWiki:Gadget-wikEd.js and -wikEd but still it isnt visible anywhere, i added line in MediaWiki:Gadgets too [07:38:25] Ruibo: http://www.mediawiki.org/wiki/Manual:FAQ#How_do_I_change_the_footer.3F you might be able to just add a new wfmsg call [07:38:45] thanks, i'll look it over [07:39:37] agaran: you added it to [[MediaWiki:Gadgets-definition]] ? [07:39:47] also, do you have the gadgets extension installed? [07:41:31] Splarka: yep, gadgets extension is installed, but i didnt knew that i have to add anything to Mediawiki::Gadgets-definition, [07:41:56] that wasnt exactly where in the footer I wanted it, its more of like below the body content. but it doesnt seem like I can add it to that part of the footer either [07:42:10] hi [07:42:11] guess I'll have to deal with it tomorrow as its getting late [07:42:18] thanks for trying tho [07:42:25] I really appreciate it [07:42:26] is there an easy way to tell how long an image has been unused? [07:42:29] Ruibo: well, that is the footer, at the bottom [07:42:36] if you didn't want it there then it wouldn't be in the "footer" [07:42:50] yes but its just easier to say footer than bottom of the content :P [07:42:53] (terminology, for the win) [07:43:07] Splarka: ok fixed, works, thank you very much [07:43:09]