[00:04:32] Question about the Group Based Access Control extension for MW... I'm trying to make pages read-only for anonymous users, by using sysops,,*(ro) (or using "anon" in place of "*"), but it only brings up the No_Access page. Does anyone know of a way to make pages read-only for anonymous users? [00:05:32] What version of mw do you have? [00:05:36] 1.11 [00:05:44] do you have access to localsettings.php? [00:05:47] yes [00:06:54] ok, then add in (sorry for multiline flood) [00:06:54] $wgGroupPermissions['*']['edit'] = false; [00:06:54] $wgGroupPermissions['user']['edit'] = true; [00:07:13] this will do it for *all* pages [00:07:17] yes [00:07:24] as opposed to the thingy, which does it for *one* page [00:07:52] (if it works) [00:07:57] or use Lockdown to do it for a namespeace :) [00:08:20] 03(NEW) Capitalization consistency in special pages - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=11986 trivial; low; MediaWiki: Special pages; (MZMcBride) [00:08:41] if you want to make only certain pages read-only for anons, just semiprotect them [00:09:50] right [00:10:51] yeah, the problem with that is that the Protect function only has 3 settings... non-protected, registered users, and sysops [00:11:21] if there was a way to add more groups into the Protect function then it would work really well... [00:11:48] there is a way [00:11:51] i forget, though :) [00:12:21] omg! I've been looking for a way to do that for a while! Couldn't find any documentation on it on the web though [00:12:45] i think i dug into the source to find out :) but it has been a while. and i need to go to bed. [00:12:51] gn8 [00:12:54] night [00:13:38] 03(NEW) Editing protected pages without ability to protect/ unprotect them - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=11987 15enhancement; normal; MediaWiki: Page editing; (skizzerz) [00:13:43] :) [00:14:07] luke123456: lemme find the way to add stuff to the protection thingy [00:14:27] that would be great Skizzerz - thanks a lot :) [00:14:54] $wgRestrictionTypes [00:15:09] !rights [00:15:09] For information on customizing user access, see . For common examples of restricting access using both rights and extensions, see . [00:15:57] did someone in here just create bug 11987? [00:16:18] yeah, that'd be me [00:16:20] Skizzerz: you can just create a new usergroup [00:16:25] and enable whatever rights you'd like [00:16:43] I did, but I want them to be able to edit the pages, but not protect/unprotect them [00:16:51] is there any documentation on that paramter Skizzerz? I'm fresh to MW and PHP, so a "how to..." goes a long way for me [00:16:58] luke123456: http://www.mediawiki.org/wiki/Manual:$wgRestrictionTypes [00:17:09] thanks heaps mate [00:17:14] np [00:18:48] huh, i'm really surprised that hasn't been added before [00:18:54] *Skizzerz is too [00:19:17] I've tried almost everything, 'editprotected' 'modifyprotected' 'sysop'... [00:19:33] yeah the meta page lists all of the rights [00:19:43] sorry, mediawiki page [00:20:02] luke123456: and http://www.mediawiki.org/wiki/Manual:$wgRestrictionLevels allows you to change what groups the levels can be assigned to [00:20:25] 14(DUP) Editing protected pages without ability to protect/ unprotect them - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=11987 +comment (10cbrown1023) [00:20:26] 03(mod) Create "editprotected" right - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=10974 +comment (10cbrown1023) [00:20:49] ok... WHY didn't my bugzilla search find that bug? [00:21:06] Skizzerz: Can you type in an example for me if I was to say, add the usergroup "Luke" to be available in the Protect page so I can see the syntax? [00:21:56] $wgRestrictionLevels = array( '', 'autoconfirmed', 'sysop', 'luke' ); [00:22:42] I'm assuming there's no pre-existing $wgRestrictionTypes within the LocalSettings.php (as I just did a search for it and it found nothing) ...? [00:23:38] it's not default in there [00:23:45] it's default in /includes/DefaultSettings.php [00:24:01] ahh ok... [00:27:01] 03(NEW) Identifier needed for MediaWiki generated pages in page source - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=11988 15enhancement; normal; MediaWiki: General/Unknown; (oq9pscb02) [00:30:51] 03(NEW) Some table nav classes don' t seem to render properly in archive pages - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=11989 normal; normal; MediaWiki: Page rendering; (FT2.wiki) [00:40:21] 03(NEW) New mailing list for bg.wikipedia.org - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=11990 normal; normal; Wikimedia: Mailing lists; (5ko) [00:56:52] any report that on some wikiproject MediaWiki:Whatlinkshere-next and MediaWiki:Whatlinkshere-prev is broken? [00:57:25] hello [00:57:50] its been awhile since i've been here... I guess thats a good thing cause our mediawiki has just been chugging along... [00:57:50] http://de.wikipedia.org/wiki/Spezial:Linkliste/Hauptseite View (previous 50) (next 50) (20 | 50 | 100 | 250 | 500) ok [00:58:01] But I have an odd question... which is: [00:58:31] When I upload a file, where does the wiki store the information about the file size? [00:58:42] http://en.wikipedia.org/wiki/Special:Whatlinkshere/Main_Page View (previous 50) (next 50) (20 | 50 | 100 | 250 | 500) [00:59:19] basically, Upload a dummy file (because the original is huge) then ssh in and replace the file with my huge one. But the image page still says the file size for the old small dummy file... [00:59:28] http://hr.wikipedia.org/wiki/Posebno:Whatlinkshere/Glavna_stranica View () (Posebno:Whatlinkshere/Glavna stranica) (20 | 50 | 100 | 250 | 500) [00:59:37] any idea? [01:15:24] 03(NEW) What links here navigation on hr wikipedia is broken - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=11991 normal; normal; MediaWiki: Special pages; (stem) [01:16:17] Hey folks... [01:16:35] Is there any way to strip categories off during a page/template transclusion? [01:17:08] For example: [01:17:15] Template:A says any page that includes it, has category:abc [01:17:23] Page B includes {{A}} and therefore has category:abc [01:17:36] but, Page C includes {{:B}}... and picks up Category:abc, which I DONT want [01:19:39] Is there a command that forces an email verifcation when creating a new user, before the account becomes active? [01:20:06] no command, but this can be config'd in LocalSettings.php I beleive [01:20:36] sorry, that's what I meant (only 4 days in to PHP/MW) [01:20:46] no worries. :) [01:20:50] I'll go through it again [01:20:55] Luke: [01:21:00] Read DefaultSettings.php [01:21:03] it will be in there [01:21:07] ah thankyou [01:21:11] ;) [01:21:18] found in /includes/ [01:24:24] No one on knows the answer to my question above? [01:25:21] please repeat Klemo [01:25:26] When I upload a file, where does the wiki store the information about the file size? [01:25:30] basically, Upload a dummy file (because the original is huge) then ssh in and replace the file with my huge one. But the image page still says the file size for the old small dummy file... [01:26:20] one sec [01:27:11] image table, image_size field [01:27:37] old images are in: oldimage, oi_size [01:28:06] sounds rediculous. I have no idea where they came up with that naming convention. ;) [01:34:29] oh so in the database... duh.. I guess I could of figured that out... [01:34:55] Isn't that what you asked? [01:34:59] yup [01:35:03] Everything is stored in the database. :) [01:35:12] (Except the images themselves...) [01:35:29] I thought maybe it checks the file on every visit to the page, but it doesnt... which means it would be in the database... [01:35:51] Yeah, that would be pretty inefficient... [01:36:06] lol now lets see if I blow away my database tring to make manual updates... [01:36:32] NOOOOOO!!! [01:36:35] KHAANNNN!! [01:36:41] (Backups.) [01:36:45] lol [01:36:47] indeed... [01:37:00] What MW version? [01:38:22] wow... impressive... our database dumps are 28mbs now... cool [01:38:23] why don't you run the rebuildImages.php maintenance script? [01:38:35] I did not know that script existed... [01:38:38] why don't i... [01:38:40] Me neither [01:38:43] ;) [01:38:59] What MW version? [01:39:07] What is your wiki URL? [01:39:09] 1.6.8 [01:39:12] ouch! [01:39:15] lol [01:39:16] upgrade time! [01:39:19] no way [01:39:22] WAY [01:39:23] is there a way to take something in a template and make spaces into underscores? I know there's a thing that makes the first letter upper or lowercase [01:39:24] if it aint broke dont fix it [01:39:32] {{#replace [01:39:39] string function extension [01:39:39] this is internal use anyway... no outside invaders are welcome [01:40:01] burried deep deep in our internal network... [01:40:03] {{#replace ? [01:40:07] like.. [01:40:09] nedscott, yup [01:40:13] {{#replace|A|B}} [01:40:41] {{#replace: HAYSTACK|NEEDLE|NEWNEEDLE}} [01:40:47] not available in default MW [01:41:09] Klemo: naughty, naughty! Are you still running Windows 95 too? [01:41:18] hell no [01:41:20] would it be something en.wiki has enabled? [01:41:24] this bitch runs on a linux box [01:41:24] nope [01:41:28] which is outdated... [01:41:29] hahaha [01:41:30] definitely not Ned [01:41:43] Linux 2.6.9-6.37.ELsmp #1 SMP Tue Mar 29 15:43:19 EST 2005 i686 i686 i386 GNU/Linux [01:42:00] Ned: you need to make the space in tags, or it won't work, too [01:42:01] hey 2.6... not soo old... [01:42:16] Linux version has nothing to do with it [01:42:18] ;) [01:42:54] I'm trying to autoformat an export url [01:43:20] su just.... #/wikidir/maintenance/ php ./rebuildImages.php [01:43:27] like http://en.wikipedia.org/w/index.php?title=Special:Export&pages=Article_Name&history=1&action=submit [01:43:34] something like that Klemo [01:43:37] lol [01:43:58] I'm trying to figure out how that url could be generated from, say, an AfD template [01:44:22] but spaces obviously break the EL [01:44:28] nope it did not update the file the file size [01:44:55] Yup, you are right Ned. You need {{PAGENAMEE}} and String Function Extension for that kind of automagic [01:45:16] Sorry Klemo... might be a bug that is fixed in a newer version of MW [01:45:17] Windows_Vista_VPN.zip (132KB, MIME type: application/x-zip) it should be in the neighbourhood of 10mb [01:45:18] Seriously [01:45:31] nerts [01:45:36] thanks anyways [01:45:38] I'l just go with plan A... update the database... [01:45:39] np [01:45:40] lol [01:45:42] thanks [01:45:49] Ned: I am doing similar stuff on my wiki. ;) [01:53:39] has anyone had their wiki hacked from the back-end? [01:54:39] wondering what security measures people put in place for that... [01:56:15] Anyone had experience installing TreeView? I've followed the instructions on the site. and I used the example usage but I'm not having luck. [01:56:47] mysql> select * from ISGWIKIimage where img_name = "Windws_Vista_VPN.zip"; [01:56:56] this returns 0 results... [01:57:42] if I do a select * from ISGWIKIimage the fields and data are there... [01:58:45] 03(mod) Categories should not be able to include themselves - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=2132 +comment (10webmaster) [02:00:22] um... where are you defining your table name in that select? [02:00:36] uh never mind... I used the time stamp... [02:00:38] thanks [02:01:42] np [02:03:13] 03gri6507 * r27508 10/trunk/extensions/TodoTasks/ (SpecialTaskList.php SpecialTaskList_body.php): v0.9.2 - minor bug fix - did not close the HTML tag. [02:05:30] YAY! Windows_Vista_VPN.zip (9766KB, MIME type: application/x-zip) [02:05:55] intresting... I put 10000000 and it rounded it down to 9,766K [02:06:05] 10000000 [02:06:07] lol [02:06:12] thanks again! [02:13:16] 03brion * r27509 10/trunk/phase3/includes/SpecialNewpages.php: Constructor parameters on a QueryPage need to be optional, so the class can be loaded in other contexts (such as running updateSpecialPages.php) [02:14:13] Duesentrieb you there? [02:25:09] 03brion * r27510 10/trunk/phase3/includes/SpecialNewpages.php: [02:25:09] * validate namespace parameter as int, not text [02:25:09] * set the hidebots setting in the links so it doesn't randomly flip off [02:25:38] !magic [02:25:42] !magicwords [02:25:42] For more information about magic words, see . [02:26:36] jamie_hari: you're aware of {{urlencode}}, etc.? [02:29:09] "validate namespace parameter as int, not text" [02:29:10] brion: ;) [02:30:07] :) [02:54:48] 03brion * r27511 10/trunk/phase3/ (4 files in 2 dirs): [02:54:48] Fix regression in r27416 -- {{REVISIONID}} was broken on save since a parsed copy from before revision-insert was used to insert into parser cache. [02:54:48] The parser now sets a "vary-revision" flag on the ParserOutput when using {{REVISIONID}}; Article then reparses the page after insertion if and only if required. [02:58:52] you know that change was in the RELEASE-NOTES [03:04:27] brion: presumably there's some important use of {{REVISIONID}} which you know about and I don't? [03:05:34] TimStarling: i'd expect it to be used mainly in cute templates for permalinks [03:05:46] occasionally for diffs or other things perhaps [03:05:50] (I've made a few. They're very cute.) [03:06:17] TimStarling: and yes i know it was in the release notes, that's why i removed it from the release notes when i fixed it :) [03:06:34] you can use the sidebar for permalinks [03:06:43] or Special:Cite [03:07:17] wouldn't it have to be substituted in a template? [03:07:36] no, because you wouldn't bloody use a template [03:09:36] brion: you added a feature which doubles the parse time on save [03:09:43] TimStarling: only for pages that use it [03:09:48] a feature which was obsolete shortly after it was introduced due to the sidebar change [03:10:09] exactly, it's a feature which, when used, doubles the parse time on save [03:10:32] well, you could fix it back to parse only after the id is known :) [03:11:18] what, just parse it twice on every edit? [03:11:44] well, you _could_ insert provisionally, then roll it back after a cancellation. [03:11:49] but i suspect it's not worth the bother. [03:11:57] you've removed the triple-parse for 99.999999% of pages [03:12:09] why bitch about a very rare double-parse that fixes a bug that that fix introduced? [03:12:31] well, you could fix it back to parse only after the id is known :) [03:12:40] ok, *how* is the ID known? [03:12:52] because I'm in a bad mood because you haven't sent the NSF recommendation letter [03:13:08] AaronSchulz: after the revision is inserted into the database, the page (was) parsed so link tables could be updated [03:13:34] that parse has now been pulled up to happen during edit filtering, so that the various spam checker plugins don't have to run separate parse passes [03:13:43] but since that's done *before* the revision is inserted, the revision id isn't available [03:13:49] right, ID not known for save [03:13:54] so I'm in a mood for bitching [03:13:56] but what uses this magic word? [03:14:14] my last change was to check if the wiki text tried to use the revision id, and then and *only* then reparse the page after revision insertion, with the id known. [03:14:19] I guess if its nothing, then the extra parse wont hurt [03:14:29] since it would only trigger for those [no] pages [03:14:30] that way the version that's saved into the parser cache at save time is consistent with later renderings [03:14:56] instead of sometimes not showing the revision id (since it was an older cached version from before save) and sometimes showing it (because it was a later parse, not cached) [03:15:29] that was a bug, yes [03:15:34] it shouldn't have shown it on the later parse [03:18:07] mm [03:19:43] brion: for a moment I was wondering, "why not just replace the magic words rather than reparse", but that would have bugs :D [03:19:56] *word [03:20:47] brion: heh, flaggedrevs has a special error for certain issues around using parserfunctions to determine transclusions, I wonder how much it will show up [03:21:12] prolly almost never, in a sane world :/ [03:21:23] that's a point, StringFunctions would destroy strip markers wouldn't it? [03:22:21] brion: heh, transclude template X if {{REVISIONID}} > z, else transclude template y [03:22:36] brion: heh, transclude template X if {{CURRENTDAY}} == z, else transclude template y [03:22:51] yes, all kinds of horrifying possibilities :D [03:22:56] brion: it's the later that I had to make the error message for [03:23:23] because the user is sent the image/template ids/keys, and it could change from the time they starting viewing till save time [03:23:56] only way to describe such scenarios is "fucked up wacky shit" [03:24:17] brion: and people wonder why Parser is not easy to port or make standalone :) [03:26:02] 03(mod) Install the StringFunctions extension - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=6455 (10tstarling) [03:26:59] I can think of some other ways to describe it [03:27:19] brion: hmms, NSF recommendation letter? [03:27:47] like dependency tracking, you can track the variables which go into deciding a template title [03:28:15] *nod* [03:30:17] 03brion * r27512 10/trunk/phase3/ (7 files in 5 dirs): [03:30:17] * Changed signature-ip to signature-anon [03:30:17] * Removed the hardcoding of the link parameter ($1), let the message control the destination. [03:30:17] * Removed the hardcoded talk page links. They can be easily re-added if desired, though, and added per-wiki. [03:30:17] * Removed the $3 and $4 parameters [03:30:18] * Fixed wfEscapeWikiText() to do "]" as well as "[", so sigs with "]" won't break links [03:30:22] * Updated a parser test case for the escaping change [03:30:59] TimStarling: i could print the letter, sign it, and scan it i guess [03:31:23] what is this magic letter? [03:32:27] their deadline was monday, and they sent me a reminder on friday afternoon [03:32:41] I said the office wasn't available to send it until tuesday, could they wait until then? [03:32:48] then i got a mail about it on... monday... which i spent doing laundry as it was a holiday :P [03:32:56] they said "yes, we can wait until wednesday at the latest" [03:33:20] yeah, I sent you an email on monday, but it was tuesday my time, I was actually thinking you could do it on tuesday [03:34:33] I could have done it myself but I don't have a scanner [03:34:36] 03(mod) Anonymous users' signatures should link to contribs page - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=11315 (10brion) [03:34:49] well tuesday was when i found you the letterhead :P [03:35:15] *AaronSchulz has a scanner/printer/fax all-in-one, it's really cute [03:35:50] anyway, yes, it would be nice if you could print it, sign it and scan it [03:36:02] yeppers [03:36:13] have i got the addy to mail it to? [03:36:14] *brion looks [03:36:28] "addy", good god [03:36:30] but I thought you were going to do it on tuesday, and then I thought you were going to do it earlier today, so I want to be really really sure this time [03:37:52] trying to fix the damn letterhead [03:38:00] "read error" doesn't look as nice as "wikimedia foundation" [03:38:06] terveen@cs.umn.edu [03:38:23] eeexcellent [03:39:20] *AaronSchulz wonders what this letter is about [03:39:47] *brion confirms the links [03:41:56] AaronSchulz: there's a research group at UMN that we are sending anonymised squid log data to [03:42:29] they recently published a paper about the impact of vandalism in terms of the number of times the "damaged" page is viewed [03:42:33] it was quite a cool paper [03:42:40] *AaronSchulz sent some bot code to someone doing their PhD [03:43:02] people really take an academic interest in wikipedia and vandalism [03:43:10] TimStarling: where are the results? [03:43:38] they're applying for a grant now, because, you know, even academics have to eat and pay the rent and what not [03:43:59] so they wanted a letter from us confirming that we helped them with their research and that their paper is pretty cool [03:44:30] gotta find my usb cables :P [03:44:30] http://www.cs.umn.edu/~reid/papers/group282-priedhorsky.pdf [03:45:13] is there any way to call a Special page by more than one name? [03:45:20] and have it act slightly differently, depending [03:45:48] like having params? [03:46:03] why would you need a fullblown special name? [03:46:06] AaronSchulz: yeah [03:46:14] i suppose that's good enough ;-) [03:46:24] 'The server at www.cs.umn.edu is taking too long to respond.' [03:49:44] jlerner: special page functions, and the execute() function of a special page class, have a subpage parameter [03:50:28] so you can have [[Special:SomeSpecialPage/param,param,etc.]] [03:51:05] http://www.kpig.org.uk/wiki/index.php?title=Science [03:51:08] holly shit [03:51:26] someone with non cologne blue/monobook default [03:51:59] that looks like nostalgia [03:52:49] version 1.4.5, so it's got some nostalgic value in mediawiki terms too [03:53:35] prolly unsafe :) [03:53:55] 'Proponents of this division use the arguments that the "soft sciences" do not use the scientific method, admit anecdotal evidence, or are not mathematical, all adding up to a "lack of rigor" in their methods.' [03:54:01] hehe, funny [03:54:58] 03nickj * r27513 10/trunk/phase3/maintenance/parserTests.txt: [03:54:58] ParserTests - Sane behavior requirement with apostrophe followed by bold in French or Italian [03:54:58] (was implied already, but this just explicitly lists this case) [03:55:57] ah, speaking of parser tests, here's the current results from my new preprocessor: [03:56:08] ok all scanified [03:56:31] brion: first "addy", now "scanified" [03:56:43] great, thanks brion [03:56:57] hmm, internet leetspeaker vs. George Bush [03:56:59] "11 previously PASSING test(s) removed o_O / 5 new PASSING test(s) :) / 17 still FAILING test(s) :( " [03:57:09] AaronSchulz: This from the person saying 'prolly'. :p [03:57:40] bye [03:58:20] I renamed some of the extractSection() tests, and removed some others [03:58:23] TimStarling: yeah - i already have it like this, vi a template: [[Special:CardAction/{{{1}}}|Claim this card]] [03:58:28] i'll add a couplemore parameters [03:58:38] because as I was saying last night, extractSection now ignores HTML-style headings [03:58:54] TimStarling: ok mailing [03:59:01] and it ignores headings with comments on the same line [04:00:04] hmmm, brion, did you see the things I said last night (earlier today my time) about my parser work? I'll PM you if not [04:00:51] i have vague memories [04:01:30] brion: I have vague memories too :) [04:03:52] actually I'll give a summary here first... [04:04:21] I've changed the parse order to put removeHTMLtags after replaceVariables, that's been my dream for a long time [04:04:53] so there are a couple of parser test changes that are direct consequences of that, such as the ability to open an HTML tag in one template and close it in another [04:05:24] I know you're not a big fan of that in particular [04:06:03] but it makes a lot of sense, because doing removeHTMLtags piecemeal, both before replaceVariables() and during substitution, makes it very hard to validate properly [04:06:35] matching tags and block elements inside inline elements and what not [04:06:55] yeah [04:07:05] it annoys me but i have to agree, the horses have already escaped the barn :) [04:07:15] so go ahead and do it right [04:08:03] I've removed split() entirely, that task is now done by an extension-tag-aware preprocessor stage, simultaneously with template expansion [04:08:45] ooooh [04:09:04] I haven't found any consequences of that in terms of wikitext syntax change, as of yet, so it's pretty solid and backwards-compatible [04:09:57] but it allows for some interesting improvements of the interface between the parser and the extension tags, such as allowing the extension tags to access the template arguments [04:10:15] because the extension tags are now expanded at a time when the template arguments are available [04:10:54] anyway, that can come later [04:10:57] sweeeeeet [04:11:28] as for extractHeadings(): there's about 20 bugs reported relating to misnumbered sections, some closed and some open [04:11:47] where the section edit link doesn't match the section in the edit box [04:12:08] I'm getting rid of that failure mode entirely [04:12:22] yay [04:13:20] instead, the preprocessor marks sections that appear in the source text with strictly correct syntax, and then both extractSections() and formatHeadings() use those same marks [04:13:51] if formatHeadings() finds a heading that's not marked by the preprocessor, it formats it as a heading but without a section edit link [04:14:11] how will that work with edit links for transcluded templates? [04:14:19] nice [04:14:31] the preprocessor marks the headings with both title and index [04:14:42] preprocessor expansion is aware of what title it is expanding [04:14:51] oh, it's going to be run after the template substitution? [04:15:30] alright, to be precise, there are two steps: a context free transformation to an XML-like document tree, and expansion of that document tree to plain wikitext [04:16:04] so template substitution occurs during expansion of the document tree [04:16:29] the XML contains heading markers, and these are translated during expansion to strip-like markers [04:16:58] the XML heading markers do not contain titles, the titles are added at the expansion stage [04:19:08] ok, so the current status is that the parserTests.php situation is good, no broken tests [04:19:39] I had to update some of the tests for interaction between templates and HTML [04:20:02] and for the extractSections() changes [04:20:33] before I commit, I'd like to try a differential fuzz test, where I compare the old parser and the new parser on random input [04:20:52] that's tough though when there are real changes [04:25:03] 03brion * r27514 10/trunk/phase3/ (3 files in 2 dirs): [04:25:03] Revert r27151 -- allows session fixation attacks. [04:25:03] Just get a user to visit a URL with the user ID and token you like in the query [04:25:03] string (say, in an referenced in a page you convince them to go to or post [04:25:03] for their review) and their login session will be replaced with the one you [04:25:03] provided. [04:30:23] hahaha [04:30:37] brion: I always love to watch you jump into code review mode [04:30:46] *AaronSchulz gets the popcorn [04:33:30] ok if there's no objection i'm going to try updating live site to svn current [04:33:31] 03(mod) Use transclusions to count articles as well - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=11868 (10fearow00) [04:34:34] brion: don't [04:34:42] hmmmmm [04:35:05] for checkuser? it looked alive when i tested it locally [04:35:16] "alive"? [04:35:32] I don't recall any syadmin running the changes [04:35:44] TimStarling: :'( [04:36:22] brion: we need some sysadmin robots for schema changes [04:36:25] it didn't break when i did a search on my local test wiki :) [04:36:41] doesn't it force the index? [04:37:01] well that's what i was thinkin [04:37:05] but maybe only sometimes.... [04:37:07] or maybe it's broke :D [04:37:32] it doesn't [04:37:48] huh, not for "get IPs", which is what matters [04:37:56] :P [04:38:23] well how expensive will it be to change this index? [04:38:37] wait...yes it does [04:39:26] FORCE INDEX(cuc_user_ip_time) [04:40:35] brion: ADD INDEX cuc_user_ip_time (cuc_user,cuc_ip,cuc_timestamp), DROP INDEX cuc_user_time; [04:42:11] brion: did you run update.php on your test wiki? [04:42:21] b/c it would have auto-updated [04:42:39] i did after i did the intiial test... [04:42:45] maybe it slipped in by mistake earlier :( [04:42:48] so off course it works ;) [04:42:58] *of [04:44:50] *AaronSchulz wonders what https://wikitech.leuksman.com/history/Squid_100_percent_cpu_headers_problem is [04:45:09] brion: you can try log_id while you are at it :D [04:45:32]