[00:01:47] longapges warning is in kilobytes (1000 B) or kibibytes (1024)? [00:02:25] Who cares? [00:02:41] messages do ;-) [00:02:49] kB != KiB [00:03:04] $this->kblength = (int)(strlen( $this->textbox1 ) / 1024); [00:03:10] thx [00:05:01] 03(mod) Provide preference-based autoformatting for unlinked dates - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=4582 (10ashtongj) [00:08:28] 03(NEW) Oversighted edits still linked to in the watchlist - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=16122 normal; normal; MediaWiki extensions: Oversight; (mikelifeguard) [00:15:53] 03(mod) Oversighted edits still linked to in the watchlist - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=16122 (10brion) [00:20:12] 03(mod) Oversighted edits still linked to in the watchlist - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=16122 (10mikelifeguard) [00:26:21] 03(mod) Oversighted edits still linked to in the watchlist - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=16122 (10brion) [00:53:36] if you have curid=130130&diff=1243711&oldid=1242545 then the diff will show the changes from oldid to diff right? [00:59:27] Why would you have such a URL? [01:01:14] well I got that bit off my watchlist [01:02:18] Seems weird, although I don't know the exact structure of diff URLs offhand. [01:02:53] 03aaron * r42588 10/trunk/extensions/FlaggedRevs/specialpages/RatingHistory_body.php: Use image/svg+xml. Also fixes chrome :) [01:02:58] so I guess the issue is that the revs between oldid and diff are gone, so there's nothing to show [01:03:07] so then the issue is that the link is shown in the first place [01:04:14] 03(mod) Oversighted edits still linked to in the watchlist - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=16122 (10mikelifeguard) [01:04:47] brion: can you sync flaggedrevs? [01:34:51] AaronSchulz: moment :) [01:40:51] hey, he's alive! ;-) [01:47:22] 03brion * r42589 10/trunk/phase3/ (RELEASE-NOTES docs/hooks.txt includes/SkinTemplate.php): [01:47:22] Revert r42584 "Hook SkinTemplateTabAction now passes $title by reference allowing it to be changed by extensions" [01:47:22] $title is used previous to this call, so that isn't an appropriate place to change it. [01:47:48] time for reverts ;-) [01:52:26] 42,584 revisions.. pretty impressive [01:53:12] and 45% of those are reversions [01:53:38] hahaha [01:53:43] evil ;-) [01:53:48] actually, anyone ever done stats on what percent of svn.wm commits contain the string /(backing\s*off|revert)/ig ? [01:54:10] quality control innit.. this is no democracy! ;-) [01:54:27] "..just as there is a best size for every animal, so the same is true for every human institution. In the Greek type of democracy all the citizens could listen to a series of orators and vote directly on questions of legislation. Hence their philosophers held that a small city was the largest possible democratic state..." [01:54:29] did you mean: this is no anarchy!? [01:54:39] im not sure now [01:55:29] 03brion * r42590 10/trunk/phase3/maintenance/updaters.inc: Update to r42587 -- use insert() DB wrapper instead of constructing raw SQL with unescaped strings (even if we assume they'll be safe... probably) [01:55:50] he updated instead of reverted: 2 drinks [01:56:17] *brion passes around the Bailey's [01:56:21] brion: did you receive my mail? [01:56:32] baileys ? that's not a mans drink sir! [01:56:37] ALE! [01:56:46] brion: did you notice my question about sql query btw? [01:56:51] juliano: yep. i'm a little behind on those :) [01:56:52] Splarka: and it wasn't a revert of my enhanced RC fixes for once :P [01:56:53] Danny_B: nope [01:57:01] brion: did you get my love letter ? [01:57:05] ok, i'll c'n'p [01:57:15] brion: is "WHERE page_title NOT REGEXP("\.(css|js)$") ok to put in source of one special page or should such narrowing be solved rather in php? [01:57:23] it's not you exobuzz, it's me. i just can't stand you ;) [01:57:28] MrZ-man: ouch, heh, maybe you should test them on test.wp progressively [01:57:36] although committing is the best way to get them noticed [01:57:39] brion: next time ill include gummy bears [01:57:45] it shoudl work now [01:57:45] :) [01:57:46] Danny_B: eww? well that might be hella slow [01:57:48] brion: ah, ok, no worries. :-) Since it has been five days, I wondered if you you maight have lost it, or I may have lost your answer [01:58:13] any other suggestion how to filter css & js pages? [01:58:14] Splarka: I decided to just not show the arrows at all for people with JS disabled [01:58:23] hah, losers [01:58:29] that way they couldn't screw everything up [01:58:36] MrZ-man: btw, did you see my new contribsrange? it uses those arrows [01:58:50] I considered having non-js fallback, but as they're generated in js, I decided not to [01:59:01] MrZ-man: i have to check what you've done with it [02:01:48] at least I got rid of all the cryptic ids it used before [02:02:09] wtf does that mean? [02:02:42] and for extra confusingness, it used RCI and RCL [02:02:54] did you keep s instead of s? [02:02:58] yes [02:03:02] great then [02:03:04] with a title attribute [02:03:16] yeah, i was translating something ;-) [02:04:41] speaking about this, is there any reasonable reason, why we use blue triangle images instead of ►▼ chars? that would save a lot of bandwidth [02:05:28] the other day i was about to change it, but got scared a bit somebody will shout on me that i've done that ;-) [02:05:52] <_mary_kate_> a lot of bandwidth? how much? [02:06:29] http://en.wikipedia.org/skins-1.5/common/images/Arr_r.png [02:06:35] 210 bytes [02:06:40] heh [02:06:49] the down arrow is 215! omg [02:06:58] wait [02:07:03] Danny_B: a) consistency of font support? b) consistency of spacing? c) our arrow images can have shadows and 3d and shit ;) [02:07:04] the HTTP headers are bigger than the image [02:07:18] you forgot about stuff [02:07:30] *brion notes if we were insane, we could use CSS sprites to reduce the http overhead of loading multiple small images ;) [02:07:42] brion: i like the c thing;-) [02:08:07] there are 778 bytes of headers for the HTTP request for http://en.wikipedia.org/skins-1.5/common/images/Arr_r.png [02:08:13] i think we should raytrace the arrow images in client-side JS [02:08:15] re b) spacing in monospace is always same [02:08:27] brion: you could generate the arrow in js using 1x1px divs [02:08:28] re a) then we can use + & - chars [02:08:28] think of the web server overhead, if just before the arrow is downloaded the keepalivetimeout/MaxKeepAliveRequests is reached! [02:08:31] <_mary_kate_> Splarka: so like 1KB the first time you load RC [02:08:55] *Splarka nods [02:09:11] cant you replace the arrow with some ascii arrows? [02:09:17] > v [02:09:19] :) [02:09:23] bigger [02:09:30] and the logo with a clever ascii representation [02:09:32] Splarka: at least use

it was simple to use classname and toggle it [02:13:32] noe rc is shitty ugly invalid table [02:13:56] Danny_B: you know a super quick way to save tons of bandwidth that no one ever considers? have all projects use a common stylepath off-domain. domain hopping on secure.wm is actually faster on dialup for me as all sites have most skins/images preloaded. some browsers of course don't like to load off-site js but fuck em [02:14:56] Splarka: i don't think it would be that big a ton; domain hoppers are going to be the minority. but yeah, i keep meaning to stick those on upload. or something [02:15:10] shaving a few trivial tags on a disused feature that no one likes isn't gonna save the US economy [02:15:12] heh [02:15:16] :/ ok rephrase... why the hell is cp "$ROOT/reference/phase3/" copying extensions and phase3 [02:15:28] if any browser doesn't support loading off-site JS, it's violating the standard :) [02:15:43] brion: is a common paranoia/security setting in opera IIRC [02:15:54] <_mary_kate_> brion: it would also be faster to serve them from upload as it doesn't run apache.. [02:16:09] can we conveniently configure it to serve the files gzipped? [02:16:16] guys, you live in countries where internet sonnection is fast, cheap and matter of course. but think about people who live in places where internet is paid for bytes transferred and/or is slow [02:16:22] <_mary_kate_> lighttp? probably.. ask domas ;) [02:16:26] :P [02:17:03] *Splarka notes wikia moved styles offsite ages ago, because domain hopping there is much more common [02:17:10] var stylepath = "http://images.wikia.com/common/releases_200810.4/skins"; [02:17:22] brion: anyway, back to the query - do you have any suggestion how to do the css/js cutoff effectively? [02:17:58] context? [02:18:54] wantedtemplates shows user js's where they eg. add button to edit bar inserting some template [02:18:59] agh, damnit... I know what's wrong [02:19:21] hmm [02:19:34] we lost the alt text on the expand/reduce arrows [02:19:38] that ain't good [02:20:15] becuase I added a title to the

[02:20:37] MrZ-man: JS on, image loading off <- my buttons are invisible [02:21:02] alt and title are totally unrelated things, dude :) [02:21:07] MrZ-man: I can't use enhanced recent changes with my computer off, plz fix [02:21:24] *Splarka hides [02:21:42] jidanni will come after you and hound you for breaking the images off case ;) [02:24:25] aaaagh... T_T cp why do you forsake me? [02:24:51] hmm, on IE7 the alt text overrides the title text [02:25:42] MrZ-man: use title on image as well [02:25:52] or only on image [02:26:13] put title on the link [02:26:15] and alt on the image [02:26:26] or...? is that the case where it gets funky? :D [02:26:31] isn't that what he's done now? [02:26:39]

[02:26:47] this should work [02:27:02] right, the "+" showing if the image isn't loaded, and title on mouseover, at least in moz [02:27:31] 03brion * r42591 10/trunk/phase3/maintenance/scapToWikimedia.php: [02:27:31] Backing out r42575 "Add quick script to update to the latest SVN version Wikimedia is using." [02:27:31] This will update backwards as well as forwards, which is a little scary. :) [02:27:31] Since it could be pretty surprising and make your life harder to recover from if [02:27:31] you run it by accident (losing mixed versions, encountering new update [02:27:34] conflicts, etc), I'd recommend backing it out for now. Also 'scap' is a term of [02:27:36] art specific to our deployment system, and is not used appropriately here. ;) [02:27:41]

;) The issue is IE, and you know how troublesome it is [02:28:41] whohoo... yay for evil shell scripts... :/ cp forced me to explicitly request that the root .svn directory be copied... [02:28:56] <_mary_kate_> DanielFriesen: did you use *? it won't match dotfiles [02:29:01] why is the main dev folder called "phase3" ? [02:29:01] btw: why did the rc switch from list to table? [02:29:10] <_mary_kate_> exobuzz: because it's the third version of the software [02:29:21] hah. ask a silly question [02:29:24] although it's mediawiki 1.x ;-) [02:29:28] just seems a little lost in all those folders [02:29:37] _mary_kate_: cp -r $ROOT/reference/phase3/* . [02:29:38] there used to be mod_wiki [02:29:46] as one of the previous versions [02:29:49] <_mary_kate_> DanielFriesen: right, that won't match files starting with . [02:29:58] aaaaaaggghhh... [02:30:00] <_mary_kate_> Danny_B: UseMod [02:30:09] DB_SLAVE will work even if I have only one server, right? [02:30:10] yeah sth like that ;-) [02:30:23] *Danny_B just remembers the "mod" [02:30:58] _mary_kate_: Any alternative other than `cp -r $ROOT/reference/phase3/* $ROOT/reference/phase3/.svn/ .` [02:31:36] `cp -r $ROOT/reference/phase3/ .` creates a phase3 folder in . which I don't want [02:31:38] Anyone know? [02:31:47] <_mary_kate_> DanielFriesen: you could use rsync: rsync -ra $ROOT/reference/phase3/ . [02:31:58] heh... [02:32:09] agh.... damnit, lame hacks... [02:32:43] screw it... I'll stick with the lame cp line... [02:32:52] cp -r $ROOT/reference/phase3/* $ROOT/reference/phase3/.* [02:32:55] The only . file I need is the .svn [02:33:16] well, that works to [02:33:32] or just cp -r $ROOT/reference/phase3 some_dir [02:33:35] <_mary_kate_> IMSoP: that will match '..', will it do what he wants? [02:33:57] _mary_kate_: meh, didn't think of that :( [02:34:30] In fact, by the sound of it: cp -r $ROOT/reference/phase3 . [02:34:44] <_mary_kate_> i think cp -r $ROOT/reference/phase3 .. might work [02:35:12] yeah, don't use globs at all, let "cp -r" handle it [02:35:35] . is $ROOT/projects/projname/ [02:35:45] Would .. not create a phase3 project [02:35:50] +directory [02:36:08] anybody seen JeLuF within last say three weeks? [02:37:38] nope; the syntax is | cp -r source destination | [02:37:56] <_mary_kate_> yes, if the current directory isn't called phase3, that won't work [02:38:41] so source is "$ROOT/reference/phase3", destination is "." (or "..") [02:38:53] $ROOT is /home/daniel/Workspace/mediawiki, under that there is trunk/phase3 and trunk/extensions which are two different svn working copies checked out under my svn user, resource/phase3 and resource/extension are similar trunk coppies, except they come from anon http://svn... [02:39:02] <_mary_kate_> DanielFriesen: what about: cd $ROOT/reference/phase3; find . | cpio -pdmv $DESTINATION [02:39:22] And there is the projects/ folder which is a set of project folders [02:40:03] I'm working on a small ./manage bash script that'll add/delete/list projects, as well as update each one to match the revision of my currently checked out trunk [02:40:19] cpio? [02:40:23] <_mary_kate_> yes [02:40:57] so why do you need to be *in* the destination? [02:41:17] i.e. why are you cp'ing to "."? [02:41:35] <_mary_kate_> i assume he doesn't want to create a directory called phase3 [02:41:48] does it make much difference whether I use . or $ROOT/projects/$PROJECT ? [02:41:56] 03mrzman * r42592 10/trunk/phase3/includes/ChangesList.php: restore image alt text on enhanced RC, add title to s so IE won't hide it with the alt text. [02:42:31] but you only get a directory called phase3 because you haven't specified a target name [02:42:38] Ya, basically when a new project is started, I want to copy what's inside of $ROOT/resource/phase3/... into $ROOT/projects/$PROJECT/ [02:42:39] Ok [02:42:52] <_mary_kate_> IMSoP: no, if you cp -r .../phase3, it'll always create a directory called phase3 in the destination [02:43:32] mkdir ~/foo; cp -r ~/foo ~/bar [02:43:40] _mary_kate_, even if the destination doesn't exist? [02:43:48] will create a directory called ~/bar, but no directory called ~/bar/foo [02:43:53] <_mary_kate_> Simetrical: well, it _does_ exist [02:44:01] <_mary_kate_> if it doesn't, okay, it won't [02:45:08] ok, my sleepy head only just cottoned on; | rm -rf $target; cp -r $source $target | ? :| [02:45:28] <_mary_kate_> but he needs to copy phase3/ and extensions/, so only the one he copies first will work [02:46:33] but not over the top of each other surely? i.e. $target will be different? [02:47:07] *DanielFriesen changes the topic to "BASH ambiguity debate channel..." [02:47:57] ok... scrapped my mkdir [02:48:01] works [02:48:32] I didn't intend to copy all of extensions anyways [02:48:36] Only what I was working on [02:48:55] going to have a different command for that [02:49:46] Now I can work on ./manage update [02:50:07] now if only I knew what funky key-combination I hit the other day that made bash generate a massive glob for the contents of the current directory [02:50:20] echo *? [02:50:34] find $PWD ? [02:51:09] ^_^ Oh, gotta love my old days of using `find / -name ...` [02:51:17] heh, locate ftw [02:51:40] no, funkier: it output something like f{oo|t{ball|bar}}|ba{r|z} and so on [02:51:50] heh [02:52:02] that sound really funky [02:52:09] *s [02:52:27] Doesn't look that hard... well, if you're not using bash [02:52:57] Just something that takes the list of files in the current directory, and breaks up groups by common leading characters [02:53:15] any more than one file matching the initial pattern turns into a {} [02:53:25] *brion-dinner wanders off for eats :) [02:53:27] 03brion * r42593 10/trunk/phase3/includes/ChangesList.php: [02:53:27] Tweak for r42592 (re-adding alt text for arrows on enhanced RC) [02:53:27] Use a non-breaking space instead of plain space for the alt text for the spacer image; the plain space seems to get lost (tested in firefox 3). :P [02:53:27] Note this was broken before, it's not a regression. :) [02:53:31] oh yeah, the point is, this wasn't a *command*, it was a bash shortcut-key; Ctrl-[ somthing I think [02:53:40] heh [02:54:06] inserted the result into the current commandline [02:57:17] bon apetite, brion [03:01:32] ^_^ ./manage update works nicely [03:01:48] though... it should handle extensions eventually [03:04:35] ^_^ If I get this bash script to work nicely, does anyone else want it? [03:09:12] Damn... one of these days I really need to get my own mailserver working [03:31:44] TimStarling: in wikimediamessages, is it possible to create new group of special pages somehow? and if not there, is it possible in initialisesettings? [03:38:37] why would you want to do that? [03:38:43] the groupings suck, you should fix them for everyone [03:56:22] http://www.newegg.com/Product/Product.aspx?Item=N82E16826100006&Tpk=nia guesses on how long until a countervandalism system is set up with one of these? [03:58:27] *snickers* [03:59:16] Actually, that's been on ncix for awhile... I almost debated buying one when I was looking for a mouse to get when I was planning on getting this laptop [04:00:27] heh... Imagine reverting an edit just by raising your right eyebrow >.< rotfl... [04:01:05] Get annoyed... automatic ban! [04:09:49] TimStarling: i'd like to create "Wikimedia" group for wmf-only extensions such as boardvote, sitematrix, global blocking, merge account [04:11:03] Danny_B, global blocking and SUL aren't supposed to be WMF-only... [04:11:08] I have a big mediawiki question [04:11:24] How does Userrights know to show (none) in the logs if the person has no rights [04:11:25] Krimpet: ok, but still board vote and sitematrix are [04:12:04] besides there may come other in future i can imagine [04:12:29] Anyone know? [04:13:00] chuck, have you checked the source? [04:14:50] gtg, just leave an answer, pls and i'll read it later, thx [04:14:57] Krimpet: I'm looking through a few things now, my terminal has frozen after it hit a bunch of message files though [04:15:15] function makeGroupNameList( $ids ) { [04:15:15] if( empty( $ids ) ) { [04:15:15] return wfMsgForContent( 'rightsnone' ); [04:15:40] what file is that? [04:15:46] Splarka, you spoiled it. :) I was hoping it'd be a zenlike revelation [04:15:51] /includes/specials/SpecialUserrights.php [04:16:00] *Splarka pokles Krimpet [04:16:04] SpecialUserrights.php, line ~300. [04:16:27] but also see LogPage.php [04:16:30] Well what the hell... [04:18:27] Why wouldn't wfMsgForContent('rightsnone'); work in this extension O_o [04:18:35] can an extension only access its own messages? [04:20:43] anyone? [04:24:35] chuck: Extensions can use any message... and any message defined by an extension can be used anywhere else... (though things will get fuzzy unless you handle wfLoadExtensionMessages correctly) [04:24:52] Nevermind, my extension is working finally!!! :D [04:24:57] *chuck is so happy [04:26:06] Hmmm... do I dare venture into the forbidden area of MediaWiki config... [04:26:12] though unless its doing exactly the same thing as what the message was intended for, messages generally shouldn't be recycled [05:01:39] How can I upload my new Extension so that it can be in the extension distributor and such? [05:04:40] chuck: ExtensionDistributor simply makes tarballs of extensions inside SVN [05:04:47] just commit it to /mediawiki/e [05:04:55] bleh, damn laptop keyboard [05:05:08] ...to /mediawiki/trunk/extensions/... [05:05:16] Ah, I don't have SVN access. Is there someone who can commit my extension and its updates? [05:05:31] I suppose I could... [05:06:00] Okay, one sec while I do a quick cleanup to make sure I didn't leave any passwords in or anything [05:09:04] Dantman: http://meta.yourwiki.net/shareduserrights.tar.gz [05:13:58] yourwiki.net? another one... and it not on my lists? [05:16:21] Oh, nevermind... 2nd sector [05:16:30] T_T But I want that domain... damn [05:16:31] O_o what? [05:17:07] ^_^ I'm big on the 3rd sector wiki hosts... one of these days I hope to start a decent one [05:17:40] What's the difference between 1st, 2nd, and 3rd? lmao [05:17:44] http://nadir-point.com/wiki/Wiki_Hosting#Wiki_Hosting_Market_Sectors [05:18:07] <_mary_kate_> chuck: some people enjoy making up terms so they sound more knowledgable ;) [05:18:47] ^_^ It's easier than having to refer to the category verbosely every time [05:18:56] _mary_kate_: :P [05:19:02] Dantman: Oh, well that's good, because we're planning on merging with Skizzerz's crack team of developers to provide a more "3rd sector" type service where users can create wikis through a special page and have SSO, etc on all of them with permissions, etc also [05:20:17] Been trying to create a good comparative list of the existing mediawiki hosts in that sector... but haven't got back to work on it [05:21:12] <_mary_kate_> Dantman: yeah, but when you use the terms like they're some kind of established terminology... it looks a bit odd. (since no one knows what you're talking about) [05:21:34] ^_^ Thats why I follow with a link [05:22:00] I was hunting it down before chuck even asked [05:25:21] http://www.mediawiki.org/wiki/Extension:SharedUserrights :D! [05:25:30] Now I just need it in SVN and I can add the links to the tarballs [05:25:48] chuck: Keep me posted, the more groups participating the better the resources that could be put together [05:27:24] I've actually got a variety of things lying around... Host names, some management extensions, scalability plans, and even business plans... [05:30:19] O_o [05:30:22] wow [05:31:35] Hmm... maybe I should talk to Simon... [05:31:58] Think he said he might help out a bit [05:32:50] Dantman: Those screenshots of CreateWiki, etc look nice :P [05:34:00] Oh, heh... [05:34:11] Right, I forgot I left those screenshots there [05:34:19] http://nadir-point.com/wiki/Wiki_Hosting/Archived_plan this also looks good too :P [05:34:41] It's branched off a little from there [05:35:45] I had a plan for a SSO system as well [05:36:19] SHID... it was a kind of hands-off SSO, which fit the model of a wiki host with diverse unrelated wiki well [05:37:59] ^_^ On Thursday I found out how easy it is to become a domain registrar... [05:38:17] The small startup company I work for actually has the license to do it. [05:38:36] Gave me an interesting idea as well... an option for in-house domain registration. [05:39:29] ie: For a simple option, people could register their domain right with the wiki host, have it automatically created, configured, and paid for with the same system for the wiki itself [05:41:16] becoming a domain registrar (at least, officially) is HARD [05:42:14] They want financial statements showing you earn more than $x/year, etc. [05:42:23] where X is about $5m [05:42:32] most people just become resellers... [05:47:08] werdnus: Sadly, most people assume reselling is the same as being a registrar (caused by the marketing teams of said resellers or registrars offering reseller accounts) [05:52:24] oh well... hopefully I can just find some way that'll either let me use some automatic system or api for registration of domain names... [05:55:09] 03(mod) Asymmetric rendering of bytes-changed in watchlists and recentchanges - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=15925 (10AlexSm) [06:00:41] Dantman: yeah, most resellers would have that. [06:00:46] strong.mw-plusminus-pos {font-weight:normal;} [06:01:27] chuck: ^_^ You do realize that $wgSharedTables exists? [06:02:13] that's a really dodgy way to give information... [06:02:38] Not everybody is extra-familiar with MediaWiki internals, and if they don't know about $wgSharedTables, they probably don't know what it does, either! [06:03:06] Easy !wg on mwbot to find that info [06:03:20] he probably doesn't know that, either. [06:03:37] chuck: you're using CentralAuth, aren't you? [06:03:44] chuck: so you can use the global groups features? [06:04:51] 03dantman * r42594 10/trunk/extensions/SharedUserrights/ (5 files): Committing SharedUserrights for Charles Melbye (chuck) on IRC [06:06:38] Actually, while it's not really relevant to that extension... It would be good if people started to understand that $wgSharedDB does not imply that users are shared anymore... [06:06:56] 03(NEW) Importing fails when using SQLite - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=16123 major; normal; MediaWiki: Export/Import; (mediawiki) [06:07:32] hmmm... wait, do I have the right props set? [06:07:59] you should have auto-props set [06:10:55] New system... [06:18:04] 03dantman * r42595 10/trunk/extensions/SharedUserrights/ (4 files): Set svn:eol-style=native props for SharedUserrights since my new laptop didn't have autoprops set. [06:34:23] Ow... looks like ScribbleWiki shutdown... [06:35:25] Hmmm... [06:38:17] Sad to see the one of best hosts in that sector shutdown... though that does mean a better edge [07:09:49] Isn't a bit of a breaking change that {{ns:3}} outputs "User talk" instead of "User_talk" now? [07:10:00] Certainly breaks templates and ParserFunctions... [07:10:52] MZMcBride: in which way? [07:11:58] VasilievVV: {{#ifeq:{{FULLPAGENAMEE}}|User_talk:{{BASEPAGENAMEE}} ... [07:12:05] Cases like that are now broken. [07:12:52] Errr.. bad example. [07:13:10] {{#ifeq:{{FULLPAGENAMEE}}|{{ns:3}}:{{BASEPAGENAMEE}} <-- Now broken. [07:14:40] not broken so much as inoperative and fixable, the bad part is an unexpected change in normal behavior [07:15:01] Right. [07:15:39] http://svn.wikimedia.org/viewvc/mediawiki?view=rev&revision=41876 ? [07:16:04] Ahh, Ilmari is to blame. [07:16:28] {{NAMESPACE}} and {{PAGENAME}} don't output underscores [07:16:47] Right. So they have {{NAMESPACEE}} and {{PAGENAMEE}} counterparts. [07:16:47] mmm... [07:17:11] *Splarka confused now [07:17:20] > Also change the output to use spaces [07:17:20] instead of underscores [07:17:36] well, bah, okay, makes sense [07:17:44] but should have been done when the word was invented, heh [07:18:08] But now an unknown number of templates have been broken (needlessly). [07:18:24] might make more sense for ifeq to treat spaces and underscores as equivalent [07:18:49] and #switch? [07:18:54] But they're not. [07:26:41] *MZMcBride posted to Ilmari's talk page. [07:40:38] "Your password is invalid or too short. It must have at least 6 characters and be different from your username." I get this when i try to set my user preferences [07:42:41] 03aaron * r42596 10/trunk/extensions/FlaggedRevs/specialpages/RatingHistory_body.php: Should be wfMsgForContent() [07:47:54] if i remove the password from "old password" field i can set my preferences again :O bug? [07:48:51] Melos: If you put in an old password then it assumes you want to change it, so if you havent put in a new one that error appears [07:48:59] You only need the old password to change your password [07:50:04] MattJohnston: when i get the first error i didn't put anything in old password field [07:51:08] Melos: That's weird - what version? [07:51:32] i get this on betawiki [07:51:49] i presume the last svn version [07:54:53] probably something's autofilling it.. [07:55:55] is there a easy way to list all pages created in my wiki? [07:57:44] Special:Allpages [07:57:53] thanks werdnus [07:59:28] ifeqe? [07:59:40] switche? [07:59:44] heh... [08:00:17] what about nse? [08:00:58] that to [08:01:11] both are good options [08:01:25] that *too* [08:01:32] heh [08:02:00] well, nse is a lot cleaner than about 10 other parserfunctions [08:02:11] ifeqe and switche or whatever you want to call them... both have the advantage that they can be used in any page context [08:02:25] not just ns [08:03:31] k [08:03:42] It's good for templates that have an argument for comparison in them, if you know the input is a page, you'll be able to handle it with looser more valid input [08:03:47] or we could install string functions [08:04:03] {{#replace:_| |{{ns:3}}}} [08:04:49] ^_^ Before that, you're going to need to replace StringFunctions with a sane extension... [08:05:07] StringFunctions is far from performance optimal... it needs a partial rewrite [08:06:25] from what I recall, brion rejected installation of it on Wikimedia cause of that [08:06:55] http://www.wallstats.com/deathandtaxes/ [08:06:56] wow [08:10:19] the guardian did something like this for the UK too - much saner numbers though =/ [08:22:29] Could anyone provide perspective as to whether APC is better than memcached for main/parser/message caching purposes on a MediaWiki install which is getting reasonable traffic but using only a single node, and similarly for PHP's own session cache vs. memcached for sessions? [08:22:31] Trying to figure out whether or not to eliminate our memcached instance since we're not looking to expand beyond one node right now and we could use the memory for other purposes. [08:26:23] memcached is only really necessery if you have >1 servers [08:26:43] otherwise APC will be fine [08:29:03] 03nad * r42597 10/trunk/extensions/PdfBook/PdfBook.php: allow format=single and notitle in query-string for exporting single pages [08:30:19] <_wooz> lo [08:45:50] 03ialex * r42598 10/trunk/extensions/Configure/ (7 files): Settings which have the same value as in DefaultSettings.php are now removed from the saved configuration [08:47:01] Splarka: What are your favourite parts of JavaScript or ECMAScript... [08:47:55] JS? I hate it, but there isn't much alternative [08:48:03] awww... [08:48:22] at the moment anyways... there is flash but that is worrrrrrrrrse [08:49:07] 03ialex * r42599 10/trunk/extensions/Configure/Configure.php: ...and forgot to bump the version [08:49:09] ^_^ My new job got me to love JavaScript... or rather, some of the underlying ideas in the ECMAScript syntax and Prototype based languages. [08:49:44] Dantman: the language as such is fine. dealing with the dom & browser bindings sucks. [08:50:09] the spec is flaky, implementations broken, and there's too much proprietary additions. [08:50:11] "language" is a bit strong [08:50:25] Splarka: naw, it's a proper language. not a bad one, either. [08:50:43] it's just not much good for dynamic content on web pages :) [08:51:12] It's a little nicer when you can focus on a single browser [08:51:27] so what good is it? server side interpreters? [08:51:38] Splarka: it's a scripting language like any other. [08:51:42] it is _for_ dynamic content on web pages, so it is basically... no good [08:51:43] or can be [08:52:08] 03(mod) Asymmetric rendering of bytes-changed in watchlists and recentchanges - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=15925 (10happy_melon) [08:52:23] Splarka: for example Java6 comes with a JS engine for scriping [08:52:28] not bad at all. [08:52:47] sounds nichey to me X_X [08:52:58] *Splarka still hates JS [08:53:04] Prototype is fairly nice... though there are also some downsides [08:53:30] first-order functions etc are also a good thing [08:53:47] Splarka: I hate it for web stuff too. but the language proper isn't so bad. [08:53:50] in theory :) [08:54:41] first order? [08:55:11] Dantman: functions as language objects. stuff you can assign, pass around, etc. [08:55:20] my main dislike is there are dozens of ways to do things, but most are broken in at least one browser [08:55:22] most "good" scripting languages have that nowadays. [08:55:24] mhmm, ^_^ gotta love that [08:55:26] not php though .=) [08:55:56] Splarka: yes, as i said, browser bindings and implementations suck. the language as such, as specified, isn't bad though :) [08:56:03] <_mary_kate_> Duesentrieb: not only first order functions, but closures [08:56:10] <_mary_kate_> Duesentrieb: a very surprising (and nice) feature to find in a language like JS [08:56:11] _mary_kate_: indeed [08:56:16] ;) Splarka, that's the DOM you hate... the language itself is quite nice and reliable [08:56:49] some level of inheritance as well [08:56:50] Dan: not _just_ the DOM, but whatever [08:56:51] _mary_kate_: not so surprising if you look at the language as such. it's fairly lispy. [08:57:18] ECMAScript is Lispy? [08:57:44] js is basically a shot in the dark "maybe it'll work for 99% of visitors, and enhance their experience" [08:58:08] ^_^ And here I was wanting a JS like language with a little more lisp like stuff... [08:58:37] here's the latest thing I've written in JS: http://en.wikipedia.org/wiki/User:Splarka/contribsrange.js [08:58:44] it'd sure be nice not to have to do that [08:59:10] Splarka: what does it do? [09:00:06] queries the API for a /16 /24 /25 /26.../32 CIDR or ABC* wildcard submitted to Special:Contributions [09:00:08] IMHO MediaWiki isn't the best of places to use JS... you can't get very close to it's full potential [09:00:08] Dantman: what lispy feature are you missing? [09:00:37] 03ialex * r42600 10/trunk/extensions/Configure/writePHP.inc: be sure that this is an array, just to avoid E_WARNING in some cases [09:01:07] ^_^ being able to redefine parts of the language itself... [09:01:12] or rather extend [09:01:44] arr.forEach() gets fairly ugly with scope sometimes when you are doing something simple [09:01:59] T_T And the builtin doesn't support a break [09:03:05] Though, I have to say JavaScript's (not ECMAScript) generators look interesting... but unfortunately of course they are Gecko only [09:03:10] can't you just assign something else to arr.forEach ? I have never tried, but i would have thought that should work. [09:03:27] well, perhaps prototypes for bultin stuff can't be modified like that. [09:03:35] but it *should* work :) [09:03:42] One should be wary of redefining things that are already defined in the browser [09:03:54] generators? with continuations? [09:04:01] i thought continuations where only in Rhino [09:05:07] https://developer.mozilla.org/en/New_in_JavaScript_1.7 https://developer.mozilla.org/en/New_in_JavaScript_1.8 [09:05:33] ^_^ And [foo, bar] = ... makes me drool... [09:06:31] python does that [09:07:36] ^_^ I know, and there is list in php... that's why seeing that ability inside of some variant of JS looks so delicious... [09:11:00] Though there are a few times, when I wished that "var foo = 'bar'; var bar = 'red'; [bar]" or some sort of other construct would evaluate to a variable [09:11:15] some sort of local scope operator or something [09:11:24] 03ialex * r42601 10/trunk/extensions/Configure/Configure.obj.php: ...and that was the wrong function *sigh* [09:11:48] It doesn't work out so well when you're dealing with anon functions and various scopes [09:12:09] ^_^ ".[foo]" would be nice [09:12:38] "var foo = 'bar'; var bar = 'red'; [foo]"* [09:15:31] ... [09:15:39] you can do ${$foo}; [09:15:48] that would do what I think you want to do [09:15:50] or just $$foo [09:16:29] ^_^ Ya, in php... [09:16:34] ;) not in JS [09:18:19] Being able to do something like (function() { var foo = 'red'; var bar = 'green'; for( var x in somelocalscopeoperator ) window[x] = somelocalscopeoperator[x]; })() wouldb [09:18:27] would be nice* [09:18:33] I think you can do window. [09:18:48] No, window is the root scope, not the local one [09:19:17] And this evaluates to the scope of the object you are in, not the function [09:20:15] There is no way to define var something = 'green';, and get 'green' back in some way from 'something' when you are in the scope of an isolated anonymous function [09:21:10] And if you are doing any real heavy JS app or library, you really want to be inside of that isolated anonymous function, or else you run the risk of having other libraries pollute your scope and break the page [09:24:03] ^_^ in something other than js... it would be interesting to be able to map foo[:bar, :baz] to something like foo.'[]'(:bar, :baz) [09:25:54] mmm, true. [09:26:21] Ooohh... someone actually thinking along my insane ideals... whohoo [09:26:34] ^_^ I've been having some fun specing out a language [09:27:06] Something taking in ideas from different language backgrounds [09:27:42] Prototype's .prototype, anonymous functions, symbols, lisp's ability to redefine parts of the language, and such [09:27:52] Closures [09:28:16] mmm, I was going to write my own language, but then I realised the sheer amount of work in it, and went back to some cool RPC stuff I've been working on. [09:28:54] heh [09:29:26] Writing a spec is a little fun... and perhaps a small interpreter [09:30:16] It's better if you can do a bit more of a lisp like one... [09:30:40]