[00:01:58] 14(DUP) Show log of previous moves when moving a page - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=12203 +comment (10raimond.spekking) [00:01:59] 03(mod) Put a move log in Special:Movepage - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=6488 +comment (10raimond.spekking) [00:02:04] ajnewbold: if you have parserfunctions installed, yes [00:02:09] !parserfunctions | ajnewbold [00:02:09] ajnewbold: "Parser functions" are a way to extend the wiki syntax. ParserFunctions is an extension that provides the basic set of parser functions (you have to install it separately!). For help using parser functions, please see . For details about the extension, see . [00:02:22] oh, cool; thanks! [00:02:33] I take it that's not part of the default WikiMedia package? [00:02:45] oh, heh, yeah [00:02:48] it says it right there [00:02:50] duh :) [00:02:52] that'S what "you have to install it separately!" means [00:02:57] :) [00:02:59] *ajnewbold is slow :) [00:03:33] grab the bundle [00:03:36] !bundles [00:03:36] unofficial bundles of extensions are at . Note that these are random snapshots and may contain a completely broken version of the extension. [00:03:48] Duesentrieb: I'd need that in order to do something like {{status|ok}} vs. {{status|pending}} and have differently-colored output, right? [00:04:17] ajnewbold: you don't *need* it for that, no [00:04:24] oh! [00:04:27] well that's even better news [00:04:58] ajnewbold: consider this:

{{{1|unknown}}}

[00:05:33] ajnewbold: then, in MediaWiki:common.css, define styles for the classes status-ok, status-pending, status-unknown, etc [00:05:49] if you want to get more fancy than that, you will need the parser functions ,though [00:05:52] ah :) [00:05:54] neat idea [00:06:01] common hack [00:06:18] I wonder if I'm just better off doing {{ok}} and {{pending}} separately though :/ [00:06:33] I'm not certain that I'll be able to modify the css or even install anything extra in this particular environment [00:06:37] you'll end up basing both on yet anotehr template [00:06:41] so, i'd say: no. [00:06:46] hmm, true [00:07:03] you can always later define {{ok}} as {{status|ok}} :) [00:07:20] ah, that's slick [00:07:32] I'm totally new to all of this, in case you can't tell ) [00:07:33] :) [00:07:38] ( aka put a redirect on the template page ;) ) [00:08:38] is there a simple way to quickly see a list of all pages that have a specific template used on them? [00:08:57] like every page that uses {{status}} or {{ok}} or whatever [00:09:09] ajnewbold: Click the What links here link [00:09:11] On the template page [00:09:14] ah! [00:09:21] I should have known that! [00:09:34] thanks :) [00:13:12] hello [00:13:25] hi [00:15:59] are there any free scripts that use wiki api? [00:16:27] oh dear, I can't seem to do a regular HTML link ( can I do an [[internal link]] that's unstyled? [00:18:34] [[Internal link]] [00:18:41] er wait, that won't work [00:19:00] style="text-decoration: none !important" ... [00:19:01] [[Internal link|link text]] [00:19:12] Or, class=plainlinks [00:19:13] thanks! [00:19:16] As that should be defined [00:19:47] amidaniel: that only removes the external icon from external links, not remove all styling from links [00:20:12] Ah, true [00:20:15] you can easily define a class the removes all styling [00:20:16] of course, you can change that in Common.css I suppose [00:20:20] classes are better to re-use [00:20:27] indeed [00:20:30] yeah [00:20:35] (and easier to use) [00:20:41] also, if using a class, you can directly select the text inside the link, so the style is not inherited, but applies directly [00:20:45] should work better [00:21:37] so, in MediaWiki:Common.css, you put down .someclass a { text-decoration: none; }, and then do the link via [[link]]... a lot easier [00:22:27] Only problem with classes is that they often get abused by people who overwrite rather than extend :( [00:42:00] Duesentrieb: Thanks heaps for your help before, much appreciated, highly valuable, cya later [01:16:41] 03shinjiman * r28171 10/trunk/phase3/languages/messages/ (4 files): [01:16:41] * Update Chinese translations [01:16:41] * Update Cantonese translations [01:16:41] * Update Old Chinese / Late Time Chinese translations [01:16:56] 03(mod) Support global (crosswiki) blocking - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=8707 (10mike.lifeguard) [01:22:13] 03shinjiman * r28172 10/trunk/extensions/MathStatFunctions/MathStatFunctions.i18n.php: fixing Chinese translations, per http://www.mediawiki.org/wiki/User_talk:Shinjiman#MathStatFunctions.i18n [01:25:21] 03david * r28173 10/trunk/extensions/LiquidThreads/ (Lqt.i18n.php LqtBaseView.php): Improved (?) behavior for fake indenting. [01:27:49] hello hello [01:27:57] wasssup :) [01:28:39] does anyone know if there is an extension for lussomo Vanilla to allow you to use mediwiki formatting in posts? [01:48:01] <[1]Apollo> does anyone in here have javascript experience? [01:48:19] <[1]Apollo> I have a quick question [01:48:25] ask [01:48:29] <[1]Apollo> k [01:48:38] <[1]Apollo> calling an onclick event using JS [01:49:05] ok...? [01:49:16] <[1]Apollo> you use javascript:yourfunction [01:49:17] <[1]Apollo> in the tag [01:49:27] and you want to know how to set up a link to call a javascript function? [01:49:51] <[1]Apollo> onclick="Javascript:myFunction [01:49:53] <[1]Apollo> that works [01:50:07] <[1]Apollo> wait, does that work I meant to say [01:50:27] <[1]Apollo> because it's capitalizing the Javascript [01:50:29] sorry, I'm confused as to exactly what you're trying to accomplish... are you clicking a link? [01:50:33] <[1]Apollo> yes [01:50:37] <[1]Apollo> but I have a tracker on it [01:50:41] ok, this is for mediawiki? [01:51:15] er, is this for a mediawiki wiki? (just have to check) [01:51:31] <[1]Apollo> for google analytics [01:51:33] <[1]Apollo> on links [01:51:44] *sigh* are you doing this on a wiki? [01:51:56] <[1]Apollo> yes [01:52:00] <[1]Apollo> and other places [01:52:05] <[1]Apollo> all I'm asking [01:52:09] <[1]Apollo> is does it make a difference [01:52:18] you cannot use or onclick= normally in a wiki [01:52:28] <[1]Apollo> if I capitalize that Javascript:, will it not work [01:52:29] the links have to be set up via js [01:53:00] <[1]Apollo> because I have a bunch of links, they have the onclick already in them, but I used Javascript:, not javascript: [01:53:09] <[1]Apollo> and I don't know if I have to go back and change them or not [01:53:21] it has to be lc... but even then I'm not sure if it'll work [01:53:23] <[1]Apollo> because they do not seem to be tracking with the capitalization [01:53:31] how are you calling the links? [01:53:33] <[1]Apollo> I have lowercase working it looks like [01:54:08] <[1]Apollo> dang, so I have to change all them to lowercase [01:54:12] yeah [01:54:18] <[1]Apollo> shoot! [01:54:28] <[1]Apollo> thanks for being patient though Skizzerz :-) [01:54:33] <[1]Apollo> very helpful! [01:54:40] yeah, I'm still a bit confused, but whatever :P [01:55:22] <[1]Apollo> me too, and I don't think it was helping any x-D [01:55:24] because I didn't think you can do that stuff normally to wiki links... unless you have RawHTML enabled [01:55:45] (which IMO is a bad thing) [01:57:31] <[1]Apollo> what if I only have admins editing articles? [01:57:43] <[1]Apollo> then the random users wouldn't make messy HTML... [01:57:54] yes, that is the way around it [01:58:00] only allow trusted people to edit [01:58:02] <[1]Apollo> :-D [01:58:11] but with default permissions, enabling raw HTML = bad [01:58:25] <[1]Apollo> yeah, you could put a redirect in [01:58:37] <[1]Apollo> or something like it [01:58:40] <[1]Apollo> nasty things [01:58:58] you could put in a script that'll track Special:Userlogin [01:58:58] <[1]Apollo> thanks for the help! [01:59:05] np [01:59:08] <[1]Apollo> ... [01:59:12] <[1]Apollo> yeah, that would not be good [01:59:26] exactly why enabling it with default permissions isn't good :P [02:43:47] 03shinjiman * r28174 10/trunk/phase3/ (RELEASE-NOTES languages/messages/MessagesZh.php): update for the generic Chinese message file, when the $wgContLanguageCode is set to 'zh', reflecting the site-wide settings when it is set to 'zh' [02:55:22] Whoah, this place is loaded with people. Anyone able to answer what I hope is a simple question? [02:55:35] Been beating my head against the same wall, and it's starting to look worse for wear. [02:56:37] I've probably just missed it because it's been returning such generic search results, but it's still giving me grief :P [02:56:58] Basically I need to replace text in a variable. Specifically, I need to turn a space into %20 [03:10:17] Pinchie: in some articles? [03:10:21] in what context [03:16:56] Sorry, been making lunch. [03:18:08] Well I'm setting up a template for people's XBox Live account info to be presented all pretty-like. However, the URL ends up as http://card.mygamercard.net/{{{gamertag}}}.png - that's fine as long as the gamertag doesn't have spaces. [03:18:20] !magic | Pinchie [03:18:20] Pinchie: For more information about creating magic words and their inner workings, see . For a list of magic words, please see . [03:18:29] i think you need {{urlencode:}} [03:18:54] Excellent, I'll read up on that. Thank you :) [03:19:01] no problem [03:25:39] Hmm, that also does it to the braces that define the gamertag instead of doing it after parsing the variable :/ [03:26:07] Now Playing: Aretha Franklin - Bridge Over Troubled Water (Cosmonica: Crazy Covers Volume 1) [5:29] [03:26:16] Wrong channel for the announce, sorry. [03:28:03] is there anyway to turn off section editing on just one section? [03:36:20] Now Playing: Gotye - Karnivool - The Only Way (Mixed Blood) [4:55] [03:41:02] 03(NEW) Enable more URL schemas/ protocols for the english wikinews (mms: and rtsp:) - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=12204 15enhancement; normal; Wikimedia: Site requests; (bawolff+wn) [03:49:43] It did work in the end. Thanks McBride :) [03:49:54] good to hear :) [03:50:08] http://live.xbox.com/en-US/profile/profile.aspx?pp=0&GamerTag={{urlencode:{{{gamertag}}}}} [03:50:21] I was parsing the whole address instead of just the variable :) [03:50:37] heh, you live, you learn [03:50:38] Thank you very very much :) [03:50:44] 03shizhao * r28175 10/trunk/tools/planet/zh/ (config.ini templates/index.html.tmpl): fix italics and case [04:55:35] hi all [04:56:17] 03(mod) New:=?UTF-8?Q?=20reduce=20the=20BiDi=20read?= =?UTF-8?Q?ability=20problem=20at=20=C2=AB?==?UTF-8?Q?=20action?== =?UTF-8?Q?credits=20=C2=BB=20if=20the=20win?= =?UTF-8?Q?dow=20size=20is=20changed?= - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=12205 (10N/A) [04:56:24] how can I use the templates on wikipedia on my offline mediawiki ? [04:56:34] does a simple copy source works ? [04:57:34] i'm setting up a small wiki. as sysop, i want notification of all changes, and new account creations. [04:58:02] should i use enotif and a wildcarded watchlist (is that possible?) [04:58:20] and how can i tell mediawiki to email the sysop when an account is created? [04:59:10] - 10http://bugzilla.wikimedia.org/show_bug.cgi?id= (10gangleri) [05:00:26] - 10http://bugzilla.wikimedia.org/show_bug.cgi?id= (10gangleri) [05:01:58] - 10http://bugzilla.wikimedia.org/show_bug.cgi?id= (10gangleri) [05:02:24] - 10http://bugzilla.wikimedia.org/show_bug.cgi?id= (10gangleri) [05:03:03] o_O [05:03:22] corduroy: I'd be looking at an extension for that [05:03:45] i don't think one exists for that particular feature [05:03:59] yeah, but it wouldn't take a huge amount of time to make one [05:04:09] I suspect there's a hook in RecentChange.php [05:15:01] 03(NEW) Vague error on captcha mismatch during login - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=12206 normal; normal; MediaWiki extensions: ConfirmEdit; (tstarling) [05:18:00] 03(mod) New:=?UTF-8?Q?=20embed=20the=20items=20at=20c?= =?UTF-8?Q?ategory=20pages=20in=20=C2=AB=20?==?UTF-8?Q?=20=C2=BB=20blocks?= - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=12207 (10N/A) [05:27:30] is there any way to get a plain text diff of a page? [05:27:52] 03(mod) New:=?UTF-8?Q?=20use=20=C2=AB=20nowrap=20=20=C2=BB=20fo?= =?UTF-8?Q?r=20items=20displayed=20on=20?==?UTF-8?Q?=C2=AB=20?=[[special: Reventchanges]]=?UTF-8?Q?=20=C2=BB=20and=20similar=20speci?= =?UTF-8?Q?al=20pages?= - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=12208 (10N/A) [06:05:33] thanks for the suggestions. i didn't realize there were hooks. i'll check it out. [06:10:25] Goddammit .. who broke the limits on the API again >:() [06:13:56] how is mediawiki support for postgresql currently? [06:14:18] Good, I believe [06:14:51] i thought we broke postgres support altogether recently [06:14:53] - 10http://bugzilla.wikimedia.org/show_bug.cgi?id= (10gangleri) [06:16:54] there was the CONCAT problem in 1.11.0 [06:17:06] the fix for that is committed [06:17:27] we should probably backport it and do a release [06:18:27] wikibugs isn't a big fan of gangleri obviously [06:19:25] heh -- I've noticed that :) [06:19:37] wikibugs keeps cutting of titles at random ... [06:20:07] i thought wikibugs always choked on weird characters.... [06:20:18] Ah, maybe that's the cause [06:20:38] I wouldn't call all characters outside ISO 8859-1 "weird" [06:21:10] some of the bugs use spanish quotation marks; wikibugs gets quite angry about that [06:21:22] *amidaniel wonders who maintains wikibugs [06:21:28] Is it just a bugzilla plugin? [06:22:08] brion [06:22:39] it [06:22:45] parses the mailing list, methinks [06:23:31] top [06:23:39] er... pardon me [06:23:50] *Werdna goes off and finds something to do [06:24:08] TimStarling: anything in particular I should look at? [06:24:45] regarding what? or are you just looking for something to do? [06:24:55] something to do. [06:25:06] *Werdna is trawling through bugs at the moment [06:26:28] would you mind taking a quick look at one for me? please? [06:27:05] oh? [06:27:18] http://bugzilla.wikimedia.org/show_bug.cgi?id=12185 [06:28:51] I could swear I enabled that when I did some autoblocking changes about this time last year [06:28:54] but I'll check the code [06:29:22] what about bug 9413? [06:31:46] MZMcBride: A look at the code indicates that this code is correct. [06:32:21] i don't understand the function; from personal experience, a user is able to vandalize, get blocked, and then just create a new account [06:32:30] 03amidaniel * r28176 10/trunk/phase3/includes/api/ (ApiQueryDeletedrevs.php ApiQueryRevisions.php): API: Standardize limits. Use ApiBase constants to avoid similar breakage and inconsistencies in the future. [06:32:52] MZMcBride: do you enable autoblock when blocking? [06:32:52] wouldn't that be a huge issue on wikis where direct db access or checkuser isn't possible [06:33:08] TimStarling: bug 9413? Arabic presentation forms? [06:33:32] i checked the "Automatically block the last IP address used by this user, and any subsequent IPs they try to edit from" box [06:33:48] and they were still able to create a new account from the same IP? [06:33:54] but the identical IP was able to create a new account according to rc_ip [06:34:05] 03amidaniel * r28177 10/trunk/phase3/RELEASE-NOTES: Always miss the RELEASE-NOTES. [06:34:05] did you see the comment in the code about schools? [06:34:24] MZMcBride: Did you check "prevent account creation" ? :) [06:34:33] when i brought the issue up in here, another user reported the same issue and filed the bug [06:35:02] yes, arabic presentation forms [06:35:22] amidaniel: (account creation disabled, e-mail blocked) from the logs [06:35:32] it's a short, fairly unusual project [06:35:36] Oh, I see, when blocking an *account* not an ip [06:35:47] yeah [06:35:53] 03(mod) Enable more URL schemas/protocols for the english wikinews (mms : and rtsp:) - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=12204 +shell (10raimond.spekking) [06:36:04] We don't actually target username blocks to ips, so it would be rather inconsistent behavior. [06:36:19] I do agree that it's a little counterintuitive to have checkmarks there that do nothing :) [06:36:28] why not? [06:36:34] well, I'll need quite a bit of guidance here (although the challenge is nice) I know that we have something somewhere that already does that kind of conversion, hmm? [06:36:47] without checkuser or db access, the user can continue to vandalize forever, right? [06:37:06] the includes/normal directory [06:38:00] it's called from WebRequest::normalizeUnicode() [06:38:18] MZMcBride: Yeah, probably :) [06:38:33] MZMcBride: Talk to Werdna .. he's the ipblocks guy :) [06:38:36] $data = UtfNormal::cleanUp( $data ); [06:38:51] I don't see why this couldn't be done with autoblocks [06:38:56] and that isn't a problem? it seems like a rather large problem to me [06:39:16] MZMcBride: I'm not saying it's not a problem, just saying it requires work to fix :) [06:39:29] ok :) [06:39:43] *amidaniel could've sworn that autoblocks prevent acct creation though ... [06:39:46] *amidaniel digs [06:39:49] TimStarling: yes, I had a quick read of the readme [06:40:25] from your comments on the bug, do I need to write code somewhere to determine whether to convert using NFC/NFCK? Or should I just add the new data to the .inc files in there? [06:40:31] Or do I need to write a decomposition routine? [06:41:13] if( $wgUser->isBlocked() ) { $this->mainLoginForm( wfMsg( 'blocked-mailpassword' ) ); [06:41:18] Shouldn't that do it? [06:41:31] *amidaniel checks User.isBlocked [06:41:54] well, the conversion we're talking about isn't NFC or NFKC, it's a hybrid [06:42:05] amidaniel: did you see the comments in the code? they seemed to indicate the function was intentional.... which really confuses me [06:42:26] right - so I /do/ need to write something to figure out whether I should use NFC or NFCK? [06:42:29] maybe you could refactor the conversion routine to allow an arbitrary code table to be passed to it [06:42:54] then make a code table for these arabic presentation forms [06:43:00] o_O [06:43:07] merged with the NFC conversions... [06:43:10] *Werdna looks at UtfNormal.php [06:43:30] MZMcBride: WORKSFORME [06:43:43] Account creation from this IP address (127.0.0.1) has been blocked by AmiDaniel. [06:43:46] The reason given by AmiDaniel is Autoblocked because your IP address has been recently used by "AmiDaniel". The reason given for AmiDaniel's block is: "" [06:44:04] did you block an IP or a user account? [06:44:12] then add a new function that does the hybrid conversion, and have UtfNormal::cleanUp call that instead of NFC [06:44:17] MZMcBride: User account [06:44:24] As the message would indicate [06:44:32] Can you show me an example of where this didn't work? [06:45:02] hmmm... not sure how to... [06:46:08] Works on enwiki too [06:46:22] Let me try with a couple different blocker settings [06:46:22] I'm going to need to understand this a little better. It looks like 'NFD' is a standard "decomposed" representation, where all of the combined letters etc are converted into some semblance of a standard form. [06:46:42] NFC is a recomposed version of NFD [06:46:45] amidaniel: are you running more current than 1.11? [06:46:51] yes, that's how the standard puts it [06:47:02] NFC is the composed form, and NFD is the decomposed form [06:47:10] so the idea seems to be that, to convert something into NFC, one must decompose it entirely, and then recompose it in a standards-compliant way [06:47:14] so it's a type of normalisation [06:47:31] you don't need to convert to NFD [06:47:32] MZMcBride: Ah, I bet I know what the problem is. In order for the autoblocker to trigger, the user has to try to make an eidt [06:47:35] edit* [06:47:35] that's just standardese [06:47:38] KD form seems to be a backwards-compatible. [06:47:45] amidaniel: no they don't. [06:47:46] If he doesn't edit, he'll be able to make an account without getting caught by the block. [06:47:53] Werdna: Yah, they do :) [06:47:55] amidaniel: ever heard of retroactive autoblock? [06:47:56] Just tested it :) [06:48:05] I wrote it back in 2006 [06:48:12] Well, then it's not working :) [06:48:20] $wgPutIPInRC needs to be on, though [06:48:35] TimStarling: but that's how UtfNormal does it currently. [06:48:56] is that so? [06:49:00] TimStarling: NFDs it, then composes the NFD'd version to produce an NFC'd version [06:49:06] static function NFC( $string ) { [06:49:06] return UtfNormal::fastCompose( UtfNormal::NFD( $string ) ); [06:49:07] } [06:49:43] yep, so it does [06:50:05] amidaniel: so you think it's an edit vs. create thing? [06:50:31] Werdna: Ah, indeed. Of course, it assumes the user has an edit / log entry in the db, and that the ip is being stored in the database. [06:50:36] but the characters in question aren't affected by either transform [06:50:38] so I'm guessing that fastCompose uses a codepage to convert NFD to NFC, and a different one to convert NFD to NFKC [06:50:43] MZMcBride: No [06:50:54] ah, I see. [06:51:17] MZMcBride: What wiki was this on that you had the problem? [06:51:35] a personal wiki running 1.11 [06:51:47] if( isset( $map[$c] ) ) { [06:51:47] $out .= $map[$c]; [06:51:54] MZMcBride: That would explain it then. [06:51:55] line 518 [06:51:58] aha [06:52:06] amidaniel: huh? [06:52:09] MZMcBride: Enable $wgPutIPInRC and get newuserlog running [06:52:20] it's enabled by default [06:52:25] O [06:52:32] that's how i was able to know the user was the same [06:52:32] fastDecompose() already has a map parameter [06:52:36] and i have newuserlog [06:52:50] so you can just make a new map for a hybrid NFD/NFKD [06:52:56] Werdna: What else do we need for retroactive autoblock? [06:53:24] then compose that and then you'll have what we what [06:53:34] MZMcBride: Well, on my local wiki with all default settings, it no worky. On enwiki, it worky :) [06:53:40] OK, so where does $map come from? [06:53:40] hybrid NFC/NFKC [06:53:40] So some configuration setting in between there :) [06:53:58] you've got a search function in your editor don't you? [06:54:09] it's not my code, I can't answer these questions off the top of my head [06:54:11] brion wrote it [06:54:14] OK. [06:54:39] Werdna: ? [06:54:50] amidaniel: on my test wiki, it worked as it was supposed to... gr... [06:55:28] MZMcBride: Okay, well, I'm marking as WORKSFORME. It's an autoblocker problem if anything. So far as I can tell it's all working according to spec [06:55:35] Except maybe Werdna's code :) [06:55:42] Werdna don't worry too much about the ICU stuff, nobody uses that [06:55:46] all right; thanks for your help [06:55:57] amidaniel: look at Block.php, line 420ish [06:56:05] TimStarling: okay [06:58:09] 14(WFM) Change "block account creation" functionality when blocking usernames - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=12185 +comment (10cannon.danielc) [07:01:41] Werdna: Hmm .. well, for some reason it didn't work for me [07:01:58] Your query on "AmiDaniel" gives me back | 127.0.0.1 | though [07:02:43] TimStarling: so I should be adding a new elseif to UtfNormal::cleanUp, to determine if I'm looking at arabic text, and call UtfNormal::fastCompose accordingly? [07:02:43] *amidaniel does some playing [07:03:01] *Werdna offers amidaniel $wgDebugLogFile [07:03:50] Hmm .. that time it did [07:04:03] To hell with you and your rational ways of debugging! [07:06:28] Werdna: I think you could do the arabic conversions unconditionally [07:06:43] it doesn't need to be arabic text [07:07:16] won't that be expensive, performance-wise? [07:07:33] don't see why, it's all pure PHP already [07:07:36] unless I write a "check that it conforms to this codepage" routine, or something? [07:08:27] there's quickIsNFCVerify(), that would need to be tweaked I suppose [07:08:32] and, if we can do them unconditionally, why shouldn't I add the arabic translations to the standard codepage anyway? [07:08:50] you can, as long as you don't call it NFC [07:08:58] because that would be confusing [07:09:14] it's currently called $utfCanonicalComp [07:09:31] 03(NEW) Invalid zh-classical interwiki in maintenance/ wikipedia-interwiki.sql - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=12209 04BLOCKER; normal; MediaWiki: Installation; (alefzet) [07:09:44] that's short for canonical composition [07:10:06] as I said on the bug report, the arabic mappings are *not* canonical mappings [07:10:43] right, so it COULD work, but it's not good practice to do it that way [07:11:27] right [07:11:35] and there's a few other little things you'd have to check, like this: [07:11:42] # Head bytes for sequences which we should do further validity checks [07:11:42] $checkit = array_flip( array_map( 'chr', [07:11:42] array( 0xc0, 0xc1, 0xe0, 0xed, 0xef, [07:11:42] 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7, [07:11:43] 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xff ) ) ); [07:11:51] is that still accurate? [07:13:14] well, by my reading of your bug report, you need to look at characters that start with fb,fc,fd,fe. All of those seem to be in that list. [07:15:16] $checkit is UTF-8, my bug report was using codepoints [07:15:47] what's the relationship? [07:16:29] I thought this might be an interesting introduction to unicode for you :) read the UTF-8 spec, it's short and simple [07:16:43] *Werdna looks [07:16:49] it'll explain it better than I could here [07:17:31] *amidaniel begins to foam at the mouth after hearing mention of utf-8 [07:26:27] Werdna: have you got the answer yet? [07:26:47] if not, I can tell you [07:27:55] I've been reading [07:28:13] 03(mod) Invalid zh-classical interwiki in maintenance/ wikipedia-interwiki.sql - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=12209 04BLOCKER->trivial (10tstarling) [07:28:18] I know what a code point is now, but I don't know how it relates to the actual internal representation of the character. [07:28:51] http://www.faqs.org/rfcs/rfc2279.html [07:28:54] (I don't know what a head byte is) [07:28:57] section 2 [07:29:01] aha [07:29:10] *Werdna was reading elsewhere [07:29:41] it has a table giving the codepoint ranges and their relationship to UTF-8 [07:30:14] TimStarling: If a diversion from strict ontopicness is allowed; have you read http://en.wikipedia.org/wiki/An_Exceptionally_Simple_Theory_of_Everything ? [07:30:55] 0800-FFFF start with binary 1110 [07:31:17] notagerbil: no [07:31:57] the other 4 bits of the "head byte", as brion calls it, are the same as the 4 most significant bits of the codepoint [07:32:35] it is tagged with a notability guideline violation tag. and it is what you used to study, is why I am asking... Maybe Michael Hardy knows... [07:32:37] the most significant nibble of the codepoint is hexadecimal F [07:33:25] 1110 is hexadecimal E [07:33:33] is that 0800 TO FFFF, or 0800 hyphen FFFF [07:33:35] though I find it kind of mind boggling that a candidate TOE would not be notable. [07:33:50] to [07:34:14] so the arabic presentation forms, when written in UTF-8, all start with 0xef [07:34:44] which is in the list already [07:34:53] yep [07:35:07] *Werdna breathes. [07:37:29] hmmm.. but it looks like these character definitions / code tables are pulled from the internet [07:39:11] they're taken from UnicodeData.txt, which comes from unicode.org [07:39:16] right. [07:39:34] so shouldn't I refrain from modifying that to add arabic stuff? [07:40:02] you have to get the arabic stuff from UnicodeData.txt, that's where it is already [07:40:53] the entries for the presentation forms in UnicodeData.txt give their equivalents in 0600-06FF, the arabic block [07:41:21] hello [07:41:23] ah Isn't that what we're after? [07:41:29] so you don't need to modify UnicodeData.txt, you just need to modify the routine that parses it [07:41:49] which is in UtfNormalGenerate.php. [07:43:25] if it's canonical.. it gets shoved in one array. If n ot, another array, by the looks [07:44:04] yeah, you want to make a hybrid array that contains all the canonical mappings plus the compatibility mappings for the presentation forms only [07:44:08] which means that I should use a third array [07:44:34] ALL the canonical mappings, or just the arabic ones? [07:44:42] just the arabic ones [07:45:41] hmmm [07:46:14] sorry, like I said the first time, ALL the canonical mappings plus the compatibility mappings for the presentation forms [07:46:26] right [07:46:55] because you want to make a single hybrid transform that operates on either arabic or non-arabic text [07:47:33] and use that instead of the normal expanding code page? [07:47:43] s/code page/transform/g [07:47:48] yes [07:49:47] so just rename the canonical transform, then? [07:49:57] (and add the arabic presentation forms to it) [07:50:25] if that's the way you want to do it [07:55:30] is there a way to do inline external image links in mediawiki sites (default install) [07:55:30] ? [07:59:27] happy chanukka btw [07:59:42] luqui: [[:Image:SomeImage.jpg]] [07:59:45] notice the starting : [07:59:56] TimStarling: how's this? http://www.devanywhere.com/ViewPub.php?id=48 [08:00:55] hmm, and SomeImage.jpg can be an external url like http://foo.com/bar.jpg ? [08:02:04] [http://some.url/path] [08:02:46] he wants to embed an external image, i.e. [08:02:47] Werdna, still getting just a [1] that links to the path. I want the image to be displayed on the pag [08:02:52] oh right [08:02:57] there's a setting somewhere [08:03:05] there's a configuration option to allow it, $wgAllowExternalImages i think [08:03:06] (I know that's not very helpful) [08:03:08] try the FAQ [08:03:13] !faq [08:03:13] Before reporting a problem or requesting assistance, please check the FAQ first. The FAQ for MediaWiki can be found at http://www.mediawiki.org/wiki/Manual:FAQ [08:03:24] ah, don't have admin on this one, so probably can't do it. ok. [08:03:42] Werdna: yes, looks good [08:04:10] are there any gotchas I need to look out for? [08:04:27] (like that list of head bytes you mentioned earlier) [08:04:34] don't think so [08:05:09] make sure you test it though [08:05:25] eugh, therein lies the problem [08:05:33] testing is half the fun with this one, yes :) [08:05:51] is there a regression test suite I can run somewhere? [08:07:20] <_wooz> lo [08:08:52] UtfNormalTest.php? [08:08:59] aha [08:09:17] but will that allow me to test if the changes are working as expected? [08:09:23] *Werdna looks [08:15:14] 03david * r28178 10/trunk/extensions/LiquidThreads/ (7 files): Show notice when threads have been edited, with smarts. [08:40:59] hi all [08:41:13] hullo [08:42:31] I was wondering if anone knows where i can find MW themes, because the default one to me gets confusing when you have more than 4 subtitiles (=x=, ==x==, ===x===, ...), After the 4th you can't really destinguish between them because they're small. [08:42:39] they're the same * [08:42:52] they're called skins. try searching for that. [08:43:54] aha, i knew somthing was wrong.. thanks ;) [08:46:24] Hello, I'm trying to rewrite URLs to http://wiki.example.com/Page_Name, but I'm getting an error in my Apache log whenever I try to access pages with a ":" - (20024)The given path misformatted or contained invalid characters: Cannot map GET /Special:Userlogin HTTP/1.1 to file, referer: http://wiki.example.com/ [08:46:40] I'm using MediaWiki on Windows. [08:47:13] (rather, on a Windows server with Apache 2.2.x [08:47:31] and PHP 5.2.5) [08:47:57] Why am I having this problem? I've never seen it before. [08:50:11] It's working fine for everything except pages in namespaces and those with a ":" in the name. [08:55:16] Hm, I gotta run now. [09:05:34] strange [09:05:46] if i gtry and go to mediawiki.org, it forwards me to www.wikipedia.org, and fails to load [09:06:02] "this page isn't redirecting propperly" [09:06:16] works for me [09:07:27] nm, it's our stupid ISA array.. [10:26:40] 03(NEW) API RC Feed Missing User - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=12210 minor; normal; MediaWiki: API; (chrisgrantmail) [10:27:48] 03(mod) API RC Feed Missing User Name - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=12210 summary (10chrisgrantmail) [11:09:16] how can i setup the base size for svg (vector) images? [11:15:49] 14(WFM) API RC Feed Missing User Name - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=12210 +comment (10Andrew) [11:19:14] 03(WONTFIX) Implement the Pure Wiki Deletion System - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=3843 +comment (10Andrew) [11:21:53]