above globalwrapper on MonoBook? [01:46:17] celeretaudax: look for a hook that'll let you do it, or add it manually into the skin [01:47:01] celeretaudax: meaning into Monobook.php [01:47:02] I tried adding it manually but I can't get it to stay above the globalwrapper [01:47:58] eh? how so? how could it go anywhere else? heh [01:48:51] I have no clue [01:48:54] 03dale * r39517 10/trunk/extensions/MetavidWiki/ (13 files in 6 dirs): some updates to the stream interface [01:49:13] 03(mod) Add rd_interwiki colum to redirect table - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=14418 (10innocentkiller) [01:49:35] what are you trying to do? [01:50:08] just put a 100% width divider above everything on the page [01:51:10] and when does it not stay? [01:51:29] did you remember to clear your browser cache? [01:51:40] you may be hitting cached pages [01:51:50] yeh I checked that, on different boxes too [01:52:13] when does it disappear? [01:52:38] it doesn't it just shows up below the top navigation tabs [01:52:56] oh. well, thats interesting [01:54:15] have you tried putting it right after the body tag? [01:54:31] yeah [01:54:36] no clue why it doesn't work [01:54:40] other than floats and z-index [01:57:07] i'd have to guess that is the reason. i tried it and I see what you mean [01:57:20] yeah, I'll have to play with it more I guess [01:57:37] could always make your own skin [01:57:52] thats a hell of a lot of work though [01:58:17] yeah I already changed a lot of monobook, might as well go all the way [02:21:13] I'm having an unusual issue with ShortURLs and wondered if anyone had ever come across it.. for some reason any URL with an apostrophe in it seems to cause an Internal Server error and I can't work out why, much as I try. [02:29:47] Orlando: do you have a link to the site? [02:29:57] or is it private? [02:30:19] Not private, just messy and chaotic right now :) [02:30:27] http://www.arkaz.come/wiki/ [02:31:57] Maybe it's more of an apache or mod_rewrite question than anything. I tried removing index.php and replacing it with a one line script that echos "test" just to see if it was anything in the scripts, but it was pretty much the same deal. [02:32:00] how are you doing the short ur;s? [02:32:04] 14(INVALID) ipblock-exempt exists where it shouldn't and isn' t listed where it should be - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=15193 +comment (10cbrown1023) [02:32:24] http://www.mediawiki.org/wiki/Manual:Short_URL/wiki/Page_title_--_no_root_access [02:32:27] That way :) [02:32:53] I removed the last line though "RewriteRule ^/*$ wiki/ [L,QSA]" as I didn't want all urls rewritten, only those with /wiki/ in them. [02:33:45] And it works fine for everything, except when the URL has an apostrophe, then it dies. I'm guessing it's an apache/mod_rewrite/server setup issue somewhere in there, more than anything. [02:35:21] yeah, thats likely [02:35:23] 04(REOPENED) ipblock-exempt exists where it shouldn't - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=15193 summary; +comment (10cbrown1023) [02:35:27] what do your logs say? [02:35:59] Shared Host, alas. Otherwise I'd have checked them ages ago. The joys of shared hosting, etc. [02:36:17] they don't give you your own logs? [02:36:17] If the hosting account hadn't been paid for for about a year, I'd have moved it from that oone [02:36:24] not the apache error_log, at least. [02:36:31] wow. thats lame [02:36:39] It does make things tricky, yea :/ [02:36:43] i have crappy shared hosting, and i have access to mine [02:37:05] I think I'm doomed to just not use short urls, lol [02:38:10] But yea, it is a little bit lame. Access to the error_log is pretty essential for stuff like this. [02:39:14] Strangely enough they do give me access to the access_log though, heh [02:39:31] yeah [02:39:44] pastebin your rewrite rules [02:40:10] Sure, one sec [02:42:22] http://www.pastebin.ca/1175232 [02:44:05] I tried their tech support, but they basically gave me a "tech support" style answer. [02:45:26] Orlando: try prepending a / to w/index.php?title=$1 on line 23, and also to wiki/ on the next line [02:45:58] *grin* Tried both already, but I'll give it another short [02:46:02] so it'd look like /w/index.php?title=$1 and /wiki/ instead [02:46:41] Orlando: arent you missing PT on your rules? [02:46:57] Same thing, alas. And it's /just/ with URLs that have an apostrophe. [02:47:10] Ryan_Lane: not according to http://www.mediawiki.org/wiki/Manual:Short_URL/wiki/Page_title_--_no_root_access at least. [02:47:15] does anyone else think Rebecca Soni is pretty darn cute? [02:47:34] Orlando: RewriteRule ^wiki/(.*)$ w/index.php?title=$1 [PT,L,QSA] [02:47:38] Schroeder: this is a tech support channel. go elsewhere. [02:47:50] Oh yea, I must have removed that from my endless messing with things. [02:48:12] i doubt that'll help though [02:48:28] Seems not, again, sadly :/ [02:48:41] I'll probably try their tech support, again, and get another tech support answer [02:49:09] Orlando: wait, I think I've found the problem [02:49:34] you need RewriteCond to detect apostrophes, rather than RewriteRule [02:50:26] What would the rewrite rule need to be changed to? [02:50:41] looking now. :) [02:52:19] hmm, actually, no [02:52:26] it's likely a server limitation. :/ [02:52:48] Joys. :D [02:53:02] I did wonder. I've had it work fine on other servers [02:53:49] I know they don't have exec disabled for php, wonder if I can just create a symlink ... [02:54:27] At least, I think they're a symlink method. I can't recall now *goes back to check the wiki* [02:58:56] Orlando: maybe you can escape the apostrophes? [02:59:29] tried that too :D [03:00:01] and tried replacing/removing them entirely [03:00:03] No joy :( [03:00:24] have you tried adding NE to the flags at the end? [03:00:31] for 'no-escape' [03:00:45] Orlando: I mean with "RewriteMap escape int:escape" [03:00:54] ie: RewriteRule ^wiki/(.*)$ w/index.php?title=$1 [PT,L,QSA,NE] ? [03:01:11] Ryan_Lane: Not sure on that method, you'll have to talk me thought it some. ;) [03:01:45] Orlando: put it right before the other rules [03:02:15] Orlando: right after the comments about "mediawiki shorturl" [03:02:29] alrighty, fingers crossed [03:02:30] *Orlando updates [03:03:21] Hm, that just leaves me with every page giving 500 errors [03:03:26] lol [03:03:40] well, I guess that didn't work so well :) [03:04:00] hmm [03:04:05] it did let me create a symlink though! [03:04:06] ack [03:04:09] i forgot something [03:04:11] It's like.. strange useless magic. [03:04:21] sorry :) [03:04:27] no worries [03:04:32] symlink was a bad idea it seems.. [03:04:41] What's the other thing I need to do? [03:05:01] RewriteRule ^wiki/(.*)$ w/index.php?title=${escape:$1} [L,QSA] [03:06:11] oh, you probably still want PT there too [03:06:22] http://www.arkaz.com/wiki/Main_Page%27test <- still errors :< [03:06:28] so: RewriteRule ^wiki/(.*)$ w/index.php?title=${escape:$1} [PT,L,QSA] [03:06:51] I may just have to mark this one down as webhosting company fail [03:06:52] ah, lame [03:06:59] maybe so [03:07:17] I've spent like.. 2 weeks on it trying various things in the evening. :D [03:07:51] yeah, it totally works on my hosted site [03:08:00] i'd blame them :) [03:08:22] The hard part is convincing them to actually look at their setup :D [03:08:38] ask them to do the alias setup for you [03:08:47] lol, I did. [03:08:50] it is far more reliable [03:08:52] oh [03:08:53] They were not much use [03:08:56] heh [03:09:02] ... unsurprisingly, I imagine [03:09:10] what service is this? [03:09:15] i'll make sure to avoid them [03:09:33] Blue Domino generally [03:10:36] They're not all bad, on the whole. They provide some useful stuff. But this one is really not awesome. [04:00:04] 03(mod) ipblock-exempt exists where it shouldn't - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=15193 (10mike.lifeguard) [04:04:19] is there a way to make an image follow your screen as you are scrolling in mediawiki [04:05:56] YEs, lemme just get it... [04:06:37]

[[Image:X.jpg]]

[04:09:03] that works except if I use [[image:x.jpg|right]] it doesnt stay to the right anymore [04:25:22] has anyone noticed that the edit involved in moves tends to get lost when a move is undone? [04:28:03] Hello? [04:28:39] hey [04:29:28] I have a question about MediaWiki namespaces [04:29:34] can you help? [04:29:39] Best to ask. ;-) [04:29:49] Sweet! [04:30:08] Well I am trying to create a notes like system to use for school...for me and my brother [04:30:21] but I goto CollegeA and he goes to CollegeB [04:30:36] and I have classes (for example) Chemistry, Math and English [04:30:50] and he has Chemistry, Physics and PreCalc [04:31:14] and then for each class I want to have a page of notes named by their date (example: August 1 2008) [04:31:24] so then....the College would be the Category [04:31:30] the Class would be a subcategory [04:31:36] and the page of notes would be a Page [04:31:57] the only problem is that I cna't have 2 august 1 2008 in the Wiki [04:32:06] so someone recommended having namespaces [04:32:17] What about using subpages? [[College/Class/August 2 2008]] ? [04:32:18] so heres my question: can I have sub-namespaces? [04:32:28] No, sub-namespaces aren't possible. [04:32:28] oooo [04:32:35] But you can have subpages of namespaces. [04:32:37] thats like exactly what I want to do MzMcBride! [04:32:43] hm... [04:32:51] They'll have a little 'breadcrumb' at the top for navigation. [04:33:09] subpages...so then the pages are seperated by like slashes (/) [04:33:17] what do you mean by breadcrumb [04:33:45] foo >> bar would be for foo/bar/baz [04:34:00] oh ok [04:34:06] yea..thats exactly what I want [04:34:11] so how hard is that to setup? [04:34:34] what's the opposite of hard? [04:34:37] ;) [04:35:14] hmm..ok [04:35:25] so is there a website that shows me how to easily set it up? [04:37:16] type it in the location bar, edit page. [04:37:16] Just create a page named "Foo/Bar" [04:37:32] If you're using the (Main) namespace, you'll need to enable subpages. [04:37:46] is the Main namespace the default one? [04:38:14] Yes. [04:38:26] http://www.mediawiki.org/wiki/Manual:$wgNamespacesWithSubpages [04:38:33] That's the variable you need to change. [04:39:03] and is there a way I can control user permissions with the subpages? because I only want my brother to edit the pages for his college [04:39:20] You can restrict editing somewhat. [04:39:23] !rights | Notes [04:39:23] --mwbot-- Notes: For information on customizing user access, see < http://www.mediawiki.org/wiki/Help:User_rights >. For common examples of restricting access using both rights and extensions, see < http://www.mediawiki.org/wiki/Manual:Preventing_access >. [04:40:04] man...you guys are awesome [04:40:08] this is really helpful [04:43:26] 03(NEW) Special:Log/Username should work - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=15194 15enhancement; normal; MediaWiki: Special pages; (MZMcBride) [04:49:55] can I create user groups? [04:50:06] so that I can allow GroupA to access this area of my wiki [04:50:14] and GroupB to NOT access that same area? [04:51:09] techdude2007: http://www.mediawiki.org/wiki/Manual:Preventing_access [04:51:59] and how do I put users into Groups? [04:52:08] Special:UserRights. [04:52:21] To create groups, you edit LocalSettings.php. [04:52:21] http://www.mediawiki.org/wiki/Manual:User_rights [04:53:05] 03(WONTFIX) Special:Log/Username should work - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=15194 +comment (10JSchulz_4587) [04:53:38] AaronSchulz: Bah. [04:53:49] AaronSchulz: https://bugzilla.wikimedia.org/show_bug.cgi?id=15146 [04:55:54] does anyone have any advice on how to correctly add line breaks or paragraphs to MediaWiki:Newarticletext ? [04:56:19] Hit return twice? [04:56:22] when I add a double line, it doesn't work. [04:56:47] MZMcBride: originally, it had single line breaks, and still split it into paragraphs [04:56:48]

[04:56:48] 03(mod) Special:Log/Username should work - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=15194 (10N/A) [04:56:51] Or

[04:57:11] MZMcBride: that's what I thought, but when I add
it puts it in a paragraph! [04:57:17] I'll try

[04:57:24] what does that Special:Log/Username do? [04:57:35] Nothing at the moment. [04:57:38] It's a feature request. [04:57:48] and I see your the creator [04:57:52] what is the plan for it to do? [04:57:57] ok,

works. [04:58:00] strange [04:58:13] techdude2007: To list logs for a specific user. [05:00:29] oh ok [05:06:39] it works now, actually. :) [05:06:40] 03krimpet * r39518 10/trunk/phase3/includes/specials/SpecialLog.php: Per request in bug 15194, add more syntax for Special:Log, similar to Special:ListUsers changes in r39269 [05:07:20] 03(FIXED) Special:Log/Username should work - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=15194 +comment (10francine) [05:20:22] how do i change the font on something i am posting [05:23:58] ... [05:24:09] s/Arial/whatever font you want. [05:24:45] thanks [05:31:11] MZMcBride where does it get these fonts from? my local pc or the server? [05:31:24] It uses the fonts on your computer. [05:32:06] nice [05:36:26] 03(mod) cat_hidden field isn't used - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=15191 (10tstarling) [05:43:23] 03(WONTFIX) Bizarre parser error - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=14376 +comment (10tstarling) [05:57:02] *Werdna slaps MZMcBride [05:57:15] :O [05:57:19] What'd I do now? [05:57:24] You should never set a font-family without a fallback to one of: sans-serif, serif, monospace [05:57:43] The browser will fallback if it's not set inline, no? [05:57:53] If not, the browser sucks. [05:57:56] Yes, but it has to know what to fall back to.. [05:58:11] MediaWiki specifies sans-serif. :P [05:58:30] Though, yes, generally one should specify. [05:59:40] 03krimpet * r39519 10/trunk/extensions/CheckUser/ (CheckUser.php CheckUser_body.php): Per bug 14839, split the 'checkuser' right into 'checkuser' and 'checkuser-log', to facilitate creation of an ombuds/oversight group; the stock CheckUser group will continue to hold both rights. [05:59:58] *Werdna tickles Krimpet [06:00:06] Bugzilla seems a bit slow... [06:00:07] *Krimpet stabs Werdna by reflex [06:00:52] Krimpet: did you want to be centralauth'd? [06:00:58] Werdna: sure. :) [06:01:14] okay, include /home/andrew/local_wiki_config.php in your LocalSettings.php [06:01:23] 03(FIXED) Split the checkuser right into checkuser and checkuser-log - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=14839 +comment (10francine) [06:01:32] Krimpet: and then run the migratePass0 script. [06:05:36] Bah. svn.wikimedia.org is slow for me as well. :/ [06:05:56] Warning: Invalid argument supplied for foreach() in /home/krimpet/public_html/lua/includes/db/LBFactory_Multi.php on line 160 [06:06:15] how do i center a header [06:06:21] like ==Blah== [06:06:37]

...

? [06:06:59] I see, is there a wiki way of doing it? [06:07:10] Not really. [06:07:14] ok thanks for the help :) [06:07:41] Werdna: see above [06:08:07] moment. [06:09:18] *Krimpet looks for more bugs to fiddle about with [06:09:27] Krimpet: try now. [06:11:08] Fatal error: Class 'CentralAuthUser' not found in /home/krimpet/public_html/lua/extensions/CentralAuth/migratePass0.php on line 38 [06:12:54] oh, did you enable the CentralAuth extension? [06:13:01] that's not done for you. [06:13:28] I did [06:14:33] *Werdna has a look [06:14:37] how would I put a table below another table? a
between the two only lowers it one line [06:14:42] DB connection error: Access denied for user 'krimpet'@'%' to database 'centralauth' (vps2.epstone.net) [06:14:56] Can you use the {{Clear}} template? [06:15:20]
should do it. :) [06:15:58] Where does one edit the text outputed for Special Pages like Special:Upload? [06:16:04] nice [06:16:08] why did that work? [06:16:52] Krimpet: worked for me. [06:17:17] well, now it did :p [06:22:16] Krimpet: you should now be able to log into epstone.net/testwiki with the same credentials. [06:36:22] <_wooz> lo [06:36:47] heya MaxSem [06:37:21] hiiii [06:40:28] TimStarling: For that LinkHook change... would it be bad to create a new parser class for experimenting on so that the Parser_DiffTest could be abused? [06:41:27] if you count that as abuse, I don't see how you could use it without abusing it [06:41:42] heh [06:41:59] but are you saying you want to check it in with a new file? [06:42:21] Perhaps? [06:43:10] I think I have to look at your proposed change again [06:45:01] Dual-passing links rather than a single-pass; Looks like the adiquate description from what it currently looks like. [06:45:30] ie: Handle [[Link]] before [[Image:...|... instead of guessing [06:45:38] you're not going to be calling a hook chain for every link, are you? [06:45:50] Hook chain? [06:45:55] wfRunHooks() [06:46:22] Nope... in fact with the complexity of this it wouldn't work out that great to use one of those [06:46:32] and you want to make a tree again? [06:46:42] Something like a tree... [06:46:51] Probably a bit of a limited level of that [06:46:53] you're going to undo all the work I just did on memory efficiency [06:47:04] and it's going to be slower as well [06:47:13] Actually, it might be possible to do that without undoing all that completely [06:47:53] The best way to test would probably be to build it, then compare speeds [06:47:56] have you checked anything in yet? [06:48:02] Nope [06:48:08] have you written anything yet? [06:49:01] Only a light bit of hooking for media links (Broken because of how | and links show up in the params; which is what I'm trying to fix); But most of that stuff is invalidated by the changes you made [06:49:37] *OverlordQ wonders if brion is abouts [06:49:42] can you explain how you want it to work? [06:50:04] Hmmm... I'm bad at this, but let's see... [06:50:04] I'm particularly interested in how you imagine the interface between the parser and the extensions will be [06:50:16] Oohh... interface [06:50:30] If it were me, I'd call a hook with the whole tree once, at the end. [06:50:48] ^_^ If you are talking interface rather than changes to the parser, then I got that [06:51:06] Firstly, there are a few cases we need to consider: [06:51:32] Werdna: and if Dantman said he was going to do that, I'd tell him to rethink [06:52:17] Dantman: I'm listening [06:52:42] A) Link alteration (Non-embed; ie: Things like our special existence cases for Special:, Media:, Image:) [06:53:10] B) Embedding (Embedding content; Such as Image:) [06:54:03] so all of these depend on namespace, right? [06:54:07] C) Patterns (Things like SMW where rather than a namespace, they use a special pattern) [06:54:15] yes, for everything cept C [06:54:16] ok, the first two then [06:55:12] so for the first two, you could have the extension register a callback for each namespace it's interested in, during firstCallInit [06:55:19] Note that A) and B) merely represent callback locations where a function would alter output [06:55:32] similar to parser functions [06:55:40] Mhmm [06:55:49] Well, not just Namespace [06:56:21] magic word prefixes too? [06:56:29] no, I mean [06:56:48] You might want NS_EMBED to have a link altering callback, and an embed callback [06:57:06] why? [06:57:09] ie: So that you can embed when [[Embed: is called, but alter existance checks when [[:Embed is called [06:57:22] Similarly to Image: [06:58:19] that sounds very much like you want Embed: to be a namespace [06:58:31] if it walks like a duck and quacks like a duck [06:58:33] yes, Embed: is a namespace... [06:58:49] I mean, not 'for each namespace' [06:59:07] You might have two different callbacks for two different purposes on the same namespace [06:59:16] links, vs. embeds [06:59:40] ^_^ Don't you love miscommunicated tangents [06:59:59] you mean initial colon and no initial colon? [07:00:14] Pretty much. [07:00:20] or are you saying you actually want to call every single link with no initial colon an "embed"? [07:00:29] No [07:00:42] Only if the namespace is embedable [07:01:03] couldn't you just have one callback, and pass the initial colon to it as an option? [07:01:25] $callback( array( 'title' => $title, 'initialColon' => true ) ) [07:01:42] There are actually two different locations inside where these are handled... but perhaps [07:01:51] Might take a bit more refactoring [07:02:41] it's really best if the way the extensions see the parser is only distantly related to the way the parser actually is [07:02:50] mhmm [07:02:51] that way you can refactor the parser without changing the interface [07:03:30] so what would the hook return? [07:04:02] there's the issue that the output of replaceInternalLinks() is mid-main-pass [07:04:03] I was actually considering at one point passing something like that as an argument and having the same function handle the cases differently... I dismissed it since it didn't seam to fit the way anything else in MW does things (fit smarty more) [07:04:08] Oh that [07:04:12] which is a kind of text which would seem very strange to an extension [07:04:27] http://dev.wiki-tools.com/wiki/LinkHook [07:05:08] When we get to the point where we can pass an extension the parameters in an embed [07:05:35] The data is actually something like "foo|bar[[link|baz]]" [07:05:49] the initial colon works differently for images and categories [07:05:56] Mhmm [07:05:56] but it works the same for every other kind of link [07:06:26] so if an extension wants to have a link type that works the same as every other kind of link: special, media, etc., they don't want to have to define two functions [07:06:36] No... [07:07:11] I had a flag... if you didn't have a media embed, you just never added a callback for that [07:07:33] it just sounds like you're too focused on a particular application [07:07:51] Mhmm... and multiples at that [07:08:00] you have the chance to make this really really general and powerful [07:08:41] Really, my important case right now is making sure that |'s don't get screwed up when extensions muddle with things [07:08:56] example? [07:09:10] Well look at the link [07:09:31] what link? [07:09:37] http://dev.wiki-tools.com/wiki/LinkHook [07:09:48] that looks like a whole page of stuff I don't understand [07:09:55] Ok... let's see [07:09:56] just give me a link right here in this channel [07:10:08] links are generally pretty short [07:10:27] [[Image:Filename.ext|Foo|Bar[[Link|Baz]]...]] [07:11:02] So we're at the point where we're going to pass an extension a callback for the Image: ns, with a set of parameters after the | [07:11:45] 03aaron * r39520 10/trunk/extensions/FlaggedRevs/FlaggedArticle.php: only show notice for users [07:12:05] The data that we have at this point, is the $nt (Title) object, and a string containing "Foo|Bar[[Link|Baz]].." [07:12:50] Currently the parser handles this by recursively parsing the links inside the text (in fact there is a note that it'll parse everything) [07:14:18] This point is actually probably the reason for the whole Image embed mess. [07:15:07] So we have 3 options: [07:16:17] Hi, [07:16:18] I use a Add-Domain for hosting my Site. I own the domain www.tamilbuddhism.com, however the site is hosted @ www.tamilcc.org. [07:16:18] The E-mailing module of Mediwaiki uses tamilcc.org domain defaultly. How should I force it use the tamilbuddhism.com Domain ? [07:16:28] hmmm... I'll write the old stuff.. though I just thought of something [07:16:45] A) Pass the callback a string with a mess of link holders (again ignoring the fact that links may actually have a different meaning within the embed/link) [07:17:43] B) Pass the callback the string verbatim (Allowing authors to likely do something screwy with a explode | and breaking links inside of their caption area) [07:18:20] C) Pass them an Array where we have already split the | in the proper locations for them [07:18:28] sorry, was distracted for a minute [07:18:34] Mhmm [07:18:49] Though, I may need to research your new LinkHolders a bit more [07:19:30] They didn't exist when I was thinking [07:19:56] In fact I didn't notice the different ways you could [07:20:00] it's the same [07:20:05] 'expand'? till a bit earlier [07:20:28] my vote is for A [07:20:29] Oh ok... n/m [07:20:47] I got confused by a few function names [07:22:08] Any idea how to handle say, a bad Image: tag inside of say, a video Embed:? [07:22:34] how bad? [07:23:34] ie: Looking at the recursive stuff now... it would basically try to embed an image inside of the caption for a Video... mkay... maybe a different question [07:23:44] hmm, I guess it could come unstuck with present use cases [07:23:51] Ok, how much data do we have inside of the link holders? [07:24:16] with my latest changes, you don't necessarily get link holders [07:24:28] you could get fully expanded links [07:24:36] Eeek... that's not good [07:24:55] Fully expanded links would pollute tooltip text? [07:25:32] I think it would be nice to let the extension decide on how to expand the links inside of the text it's given [07:25:34] ie: [07:26:52] Back to WikiText (Links aren't valid inside of the caption, don't use them), back to plaintext (For tooltips where we don't want [[link]] or an actual link), or even a case where all links even those like Image: would be simple links [07:27:47] thinking [07:27:48] ie: [[Embed:Video|[[Image:Foobar.png]]]] and the callback for the embed tag could decide that Image: should actually be link rather than an image embed [07:29:32] Parser::makeImage() calls both $holders->replaceText() and Sanitizer::stripAllTags() on the alt text [07:29:45] so presumably that will give you roughly the same text whether it's been expanded or not [07:30:31] for an image inside an image: maybe not [07:30:56] Oh... also remember SMW [07:31:17] no... an image inside an image is OK [07:31:22] because image links don't make link holders [07:31:36] O_o eeek [07:31:54] so it's still invariant [07:34:39] the problem with option B is that it can take O(N^2) time [07:35:04] Eh? Do you mean C? [07:35:15] no, B, passing the string verbatim [07:35:26] because you have to find the end of the link, which is O(N) [07:35:37] We already do that [07:35:41] then the next recursion level has to find the end of the link too, which is another O(N) [07:35:59] Remember what I said about the current state [07:36:18] The data we have is the $nt object, and the string [07:36:38] Believe it or not, but it does contain something like "foo|bar[[link|baz]]" [07:37:18] really... it makes me think... "Why are we using a recursive call, instead of looping twice?" [07:37:21] we don't do it already, because we limit the recursion level [07:37:44] the O(N^2) comes from N per level times N levels [07:37:53] if you only have 2 levels, then it's O(N) [07:38:44] ^_^ And I guess that's why An image inside of an image, with a link inside breaks utterly? [07:38:48] avoiding O(N^2) time in this case was actually one of the motivations for using a tree in the preprocessor [07:39:29] you could explicitly limit the level, to say 5 [07:39:37] that'd fix the time order [07:41:35] the same time order issue is present with option C unless you construct the full tree [07:41:46] 03(mod) Image, category pages list both draft and stable links, mixed together - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=14557 (10JSchulz_4587) [07:42:02] but if you do have a full tree, that makes it more complex for extension authors, because they don't understand trees [07:42:31] limiting the recursion level does seem like a reasonable solution [07:42:32] You did fairly well with ParserFunctions [07:43:07] we don't actually want links nested to 5000 levels do we? [07:43:14] heh... no [07:43:43] and actually there's a few types of recursion depth limit in PPFrame::expand() [07:43:52] Of course [07:44:53] re ParserFunctions: yes, I understand trees [07:45:09] 03(mod) Image, category pages list both draft and stable links, mixed together - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=14557 (10JSchulz_4587) [07:45:12] but you don't see all the other setFunctionHook() users migrating to SFH_OBJECT_ARGS do you? [07:45:32] and when you do use SFH_OBJECT_ARGS, the code is longer and uglier [07:45:46] ^_^ Really, I kinda like it... heh [07:46:13] Then again... half my functions need control over expansion [07:47:44] Hmmm... I should experiment with a little thought I just had [07:48:13] ok, so I think B or C are feasible, but with limited recursion depth and probably limited total node count [07:48:24] Mhmm [07:48:35] Well, scratch B [07:48:46] B isn't really what you think I meant it as [07:49:16] I meant just passing them the string we already have and letting them do all the ugly stuff we do [07:49:32] ^_^ I wrote that one off from the beginning [07:49:37] if we can avoid constructing an array at all at the top level, that would be great [07:49:49] Which array? [07:50:09] an array of all links in the document [07:50:17] Hmmm... [07:50:29]