[00:19:44] 03(NEW) Implement Japanese and North Korean calendars - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18849 enhancement; Normal; MediaWiki: Internationalization; (jidanni) [00:20:47] 03(ASSIGNED) SQL syntax error - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18622 +comment (10bugzilla) [00:21:54] I wonder how many MediaWiki users there are in North Korea [00:22:28] idk [00:28:16] Hey guys, on a wiki I maintain whenever I save a page after editting it, Firefox waits for a very long time, then returns an empty index.php to me. Nothing is logged in the Apache error log. Any suggestions on where to look? [00:30:01] how empty? does it have headers? [00:30:25] I just need to wait for it to time out this time. [00:30:49] with no headers, firefox should show an error message [00:30:56] with headers, it would show a blank page [00:31:24] I'm getting a dialog to save index.php which if I safe it is zero bytes. [00:34:43] you should probably treat it as a PHP fatal [00:35:07] Hmm, and I can't view the headers. [00:35:40] in php.ini, set error_reporting=E_ALL | E_STRICT, and display_errors=on [00:35:52] see if that makes a difference [00:36:11] you will have to restart apache for it to take effect [00:36:48] actually come to think of it, maybe I have seen that save dialog for a segfault [00:36:57] Okay, I've changed the reporting level (was E_ALL | & ~E_NOTICE [00:37:24] retrying the save now [00:37:25] it's linux, right? [00:37:28] yes [00:37:47] you should check the syslog for apache errors [00:38:21] nope, no errors in syslog either [00:40:13] if display_errors didn't help, we should probably assume it's a segfault and get out the big guns [00:40:26] yeah? [00:40:35] it isn't working [00:40:44] run apache under gdb [00:40:52] oh joy [00:41:10] so it'll typically be: gdb apache2 [00:41:16] yeah [00:41:18] then: run -X [00:41:22] for single-threaded [00:42:08] okay, just installed gdb [00:42:14] installing that it is. [00:42:40] then reproduce the problem, and with a bit of luck it'll drop out into gdb when it segfaults [00:42:42] 03(NEW) Retire 'Wikipedia Day' from Preferences Date and Time page, and just use current time - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18850 enhancement; Normal; MediaWiki: User preferences; (jidanni) [00:43:04] okay, I'll report back once I have that. [00:45:52] ping mdale, any news on upload branch? i'd like to get going on that [00:48:31] 03(mod) Implement Japanese and North Korean calendars - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18849 (10jidanni) [00:49:28] TimStarling: using run -X I can't even view a basic page! [00:50:43] well, since it's single-threaded, it might not play nicely when the browser requests multiple pages at once [00:51:05] you should have just clicked "save" from the edit page you already had open [00:51:20] alternatively... [00:53:08] "set follow-fork child" the run without the -X, but that doesn't seem to be ideal [00:54:02] it'll probably give you a random child process, and you'll then have to hit save a lot of times until you get the right one [00:54:21] yeah, I've hit save on the edit page, but it isn't returning. [00:54:51] so you ran it without the -X? [00:55:26] no, not yet [00:55:40] so with -X, you hit save, then what? [00:55:45] it shows the download box? [00:55:56] not yet, still waiting. It takes a while. [00:57:49] 03(mod) Free up "Project" as a namespace (for Wikipedia:Projects) by using an alternate token - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18793 (10innocentkiller) [00:58:29] is CPU usage high, while you're waiting? [00:59:25] no, not really and I just received a dialog window. No output in gdb. [01:00:09] so if you run "top" while the request is going, do you see apache2 pegged at 100% CPU? [01:00:48] no, it doesn't seem to be [01:00:59] load average 0.01. [01:01:11] And there is other unrelated things running on that box. [01:01:27] alright, strace [01:01:32] :) [01:01:34] sure [01:02:40] you should run apache2 -X in one window, and then attach to it with strace -p in another window, to avoid being spammed with startup syscalls [01:02:56] strace apache2 -X &>str & [01:03:11] the low CPU indicates that it's either waiting or dead [01:03:13] close 'nug [01:03:22] if it's waiting, strace will tell us what it's waiting for [01:03:28] aye [01:03:50] so the aim here is not to wait for the timeout, you should just look at what syscall is running for the bulk of the time [01:07:40] <^demon> TimStarling: When you get a chance, a few revs of mine you pegged as needing followup should be resolved (47900, 50313, 50635). No rush though. [01:08:40] you put comments on CodeReview right? it notifies me when you do that [01:09:20] <^demon> Yeah I did. Should all have links to the relevent followup revs. [01:14:01] thanks for that, I'll check them a bit later [01:14:31] <^demon> No problem. Should be pretty straightfoward. I've got a few other fixme's I'm hoping to commit this evening as well. [01:19:22] 03(mod) Add setting to hide a list of preferences - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18761 (10jidanni) [01:26:56] hello I am trying to revert spam and I get "The edit could not be undone due to conflicting intermediate edits." [01:27:05] How do I revert -f [01:27:31] if you have rollback, use it [01:27:51] if not, or if there are multiple spammers per page, go into the page history, find the last good revision, edit it, and save [01:28:52] Great, thanks [01:29:07] and what is the recommended solution for spam? [01:29:35] !e SpamBlacklist [01:29:35] --mwbot-- http://www.mediawiki.org/wiki/Extension:SpamBlacklist [01:30:08] great, thank you! [01:32:31] TimStarling: I've worked out the issue, the MTU on a tunnel was bad. So the page that was returned wasn't getting returned correctly. [01:34:10] mmm, should have guessed [01:34:37] so strace would have shown you apache waiting to write to the network [01:35:41] Well, trying to write to the network, but the socket was closed. [01:36:52] But it lead me in the right direction. I've had issues with MTU on this tunnel before. But I'm sure that I'd tried editting some stuff while at the other end of the tunnel and had issues. [01:36:57] I guess I'm remembering wrong. [01:38:31] spam blacklist will fail. all edits come from 127.0.0.1 due to whatever special setup sourceforge has [01:38:42] any other recommendations? [01:38:59] well, it was a productive debugging session anyway [01:39:06] fulld: I don't see how, SpamBlacklist doesn't care where the edits come from [01:39:16] it blocks the domains for /everyone/ [01:39:27] ah [01:39:30] it's always easy when you have a user who knows stuff [01:40:04] TimStarling: cool, thanks for your help! [01:40:13] no problem [01:53:54] 03shinjiman * r50804 10/trunk/phase3/ (5 files in 3 dirs): * (bug 18849) Added Japanese and North Korean calendars support [01:54:33] 03(FIXED) Implement Japanese and North Korean calendars - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18849 +comment (10shinjiman) [01:59:56] 03shinjiman * r50805 10/trunk/extensions/LocalisationUpdate/ (LocalisationUpdate.i18n.php LocalisationUpdate.php): Localisation updates Cantonese and Chinese for extension messages [02:05:05] 03(mod) Must check $wgEnableEmail before showing ANY email related preferences - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18760 (10jidanni) [02:05:45] 03(mod) Must check $wgEnableEmail before showing ANY email related preferences - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18760 +patch (10jidanni) [02:23:56] morning folks! [02:24:47] how can i debug a mediawiki extension? I get errors under a specific set of circumstances, would like to output value of certain vars deep in a hook's code. [02:25:06] !debug [02:25:06] --mwbot-- For information on debugging (including viewing errors), see . [02:25:24] OverlordQ: thx! [02:25:56] basically enabled logging toa file and sprinkle in some wfDebug calls :) [02:27:54] 03(mod) Implement Japanese and North Korean calendars - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18849 (10jidanni) [02:43:13] so i find that under Apache (on win32) all is good. under IIS, I see "undefined offset" from: explode('|', htmlspecialchars($input)) Is there simply a way for me to gobble the notices under MW ? [02:44:16] btw: these only display when using "show preview", not when a page is saved/viewed or even initially edited. [02:45:58] are you debugging your own extension, or someone else's? [02:46:11] someone else's. VideoFlash [02:47:02] can you show me the complete line that is giving the notice? [02:47:17] list($id,$width,$height,$style) = explode('|',htmlspecialchars($input)); [02:47:24] heh [02:47:26] i guess it is the list() that is acutally throwing the problem [02:47:32] yes [02:48:03] The input is the same regardless of "show preview" or simply "save page" [02:48:10] it's lazy code, the response when $input doesn't have enough parameters should be treated explicitly [02:48:12] in preview, notices. [02:48:19] o i c [02:48:31] like... $args = explode('|',htmlspecialchars($input)); [02:48:39] if ( count( $args ) < 4 ) { ... [02:48:46] and then handle $args[0], etc....i get it. [02:48:47] thanks! [02:48:49] yes [02:49:11] I'm a Perl guy at heart.... :) [02:49:48] it's silly because they have conditional clauses *after* the fact to deal with the values not being set. [02:50:01] well yeah [02:50:03] is this a behaviour change in newer PHP builds? [02:50:23] the developer probably doesn't have E_NOTICE enabled, so they don't realise it's a problem [02:50:40] i see [02:50:57] we recommend that all developers enable it, it's very useful for detecting references to undefined variables [02:51:06] so i could simply unset that in php.ini...but then i'd be running iwth not-so-good code [02:51:20] cool. you have learned me good. [02:54:07] anyone here used Special:Export ? [02:54:17] because I tried the examples from here http://www.mediawiki.org/wiki/Manual:Parameters_to_Special:Export [02:54:36] and they pulled out the entire history of the article(all the revisons) instead of just what I asked for [02:54:54] I thought it would take just 10 revisions starting from the offset [02:55:03] 03aude * r50806 10/trunk/tools/planet/en/config.ini: adjust Witty Lama's feed link to make it work [02:55:05] can anyone explain what's going on ? [02:58:07] what would be a good regex that matches a mediawiki-style date [02:58:27] for example, 21:27, 19 May 2009 (UTC [02:58:29] ) [02:58:41] harej: hang on, I know where I can find one [02:58:47] thank you [03:01:36] ([0-2]\d):([0-5]\d),\s(\d{1,2})\s(\w*)\s(\d{4})\s$UTC$ [03:02:13] :D [03:02:34] It looks really confusing, but it works [03:02:54] It only accepts valid dates, too [03:02:59] 03shinjiman * r50807 10/trunk/phase3/languages/messages/MessagesZh_tw.php: additional date format for Taiwan Chinese language, per code review on r50778. [03:03:41] a few invalid ones, but who's counting [03:03:52] oh wait nm [03:03:55] damn colon [03:04:29] Oh right, the colon needs escaping [03:13:55] does anyone here know anything about the mediawiki debian package? i just installed it and i can't figure out how to access mediawiki. Going to http://localhost/mediawiki isn't doing it, and i see nothing in my /var/www/ that indicates mediawiki is there [03:15:10] DeFender1031: I would send an email to the maintainer of the package [03:15:26] hmmmmmm [03:15:30] now my bod wants to edit the main page [03:15:58] chuck, i don't think it's a bug or a package error, i think it's probably something i'm doing wrong. i was just wondering if anyone knew what it was [03:16:08] bot* [03:16:52] <^demon> DeFender1031: I just tried on Ubuntu, and it installed to /var/lib. Might be similar for you. [03:17:47] ^demon, that's where most of it ended up for me as well [03:18:15] DeFender1031: the package adds a config to apache's conf.d folder to create an Alias [03:18:17] try restarting apache [03:18:21] ah [03:18:41] and try /wiki as well I can't remember the exact url it uses [03:18:47] remind me how to restart apace? i'm sort of rusty [03:19:04] OverlordQ, it's definitely /mediawiki, i have a friend who did this before [03:19:36] sudo /etc/init.d/apache2 restart [03:19:41] thank you guys :] [03:19:45] anyone here knows anything about Special:Export ? [03:19:55] does that API still exist in mediawiki ? [03:20:06] has it changed since http://www.mediawiki.org/wiki/Manual:Parameters_to_Special:Export [03:20:06] thanks [03:20:12] has it changed alot ? can it still be used ? [03:20:22] what is different from when they wrote that stuff ? [03:21:03] hmm... still not working... [03:21:53] <^demon> DeFender1031: Check out /etc/apache2/conf.d/mediawiki.conf. It had the Alias commented out by default. [03:21:53] (and yes, i tried /wiki as well) [03:22:03] <^demon> Then restart apache again [03:22:40] hooray! [03:23:02] thanks OverlordQ, chuck, and ^demon, you guys are great [03:24:07] <^demon> At least we know the Debian/Ubuntu package more or less installs sanely :) [03:29:30] 03shinjiman * r50808 10/trunk/phase3/languages/messages/ (MessagesKo.php MessagesZh_tw.php): * (bug 18849) follow up r50778 and r50804, additional date format for Taiwan Chinese language and Korean language. [03:30:31] ^demon, any idea how to get parserfunctions working? [03:31:38] !parserfunctions | DeFender1031 [03:31:38] --mwbot-- DeFender1031: "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 . [03:31:42] i installed the extensions package along with it [03:32:04] chuck, this is specifically in debian, i remember when my friend did it, the procedure was different [03:32:10] <^demon> The extensions package seems to have thrown a bunch of stuff into /etc/mediawiki-extensions/ [03:32:16] <^demon> Hell if I know why. [03:32:22] That's really odd [03:32:33] <^demon> and /usr/share/ [03:33:08] I guess Debian is too cool to install extensions just like every other MediaWiki site on the internet does [03:34:20] <^demon> If the mediawiki package is installing to /var/mediawiki (which is ok, the conf files make it all work just fine), it would be nice if mediawiki-extensions would install to /var/mediawiki/extensions like a sane install. [03:34:26] chuck, they are... they want it to be able to be installed through their package manager with no muddling on the part of the user [03:35:18] <^demon> DeFender1031: My advice is to not use the mediawiki-extensions package and install them by hand. If mediawiki is installed at /var/mediawiki, then put the extensions in the 'extensions' subdirectory of that. It's where they should normally go. [03:35:23] <^demon> Stupid package. [03:35:43] Or, email the package maintainer and figure out what they meant when they originally created the package [03:36:05] <^demon> That would be nice in the long term, yes. [03:36:17] ^demon, i would, but i seem to remember that following the regular instructions didn't work when my friend did it, apparently debian modifies sometyhing slightly to make it work with their setup [03:37:18] where did it install them to? [03:37:54] <^demon> /etc/mediawiki-extensions and /usr/share/mediawiki-extensions. The mediawiki package installed to /var/mediawiki [03:38:05] <^demon> It would follow that extensions should go in /var/mediawiki/extensions [03:38:25] some packages shove crap into /usr/share and then symlink iirc [03:42:02] ^demon, debian's system is typing into a console "mwenext " [03:42:31] i got it [03:42:37] thanks for all the help, guys [03:42:43] What's the most resource conservative way of running MediaWiki and PHP? [03:43:57] hmm... how do i make it not force capitalized page names? [03:44:16] <^demon> DeFender1031: http://www.mediawiki.org/wiki/Manual:$wgCapitalLinks [03:45:03] what file is that var in? [03:45:13] LocalSettings.php [03:45:24] At least, that's the file in a regular MediaWiki installation. [03:45:37] <^demon> Do all your configuration in /var/mediawiki/LocalSettings.php [03:47:30] thanks [03:47:56] so noone knows anything about Special:Export ? [03:48:52] <^demon> spx2: It hasn't changed too much. That API should still work. In addition, there is now export in the normal mw api.php, so see those docs if you want that too. [04:08:18] <^demon> Anyone else with commit access around? [04:10:31] *^demon really needs to get his key set up on the other computer :-\ [04:19:26] ^demon: What's up? [04:22:04] Ugh, why is MediaWiki's Main Page taking FOREVER to load, when other pages load instantlyl... [04:22:07] <^demon> chuck: I need a patch applied to trunk mediawiki, can't commit right now. [04:22:07] *instantly [04:22:09] So peculiar [04:22:21] <^demon> http://pastey.net/114578 [04:22:22] ^demon: I can apply it if you'd like [04:23:03] <^demon> It's a followup to r50776, retaining the old method names for b/c, but marking them as deprecated. [04:23:20] Who is the patch by? Yourself? [04:23:25] <^demon> Yes. [04:23:40] <^demon> Damn key issues :( [04:26:55] <^demon> TimStarling: Are you about? [04:28:13] yes [04:28:30] I can add an extra key for you if you like [04:29:29] <^demon> That'd be great. http://pastey.net/114580 [04:29:34] ^demon: How should you be attributed in the commit message? ^demon, your real name? [04:29:51] <^demon> Just say patch by ^demon, or just wait for Tim to add my new key :) [04:30:05] Heh, sorry it took a while, I had to svn up [04:30:14] but, wait for it... [04:30:16] 03charlie * r50809 10/trunk/phase3/includes/AuthPlugin.php: [04:30:16] Follow-up to r50776. Retaining old method names for backwards compatibility, [04:30:16] but making them as deprecated. Patch by ^demon. [04:30:22] :P [04:31:11] done [04:31:14] <^demon> Thanks. [05:16:20] morning [05:16:30] rar [05:20:11] !rar [05:20:11] --mwbot-- rar [05:21:55] morning [05:23:36] will brion ever catch up with code review ? [05:24:05] the scapmap looks yellowish [05:24:39] that's because I reset it all back to new [05:25:19] oh i see [05:25:39] was that on purpose ? [05:25:46] yes [05:25:46] Tim: do you review in order, or do you wait for a dev to be around IRC and try to focus on theirs? [05:26:15] + * @deprecated Use allowPropChange() [05:26:19] wfDeprecated anyone? [05:26:24] I usually do it author by author [05:26:35] there's more continuity that way [05:26:55] and you get a better sense of the dev's coding style and what they know and don't know [05:27:15] I don't wait for them to be on IRC [05:27:17] hmm... why is it that {{#time: G:i:s|{{{1}}}}} produces inconsistant results if a time is not actually specified? example: different results for 1=1995 than it does for 1=10/25/76 [05:27:57] timezone? [05:28:35] I set them back to new because there was no way to tell what was properly reviewed and what wasn't [05:28:43] Nikerabbit, doesn't matter [05:28:57] TimStarling: so you'll review everything again ? [05:29:11] some people were making huge commits, and then marking them as "resolved" because they fixed some tiny part of it [05:29:58] the status log is still there, I sometimes will just respect whatever is in it [05:30:07] depending on who did the status change and what it was [05:30:14] Perhaps reviewing ability should be restricted or split? [05:30:45] maybe some day [05:34:09] DeFender1031: 1995 is interpreted as 19:95 [05:34:18] er, if it is a valid time that is, heh [05:34:18] TimStarling: which brings me, any idea how to make wfDeprecated useful? [05:34:23] Splarka, no it's not [05:34:30] 1959 as 19:59, 1960 as 1960 [05:35:08] it's outputting the right year [05:35:16] just... the time is inconsistant [05:35:32] are there some functions which keep being used in new code despite being marked deprecated? [05:35:32] you'll notice that i'm outputting hours, minutes, and seconds, and that neither of the inputs contains that data. i would expect that in such a case it would default to either the current time, or 0:00:00, but it seems to be doing neither in some cases and current time in others [05:36:04] It outputs differently on which wiki? [05:36:07] A personal wiki? [05:36:35] TimStarling: nobody knows, since it's not throwing notices anymore (the function documentation is wrong too now) [05:37:28] there are a number of things that could be dealt with in a pre-commit or post-commit hook [05:37:37] hmm [05:37:40] parse errors, unicode BOMs, eol-style [05:37:52] Marybelle, it's a local wiki, yes [05:37:52] searching for deprecated functions could be done too [05:38:08] using parsekit or token_get_all() [05:38:14] and i don't have port forwarding available where i am. if i was at home, i could make it available to you [05:38:25] SVN needs an AbuseFilter? [05:38:45] {{#time: Y-n-j G:i:s|1995}} - {{#time: Y-n-j G:i:s|10/25/95}} -> 1995-5-20 5:38:18 - 1995-10-25 0:00:00 [05:38:57] so I think designing and writing a commit hook would be a nice little project to deal with this sort of thing generally [05:39:05] looks like if you just specify the year, it gives you the time now of that year, but if you specify a date it gives you a zero-time for that date [05:39:34] Huh. [05:39:39] go bug ##php [05:40:09] http://uk3.php.net/manual/en/function.strtotime.php [05:41:00] Splarka: http://www.mediawiki.org/wiki/Help:Extension:ParserFunctions#.23time: :o [05:41:01] Splarka, i'm getting 4:00:00 for certain things, 5:00:00 for other things, and current time for still other things [05:41:26] Nikerabbit: have read it, what about it? [05:41:30] and i kind of need a consistant way to read time, because the template i'm writing needs to be able to optionally take a time [05:41:37] Warning: The fill-in feature is not consistent; some parts are filled in using the current values, others are not: [05:41:42] yah [05:41:53] so don't feed crap to it :) [05:41:54] not the fault of MediaWiki [05:41:55] so then how do i parse dates and times consistantly? [05:42:19] Rewrite MediaWiki is a different language? [05:42:24] In, even. [05:42:31] i would disagree. putting in your documentation "we don't know how to write this function correctly" doesn't make it not your fault. [05:42:37] heh [05:42:42] ^ go bug ##php [05:42:49] oh right... [05:42:58] DeFender1031: it's not a MediaWIki problem it's a PHP problem [05:43:00] #time builds off of the php time function... [05:43:04] i forgot [05:43:08] Functionarionista. [05:43:09] builds off -> shells out to [05:43:12] heh, so it really ISN'T your fault, eh? [05:43:24] well, lets see who is NOT to blame [05:43:29] yeah well, php is full of crappy code, what can ya do? [05:43:39] port it to ruby [05:43:42] I thing somehow Ballmer is involved [05:43:46] Shouting in the MIDDLE of sentences makes for awkward conversation. [05:43:58] #time sucks anyway, it only speaks english [05:44:10] BRIAN BLESSED was here [05:44:28] I think it accepts food. [05:44:32] isn't not middle brian blessed? [05:44:44] People can be blocked for 1 chicken. [05:45:32] 03nikerabbit * r50810 10/trunk/phase3/includes/AuthPlugin.php: Added wfDeprecated too [05:45:38] http://svn.wikimedia.org/viewvc/mediawiki/trunk/extensions/ParserFunctions/ParserFunctions.php?view=annotate [05:46:03] it is mostly brion and tim's not-fault [05:46:20] so do don't blame them [05:46:23] for one chicken?? [05:46:29] *apergos must be hallucinating [05:50:43] Splarka, i take back blaming MW [05:50:50] i blame php. [05:51:07] welcome to the internet! [05:51:44] thanks, glad to be here. ;) [05:52:47] Is it possible to give not logged in users (* i believe) read only access to ONLY index.php/Main_Page and nothing else? [05:53:33] http://www.mediawiki.org/wiki/Manual:$wgWhitelistRead [05:53:36] !access [05:53:36] --mwbot-- For information on customizing user access, see . For common examples of restricting access using both rights and extensions, see . [05:54:04] oh_noes: you'll also wanna give access to Special:UserLogin or you won't be able to, heh [05:57:07] login, logout, password reset, etc. are automatically whitelisted [05:57:48] Splarka, after discussing it with someone, i take back blaming php and blame MW again. apparently the function is not intended to be used the same way as the {{#time}} function is [05:58:16] DeFender1031: well, but #time has clear warning that it is inconsistent [05:58:53] *Splarka hands DeFender1031 over to Tim... [05:59:20] Nikerabbit, true, but as i said before, putting in your documentation "we don't know how to write this function correctly" doesn't make it not your fault. Granted it's more like "we're not USING this function correctly" but still... [05:59:58] i can't really blame you though, because without using strtotime you wouldn't have a {{#time}} function at all so... [06:00:01] so, you keep quoting that, where does it say "we don't know how to write this function correctly" [06:00:24] and who do you mean by "we", any mediawiki users who happen to use it, whoever wrote that page, or whoever edited the relevant parts of ParserFunctions? [06:00:46] Splarka, i was paraphrasing the "Warning: The fill-in feature is not consistent; some parts are filled in using the current values, others are not" [06:01:06] quite a leap [06:01:39] DeFender1031: did you get that from ##php? [06:01:53] TimStarling, no [06:02:15] *OverlordQ grabs popcorn [06:02:25] *Splarka sells funnelcakes [06:02:27] *charitwo steals OverlordQ's popcorn [06:02:34] *charitwo also buys a Splarka funnelcake [06:02:39] but i'm told that strtotime is meant to interpret what you most likely mean, so depending on how your DATE is formatted, your TIME will come out differently [06:03:22] caveats on mw:Help:Extension:ParserFunctions#.23time: are not meant to be interpreted as "the devs are lazy and stupid", but are meant to warn you that strtotime sucks [06:05:38] apparently it doesn't suck, but tries to fill in the time as best for the date it's given. if you give it an american format, it'll do 0 time in DST, if you give it a european format it'll do 0 time without DST, and if you say "next tuesday" it'll do current time, because that's what "makes the most sense" in each of those cases [06:05:58] there IS a certain logic to that [06:05:59] it's nice that we have strtotime() and parse_date.c [06:06:39] it's just a pity that Derek is so combative and hard to talk to about any bugs or features [06:07:10] DeFender1031: it /does/ suck, as Nikerabbit said above [06:08:12] we will just have to make do with what we're given, I think [06:08:24] whatever the case, can you guys think of any way that with the current parserfunctions and the current {{#time}} i could write a template that has the interpreting functionality and deals with absent times consistantly? [06:08:32] so you can't write your own function if PHP's doesn't do what you want? [06:08:45] who limits themselves to the built in libraries? [06:08:48] you've seen the size of the function right? [06:09:11] it's one of the most massive things in the whole of PHP [06:09:30] DeFender1031: if you need consistency, put in a dummy time, although be careful not to use a dummy time from midnight to 3 AM [06:09:41] I don't think it makes sense to reimplement it unless there's some *really* good reason [06:10:26] DeFender1031: maybe add some tricks like {{#ifeq:{{#time:G:i:s|{{{1}}}}}|0:00:00|define time|do nothing}} [06:10:36] inside another #time [06:10:38] insane_coder, unfortunately, i can't do that, as i'm writing a template. the user of the template would have to be the one to include the dummy time, and that's exactly what i want to avoid [06:10:42] you do not want to use 0:00:00 [06:11:11] that is just checking /against/ 0:00:00 [06:11:37] Splarka, as i said, it's not outputting 0:00:00 for me anyway [06:11:49] it's this 1700 line file: http://cvs.php.net/viewvc.cgi/php-src/ext/date/lib/parse_date.re?revision=1.26.2.27.2.15&view=markup [06:11:57] tricks /like/ [06:12:02] which generates these 24000 lines of C: http://cvs.php.net/viewvc.cgi/php-src/ext/date/lib/parse_date.c?revision=1.102&view=markup [06:12:04] substituting in whatever funky times you're getting [06:12:19] which is, IIRC, by far the biggest C file in PHP [06:12:34] do you have a custom $wgLocaltimezone ? [06:12:49] Splarka, except is the time actually IS the funky time, it'll kill the templat [06:13:12] picky picky [06:13:33] not picky, i just like stuff to work right [06:13:54] *Splarka wonders who will say "then don't rely on #time" or "don't use mediawiki then!" [06:15:04] http://www.filmjunk.com/images/weblog/2009/05/treknobabble65.jpg [06:16:51] hmm [06:17:08] DeFender1031: maybe {{#ifeq:{{#time:n-j G:i:s|{{{1|1995}}}}}|{{#time:n-j G:i:s}}|dummy year entered|good date}} [06:17:13] Splarka, how about this. if i offered to write a function in php that was similar to strtotime (such that it supported everything that strtotime supports and maybe more) yet was consistant about fillers, would you guys include it? [06:18:06] *Splarka hides behind the funnel cake machine [06:18:48] TimStarling, do you have an answer or do you too have a sudden craving for funnel cakes? [06:19:20] yes [06:19:22] How would you write a better function? [06:19:49] I think that would be very useful [06:20:12] and localizable inputs? [06:20:14] but like I say, it's a big project [06:20:25] thousands of lines [06:20:28] is that "yes i would include it" or "yes i want funnel cakes"? or both? ;) [06:20:42] en-gb vs en-us for 1/2/2009 vs 2/1/2009 [06:20:52] yes, I would include it, assuming it's of an appropriate quality [06:21:25] alright... well, i'll see what i can do with it [06:21:50] Marybelle, i have a bit of a knack for parsing [06:21:52] Splarka: You're a Debbie Downer. [06:22:16] I have a knack for naps. [06:22:17] I prefer to see the glass as half full... [06:22:19] of arsenic [06:22:34] I see it as the glass is too big [06:22:46] hahahahaha [06:26:31] Splarka: x/y/z is always US [06:26:39] sigh [06:26:44] which is why everyone should use y/m/d [06:27:11] 02-01-90 [06:27:26] insane_coder, i've seen it like that elsewhere, only the month and day are reversed [06:27:40] 3/4/2005 == March 4 2005 - 12:00:00 AM [06:27:41] 3-4-2005 == April 3 2005 - 12:00:00 AM [06:27:43] 3.4.2005 == April 3 2005 - 12:00:00 AM [06:28:23] if the US uses m/d/y, why do the green visa waiver cards ask for a date in d/m/y format? [06:28:25] the point is, if DeFender1031 is going to improve strtotime for mediawiki, it should understand non-American [06:28:58] strtotime() already understand non-american [06:29:09] the question is, do you know how to enter a non-american date? [06:29:37] is it because the only people filling out those cards are not american citizens? [06:29:44] Splarka, well, here's the thing, in a case where one number is greater than 12, it's obvious what the format is. in a case where it's not, the function has to choose what format to go with [06:29:48] and if so, do the customs cards use m/d/y? I don't remember [06:30:15] TimStarling: an international company can do whatever the heck it wants, and generally they try to follow their biggest customer base [06:30:18] DeFender1031: and it should probably choose based on site content language, like formatnum|R does [06:30:27] but if the language is 'en' it can't decide that way [06:30:43] Anything internationalized will use d/m [06:31:15]