[01:01:04] is it possible to search all contributions from a particular user for a particular string? Or, going the other way, look for all articles with a string and then check if those articles include a particular contributor? [01:03:29] It's possible [01:03:36] But isn't going to be a quick thing [01:06:32] well, yes, in both cases I can get the initial list, but then it seems like I'd have to manually visit the pages to look for the second half .. I was looking for a search tool that would use both criteria [01:08:53] With a database query you can do it as one command [01:10:00] is there some way I can perform the database query? [01:10:29] Yes... [01:10:35] What wiki are you wanting to do it for? [01:10:51] wikipedia [01:11:50] that's obviously not one wiki. But against a specific language project, you can do that either on toolserver (deprecated) or via wmflabs [01:12:20] ok .. wikipedia-en [01:13:16] csteipp_afk, anomie|away: for OAuth, how often is the version number going to change? it's weird to have two tags for the same applicaiton: https://www.wikidata.org/wiki/Special:Tags [01:14:35] legoktm: It would depend on how often the developer has to make a new consumer request, I suppose. [01:16:39] what would cause the developer to need to make a new request? [01:17:46] it's gonna be confusing if the tag keeps changing [01:18:35] Having to change any of the fields on the consumer registration page, most likely either the OAuth "callback" URL or the set of grants being requested. [01:29:12] reedy - ok, I'm at tools.wmflabs.org .. was there a particular tool you were thinking of? I don't see anything that would do it, although usersearch looked like it had potential [01:34:14] slvrstn_: maybe "dbreps" ? i'd suggest #wikimedia-labs from here [01:37:16] mutante - thanks, looked like dbreps were pre-defined reports, not something I could run a query from [01:37:27] appreciate the help though [01:38:05] slvrstn_: i know there are db replicas for labs ..somewhere..just not sure about the project name either [06:50:58] It seems every time I load a file description page on Commons, the image refreshes. [06:51:39] It's kind of annoying. [14:32:06] Reedy: Hey, could you force a force a localisation cache rebuild for "ml"? See bottom of https://bugzilla.wikimedia.org/show_bug.cgi?id=58074#c12 [16:20:16] before [16:20:16] reedy@tin:/a/common/php-1.23wmf6/cache/l10n$ ls -al l10n_cache-ml.cdb [16:20:16] -rw-r--r-- 1 l10nupdate l10nupdate 3098181 Dec 11 02:06 l10n_cache-ml.cdb [16:20:23] after [16:20:24] -rw-r--r-- 1 l10nupdate l10nupdate 3098181 Dec 11 16:19 l10n_cache-ml.cdb [16:21:29] heh [16:23:49] -rw-r--r-- 1 l10nupdate l10nupdate 3095899 Dec 11 16:23 l10n_cache-ml.cdb [16:23:49] -rw-r--r-- 1 l10nupdate l10nupdate 3098181 Dec 11 02:06 l10n_cache-ml.cdb.old [16:23:54] Forced rebuild and it gets smaller [16:24:06] * Reedy syncs it for the hell of it [16:24:46] Reedy: that's https://bugzilla.wikimedia.org/show_bug.cgi?id=58074 ? Thanks! [16:25:34] I'm not sure it's going to help [16:26:16] api output looks exactly the same against siebrands paste [16:26:25] :/ [16:30:05] Though [16:30:12] mlwiki isn't using wmf6 yet [16:30:16] * Reedy does the same for wmf5 [16:36:43] Reedy: any luck? [16:41:12] Same size after a forced rebuild [16:48:19] greg-g, Reedy: On my local system when I set the language to ml, I don't see any of the namespace aliases defined in MessagesMl.php until I revert I933f81782f3482334c9a8d820d4ea2b1120f5c59. Then they all show up. [16:51:40] * anomie now tries to see how $namespaceNames is handled differently from $namespaceAliases [16:54:41] Aren't they usually in seperate files? [16:56:39] anomie: It looks like Scribunto is the only extension to do that [16:56:50] SemanticMediaWiki does some magic [16:56:54] Reedy: I was just coming to that conclusion, too. [16:56:56] And that one. [16:57:14] Pastebin of that is http://p.defau.lt/?_P9SLGra7gcxOa7qUIy0_w [16:57:26] * Reedy reverts I933f81782f3482334c9a8d820d4ea2b1120f5c59 as broken [16:59:38] anomie: Any idea if we can just update to master? Or do we want to create wmf/1.23wmf5 and wmf/1.23wmf6 branches and cherry pick onto it? [17:00:19] Reedy: Let me look [17:00:21] Thanks [17:01:06] I note the commit before the ml one adds aliases to the namespaceNames array [17:01:07] https://git.wikimedia.org/commitdiff/mediawiki%2Fextensions%2FScribunto.git/6126644d92273ec8797ff599cd789b7a74dec848 [17:03:59] namespaceNames is fine, it's just that namespaceAliases array that's borken. [17:05:00] Moving Scribunto to head would deploy 5 other code changes (see git log --oneline 6a1b0d5..HEAD). Nothing should break by deploying them, but... [17:05:10] * anomie has been busy this week [17:07:56] we can branch then [17:08:04] not like it's loads more work to do it [17:10:50] As for the different handling of namespaceNames and namespaceAliases, it looks like it's because 'namespaceNames' is listed in LocalisationCache::$mergeableMapKeys and 'namespaceAliases' isn't. But adding it there would also make it pick up all the aliases from the whole fallback chain, so that's probably not something we'd want to do. [19:16:51] some servers for tools seem have disturbances [19:30:20] also wiki pages load slowly [20:25:02] did you mean no longer expands to the full NS name? [20:25:13] > Did you mean: [-1]:Expand templates [20:25:32] ^d: manybubbles : ^ [20:26:05] jeremyb: I see what you are saying but I'm not sure what you mean [20:26:37] manybubbles: it should say: Did you mean: Special:Expand templates [20:27:01] in this case i was on commons fwiw [20:27:25] jeremyb: haven't changed commons (yet), but can you send me a url so I can look at it? [20:27:56] jeremyb: not in production, I mean. in beta, cirrus is on commons [20:28:10] i'm in prod [20:28:47] https://commons.wikimedia.org/w/index.php?title=Special%3ASearch&search=Special%3AExpandTemplates&fulltext=Search [20:29:22] impressive [20:29:48] ^d: I wonder if something we did caused something in mwsearch to go a bit crazy: https://commons.wikimedia.org/w/index.php?title=Special%3ASearch&search=Special%3AExpandTemplates&fulltext=Search [20:30:02] <^d> I see. [20:30:03] <^d> Hmm [20:30:06] <^d> Don't think so. [20:30:22] fwiw ExpandTemplates was recently merged into core [20:30:43] and it might still also be deployed as an extension, I forget. [20:31:01] <^d> Good times :) [20:31:24] legoktm: it still is, i think [20:31:29] i filed a bug requesting that it be undeployed [20:32:20] ^d: it looks like it does it on other searches, like Special:Abuse Filte [20:32:30] it'll suggest [-1]:Abuse *Filter* [20:32:34] helpful [20:33:00] <^d> Solution is obvious: move all the special pages to new names :p [20:33:03] legoktm: i know, good point. i can reproduce with a search for special:searc [20:33:34] lolol [20:34:37] legoktm: maybe you want to help dominic with the issue i was thinking about using expandtemplates for? :) lua error @ [[commons:File:Howard P. Perry, the first African-American US Marine Corps recruit, 1942.tif]] [20:35:02] * legoktm clicks [20:35:13] i was hoping xml tree or something would be helpful but didn't give any clues [20:35:27] > local vowels = 'aeiouyąăẵằẳặȃắâẫấầẩậãäǟāáàȁǎảẚåǻḁạǡæǣǽĕȇêễếềểệḙẽḛëēḕéḗèȅěẻẹęȩḝǝĭȋîĩḭïḯīíìȉǐỉịįıŏȏôỗốồổộõṏṍöōṑóṓòȍǒỏọǫǭơỡớờởợøǿŭȗûṷũṻṹṵüǖǘǜǚṳūúùȕǔủůụųưữứừửựŷỹÿȳýỳỷẙỵ' [20:35:34] hahaha [20:35:37] is there a doc on how to debug lua script issues? [20:36:27] like normal templates? [20:37:36] {{Information field|name={{int:wm-license-information-author}}|value={{#if: {{{Author|1}}} [20:37:36] | {{{Author|''Author's name''}}} | {{conj|type=or|{{unknown|author}}|{{not provided}} }} }} }} [20:37:36] {{#if: {{{Creator|1}}} |{{Information field|name={{NARA-image-full/I18n|Creator}}|value={{{Creator|''Creator's name''}}} }} }} [20:37:49] ^d and jeremyb: honestly it kinda looks like it could have been doing this all along. not saying it is, but that at first glance I can see why it would and nothing to stop it [20:38:29] jeremyb: the issue is in {{conj}} [20:38:42] manybubbles: are we on cirrus or lucene now? [20:39:14] default is still lucene [20:39:15] manybubbles: i feel like there's an 80+% chance that I would have doen this before today (and noticed if it was like that) [20:39:34] but can't be sure [20:39:37] jeremyb: commons is lucene. cirrus doesn't display this behavior. it actually rips the namespace off in the suggestion which may not be right either [20:39:46] jeremyb: I understand [20:41:00] manybubbles: anyway, in cases where the searched name is an existing page, maybe the link to that page should be more prominent than spelling suggestions [20:41:52] jeremyb: understood. I'm meeting with designers in 20 minutes and will bring it up. [20:41:55] jeremyb: also, https://bugzilla.wikimedia.org/show_bug.cgi?id=58348 [20:42:02] manybubbles: good timing :-) [20:45:00] ^d: well, it's rather silly that they are CamelCased [20:45:11] i think just about all translations use Normal case [20:45:33] <^d> PeopleShouldUseCamelCaseInEverydayConversation. [20:45:34] (in page names too, not just their titles) [20:46:40] you know, i might file a bug on that [20:47:06] and maybe cc a few designers to make it more unpleasant for everyone :> [20:48:11] manybubbles: btw, seems like there's some kind of race where [[special:search/foo]] and typing "foo" in the box and hitting enter both will usually find and take you to a match even if it's a different case. and hitting the "search" button will take you to search results. *but* if the timing's wrong (i guess loading some JS or ajax) then hitting enter takes you to search results instead of straight to the page [20:48:36] <^d> MatmaRex: Please save my remaining bits of sanity by not CC'ing me on that bug :) [20:49:07] ^d: yes. I think that is the case. If you mash enter before the ajax has come back I'm pretty sure you just get the search results. [20:49:12] jeremyb: ^^ [20:49:14] not ^d [20:49:33] <^d> I was like huh? [20:49:34] <^d> :p [20:49:41] I can think of worse ways to do it, but I may as well bring it up in 10 minutes [20:49:49] manybubbles: why not just behave like [[special:search/foo]] ? [20:50:30] * jeremyb wonders who's in the meeting? [20:51:48] jeremyb: Can you file a bug against mw's search saying exactly what you think should happen? I think it'd be simpler for me to talk about. and I'll make sure I understand you 100% [20:51:56] jeremyb: dan garry and other people I don't yet know [20:52:37] manybubbles: i'm not quite sure i know what i think. or that i'm the best person to decide. besides that "it shouldn't matter whether or not the ajax had arrived yet" [20:52:49] i.e. there shouldn't be a race [20:52:59] races are surprising [20:53:34] you mean to the user? [20:55:10] anyway, have fun @ mtg... :) i have to run