[12:07:01] thare are problems with genereting thumbnails from multi-page large tif: https://commons.wikimedia.org/w/index.php?title=File:UBBasel_Map_1569_Kartenslg_AA_3-5.tif&page=2 [12:07:18] message is "Sorry, the servers are overloaded at the moment" [12:07:26] is it a temporary problem? [12:12:20] ankry: does it occur just for that url on commons for you? [12:13:19] <_joe_> how big is that tif file? [12:21:02] <_joe_> ankry: the tiff file is huge, I think it's a bit too big for our infrastructure to handle smoothly [12:21:12] <_joe_> (1.5 gb huge) [12:24:57] Pretty sure it's the file size. [12:26:46] looks like someone ticket the 'don't compress' option in the system generating this tiff, thinking you need that to create a lossless image. [12:30:35] i'm glad i have 16GB of RAM..... [12:35:41] Is this the right place for getting help with Quarry? [12:37:29] some folks might know a few things. Or you can try #wikimedia-labs I guess [12:38:03] It's probably quite simple - I'm trying to run the top query here: https://phabricator.wikimedia.org/T130058#2208278d [12:38:17] But I get: SELECT command denied to user for table 'externallinks' [12:40:10] use enwiki_p [12:40:23] Ah great, that [12:40:27] that's working, thanks :) [12:41:04] Samwalton9: _p for public [12:44:06] ankry: wow, it really does seem like it's 1.5 GB compressed. yikes [12:44:32] ankry: please file a bugreport in phabricator [12:45:36] with the tag pagedtiffhandler [13:10:12] thedj: https://phabricator.wikimedia.org/T134169 [14:14:47] hey, dewiki is dead [14:14:55] hey, dewiki is dead [14:14:56] All wikis are dead? [14:15:04] en.wiki, commons, mw.o [14:15:12] Yes, we have 503s all over the place. Also in Phabricator. [14:15:15] writing by API is possible [14:15:18] otrs also [14:15:19] #wikimedia-operations is aware. [14:15:32] reading by API, too [14:15:36] ok, thanks [14:15:38] Might only be in Europe (esams) [14:15:46] :( [14:16:27] It's also in Europe when it is time to pick kids from schools, so... see you later. :) [14:16:57] Lucky you qgil [14:19:53] how can such a problem be possible? [14:20:18] doctaxon: was that a serious question? :) [14:20:56] IP block? [14:20:58] Recovering now, it seems. [14:21:43] doctaxon: See https://wikitech.wikimedia.org/wiki/Incident_documentation for previous downtimes and their various reasons [14:21:55] thx [14:26:50] can't find a EuropeWikiOutage there [14:28:18] doctaxon, look at esams on the page [14:28:55] doctaxon: or pick a random other one, like e.g. https://wikitech.wikimedia.org/wiki/Incident_documentation/20151127-PartialNetworkOutage or whatever :) [14:31:13] Trizek: what is an esams [14:32:05] That's the name of the server located in Amsterdam, doctaxon. [14:32:38] But don't ask me more, you have seen the limit of my knowledge in my previous answer! :) [14:32:40] ah okay [14:33:03] yes, it's okay [14:34:39] doctaxon, https://wikitech.wikimedia.org/wiki/Esams_cluster [15:45:30] MatmaRex_, thx, could you point me to the exact lines of the compat script? [15:45:54] yurik: see that patch i linked [15:45:55] it's just [15:45:58] MatmaRex_, window.moment = module.exports; ? [15:45:59] window.whatever = module.exports [15:46:04] ah, gotcha, thx [15:46:08] and add that to the RL module definitio [15:46:10] n [15:47:06] MatmaRex_, rl module def? you mean some flag in extension.json? [15:48:25] yurik: https://phabricator.wikimedia.org/rMWbc4e07b6f63b0865a14aef981366a79d10329c87#cb95de99 [15:48:49] put that in a separate .js file and add that file to the 'script's key on module definition [15:49:14] ah, its a new file, thx [15:50:21] still, it sucks that we had to break it before fixing it [15:50:33] i wonder how we can automate unit testing of this [15:56:43] MatmaRex_, are you sure that method will work? Take a look at https://github.com/wikimedia/mediawiki-extensions-Graph/blob/master/extension.json#L69 [15:56:51] vega uses multiple libraries [15:58:41] yurik: ugh, well, that's a bit uglier then, but no reason why it shouldn't work [15:58:49] you'll just need something like [15:58:54] "lib/d3.js", [15:58:54] "src/d3-compat.js", [15:58:54] "lib/topojson.js", [15:58:54] "src/topojson-compat.js", [15:58:54] "lib/vega2/vega.js", [15:58:55] "src/vega2-compat.js", [15:58:55] "lib/graph2.compiled.js", [15:58:56] "src/graph2-compat.js" [15:59:17] (unless any of these don't use module.exports) [15:59:56] yurik: or you could always change the library JS files not to export stuff. but that's going to be annoying when upgrading or whatever [16:00:19] MatmaRex_, i really don't want to touch upstream libraries like that [16:00:29] yeah [16:00:54] MatmaRex_, so i'm a bit confused - how will appending all these scripts one after another would automagically resolve all the module.exports? [16:02:02] MatmaRex_, also, graph2.compiled.js has been browserified, and it uses exports internally [16:02:50] d3 has this magical thingy: [16:02:50] if (typeof define === "function" && define.amd) this.d3 = d3, define(d3); else if (typeof module === "object" && module.exports) module.exports = d3; else this.d3 = d3; [16:05:14] um [16:05:52] `/topic [16:06:00] yurik: that is standard-ish for JS libraries [16:06:09] there's some name for this pattern that i can't recall [16:06:53] yurik: but the only thing you need to care about, it that the library will attach itself to module.exports if 'module' is present, or to window.d3 otherwise [16:07:05] since 'module' is not present, it stopped attaching tiself to window.d3 [16:07:11] so you have to attach it there yourself [16:07:27] stats.wikimedia.org is still not available, at least from EU; other domains seem to work [16:07:32] MatmaRex_, ok. So should i add the shim for a browserified code - https://github.com/wikimedia/mediawiki-extensions-Graph/blob/master/lib/graph2.compiled.js [16:08:15] yurik: dude, how would i know [16:08:18] try if it still works [16:08:23] if it doesn't, yes, you need the shim [16:08:24] :D [16:08:26] heh, the best solution :) [16:08:30] thx though! [16:08:42] i will poke jdlrobson repeatedly for breaking it! [19:08:01] MatmaRex: UMD? [19:08:13] perhaps [19:08:43] (the name of the pattern to export something in a way that works in AMD, CommonJS/Require/Node and browser plain) [19:09:05] Although afiak usually without the 'this' assignment in the define condition. [19:09:56] But yeah, since we now support module.exports, you'll need to add a little file that assigns window.MHY_NAME = module.exports to the RL module if you want to keep the global for back-compat. [19:10:06] This was done in core for moment.js and oo.js [19:25:10] jdlrobson, are you back? [21:26:47] Feels so much like our lists http://xkcd.com/1675/ [21:40:26] maybe we should alter the mailing list footers to make it clearer