[00:46:25] Hi [00:46:44] Is it possible to port [[tswiki:MySQL_queries#List_of_interwiki_links_from_all_wikis]] to toollabs? [00:46:59] https://wiki.toolserver.org/view/MySQL_queries#List_of_interwiki_links_from_all_wikis [01:25:28] hello? [02:02:00] petan, ping [02:03:58] Cyberpower678: do you know SQL? [02:04:15] PiRSquared, why do you ask? [02:14:20] Cyberpower678: Is it possible to port [[tswiki:MySQL_queries#List_of_interwiki_links_from_all_wikis]] to toollabs? [02:14:45] Not that I know of. [02:16:18] I don't have SQL access right now. [02:16:29] So I can't give you a positive answer [02:39:27] PiRSquared: At a glance, it looks like the only hard part would be the lack of a "namespacename" table to get the localized namespace names from. [04:35:50] Coren: Are you here? [04:36:04] SigmaWP: Semi. What's up? [04:36:34] Coren: If you run locale -a, you will see that we're missing quite a few languages [04:36:58] According to http://stackoverflow.com/questions/14547631/python-locale-error-unsupported-locale-setting you have to install locales individually and then run a command [04:37:32] We're not missing them per se, tools currently has the union of the prodution defaults + whatever has been requested to date. [04:37:42] Ah. [04:37:56] Then in that case, can I request every language that has a Wikipedia? [04:42:26] Coren: Eh? [04:42:41] * Coren ponders. [04:43:05] I'm not entirely sure that there even /are/ locale files for all of those languages, certainly no complete ones. [04:43:25] Ah. But we can do our best to cover them all [04:43:36] as many as possible, I mean [04:44:36] Hm; provided the quality is sufficient. I wouldn't even be able to tell you if Tegulu collation was working at all -- let alone correct -- for instance. :-) Mind you, making this a movement-wide project would be a worthwhile endeavor in itself. [04:45:25] Perhaps there already /is/ such a project somewhere for that matter. [04:45:28] * Coren googles. [04:45:47] Is Tegulu even a language? [04:46:15] * SigmaWP not to be confused with "Telugu"... [04:49:57] It's a typo. :-) [04:51:29] Coren: Anyway, thing is, I'm writing a tool that needs to localise timestamps, and I think we're seriously lacking in internationalisation when we don't even have Chinese or Spanish [04:51:34] At first glance, I have partial support for ~124 language, but only about 50 of those seem mostly complete. [04:52:09] SigmaWP: Which language is it written in? Because most have (more complete) i18n libraries available. [04:52:18] Python [04:52:43] strftime() localises strings according to $LANG [04:55:58] Yeah, if you're only doing timestamps changes are locale support should be enough. Most python coders I've seen use Babel for i18n though. [04:56:27] But what do we lose by installing the language packs? [04:56:39] You'd better not say "hard drive space" [04:57:48] Heh. Disk space, actually; compiling all of them takes nearly 8G of OS disk space, and that is actually at rather the premium. I couldn't even do it on the current images. [04:58:26] I could probably perform some trickery with symlinks at need and stuff those in project storage, but that'd have quite the performance impact. [04:59:38] Hmm. [04:59:46] A couple dozen would have little impact, though. [05:01:35] How about, the languages of the UN + languages with > 30 million speakers, taken from https://en.wikipedia.org/wiki/List_of_languages_by_number_of_native_speakers ? [05:02:07] That'd fit. [05:03:00] Barely, but it would. [05:05:37] Open a bz with a list of the locales that'd cover, and I'll make a puppet patch for it Monday. Keep in mind that over 30 is pushing it and 40 is pretty much the upper limit. [05:28:17] Coren: Hmm, on second thought, looking more closely, the languages in the 30-50 million range are spoken by 1 or fewer countries [05:29:04] Which means 25 in total, including the ones we already have [15:46:31] is there a way to access the public-html folder of a tool in tools-dev server using a web browser? [16:04:06] fale: that's just the same folder as on any other server in the tools cluster [16:04:15] so http://tools.wmflabs.org/toolname/filename should work [16:04:53] Coren: don't want to disturb your holy sunday :), just let you know: I traced back and prepared all current database and table definitions from logs. [16:05:18] hedonil: And indeed that sounds like our best bet. Give me a minute and I'll ready this for you. [16:05:35] Coren: fine [16:06:30] hedonil: Allright, go ahead and create the schema. [16:06:49] Coren: thx. will report when ready. [16:20:16] valhallasw: I see :( So it's not possible to test the dev version of a tool while running the stable version? [16:20:56] fale: not in that way, no. You can, however, make a second project, or a subdirectory in your public_html [16:20:59] fale: tools-dev is just another copy of tools-login, it's not distinct. [16:21:49] Coren, valhallasw: I see, thanks :). Is possible with toolsbeta? [16:22:37] fale: Yes, albeit that project is more for experimenting with the infrastructure itself. The "best" way to do what you want is to have a second project for the dev version. [16:25:13] Coren: I see :) thanks for the suggestion. I'll open a new project with the -dev name for development purpouse :) [16:44:07] Coren: Done. rebuilt all db objects, show table status reports all views and tables in maiden state, finally logged out. [16:44:24] Allright, let's try this. Knock on wood. [16:44:34] Coren: yeah. [16:48:57] hello all! I have a issue with the grid engine running python scripts - some help here?? [16:49:33] DrTrigon: what are you doing, and how? [16:50:01] most issues with SGE are path issues, or problems with inconsistent packages on the different execution nodes [16:50:35] running my bot via jsub and that fails - please confer https://bugzilla.wikimedia.org/show_bug.cgi?id=58181 [16:51:06] looks too my like some unicode locale issue - but thats just a guess... [16:51:41] DrTrigon: you need to encode your filename [16:51:57] the script running through the console is basically having good luck [16:53:59] hedonil: I'm sorry, but it's not working. I do have all of the .idb of your tables, but tying them back in is beyond my skill with mysql. In a few hours, our "real" DBA should be back to work (he's in Australia) and we can ask him to attempt it? [16:54:22] Coren: ok. thanks. [16:54:24] DrTrigon: It's almost certainly a local issue; make certain that $LANG is explicitly set to a suitable value. [16:54:47] locale* [16:55:19] vallhallasw: so what to change here: https://git.wikimedia.org/blob/pywikibot%2Fcompat/341ed66bd3ec1ac604aa06598b9f58144891e401/sum_disc.py#L525 ??? [16:55:49] DrTrigon: huh? [16:56:26] oh, it's not the filename, it's the text you're writing [16:56:30] or? [16:57:09] ok, you probably want https://bugzilla.wikimedia.org/show_bug.cgi?id=58181 [16:57:12] er [16:57:13] valhallasw: is this filename not properly unicode encoded? as I see it should be... [16:57:17] pywikibot.config.datafilepath(....).encode('utf-8') [16:57:23] there is no such thing as 'unicode encoded' [16:57:35] http://www.joelonsoftware.com/articles/Unicode.html < read this. [16:58:00] you have a *unicode string*, which python tries to convert to a *bytestring* using the ascii encoding automatically [16:58:39] why does i work on my local machine, the toolserver and labs console?? [16:58:59] all "just by accident"... I don't think so... [16:59:04] because for some reason, python uses 'utf-8' as standard encoding there. [16:59:18] if you think that's not 'just by accident', read the link I just pasted [17:00:33] I think it the locale setting - I will look into this first before starting to think about all the unicode stuff (again)... ;) [17:01:04] ...but if I am wrong I will come back to you - thanks! [17:05:32] Is there something like a "default" or preferred python web framework on tool-labs?? [17:06:29] there are a few people around with Flask experience on tool labs [17:06:37] but in general, any framework should be usable [17:07:28] Actually "any framework should be usable" is exactly my issue... ;) which one to choose...? ;)) [17:07:58] Do some googling, see which one you like. [17:08:14] http://flask.pocoo.org/ looks intressting indeed! Do you know how many guys are using it? [17:09:16] "Do some googling, see which one you like." - ooo man - I did for hours and hours several times... there are a lot which look good... [17:10:29] Sumana is using it, surenm, josh parris, myself [17:11:13] Sumana and you - at least two important ones... ;)) [17:14:11] THANKS TO TWO!!! good night! [20:58:09] Hey all. What other inormation should I be giving? https://bugzilla.wikimedia.org/show_bug.cgi?id=58516 [20:58:13] *information [21:00:04] jarry1250: based on 'being able to emulate production is extremely important', I guess you really want the version used in production [21:00:12] but I think the relevant info is there [21:01:52] Yes. I'm not actually sure what version that is; but apparently it's in maintenance-only mode as a package anyway [21:01:54] So *meh* [21:06:23] jarry1250: apt-get install rsvg would've essentially sufficed [21:06:39] It's not like you're requesting some obscure tool [21:08:49] Reedy: I can run apt-get? [21:09:39] I was more meaning the elabouration in the bug wasn't really needed [21:10:29] Oh, I see. Well, too much information can't hurt, that's the way I figure it. Don't want to be held up, I've got a dozen tools to migrate here :)