[06:01:58] <LA2>	 call me old or stupid, but this help page would benefit from explaining what's meant with "instance" and "node", before using the terms, https://labsconsole.wikimedia.org/wiki/Help:Access
[06:06:32] <LA2>	 and here https://labsconsole.wikimedia.org/wiki/Instances "instances are not projects, see..." and the link goes to a blank page
[06:12:53] <Ryan_Lane>	 LA2: well, it seems the semantic search is broken
[06:13:20] <Ryan_Lane>	 also, feel free to link to this page anywhere you see fit: https://labsconsole.wikimedia.org/wiki/Help:Terminology
[06:14:00] <LA2>	 yes, I've linked to that. Thanks.
[06:14:15] <LA2>	 I'm coming to Berlin, so this gives me a push into the Labs environment.
[06:14:32] * Ryan_Lane  nods

[06:15:21] <Ryan_Lane>	 well, it's having some performance issues as of late, so expect a slightly poor experience until we make some changes
[06:17:18] <LA2>	 If all I wanted to do was to download an XML dump and run some Perl scripts on it, but I would like to share this with a group of other users, making my private computer unsuitable. Can I do this as a Labs instance? Once I thought that the German toolserver would be a good place to do things like that, but there are practical details in bureaucracy (quotas etc.) that stops me.
[06:20:06] <LA2>	 the definition of "instance" here implies that I want to run things that are web tools and/or have database access
[06:21:47] <Ryan_Lane>	 yes
[06:21:52] <Ryan_Lane>	 but you don't need to download an xml dump
[06:22:03] <Ryan_Lane>	 they are accessible to all instances
[06:22:07] <LA2>	 fine
[06:22:11] <Ryan_Lane>	 via nfs
[06:22:29] <Ryan_Lane>	 there isn't database access in labs right now
[06:22:36] <Ryan_Lane>	 not replicated copies, anyway
[06:26:57] <LA2>	 for some XML dump analysis jobs, I found tables such as 'namespacename' are very useful to have. This changes on a monthly basis, if ever, and doesn't need live replication
[06:31:57] <Ryan_Lane>	 well, all databases are user created right now
[06:32:29] <Ryan_Lane>	 basically what are giving people right now is a way to create virtual machines that can easily shared and configured as a community
[06:32:50] <Ryan_Lane>	 as time goes on, we'll be adding more and more services
[06:33:52] <LA2>	 in an alternative universe, namespacenames would be software constants that I get through git, rather than live replicated database tables
[06:49:29] <Ryan_Lane>	 umm
[06:49:30] <Ryan_Lane>	 no
[06:49:56] <Ryan_Lane>	 that's the opposite direction they should go
[06:50:32] <Ryan_Lane>	 they'll eventually move more into the database than they currently are
[06:52:15] <apergos>	 ewwww software tables in git
[06:52:17] * apergos  shudders

[06:52:50] <apergos>	 (that's me as the xml dump maintainer shuddering btw)
[06:52:57] <Ryan_Lane>	 ok. wtf
[06:52:59] <Ryan_Lane>	 fatal: reference is not a tree: 490713ad6dca497e1683fac74b05f41c8f768d84
[06:52:59] <Ryan_Lane>	 Unable to checkout '490713ad6dca497e1683fac74b05f41c8f768d84' in submodule path 'extensions/StringFunctionsEscaped'
[06:53:10] <Ryan_Lane>	 did someone break the extension's history?
[06:53:33] <Ryan_Lane>	 this is the wmf/1.20wmf3 branch
[07:01:15] <LA2>	 apergos: I'm reading page.sql.gz with a Perl script. I see namespace numbers, but I need namespace names. How should I get them?
[07:01:58] <LA2>	 a) live replicated database, b) API to live website, c) is there a dump file for this?
[07:02:56] <apergos>	 if you look at the stubs file for the dump you'll see at the beginning a list of namespaces by name and number
[07:03:34] <apergos>	 the stubs files and the page content files all have them
[07:04:42] <LA2>	 That's a mapping from number to canonical namespace name, i.e. 12 > Hjälp (for svwiki)
[07:05:02] <LA2>	 but I need a mapping from English namespace alias (Help) to number
[07:06:15] <LA2>	 alternatively, with the new definition of "content" namespaces, I need to know which they are
[07:06:54] <apergos>	 for the fixed namespaces (below 100) the mapping doesn't change across wikis so you can look at any en dump for that
[07:07:16] <apergos>	 for 100 on up these are wiki specific, so there is not a canonical en mapping necessarily
[07:08:20] <LA2>	 API (siteinfo) says:       <ns id="106" case="first-letter" canonical="Författare" content="" xml:space="preserve">Författare</ns>
[07:08:58] <LA2>	 hm... shouldn't canonical be "Author" here...?
[07:09:54] <apergos>	 I think canonical in those cases is the content language of the wiki, (my opinion)
[07:10:19] <LA2>	 for sub-100, canonical= the ENglish name, e.g. <ns id="2" case="first-letter" subpages="" canonical="User" xml:space="preserve">Användare</ns>
[07:10:56] <LA2>	 anyway, the API-siteinfo looks different from the XML dump intro
[07:11:09] <apergos>	 for sub 100 sure
[07:11:45] <LA2>	 XML dump says: <namespace key="2" case="first-letter">Användare</namespace>
[07:12:02] <LA2>	 if that had content="" and canonical="User", all would be fine
[07:12:04] <apergos>	 yes it does
[07:12:21] <apergos>	 it produces the namespaces in the content language of the wiki
[07:12:31] <apergos>	 since that's how they will be listed in links and every other thing
[07:14:50] <LA2>	 there is another use case, though: if I want to analyze all "Portal:" pages, I don't know what the name of the "Portal:" namespace is in every language
[07:16:00] <apergos>	 φορ 100 ανδ θπ ιτ'σ γοινγ το βε α προβλεμ
[07:16:01] <apergos>	 er
[07:16:08] <apergos>	 for 100 and up it's going to be a problem
[07:16:21] <apergos>	 what api query are you using to get the list?
[07:16:41] <apergos>	 http://el.wikipedia.org/w/api.php?action=query&meta=siteinfo&siprop=general|namespaces|namespacealiases|statistics    this shows me the en names for below 100
[07:16:44] <LA2>	 so far, I used SQL on the German toolserver, table namespacename
[07:17:02] <LA2>	 yes, this is the API call, http://sv.wikisource.org/w/api.php?action=query&meta=siteinfo&siprop=general|namespaces|namespacealiases|statistics
[07:17:44] <apergos>	 so where somenoe has defined an alias it's availabl
[07:17:45] <apergos>	 e
[07:18:06] <LA2>	 now that content="" is in the API, I could use this, so I don't really need the names of namespaces
[07:18:24] <LA2>	 but it would be fine if content="" was also in the XML dump
[07:18:58] <LA2>	 but that's my application, where I care about content namespaces
[07:19:18] <LA2>	 other users might still need to find the Portal: namespace without knowing its name in the local language
[07:19:44] <LA2>	 if canonical= were always in English, as it is for sub-100, that would help
[07:20:47] <LA2>	 ah, the API also has <namespacealiases> with the English names
[07:23:27] <apergos>	 yes, that's what I was saying
[07:26:55] <apergos>	 I would really like to see the namespace info move into the db and out of the config files, hope they do so soon
[07:27:30] <apergos>	 give crats the ability to add/change those.  have to think about dup page titles and all that but still
[14:38:11] <toniher_casa>	 Hi, I've recently got an extension migrated https://gerrit.wikimedia.org/r/gitweb?p=mediawiki/extensions/UserFunctions.git;a=tree . But I cannot push anything. What should I do? Any hint in the docs?
[14:39:55] <Hydriz>	 toniher_casa: Gerrit is different from other Git systems :)
[14:40:52] <Hydriz>	 Do read through this: https://www.mediawiki.org/wiki/Git/Workflow
[14:40:56] <toniher_casa>	 Hydriz, hehe. I've seen. But I don't have any gitreview in my repo
[14:41:03] <toniher_casa>	 .gitreview
[14:41:10] <toniher_casa>	 should I ask someone to add it?
[14:41:35] * Hydriz  shrugs, and goes back to studying all about Git

[14:44:18] <toniher_casa>	 After gitreview -s Message: Problem running 'git remote update gerrit' Fetching gerrit fatal: '/mediawiki/Extensions/UserFunctions.git': not a Gerrit project fatal: The remote end hung up unexpectedly error: Could not fetch gerrit
[14:44:27] <toniher_casa>	 I guess something is missing
[14:52:49] * toniher_casa  is going to write to Chad (^demon) Thanks Hydriz

[14:53:04] <Hydriz>	 lol np :)
[21:55:55] <kaldari>	 When I do "git branch --track wmf/1.20wmf3 origin/wmf/1.20wmf3", it says "fatal: Not a valid object name: 'origin/wmf/1.20wmf3'."
[21:56:23] <saper>	 kaldari: does "git remote update" help?
[21:56:47] <kaldari>	 let's see...
[21:56:59] <kaldari>	 yes! thank you
[21:57:19] <Reedy>	 1.20wmf4 starts tomorrow ;)
[22:03:49] <kaldari>	 I'm just practicing :)
[22:05:42] <aude>	 is content handler making it into 1.20wmf4 ?
[22:13:30] <Reedy>	 Nope...
[22:13:48] <Reedy>	 I think the plan was to merge it sometime soon after, giving it a couple of weeks to further stabilizwe
[22:18:47] <aude>	 Reedy: :(
[22:19:07] <aude>	 it's blocking us
[22:20:12] <Reedy>	 Indeed
[22:20:35] <aude>	 more review would be good
[22:20:51] <Reedy>	 +4798, -339 does not a trivial change make
[22:21:45] <saper>	 aude: which rXXX is this currently?
[22:21:57] <Reedy>	 https://gerrit.wikimedia.org/r/#/c/7716/
[22:22:20] <Reedy>	 if it gets merged for wmf5, it'll start becoming live within 2 weeks
[22:22:35] <aude>	 we're working in a branch now
[22:22:55] <aude>	 more stuff to merge in
[22:22:59] <aude>	 after this one
[22:34:06] <saper>	 aude: -1 for me for ignoring other RDBMS's
[22:34:27] <Reedy>	 Ouch
[22:35:24] <aude>	 saper: yikes
[22:36:16] <saper>	 need to check my remarks to the old change (don't know why it got abandoned, but fine)
[22:36:34] <Reedy>	 I think it was just due to teh mass of changes
[22:36:36] <Reedy>	 to keep it tidier
[22:37:46] <saper>	 well I had a problem with use of MySQL-specific data types before, need to check if it got better now.
[22:55:35] <saper>	 aude: ah and this https://bugzilla.wikimedia.org/show_bug.cgi?id=36759 <-- one change done as one of the previous changes related to wikidata, still waiting for investigation... I suppose problem with transaction isolation, if yes we should check for possible sources of trouble for other code as well
[22:57:59] <aude>	 saper: ok