[04:12:43] $popts->enableLimitReport( !$params['disablepp'] ); [04:13:07] API doesn't allow boolean parameters that are default true, so they're all "disablefoo" [04:13:31] it's the "pp" bit that is making me say wtf [04:13:37] "apihelp-parse-param-disablepp": "Disable the PP Report from the parser output.", [04:13:50] as if whoever introduced this parameter didn't know what PP stood for [04:14:16] I mean, it is the limit report, not the PP report [04:14:53] I just put NewPP in there so that its output could be distinguished from the one generated by the pre-MW 1.12 parser [04:15:31] there was a transition period where you could select which parser you wanted to use [04:15:51] that's why it is enableLimitReport(), not enablePPReport() [04:16:42] ah [04:16:42] lol [04:16:42] gitblame points to Reedy [04:16:58] https://www.mediawiki.org/wiki/Special:Code/MediaWiki/70450 [04:19:44] I need to make a batch parse interface of some description [04:20:45] I was hoping to be able to put it in an extension but I'm not sure if that is feasible [04:22:00] maybe I should just make a new module instead of overloading the two current parser modules [04:23:16] input format probably needs to be JSON, since that is fashionable [04:23:33] in an extension I might have had the flexibility to choose it [04:27:19] legoktm: do you remember any talk about a generic API batch interface, maybe at lyon? [04:27:58] I couldn't find any documents talking about such a thing [04:29:39] TimStarling: I remember talking about it and a lua thing, I don't know if we took any notes... [04:29:53] TimStarling: https://etherpad.wikimedia.org/p/api.php [04:30:53] ah yes, thanks [04:33:45] it would be better to avoid a lua dependency in this case, if the idea is to support all wikis that parsoid supports [04:40:56] it is daunting [04:41:02] maybe I should just make a new extension for it [07:11:11] input format probably needs to be JSON, since that is fashionable [07:11:13] hahaha [07:12:04] your grudge against JSON is so bizarrely intense and specific [07:12:32] ;) [07:14:37] how's it going, aaron? [07:14:53] i met peter in person finally, he seems great [07:16:15] * AaronSchulz listens to some Owl City [07:16:42] oh cool, haven't heard of him [07:16:47] listening to 'fireflies' now [07:17:19] good choice [07:17:25] oh, i know this song [07:17:28] it sounds like the postal service [07:17:34] lol [07:18:12] like https://www.youtube.com/watch?v=0wrsZog8qXg [07:18:48] my chest still aches as punishment for having come back to the gym after months of laziness [07:20:20] don't overdo it! [07:20:33] https://www.youtube.com/watch?v=KGEyqP0744c [07:20:51] yeah that's a good song [07:23:36] * AaronSchulz awaits more cat ears [07:25:09] Etsy can be addictive you know [07:27:02] they have a great performance engineering team, by the way [07:27:28] https://codeascraft.com/2015/07/13/q2-2015-site-performance-report/ [07:30:00] yeah if I ever left here, that would be good choice probably [07:31:30] yeah it's a nice place to work at [07:31:41] they have a very gorgeous office as you might expect in a nice part of brooklyn [07:32:05] i still think we're cooler :P [07:32:32] but the quarterly site performance report is totally something we should do [08:11:18] did you know JSON.stringify() can't be used to generate JavaScript code fragments containing arbitrary strings? [08:12:47] there are two unicode characters that it screws up, and PEG.js uses both of them in its example parser to parse JavaScript code [08:13:29] I wonder if there is something that it is actually good for [13:17:45] TimStarling: "batch parse interface", as in passing multiple fragments to be parsed individually? I've been trying to think of a good generic way to support multi-value in the API when the values might contain '|', for parse/expandtemplates since you mentioned it once before, and for action=options, and maybe for other params too. Options include: PHP foo[] syntax or some variation, fookeys=foo1|foo2|foo3&foo1=…1…&foo2=…2…&fo [13:17:45] o3=…3…, or fooseparator=🙌&foo=…1…🙌…2…🙌…3…. I don't much like JSON blobs as input parameters to the API since it leads to a single parameter with documentation that boils down to "pass an undocumented JSON blob here". [14:26:41] MatmaRex: What's a good way to locally load the latest oojs-ui in MW, instead of waiting for +2s and releases? [14:27:29] anomie: Once you've built it in the OOUI repo, you can just `cp -r dist/` it to MW core's resources/lib/oojs-ui [14:27:37] anomie: run `grunt build, then copy entire oojs/ui/dist directory over to resources/lib/oojs-ui [14:27:45] Thanks! [14:27:46] anomie: (If you're just after JS; if you need the PHP too, you need to add to /vendor.) [14:28:05] anomie: do not remove the additional JSON files that are there, or copy them from oojs/ui/src/somethingsomething [14:28:33] (oojs/ui/src/themes/mediawiki) [14:28:36] Yeah, now that https://gerrit.wikimedia.org/r/#/c/216909/ and https://gerrit.wikimedia.org/r/#/c/224966/ exist I'm going to look at having my WIP ApiSandbox code use them. [14:29:34] anomie: the script used for the regular updates is /maintenance/resources/update-oojs-ui.sh , perhaps we should extract the middle part which does that actual copying to something standalone. [14:30:00] MatmaRex: Not before we fix it. [16:01:20] AaronSchulz: Aaron Schulz (talk | contribs)‏‎ (edit filter manager, administrator) (Created on 15 July 2005 at 19:05) [16:01:39] AaronSchulz: happy wikipedia decennary [17:22:12] legoktm: https://gerrit.wikimedia.org/r/#/c/225012/2 [17:25:04] AaronSchulz: is that class already using the WAN stuff? [17:25:20] yes, that's the problem [17:25:29] it has for some time [17:25:35] +2'd [17:33:12] legoktm: I guess there is https://gerrit.wikimedia.org/r/#/c/225013/ while at it [18:02:57] AaronSchulz: will the cache key change? [18:06:41] legoktm: indirectly, yes, via a prefix [18:19:53] AaronSchulz: so how is that going to work with two wmf branches? do we need to deploy it specially? [18:23:56] legoktm: it can delete both, though I missed a spot in invalidateCache() [18:25:05] actually that should just wrap quickInvalidateCache anyway [18:28:47] * AaronSchulz updates [19:36:14] SMalyshev: http://www.amazon.com/Graph-Databases-Opportunities-Connected-Data/dp/1491930896 [19:36:34] (just published, I found out about it via http://anynewbooks.com/) [19:54:25] anomie: alright, i just finished with things i had half-done and recent breakages, and now i'll finally do some of the things you need me to do. which do you want first? :D i'll grab a dinner, and then do that. [19:55:50] MatmaRex: Either https://gerrit.wikimedia.org/r/#/c/216909/ or https://gerrit.wikimedia.org/r/#/c/224966/, I'll need both to get ApiSandbox moving. [19:56:36] ori: thanks! [19:58:20] ori: judging by mentioning Cypher, it's probably Neo4j-oriented, but still can be interesting [21:41:15] ostriches: https://gerrit.wikimedia.org/r/#/c/224735/ [23:24:51] anomie: i dumped another load of comments on the capsule patch. that's it from me for today, good night :)