Page MenuHomePhabricator
Paste P3179

2016-05-25 ArchCom-RFC triage meeting (#wikimedia-office)

Authored by RobLa-WMF on May 25 2016, 10:12 PM.
21:00:44 <robla> #startmeeting E187 ArchCom-RFC triage meeting
21:00:44 <wm-labs-meetbot> Meeting started Wed May 25 21:00:44 2016 UTC and is due to finish in 60 minutes. The chair is robla. Information about MeetBot at
21:00:44 <wm-labs-meetbot> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:00:44 <wm-labs-meetbot> The meeting name has been set to 'e187_archcom_rfc_triage_meeting'
21:01:33 <robla> #topic E187 ArchCom-RFC meeting - Please note: Channel is logged and publicly posted (DO NOT REMOVE THIS NOTE) | Logs:
21:02:04 <robla> hi folks!
21:02:44 <Scott_WUaS> Hi
21:03:07 <Scott_WUaS> :)
21:04:45 <robla> #link
21:05:16 <TimStarling> is the idea to assign priorities?
21:05:38 <robla> TimStarling: yeah, I think there's a couple of things I'd like to accomplish:
21:05:55 <robla> 1. Assign priorities for the "Backlog" items that say "need triage"
21:06:07 <robla> 2. Pick a good candidate for next week's IRC meeting
21:06:22 <robla> so....let's start with the top of the list I just linked to:
21:06:35 <robla> T128602
21:06:35 <stashbot> T128602: Create and deploy an extension that implements an authenticated key-value store. -
21:07:45 <robla> TimStarling: gwicke : Krinkle : is T128602 one that we just discussed in our meeting in the last hour today?
21:07:45 <stashbot> T128602: Create and deploy an extension that implements an authenticated key-value store. -
21:08:00 <TimStarling> don't think so
21:08:33 * gwicke does not think so either
21:08:56 <TimStarling> all the discussion was in march apparently
21:09:34 <TimStarling> is brion here?
21:09:52 <TimStarling> or tgr, Krenair, dbrant
21:11:01 <robla> perhaps we can set it to "normal" or "low" priority, and wait for objection?
21:11:02 <TimStarling> I think this RFC would not be controversial if it were discussed
21:11:14 <TimStarling> yeah, low priority
21:11:46 <TimStarling> it needs a committed client, once it has one of those, the priority can be higher
21:11:54 <robla> on the list: T120380
21:11:54 <stashbot> T120380: RFC: Allow JSON values to be included in the API results -
21:12:55 <MaxSem> I'll summon yurik
21:13:16 <MaxSem> ...and he's not online :P
21:13:42 <robla> ok....normal? low? high?
21:14:18 <TimStarling> I think normal? there is conflict, which probably makes it more our responsibility than the implementor's
21:14:42 <TimStarling> anomie seems to be dead against it
21:14:44 <robla> normal it is
21:15:17 <robla> ok, next up: T119050
21:15:17 <stashbot> T119050: Parametric JSON builder -
21:15:21 <robla> same story with that one?
21:16:11 * robla prepares to say "yes" and move on
21:16:47 <robla> ok.... normal for 119050. Next up: T117550
21:16:48 <stashbot> T117550: [RFC] Content bundler -
21:17:06 <robla> that one is hexmode 's. Is he around?
21:18:44 <TimStarling> can you remind me what the criteria are for choosing the priority?
21:19:35 <robla> TimStarling: good point. I'm thinking priority should have a bit of a temporal element to it for our purposes
21:19:58 <TimStarling> something which needs to be discussed soon?
21:20:37 <TimStarling> I think this is low priority
21:21:02 <robla> yeah, so basically we should say this: "High" means "let's dedicate a meeting in the next 1-3 months (if not next week)
21:21:14 <TimStarling> on the basis that it may be superseded by shadow namespaces, and nobody is really pushing for it
21:21:29 <robla> "Medium" means "we really should resolve this by the end of the calendar year"
21:21:41 <TimStarling> we can discuss whether the requirements were met after shadow namespaces is done
21:22:36 <robla>, 117550 we can say is "low" priority, pending shadow namespaces
21:22:57 <robla> Next up: T114640
21:22:58 <stashbot> T114640: RFC: make Parser::getTargetLanguage aware of multilingual wikis -
21:22:58 <gwicke> there is also a connection to OCG's bundler stuff
21:23:57 <TimStarling> high? seeing as daniel keeps talking about it, and cscott is promising to prototype it (as of a month ago)
21:24:28 * cscott delurks
21:24:53 * robla waves at cscott
21:25:28 <robla> high priority for 114640 then? Sounds good to me
21:25:34 <robla> next up....
21:25:37 <cscott> oh, right, the {{#lang}} thing. Yeah, I could totally whip up a prototype for that....
21:25:43 <TimStarling> cscott: we are choosing priorities for RFCs, where priority has some kind of relationship to meeting scheduling
21:26:05 <cscott> one way or another it would be good to straighten out this whole "user interface language" thing.
21:26:23 <gwicke> I'm starting to think that we generally need to come up with a high-level strategy for how multi-lingual content should interact with caching & parsing
21:26:24 <TimStarling> so high = a relatively high chance of being scheduled in the next few months
21:26:36 <cscott> sure, just give me, say, 2 weeks notice before we schedule 114640 for discussion and i ought to be able to whip up a prototype to discuss.
21:26:39 <gwicke> these RFCs that focus on a part of the problem seem to be too narrow
21:27:04 <gwicke> we have had that come up repeatedly in the last weeks
21:27:13 <TimStarling> cscott: do you have any RFCs you would like to see discussed soon, e.g. next week?
21:27:14 <cscott> gwicke: sure, the point of the prototype is mostly to learn-by-doing and understand the problem better. in my case at least.
21:27:18 <TimStarling> next week's slot is empty
21:27:40 <cscott> TimStarling: i won't be able to attend the meeting next week, alas.
21:27:45 <cscott> i will be in bermuda
21:27:54 <TimStarling> poor you
21:27:55 <cscott> on a big boat
21:28:08 <TimStarling> ;.(
21:28:41 <robla> next week's meeting:
21:29:14 <robla> (it's boilerplate right now....let's figure out a candidate for next week)
21:29:51 <cscott> gwicke: without going too off-topic for your scheduling meeting, i agree that there are large issues related to how commons and wikidata handle multilingual content... and how they mix UX (via templates) and "content".
21:30:11 <robla> anyway, it looks like 114640 (or any of cscott's) are bad choices. We'll set T114640 to high, and maybe move on
21:30:11 <stashbot> T114640: RFC: make Parser::getTargetLanguage aware of multilingual wikis -
21:30:14 <gwicke> cscott: the elephant in the room is caching & fragmentation
21:30:47 <gwicke> currently, none of the proposals seem to address that seriously
21:30:49 <cscott> yeah, it may be worth trying to build new tools that let commons/wikidata/etc construct UX in a caching-friendly way.
21:31:37 <robla> cscott: next up: T114454
21:31:38 <stashbot> T114454: [RFC] Visual Templates: Authoring templates with Visual Editor -
21:31:38 <cscott> some sort of template system that lives between the existing handlebars templates and the existing wikitext templates, perhaps.
21:31:45 <cscott> oh, speak of the devil! ;)
21:32:19 <robla> cscott: we're using this as the list, and you're name is on there a lot :-)
21:32:42 <cscott> here maybe i'll take gwicke's point and say it would be good from an RFC perspective to have a higher-level discussion about templates and components and caching.
21:32:57 <TimStarling> pretty obvious who the author is when it talks about templates being "hygenic by default"
21:33:29 <cscott> oh, i thought i'd scrubbed the word "hygenic" from all my proposals
21:33:38 <robla> :-)
21:34:03 <TimStarling> I think this is low? pretty old and not recently discussed in the parsing team
21:34:29 <robla> priority of T114454 , "low", plus a comment "please file a bigger picture RFC so that we can block this RFC with that RFC?
21:34:29 <stashbot> T114454: [RFC] Visual Templates: Authoring templates with Visual Editor -
21:34:34 <cscott> one of those things I think i want to take an afternoon to implement before discussing it further
21:34:46 <cscott> robla: i'm pretty sure gwicke already has a bigger picture rfc for that
21:35:01 <robla> cscott: which one?
21:35:10 <cscott> "page components"/templates
21:35:30 <gwicke> that's more focused on content, and not so much on UI
21:35:54 <gwicke> the client-side front-end one is the UI complement
21:36:02 <robla> cscott: Phab task?
21:36:12 <cscott> i had an unconference session at the all-hands on the visual templates stuff and it went terribly, turns out everyone who signed up assumed i was going to teach them to use mediawiki's existing template system, not propose some new doesn't-exist-yet hotness
21:36:24 <gwicke>
21:36:31 <cscott> so in the unconference spirit the session got redirected to match what people actually wanted
21:36:54 <Scott_WUaS> (gwicke: can you please share some URLs that currently focus "multi-lingual content should interact with caching & parsing"?)
21:36:56 <gwicke> it's not specifically targeted at i18n currently
21:37:09 <cscott> Scott_WUaS: that's daniel's RFCs
21:37:28 <cscott> Scott_WUaS: T114662 and T114640
21:37:28 <stashbot> T114662: RFC: Per-language URLs for multilingual wiki pages -
21:37:28 <stashbot> T114640: RFC: make Parser::getTargetLanguage aware of multilingual wikis -
21:37:42 <cscott> gwicke has commented on both of those, i believe, bringing up the caching issue
21:37:48 <Scott_WUaS> Thnx:)
21:37:58 <TimStarling> T114432 is next on the list, I thought it was a nice idea, but reception at the dev summit was mixed IIRC, and then there was no further discussion after that, so I guess it is low
21:37:58 <stashbot> T114432: [RFC] Heredoc arguments for templates (aka "hygenic" or "long" arguments) -
21:38:12 <gwicke> has some high-level thoughts on the general problem area
21:39:35 <cscott> I think T114432 actually had a meh reaction more than a negative one. i don't think people were opposed, but there is a very vocal contingent who reacts negatively whenever anything is said "to make things easier for visual editor"
21:39:36 <stashbot> T114432: [RFC] Heredoc arguments for templates (aka "hygenic" or "long" arguments) -
21:40:15 <robla> cscott: right, but are you disagreeing with TimStarling's proposed "low" priority?
21:40:18 <SMalyshev> it looks like pretty arcane addition to already quite arcane syntax
21:40:36 <gwicke> it might be useful to think through & write down how an API-driven frontend could help i18n specifically
21:41:12 <cscott> robla: "low" seems right to me, i'm not blocked on further rfc discussion
21:41:36 <robla> thanks cscott - ok, next up:
21:41:50 <robla> T114421
21:41:50 <stashbot> T114421: [RFC] Optional Travis integration for Jenkins -
21:42:00 <robla> also cscott :-)
21:42:35 <cscott> i still use my npm-travis tool, but ops seems violently opposed.
21:42:54 <cscott> containerized test jobs is always right around the corner, and will eliminate any need for travis, i'm told.
21:43:14 <robla> cscott: is that going to be in perpetual "agree to disagree" state, or do you see a way forward?
21:43:56 <cscott> well, i don't have the appetite to fight that particular battle myself. it seems that teams are just going around ops and using travis if they need it.
21:44:22 <cscott> that seems dysfunctional, but it's a dysfunction i'm not in a good place to address. it's not entirely clear that the RFC process is a good way to address it either.
21:44:22 <gwicke> cscott: is *ops* really opposed? I only see releng on the task
21:44:27 <cscott> so i'd say status "stalled"
21:45:02 <robla> priority is "low" for ArchCom, I may even be one that we take ourselves off of
21:45:26 <gwicke> hehe
21:45:26 <cscott> yeah.
21:45:45 <robla> ok, next up
21:45:56 <robla> T114394
21:45:56 <stashbot> T114394: RFC: PageLookup service and PageRecord object -
21:46:04 <TimStarling> with 15 minutes left, should we soon talk about scheduling for next week?
21:46:05 <cscott> at the time in the flush of early excitement about the rfc process i may have abused it to try to settle impasses ;)
21:46:19 <cscott> i think my rfcs are done, right?
21:46:38 <robla> cscott: yup
21:46:38 <TimStarling> cscott: yes
21:47:17 <cscott> recapping from my perspective only, i think prototyping {{#lang}} (or whatever) was the only thing that rose to 'high' priority?
21:47:24 <robla> TimStarling: good point. So....we didn't hit paydirt on anything we talked about yet as a candidate for next week
21:47:56 <robla> anyone have a nomination for next week?
21:48:05 <cscott> T91137?
21:48:05 <stashbot> T91137: RFC: Support a branching content history model -
21:48:12 <cscott> although i'd like to be there for that one as well, i guess.
21:48:37 <cscott> T351 is mine too, i think, don't know why Qgil's name is associated.
21:48:37 <stashbot> T351: RfC: Square bounding boxes -
21:48:56 <cscott> what about T96384? that seems like it should be straightforward perhaps?
21:48:56 <stashbot> T96384: Integrate file revisions with description page history -
21:49:03 <TimStarling> the low-numbered RFCs were copied from the wiki by qgil
21:49:06 <cscott> that's a @daniel rfc
21:49:08 <robla> T382 looks like an interesting choice
21:49:09 <stashbot> T382: RfC: Server-side Javascript error logging -
21:49:31 <robla> cscott: we probably shouldn't choose one of yours if you're out next week
21:50:43 <cscott> robla: yeah, that's why i nominated a T96384, a @daniel RFC. dunno his availability.
21:50:43 <stashbot> T96384: Integrate file revisions with description page history -
21:51:48 <cscott> hm, there's a @csteipp RFC as well, maybe that's worth getting to before he transitions out of the WMF headspace? T75953
21:51:48 <stashbot> T75953: RFC: MediaWiki HTTPS policy -
21:52:11 <cscott> it will probably be harder to get his participation if that if deferred for a few months
21:53:41 <robla> ok....maybe in general I can take it as an action to figure out a security related RFC for next week. We also have T135963 as a recent one that has big implications worth discussing
21:53:41 <stashbot> T135963: Add support for Content-Security-Policy (CSP) headers in MediaWiki -
21:54:05 <TimStarling> sounds good
21:54:25 <robla> Krinkle basically advised that we hold off a little bit on that one, though. He's going to be shepherding it. Still I'll be happy to own figuring this out
21:54:41 <TimStarling> maybe csteipp will have an idea for a task which is not even tagged as an RFC yet
21:54:51 <dapatrick> robla, why did Krinkle advise holding off?
21:55:26 <TimStarling> he has 53 tasks assigned to him
21:55:28 <robla> dapatrick - I think just because he had only just volunteered to shepherd, and I had just put him on the spot with "how about next week?"
21:55:35 <Krinkle> dapatrick: I'm not advising to hold off on the project in general. Merely holding off on scheduling the IRC meeting as I familiarise myself with the task.
21:55:40 <Krinkle> 1 week :)
21:55:48 <TimStarling> authored 100
21:56:07 <TimStarling> csteipp authored 100 open tasks
21:56:21 <dapatrick> robla, Krinkle I see.
21:56:24 <cscott> +1 to robla's plan to triage csteipp
21:56:55 <cscott> and maybe T135963 for the week after that, it can be a security fest ;)
21:56:55 <stashbot> T135963: Add support for Content-Security-Policy (CSP) headers in MediaWiki -
21:57:11 <robla> think we're done with the realtime portion of the meeting. However, we also have Z425 as a place to continue this conversation in a less synchronous fashion
21:57:33 <robla> #link
21:57:53 <robla> #info triage discussion can continue on Z425
21:58:13 <Scott_WUaS> (Great process emerging here re Phabricator - thanks, all!)
21:58:28 <robla> any last minute comments/questions/concerns before I officially end this meeting?
21:59:21 * robla might actually end it *dozens of seconds* early
21:59:38 <robla> #endmeeting