Architecture meetings/RFC review 2014-08-13

From mediawiki.org
Jump to navigation Jump to search

2100-2200 UTC on Wednesday, 13 Aug 2014, at #wikimedia-office connect.

Requests for Comment to review[edit]

  1. Who is now the point person for CentralNotice Caching Overhaul - Frontend Proxy?
  2. Followup from 23 July: have we done security update planning re Composer managed libraries for use on WMF cluster?
  3. Followup from 21 May: has cscott run the requested stats on Square bounding boxes?
  4. Followup from 16 July: does Brion know the next steps re CSSJanus on vertical writing support?
  5. Followup from 11 June: how has the bandwidth savings been on Reducing image quality for mobile? Will we change regular/original images as well as thumbnails?

Also, it would be good to have a Wikimania followup/reportout of any big decisions made there.

Summary and logs[edit]

Meeting summary[edit]

Meeting started by sumanah at 21:04:21 UTC. The full logs are available at https://tools.wmflabs.org/meetbot/wikimedia-office/2014/wikimedia-office.2014-08-13-21.04.log.html .



Meeting ended at 21:49:59 UTC.


Action items[edit]

  • csteipp to check with Bryan re Composer managed libraries for use on WMF cluster security planning


Action items, by person[edit]

  • csteipp
    • csteipp to check with Bryan re Composer managed libraries for use on WMF cluster security planning


People present (lines said)[edit]

  • sumanah (77)
  • brion (25)
  • yurikR (12)
  • superm401 (9)
  • AndyRussG (8)
  • ghostsumana (7)
  • awight (7)
  • mark (5)
  • csteipp (5)
  • milimetric (5)
  • wm-labs-meetbot (4)
  • ori (3)
  • awjr (3)
  • cscott (2)
  • Samoyed (2)
  • atgo (2)
  • Thehelpfulone (2)
  • parent5446 (1)
  • Emufarmers (1)
  • Tim-away (0)


Generated by MeetBot 0.1.4 (http://wiki.debian.org/MeetBot)

Full log[edit]

Meeting logs


21:04:21 <sumanah> #startmeeting RfC catchup | Channel is logged and publicly posted (DO NOT REMOVE THIS NOTE). | https://meta.wikimedia.org/wiki/IRC_office_hours | Logs: http://bots.wmflabs.org/~wm-bot/logs/%23wikimedia-office/
21:04:21 <wm-labs-meetbot> Meeting started Wed Aug 13 21:04:21 2014 UTC and is due to finish in 60 minutes.  The chair is sumanah. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:04:21 <wm-labs-meetbot> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:04:21 <wm-labs-meetbot> The meeting name has been set to 'rfc_catchup___channel_is_logged_and_publicly_posted__do_not_remove_this_note_____https___meta_wikimedia_org_wiki_irc_office_hours___logs__http___bots_wmflabs_org__wm_bot_logs__23wikimedia_office_'
21:04:21 <awight> she's back!
21:04:33 <sumanah> #chair brion Tim-away mark sumanah
21:04:33 <wm-labs-meetbot> Current chairs: Tim-away brion mark sumanah
21:04:41 <sumanah> How was Wikimania? :)
21:04:53 * sumanah was not there
21:04:55 <brion> good fun :) still in UK
21:05:12 <sumanah> 2 topics today - basically, 1) anything from Wikimania folks need to know about?
21:05:14 <sumanah> point to notes, etc
21:05:23 <sumanah> and 2) some followup/next steps from previous meetings
21:05:30 <sumanah> #link https://www.mediawiki.org/wiki/Architecture_meetings
21:05:34 <sumanah> #topic Wikimania
21:05:58 <sumanah> so, brion aude awjr cscott greg-g and all the rest of y'all who were at Wikimania
21:06:03 <sumanah> any architectural or design decisions?
21:06:10 * awjr waves
21:06:14 <awjr> i wasn't actually at wikimania :)
21:06:20 <awjr> im just lurking
21:06:56 <brion> no huge decisions yet. were talking about expanding/delegating from the arch committee but nothing's finalized yet
21:07:32 <brion> but that should help with some more rfc work later :)
21:07:53 <superm401> Is the arch committee the same as the three Architects?
21:08:00 <sumanah> hahahaha
21:08:00 <sumanah> TrevorParscal: James_F I think you were there, right?
21:08:00 <sumanah> any decisions that need reporting back that haven't already made it to wikitech-l?
21:08:10 <brion> for the moment yes its the three of us
21:08:20 <sumanah> but soon: moar!
21:08:30 <brion> :)
21:09:51 <brion> Ill make sure we get a few things summarized I have notes to go back over
21:10:57 <brion> but nothing earth shattering. the main controversies right now are outside arch committees purvue :P
21:12:23 <ghostsumana> #info there's discussion of how to expand/delegate from the arch committee but nothing's finalized yet
21:12:40 <ghostsumana> brion: while I fuss with IRC will you switch #topic to RfC followup?
21:12:47 <parent5446> whoah, what is ghostsumana?
21:12:54 <ghostsumana> Whooooo I'm a ghost!
21:13:11 <brion> #topic RfC followuo
21:13:16 <ghostsumana> #link https://www.mediawiki.org/wiki/Architecture_meetings/RFC_review_2014-08-13
21:13:21 <Emufarmers> she has come to coordinate this meeting from beyoooooond the graaaaaaaaaaaaaave
21:13:26 <brion> woopa I'm on my phone and can't spell
21:13:31 <sumanah> hahahaha
21:13:37 <sumanah> it's fine
21:13:46 <ghostsumana> IT'S NOT FINE. I AM A GHOST.
21:13:55 * sumanah is laughing too hard
21:13:55 <brion> :)
21:13:57 <sumanah> ok!
21:13:59 <sumanah> AndyRussG:
21:14:09 <sumanah> CentralNotice proxy! fundraising! thing! who is the point person now?
21:14:19 <sumanah> https://www.mediawiki.org/wiki/Requests_for_comment/CentralNotice_Caching_Overhaul_-_Frontend_Proxy
21:14:27 <sumanah> Is it awight?
21:14:32 * awight points to AndyRussG
21:14:39 <AndyRussG> I think it's me :)
21:14:40 <awight> I can be a loyal "yes" person though
21:14:48 <brion> excellent
21:15:18 <AndyRussG> Things always work out best if there's a disloyal "no" person and a loyal "yes" person
21:15:26 <sumanah> #agreed that AndyRussG is the new CentralNotice Caching Overhaul - Frontend Proxy point person, and awight helps out friendly-like
21:15:29 <sumanah> ok, moving on!
21:15:34 <sumanah> Followup from 23 July: have we done security update planning re Composer managed libraries for use on WMF cluster?
21:15:35 <AndyRussG> Quick question first?
21:15:42 <sumanah> oh go ahead AndyRussG
21:15:43 <AndyRussG> Or questions?
21:16:01 <ghostsumana> (these are the #links I forged in life)
21:16:14 <brion> heh
21:16:17 <AndyRussG> 1) Timeline needed/expected: how soon does everyone need aciton on the CentralNotice RfC?
21:16:56 <sumanah> #info let's look at https://www.mediawiki.org/wiki/Architecture_meetings/RFC_review_2014-07-30 for a bit of context
21:17:02 <AndyRussG> And 2) just to confirm that ESI is still a ways off?
21:17:11 <mark> ESI is still a fair bit out
21:17:30 <sumanah> Jeff_Green: ^
21:17:33 <awight> ... and whether leaning on the ESI issue might accelerate things :p
21:17:34 <sumanah> (just for heads-up)
21:18:30 <brion> I'm going to lose signal be right back
21:18:36 <sumanah> ok
21:18:59 <sumanah> AndyRussG: I think your product manager-type person is the one to talk with about schedules. Anne?
21:18:59 <awight> mark: is the blocker for ESI the Varnish 3.* upgrade, or something else?
21:19:04 <awight> atgo_: ^^
21:19:09 <mark> at least the varnish 4 upgrade
21:19:14 <mark> and who knows if it's reliable then
21:19:46 <awight> ok we can chat offline, that's news at least!
21:20:12 <atgo> hey all. getting caught up here - major computer/IRC issues today
21:21:26 <sumanah> ok atgo - AndyRussG asked: "Timeline needed/expected: how soon does everyone need aciton on the CentralNotice RfC?"
21:21:41 <AndyRussG> K I guess we could look into timeline for the CentralNotice RfC in more detail later, and get back to everyone maybe via Wikitech
21:21:55 <sumanah> ok!
21:22:03 <atgo> sounds like a good plan AndyRussG
21:22:08 <sumanah> csteipp: Followup from 23 July: have we done security update planning re Composer managed libraries for use on WMF cluster?
21:22:09 <AndyRussG> :)
21:22:14 <sumanah> bd808|BUFFER:
21:22:30 <sumanah> also "bd808 to get repo renamed from mw/core/vendor to mw/vendor" - happened?
21:22:33 <csteipp> Not really?
21:23:09 <csteipp> I know bd808|BUFFER is sponsoring the monolog bit, so he's the person to watch for updates, and will delegate someone else if when he's no longer able to
21:23:30 <csteipp> But the process hasn't been documented anywhere that I'm aware of
21:23:42 <Thehelpfulone> oh mark whilst you're here YGM from me about a mailman upgrade
21:23:46 <sumanah> "Look into https://security.sensiolabs.org/check and https://github.com/sensiolabs/security-advisories for vulnerability tracking" - can we assign this to Bryan as well?
21:23:55 <mark> Thehelpfulone: yes, sorry, i'll respond tomorrow
21:24:13 <Thehelpfulone> thanks
21:24:46 <csteipp> Right, I was on vacation for that meeting... no wonder it didn't look familiar. That sensiolabs thing looks interesting.
21:25:41 <csteipp> I'll check with Bryan when he gets back and see if he's set it up anywhere.
21:26:05 <sumanah> #action csteipp to check with Bryan re Composer managed libraries for use on WMF cluster security planning
21:26:23 <sumanah> ok, next Followup from 21 May: has cscott run the requested stats on Square bounding boxes?
21:26:32 <sumanah> cscott: https://www.mediawiki.org/wiki/Architecture_meetings/RFC_review_2014-05-21 and https://www.mediawiki.org/wiki/Requests_for_comment/Square_bounding_boxes
21:27:26 <sumanah> I know it's past the end of the working day on the East Coast really
21:27:30 <sumanah> so I'll move on
21:27:38 <mark> or europe :p
21:27:41 <sumanah> brion: are you pingable?
21:27:42 <sumanah> certainly!
21:27:44 <brion> yo
21:27:49 <sumanah> brion: Followup from 16 July: does Brion know the next steps re CSSJanus on vertical writing support?
21:27:58 <sumanah> https://www.mediawiki.org/wiki/Architecture_meetings/RFC_review_2014-07-16 and https://www.mediawiki.org/wiki/Requests_for_comment/Vertical_writing_support
21:28:23 <brion> sumanah: if nobody else wants to take it it’s still on my research todo list, but i have to follow up on it further
21:29:50 <sumanah> brion: well it was assigned to you which is why I think people thought it was your task
21:29:57 <brion> yes :)
21:29:59 <sumanah> :)
21:30:15 <brion> just haven’t gotten to it yet, it’s still in my queue
21:30:18 <sumanah> got it
21:30:30 <sumanah> np, if the RfC authors want more activity happening, they have their own TODOs
21:31:07 <sumanah> ok, next, yurikR
21:31:08 <sumanah> Followup from 11 June: how has the bandwidth savings been on Reducing image quality for mobile? Will we change regular/original images as well as thumbnails?
21:31:14 <sumanah> https://www.mediawiki.org/wiki/Architecture_meetings/RFC_review_2014-06-11
21:31:17 <sumanah> https://www.mediawiki.org/wiki/Requests_for_comment/Reducing_image_quality_for_mobile
21:32:30 <superm401> From the notes, it seems like the idea was to have a user option for changing the quality of full-size (i.e. originals as of now) images.
21:32:42 <superm401> I assume it would still keep the originally uploaded image for a couple reasons:
21:32:47 <yurikR> sumanah, we have deployed it for a limitted number of partners, but we don't have an easy way to measure it yet. I will try to come up with some data from zero logs, and we should probably ping analytics
21:33:07 <brion> superm401: yeah it’s an alternate version of the thumbnail
21:33:24 <brion> the originals are unchanged
21:33:54 <superm401> Thanks for clarifying, brion.
21:33:56 <yurikR> i think the biggest partner, russian beeline, is on the new shrinkimg as of two weeks ago
21:34:59 <yurikR> so now we need a way to evaluate the bandwidth.  Does anyoen know if there is a way to monitor bandwidth like that?
21:35:17 <brion> yurikR: count the hits and multiply by the file size, i suppose :D
21:35:34 <sumanah> Analytics people?
21:35:42 <sumanah> milimetric: ^
21:35:51 <milimetric> hi
21:35:54 <yurikR> brion, you mean to basically analyze zero logs which have some sort of a file size
21:35:57 * milimetric reads up
21:36:12 <superm401> yurikR, I think there is now page view data on images, so if nothing else, you could same some image hits, calculate the bandwidth used, then check the equivalent without the lower-qualityversion.
21:36:36 <brion> yurikR: with luck the log entries include file size. without luck, you have to look up the files to get the number
21:36:57 <superm401> brion, to be clear I was talking about, "maybe someday we could also change original/regular images, not just thumbnails, if the user wants to save bandwidth there, too, but let's see how thumbs go first." from https://www.mediawiki.org/wiki/Architecture_meetings/RFC_review_2014-06-11
21:37:02 <yurikR> brion, yes, that's what i was planning to do, ignoring all the 304s. superm401 any links?
21:37:28 <superm401> https://www.mediawiki.org/wiki/Hit_stats_aggregation#Image_views ?
21:37:32 <superm401> Just found that.
21:37:35 <milimetric> yurikR: lemme check real quick, I know the hadoop logs have the response size, but not sure about the other logs
21:37:52 <brion> superm401: ah i think that’s for when you zoom in rather than modifying the original file
21:38:05 <ori> just curl the urls and measure the file size that way
21:38:06 <yurikR> milimetric, i thought we didn't have a functioning hadoop
21:38:08 <ori> if you do it for a sample
21:38:11 <ori> that's fine
21:38:30 <superm401> Yeah, I assume the same would go (only affects display, original is unaffected), but someone should correct me if there has been discussion of lossy changes to the actual original.
21:38:31 <milimetric> yurikR: not for public consumption yet, but it's there to poke around
21:39:57 <yurikR> milimetric, will it have X-Analytics field?
21:41:28 <sumanah> ok, yurikR do you have the answers you need?
21:41:30 <sumanah> any action items?
21:42:07 <yurikR> sumanah, just checked in the zero logs, it seems multimedia is not being marked with X-Analytics (zero) tags
21:42:10 <milimetric> yurikR: let's talk further in -analytics if you want
21:42:18 <yurikR> milimetric, ok, thx
21:42:38 <sumanah> that was the last question I had, but brion mark and anyone else, if you have more RfC questions please speak up
21:42:44 <sumanah> otherwise we can wrap up
21:42:52 <brion> i’m good for now
21:42:58 <sumanah> cscott: ping in case you have Square bounding box thoughts :)
21:43:15 * sumanah looks at https://www.mediawiki.org/wiki/Requests_for_comment/Standardized_thumbnails_sizes
21:43:37 <yurikR> brion, will be happy about https://www.mediawiki.org/wiki/Extension:Graph
21:43:49 <cscott> sorry, i'm overwhelmed with PDF stuff, haven't had any brain cycles to spare for fixing images
21:43:51 <brion> \o/
21:43:54 <yurikR> almost ready to deply ;)
21:44:00 <yurikR> pending sec review
21:44:07 <cscott> also, i don't have the appetite for breaking the world at the moment. ;)  enough turmoil already atm.
21:44:17 <sumanah> gi11es and Ops ( mark ?) let me update https://www.mediawiki.org/wiki/Requests_for_comment/Standardized_thumbnails_sizes with what they had said
21:44:42 <sumanah> #info square bounding boxes still delayed and waiting on cscott's time and community appetite for change :)
21:45:35 <sumanah> got it cscott
21:46:05 <sumanah> mark: re https://www.mediawiki.org/wiki/Talk:Requests_for_comment/Standardized_thumbnails_sizes did you end up organizing the sprint you suggested?
21:46:21 <sumanah> "As you know there�s a fair amount of desire to change the way thumb handling & storage works, support more variants (sizes, quality), etc by various teams. I�m starting to think it would be good for us to organise a sprint on this very topic, get the multimedia team, relevant Ops people and some other interested developers from other teams and really dive into these problems."
21:46:27 <sumanah> (after this I'm gonna close it up)
21:46:30 <sumanah> (the meeting I mean)
21:46:47 <sumanah> ghostsumana I think your time wandering this earth is coming to an end
21:47:08 <ghostsumana> I only came here to fulfill the need to make this meeting happen; now I am at peace
21:47:15 <Samoyed> wait what's happening
21:47:18 <Samoyed> what just happened
21:47:23 <brion> :)
21:47:45 <sumanah> Samoyed: (My IRC connection broke for a few min so I opened up a webchat thing as "ghostsumana" to kick off the rfc review)
21:48:01 <sumanah> harej: ^
21:48:16 <sumanah> bringin' a little Tumblr spirit to these meetings
21:49:19 <sumanah> ok, so I'm gonna read that as a no, no sprint planned yet
21:49:55 <sumanah> thanks everyone!
21:49:59 <sumanah> #endmeeting