IRC log of webperf on 2015-05-27

Timestamps are in UTC.

19:00:37 [RRSAgent]
RRSAgent has joined #webperf
19:00:37 [RRSAgent]
logging to http://www.w3.org/2015/05/27-webperf-irc
19:00:39 [trackbot]
RRSAgent, make logs world
19:00:41 [trackbot]
Zakim, this will be web-per
19:00:41 [Zakim]
ok, trackbot, I see RWC_web-per()3:00PM already started
19:00:42 [trackbot]
Meeting: Web Performance Working Group Teleconference
19:00:42 [trackbot]
Date: 27 May 2015
19:00:59 [Zakim]
+Plh
19:01:09 [plh]
zakim, drop plh
19:01:09 [Zakim]
Plh is being disconnected
19:01:10 [Zakim]
-Plh
19:01:12 [Zakim]
+Plh
19:01:55 [plh]
zakim, ??p1 is Ilya
19:01:55 [Zakim]
+Ilya; got it
19:02:40 [Zakim]
+ +33.1.56.69.aaaa
19:02:57 [yoav]
Zakim, aaaa is yoav
19:02:57 [Zakim]
+yoav; got it
19:03:27 [plh]
agenda?
19:04:51 [plh]
Item 8 will be done by email
19:05:00 [plh]
zaki, drop agendum 8
19:05:04 [plh]
zakim, drop agendum 8
19:05:04 [Zakim]
agendum 8, Clarify "zero time" in Workers, dropped
19:05:14 [plh]
zakim, move to next agendum
19:05:14 [Zakim]
agendum 1. "attribute filtering" taken up [from plh]
19:06:00 [plh]
Present+ Justin
19:06:31 [plh]
Justin: using a filter object, we filter the entries. by people want to use a generic object
19:06:43 [plh]
... but it doesn't mean all properties are relevant
19:06:53 [plh]
... so we try to specifiy the property are important
19:07:09 [plh]
s/ty are/ties that are/
19:07:25 [plh]
... so people will make mistake, like passing a window object
19:07:31 [plh]
... bad things can happen
19:08:05 [plh]
... when we define a dictionary, we define a specific set of properties
19:08:34 [plh]
... we would then allow the user to use the returned array and use filter() on it
19:08:46 [plh]
... counter proposal is performance
19:09:01 [plh]
... we don't know any perf issue in IE
19:09:11 [plh]
s/proposal/argument/
19:09:54 [plh]
Todd: in perf observers, we use lazy evaluation. we don't marshall it unless it's needed
19:10:52 [plh]
Justin: WebIDL doesn't support generic property bag. we would have to define the algorithm to access the properties of the object.
19:11:15 [plh]
Ilya: the only question was interop with User Timing
19:11:40 [plh]
... some request to extend User Timing with arbitrary attributes
19:12:02 [plh]
... if we go down that path, we'll have custom attributes
19:12:17 [plh]
... would be nice for users to filter those
19:12:44 [plh]
... if we namespace them however, the user would have to filter manually
19:13:19 [plh]
Justin: we could have a filter object to be strongly type, with custom keys and values as DOMString
19:17:20 [plh]
Plh: only DOMString for values?
19:17:32 [plh]
Justin: yes, the convertion should work fine
19:18:51 [plh]
Ilya: sounds like we're ok with typed dictionary, with current defined attributes of all entry types
19:19:22 [plh]
... for the sequence, won't add it yet but we could extend with .keys/.values in the future if needed
19:19:37 [plh]
... I'll update the pull request
19:19:57 [plh]
zakim, move to next agendum
19:19:57 [Zakim]
agendum 2. "setImmediate usage" taken up [from plh]
19:20:10 [plh]
Todd: we're got telemetry for setImmediate
19:20:18 [plh]
... 4% for navigation in the past 2 weeks
19:20:33 [plh]
... higher that we expected in fact
19:20:38 [plh]
... so clearly it's used
19:21:42 [plh]
Ilya: should we take it to the list?
19:21:52 [plh]
Todd: ok, will follow up there
19:21:59 [plh]
zakim, move to next agendum
19:21:59 [Zakim]
agendum 3. "frame timing + :visited" taken up [from plh]
19:22:30 [newtron]
newtron has joined #webperf
19:25:26 [plh]
Plh: [3 path forward]
19:25:34 [plh]
Justin: people timing differences
19:25:51 [plh]
... if we filter out, they may still be able to detect
19:26:23 [plh]
Todd: 2 timing info: render and composite
19:27:34 [plh]
Justin: you can imagine forcing complex style and determine timing information
19:27:45 [plh]
.., like transition state change
19:29:07 [plh]
Yoav: so this applies to new links added to the page anyway
19:29:36 [plh]
Ilya: if you try to apply complex style to generate timing attack, we'll filter those out
19:29:52 [plh]
Justin: we can provide a list of our filtered properties
19:30:09 [plh]
... anything that change shapes will be suppressed
19:30:25 [plh]
... dotted borders, etc.
19:30:45 [plh]
... will have to go back to check
19:31:23 [plh]
Todd: would be helpful to mention that
19:31:43 [plh]
Ilya: so IE does not update the link if href changes
19:31:59 [plh]
... did it cause any dev feedback?
19:32:19 [plh]
Justin: depending on our you recycle the DOM, you'll get the update
19:33:08 [plh]
... for visited links, we do not render visited and then we go back and rerender them
19:33:39 [plh]
... we can provide more details
19:35:54 [plh]
... if you always rerender the links, you can put 10,000 of them and measure the difference
19:37:11 [plh]
Ilya: one solution was to disable link painting if frame timing is enabled, but it was ruled out because FT should desactive other features
19:39:34 [plh]
Justin: [on the use cases for :visited] it doesn't seem unreasonnable that, when the page is live, the painting based on user history is disabled
19:41:29 [plh]
.. if you click on a link, we'll set the visited flag on the click action
19:42:36 [plh]
Ilya: so, depending on the DOM update path, you take the history into account?
19:43:23 [plh]
Justin: the only thing that would update the link state would be user update interaction
19:44:00 [plh]
... we're already disabling those state in private mode
19:45:42 [plh]
Todd: we'll follow up in the github issue
19:47:21 [plh]
zakim, move to next agendum
19:47:21 [Zakim]
agendum 4. "is domLoading useful?" taken up [from plh]
19:47:51 [plh]
Todd: domLoading is internal to implementation
19:47:59 [plh]
... you don't learn much from this value
19:48:12 [plh]
... I would proposing to get rid of it
19:48:24 [plh]
... but Ilya indicated that it will cause some pain
19:48:37 [plh]
... so proposal is to update the spec with a Note
19:48:49 [plh]
... ie deprecate via spec
19:49:14 [plh]
Ilya: that's the minimum pain. doesn't solve the issue but it will be documented
19:49:38 [plh]
Yoav: I'm ok with spec update but should we add console messages as well?
19:50:01 [plh]
Ilya: might be difficult because you're getting a bag of values
19:50:53 [Zakim]
+ +1.310.310.aabb
19:51:50 [mpb]
mpb has joined #webperf
19:52:01 [plh]
Ilya: we should add the note in the spec and the UA is free to add the console warning
19:52:09 [plh]
plh: I'll propose a note
19:52:34 [plh]
Todd: in Edge, it will be the same as responseStart
19:53:00 [newtron]
newtron has joined #webperf
19:53:53 [plh]
Ilya: I'm curious how it will impact developers
19:54:02 [plh]
... we should announce the change in Edge on the list
19:54:17 [plh]
... see if people scream on the list, and revisit depending on feedback
19:54:38 [newtron_]
newtron_ has joined #webperf
19:54:46 [plh]
agenda?
19:55:14 [plh]
zakim, take up agendum 6
19:55:14 [Zakim]
agendum 6. "preconnect - shipping the non-contentious parts" taken up [from plh]
19:55:33 [plh]
Yoav: currently, we'
19:56:07 [plh]
... maybe we shouldn't block shipping preconnect for non-anonymous connection
19:56:31 [nwtn]
nwtn has joined #webperf
19:57:19 [plh]
Ilya: preconnect is a hint. as such, you're not guaranteed anything, ie not breaking contracts
19:57:34 [plh]
... my concern is creating unnecessary confusion
19:57:52 [plh]
... like if a dev uses for webfont, it won't work and folks will be confused
19:58:17 [plh]
... firefox has plan to ship preconnect without cross origin in 39
19:58:22 [newtron_work]
newtron_work has joined #webperf
19:58:28 [plh]
... I pinged them and waiting for a reply
19:58:45 [plh]
... so confusion and perception issue on my side
19:58:53 [plh]
Yoav: I think we can mitigate it
19:59:09 [plh]
... by making is clear it doesn't work for fonts
20:00:02 [newtron__]
newtron__ has joined #webperf
20:00:05 [plh]
Ilya: how long do we need to resolve the privacy link-ability question?
20:00:34 [yoav]
yoav has joined #webperf
20:00:34 [plh]
... I'll nodge Ryan and see if we can get something up
20:01:36 [newtron]
newtron has joined #webperf
20:03:25 [plh]
Ilya: for v1, we could land prefetch, and have a separate issue for preconnect
20:03:27 [newtron_]
newtron_ has joined #webperf
20:04:48 [Zakim]
-[Microsoft]
20:04:50 [Zakim]
-yoav
20:04:50 [Zakim]
- +1.310.310.aabb
20:04:51 [Zakim]
-Plh
20:04:53 [Zakim]
-Ilya
20:04:55 [Zakim]
RWC_web-per()3:00PM has ended
20:04:55 [Zakim]
Attendees were [Microsoft], Plh, Ilya, +33.1.56.69.aaaa, yoav, +1.310.310.aabb
20:04:57 [plh]
rrsagent, generate minutes
20:04:57 [RRSAgent]
I have made the request to generate http://www.w3.org/2015/05/27-webperf-minutes.html plh
20:05:16 [nwtn]
nwtn has joined #webperf
20:07:05 [newtron_work]
newtron_work has joined #webperf
20:08:47 [newtron]
newtron has joined #webperf
20:10:42 [newtron_]
newtron_ has joined #webperf
20:12:30 [newtron__]
newtron__ has joined #webperf
20:14:13 [nwtn]
nwtn has joined #webperf
20:16:01 [newtron_work]
newtron_work has joined #webperf
20:17:44 [newtron]
newtron has joined #webperf
20:19:33 [newtron_]
newtron_ has joined #webperf
20:21:21 [nwtn]
nwtn has joined #webperf
20:23:11 [newtron_work]
newtron_work has joined #webperf
20:24:55 [newtron]
newtron has joined #webperf
20:26:46 [newtron_]
newtron_ has joined #webperf
20:28:34 [nwtn]
nwtn has joined #webperf
20:30:23 [newtron_work]
newtron_work has joined #webperf
20:32:13 [newtron__]
newtron__ has joined #webperf
20:33:52 [newtron]
newtron has joined #webperf
20:35:40 [newtron_]
newtron_ has joined #webperf
20:37:30 [nwtn]
nwtn has joined #webperf
20:39:20 [newtron_work]
newtron_work has joined #webperf
20:41:12 [newtron]
newtron has joined #webperf
20:42:55 [newtron_]
newtron_ has joined #webperf
20:44:40 [nwtn]
nwtn has joined #webperf
20:46:30 [newtron_work]
newtron_work has joined #webperf
20:47:05 [newtron_work]
newtron_work has joined #webperf
21:22:06 [yoav]
yoav has joined #webperf