Minutes (was: Re: WebPerf (triage) group call - August 20th 11am PST)

Available at:
https://www.w3.org/2018/08/20-webperf-minutes.html

Also in plain text as:

WebPerf Group Call
20 Aug 2018
Attendees
Present
yoav, igrigorik, todd, Philip, Charles, Nic, Nicolas, ryosuke
Regrets
Chair
yoav
Scribe
igrigorik
Contents
Topics
TPAC
Distributed Tracing group comments
next call
page visibility
resource timing
Summary of Action Items
Summary of Resolutions
TPAC
yoav: TPAC, we're meeting thursday and friday
... do we want to follow similar setup as last TPAC, with separation for 
design vs triage days?

todd: sg, main hurdle is scheduling vs other groups

who's coming? Philip, Charles, Nic, Nicolas, Todd, Yoav

<scribe> ACTION: chairs to start planning doc for TPAC: who's coming, 
agenda ideas

Distributed Tracing group comments
Charlie: Distributed Tracing group contacted us way back

<scribe> ACTION: Yoav to followup with DT group

next call
next design call: Sept 6th @ 11AM PST

page visibility
<yoav> https://github.com/w3c/page-visibility/pull/38

<scribe> ACTION: Yoav & Todd to review ^

hidden when fully obscured: 
https://github.com/w3c/page-visibility/pull/38

<scribe> ACTION: Todd to review tomorrow PV pull 38

Next: support waitUntil semantics in pagevisibility?

Philip: we can close it,

Yoav: sgtm, no further action needed

<yoav_> https://github.com/w3c/page-visibility/issues/40

Next: Provide a clear way to hook into "document becomes visible/hidden"

https://github.com/w3c/page-visibility/issues/40

yoav: it's an editorial issue, unclear if there are upstream use cases

todd: none of the currently open issues block us from CR
... re #34, I propose we move that to L3 because we don't have an 
implementation that conforms to this

ig: sgtm

<yoav_> https://github.com/w3c/page-visibility/issues/39

Philip: the order is defined in HTML spec

meta: should Pagelifecyle API assume L3 features?

todd: we should complete L2 PV;

yoav: ditto
... for L3 we should chat w/ Page Lifecycle editors on mering

ryosuke: Safari does implement #34 when fully obscured

todd: what do we need to clear up in L2
... current language does not talk about obscured use case
... the problem with current language is that it's a must
... it's easy to fix, I'll tackle it tomorrow

Hooks (#40): let's investigate for L2

todd: I'll take a run at this one

Re, https://github.com/w3c/page-visibility/issues/39:

todd: this is not in PV, we should move this to HTML repo

philip: yes, will do

resource timing
<yoav_> https://github.com/w3c/resource-timing/issues/70

ig: if Todd resolves the two issues, we should be able to republish PV

yoav: URL is currently resolved in SW
... s/resolved/exposed
... what is our stance on this wrt exposing timing information?
... SW today gets the fetch event for the resource, which allows it to 
observe start and end

rniwa: given TAO, this seems fine

Yoav: I'll followup on the issue; I need to check-in w/ current plans 
Mozilla has here
... I'll summarize the discussion here on the thread

<yoav_> https://github.com/w3c/resource-timing/issues/82

<yoav_> https://github.com/w3c/resource-timing/issues/141

yoav: rniwa objected to current language for synchronous event for 
bufferfull
... in Safari it's fired asynchronously; the implementation is: when RT 
buffer is full, we schedule an event to fire, in the meantime we buffer 
new events into secondary buffer. when event does fire, we iterate 
through secondary buffer to append if there's new space

rniwa@: ^ correct

npm/yoav: this behavior does not affect PO

yoav: this behavior makes sense, I'd like to investigate integrating it 
into the spec if there's support from other vendors

nic: afaik, RT is the only place with buffering logic

todd: we haven't implemented onresourcebufferfull in Edge; we just 
raised our limit to 500
... we don't have a strong opinion on this one

<scribe> ACTION: Yoav to propose draft language, so we can sanity check 
w/ implementers

<yoav_> https://github.com/w3c/resource-timing/issues/96

yoav: this is a bug in processing model

rniwa: this behavior is what triggered the issues above

todd/yoav: we can resolve this in same patch

--- thanks all! ---

Summary of Action Items
[NEW] ACTION: chairs to start planning doc for TPAC: who's coming, 
agenda ideas
[NEW] ACTION: Todd to review tomorrow PV pull 38
[NEW] ACTION: Yoav & Todd to review ^
[NEW] ACTION: Yoav to followup with DT group
[NEW] ACTION: Yoav to propose draft language, so we can sanity check w/ 
implementers

Summary of Resolutions
[End of minutes]


On 2018-08-09 16:12, Yoav Weiss wrote:
> Hey all!
> 
> Setting a group call for Monday, August 20th.
> Hangout:
> https://hangouts.google.com/hangouts/_/chromium.org/webperf-wg
> 
> Current tentative agenda:
> 
> 	* Page Visibility
> 
> 	* clarify UA may be hidden when fully obscured [1]
> 	* support waitUntil semantics in pagevisibility? [2]
> 	* Provide a clear way to hook into "document becomes visible/hidden"
> [3]
> 
> 	* Resource Timing
> 
> 	* Exposing CSS subresource URLs [4]
> 	* Clarifications on when entries are added to the buffer, surfaced to
> PO and trigger RTbufferfull events [1 [5]] [2 [6]]
> 	* setResourceTimingBufferSize behavior changes [7]
> 
> Please add any other topics you'd like to discuss to the agenda
> document [8].
> 
> See you there! :)
> Yoav
> 
> Links:
> ------
> [1] https://github.com/w3c/page-visibility/pull/38
> [2] https://github.com/w3c/page-visibility/issues/36
> [3] https://github.com/w3c/page-visibility/issues/40
> [4] https://github.com/w3c/resource-timing/issues/70
> [5] https://github.com/w3c/resource-timing/issues/82
> [6] https://github.com/w3c/resource-timing/issues/141
> [7] https://github.com/w3c/resource-timing/issues/96
> [8]
> https://docs.google.com/document/d/10dz_7QM5XCNsGeI63R864lF9gFqlqQD37B4q8Q46LMM/edit?pli=1#

Received on Tuesday, 21 August 2018 14:31:16 UTC