Available at
https://www.w3.org/2016/03/03-webperf-minutes.html
Text version:
Web Performance Working Group Teleconference
03 Mar 2016
See also: [2]IRC log
[2] http://www.w3.org/2016/03/03-webperf-irc
Attendees
Present
Ilya, yoav, plh, todd
Regrets
Chair
Ilya
Scribe
plh
Contents
* [3]Topics
1. [4]Resource Timing
2. [5]Beacon
3. [6]Preload
4. [7]s/type/as/ attribute
5. [8]Reporting
6. [9]hr-time
7. [10]Velocity
8. [11]Next meeting
* [12]Summary of Action Items
* [13]Summary of Resolutions
__________________________________________________________
Resource Timing
[14]https://github.com/w3c/resource-timing/pull/51
[14] https://github.com/w3c/resource-timing/pull/51
<igrigorik> preview:
[15]https://rawgit.com/w3c/resource-timing/continuous-idl/index
.html
[15] https://rawgit.com/w3c/resource-timing/continuous-idl/index.html
Ilya: looked an interesting approach
... left some comments
Todd: *bodySize and nextHopProtocol doesn't seem to be
implemented
Ilya: maybe in Firefox
<igrigorik>
[16]https://bugzilla.mozilla.org/show_bug.cgi?id=1154309
[16] https://bugzilla.mozilla.org/show_bug.cgi?id=1154309
plh: stable is 44
Ilya: it's marked for 45
plh: cn we ship HRT, PT, RT, NT, and UT as a batch?
... at least a subset of it
Todd: we'd like all UAs to implement the most recent RECs, but
they'll implement whatever they can
plh: I'll make a proposal
Beacon
[17]https://github.com/w3c/beacon/pull/27#issuecomment-18740407
5
[17] https://github.com/w3c/beacon/pull/27#issuecomment-187404075
<toddreifsteck> [18]https://github.com/w3c/beacon/pull/27
[18] https://github.com/w3c/beacon/pull/27
Todd: I'm ok with fetchBeacon
Ilya: method = POST, body nil as default
... headers?
... make sense to have but still needs to think
... fetchBeacon doesn't return anything...
... return true/false
... which indicates if it was queued
Todd: yes, would be exact same. we validate method/headers/body
and return true if ok
Ilya: would developers expect a Promise, ie should we return a
resolved Promise?
Todd: added async code on unload seems like a bad idea
Ilya: should we cc Domenic to look at this?
... or Anne
... we should consider aligning with Fetch on the return value
Todd: ok
... re headers, should we keep it?
Ilya: I think it's fine
... Fetch will set some of the headers for us
... so should be optional
and we're passing it through
Todd: I'll sort it out later today
Preload
[19]https://github.com/w3c/preload/issues/32#issuecomment-19024
3646
[19] https://github.com/w3c/preload/issues/32#issuecomment-190243646
Yoav: while implementing and testing preload, ran into fonts
... crossorigin attribute got in the way
Ilya: you're trying to preload a font
... link ref=preload crossorigin href=[20]http://same-origin
... crossorigin is meaningless here
[20] http://same-origin/
Yoav: not sure about that
... let's assume that for now
Ilya: if the browser issues a 3xxx to another origin
... then the crossorigin becomes relevant
Yoav: agreed
... for non-redirect, they should match
Ilya: if the rec is to always have crossorigin attribute and
the response from same-origin
... will that match my @font-face?
Yoav: asking to put crossorigin on same-origin is weird.
Ilya: this may trigger different cookies behaviors
s/type/as/ attribute
<igrigorik> [21]https://github.com/w3c/preload/pull/59
[21] https://github.com/w3c/preload/pull/59
Ilya: I got confused while trying to fix the spec and I'd like
to restore the old logic and introduces as
Yoav: if you change from one image format to an other, should
we trigger a new request?
... because we're changing the client logic?
Ilya: ok. type only afects if it got triggers in the first
place
Yoav: if you want to trgigger a different download, you should
change something else, type won't be enough
... if it was fetched type won't matter
Ilya: checking the html spec...
Yoav: I filed some bugs around that
Ilya: html has 2 items for type
... and we're mirroring that
<igrigorik>
[22]https://html.spec.whatwg.org/#concept-link-obtain
[22] https://html.spec.whatwg.org/#concept-link-obtain
<igrigorik>
[23]https://html.spec.whatwg.org/#link-type-stylesheet
[23] https://html.spec.whatwg.org/#link-type-stylesheet
Ilya: ok, will drop the first point from the PR
... but we'll keep both type/as
Tp[oc: Reporting
Reporting
[24]http://w3c.github.io/reporting/
[24] http://w3c.github.io/reporting/
Plh: I'll set up the repo for automatic bikeshed generation
ILya: new CSP spec is using it
... I'd like to push out the first draft to get feedback
Todd: it's a subset of NEL
plh: are ok to adop and publish a FPWD?
Todd: looks fine
Yoav: fine
RESOLUTION: publish as FPWD
hr-time
Todd: when you're up in Seattle and be around a whiteboard
Ilya: ok
Velocity
Ilya: we are currently scheduled for second day for keynote
... looking to reschedule on first day
... full day meetup thing seems unlikely
... option: move keynote on first day and do a BOF in the
evening
... option 2: during lunch hour, they have office hours. we
could do office hours after our keynote
... but setup isn't the best
... option 3: find a meeting room but we would compete with
other scheduled meetings
yoav: 1 & 3 sounds reasonnable. 1 sounds the best
... 2 doesn't sound productive
Todd: +1 to Yoav
Plh: ditto
ok, preference is 1, 3, 2
Next meeting
March 23rd, 10am PT?
plh to check
Summary of Action Items
Summary of Resolutions
1. [25]publish as FPWD
[End of minutes]
