TOPIC: WebApps WG TPAC 19 Sep

rrsagent, set logs world-visible

Meeting structure w3c/webappswg #10

Agenda: w3c/webappswg #10

meeting: WebApps WG TPAC Meeting

present+

present+

present+

present+

present+

present+

present+ Mustaq Ahmed

present+

present+

present+ Olli_Pettay

`present+ Léonie (tink)

present+ NavidZ

present+ Mike_Wasserman (Google)

Present+ Laszlo_Gombos

present+

present+

present+

present+

present +

present+ Oriol Brufau (Igalia)

present+ Kangchan Lee (ETRI)

present-

present+ Daniel Murphy (Google)

Code of conduct: w3.org/Consortium/cepc/cepc-20141022/Overview.html

present

present

present+

+garykac

present+

code of conduct: w3.org/Consortium/cepc/cepc-20151007

webapps charter: w3.org/2019/05/webapps-charter.html

Mek: not much progress, but we added new promisy methods. Don't expect to see many more changes but we need cleanup

present+

Mek: some mostly need to talk to annevk

present+

present+

present+

present+

present+

MC: anything blocking?

Mek: just time

Let's move onto Gamepad API

steve: we are trying to get to CR. We have a privacy issue and a proposal.

steve: we have some tests in WPT.

present+

MC: it would good to see how much test coverage we have .

present+

Indexed Database API

present+

IndexDB

dmurph: three major changes

dmurph: 1) IDBFactory databases

2) IDBTransaction commit() -- explicit transaction commit

dmurph: 3) IDBCursor request attribute (?)

dmurph: Better controls over transaction lifetime and status

present+

dmurph: Maybe not enough support to move to CR due to lack of adoption

dmurph: Need to chat with Mozilla people

dmurph: Seeking opinions from Apple; Brady/Apple is the person, but not present

IntersectionObserver tl;dr -- basically ready for CR, implemented in all major browsers

szager evangelizing for IntersectionObserver V2 (occlusion detection)

PointerLock

Issues due to OS display scaling affecting cursor location

There's a path forward on Windows, looking for Apple people to collaborate

Looking for usability feedback prior to CR

present+

Push API

Concern about testability

Improvements in interop over last 12 months

Editor(s) unsure whether it's worthwhile to move to CR

API may undergo additional churn, so CR may be premature

Screen Orientation API

Spec overhauled over last 12 months

Added lots of WPT

Testing on mobile is tedious

Fun issues around lock() reentry

Still lots of work to do

Seeking input from WebKIt

Web Manifest spec

Limited progress; wanting for feedback from browsers

present+ Dan Appelquist

Testing the feature is hard

Shout out to Maximiliano Firtman (@firt)

Just got into Firefox preview, waiting for issues/feedback

concerns around BeforeInstallPrompt()

aarongu: Want to add "shortcuts" to manifest, to enable long-press type shortcuts for install

aarongu: Initial feedback is positive, more discussion necessary

shortcuts = context menu like capability

Web Share API

shortcuts background/discussion w3c/manifest #582

This is an example of the standards process working well

shipped in chromium and webkit, being implemented in firefox

CR process should be smooth sailing

Do we want to support sharing files? Security concerns.

How to do feature detection for level-2 features?

Here come the editing specs..

UIEvents

present+

RRSAgent, make logs public

Long crufty history of this spec

tantek: Need significant rewrites of old spec prior to CR

s/tantek/???

Anybody: who is the speaker?

garykac: This is grungy unglamorous work, but the current state is limiting other API's

Clipboard API

hello

present+

garykac: implemented text and image support, spec work to follow

garykac: Long list of desired features

garykac: Async text clipboard API was first step; step two adds image support

garykac: Seeking feedback

present+

present+

jsbell will talk about IndexDB cursor requests

jsbell: cursor->request bookkeeping was hard, now it's easy

jsbell: \o/

Input Events API

Johannes Wil,

Wilm

jw

Johannes Will: Two levels of API

present+

Level 1 implemented in chromium and WK

Level 2 only in @K

WK

present+

Seeking feedback from chromium

present+

Input Events and Contenteditable -> w3c/editing #199

marcosc_: Maybe time to revisit two-implementor policy for spec progress

Johanness Will (JW): Level 2 was original proposal, but chromium discovered problems and backed off original proposal, leading to schism

JW: Onus on chromium to suggest path forward

Selection API

Postpone Selection API pending rniwa

ContentEditable

JW: All browsers have legacy implementations without formal spec

JW: Intention is to revisit after Input Events API

JW: execCommand is buggiest part

IDBCursor request video: youtube.com/watch?v=y1TEPQz02iU

jsbell: Please subscribe to my YT channel!

Pthbbbbt

Accessibility specs...

Editors not present, can anyone speak?

Leonie Watson (LW): Lack of tests is primary impediment

present+ Victor Costan (Google)

LW: Extracting old accessibility from HTML5 spec created lack of implementors

LW: There are active participants from Apple and Mozilla

q+ to ask about the things that have come over from WICG?

RRSAgent, make minutes

RRSAgent, make logs world-visible

Editing Meeting -> w3c/editing #198

I'd like to discuss text range references with anyone interested, e.g. indieweb.org/fragmentions

aarongu here's one pointer w3c/payment-method-manifest #27

present+

present+ Staphany Park (Google)

present+ Victor Costan (Google)

can people announce what and when they are going to talk about in the web apps room in this IRC channel?

Going to try and talk about IndexedDB in the next few minutes

topic: Image Resource

marcos: Things to be careful of: purpose, things that are weird in different contexts, overly specific

webmanifest we're going to convert into info types

we can redefine the whole manifest spec in terms of that.

aaron: Does that make it so that image resource would be extensible within webappmanifest

marcos: yes, where possible we'd try and share

rayan: We're using the image resource in background fetch and content indexing

beverloo: Matt just raised an issue on the image spec adding 50 new purposes

marcos: Other things to keep in mind are that it isn't just the members, its the processing model: needs to be generalized correctly.

Would the maskable stuff stay in here?

rayan: Maskable is very specific to manifest

aaron: Usable in the context of the icon and shortcut icon

beverloo: About to ship on Android for WebAPKs

rayan: platform and purpose are very manifest specific.

marcos: Probably need to talk to jake

aaron: If we migrate it to a new repo, do we migrate issues?

marcos: yes.

aaron: Wanted to bring up color scheme: Image is for a specific theme

??: 3D image resources

marcos: question in my mind is is it just an image type?

??: Want a new mimetype

s/??/hernicus

aaron: Favicon is separate from this spec

user agent can determine if it wants to use it

hernicus: Only new thing then would be size (we have 3 dimensions)

marcos: If there are 2 x's then we could assume it's a 3D size.

marcos: Probably needs to go through the html spec

we will pick it up automatically afterwards.

aaron: Good in our book.

topic: Color Schemes

aaron: Different icons for different color schemes

maybe works with highncontrast?

some colors don't look good in darkmode

the vscode nightly icon looks bad in night mode --> Nice to have options

raymes: What is exposed to the website

aaron: The say in the manifest what scheme the icon is for (by default all)

beverloo: I like it. It's probably hard to support on Android, but I'm for it.

aaron: Wouldn't dynamically switch on Android

beverloo: This is tricky, we can maybe do something about it?

might be hard on Firefox, because you use shortcuts, which can't be updated.

beverloo: I'm a bit concerned, we're worried about the default case working well

Devs will have to test, probably not top of mind

too easy to just set it to light mode and forget.

aaron: No preference would be the default, so it would probably work okayish

we should copy what the color scheme folks are doing.

leaves it open to the OS to convert the icon (if that's the kind of thing it wants to do)

marcos: Who wants to do the work

aaron: I'll do it.

I wish agendas wouldn't link to non-open Google docs

garykac: is raw clipboard stuff in github.com/w3c/editing/blob/gh-pages/TPAC/2019/Agenda.md perhaps something you have access to?

smaug: I pinged the author to make it publi

thanks

smaug: should be fixed