W3C

– DRAFT –
Web Editing Working Group 22-04

14 April 2022

Attendees

Present
BoCupp, comandeer, johanneswilm
Regrets
-
Chair
johanneswilm
Scribe
johanneswilm

Meeting minutes

set logs world-visible

Zakim: this will be the Web Editing Group 22-04 meeting

<whsieh> if we have time today, should we discuss how we should proceed with https://github.com/web-platform-tests/interop-2022/issues/17?

ok, let'

s add that to the end then.

<whsieh> 👍🏻

https://github.com/w3c/clipboard-apis/issues/170

https://github.com/w3c/clipboard-apis/issues/170

Bo: we are in favor of something like that. We ca see how a custom format can turn into wellknown format over time. thumbs up.

whsieh: and that is synchronous?

BoCupp: yes, just a hardcode list

whsieh: we are in support for it as well

BoCupp: is clipboarditem the right place?

whsieh: only alternative method is to have method on clipboard and then have mwethod on it handing it an item.

BoCupp: I favor clipboarditem

RESOLUTION: add synchronous method on clipboarditem to query supported wellknown formats and whether web custom formats are supports.

https://github.com/w3c/clipboard-apis/pull/164

BoCupp: Our objection is that removal of delegation mechanism favors one particular user interface like in safari ... firefox plans to also support similar UI. we don't.
… Problems showing up when one uses an iframe and doesn't do particular type of UI.

BoCupp: Userexperience should be an area where browsers can differ. SO we would be against applying PR removing this.

whsieh: So is permissions prompt currently required?

BoCupp: No, this is when you do have permissions prompt and iframes. That's when you need a delegation model. Can we leave it in the spec. We'll need it even if it isn't in there.

johanneswilm: so having this in spec will not affect firefox/safari?

BoCupp: correct.

BoCupp: we have not really discussed with PR stakeholders.

ACTION: BoCupp will communicate with authors of PR.

https://github.com/w3c/clipboard-apis/issues/75

johanneswilm: user gestures are problematic for remote desktop clients.

anupam: We discussed async. Issue should be resolved.

whsieh: how specifically should we relax user gesture requirement?

BoCupp: is timing only applying to write?

whsieh: yes

BoCupp: so how about tab not being active anymore, or a second write being initiated....

whsieh: Problem is that we don't even know if something should be written to clipboard. So we just extend the time.

BoCupp: Problem is that it may take longer than 5 seconds to receive an answer from server and knwo whether something should eb added to clipboard.

BoCupp: So we could extend time limit.

BoCupp: Firefox uses 5 seconds.

whsieh: We use 1 second currently settimeout and fetch to propagatae user gestures. This is what we we wanted to use universally.

ACTION: whsieh will check user activation spec and come up with a proposal for a number of seconds.

BoCupp: no amount of time will solve the problem. 1 second seems reasonable.

https://github.com/w3c/editing/issues/393

BoCupp: We will discuss async.

https://github.com/w3c/virtual-keyboard/issues/15

BoCupp: I have seen the answer, but we have not time to look at it.

BoCupp: The question is whether the automatic lifting to doing so manually influence screenreaders?

whsieh: talked to accessibility group.

whsieh: The way we deal with having fixed element at bottom is that we allow scrolling so that one have the keyboard below the lowest element on the page.

BoCupp: I understand the concern, but we need to experiment.

BoCupp: cannot commit someone to work on it right now. Cannot [promise answer in a month.

https://github.com/web-platform-tests/interop-2022/issues/17

<whsieh> https://github.com/w3c/editing/issues/396

https://github.com/w3c/editing/issues/396

whsieh: Jen Simmons has been coordinating interop 2022 from our side

whsieh: pat of that is making web platform tests universal.

whsieh: many failures in editing is that execCommand behaves differently in different browsers.

whsieh: so what do we do? we historically tried to make them consistent with native UI behavior.

BoCupp: This is not the first time this comes up. We have not tried to align similar behavior.

BoCupp: Becomes large editors don't use DOM. Even editors using DOM use their own commands (even if they use HTML and not their own document model).

BoCupp: It's not very extensible.

BoCupp: There is not much of a surface for web platform developers.

BoCupp: It seesm to me that unless you want to have a ribbon at the top of an editor.... it's not that usable with execCommand.

whsieh: I think we all agree that execCommand goes the way of the dinosaurs.

whsieh: The question is what do we do with the tests? Do we just discard them?

whsieh: The problem is there are still web pages that use them?

BoCupp: Are there some tests that do pass everywhere that we can use as anti-regression tests?

BoCupp: Are there areas where extra tests would make sense?

<whsieh> whsieh: yeah, standardizing selection behavior is at least as hard as execCommand

<snianu> Blink and Webkit have very similar behavior last I checked, but FF differs a lot in the editing commands. me says:I think selection behavior is a much harder problem as there are commands like Word movement and stuff(that depends on the ICU lib) that affects selection and standardizing that is not an easy task me says:and there are platform specific word break behavior as well that complicates it a bit

johanneswilm: For editors that use the native selection (caret), it would be helpful to get more standardization on how that caret behaves.

whsieh: Some of that is currently hard to test.

BoCupp: [multiple examples...] some of the differences are platform dependent.

BoCupp: We have come a long way in the area of input events and target ranges.

BoCupp: It may sometimes be difficult to say how they behave in case of eevry execComnmand.

<whsieh> I think we could all feasibly align on async clipboard API WPT as well

BoCupp: It may be a point to make sure that input text events are behaving the same. There is still an issue with the last event in IME being canceleable.

whsieh: We had some differences in the area of pickling, etc. but taking a step back some more general areas are fairly standardized. If we can tests user gestures.

whsieh: We don't have gesture requirement for same-origin clipboard events, so they could just add those themselves...

BoCupp: So input events has best shot

BoCupp: agree

ACTION: BoCupp will tag implementers + johanneswilm in issue on which parts of editing make sense in interop 2022.

Summary of action items

  1. BoCupp will communicate with authors of PR.
  2. whsieh will check user activation spec and come up with a proposal for a number of seconds.
  3. BoCupp will tag implementers + johanneswilm in issue on which parts of editing make sense in interop 2022.

Summary of resolutions

  1. add synchronous method on clipboarditem to query supported wellknown formats and whether web custom formats are supports.
Minutes manually created (not a transcript), formatted by scribe.perl version 185 (Thu Dec 2 18:51:55 2021 UTC).

Diagnostics

No scribenick or scribe found. Guessed: johanneswilm

Maybe present: anupam, Bo, whsieh, Zakim