[cssom-view] Change `document.caretPositionFromPoint`'s `shadowRoots` parameter from rest parameter to an array

[siliu1]

The shadowRoots parameter is defined in the spec as a rest parameter of type ShadowRoot.

We should change it from rest parameter to a single array of shadow roots. Similar to https://html.spec.whatwg.org/#gethtmloptions. It'd be good to be consistent.

[siliu1]

@mfreed7 FYI I opened this issue to update the spec for document.caretPositionFromPoint to take a single array of shadow roots instead of using a rest parameter.

[siliu1]

@annevk @zcorpan @smaug---- FYI.

[annevk]

Doesn't seem unreasonable, but it might be even better to make it a dictionary argument so we can add more arguments there in the future.

I probably wouldn't go as far as https://w3ctag.github.io/design-principles/#prefer-dictionaries and also make the mandatory arguments part of the dictionary as the surrounding methods don't match that pattern and "x, y" is pretty established, but for the third argument it seems like a good idea.

[mfreed7]

@mfreed7 FYI I opened this issue to update the spec for document.caretPositionFromPoint to take a single array of shadow roots instead of using a rest parameter.

+1. That aligns with getHTML() and also allows additional arguments to be added later.

Doesn't seem unreasonable, but it might be even better to make it a dictionary argument so we can add more arguments there in the future.

+1 to that also. It makes the name of that argument explicit, which is nice, and allows additional arguments in an even better way.

By the way, I think we should make the same change to getComposedRanges() for all of the same reasons.

[siliu1]

Proposed solution: Put the array of shadow roots into a dictionary which is an argument of document.caretPositionFromPoint API.

document.caretPositionFromPoint(double x, double y, optional CaretPositionFromPointOptions options = {});

dictionary CaretPositionFromPointOptions {
  sequence<ShadowRoot> shadowRoots = [];
};

[siliu1]

@mfreed7 FYI I opened this issue to update the spec for document.caretPositionFromPoint to take a single array of shadow roots instead of using a rest parameter.

+1. That aligns with getHTML() and also allows additional arguments to be added later.

Doesn't seem unreasonable, but it might be even better to make it a dictionary argument so we can add more arguments there in the future.

+1 to that also. It makes the name of that argument explicit, which is nice, and allows additional arguments in an even better way.

By the way, I think we should make the same change to getComposedRanges() for all of the same reasons.

Filed w3c/selection-api#176

[siliu1]

Proposed solution: Put the array of shadow roots into a dictionary which is an argument of document.caretPositionFromPoint API.

document.caretPositionFromPoint(double x, double y, optional CaretPositionFromPointOptions options = {});

dictionary CaretPositionFromPointOptions {
  sequence<ShadowRoot> shadowRoots = [];
};

@astearns Can we get to resolution of this issue asynchronously with the proposed solution? Thanks!

[astearns]

The CSSWG will automatically accept this resolution one week from now if no objections are raised here. Anyone can add an emoji to this comment to express support. If you do not support this resolution, please add a new comment.

Proposed Resolution: Add an array of shadow roots to a dictionary which is an argument of the document.caretPositionFromPoint API as described in #10345 (comment)