Completely agree that caret position shouldn't be live or backed by a live range, because of the perf concerns you called out. I think the spec should be updated to explicitly mention that.

Yeah, I think the current definition is broken as someone could mutate the range and that would impact the result, which isn't what you want to happen.

@smaug---- @zcorpan Any thoughts?

Firefox is the only browser that ships this API at the moment and even Firefox doesn't appear to use a live range in its implementation: https://searchfox.org/mozilla-central/source/dom/base/nsDOMCaretPosition.h#27. So I think we should update the spec, remove the "caret range" concept and explicitly mention that CaretPosition shouldn't be backed by a live range under the hood.

yeah, the spec doesn't really make sense. Gecko does internally create a temporary Range just to call getBoundingClientRect and after that the Range is released, so it won't observe mutations to DOM.

I think this stuff comes initially from 532723f

I don't see that being discussed in
https://www.w3.org/Bugs/Public/buglist.cgi?product=CSS&component=CSSOM%20View&resolution=---
So perhaps it came from the mailing list.

Here https://lists.w3.org/Archives/Public/www-style/2009Nov/0244.html
I didn't try to find the relevant TPAC meeting minutes.

Oh, and looks like I complained about the range already 13 years ago :) https://lists.w3.org/Archives/Public/www-style/2011May/0051.html

Thanks for the input. It seems that @smaug---- from Mozilla, @annevk from Apple, and @sanketj and myself from Microsoft all agree that CaretPosition should NOT be backed by a live range.

@astearns Is it possible to get to resolution of this issue without bringing up to CSSWG meeting? Thanks.

cc'ing Googler @mfreed7.

cc'ing Googler @mfreed7.

LGTM. Incidentally, I'm also in favor of going back in time and making every range a static range, but that's a tougher lift.

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: Remove caret range concept (and live ranges) from CaretPosition interface.

RESOLVED: Remove caret range concept (and live ranges) from CaretPosition interface.