[CSS-UI] caret-shape: block/underscore and zero width space

[ziransun]

The spec for caret-shape says for block caret shape "The UA must render the insertion caret as a rectangle overlapping the next visible character following the insertion point".

The question is, what we do with the case of zero-width space? In Chromium, by default (bar shape), it displays a bar caret. I'd think that for caret-shape: block and caret-shape: underscore, we probably should display the caret with the width of 1px ( the width of bar shape), which is reflected in this CL .

Please let me know what you think. And if this proposal is reasonable, I wonder if the spec needs a bit update? Thanks!

[ziransun]

@frivoal @schenney-chromium

[ziransun]

Just to clarify, for block shape, it would be 1px width and height stays the same as that of the block. For underscore, the width is 1px too, and the height stays the same as that of the underscore (normally 1px too).

[frivoal]

In the case of a character that is non visible, the spec says that the width should be 1ch. I think that's reasonable when at the end of the line / string, but this may be less than ideal when in the middle of a line, where there are other characters (with non-zero advance) after the zero-width one. In that case, 1px might be preferable.

Agenda+ to propose adopting this refinement.

[jfkthame]

I don't think a 1px underscore caret is a good idea -- it'll be virtually invisible.

In the case of a character that is non visible, the spec says that the width should be 1ch

My reading of the spec is that if there is a non-visible (zero-width) character or characters following the insertion point, the caret ignores these and is drawn with the width of the next visible character:

The width of the block and underscore carets should be the advance measure of the next visible character after the insertion point

I think that's reasonable. The 1ch fallback only comes into effect "if there is no next visible character".

I suppose letting the block caret become 1px if the following character is invisible would be OK (though I'm not convinced it's an improvement), but we should not do it for underscore, at least.

[ziransun]

Looking up the spec for caret-shape , I found the following statements

• for block and underscore as " If there is no visible character after the insertion point, the UA must render the caret after the last visible character. "
• Later, it says that "The width of the block and underscore ..., or 1ch if there is no next visible character or if this information is impractical to determine."

The zero space character in the middle of a line is a bit tricky. Somehow, I feel that it might makes sense that we follow the same "width" rule (either 1px or 1ch or the width of the previous visible character) for both block and underscore...

It would be very helpful to move this discussion forward so I can update the implementation accordingly. Thanks!

[ziransun]

JFYI, I updated Chromium implementation at https://chromium-review.googlesource.com/c/chromium/src/+/6973387. Basically, this CL follows "the width of the block and underscore carets should be 1ch if there is no next visible character or if this information or is impractical to determine."

I'd be very happy to update implementations to address any updates on the specs. Thank you!