Input Interaction: fix buffer for the triggering of multi-select

[ellatrix]

Description

Fixes #12322.

Fixes the buffer for calculating whether, based on the text selection, multi-selection should be triggered or not. This buffer was originally added with only a collapsed selection in mind, so calculating the line height was a simple matter of taking the height of the selection rectangle.

Solution: use getComputedStyle to get the line height from the browser.

Why is this buffer needed in the first place? Depending on the browser, there may be a difference between the height of the selection rectangles and the line height. The selection rectangle may not reach all the way to the line edge. This is why we add a buffer that is the line height divided by two.

How has this been tested?

See #12322. Put the caret at the start of a paragraph that is at least 4 lines long. Press shift+arrowDown and ensure multi-block selection only starts after the last line has been selected.

Screenshots

Types of changes

Checklist:

• My code is tested.
• My code follows the WordPress code style.
• My code follows the accessibility standards.
• My code has proper inline documentation.
• I've included developer documentation if appropriate.

[ellatrix]

I'll see if I can add a test case.

[aduth]

The implementation is consistent with what I would expect to be an improvement based on the hypothesis in #12322 (comment) . The need for buffer I still find to be unfortunate, though would be happier to see an improvement than delay on it further.

The end-to-end test is great.

[oandregal]

Fixes the issue and the code is fine. Thanks for the e2e test!

[ellatrix]

Thanks for the reviews!

@aduth Yeah the buffer is unfortunate, though I see no better way to address the issue. :/

[aduth]

@aduth Yeah the buffer is unfortunate, though I see no better way to address the issue. :/

Right, if I recall correctly, the issue is that the selection bounding box doesn't align neatly with the edges of the input it's in? I remember exploring this and finding it exceedingly difficult to measure that difference accurately, even going so far as to explore possible effects of various text metrics characteristics.