On: DuckDuckGo \u202E

[kroltan]

It's intentional, if you inspect the `innerText` you'll see it's reversed there too:

    zero_click_wrapper.innerText.codePointAt(0)

Evaluates to 32. And if you think 32 = 0x20 could mean the next one would be 0x2E, then no, codePointAt(1) is 0x55.

[nneonneo → kroltan]

`innerText` doesn't include the RTL marker, probably due to the fact that it is supposed to reflect the "rendered" appearance of the element (i.e. deleting certain invisible characters). However, `textContent` shows the RTL marker as expected.

I'm on the side of this being an unintentional effect.

[missblit → nneonneo]

> `innerText` doesn't include the RTL marker

I'm too under the weather to dig into this, but this might be a mismatch between Firefox and the spec. I don't see in the spec [1][2] where this character could be removed since it shouldn't count as whitespace for whitespace processing.

It looks like in Chrome `innerText` contains the override. And the innerText spec is only 6 or so years old (!) so it wouldn't be too surprising if there were was a lingering incompatibility.

[1] https://html.spec.whatwg.org/multipage/dom.html#the-innertex... [2] https://drafts.csswg.org/css-text/#white-space-processing