refactor(ext/web): Don't rely on NaN comparisons in TextEncoderStream

[andreubotella]

In the transform function to TextEncoderStream's internal TransformStream, if chunk is the empty string and this.#pendingHighSurrogate is null, then lastCodeUnit will be NaN. As it turns out, this does not cause a bug because the comparison to check for lone surrogates turns out to be false for NaN, but to rely on it makes the code brittle.

[bartlomieju]

Is there a test that could be used to check this change?

[andreubotella]

The NaN comparison wasn't actually causing any difference in behavior from what would be expected, since the comparison to check whether chunk ends with a lead surrogate (0xD800 <= lastCodeUnit && lastCodeUnit <= 0xDBFF) turns out to be false if lastCodeUnit is NaN, which is what you'd expect if chunk is empty. As the commit message says, this refactoring removes the NaN comparison to make the code less brittle.

Returning early if chunk is the empty string but this.#pendingHighSurrogate is not null is also not a behavior difference, since this.#pendingHighSurrogate is either null or a string containing a single lead surrogate code unit, so that code unit would be prepended to chunk and then removed and stored into this.#pendingHighSurrogate again.

[bartlomieju]

Thanks Andreu