skip_whitespace is inefficient for non-whitespace tokens

[jdm]

Depending on how often we call skip_whitespace when non-whitespace tokens are expected (which is actually quite common, IIRC) we can probably make it more efficient by adding a case to the top of the match at https://github.com/servo/rust-cssparser/blob/master/src/tokenizer.rs#L459 that catches any byte after / and immediately returns.

[SimonSapin]

Line 459 is currently b' ' | b'\t' => {, I don’t understand what you mean by "any byte after /".

[jdm]

I mean that / has the highest ascii value, so any byte value greater than it is clearly not whitespace. Adding a check for that case before any other might be more efficient.

[SimonSapin]

Oh I see. So something like byte if byte > b'/' => return, ?

[jdm]

Yeah.

[emilio]

We use match_byte for this now, which uses a lookup table, which seems to perform well