Parse escaped content as part of all literals and values

[kitten]

In CSS any literal or value can contain escaped content, a simple example being content: "\""; Here we escape another quote inside the string with a backslash.

The best article on what's possible is here: https://mathiasbynens.be/notes/css-escapes

But the tl;dr is, any character can be escaped either using itself: "\+" = "+" or using a hex code: "\2b " = "+" or "\00002b" = "+". Note that if the hex escape sequence has an ambiguous length (anything else than exactly 6 hex chars) then a space must follow it.

A simpler detection of escaped content would be to allow either one char, or one or more hex chars (/[a-f][0-9]/i) with an optional space after them. This is easier and more efficient to parse. (Used by PostCSS as a heuristic for instance)

As soon as a character is escaped it can become part of a literal or a value, but it can never be part of any other token, meaning that in #\#some-id { the hash symbol doesn't stand for CSS’ hash token.

A problem for bredon in particular is that since any literal's char can be escaped, that also means that for 100% those would need to be normalised for literals. A simple example is font-\size which is still detected as font-size in the browser. So for correct detection and to make plugins work correctly this technically needs to be normalised. (This normalisation is not done in most CSS parsers however)