Module lex 

The lexer: the single cssToken matcher, the scanners it reads spans with, and the token buffer the rule machine peeks into.

CSS is context-sensitive — the same characters can begin a selector, a property or a value — so the lexer owns the hard tokenisation and the grammar only assembles typed nodes from what it emits. Which token comes out depends on the rule that is active when the character is read, which is why Lex::peek takes a rule name.

See ts/src/css.ts for the canonical commentary; the three ports are kept in step.

Structs

Error

A parse failure.

Lex

The lexer, plus the token buffer the rule machine peeks into.

Point

The lexer’s scan position.

Token

One lexed token.

Enums

Tin

A token kind.

Functions

end_pos

The position just after a token’s last character.

es_is_whitespace

Whether c is whitespace as ECMAScript defines it, which is NOT what char::is_whitespace answers.

es_trim

Trim as JavaScript’s String.prototype.trim does. See es_is_whitespace for why this is not str::trim.

split_selectors

Split a selector-group prelude on its top-level commas (commas inside strings, () or [] are part of a single selector), stripping comments and trimming each selector. An empty prelude yields an empty list.

start_pos

A token’s first-character position (1-based line and column).

strip_comments

Remove /* ... */ comments from a selector or value run, leaving quoted strings (which may contain /*) untouched.

vendor_prefix

^(-[a-z]+-): a vendor prefix on an at-keyword, e.g. -webkit-keyframes.