elinks/src/dom/css/scanner.h

#ifndef EL__DOM_CSS_SCANNER_H
#define EL__DOM_CSS_SCANNER_H

#include "dom/scanner.h"

/* The various token types and what they contain. Patterns taken from
 * the flex scanner declarations in the CSS 2 Specification. */
enum css_token_type {
	/* Char tokens: */

	/* Char tokens range from 1 to 255 and have their char value as type */
	/* meaning non char tokens have values from 256 and up. */

	/* Low level string tokens: */

	/* {...} means char group, <...> means token */
	/* {identstart}	[a-z_]|{nonascii} */
	/* {ident}	[a-z0-9_-]|{nonascii} */
	/* <ident>	{identstart}{ident}* */
	/* <name>	{ident}+ */
	/* <number>	[0-9]+|[0-9]*"."[0-9]+ */

	/* Percentage is put because although it looks like being composed of
	 * <number> and '%' floating point numbers are really not allowed but
	 * strtol() will round it down for us ;) */
	CSS_TOKEN_IDENT = 256,	/* <ident> */
	CSS_TOKEN_NUMBER,	/* <number> */
	CSS_TOKEN_PERCENTAGE,	/* <number>% */
	CSS_TOKEN_STRING,	/* Char sequence delimted by matching ' or " */

	/* High level string tokens: */

	/* The various number values; dimension being the most generic */
	CSS_TOKEN_ANGLE,	/* <number>rad, <number>grad or <number>deg */
	CSS_TOKEN_DIMENSION,	/* <number><ident> */
	CSS_TOKEN_EM,		/* <number>em */
	CSS_TOKEN_EX,		/* <number>ex */
	CSS_TOKEN_FREQUENCY,	/* <number>Hz or <number>kHz */
	CSS_TOKEN_LENGTH,	/* <number>{px,cm,mm,in,pt,pc} */
	CSS_TOKEN_TIME,		/* <number>ms or <number>s */

	/* XXX: CSS_TOKEN_HASH conflicts with CSS_TOKEN_HEX_COLOR. Generating
	 * hex color tokens has precedence and the hash token user have to
	 * treat CSS_TOKEN_HASH and CSS_TOKEN_HEX_COLOR alike. */
	CSS_TOKEN_HASH,		/* #<name> */
	CSS_TOKEN_HEX_COLOR,	/* #[0-9a-f]\{3,6} */

	/* For all unknown functions we generate on token contain both function name
	 * and args so scanning/parsing is easier. Besides we already check for
	 * ending ')'. */
	/* For known functions where we need several args [like rgb()] we want
	 * to generate tokens for every arg and arg delimiter ( ',' or ')' ).
	 * Because url() is a bit triggy: it can contain both <string> and some
	 * chars that would other wise make the scanner probably choke we also
	 * include the arg in that token. Besides it will make things like
	 * 'background' property parsing easier. */
	CSS_TOKEN_FUNCTION,	/* <ident>(<args>) */
	CSS_TOKEN_RGB,		/* rgb( */
	CSS_TOKEN_URL,		/* url(<arg>) */

	/* @-rule symbols */
	CSS_TOKEN_AT_KEYWORD,	/* @<ident> */
	CSS_TOKEN_AT_CHARSET,	/* @charset */
	CSS_TOKEN_AT_FONT_FACE,	/* @font-face */
	CSS_TOKEN_AT_IMPORT,	/* @import */
	CSS_TOKEN_AT_MEDIA,	/* @media */
	CSS_TOKEN_AT_PAGE,	/* @page */

	CSS_TOKEN_IMPORTANT,	/* !<whitespace>important */

	/* TODO: Selector stuff: */
	CSS_TOKEN_SELECT_SPACE_LIST,	/* ~= */
	CSS_TOKEN_SELECT_HYPHEN_LIST,	/* |= */
	CSS_TOKEN_SELECT_BEGIN,		/* ^= */
	CSS_TOKEN_SELECT_END,		/* $= */
	CSS_TOKEN_SELECT_CONTAINS,	/* *= */

	/* Special tokens: */

	/* A special token for unrecognized strings */
	CSS_TOKEN_GARBAGE,

	/* Token type used internally when scanning to signal that the token
	 * should not be recorded in the scanners token table. */
	CSS_TOKEN_SKIP,

	/* Another internal token type used both to mark unused tokens in the
	 * scanner table as invalid or when scanning to signal that the
	 * scanning should end. */
	CSS_TOKEN_NONE = 0,
};

extern struct dom_scanner_info dom_css_scanner_info;

#define skip_css_tokens(scanner, type) \
	skip_dom_scanner_tokens(scanner, type, get_css_precedence(type))

#define get_css_precedence(token_type) \
	((token_type) == '}' ? (1 << 10) : \
	 (token_type) == '{' ? (1 <<  9) : \
	 (token_type) == ';' ? (1 <<  8) : \
	 (token_type) == ')' ? (1 <<  7) : 0)

/* Check whether it is safe to skip the @token when looking for @skipto. */
static inline int
check_css_precedence(int type, int skipto)
{
	return get_css_precedence(type) < get_css_precedence(skipto);
}

#endif