Bug 1154809 - rewrite output-parser.js to use CSSLexer; r=pbrosset

This commit is contained in:
Tom Tromey 2015-05-19 08:56:16 -07:00
Родитель 109c223159
Коммит 2a621cdafe
1 изменённых файлов: 122 добавлений и 163 удалений

Просмотреть файл

@ -11,28 +11,18 @@ const {Services} = Cu.import("resource://gre/modules/Services.jsm", {});
const HTML_NS = "http://www.w3.org/1999/xhtml";
const MAX_ITERATIONS = 100;
const REGEX_QUOTES = /^".*?"|^".*|^'.*?'|^'.*/;
const REGEX_WHITESPACE = /^\s+/;
const REGEX_FIRST_WORD_OR_CHAR = /^\w+|^./;
const REGEX_CUBIC_BEZIER = /^linear|^ease-in-out|^ease-in|^ease-out|^ease|^cubic-bezier\(([0-9.\- ]+,){3}[0-9.\- ]+\)/;
// CSS variable names are identifiers which the spec defines as follows:
// In CSS, identifiers (including element names, classes, and IDs in
// selectors) can contain only the characters [a-zA-Z0-9] and ISO 10646
// characters U+00A0 and higher, plus the hyphen (-) and the underscore (_).
const REGEX_CSS_VAR = /^\bvar\(\s*--[-_a-zA-Z0-9\u00A0-\u10FFFF]+\s*\)/;
const BEZIER_KEYWORDS = ["linear", "ease-in-out", "ease-in", "ease-out", "ease"];
/**
* This regex matches:
* - #F00
* - #FF0000
* - hsl()
* - hsla()
* - rgb()
* - rgba()
* - color names
*/
const REGEX_ALL_COLORS = /^#[0-9a-fA-F]{3}\b|^#[0-9a-fA-F]{6}\b|^hsl\(.*?\)|^hsla\(.*?\)|^rgba?\(.*?\)|^[a-zA-Z-]+/;
// Functions that accept a color argument.
const COLOR_TAKING_FUNCTIONS = ["linear-gradient",
"-moz-linear-gradient",
"repeating-linear-gradient",
"-moz-repeating-linear-gradient",
"radial-gradient",
"-moz-radial-gradient",
"repeating-radial-gradient",
"-moz-repeating-radial-gradient"];
loader.lazyGetter(this, "DOMUtils", function () {
return Cc["@mozilla.org/inspector/dom-utils;1"].getService(Ci.inIDOMUtils);
@ -94,29 +84,44 @@ OutputParser.prototype = {
},
/**
* Matches the beginning of the provided string to a css background-image url
* and return both the whole url(...) match and the url itself.
* This isn't handled via a regular expression to make sure we can match urls
* that contain parenthesis easily
* Given an initial FUNCTION token, read tokens from |tokenStream|
* and collect all the (non-comment) text. Return the collected
* text. The function token and the close paren are included in the
* result.
*
* @param {CSSToken} initialToken
* The FUNCTION token.
* @param {String} text
* The original CSS text.
* @param {CSSLexer} tokenStream
* The token stream from which to read.
* @return {String}
* The text of body of the function call.
*/
_matchBackgroundUrl: function(text) {
let startToken = "url(";
if (text.indexOf(startToken) !== 0) {
return null;
_collectFunctionText: function(initialToken, text, tokenStream) {
let result = text.substring(initialToken.startOffset,
initialToken.endOffset);
let depth = 1;
while (depth > 0) {
let token = tokenStream.nextToken();
if (!token) {
break;
}
let uri = text.substring(startToken.length).trim();
let quote = uri.substring(0, 1);
if (quote === "'" || quote === '"') {
uri = uri.substring(1, uri.search(new RegExp(quote + "\\s*\\)")));
} else {
uri = uri.substring(0, uri.indexOf(")"));
quote = "";
if (token.tokenType === "comment") {
continue;
}
let end = startToken + quote + uri;
text = text.substring(0, text.indexOf(")", end.length) + 1);
return [text, uri.trim()];
result += text.substring(token.startOffset, token.endOffset);
if (token.tokenType === "symbol") {
if (token.text === "(") {
++depth;
} else if (token.text === ")") {
--depth;
}
} else if (token.tokenType === "function") {
++depth;
}
}
return result;
},
/**
@ -133,142 +138,96 @@ OutputParser.prototype = {
_parse: function(text, options={}) {
text = text.trim();
this.parsed.length = 0;
let i = 0;
while (text.length > 0) {
let matched = null;
if (options.expectFilter) {
this._appendFilter(text, options);
} else {
let tokenStream = DOMUtils.getCSSLexer(text);
let i = 0;
while (true) {
let token = tokenStream.nextToken();
if (!token)
break;
if (token.tokenType === "comment") {
continue;
}
// Prevent this loop from slowing down the browser with too
// many nodes being appended into output. In practice it is very unlikely
// that this will ever happen.
i++;
if (i > MAX_ITERATIONS) {
this._appendTextNode(text);
text = "";
this._appendTextNode(text.substring(token.startOffset,
token.endOffset));
continue;
}
switch (token.tokenType) {
case "function": {
if (COLOR_TAKING_FUNCTIONS.indexOf(token.text) >= 0) {
// The function can accept a color argument, and we know
// it isn't special in some other way. So, we let it
// through to the ordinary parsing loop so that colors
// can be handled in a single place.
this._appendTextNode(text.substring(token.startOffset,
token.endOffset));
} else {
let functionText = this._collectFunctionText(token, text,
tokenStream);
if (options.expectCubicBezier && token.text === "cubic-bezier") {
this._appendCubicBezier(functionText, options);
} else if (options.supportsColor &&
DOMUtils.isValidCSSColor(functionText)) {
this._appendColor(functionText, options);
} else {
this._appendTextNode(functionText);
}
}
break;
}
if (options.expectFilter) {
this._appendFilter(text, options);
case "ident":
if (options.expectCubicBezier &&
BEZIER_KEYWORDS.indexOf(token.text) >= 0) {
this._appendCubicBezier(token.text, options);
} else if (options.supportsColor &&
DOMUtils.isValidCSSColor(token.text)) {
this._appendColor(token.text, options);
} else {
this._appendTextNode(text.substring(token.startOffset,
token.endOffset));
}
break;
case "id":
case "hash": {
let original = text.substring(token.startOffset, token.endOffset);
if (options.supportsColor && DOMUtils.isValidCSSColor(original)) {
this._appendColor(original, options);
} else {
this._appendTextNode(original);
}
break;
}
matched = text.match(REGEX_QUOTES);
if (matched) {
let match = matched[0];
case "url":
case "bad_url":
this._appendURL(text.substring(token.startOffset, token.endOffset),
token.text, options);
break;
text = this._trimMatchFromStart(text, match);
this._appendTextNode(match);
continue;
default:
this._appendTextNode(text.substring(token.startOffset,
token.endOffset));
break;
}
matched = text.match(REGEX_CSS_VAR);
if (matched) {
let match = matched[0];
text = this._trimMatchFromStart(text, match);
this._appendTextNode(match);
continue;
}
matched = text.match(REGEX_WHITESPACE);
if (matched) {
let match = matched[0];
text = this._trimMatchFromStart(text, match);
this._appendTextNode(match);
continue;
}
matched = this._matchBackgroundUrl(text);
if (matched) {
let [match, url] = matched;
text = this._trimMatchFromStart(text, match);
this._appendURL(match, url, options);
continue;
}
if (options.expectCubicBezier) {
matched = text.match(REGEX_CUBIC_BEZIER);
if (matched) {
let match = matched[0];
text = this._trimMatchFromStart(text, match);
this._appendCubicBezier(match, options);
continue;
}
}
if (options.supportsColor) {
let dirty;
[text, dirty] = this._appendColorOnMatch(text, options);
if (dirty) {
continue;
}
}
// This test must always be last as it indicates use of an unknown
// character that needs to be removed to prevent infinite loops.
matched = text.match(REGEX_FIRST_WORD_OR_CHAR);
if (matched) {
let match = matched[0];
text = this._trimMatchFromStart(text, match);
this._appendTextNode(match);
}
}
return this._toDOM();
},
/**
* Convenience function to make the parser a little more readable.
*
* @param {String} text
* Main text
* @param {String} match
* Text to remove from the beginning
*
* @return {String}
* The string passed as 'text' with 'match' stripped from the start.
*/
_trimMatchFromStart: function(text, match) {
return text.substr(match.length);
},
/**
* Check if there is a color match and append it if it is valid.
*
* @param {String} text
* Main text
* @param {Object} options
* Options object. For valid options and default values see
* _mergeOptions().
*
* @return {Array}
* An array containing the remaining text and a dirty flag. This array
* is designed for deconstruction using [text, dirty].
*/
_appendColorOnMatch: function(text, options) {
let dirty;
let matched = text.match(REGEX_ALL_COLORS);
if (matched) {
let match = matched[0];
if (this._appendColor(match, options)) {
text = this._trimMatchFromStart(text, match);
dirty = true;
}
} else {
dirty = false;
}
return [text, dirty];
},
/**
* Append a cubic-bezier timing function value to the output
*