Skip to content
This repository was archived by the owner on Jan 23, 2020. It is now read-only.

Avoided string split() and join() in HTML Canonicalization. #41

Open
adon-at-work wants to merge 2 commits into
base: master
Choose a base branch
from
Open

Avoided string split() and join() in HTML Canonicalization. #41

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
c4c4244
Avoided string split() and join()
Oct 8, 2015
5a27daa
enhanced benchmark scripts
Oct 8, 2015
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
14 changes: 11 additions & 3 deletions bin/benchmark
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -20,10 +20,17 @@ var allParsers = parsers.concat([
['HtmlParser2', 'htmlparser2', 'Parser', 'end'],
]);

console.log("Usage: benchmark [all]");
var selfParsers = [
['Context Parser (FastParser)', '../src/context-parser', 'FastParser', 'contextualize', {}],
['Context Parser (Parser)', '../src/context-parser', 'Parser', 'contextualize', {}],
['Context Parser (Parser,stateTracking=Off)', '../src/context-parser', 'Parser', 'contextualize', {enableStateTracking: false}],
['Context Parser (Parser,stateTracking=Off,canonicalization=On)', '../src/context-parser', 'Parser', 'contextualize', {enableStateTracking: false, enableCanonicalization: true}]
];

console.log("Usage: benchmark [all|self]");

if ( process.argv.length > 2 ) {
parsers = allParsers;
parsers = process.argv[3] === 'all' ? allParsers : selfParsers;
}

parsers.forEach(function(parser) {
Expand All @@ -33,13 +40,14 @@ parsers.forEach(function(parser) {
var classname = parser[1];
var name = parser[2];
var method = parser[3];
var config = parser[4];

try {
if ( name || method ) {
var Parser = name ? require(classname)[name] : require(classname);
start = +new Date();
for(var i=0; i<10; i++) {
var parser = new Parser();
var parser = config ? new Parser(config) : new Parser();
parser[method](html);
}
end = +new Date();
Expand Down
145 changes: 32 additions & 113 deletions src/context-parser.js
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -524,14 +524,18 @@ Parser.prototype._getNextState = function (state, i, endsWithEOF) {
};

/**
* @function Parser._convertString2Array
* @function Parser._inputSplice
* @param {integer} start - Index at which to start changing the this.input
* @param {integer} deleteCount - An integer indicating the number of characters to remove from this.input
* @param {string} addChars - A string to insert into the this.input
*
* @description
* Convert the immutable this.input to array type for Strict Context Parser processing (lazy conversion).
*
* This provides a similar splice interface to act on this.input as if it's an array. In addition, it updates the new length to this.inputLen
*/
Parser.prototype._convertString2Array = function () {
if (typeof this.input === "string") this.input = this.input.split('');
Parser.prototype._inputSplice = function(start, deleteCount, addChars) {
var str = this.input;
this.input = str.substr(0, start) + addChars + str.substr(start + deleteCount);
this.inputLen += addChars.length - deleteCount;
};

/**
Expand Down Expand Up @@ -567,20 +571,9 @@ Parser.prototype.fork = function() {
*/
Parser.prototype.contextualize = function (input, endsWithEOF) {
FastParser.prototype.contextualize.call(this, input, endsWithEOF);
return this.getModifiedInput();
return this.input;
};

/**
* @function Parser#getModifiedInput
*
* @description
* Get the modified input due to Strict Context Parser processing.
*
*/
Parser.prototype.getModifiedInput = function() {
// TODO: it is not defensive enough, should use Array.isArray, but need polyfill
return (typeof this.input === "string")? this.input:this.input.join('');
};

/**
* @function Parser#setCurrentState
Expand Down Expand Up @@ -675,12 +668,8 @@ Parser.prototype.getLastState = function() {
*/

function ConvertBogusCommentToComment(i) {
// for lazy conversion
this._convertString2Array();

// convert !--. i.e., from <* to <!--*
this.input.splice(i, 0, '!', '-', '-');
this.inputLen += 3;
this._inputSplice(i, 0, '!--');

// convert the next > to -->
this.on('preCanonicalize', PreCanonicalizeConvertBogusCommentEndTag);
Expand All @@ -691,12 +680,8 @@ function PreCanonicalizeConvertBogusCommentEndTag(state, i, endsWithEOF) {
// remove itself from the listener list
this.off('preCanonicalize', PreCanonicalizeConvertBogusCommentEndTag);

// for lazy conversion
this._convertString2Array();

// convert [>] to [-]->
this.input.splice(i, 0, '-', '-');
this.inputLen += 2;
this._inputSplice(i, 0, '--');

this.emit(this.listeners.bogusCommentCoverted, [state, i, endsWithEOF]);
}
Expand Down Expand Up @@ -730,44 +715,34 @@ function Canonicalize(state, i, endsWithEOF) {
// batch replacement of NULL with \uFFFD would violate the spec
// - for example, NULL is untouched in CDATA section state
if (chr === '\x00' && statesRequiringNullReplacement[state]) {
// for lazy conversion
this._convertString2Array();
this.input[i] = '\uFFFD';
this._inputSplice(i, 1, '\uFFFD');
}
// encode < into &lt; for [<]* (* is non-alpha) in STATE_DATA, [<]% and [<]! in STATE_RCDATA and STATE_RAWTEXT
else if ((potentialState === htmlState.STATE_TAG_OPEN && nextPotentialState === htmlState.STATE_DATA) || // [<]*, where * is non-alpha
((state === htmlState.STATE_RCDATA || state === htmlState.STATE_RAWTEXT) && // in STATE_RCDATA and STATE_RAWTEXT
chr === '<' && (nextChr === '%' || nextChr === '!'))) { // [<]% or [<]!
// for lazy conversion
this._convertString2Array();

// [<]*, [<]%, [<]!
this.input.splice(i, 1, '&', 'l', 't', ';');
this.inputLen += 3;
this._inputSplice(i, 1, '&lt;');
}
// enforce <!doctype html>
// + convert bogus comment or unknown doctype to the standard html comment
else if (potentialState === htmlState.STATE_MARKUP_DECLARATION_OPEN) { // <[!]***
reCanonicalizeNeeded = false;

// for lazy conversion
this._convertString2Array();

// context-parser treats the doctype and [CDATA[ as resulting into STATE_BOGUS_COMMENT
// so, we need our algorithm here to extract and check the next 7 characters
var commentKey = this.input.slice(i + 1, i + 8).join('');
var commentKey = this.input.slice(i + 1, i + 8);

// enforce <!doctype html>
if (commentKey.toLowerCase() === 'doctype') { // <![d]octype
// extract 6 chars immediately after <![d]octype and check if it's equal to ' html>'
if (this.input.slice(i + 8, i + 14).join('').toLowerCase() !== ' html>') {
if (this.input.slice(i + 8, i + 14).toLowerCase() !== ' html>') {

// replace <[!]doctype xxxx> with <[!]--!doctype xxxx--><doctype html>
ConvertBogusCommentToComment.call(this, i);

this.once('bogusCommentCoverted', function (state, i) {
[].splice.apply(this.input, [i + 3, 0].concat('<!doctype html>'.split('')));
this.inputLen += 15;
this._inputSplice(i + 3, 0, '<!doctype html>');
});

reCanonicalizeNeeded = true;
Expand Down Expand Up @@ -809,10 +784,7 @@ function Canonicalize(state, i, endsWithEOF) {
// htmlState.STATE_BEFORE_ATTRIBUTE_NAME, // <a [/]* replaced with <a [ ]*
// htmlState.STATE_AFTER_ATTRIBUTE_VALUE_QUOTED].indexOf(state) !== -1) // <a abc=""[/]* replaced with <a abc=""[ ]*

// for lazy conversion
this._convertString2Array();

this.input[i] = ' ';
this._inputSplice(i, 1, ' ');

// given this.input[i] was '/', nextPotentialState was htmlState.STATE_BEFORE_ATTRIBUTE_NAME
// given this.input[i] is now ' ', nextPotentialState becomes STATE_BEFORE_ATTRIBUTE_NAME if current state is STATE_ATTRIBUTE_NAME or STATE_AFTER_ATTRIBUTE_NAME
Expand All @@ -821,29 +793,20 @@ function Canonicalize(state, i, endsWithEOF) {
if (state === htmlState.STATE_ATTRIBUTE_NAME || // <a abc[/]=abc replaced with <a abc[ ]*
state === htmlState.STATE_AFTER_ATTRIBUTE_NAME) { // <a abc [/]=abc replaced with <a abc [ ]*
for (var j = i + 1; j < this.inputLen && this.input[j] === '='; j++) {
this.input.splice(j, 1);
this.inputLen--;
this._inputSplice(j, 1, '');
}
}
}
// remove unnecessary equal signs, hence <input checked[=]> become <input checked[>], or <input checked [=]> become <input checked [>]
else if (potentialState === htmlState.STATE_BEFORE_ATTRIBUTE_VALUE && // only from STATE_ATTRIBUTE_NAME or STATE_AFTER_ATTRIBUTE_NAME
nextPotentialState === htmlState.STATE_DATA) { // <a abc[=]> or <a abc [=]>
// for lazy conversion
this._convertString2Array();

this.input.splice(i, 1);
this.inputLen--;
this._inputSplice(i, 1, '');
}
// insert a space for <a abc="***["]* or <a abc='***[']* after quoted attribute value (i.e., <a abc="***["] * or <a abc='***['] *)
else if (potentialState === htmlState.STATE_AFTER_ATTRIBUTE_VALUE_QUOTED && // <a abc=""[*] where * is not SPACE (\t,\n,\f,' ')
nextPotentialState === htmlState.STATE_BEFORE_ATTRIBUTE_NAME &&
this._getSymbol(i + 1) !== stateMachine.Symbol.SPACE) {
// for lazy conversion
this._convertString2Array();

this.input.splice(i + 1, 0, ' ');
this.inputLen++;
this._inputSplice(i + 1, 0, ' ');
}
// else here means no special pattern was found requiring rewriting
else {
Expand All @@ -854,23 +817,15 @@ function Canonicalize(state, i, endsWithEOF) {
switch (potentialState) {
case htmlState.STATE_BEFORE_ATTRIBUTE_NAME: // remove ambigious symbols in <a [*]href where * is ", ', <, or =
if (nextChr === "=") {
// for lazy conversion
this._convertString2Array();

this.input.splice(i + 1, 1);
this.inputLen--;
this._inputSplice(i + 1, 1, '');
reCanonicalizeNeeded = true;
break;
}
/* falls through */
case htmlState.STATE_ATTRIBUTE_NAME: // remove ambigious symbols in <a href[*] where * is ", ', or <
case htmlState.STATE_AFTER_ATTRIBUTE_NAME: // remove ambigious symbols in <a href [*] where * is ", ', or <
if (nextChr === '"' || nextChr === "'" || nextChr === '<') {
// for lazy conversion
this._convertString2Array();

this.input.splice(i + 1, 1);
this.inputLen--;
this._inputSplice(i + 1, 1, '');
reCanonicalizeNeeded = true;
}
break;
Expand All @@ -884,74 +839,42 @@ function Canonicalize(state, i, endsWithEOF) {
// escape " ' < = ` to avoid raising parse errors for unquoted value
case htmlState.STATE_ATTRIBUTE_VALUE_UNQUOTED:
if (chr === '"') {
// for lazy conversion
this._convertString2Array();

this.input.splice(i, 1, '&', 'q', 'u', 'o', 't', ';');
this.inputLen += 5;
this._inputSplice(i, 1, '&quot;');
break;
} else if (chr === "'") {
// for lazy conversion
this._convertString2Array();

this.input.splice(i, 1, '&', '#', '3', '9', ';');
this.inputLen += 4;
this._inputSplice(i, 1, '&#39;');
break;
}
/* falls through */
case htmlState.STATE_BEFORE_ATTRIBUTE_VALUE: // treat < = ` as if they are in STATE_ATTRIBUTE_VALUE_UNQUOTED
if (chr === '<') {
// for lazy conversion
this._convertString2Array();

this.input.splice(i, 1, '&', 'l', 't', ';');
this.inputLen += 3;
this._inputSplice(i, 1, '&lt;');
} else if (chr === '=') {
// for lazy conversion
this._convertString2Array();

this.input.splice(i, 1, '&', '#', '6', '1', ';');
this.inputLen += 4;
this._inputSplice(i, 1, '&#61;');
} else if (chr === '`') {
// for lazy conversion
this._convertString2Array();

this.input.splice(i, 1, '&', '#', '9', '6', ';');
this.inputLen += 4;
this._inputSplice(i, 1, '&#96;');
}
break;

// add hyphens to complete <!----> to avoid raising parsing errors
// replace <!--[>] with <!--[-]->
case htmlState.STATE_COMMENT_START:
if (chr === '>') { // <!--[>]
// for lazy conversion
this._convertString2Array();

this.input.splice(i, 0, '-', '-');
this.inputLen += 2;
this._inputSplice(i, 0, '--');
// reCanonicalizeNeeded = true; // not need due to no where to treat its potential states
}
break;
// replace <!---[>] with <!---[-]>
case htmlState.STATE_COMMENT_START_DASH:
if (chr === '>') { // <!---[>]
// for lazy conversion
this._convertString2Array();

this.input.splice(i, 0, '-');
this.inputLen++;
this._inputSplice(i, 0, '-');
// reCanonicalizeNeeded = true; // not need due to no where to treat its potential states
}
break;
// replace --[!]> with --[>]
case htmlState.STATE_COMMENT_END:
if (chr === '!' && nextChr === '>') {
// for lazy conversion
this._convertString2Array();

this.input.splice(i, 1);
this.inputLen--;
this._inputSplice(i, 1, '');
// reCanonicalizeNeeded = true; // not need due to no where to treat its potential states
}
// if (chr === '-'), ignored this parse error. TODO: consider stripping n-2 hyphens for ------>
Expand All @@ -966,11 +889,7 @@ function Canonicalize(state, i, endsWithEOF) {
// remove IE conditional comments
function DisableIEConditionalComments(state, i){
if (state === htmlState.STATE_COMMENT && this.input[i] === ']' && this.input[i+1] === '>') {
// for lazy conversion
this._convertString2Array();

this.input.splice(i + 1, 0, ' ');
this.inputLen++;
this._inputSplice(i + 1, 0, ' ');
}
}

Expand Down