blob: a8fc8d4718ba2b2ee2fa5b2a33d180b21f9abb07 [file] [log] [blame]
// Copyright 2006-2009 the V8 project authors. All rights reserved.
// Redistribution and use in source and binary forms, with or without
// modification, are permitted provided that the following conditions are
// met:
//
// * Redistributions of source code must retain the above copyright
// notice, this list of conditions and the following disclaimer.
// * Redistributions in binary form must reproduce the above
// copyright notice, this list of conditions and the following
// disclaimer in the documentation and/or other materials provided
// with the distribution.
// * Neither the name of Google Inc. nor the names of its
// contributors may be used to endorse or promote products derived
// from this software without specific prior written permission.
//
// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
// A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
// OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
// LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
// DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
// THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
// This file relies on the fact that the following declaration has been made
// in runtime.js:
// const $String = global.String;
// const $NaN = 0/0;
// Set the String function and constructor.
%SetCode($String, function(x) {
var value = %_ArgumentsLength() == 0 ? '' : TO_STRING_INLINE(x);
if (%_IsConstructCall()) {
%_SetValueOf(this, value);
} else {
return value;
}
});
%FunctionSetPrototype($String, new $String());
// ECMA-262 section 15.5.4.2
function StringToString() {
if (!IS_STRING(this) && !IS_STRING_WRAPPER(this))
throw new $TypeError('String.prototype.toString is not generic');
return %_ValueOf(this);
}
// ECMA-262 section 15.5.4.3
function StringValueOf() {
if (!IS_STRING(this) && !IS_STRING_WRAPPER(this))
throw new $TypeError('String.prototype.valueOf is not generic');
return %_ValueOf(this);
}
// ECMA-262, section 15.5.4.4
function StringCharAt(pos) {
var char_code = %_FastCharCodeAt(this, pos);
if (!%_IsSmi(char_code)) {
var subject = TO_STRING_INLINE(this);
var index = TO_INTEGER(pos);
if (index >= subject.length || index < 0) return "";
char_code = %StringCharCodeAt(subject, index);
}
return %CharFromCode(char_code);
}
// ECMA-262 section 15.5.4.5
function StringCharCodeAt(pos) {
var fast_answer = %_FastCharCodeAt(this, pos);
if (%_IsSmi(fast_answer)) {
return fast_answer;
}
var subject = TO_STRING_INLINE(this);
var index = TO_INTEGER(pos);
return %StringCharCodeAt(subject, index);
}
// ECMA-262, section 15.5.4.6
function StringConcat() {
var len = %_ArgumentsLength();
var this_as_string = TO_STRING_INLINE(this);
if (len === 1) {
return this_as_string + %_Arguments(0);
}
var parts = new $Array(len + 1);
parts[0] = this_as_string;
for (var i = 0; i < len; i++) {
var part = %_Arguments(i);
parts[i + 1] = TO_STRING_INLINE(part);
}
return %StringBuilderConcat(parts, len + 1, "");
}
// Match ES3 and Safari
%FunctionSetLength(StringConcat, 1);
// ECMA-262 section 15.5.4.7
function StringIndexOf(searchString /* position */) { // length == 1
var subject_str = TO_STRING_INLINE(this);
var pattern_str = TO_STRING_INLINE(searchString);
var subject_str_len = subject_str.length;
var pattern_str_len = pattern_str.length;
var index = 0;
if (%_ArgumentsLength() > 1) {
var arg1 = %_Arguments(1); // position
index = TO_INTEGER(arg1);
}
if (index < 0) index = 0;
if (index > subject_str_len) index = subject_str_len;
if (pattern_str_len + index > subject_str_len) return -1;
return %StringIndexOf(subject_str, pattern_str, index);
}
// ECMA-262 section 15.5.4.8
function StringLastIndexOf(searchString /* position */) { // length == 1
var sub = TO_STRING_INLINE(this);
var subLength = sub.length;
var pat = TO_STRING_INLINE(searchString);
var patLength = pat.length;
var index = subLength - patLength;
if (%_ArgumentsLength() > 1) {
var position = ToNumber(%_Arguments(1));
if (!$isNaN(position)) {
position = TO_INTEGER(position);
if (position < 0) {
position = 0;
}
if (position + patLength < subLength) {
index = position
}
}
}
if (index < 0) {
return -1;
}
return %StringLastIndexOf(sub, pat, index);
}
// ECMA-262 section 15.5.4.9
//
// This function is implementation specific. For now, we do not
// do anything locale specific.
function StringLocaleCompare(other) {
if (%_ArgumentsLength() === 0) return 0;
var this_str = TO_STRING_INLINE(this);
var other_str = TO_STRING_INLINE(other);
return %StringLocaleCompare(this_str, other_str);
}
// ECMA-262 section 15.5.4.10
function StringMatch(regexp) {
if (!IS_REGEXP(regexp)) regexp = new $RegExp(regexp);
var subject = TO_STRING_INLINE(this);
if (!regexp.global) return regexp.exec(subject);
%_Log('regexp', 'regexp-match,%0S,%1r', [subject, regexp]);
// lastMatchInfo is defined in regexp.js.
return %StringMatch(subject, regexp, lastMatchInfo);
}
// SubString is an internal function that returns the sub string of 'string'.
// If resulting string is of length 1, we use the one character cache
// otherwise we call the runtime system.
function SubString(string, start, end) {
// Use the one character string cache.
if (start + 1 == end) {
var char_code = %_FastCharCodeAt(string, start);
if (!%_IsSmi(char_code)) {
char_code = %StringCharCodeAt(string, start);
}
return %CharFromCode(char_code);
}
return %_SubString(string, start, end);
}
// This has the same size as the lastMatchInfo array, and can be used for
// functions that expect that structure to be returned. It is used when the
// needle is a string rather than a regexp. In this case we can't update
// lastMatchArray without erroneously affecting the properties on the global
// RegExp object.
var reusableMatchInfo = [2, "", "", -1, -1];
// ECMA-262, section 15.5.4.11
function StringReplace(search, replace) {
var subject = TO_STRING_INLINE(this);
// Delegate to one of the regular expression variants if necessary.
if (IS_REGEXP(search)) {
%_Log('regexp', 'regexp-replace,%0r,%1S', [search, subject]);
if (IS_FUNCTION(replace)) {
return StringReplaceRegExpWithFunction(subject, search, replace);
} else {
return StringReplaceRegExp(subject, search, replace);
}
}
// Convert the search argument to a string and search for it.
search = TO_STRING_INLINE(search);
var start = %StringIndexOf(subject, search, 0);
if (start < 0) return subject;
var end = start + search.length;
var builder = new ReplaceResultBuilder(subject);
// prefix
builder.addSpecialSlice(0, start);
// Compute the string to replace with.
if (IS_FUNCTION(replace)) {
builder.add(replace.call(null, search, start, subject));
} else {
reusableMatchInfo[CAPTURE0] = start;
reusableMatchInfo[CAPTURE1] = end;
replace = TO_STRING_INLINE(replace);
ExpandReplacement(replace, subject, reusableMatchInfo, builder);
}
// suffix
builder.addSpecialSlice(end, subject.length);
return builder.generate();
}
// Helper function for regular expressions in String.prototype.replace.
function StringReplaceRegExp(subject, regexp, replace) {
replace = TO_STRING_INLINE(replace);
return %StringReplaceRegExpWithString(subject,
regexp,
replace,
lastMatchInfo);
};
// Expand the $-expressions in the string and return a new string with
// the result.
function ExpandReplacement(string, subject, matchInfo, builder) {
var next = %StringIndexOf(string, '$', 0);
if (next < 0) {
builder.add(string);
return;
}
// Compute the number of captures; see ECMA-262, 15.5.4.11, p. 102.
var m = NUMBER_OF_CAPTURES(matchInfo) >> 1; // Includes the match.
if (next > 0) builder.add(SubString(string, 0, next));
var length = string.length;
while (true) {
var expansion = '$';
var position = next + 1;
if (position < length) {
var peek = %_FastCharCodeAt(string, position);
if (!%_IsSmi(peek)) {
peek = %StringCharCodeAt(string, position);
}
if (peek == 36) { // $$
++position;
builder.add('$');
} else if (peek == 38) { // $& - match
++position;
builder.addSpecialSlice(matchInfo[CAPTURE0],
matchInfo[CAPTURE1]);
} else if (peek == 96) { // $` - prefix
++position;
builder.addSpecialSlice(0, matchInfo[CAPTURE0]);
} else if (peek == 39) { // $' - suffix
++position;
builder.addSpecialSlice(matchInfo[CAPTURE1], subject.length);
} else if (peek >= 48 && peek <= 57) { // $n, 0 <= n <= 9
++position;
var n = peek - 48;
if (position < length) {
peek = %_FastCharCodeAt(string, position);
if (!%_IsSmi(peek)) {
peek = %StringCharCodeAt(string, position);
}
// $nn, 01 <= nn <= 99
if (n != 0 && peek == 48 || peek >= 49 && peek <= 57) {
var nn = n * 10 + (peek - 48);
if (nn < m) {
// If the two digit capture reference is within range of
// the captures, we use it instead of the single digit
// one. Otherwise, we fall back to using the single
// digit reference. This matches the behavior of
// SpiderMonkey.
++position;
n = nn;
}
}
}
if (0 < n && n < m) {
addCaptureString(builder, matchInfo, n);
} else {
// Because of the captures range check in the parsing of two
// digit capture references, we can only enter here when a
// single digit capture reference is outside the range of
// captures.
builder.add('$');
--position;
}
} else {
builder.add('$');
}
} else {
builder.add('$');
}
// Go the the next $ in the string.
next = %StringIndexOf(string, '$', position);
// Return if there are no more $ characters in the string. If we
// haven't reached the end, we need to append the suffix.
if (next < 0) {
if (position < length) {
builder.add(SubString(string, position, length));
}
return;
}
// Append substring between the previous and the next $ character.
builder.add(SubString(string, position, next));
}
};
// Compute the string of a given regular expression capture.
function CaptureString(string, lastCaptureInfo, index) {
// Scale the index.
var scaled = index << 1;
// Compute start and end.
var start = lastCaptureInfo[CAPTURE(scaled)];
var end = lastCaptureInfo[CAPTURE(scaled + 1)];
// If either start or end is missing return undefined.
if (start < 0 || end < 0) return;
return SubString(string, start, end);
};
// Add the string of a given regular expression capture to the
// ReplaceResultBuilder
function addCaptureString(builder, matchInfo, index) {
// Scale the index.
var scaled = index << 1;
// Compute start and end.
var start = matchInfo[CAPTURE(scaled)];
var end = matchInfo[CAPTURE(scaled + 1)];
// If either start or end is missing return.
if (start < 0 || end <= start) return;
builder.addSpecialSlice(start, end);
};
// Helper function for replacing regular expressions with the result of a
// function application in String.prototype.replace. The function application
// must be interleaved with the regexp matching (contrary to ECMA-262
// 15.5.4.11) to mimic SpiderMonkey and KJS behavior when the function uses
// the static properties of the RegExp constructor. Example:
// 'abcd'.replace(/(.)/g, function() { return RegExp.$1; }
// should be 'abcd' and not 'dddd' (or anything else).
function StringReplaceRegExpWithFunction(subject, regexp, replace) {
var matchInfo = DoRegExpExec(regexp, subject, 0);
if (IS_NULL(matchInfo)) return subject;
var result = new ReplaceResultBuilder(subject);
// There's at least one match. If the regexp is global, we have to loop
// over all matches. The loop is not in C++ code here like the one in
// RegExp.prototype.exec, because of the interleaved function application.
// Unfortunately, that means this code is nearly duplicated, here and in
// jsregexp.cc.
if (regexp.global) {
var numberOfCaptures = NUMBER_OF_CAPTURES(matchInfo) >> 1;
var previous = 0;
do {
var startOfMatch = matchInfo[CAPTURE0];
result.addSpecialSlice(previous, startOfMatch);
previous = matchInfo[CAPTURE1];
if (numberOfCaptures == 1) {
var match = SubString(subject, startOfMatch, previous);
// Don't call directly to avoid exposing the built-in global object.
result.add(replace.call(null, match, startOfMatch, subject));
} else {
result.add(ApplyReplacementFunction(replace, matchInfo, subject));
}
// Can't use matchInfo any more from here, since the function could
// overwrite it.
// Continue with the next match.
// Increment previous if we matched an empty string, as per ECMA-262
// 15.5.4.10.
if (previous == startOfMatch) {
// Add the skipped character to the output, if any.
if (previous < subject.length) {
result.addSpecialSlice(previous, previous + 1);
}
previous++;
}
// Per ECMA-262 15.10.6.2, if the previous index is greater than the
// string length, there is no match
matchInfo = (previous > subject.length)
? null
: DoRegExpExec(regexp, subject, previous);
} while (!IS_NULL(matchInfo));
// Tack on the final right substring after the last match, if necessary.
if (previous < subject.length) {
result.addSpecialSlice(previous, subject.length);
}
} else { // Not a global regexp, no need to loop.
result.addSpecialSlice(0, matchInfo[CAPTURE0]);
var endOfMatch = matchInfo[CAPTURE1];
result.add(ApplyReplacementFunction(replace, matchInfo, subject));
// Can't use matchInfo any more from here, since the function could
// overwrite it.
result.addSpecialSlice(endOfMatch, subject.length);
}
return result.generate();
}
// Helper function to apply a string replacement function once.
function ApplyReplacementFunction(replace, matchInfo, subject) {
// Compute the parameter list consisting of the match, captures, index,
// and subject for the replace function invocation.
var index = matchInfo[CAPTURE0];
// The number of captures plus one for the match.
var m = NUMBER_OF_CAPTURES(matchInfo) >> 1;
if (m == 1) {
var s = CaptureString(subject, matchInfo, 0);
// Don't call directly to avoid exposing the built-in global object.
return replace.call(null, s, index, subject);
}
var parameters = $Array(m + 2);
for (var j = 0; j < m; j++) {
parameters[j] = CaptureString(subject, matchInfo, j);
}
parameters[j] = index;
parameters[j + 1] = subject;
return replace.apply(null, parameters);
}
// ECMA-262 section 15.5.4.12
function StringSearch(re) {
var regexp = new $RegExp(re);
var s = TO_STRING_INLINE(this);
var last_idx = regexp.lastIndex; // keep old lastIndex
regexp.lastIndex = 0; // ignore re.global property
var result = regexp.exec(s);
regexp.lastIndex = last_idx; // restore lastIndex
if (result == null)
return -1;
else
return result.index;
}
// ECMA-262 section 15.5.4.13
function StringSlice(start, end) {
var s = TO_STRING_INLINE(this);
var s_len = s.length;
var start_i = TO_INTEGER(start);
var end_i = s_len;
if (end !== void 0)
end_i = TO_INTEGER(end);
if (start_i < 0) {
start_i += s_len;
if (start_i < 0)
start_i = 0;
} else {
if (start_i > s_len)
start_i = s_len;
}
if (end_i < 0) {
end_i += s_len;
if (end_i < 0)
end_i = 0;
} else {
if (end_i > s_len)
end_i = s_len;
}
var num_c = end_i - start_i;
if (num_c < 0)
num_c = 0;
return SubString(s, start_i, start_i + num_c);
}
// ECMA-262 section 15.5.4.14
function StringSplit(separator, limit) {
var subject = TO_STRING_INLINE(this);
limit = (IS_UNDEFINED(limit)) ? 0xffffffff : TO_UINT32(limit);
if (limit === 0) return [];
// ECMA-262 says that if separator is undefined, the result should
// be an array of size 1 containing the entire string. SpiderMonkey
// and KJS have this behaviour only when no separator is given. If
// undefined is explicitly given, they convert it to a string and
// use that. We do as SpiderMonkey and KJS.
if (%_ArgumentsLength() === 0) {
return [subject];
}
var length = subject.length;
if (!IS_REGEXP(separator)) {
separator = TO_STRING_INLINE(separator);
var separator_length = separator.length;
// If the separator string is empty then return the elements in the subject.
if (separator_length === 0) {
var result = $Array(length);
for (var i = 0; i < length; i++) result[i] = subject[i];
return result;
}
var result = [];
var start_index = 0;
var index;
while (true) {
if (start_index + separator_length > length ||
(index = %StringIndexOf(subject, separator, start_index)) === -1) {
result.push(SubString(subject, start_index, length));
break;
}
if (result.push(SubString(subject, start_index, index)) === limit) break;
start_index = index + separator_length;
}
return result;
}
%_Log('regexp', 'regexp-split,%0S,%1r', [subject, separator]);
if (length === 0) {
if (splitMatch(separator, subject, 0, 0) != null) return [];
return [subject];
}
var currentIndex = 0;
var startIndex = 0;
var result = [];
while (true) {
if (startIndex === length) {
result[result.length] = subject.slice(currentIndex, length);
return result;
}
var matchInfo = splitMatch(separator, subject, currentIndex, startIndex);
if (IS_NULL(matchInfo)) {
result[result.length] = subject.slice(currentIndex, length);
return result;
}
var endIndex = matchInfo[CAPTURE1];
// We ignore a zero-length match at the currentIndex.
if (startIndex === endIndex && endIndex === currentIndex) {
startIndex++;
continue;
}
result[result.length] = SubString(subject, currentIndex, matchInfo[CAPTURE0]);
if (result.length === limit) return result;
var num_captures = NUMBER_OF_CAPTURES(matchInfo);
for (var i = 2; i < num_captures; i += 2) {
var start = matchInfo[CAPTURE(i)];
var end = matchInfo[CAPTURE(i + 1)];
if (start != -1 && end != -1) {
result[result.length] = SubString(subject, start, end);
} else {
result[result.length] = void 0;
}
if (result.length === limit) return result;
}
startIndex = currentIndex = endIndex;
}
}
// ECMA-262 section 15.5.4.14
// Helper function used by split. This version returns the matchInfo
// instead of allocating a new array with basically the same information.
function splitMatch(separator, subject, current_index, start_index) {
var matchInfo = DoRegExpExec(separator, subject, start_index);
if (matchInfo == null) return null;
// Section 15.5.4.14 paragraph two says that we do not allow zero length
// matches at the end of the string.
if (matchInfo[CAPTURE0] === subject.length) return null;
return matchInfo;
}
// ECMA-262 section 15.5.4.15
function StringSubstring(start, end) {
var s = TO_STRING_INLINE(this);
var s_len = s.length;
var start_i = TO_INTEGER(start);
if (start_i < 0) {
start_i = 0;
} else if (start_i > s_len) {
start_i = s_len;
}
var end_i = s_len;
if (!IS_UNDEFINED(end)) {
end_i = TO_INTEGER(end);
if (end_i > s_len) {
end_i = s_len;
} else {
if (end_i < 0) end_i = 0;
if (start_i > end_i) {
var tmp = end_i;
end_i = start_i;
start_i = tmp;
}
}
}
return SubString(s, start_i, end_i);
}
// This is not a part of ECMA-262.
function StringSubstr(start, n) {
var s = TO_STRING_INLINE(this);
var len;
// Correct n: If not given, set to string length; if explicitly
// set to undefined, zero, or negative, returns empty string.
if (n === void 0) {
len = s.length;
} else {
len = TO_INTEGER(n);
if (len <= 0) return '';
}
// Correct start: If not given (or undefined), set to zero; otherwise
// convert to integer and handle negative case.
if (start === void 0) {
start = 0;
} else {
start = TO_INTEGER(start);
// If positive, and greater than or equal to the string length,
// return empty string.
if (start >= s.length) return '';
// If negative and absolute value is larger than the string length,
// use zero.
if (start < 0) {
start += s.length;
if (start < 0) start = 0;
}
}
var end = start + len;
if (end > s.length) end = s.length;
return SubString(s, start, end);
}
// ECMA-262, 15.5.4.16
function StringToLowerCase() {
return %StringToLowerCase(TO_STRING_INLINE(this));
}
// ECMA-262, 15.5.4.17
function StringToLocaleLowerCase() {
return %StringToLowerCase(TO_STRING_INLINE(this));
}
// ECMA-262, 15.5.4.18
function StringToUpperCase() {
return %StringToUpperCase(TO_STRING_INLINE(this));
}
// ECMA-262, 15.5.4.19
function StringToLocaleUpperCase() {
return %StringToUpperCase(TO_STRING_INLINE(this));
}
// ES5, 15.5.4.20
function StringTrim() {
return %StringTrim(TO_STRING_INLINE(this), true, true);
}
function StringTrimLeft() {
return %StringTrim(TO_STRING_INLINE(this), true, false);
}
function StringTrimRight() {
return %StringTrim(TO_STRING_INLINE(this), false, true);
}
// ECMA-262, section 15.5.3.2
function StringFromCharCode(code) {
var n = %_ArgumentsLength();
if (n == 1) return %CharFromCode(ToNumber(code) & 0xffff)
// NOTE: This is not super-efficient, but it is necessary because we
// want to avoid converting to numbers from within the virtual
// machine. Maybe we can find another way of doing this?
var codes = new $Array(n);
for (var i = 0; i < n; i++) codes[i] = ToNumber(%_Arguments(i));
return %StringFromCharCodeArray(codes);
}
// Helper function for very basic XSS protection.
function HtmlEscape(str) {
return TO_STRING_INLINE(str).replace(/</g, "&lt;")
.replace(/>/g, "&gt;")
.replace(/"/g, "&quot;")
.replace(/'/g, "&#039;");
};
// Compatibility support for KJS.
// Tested by mozilla/js/tests/js1_5/Regress/regress-276103.js.
function StringLink(s) {
return "<a href=\"" + HtmlEscape(s) + "\">" + this + "</a>";
}
function StringAnchor(name) {
return "<a name=\"" + HtmlEscape(name) + "\">" + this + "</a>";
}
function StringFontcolor(color) {
return "<font color=\"" + HtmlEscape(color) + "\">" + this + "</font>";
}
function StringFontsize(size) {
return "<font size=\"" + HtmlEscape(size) + "\">" + this + "</font>";
}
function StringBig() {
return "<big>" + this + "</big>";
}
function StringBlink() {
return "<blink>" + this + "</blink>";
}
function StringBold() {
return "<b>" + this + "</b>";
}
function StringFixed() {
return "<tt>" + this + "</tt>";
}
function StringItalics() {
return "<i>" + this + "</i>";
}
function StringSmall() {
return "<small>" + this + "</small>";
}
function StringStrike() {
return "<strike>" + this + "</strike>";
}
function StringSub() {
return "<sub>" + this + "</sub>";
}
function StringSup() {
return "<sup>" + this + "</sup>";
}
// ReplaceResultBuilder support.
function ReplaceResultBuilder(str) {
this.elements = new $Array();
this.special_string = str;
}
ReplaceResultBuilder.prototype.add = function(str) {
str = TO_STRING_INLINE(str);
if (str.length > 0) {
var elements = this.elements;
elements[elements.length] = str;
}
}
ReplaceResultBuilder.prototype.addSpecialSlice = function(start, end) {
var len = end - start;
if (len == 0) return;
var elements = this.elements;
if (start < 0x80000 && len < 0x800) {
elements[elements.length] = (start << 11) + len;
} else {
// 0 < len <= String::kMaxLength and Smi::kMaxValue >= String::kMaxLength,
// so -len is a smi.
elements[elements.length] = -len;
elements[elements.length] = start;
}
}
ReplaceResultBuilder.prototype.generate = function() {
var elements = this.elements;
return %StringBuilderConcat(elements, elements.length, this.special_string);
}
function StringToJSON(key) {
return CheckJSONPrimitive(this.valueOf());
}
// -------------------------------------------------------------------
function SetupString() {
// Setup the constructor property on the String prototype object.
%SetProperty($String.prototype, "constructor", $String, DONT_ENUM);
// Setup the non-enumerable functions on the String object.
InstallFunctions($String, DONT_ENUM, $Array(
"fromCharCode", StringFromCharCode
));
// Setup the non-enumerable functions on the String prototype object.
InstallFunctionsOnHiddenPrototype($String.prototype, DONT_ENUM, $Array(
"valueOf", StringValueOf,
"toString", StringToString,
"charAt", StringCharAt,
"charCodeAt", StringCharCodeAt,
"concat", StringConcat,
"indexOf", StringIndexOf,
"lastIndexOf", StringLastIndexOf,
"localeCompare", StringLocaleCompare,
"match", StringMatch,
"replace", StringReplace,
"search", StringSearch,
"slice", StringSlice,
"split", StringSplit,
"substring", StringSubstring,
"substr", StringSubstr,
"toLowerCase", StringToLowerCase,
"toLocaleLowerCase", StringToLocaleLowerCase,
"toUpperCase", StringToUpperCase,
"toLocaleUpperCase", StringToLocaleUpperCase,
"trim", StringTrim,
"trimLeft", StringTrimLeft,
"trimRight", StringTrimRight,
"link", StringLink,
"anchor", StringAnchor,
"fontcolor", StringFontcolor,
"fontsize", StringFontsize,
"big", StringBig,
"blink", StringBlink,
"bold", StringBold,
"fixed", StringFixed,
"italics", StringItalics,
"small", StringSmall,
"strike", StringStrike,
"sub", StringSub,
"sup", StringSup,
"toJSON", StringToJSON
));
}
SetupString();