update narcissus. fix #365

This commit is contained in:
Fabian Jakobs 2011-08-04 12:11:51 +02:00
commit 1b02f2895e
4 changed files with 2817 additions and 1948 deletions

View file

@ -87,7 +87,7 @@ var Document = function(text) {
this.$detectNewLine = function(text) {
var match = text.match(/^.*?(\r?\n)/m);
var match = text.match(/^.*?(\r\n|\r|\n)/m);
if (match) {
this.$autoNewLine = match[1];
} else {
@ -111,7 +111,8 @@ var Document = function(text) {
this.$autoNewLine = "\n";
this.$newLineMode = "auto";
this.setNewLineMode = function(newLineMode) {
if (this.$newLineMode === newLineMode) return;
if (this.$newLineMode === newLineMode)
return;
this.$newLineMode = newLineMode;
};

View file

@ -52,324 +52,628 @@
define(function(require, exports, module) {
exports.options = {
version: 185,
};
(function() {
exports.hostGlobal = this
})();
var tokens = [
// End of source.
"END",
// Operators and punctuators. Some pair-wise order matters, e.g. (+, -)
// and (UNARY_PLUS, UNARY_MINUS).
"\n", ";",
",",
"=",
"?", ":", "CONDITIONAL",
"||",
"&&",
"|",
"^",
"&",
"==", "!=", "===", "!==",
"<", "<=", ">=", ">",
"<<", ">>", ">>>",
"+", "-",
"*", "/", "%",
"!", "~", "UNARY_PLUS", "UNARY_MINUS",
"++", "--",
".",
"[", "]",
"{", "}",
"(", ")",
// Nonterminal tree node type codes.
"SCRIPT", "BLOCK", "LABEL", "FOR_IN", "CALL", "NEW_WITH_ARGS", "INDEX",
"ARRAY_INIT", "OBJECT_INIT", "PROPERTY_INIT", "GETTER", "SETTER",
"GROUP", "LIST", "LET_BLOCK", "ARRAY_COMP", "GENERATOR", "COMP_TAIL",
// Terminals.
"IDENTIFIER", "NUMBER", "STRING", "REGEXP",
// Keywords.
"break",
"case", "catch", "const", "continue",
"debugger", "default", "delete", "do",
"else",
"false", "finally", "for", "function",
"if", "in", "instanceof",
"let",
"new", "null",
"return",
"switch",
"this", "throw", "true", "try", "typeof",
"var", "void",
"yield",
"while", "with",
];
var statementStartTokens = [
"break",
"const", "continue",
"debugger", "do",
"for",
"if",
"return",
"switch",
"throw", "try",
"var",
"yield",
"while", "with",
];
// Operator and punctuator mapping from token to tree node type name.
// NB: because the lexer doesn't backtrack, all token prefixes must themselves
// be valid tokens (e.g. !== is acceptable because its prefixes are the valid
// tokens != and !).
var opTypeNames = {
'\n': "NEWLINE",
';': "SEMICOLON",
',': "COMMA",
'?': "HOOK",
':': "COLON",
'||': "OR",
'&&': "AND",
'|': "BITWISE_OR",
'^': "BITWISE_XOR",
'&': "BITWISE_AND",
'===': "STRICT_EQ",
'==': "EQ",
'=': "ASSIGN",
'!==': "STRICT_NE",
'!=': "NE",
'<<': "LSH",
'<=': "LE",
'<': "LT",
'>>>': "URSH",
'>>': "RSH",
'>=': "GE",
'>': "GT",
'++': "INCREMENT",
'--': "DECREMENT",
'+': "PLUS",
'-': "MINUS",
'*': "MUL",
'/': "DIV",
'%': "MOD",
'!': "NOT",
'~': "BITWISE_NOT",
'.': "DOT",
'[': "LEFT_BRACKET",
']': "RIGHT_BRACKET",
'{': "LEFT_CURLY",
'}': "RIGHT_CURLY",
'(': "LEFT_PAREN",
')': "RIGHT_PAREN"
};
// Hash of keyword identifier to tokens index. NB: we must null __proto__ to
// avoid toString, etc. namespace pollution.
var keywords = {__proto__: null};
// Define const END, etc., based on the token names. Also map name to index.
var tokenIds = {};
// Building up a string to be eval'd in different contexts.
var consts = "const ";
for (var i = 0, j = tokens.length; i < j; i++) {
if (i > 0)
consts += ", ";
var t = tokens[i];
var name;
if (/^[a-z]/.test(t)) {
name = t.toUpperCase();
keywords[t] = i;
} else {
name = (/^\W/.test(t) ? opTypeNames[t] : t);
}
consts += name + " = " + i;
tokenIds[name] = i;
tokens[t] = i;
}
consts += ";";
var isStatementStartCode = {__proto__: null};
for (i = 0, j = statementStartTokens.length; i < j; i++)
isStatementStartCode[keywords[statementStartTokens[i]]] = true;
// Map assignment operators to their indexes in the tokens array.
var assignOps = ['|', '^', '&', '<<', '>>', '>>>', '+', '-', '*', '/', '%'];
for (i = 0, j = assignOps.length; i < j; i++) {
t = assignOps[i];
assignOps[t] = tokens[t];
}
function defineGetter(obj, prop, fn, dontDelete, dontEnum) {
Object.defineProperty(obj, prop,
{ get: fn, configurable: !dontDelete, enumerable: !dontEnum });
}
function defineProperty(obj, prop, val, dontDelete, readOnly, dontEnum) {
Object.defineProperty(obj, prop,
{ value: val, writable: !readOnly, configurable: !dontDelete,
enumerable: !dontEnum });
}
// Returns true if fn is a native function. (Note: SpiderMonkey specific.)
function isNativeCode(fn) {
// Relies on the toString method to identify native code.
return ((typeof fn) === "function") && fn.toString().match(/\[native code\]/);
}
function getPropertyDescriptor(obj, name) {
while (obj) {
if (({}).hasOwnProperty.call(obj, name))
return Object.getOwnPropertyDescriptor(obj, name);
obj = Object.getPrototypeOf(obj);
}
}
function getOwnProperties(obj) {
var map = {};
for (var name in Object.getOwnPropertyNames(obj))
map[name] = Object.getOwnPropertyDescriptor(obj, name);
return map;
}
function makePassthruHandler(obj) {
// Handler copied from
// http://wiki.ecmascript.org/doku.php?id=harmony:proxies&s=proxy%20object#examplea_no-op_forwarding_proxy
return {
getOwnPropertyDescriptor: function(name) {
var desc = Object.getOwnPropertyDescriptor(obj, name);
// a trapping proxy's properties must always be configurable
desc.configurable = true;
return desc;
var narcissus = {
options: {
version: 185,
// Global variables to hide from the interpreter
hiddenHostGlobals: { Narcissus: true },
// Desugar SpiderMonkey language extensions?
desugarExtensions: false
},
getPropertyDescriptor: function(name) {
var desc = getPropertyDescriptor(obj, name);
// a trapping proxy's properties must always be configurable
desc.configurable = true;
return desc;
},
getOwnPropertyNames: function() {
return Object.getOwnPropertyNames(obj);
},
defineProperty: function(name, desc) {
Object.defineProperty(obj, name, desc);
},
"delete": function(name) { return delete obj[name]; },
fix: function() {
if (Object.isFrozen(obj)) {
return getOwnProperties(obj);
hostSupportsEvalConst: (function() {
try {
return eval("(function(s) { eval(s); return x })('const x = true;')");
} catch (e) {
return false;
}
// As long as obj is not frozen, the proxy won't allow itself to be fixed.
return undefined; // will cause a TypeError to be thrown
},
has: function(name) { return name in obj; },
hasOwn: function(name) { return ({}).hasOwnProperty.call(obj, name); },
get: function(receiver, name) { return obj[name]; },
// bad behavior when set fails in non-strict mode
set: function(receiver, name, val) { obj[name] = val; return true; },
enumerate: function() {
var result = [];
for (name in obj) { result.push(name); };
return result;
},
keys: function() { return Object.keys(obj); }
})(),
hostGlobal: this
};
}
Narcissus = narcissus;
// default function used when looking for a property in the global object
function noPropFound() { return undefined; }
var tokens = [
// End of source.
"END",
var hasOwnProperty = ({}).hasOwnProperty;
// Operators and punctuators. Some pair-wise order matters, e.g. (+, -)
// and (UNARY_PLUS, UNARY_MINUS).
"\n", ";",
",",
"=",
"?", ":", "CONDITIONAL",
"||",
"&&",
"|",
"^",
"&",
"==", "!=", "===", "!==",
"<", "<=", ">=", ">",
"<<", ">>", ">>>",
"+", "-",
"*", "/", "%",
"!", "~", "UNARY_PLUS", "UNARY_MINUS",
"++", "--",
".",
"[", "]",
"{", "}",
"(", ")",
function StringMap() {
this.table = Object.create(null, {});
this.size = 0;
}
// Nonterminal tree node type codes.
"SCRIPT", "BLOCK", "LABEL", "FOR_IN", "CALL", "NEW_WITH_ARGS", "INDEX",
"ARRAY_INIT", "OBJECT_INIT", "PROPERTY_INIT", "GETTER", "SETTER",
"GROUP", "LIST", "LET_BLOCK", "ARRAY_COMP", "GENERATOR", "COMP_TAIL",
StringMap.prototype = {
has: function(x) { return hasOwnProperty.call(this.table, x); },
set: function(x, v) {
if (!hasOwnProperty.call(this.table, x))
this.size++;
this.table[x] = v;
},
get: function(x) { return this.table[x]; },
getDef: function(x, thunk) {
if (!hasOwnProperty.call(this.table, x)) {
this.size++;
this.table[x] = thunk();
// Terminals.
"IDENTIFIER", "NUMBER", "STRING", "REGEXP",
// Keywords.
"break",
"case", "catch", "const", "continue",
"debugger", "default", "delete", "do",
"else", "export",
"false", "finally", "for", "function",
"if", "import", "in", "instanceof",
"let", "module",
"new", "null",
"return",
"switch",
"this", "throw", "true", "try", "typeof",
"var", "void",
"yield",
"while", "with",
];
var statementStartTokens = [
"break",
"const", "continue",
"debugger", "do",
"for",
"if",
"return",
"switch",
"throw", "try",
"var",
"yield",
"while", "with",
];
// Whitespace characters (see ECMA-262 7.2)
var whitespaceChars = [
// normal whitespace:
"\u0009", "\u000B", "\u000C", "\u0020", "\u00A0", "\uFEFF",
// high-Unicode whitespace:
"\u1680", "\u180E",
"\u2000", "\u2001", "\u2002", "\u2003", "\u2004", "\u2005", "\u2006",
"\u2007", "\u2008", "\u2009", "\u200A",
"\u202F", "\u205F", "\u3000"
];
var whitespace = {};
for (var i = 0; i < whitespaceChars.length; i++) {
whitespace[whitespaceChars[i]] = true;
}
// Operator and punctuator mapping from token to tree node type name.
// NB: because the lexer doesn't backtrack, all token prefixes must themselves
// be valid tokens (e.g. !== is acceptable because its prefixes are the valid
// tokens != and !).
var opTypeNames = {
'\n': "NEWLINE",
';': "SEMICOLON",
',': "COMMA",
'?': "HOOK",
':': "COLON",
'||': "OR",
'&&': "AND",
'|': "BITWISE_OR",
'^': "BITWISE_XOR",
'&': "BITWISE_AND",
'===': "STRICT_EQ",
'==': "EQ",
'=': "ASSIGN",
'!==': "STRICT_NE",
'!=': "NE",
'<<': "LSH",
'<=': "LE",
'<': "LT",
'>>>': "URSH",
'>>': "RSH",
'>=': "GE",
'>': "GT",
'++': "INCREMENT",
'--': "DECREMENT",
'+': "PLUS",
'-': "MINUS",
'*': "MUL",
'/': "DIV",
'%': "MOD",
'!': "NOT",
'~': "BITWISE_NOT",
'.': "DOT",
'[': "LEFT_BRACKET",
']': "RIGHT_BRACKET",
'{': "LEFT_CURLY",
'}': "RIGHT_CURLY",
'(': "LEFT_PAREN",
')': "RIGHT_PAREN"
};
// Hash of keyword identifier to tokens index. NB: we must null __proto__ to
// avoid toString, etc. namespace pollution.
var keywords = {__proto__: null};
// Define const END, etc., based on the token names. Also map name to index.
var tokenIds = {};
// Building up a string to be eval'd in different contexts.
var consts = Narcissus.hostSupportsEvalConst ? "const " : "var ";
for (var i = 0, j = tokens.length; i < j; i++) {
if (i > 0)
consts += ", ";
var t = tokens[i];
var name;
if (/^[a-z]/.test(t)) {
name = t.toUpperCase();
keywords[t] = i;
} else {
name = (/^\W/.test(t) ? opTypeNames[t] : t);
}
return this.table[x];
},
forEach: function(f) {
var table = this.table;
for (var key in table)
f.call(this, key, table[key]);
},
toString: function() { return "[object StringMap]" }
};
consts += name + " = " + i;
tokenIds[name] = i;
tokens[t] = i;
}
consts += ";";
// non-destructive stack
function Stack(elts) {
this.elts = elts || null;
}
var isStatementStartCode = {__proto__: null};
for (i = 0, j = statementStartTokens.length; i < j; i++)
isStatementStartCode[keywords[statementStartTokens[i]]] = true;
Stack.prototype = {
push: function(x) {
return new Stack({ top: x, rest: this.elts });
},
top: function() {
if (!this.elts)
throw new Error("empty stack");
return this.elts.top;
},
isEmpty: function() {
return this.top === null;
},
find: function(test) {
for (var elts = this.elts; elts; elts = elts.rest) {
if (test(elts.top))
return elts.top;
}
return null;
},
has: function(x) {
return Boolean(this.find(function(elt) { return elt === x }));
},
forEach: function(f) {
for (var elts = this.elts; elts; elts = elts.rest) {
f(elts.top);
// Map assignment operators to their indexes in the tokens array.
var assignOps = ['|', '^', '&', '<<', '>>', '>>>', '+', '-', '*', '/', '%'];
for (i = 0, j = assignOps.length; i < j; i++) {
t = assignOps[i];
assignOps[t] = tokens[t];
}
function defineGetter(obj, prop, fn, dontDelete, dontEnum) {
Object.defineProperty(obj, prop,
{ get: fn, configurable: !dontDelete, enumerable: !dontEnum });
}
function defineGetterSetter(obj, prop, getter, setter, dontDelete, dontEnum) {
Object.defineProperty(obj, prop, {
get: getter,
set: setter,
configurable: !dontDelete,
enumerable: !dontEnum
});
}
function defineMemoGetter(obj, prop, fn, dontDelete, dontEnum) {
Object.defineProperty(obj, prop, {
get: function() {
var val = fn();
defineProperty(obj, prop, val, dontDelete, true, dontEnum);
return val;
},
configurable: true,
enumerable: !dontEnum
});
}
function defineProperty(obj, prop, val, dontDelete, readOnly, dontEnum) {
Object.defineProperty(obj, prop,
{ value: val, writable: !readOnly, configurable: !dontDelete,
enumerable: !dontEnum });
}
// Returns true if fn is a native function. (Note: SpiderMonkey specific.)
function isNativeCode(fn) {
// Relies on the toString method to identify native code.
return ((typeof fn) === "function") && fn.toString().match(/\[native code\]/);
}
function getPropertyDescriptor(obj, name) {
while (obj) {
if (({}).hasOwnProperty.call(obj, name))
return Object.getOwnPropertyDescriptor(obj, name);
obj = Object.getPrototypeOf(obj);
}
}
};
exports.tokens = tokens;
exports.opTypeNames = opTypeNames;
exports.keywords = keywords;
exports.isStatementStartCode = isStatementStartCode;
exports.tokenIds = tokenIds;
exports.consts = consts;
exports.assignOps = assignOps;
exports.defineGetter = defineGetter;
exports.defineProperty = defineProperty;
exports.isNativeCode = isNativeCode;
exports.makePassthruHandler = makePassthruHandler;
exports.noPropFound = noPropFound;
exports.StringMap = StringMap;
exports.Stack = Stack;
function getPropertyNames(obj) {
var table = Object.create(null, {});
while (obj) {
var names = Object.getOwnPropertyNames(obj);
for (var i = 0, n = names.length; i < n; i++)
table[names[i]] = true;
obj = Object.getPrototypeOf(obj);
}
return Object.keys(table);
}
function getOwnProperties(obj) {
var map = {};
for (var name in Object.getOwnPropertyNames(obj))
map[name] = Object.getOwnPropertyDescriptor(obj, name);
return map;
}
function blacklistHandler(target, blacklist) {
var mask = Object.create(null, {});
var redirect = StringMap.create(blacklist).mapObject(function(name) { return mask; });
return mixinHandler(redirect, target);
}
function whitelistHandler(target, whitelist) {
var catchall = Object.create(null, {});
var redirect = StringMap.create(whitelist).mapObject(function(name) { return target; });
return mixinHandler(redirect, catchall);
}
function mirrorHandler(target, writable) {
var handler = makePassthruHandler(target);
var defineProperty = handler.defineProperty;
handler.defineProperty = function(name, desc) {
if (!desc.enumerable)
throw new Error("mirror property must be enumerable");
if (!desc.configurable)
throw new Error("mirror property must be configurable");
if (desc.writable !== writable)
throw new Error("mirror property must " + (writable ? "" : "not ") + "be writable");
defineProperty(name, desc);
};
handler.fix = function() { };
handler.getOwnPropertyDescriptor = handler.getPropertyDescriptor;
handler.getOwnPropertyNames = getPropertyNames.bind(handler, target);
handler.keys = handler.enumerate;
handler["delete"] = function() { return false; };
handler.hasOwn = handler.has;
return handler;
}
/*
* Mixin proxies break the single-inheritance model of prototypes, so
* the handler treats all properties as own-properties:
*
* X
* |
* +------------+------------+
* | O |
* | | |
* | O O O |
* | | | | |
* | O O O O |
* | | | | | |
* | O O O O O |
* | | | | | | |
* +-(*)--(w)--(x)--(y)--(z)-+
*/
function mixinHandler(redirect, catchall) {
function targetFor(name) {
return hasOwn(redirect, name) ? redirect[name] : catchall;
}
function getMuxPropertyDescriptor(name) {
var desc = getPropertyDescriptor(targetFor(name), name);
if (desc)
desc.configurable = true;
return desc;
}
function getMuxPropertyNames() {
var names1 = Object.getOwnPropertyNames(redirect).filter(function(name) {
return name in redirect[name];
});
var names2 = getPropertyNames(catchall).filter(function(name) {
return !hasOwn(redirect, name);
});
return names1.concat(names2);
}
function enumerateMux() {
var result = Object.getOwnPropertyNames(redirect).filter(function(name) {
return name in redirect[name];
});
for (name in catchall) {
if (!hasOwn(redirect, name))
result.push(name);
};
return result;
}
function hasMux(name) {
return name in targetFor(name);
}
return {
getOwnPropertyDescriptor: getMuxPropertyDescriptor,
getPropertyDescriptor: getMuxPropertyDescriptor,
getOwnPropertyNames: getMuxPropertyNames,
defineProperty: function(name, desc) {
Object.defineProperty(targetFor(name), name, desc);
},
"delete": function(name) {
var target = targetFor(name);
return delete target[name];
},
// FIXME: ha ha ha
fix: function() { },
has: hasMux,
hasOwn: hasMux,
get: function(receiver, name) {
var target = targetFor(name);
return target[name];
},
set: function(receiver, name, val) {
var target = targetFor(name);
target[name] = val;
return true;
},
enumerate: enumerateMux,
keys: enumerateMux
};
}
function makePassthruHandler(obj) {
// Handler copied from
// http://wiki.ecmascript.org/doku.php?id=harmony:proxies&s=proxy%20object#examplea_no-op_forwarding_proxy
return {
getOwnPropertyDescriptor: function(name) {
var desc = Object.getOwnPropertyDescriptor(obj, name);
// a trapping proxy's properties must always be configurable
desc.configurable = true;
return desc;
},
getPropertyDescriptor: function(name) {
var desc = getPropertyDescriptor(obj, name);
// a trapping proxy's properties must always be configurable
desc.configurable = true;
return desc;
},
getOwnPropertyNames: function() {
return Object.getOwnPropertyNames(obj);
},
defineProperty: function(name, desc) {
Object.defineProperty(obj, name, desc);
},
"delete": function(name) { return delete obj[name]; },
fix: function() {
if (Object.isFrozen(obj)) {
return getOwnProperties(obj);
}
// As long as obj is not frozen, the proxy won't allow itself to be fixed.
return undefined; // will cause a TypeError to be thrown
},
has: function(name) { return name in obj; },
hasOwn: function(name) { return ({}).hasOwnProperty.call(obj, name); },
get: function(receiver, name) { return obj[name]; },
// bad behavior when set fails in non-strict mode
set: function(receiver, name, val) { obj[name] = val; return true; },
enumerate: function() {
var result = [];
for (name in obj) { result.push(name); };
return result;
},
keys: function() { return Object.keys(obj); }
};
}
var hasOwnProperty = ({}).hasOwnProperty;
function hasOwn(obj, name) {
return hasOwnProperty.call(obj, name);
}
function StringMap(table, size) {
this.table = table || Object.create(null, {});
this.size = size || 0;
}
StringMap.create = function(table) {
var init = Object.create(null, {});
var size = 0;
var names = Object.getOwnPropertyNames(table);
for (var i = 0, n = names.length; i < n; i++) {
var name = names[i];
init[name] = table[name];
size++;
}
return new StringMap(init, size);
};
StringMap.prototype = {
has: function(x) { return hasOwnProperty.call(this.table, x); },
set: function(x, v) {
if (!hasOwnProperty.call(this.table, x))
this.size++;
this.table[x] = v;
},
get: function(x) { return this.table[x]; },
getDef: function(x, thunk) {
if (!hasOwnProperty.call(this.table, x)) {
this.size++;
this.table[x] = thunk();
}
return this.table[x];
},
forEach: function(f) {
var table = this.table;
for (var key in table)
f.call(this, key, table[key]);
},
map: function(f) {
var table1 = this.table;
var table2 = Object.create(null, {});
this.forEach(function(key, val) {
table2[key] = f.call(this, val, key);
});
return new StringMap(table2, this.size);
},
mapObject: function(f) {
var table1 = this.table;
var table2 = Object.create(null, {});
this.forEach(function(key, val) {
table2[key] = f.call(this, val, key);
});
return table2;
},
toObject: function() {
return this.mapObject(function(val) { return val; });
},
choose: function() {
return Object.getOwnPropertyNames(this.table)[0];
},
remove: function(x) {
if (hasOwnProperty.call(this.table, x)) {
this.size--;
delete this.table[x];
}
},
copy: function() {
var table = Object.create(null, {});
for (var key in this.table)
table[key] = this.table[key];
return new StringMap(table, this.size);
},
toString: function() { return "[object StringMap]" }
};
// an object-key table with poor asymptotics (replace with WeakMap when possible)
function ObjectMap(array) {
this.array = array || [];
}
function searchMap(map, key, found, notFound) {
var a = map.array;
for (var i = 0, n = a.length; i < n; i++) {
var pair = a[i];
if (pair.key === key)
return found(pair, i);
}
return notFound();
}
ObjectMap.prototype = {
has: function(x) {
return searchMap(this, x, function() { return true }, function() { return false });
},
set: function(x, v) {
var a = this.array;
searchMap(this, x,
function(pair) { pair.value = v },
function() { a.push({ key: x, value: v }) });
},
get: function(x) {
return searchMap(this, x,
function(pair) { return pair.value },
function() { return null });
},
getDef: function(x, thunk) {
var a = this.array;
return searchMap(this, x,
function(pair) { return pair.value },
function() {
var v = thunk();
a.push({ key: x, value: v });
return v;
});
},
forEach: function(f) {
var a = this.array;
for (var i = 0, n = a.length; i < n; i++) {
var pair = a[i];
f.call(this, pair.key, pair.value);
}
},
choose: function() {
return this.array[0].key;
},
get size() {
return this.array.length;
},
remove: function(x) {
var a = this.array;
searchMap(this, x,
function(pair, i) { a.splice(i, 1) },
function() { });
},
copy: function() {
return new ObjectMap(this.array.map(function(pair) {
return { key: pair.key, value: pair.value }
}));
},
clear: function() {
this.array = [];
},
toString: function() { return "[object ObjectMap]" }
};
// non-destructive stack
function Stack(elts) {
this.elts = elts || null;
}
Stack.prototype = {
push: function(x) {
return new Stack({ top: x, rest: this.elts });
},
top: function() {
if (!this.elts)
throw new Error("empty stack");
return this.elts.top;
},
isEmpty: function() {
return this.top === null;
},
find: function(test) {
for (var elts = this.elts; elts; elts = elts.rest) {
if (test(elts.top))
return elts.top;
}
return null;
},
has: function(x) {
return Boolean(this.find(function(elt) { return elt === x }));
},
forEach: function(f) {
for (var elts = this.elts; elts; elts = elts.rest) {
f(elts.top);
}
}
};
module.exports = {
tokens: tokens,
whitespace: whitespace,
opTypeNames: opTypeNames,
keywords: keywords,
isStatementStartCode: isStatementStartCode,
tokenIds: tokenIds,
consts: consts,
assignOps: assignOps,
defineGetter: defineGetter,
defineGetterSetter: defineGetterSetter,
defineMemoGetter: defineMemoGetter,
defineProperty: defineProperty,
isNativeCode: isNativeCode,
mirrorHandler: mirrorHandler,
mixinHandler: mixinHandler,
whitelistHandler: whitelistHandler,
blacklistHandler: blacklistHandler,
makePassthruHandler: makePassthruHandler,
StringMap: StringMap,
ObjectMap: ObjectMap,
Stack: Stack
};
});

View file

@ -47,416 +47,503 @@
* Lexical scanner.
*/
define(function(require, exports, module) {
define(function(require, exports, module) {
var definitions = require("ace/narcissus/jsdefs");
var definitions = require("ace/narcissus/jsdefs");
// Set constants in the local scope.
eval(definitions.consts);
// Set constants in the local scope.
eval(definitions.consts);
// Build up a trie of operator tokens.
var opTokens = {};
for (var op in definitions.opTypeNames) {
if (op === '\n' || op === '.')
continue;
// Banned keywords by language version
const blackLists = { 160: {}, 185: {}, harmony: {} };
blackLists[160][LET] = true;
blackLists[160][MODULE] = true;
blackLists[160][YIELD] = true;
blackLists[185][MODULE] = true;
var node = opTokens;
for (var i = 0; i < op.length; i++) {
var ch = op[i];
if (!(ch in node))
node[ch] = {};
node = node[ch];
node.op = op;
// Build up a trie of operator tokens.
var opTokens = {};
for (var op in definitions.opTypeNames) {
if (op === '\n' || op === '.')
continue;
var node = opTokens;
for (var i = 0; i < op.length; i++) {
var ch = op[i];
if (!(ch in node))
node[ch] = {};
node = node[ch];
node.op = op;
}
}
}
/*
* Tokenizer :: (source, filename, line number) -> Tokenizer
*/
function Tokenizer(s, f, l) {
this.cursor = 0;
this.source = String(s);
this.tokens = [];
this.tokenIndex = 0;
this.lookahead = 0;
this.scanNewlines = false;
this.unexpectedEOF = false;
this.filename = f || "";
this.lineno = l || 1;
}
Tokenizer.prototype = {
get done() {
// We need to set scanOperand to true here because the first thing
// might be a regexp.
return this.peek(true) === END;
},
get token() {
return this.tokens[this.tokenIndex];
},
match: function (tt, scanOperand) {
return this.get(scanOperand) === tt || this.unget();
},
mustMatch: function (tt) {
if (!this.match(tt)) {
throw this.newSyntaxError("Missing " +
definitions.tokens[tt].toLowerCase());
}
return this.token;
},
peek: function (scanOperand) {
var tt, next;
if (this.lookahead) {
next = this.tokens[(this.tokenIndex + this.lookahead) & 3];
tt = (this.scanNewlines && next.lineno !== this.lineno)
? NEWLINE
: next.type;
} else {
tt = this.get(scanOperand);
this.unget();
}
return tt;
},
peekOnSameLine: function (scanOperand) {
this.scanNewlines = true;
var tt = this.peek(scanOperand);
/*
* Tokenizer :: (source, filename, line number) -> Tokenizer
*/
function Tokenizer(s, f, l) {
this.cursor = 0;
this.source = String(s);
this.tokens = [];
this.tokenIndex = 0;
this.lookahead = 0;
this.scanNewlines = false;
return tt;
},
this.unexpectedEOF = false;
this.filename = f || "";
this.lineno = l || 1;
this.blackList = blackLists[Narcissus.options.version];
this.blockComments = null;
}
// Eat comments and whitespace.
skip: function () {
var input = this.source;
for (;;) {
var ch = input[this.cursor++];
var next = input[this.cursor];
if (ch === '\n' && !this.scanNewlines) {
this.lineno++;
} else if (ch === '/' && next === '*') {
this.cursor++;
for (;;) {
ch = input[this.cursor++];
if (ch === undefined)
throw this.newSyntaxError("Unterminated comment");
Tokenizer.prototype = {
get done() {
// We need to set scanOperand to true here because the first thing
// might be a regexp.
return this.peek(true) === END;
},
if (ch === '*') {
get token() {
return this.tokens[this.tokenIndex];
},
match: function (tt, scanOperand) {
return this.get(scanOperand) === tt || this.unget();
},
mustMatch: function (tt) {
if (!this.match(tt)) {
throw this.newSyntaxError("Missing " +
definitions.tokens[tt].toLowerCase());
}
return this.token;
},
peek: function (scanOperand) {
var tt, next;
if (this.lookahead) {
next = this.tokens[(this.tokenIndex + this.lookahead) & 3];
tt = (this.scanNewlines && next.lineno !== this.lineno)
? NEWLINE
: next.type;
} else {
tt = this.get(scanOperand);
this.unget();
}
return tt;
},
peekOnSameLine: function (scanOperand) {
this.scanNewlines = true;
var tt = this.peek(scanOperand);
this.scanNewlines = false;
return tt;
},
lastBlockComment: function() {
var length = this.blockComments.length;
return length ? this.blockComments[length - 1] : null;
},
// Eat comments and whitespace.
skip: function () {
var input = this.source;
this.blockComments = [];
for (;;) {
var ch = input[this.cursor++];
var next = input[this.cursor];
// handle \r, \r\n and (always preferable) \n
if (ch === '\r') {
// if the next character is \n, we don't care about this at all
if (next === '\n') continue;
// otherwise, we want to consider this as a newline
ch = '\n';
}
if (ch === '\n' && !this.scanNewlines) {
this.lineno++;
} else if (ch === '/' && next === '*') {
var commentStart = ++this.cursor;
for (;;) {
ch = input[this.cursor++];
if (ch === undefined)
throw this.newSyntaxError("Unterminated comment");
if (ch === '*') {
next = input[this.cursor];
if (next === '/') {
var commentEnd = this.cursor - 1;
this.cursor++;
break;
}
} else if (ch === '\n') {
this.lineno++;
}
}
this.blockComments.push(input.substring(commentStart, commentEnd));
} else if (ch === '/' && next === '/') {
this.cursor++;
for (;;) {
ch = input[this.cursor++];
next = input[this.cursor];
if (next === '/') {
this.cursor++;
if (ch === undefined)
return;
if (ch === '\r') {
// check for \r\n
if (next !== '\n') ch = '\n';
}
if (ch === '\n') {
if (this.scanNewlines) {
this.cursor--;
} else {
this.lineno++;
}
break;
}
} else if (ch === '\n') {
this.lineno++;
}
} else if (!(ch in definitions.whitespace)) {
this.cursor--;
return;
}
} else if (ch === '/' && next === '/') {
}
},
// Lex the exponential part of a number, if present. Return true iff an
// exponential part was found.
lexExponent: function() {
var input = this.source;
var next = input[this.cursor];
if (next === 'e' || next === 'E') {
this.cursor++;
for (;;) {
ch = input[this.cursor++];
if (ch === '+' || ch === '-')
ch = input[this.cursor++];
if (ch === undefined)
return;
if (ch === '\n') {
this.lineno++;
break;
}
}
} else if (ch !== ' ' && ch !== '\t') {
this.cursor--;
return;
}
}
},
if (ch < '0' || ch > '9')
throw this.newSyntaxError("Missing exponent");
// Lex the exponential part of a number, if present. Return true iff an
// exponential part was found.
lexExponent: function() {
var input = this.source;
var next = input[this.cursor];
if (next === 'e' || next === 'E') {
this.cursor++;
ch = input[this.cursor++];
if (ch === '+' || ch === '-')
ch = input[this.cursor++];
if (ch < '0' || ch > '9')
throw this.newSyntaxError("Missing exponent");
do {
ch = input[this.cursor++];
} while (ch >= '0' && ch <= '9');
this.cursor--;
return true;
}
return false;
},
lexZeroNumber: function (ch) {
var token = this.token, input = this.source;
token.type = NUMBER;
ch = input[this.cursor++];
if (ch === '.') {
do {
ch = input[this.cursor++];
} while (ch >= '0' && ch <= '9');
this.cursor--;
this.lexExponent();
token.value = parseFloat(token.start, this.cursor);
} else if (ch === 'x' || ch === 'X') {
do {
ch = input[this.cursor++];
} while ((ch >= '0' && ch <= '9') || (ch >= 'a' && ch <= 'f') ||
(ch >= 'A' && ch <= 'F'));
this.cursor--;
token.value = parseInt(input.substring(token.start, this.cursor));
} else if (ch >= '0' && ch <= '7') {
do {
ch = input[this.cursor++];
} while (ch >= '0' && ch <= '7');
this.cursor--;
token.value = parseInt(input.substring(token.start, this.cursor));
} else {
this.cursor--;
this.lexExponent(); // 0E1, &c.
token.value = 0;
}
},
lexNumber: function (ch) {
var token = this.token, input = this.source;
token.type = NUMBER;
var floating = false;
do {
ch = input[this.cursor++];
if (ch === '.' && !floating) {
floating = true;
ch = input[this.cursor++];
}
} while (ch >= '0' && ch <= '9');
this.cursor--;
var exponent = this.lexExponent();
floating = floating || exponent;
var str = input.substring(token.start, this.cursor);
token.value = floating ? parseFloat(str) : parseInt(str);
},
lexDot: function (ch) {
var token = this.token, input = this.source;
var next = input[this.cursor];
if (next >= '0' && next <= '9') {
do {
ch = input[this.cursor++];
} while (ch >= '0' && ch <= '9');
this.cursor--;
this.lexExponent();
token.type = NUMBER;
token.value = parseFloat(token.start, this.cursor);
} else {
token.type = DOT;
token.assignOp = null;
token.value = '.';
}
},
lexString: function (ch) {
var token = this.token, input = this.source;
token.type = STRING;
var hasEscapes = false;
var delim = ch;
if (input.length <= this.cursor)
throw this.newSyntaxError("Unterminated string literal");
while ((ch = input[this.cursor++]) !== delim) {
if (this.cursor == input.length)
throw this.newSyntaxError("Unterminated string literal");
if (ch === '\\') {
hasEscapes = true;
if (++this.cursor == input.length)
throw this.newSyntaxError("Unterminated string literal");
}
}
token.value = hasEscapes
? eval(input.substring(token.start, this.cursor))
: input.substring(token.start + 1, this.cursor - 1);
},
lexRegExp: function (ch) {
var token = this.token, input = this.source;
token.type = REGEXP;
do {
ch = input[this.cursor++];
if (ch === '\\') {
this.cursor++;
} else if (ch === '[') {
do {
if (ch === undefined)
throw this.newSyntaxError("Unterminated character class");
if (ch === '\\')
this.cursor++;
ch = input[this.cursor++];
} while (ch !== ']');
} else if (ch === undefined) {
throw this.newSyntaxError("Unterminated regex");
} while (ch >= '0' && ch <= '9');
this.cursor--;
return true;
}
} while (ch !== '/');
do {
return false;
},
lexZeroNumber: function (ch) {
var token = this.token, input = this.source;
token.type = NUMBER;
ch = input[this.cursor++];
} while (ch >= 'a' && ch <= 'z');
if (ch === '.') {
do {
ch = input[this.cursor++];
} while (ch >= '0' && ch <= '9');
this.cursor--;
this.cursor--;
this.lexExponent();
token.value = parseFloat(token.start, this.cursor);
} else if (ch === 'x' || ch === 'X') {
do {
ch = input[this.cursor++];
} while ((ch >= '0' && ch <= '9') || (ch >= 'a' && ch <= 'f') ||
(ch >= 'A' && ch <= 'F'));
this.cursor--;
token.value = eval(input.substring(token.start, this.cursor));
},
token.value = parseInt(input.substring(token.start, this.cursor));
} else if (ch >= '0' && ch <= '7') {
do {
ch = input[this.cursor++];
} while (ch >= '0' && ch <= '7');
this.cursor--;
lexOp: function (ch) {
var token = this.token, input = this.source;
token.value = parseInt(input.substring(token.start, this.cursor));
} else {
this.cursor--;
this.lexExponent(); // 0E1, &c.
token.value = 0;
}
},
// A bit ugly, but it seems wasteful to write a trie lookup routine
// for only 3 characters...
var node = opTokens[ch];
var next = input[this.cursor];
if (next in node) {
node = node[next];
this.cursor++;
next = input[this.cursor];
lexNumber: function (ch) {
var token = this.token, input = this.source;
token.type = NUMBER;
var floating = false;
do {
ch = input[this.cursor++];
if (ch === '.' && !floating) {
floating = true;
ch = input[this.cursor++];
}
} while (ch >= '0' && ch <= '9');
this.cursor--;
var exponent = this.lexExponent();
floating = floating || exponent;
var str = input.substring(token.start, this.cursor);
token.value = floating ? parseFloat(str) : parseInt(str);
},
lexDot: function (ch) {
var token = this.token, input = this.source;
var next = input[this.cursor];
if (next >= '0' && next <= '9') {
do {
ch = input[this.cursor++];
} while (ch >= '0' && ch <= '9');
this.cursor--;
this.lexExponent();
token.type = NUMBER;
token.value = parseFloat(token.start, this.cursor);
} else {
token.type = DOT;
token.assignOp = null;
token.value = '.';
}
},
lexString: function (ch) {
var token = this.token, input = this.source;
token.type = STRING;
var hasEscapes = false;
var delim = ch;
if (input.length <= this.cursor)
throw this.newSyntaxError("Unterminated string literal");
while ((ch = input[this.cursor++]) !== delim) {
if (this.cursor == input.length)
throw this.newSyntaxError("Unterminated string literal");
if (ch === '\\') {
hasEscapes = true;
if (++this.cursor == input.length)
throw this.newSyntaxError("Unterminated string literal");
}
}
token.value = hasEscapes
? eval(input.substring(token.start, this.cursor))
: input.substring(token.start + 1, this.cursor - 1);
},
lexRegExp: function (ch) {
var token = this.token, input = this.source;
token.type = REGEXP;
do {
ch = input[this.cursor++];
if (ch === '\\') {
this.cursor++;
} else if (ch === '[') {
do {
if (ch === undefined)
throw this.newSyntaxError("Unterminated character class");
if (ch === '\\')
this.cursor++;
ch = input[this.cursor++];
} while (ch !== ']');
} else if (ch === undefined) {
throw this.newSyntaxError("Unterminated regex");
}
} while (ch !== '/');
do {
ch = input[this.cursor++];
} while (ch >= 'a' && ch <= 'z');
this.cursor--;
token.value = eval(input.substring(token.start, this.cursor));
},
lexOp: function (ch) {
var token = this.token, input = this.source;
// A bit ugly, but it seems wasteful to write a trie lookup routine
// for only 3 characters...
var node = opTokens[ch];
var next = input[this.cursor];
if (next in node) {
node = node[next];
this.cursor++;
next = input[this.cursor];
if (next in node) {
node = node[next];
this.cursor++;
next = input[this.cursor];
}
}
}
var op = node.op;
if (definitions.assignOps[op] && input[this.cursor] === '=') {
this.cursor++;
token.type = ASSIGN;
token.assignOp = definitions.tokenIds[definitions.opTypeNames[op]];
op += '=';
} else {
token.type = definitions.tokenIds[definitions.opTypeNames[op]];
token.assignOp = null;
}
var op = node.op;
if (definitions.assignOps[op] && input[this.cursor] === '=') {
this.cursor++;
token.type = ASSIGN;
token.assignOp = definitions.tokenIds[definitions.opTypeNames[op]];
op += '=';
} else {
token.type = definitions.tokenIds[definitions.opTypeNames[op]];
token.assignOp = null;
}
token.value = op;
},
token.value = op;
},
// FIXME: Unicode escape sequences
// FIXME: Unicode identifiers
lexIdent: function (ch) {
var token = this.token, input = this.source;
// FIXME: Unicode escape sequences
lexIdent: function (ch) {
var token = this.token;
var id = ch;
do {
ch = input[this.cursor++];
} while ((ch >= 'a' && ch <= 'z') || (ch >= 'A' && ch <= 'Z') ||
(ch >= '0' && ch <= '9') || ch === '$' || ch === '_');
while ((ch = this.getValidIdentifierChar(false)) !== null) {
id += ch;
}
this.cursor--; // Put the non-word character back.
token.type = definitions.keywords[id] || IDENTIFIER;
if (token.type in this.blackList) {
// banned keyword, this is an identifier
token.type = IDENTIFIER;
}
token.value = id;
},
var id = input.substring(token.start, this.cursor);
token.type = definitions.keywords[id] || IDENTIFIER;
token.value = id;
},
/*
* Tokenizer.get :: void -> token type
*
* Consume input *only* if there is no lookahead.
* Dispatch to the appropriate lexing function depending on the input.
*/
get: function (scanOperand) {
var token;
while (this.lookahead) {
--this.lookahead;
this.tokenIndex = (this.tokenIndex + 1) & 3;
token = this.tokens[this.tokenIndex];
if (token.type !== NEWLINE || this.scanNewlines)
return token.type;
}
this.skip();
/*
* Tokenizer.get :: void -> token type
*
* Consume input *only* if there is no lookahead.
* Dispatch to the appropriate lexing function depending on the input.
*/
get: function (scanOperand) {
var token;
while (this.lookahead) {
--this.lookahead;
this.tokenIndex = (this.tokenIndex + 1) & 3;
token = this.tokens[this.tokenIndex];
if (token.type !== NEWLINE || this.scanNewlines)
return token.type;
}
if (!token)
this.tokens[this.tokenIndex] = token = {};
this.skip();
var input = this.source;
if (this.cursor >= input.length)
return token.type = END;
this.tokenIndex = (this.tokenIndex + 1) & 3;
token = this.tokens[this.tokenIndex];
if (!token)
this.tokens[this.tokenIndex] = token = {};
token.start = this.cursor;
token.lineno = this.lineno;
var input = this.source;
if (this.cursor === input.length)
return token.type = END;
var ich = this.getValidIdentifierChar(true);
var ch = (ich === null) ? input[this.cursor++] : null;
if (ich !== null) {
this.lexIdent(ich);
} else if (scanOperand && ch === '/') {
this.lexRegExp(ch);
} else if (ch in opTokens) {
this.lexOp(ch);
} else if (ch === '.') {
this.lexDot(ch);
} else if (ch >= '1' && ch <= '9') {
this.lexNumber(ch);
} else if (ch === '0') {
this.lexZeroNumber(ch);
} else if (ch === '"' || ch === "'") {
this.lexString(ch);
} else if (this.scanNewlines && (ch === '\n' || ch === '\r')) {
// if this was a \r, look for \r\n
if (ch === '\r' && input[this.cursor] === '\n') this.cursor++;
token.type = NEWLINE;
token.value = '\n';
this.lineno++;
} else {
throw this.newSyntaxError("Illegal token");
}
token.start = this.cursor;
token.lineno = this.lineno;
token.end = this.cursor;
return token.type;
},
var ch = input[this.cursor++];
if ((ch >= 'a' && ch <= 'z') || (ch >= 'A' && ch <= 'Z') || ch === '$' || ch === '_') {
this.lexIdent(ch);
} else if (scanOperand && ch === '/') {
this.lexRegExp(ch);
} else if (ch in opTokens) {
this.lexOp(ch);
} else if (ch === '.') {
this.lexDot(ch);
} else if (ch >= '1' && ch <= '9') {
this.lexNumber(ch);
} else if (ch === '0') {
this.lexZeroNumber(ch);
} else if (ch === '"' || ch === "'") {
this.lexString(ch);
} else if (this.scanNewlines && ch === '\n') {
token.type = NEWLINE;
token.value = '\n';
this.lineno++;
} else {
throw this.newSyntaxError("Illegal token");
}
/*
* Tokenizer.unget :: void -> undefined
*
* Match depends on unget returning undefined.
*/
unget: function () {
if (++this.lookahead === 4) throw "PANIC: too much lookahead!";
this.tokenIndex = (this.tokenIndex - 1) & 3;
},
token.end = this.cursor;
return token.type;
},
newSyntaxError: function (m) {
m = (this.filename ? this.filename + ":" : "") + this.lineno + ": " + m;
var e = new SyntaxError(m, this.filename, this.lineno);
e.source = this.source;
e.cursor = this.lookahead
? this.tokens[(this.tokenIndex + this.lookahead) & 3].start
: this.cursor;
return e;
},
/*
* Tokenizer.unget :: void -> undefined
*
* Match depends on unget returning undefined.
*/
unget: function () {
if (++this.lookahead === 4) throw "PANIC: too much lookahead!";
this.tokenIndex = (this.tokenIndex - 1) & 3;
},
/* Gets a single valid identifier char from the input stream, or null
* if there is none.
* Since JavaScript provides no convenient way to determine if a
* character is in a particular Unicode category, we use
* metacircularity to accomplish this (oh yeaaaah!) */
getValidIdentifierChar: function(first) {
var input = this.source;
if (this.cursor >= input.length) return null;
var ch = input[this.cursor];
newSyntaxError: function (m) {
var e = new SyntaxError(m, this.filename, this.lineno);
e.source = this.source;
e.lineno = this.lineno;
e.cursor = this.lookahead
? this.tokens[(this.tokenIndex + this.lookahead) & 3].start
: this.cursor;
return e;
},
};
// first check for \u escapes
if (ch === '\\' && input[this.cursor+1] === 'u') {
// get the character value
try {
ch = String.fromCharCode(parseInt(
input.substring(this.cursor + 2, this.cursor + 6),
16));
} catch (ex) {
return null;
}
this.cursor += 5;
}
exports.Tokenizer = Tokenizer;
// check directly for ASCII
if (ch <= "\u007F") {
if ((ch >= 'a' && ch <= 'z') || (ch >= 'A' && ch <= 'Z') || ch === '$' || ch === '_' ||
(!first && (ch >= '0' && ch <= '9'))) {
this.cursor++;
return ch;
}
return null;
}
// create an object to test this in
var x = {};
x["x"+ch] = true;
x[ch] = true;
// then use eval to determine if it's a valid character
var valid = false;
try {
valid = (Function("x", "return (x." + (first?"":"x") + ch + ");")(x) === true);
} catch (ex) {}
if (valid) this.cursor++;
return (valid ? ch : null);
},
};
});
return { Tokenizer: Tokenizer };
});

File diff suppressed because it is too large Load diff