1 /*
2  *  Copyright (C) 1999-2002 Harri Porten (porten@kde.org)
3  *  Copyright (C) 2001 Peter Kelly (pmk@post.com)
4  *  Copyright (C) 2003, 2004, 2005, 2006, 2007, 2008, 2009 Apple Inc. All rights reserved.
5  *  Copyright (C) 2007 Cameron Zwarich (cwzwarich@uwaterloo.ca)
6  *  Copyright (C) 2007 Maks Orlovich
7  *
8  *  This library is free software; you can redistribute it and/or
9  *  modify it under the terms of the GNU Library General Public
10  *  License as published by the Free Software Foundation; either
11  *  version 2 of the License, or (at your option) any later version.
12  *
13  *  This library is distributed in the hope that it will be useful,
14  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
15  *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
16  *  Library General Public License for more details.
17  *
18  *  You should have received a copy of the GNU Library General Public License
19  *  along with this library; see the file COPYING.LIB.  If not, write to
20  *  the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
21  *  Boston, MA 02110-1301, USA.
22  *
23  */
24 
25 #include "config.h"
26 #include "JSGlobalObjectFunctions.h"
27 
28 #include "CallFrame.h"
29 #include "Interpreter.h"
30 #include "JSGlobalObject.h"
31 #include "JSString.h"
32 #include "JSStringBuilder.h"
33 #include "Lexer.h"
34 #include "LiteralParser.h"
35 #include "Nodes.h"
36 #include "Parser.h"
37 #include "UStringBuilder.h"
38 #include "dtoa.h"
39 #include <stdio.h>
40 #include <stdlib.h>
41 #include <wtf/ASCIICType.h>
42 #include <wtf/Assertions.h>
43 #include <wtf/MathExtras.h>
44 #include <wtf/StringExtras.h>
45 #include <wtf/unicode/UTF8.h>
46 
47 using namespace WTF;
48 using namespace Unicode;
49 
50 namespace JSC {
51 
encode(ExecState * exec,const char * doNotEscape)52 static JSValue encode(ExecState* exec, const char* doNotEscape)
53 {
54     UString str = exec->argument(0).toString(exec);
55     CString cstr = str.utf8(true);
56     if (!cstr.data())
57         return throwError(exec, createURIError(exec, "String contained an illegal UTF-16 sequence."));
58 
59     JSStringBuilder builder;
60     const char* p = cstr.data();
61     for (size_t k = 0; k < cstr.length(); k++, p++) {
62         char c = *p;
63         if (c && strchr(doNotEscape, c))
64             builder.append(c);
65         else {
66             char tmp[4];
67             snprintf(tmp, sizeof(tmp), "%%%02X", static_cast<unsigned char>(c));
68             builder.append(tmp);
69         }
70     }
71     return builder.build(exec);
72 }
73 
decode(ExecState * exec,const char * doNotUnescape,bool strict)74 static JSValue decode(ExecState* exec, const char* doNotUnescape, bool strict)
75 {
76     JSStringBuilder builder;
77     UString str = exec->argument(0).toString(exec);
78     int k = 0;
79     int len = str.length();
80     const UChar* d = str.characters();
81     UChar u = 0;
82     while (k < len) {
83         const UChar* p = d + k;
84         UChar c = *p;
85         if (c == '%') {
86             int charLen = 0;
87             if (k <= len - 3 && isASCIIHexDigit(p[1]) && isASCIIHexDigit(p[2])) {
88                 const char b0 = Lexer::convertHex(p[1], p[2]);
89                 const int sequenceLen = UTF8SequenceLength(b0);
90                 if (sequenceLen != 0 && k <= len - sequenceLen * 3) {
91                     charLen = sequenceLen * 3;
92                     char sequence[5];
93                     sequence[0] = b0;
94                     for (int i = 1; i < sequenceLen; ++i) {
95                         const UChar* q = p + i * 3;
96                         if (q[0] == '%' && isASCIIHexDigit(q[1]) && isASCIIHexDigit(q[2]))
97                             sequence[i] = Lexer::convertHex(q[1], q[2]);
98                         else {
99                             charLen = 0;
100                             break;
101                         }
102                     }
103                     if (charLen != 0) {
104                         sequence[sequenceLen] = 0;
105                         const int character = decodeUTF8Sequence(sequence);
106                         if (character < 0 || character >= 0x110000)
107                             charLen = 0;
108                         else if (character >= 0x10000) {
109                             // Convert to surrogate pair.
110                             builder.append(static_cast<UChar>(0xD800 | ((character - 0x10000) >> 10)));
111                             u = static_cast<UChar>(0xDC00 | ((character - 0x10000) & 0x3FF));
112                         } else
113                             u = static_cast<UChar>(character);
114                     }
115                 }
116             }
117             if (charLen == 0) {
118                 if (strict)
119                     return throwError(exec, createURIError(exec, "URI error"));
120                 // The only case where we don't use "strict" mode is the "unescape" function.
121                 // For that, it's good to support the wonky "%u" syntax for compatibility with WinIE.
122                 if (k <= len - 6 && p[1] == 'u'
123                         && isASCIIHexDigit(p[2]) && isASCIIHexDigit(p[3])
124                         && isASCIIHexDigit(p[4]) && isASCIIHexDigit(p[5])) {
125                     charLen = 6;
126                     u = Lexer::convertUnicode(p[2], p[3], p[4], p[5]);
127                 }
128             }
129             if (charLen && (u == 0 || u >= 128 || !strchr(doNotUnescape, u))) {
130                 c = u;
131                 k += charLen - 1;
132             }
133         }
134         k++;
135         builder.append(c);
136     }
137     return builder.build(exec);
138 }
139 
isStrWhiteSpace(UChar c)140 bool isStrWhiteSpace(UChar c)
141 {
142     switch (c) {
143         // ECMA-262-5th 7.2 & 7.3
144         case 0x0009:
145         case 0x000A:
146         case 0x000B:
147         case 0x000C:
148         case 0x000D:
149         case 0x0020:
150         case 0x00A0:
151         case 0x2028:
152         case 0x2029:
153         case 0xFEFF:
154             return true;
155         default:
156             return c > 0xff && isSeparatorSpace(c);
157     }
158 }
159 
parseDigit(unsigned short c,int radix)160 static int parseDigit(unsigned short c, int radix)
161 {
162     int digit = -1;
163 
164     if (c >= '0' && c <= '9')
165         digit = c - '0';
166     else if (c >= 'A' && c <= 'Z')
167         digit = c - 'A' + 10;
168     else if (c >= 'a' && c <= 'z')
169         digit = c - 'a' + 10;
170 
171     if (digit >= radix)
172         return -1;
173     return digit;
174 }
175 
parseIntOverflow(const char * s,int length,int radix)176 double parseIntOverflow(const char* s, int length, int radix)
177 {
178     double number = 0.0;
179     double radixMultiplier = 1.0;
180 
181     for (const char* p = s + length - 1; p >= s; p--) {
182         if (radixMultiplier == Inf) {
183             if (*p != '0') {
184                 number = Inf;
185                 break;
186             }
187         } else {
188             int digit = parseDigit(*p, radix);
189             number += digit * radixMultiplier;
190         }
191 
192         radixMultiplier *= radix;
193     }
194 
195     return number;
196 }
197 
parseIntOverflow(const UChar * s,int length,int radix)198 double parseIntOverflow(const UChar* s, int length, int radix)
199 {
200     double number = 0.0;
201     double radixMultiplier = 1.0;
202 
203     for (const UChar* p = s + length - 1; p >= s; p--) {
204         if (radixMultiplier == Inf) {
205             if (*p != '0') {
206                 number = Inf;
207                 break;
208             }
209         } else {
210             int digit = parseDigit(*p, radix);
211             number += digit * radixMultiplier;
212         }
213 
214         radixMultiplier *= radix;
215     }
216 
217     return number;
218 }
219 
parseInt(const UString & s,int radix)220 static double parseInt(const UString& s, int radix)
221 {
222     int length = s.length();
223     const UChar* data = s.characters();
224     int p = 0;
225 
226     while (p < length && isStrWhiteSpace(data[p]))
227         ++p;
228 
229     double sign = 1;
230     if (p < length) {
231         if (data[p] == '+')
232             ++p;
233         else if (data[p] == '-') {
234             sign = -1;
235             ++p;
236         }
237     }
238 
239     if ((radix == 0 || radix == 16) && length - p >= 2 && data[p] == '0' && (data[p + 1] == 'x' || data[p + 1] == 'X')) {
240         radix = 16;
241         p += 2;
242     } else if (radix == 0) {
243         if (p < length && data[p] == '0')
244             radix = 8;
245         else
246             radix = 10;
247     }
248 
249     if (radix < 2 || radix > 36)
250         return NaN;
251 
252     int firstDigitPosition = p;
253     bool sawDigit = false;
254     double number = 0;
255     while (p < length) {
256         int digit = parseDigit(data[p], radix);
257         if (digit == -1)
258             break;
259         sawDigit = true;
260         number *= radix;
261         number += digit;
262         ++p;
263     }
264 
265     if (number >= mantissaOverflowLowerBound) {
266         if (radix == 10)
267             number = WTF::strtod(s.substringSharingImpl(firstDigitPosition, p - firstDigitPosition).utf8().data(), 0);
268         else if (radix == 2 || radix == 4 || radix == 8 || radix == 16 || radix == 32)
269             number = parseIntOverflow(s.substringSharingImpl(firstDigitPosition, p - firstDigitPosition).utf8().data(), p - firstDigitPosition, radix);
270     }
271 
272     if (!sawDigit)
273         return NaN;
274 
275     return sign * number;
276 }
277 
278 static const int SizeOfInfinity = 8;
279 
isInfinity(const UChar * data,const UChar * end)280 static bool isInfinity(const UChar* data, const UChar* end)
281 {
282     return (end - data) >= SizeOfInfinity
283         && data[0] == 'I'
284         && data[1] == 'n'
285         && data[2] == 'f'
286         && data[3] == 'i'
287         && data[4] == 'n'
288         && data[5] == 'i'
289         && data[6] == 't'
290         && data[7] == 'y';
291 }
292 
293 // See ecma-262 9.3.1
jsHexIntegerLiteral(const UChar * & data,const UChar * end)294 static double jsHexIntegerLiteral(const UChar*& data, const UChar* end)
295 {
296     // Hex number.
297     data += 2;
298     const UChar* firstDigitPosition = data;
299     double number = 0;
300     while (true) {
301         number = number * 16 + toASCIIHexValue(*data);
302         ++data;
303         if (data == end)
304             break;
305         if (!isASCIIHexDigit(*data))
306             break;
307     }
308     if (number >= mantissaOverflowLowerBound)
309         number = parseIntOverflow(firstDigitPosition, data - firstDigitPosition, 16);
310 
311     return number;
312 }
313 
314 // See ecma-262 9.3.1
jsStrDecimalLiteral(const UChar * & data,const UChar * end)315 static double jsStrDecimalLiteral(const UChar*& data, const UChar* end)
316 {
317     ASSERT(data < end);
318 
319     // Copy the sting into a null-terminated byte buffer, and call strtod.
320     Vector<char, 32> byteBuffer;
321     for (const UChar* characters = data; characters < end; ++characters) {
322         UChar character = *characters;
323         byteBuffer.append(isASCII(character) ? character : 0);
324     }
325     byteBuffer.append(0);
326     char* endOfNumber;
327     double number = WTF::strtod(byteBuffer.data(), &endOfNumber);
328 
329     // Check if strtod found a number; if so return it.
330     ptrdiff_t consumed = endOfNumber - byteBuffer.data();
331     if (consumed) {
332         data += consumed;
333         return number;
334     }
335 
336     // Check for [+-]?Infinity
337     switch (*data) {
338     case 'I':
339         if (isInfinity(data, end)) {
340             data += SizeOfInfinity;
341             return Inf;
342         }
343         break;
344 
345     case '+':
346         if (isInfinity(data + 1, end)) {
347             data += SizeOfInfinity + 1;
348             return Inf;
349         }
350         break;
351 
352     case '-':
353         if (isInfinity(data + 1, end)) {
354             data += SizeOfInfinity + 1;
355             return -Inf;
356         }
357         break;
358     }
359 
360     // Not a number.
361     return NaN;
362 }
363 
364 // See ecma-262 9.3.1
jsToNumber(const UString & s)365 double jsToNumber(const UString& s)
366 {
367     unsigned size = s.length();
368 
369     if (size == 1) {
370         UChar c = s.characters()[0];
371         if (isASCIIDigit(c))
372             return c - '0';
373         if (isStrWhiteSpace(c))
374             return 0;
375         return NaN;
376     }
377 
378     const UChar* data = s.characters();
379     const UChar* end = data + size;
380 
381     // Skip leading white space.
382     for (; data < end; ++data) {
383         if (!isStrWhiteSpace(*data))
384             break;
385     }
386 
387     // Empty string.
388     if (data == end)
389         return 0.0;
390 
391     double number;
392     if (data[0] == '0' && data + 2 < end && (data[1] | 0x20) == 'x' && isASCIIHexDigit(data[2]))
393         number = jsHexIntegerLiteral(data, end);
394     else
395         number = jsStrDecimalLiteral(data, end);
396 
397     // Allow trailing white space.
398     for (; data < end; ++data) {
399         if (!isStrWhiteSpace(*data))
400             break;
401     }
402     if (data != end)
403         return NaN;
404 
405     return number;
406 }
407 
parseFloat(const UString & s)408 static double parseFloat(const UString& s)
409 {
410     unsigned size = s.length();
411 
412     if (size == 1) {
413         UChar c = s.characters()[0];
414         if (isASCIIDigit(c))
415             return c - '0';
416         return NaN;
417     }
418 
419     const UChar* data = s.characters();
420     const UChar* end = data + size;
421 
422     // Skip leading white space.
423     for (; data < end; ++data) {
424         if (!isStrWhiteSpace(*data))
425             break;
426     }
427 
428     // Empty string.
429     if (data == end)
430         return NaN;
431 
432     return jsStrDecimalLiteral(data, end);
433 }
434 
globalFuncEval(ExecState * exec)435 EncodedJSValue JSC_HOST_CALL globalFuncEval(ExecState* exec)
436 {
437     JSObject* thisObject = exec->hostThisValue().toThisObject(exec);
438     JSObject* unwrappedObject = thisObject->unwrappedObject();
439     if (!unwrappedObject->isGlobalObject() || static_cast<JSGlobalObject*>(unwrappedObject)->evalFunction() != exec->callee())
440         return throwVMError(exec, createEvalError(exec, "The \"this\" value passed to eval must be the global object from which eval originated"));
441 
442     JSValue x = exec->argument(0);
443     if (!x.isString())
444         return JSValue::encode(x);
445 
446     UString s = x.toString(exec);
447 
448     LiteralParser preparser(exec, s, LiteralParser::NonStrictJSON);
449     if (JSValue parsedObject = preparser.tryLiteralParse())
450         return JSValue::encode(parsedObject);
451 
452     EvalExecutable* eval = EvalExecutable::create(exec, makeSource(s), false);
453     JSObject* error = eval->compile(exec, static_cast<JSGlobalObject*>(unwrappedObject)->globalScopeChain());
454     if (error)
455         return throwVMError(exec, error);
456 
457     return JSValue::encode(exec->interpreter()->execute(eval, exec, thisObject, static_cast<JSGlobalObject*>(unwrappedObject)->globalScopeChain()));
458 }
459 
globalFuncParseInt(ExecState * exec)460 EncodedJSValue JSC_HOST_CALL globalFuncParseInt(ExecState* exec)
461 {
462     JSValue value = exec->argument(0);
463     int32_t radix = exec->argument(1).toInt32(exec);
464 
465     if (radix != 0 && radix != 10)
466         return JSValue::encode(jsNumber(parseInt(value.toString(exec), radix)));
467 
468     if (value.isInt32())
469         return JSValue::encode(value);
470 
471     if (value.isDouble()) {
472         double d = value.asDouble();
473         if (isfinite(d))
474             return JSValue::encode(jsNumber((d > 0) ? floor(d) : ceil(d)));
475         if (isnan(d) || isinf(d))
476             return JSValue::encode(jsNaN());
477         return JSValue::encode(jsNumber(0));
478     }
479 
480     return JSValue::encode(jsNumber(parseInt(value.toString(exec), radix)));
481 }
482 
globalFuncParseFloat(ExecState * exec)483 EncodedJSValue JSC_HOST_CALL globalFuncParseFloat(ExecState* exec)
484 {
485     return JSValue::encode(jsNumber(parseFloat(exec->argument(0).toString(exec))));
486 }
487 
globalFuncIsNaN(ExecState * exec)488 EncodedJSValue JSC_HOST_CALL globalFuncIsNaN(ExecState* exec)
489 {
490     return JSValue::encode(jsBoolean(isnan(exec->argument(0).toNumber(exec))));
491 }
492 
globalFuncIsFinite(ExecState * exec)493 EncodedJSValue JSC_HOST_CALL globalFuncIsFinite(ExecState* exec)
494 {
495     double n = exec->argument(0).toNumber(exec);
496     return JSValue::encode(jsBoolean(!isnan(n) && !isinf(n)));
497 }
498 
globalFuncDecodeURI(ExecState * exec)499 EncodedJSValue JSC_HOST_CALL globalFuncDecodeURI(ExecState* exec)
500 {
501     static const char do_not_unescape_when_decoding_URI[] =
502         "#$&+,/:;=?@";
503 
504     return JSValue::encode(decode(exec, do_not_unescape_when_decoding_URI, true));
505 }
506 
globalFuncDecodeURIComponent(ExecState * exec)507 EncodedJSValue JSC_HOST_CALL globalFuncDecodeURIComponent(ExecState* exec)
508 {
509     return JSValue::encode(decode(exec, "", true));
510 }
511 
globalFuncEncodeURI(ExecState * exec)512 EncodedJSValue JSC_HOST_CALL globalFuncEncodeURI(ExecState* exec)
513 {
514     static const char do_not_escape_when_encoding_URI[] =
515         "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
516         "abcdefghijklmnopqrstuvwxyz"
517         "0123456789"
518         "!#$&'()*+,-./:;=?@_~";
519 
520     return JSValue::encode(encode(exec, do_not_escape_when_encoding_URI));
521 }
522 
globalFuncEncodeURIComponent(ExecState * exec)523 EncodedJSValue JSC_HOST_CALL globalFuncEncodeURIComponent(ExecState* exec)
524 {
525     static const char do_not_escape_when_encoding_URI_component[] =
526         "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
527         "abcdefghijklmnopqrstuvwxyz"
528         "0123456789"
529         "!'()*-._~";
530 
531     return JSValue::encode(encode(exec, do_not_escape_when_encoding_URI_component));
532 }
533 
globalFuncEscape(ExecState * exec)534 EncodedJSValue JSC_HOST_CALL globalFuncEscape(ExecState* exec)
535 {
536     static const char do_not_escape[] =
537         "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
538         "abcdefghijklmnopqrstuvwxyz"
539         "0123456789"
540         "*+-./@_";
541 
542     JSStringBuilder builder;
543     UString str = exec->argument(0).toString(exec);
544     const UChar* c = str.characters();
545     for (unsigned k = 0; k < str.length(); k++, c++) {
546         int u = c[0];
547         if (u > 255) {
548             char tmp[7];
549             snprintf(tmp, sizeof(tmp), "%%u%04X", u);
550             builder.append(tmp);
551         } else if (u != 0 && strchr(do_not_escape, static_cast<char>(u)))
552             builder.append(c, 1);
553         else {
554             char tmp[4];
555             snprintf(tmp, sizeof(tmp), "%%%02X", u);
556             builder.append(tmp);
557         }
558     }
559 
560     return JSValue::encode(builder.build(exec));
561 }
562 
globalFuncUnescape(ExecState * exec)563 EncodedJSValue JSC_HOST_CALL globalFuncUnescape(ExecState* exec)
564 {
565     UStringBuilder builder;
566     UString str = exec->argument(0).toString(exec);
567     int k = 0;
568     int len = str.length();
569     while (k < len) {
570         const UChar* c = str.characters() + k;
571         UChar u;
572         if (c[0] == '%' && k <= len - 6 && c[1] == 'u') {
573             if (isASCIIHexDigit(c[2]) && isASCIIHexDigit(c[3]) && isASCIIHexDigit(c[4]) && isASCIIHexDigit(c[5])) {
574                 u = Lexer::convertUnicode(c[2], c[3], c[4], c[5]);
575                 c = &u;
576                 k += 5;
577             }
578         } else if (c[0] == '%' && k <= len - 3 && isASCIIHexDigit(c[1]) && isASCIIHexDigit(c[2])) {
579             u = UChar(Lexer::convertHex(c[1], c[2]));
580             c = &u;
581             k += 2;
582         }
583         k++;
584         builder.append(*c);
585     }
586 
587     return JSValue::encode(jsString(exec, builder.toUString()));
588 }
589 
590 #ifndef NDEBUG
globalFuncJSCPrint(ExecState * exec)591 EncodedJSValue JSC_HOST_CALL globalFuncJSCPrint(ExecState* exec)
592 {
593     CString string = exec->argument(0).toString(exec).utf8();
594     puts(string.data());
595     return JSValue::encode(jsUndefined());
596 }
597 #endif
598 
599 } // namespace JSC
600