/[jscoverage]/trunk/js/jsstr.h
ViewVC logotype

Contents of /trunk/js/jsstr.h

Parent Directory Parent Directory | Revision Log Revision Log


Revision 460 - (show annotations)
Sat Sep 26 23:15:22 2009 UTC (10 years, 1 month ago) by siliconforks
File MIME type: text/plain
File size: 29712 byte(s)
Upgrade to SpiderMonkey from Firefox 3.5.3.

1 /* -*- Mode: C; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 4 -*-
2 *
3 * ***** BEGIN LICENSE BLOCK *****
4 * Version: MPL 1.1/GPL 2.0/LGPL 2.1
5 *
6 * The contents of this file are subject to the Mozilla Public License Version
7 * 1.1 (the "License"); you may not use this file except in compliance with
8 * the License. You may obtain a copy of the License at
9 * http://www.mozilla.org/MPL/
10 *
11 * Software distributed under the License is distributed on an "AS IS" basis,
12 * WITHOUT WARRANTY OF ANY KIND, either express or implied. See the License
13 * for the specific language governing rights and limitations under the
14 * License.
15 *
16 * The Original Code is Mozilla Communicator client code, released
17 * March 31, 1998.
18 *
19 * The Initial Developer of the Original Code is
20 * Netscape Communications Corporation.
21 * Portions created by the Initial Developer are Copyright (C) 1998
22 * the Initial Developer. All Rights Reserved.
23 *
24 * Contributor(s):
25 *
26 * Alternatively, the contents of this file may be used under the terms of
27 * either of the GNU General Public License Version 2 or later (the "GPL"),
28 * or the GNU Lesser General Public License Version 2.1 or later (the "LGPL"),
29 * in which case the provisions of the GPL or the LGPL are applicable instead
30 * of those above. If you wish to allow use of your version of this file only
31 * under the terms of either the GPL or the LGPL, and not to allow others to
32 * use your version of this file under the terms of the MPL, indicate your
33 * decision by deleting the provisions above and replace them with the notice
34 * and other provisions required by the GPL or the LGPL. If you do not delete
35 * the provisions above, a recipient may use your version of this file under
36 * the terms of any one of the MPL, the GPL or the LGPL.
37 *
38 * ***** END LICENSE BLOCK ***** */
39
40 #ifndef jsstr_h___
41 #define jsstr_h___
42 /*
43 * JS string type implementation.
44 *
45 * A JS string is a counted array of unicode characters. To support handoff
46 * of API client memory, the chars are allocated separately from the length,
47 * necessitating a pointer after the count, to form a separately allocated
48 * string descriptor. String descriptors are GC'ed, while their chars are
49 * allocated from the malloc heap.
50 */
51 #include <ctype.h>
52 #include "jspubtd.h"
53 #include "jsprvtd.h"
54
55 JS_BEGIN_EXTERN_C
56
57 /*
58 * The GC-thing "string" type.
59 *
60 * When the JSSTRFLAG_DEPENDENT bit of the length field is unset, the u.chars
61 * field points to a flat character array owned by its GC-thing descriptor.
62 * The array is terminated at index length by a zero character and the size of
63 * the array in bytes is (length + 1) * sizeof(jschar). The terminator is
64 * purely a backstop, in case the chars pointer flows out to native code that
65 * requires \u0000 termination.
66 *
67 * A flat string with JSSTRFLAG_MUTABLE set means that the string is accessible
68 * only from one thread and it is possible to turn it into a dependent string
69 * of the same length to optimize js_ConcatStrings. It is also possible to grow
70 * such a string, but extreme care must be taken to ensure that no other code
71 * relies on the original length of the string.
72 *
73 * A flat string with JSSTRFLAG_ATOMIZED set means that the string is hashed as
74 * an atom. This flag is used to avoid re-hashing the already-atomized string.
75 *
76 * Any string with JSSTRFLAG_DEFLATED set means that the string has an entry
77 * in the deflated string cache. The GC uses this flag to optimize string
78 * finalization and avoid an expensive cache lookup for strings that were
79 * never deflated.
80 *
81 * When JSSTRFLAG_DEPENDENT is set, the string depends on characters of another
82 * string strongly referenced by the u.base field. The base member may point to
83 * another dependent string if JSSTRING_CHARS has not been called yet.
84 *
85 * JSSTRFLAG_PREFIX determines the kind of the dependent string. When the flag
86 * is unset, the length field encodes both starting position relative to the
87 * base string and the number of characters in the dependent string, see
88 * JSSTRDEP_START_MASK and JSSTRDEP_LENGTH_MASK macros below for details.
89 *
90 * When JSSTRFLAG_PREFIX is set, the dependent string is a prefix of the base
91 * string. The number of characters in the prefix is encoded using all non-flag
92 * bits of the length field and spans the same 0 .. SIZE_T_MAX/4 range as the
93 * length of the flat string.
94 *
95 * NB: Always use the JSSTRING_LENGTH and JSSTRING_CHARS accessor macros.
96 */
97 struct JSString {
98 size_t length;
99 union {
100 jschar *chars;
101 JSString *base;
102 } u;
103 };
104
105 /*
106 * Definitions for flags stored in the high order bits of JSString.length.
107 * JSSTRFLAG_PREFIX and JSSTRFLAG_MUTABLE are two aliases for the same value.
108 * JSSTRFLAG_PREFIX should be used only if JSSTRFLAG_DEPENDENT is set and
109 * JSSTRFLAG_MUTABLE should be used only if the string is flat.
110 * JSSTRFLAG_ATOMIZED is used only with the flat immutable strings.
111 */
112 #define JSSTRFLAG_DEPENDENT JSSTRING_BIT(JS_BITS_PER_WORD - 1)
113 #define JSSTRFLAG_PREFIX JSSTRING_BIT(JS_BITS_PER_WORD - 2)
114 #define JSSTRFLAG_MUTABLE JSSTRFLAG_PREFIX
115 #define JSSTRFLAG_ATOMIZED JSSTRING_BIT(JS_BITS_PER_WORD - 3)
116 #define JSSTRFLAG_DEFLATED JSSTRING_BIT(JS_BITS_PER_WORD - 4)
117
118 #define JSSTRING_LENGTH_BITS (JS_BITS_PER_WORD - 4)
119 #define JSSTRING_LENGTH_MASK JSSTRING_BITMASK(JSSTRING_LENGTH_BITS)
120
121 /* Universal JSString type inquiry and accessor macros. */
122 #define JSSTRING_BIT(n) ((size_t)1 << (n))
123 #define JSSTRING_BITMASK(n) (JSSTRING_BIT(n) - 1)
124 #define JSSTRING_HAS_FLAG(str,flg) ((str)->length & (flg))
125 #define JSSTRING_IS_DEPENDENT(str) JSSTRING_HAS_FLAG(str, JSSTRFLAG_DEPENDENT)
126 #define JSSTRING_IS_FLAT(str) (!JSSTRING_IS_DEPENDENT(str))
127 #define JSSTRING_IS_MUTABLE(str) (((str)->length & (JSSTRFLAG_DEPENDENT | \
128 JSSTRFLAG_MUTABLE)) == \
129 JSSTRFLAG_MUTABLE)
130 #define JSSTRING_IS_ATOMIZED(str) (((str)->length & (JSSTRFLAG_DEPENDENT | \
131 JSSTRFLAG_ATOMIZED)) ==\
132 JSSTRFLAG_ATOMIZED)
133
134 #define JSSTRING_CHARS(str) (JSSTRING_IS_DEPENDENT(str) \
135 ? JSSTRDEP_CHARS(str) \
136 : JSFLATSTR_CHARS(str))
137 #define JSSTRING_LENGTH(str) (JSSTRING_IS_DEPENDENT(str) \
138 ? JSSTRDEP_LENGTH(str) \
139 : JSFLATSTR_LENGTH(str))
140
141 JS_STATIC_ASSERT(sizeof(size_t) == sizeof(jsword));
142
143 #define JSSTRING_IS_DEFLATED(str) ((str)->length & JSSTRFLAG_DEFLATED)
144
145 #define JSSTRING_SET_DEFLATED(str) \
146 JS_ATOMIC_SET_MASK((jsword*)&(str)->length, JSSTRFLAG_DEFLATED)
147
148 #define JSSTRING_CHARS_AND_LENGTH(str, chars_, length_) \
149 ((void)(JSSTRING_IS_DEPENDENT(str) \
150 ? ((length_) = JSSTRDEP_LENGTH(str), \
151 (chars_) = JSSTRDEP_CHARS(str)) \
152 : ((length_) = JSFLATSTR_LENGTH(str), \
153 (chars_) = JSFLATSTR_CHARS(str))))
154
155 #define JSSTRING_CHARS_AND_END(str, chars_, end) \
156 ((void)((end) = JSSTRING_IS_DEPENDENT(str) \
157 ? JSSTRDEP_LENGTH(str) + ((chars_) = JSSTRDEP_CHARS(str)) \
158 : JSFLATSTR_LENGTH(str) + ((chars_) = JSFLATSTR_CHARS(str))))
159
160 /* Specific flat string initializer and accessor macros. */
161 #define JSFLATSTR_INIT(str, chars_, length_) \
162 ((void)(JS_ASSERT(((length_) & ~JSSTRING_LENGTH_MASK) == 0), \
163 (str)->length = (length_), (str)->u.chars = (chars_)))
164
165 #define JSFLATSTR_LENGTH(str) \
166 (JS_ASSERT(JSSTRING_IS_FLAT(str)), (str)->length & JSSTRING_LENGTH_MASK)
167
168 #define JSFLATSTR_CHARS(str) \
169 (JS_ASSERT(JSSTRING_IS_FLAT(str)), (str)->u.chars)
170
171 /*
172 * Special flat string initializer that preserves the JSSTR_DEFLATED flag.
173 * Use this macro when reinitializing an existing string (which may be
174 * hashed to its deflated bytes. Newborn strings must use JSFLATSTR_INIT.
175 */
176 #define JSFLATSTR_REINIT(str, chars_, length_) \
177 ((void)(JS_ASSERT(((length_) & ~JSSTRING_LENGTH_MASK) == 0), \
178 (str)->length = ((str)->length & JSSTRFLAG_DEFLATED) | \
179 (length_ & ~JSSTRFLAG_DEFLATED), \
180 (str)->u.chars = (chars_)))
181
182 /*
183 * Macros to manipulate atomized and mutable flags of flat strings. It is safe
184 * to use these without extra locking due to the following properties:
185 *
186 * * We do not have a macro like JSFLATSTR_CLEAR_ATOMIZED as a string
187 * remains atomized until the GC collects it.
188 *
189 * * A thread may call JSFLATSTR_SET_MUTABLE only when it is the only thread
190 * accessing the string until a later call to JSFLATSTR_CLEAR_MUTABLE.
191 *
192 * * Multiple threads can call JSFLATSTR_CLEAR_MUTABLE but the macro
193 * actually clears the mutable flag only when the flag is set -- in which
194 * case only one thread can access the string (see previous property).
195 *
196 * Thus, when multiple threads access the string, JSFLATSTR_SET_ATOMIZED is
197 * the only macro that can update the length field of the string by changing
198 * the mutable bit from 0 to 1. We call the macro only after the string has
199 * been hashed. When some threads in js_ValueToStringId see that the flag is
200 * set, it knows that the string was atomized.
201 *
202 * On the other hand, if the thread sees that the flag is unset, it could be
203 * seeing a stale value when another thread has just atomized the string and
204 * set the flag. But this can lead only to an extra call to js_AtomizeString.
205 * This function would find that the string was already hashed and return it
206 * with the atomized bit set.
207 */
208 #define JSFLATSTR_SET_ATOMIZED(str) \
209 JS_BEGIN_MACRO \
210 JS_ASSERT(JSSTRING_IS_FLAT(str) && !JSSTRING_IS_MUTABLE(str)); \
211 JS_ATOMIC_SET_MASK((jsword*) &(str)->length, JSSTRFLAG_ATOMIZED); \
212 JS_END_MACRO
213
214 #define JSFLATSTR_SET_MUTABLE(str) \
215 ((void)(JS_ASSERT(JSSTRING_IS_FLAT(str) && !JSSTRING_IS_ATOMIZED(str)), \
216 (str)->length |= JSSTRFLAG_MUTABLE))
217
218 #define JSFLATSTR_CLEAR_MUTABLE(str) \
219 ((void)(JS_ASSERT(JSSTRING_IS_FLAT(str)), \
220 JSSTRING_HAS_FLAG(str, JSSTRFLAG_MUTABLE) && \
221 ((str)->length &= ~JSSTRFLAG_MUTABLE)))
222
223 /* Specific dependent string shift/mask accessor and mutator macros. */
224 #define JSSTRDEP_START_BITS (JSSTRING_LENGTH_BITS-JSSTRDEP_LENGTH_BITS)
225 #define JSSTRDEP_START_SHIFT JSSTRDEP_LENGTH_BITS
226 #define JSSTRDEP_START_MASK JSSTRING_BITMASK(JSSTRDEP_START_BITS)
227 #define JSSTRDEP_LENGTH_BITS (JSSTRING_LENGTH_BITS / 2)
228 #define JSSTRDEP_LENGTH_MASK JSSTRING_BITMASK(JSSTRDEP_LENGTH_BITS)
229
230 #define JSSTRDEP_IS_PREFIX(str) JSSTRING_HAS_FLAG(str, JSSTRFLAG_PREFIX)
231
232 #define JSSTRDEP_START(str) (JSSTRDEP_IS_PREFIX(str) ? 0 \
233 : (((str)->length \
234 >> JSSTRDEP_START_SHIFT) \
235 & JSSTRDEP_START_MASK))
236 #define JSSTRDEP_LENGTH(str) ((str)->length \
237 & (JSSTRDEP_IS_PREFIX(str) \
238 ? JSSTRING_LENGTH_MASK \
239 : JSSTRDEP_LENGTH_MASK))
240
241 #define JSSTRDEP_INIT(str,bstr,off,len) \
242 ((str)->length = JSSTRFLAG_DEPENDENT \
243 | ((off) << JSSTRDEP_START_SHIFT) \
244 | (len), \
245 (str)->u.base = (bstr))
246
247 /* See JSFLATSTR_INIT. */
248 #define JSSTRDEP_REINIT(str,bstr,off,len) \
249 ((str)->length = JSSTRFLAG_DEPENDENT \
250 | ((str->length) & JSSTRFLAG_DEFLATED) \
251 | ((off) << JSSTRDEP_START_SHIFT) \
252 | (len), \
253 (str)->u.base = (bstr))
254
255 #define JSPREFIX_INIT(str,bstr,len) \
256 ((str)->length = JSSTRFLAG_DEPENDENT | JSSTRFLAG_PREFIX | (len), \
257 (str)->u.base = (bstr))
258
259 /* See JSFLATSTR_INIT. */
260 #define JSPREFIX_REINIT(str,bstr,len) \
261 ((str)->length = JSSTRFLAG_DEPENDENT | JSSTRFLAG_PREFIX | \
262 ((str->length) & JSSTRFLAG_DEFLATED) | (len), \
263 (str)->u.base = (bstr))
264
265 #define JSSTRDEP_BASE(str) ((str)->u.base)
266 #define JSPREFIX_BASE(str) JSSTRDEP_BASE(str)
267 #define JSPREFIX_SET_BASE(str,bstr) ((str)->u.base = (bstr))
268
269 #define JSSTRDEP_CHARS(str) \
270 (JSSTRING_IS_DEPENDENT(JSSTRDEP_BASE(str)) \
271 ? js_GetDependentStringChars(str) \
272 : JSFLATSTR_CHARS(JSSTRDEP_BASE(str)) + JSSTRDEP_START(str))
273
274 extern size_t
275 js_MinimizeDependentStrings(JSString *str, int level, JSString **basep);
276
277 extern jschar *
278 js_GetDependentStringChars(JSString *str);
279
280 extern const jschar *
281 js_GetStringChars(JSContext *cx, JSString *str);
282
283 extern JSString * JS_FASTCALL
284 js_ConcatStrings(JSContext *cx, JSString *left, JSString *right);
285
286 extern const jschar *
287 js_UndependString(JSContext *cx, JSString *str);
288
289 extern JSBool
290 js_MakeStringImmutable(JSContext *cx, JSString *str);
291
292 extern JSString* JS_FASTCALL
293 js_toLowerCase(JSContext *cx, JSString *str);
294
295 extern JSString* JS_FASTCALL
296 js_toUpperCase(JSContext *cx, JSString *str);
297
298 typedef struct JSCharBuffer {
299 size_t length;
300 jschar *chars;
301 } JSCharBuffer;
302
303 struct JSSubString {
304 size_t length;
305 const jschar *chars;
306 };
307
308 extern jschar js_empty_ucstr[];
309 extern JSSubString js_EmptySubString;
310
311 /* Unicode character attribute lookup tables. */
312 extern const uint8 js_X[];
313 extern const uint8 js_Y[];
314 extern const uint32 js_A[];
315
316 /* Enumerated Unicode general category types. */
317 typedef enum JSCharType {
318 JSCT_UNASSIGNED = 0,
319 JSCT_UPPERCASE_LETTER = 1,
320 JSCT_LOWERCASE_LETTER = 2,
321 JSCT_TITLECASE_LETTER = 3,
322 JSCT_MODIFIER_LETTER = 4,
323 JSCT_OTHER_LETTER = 5,
324 JSCT_NON_SPACING_MARK = 6,
325 JSCT_ENCLOSING_MARK = 7,
326 JSCT_COMBINING_SPACING_MARK = 8,
327 JSCT_DECIMAL_DIGIT_NUMBER = 9,
328 JSCT_LETTER_NUMBER = 10,
329 JSCT_OTHER_NUMBER = 11,
330 JSCT_SPACE_SEPARATOR = 12,
331 JSCT_LINE_SEPARATOR = 13,
332 JSCT_PARAGRAPH_SEPARATOR = 14,
333 JSCT_CONTROL = 15,
334 JSCT_FORMAT = 16,
335 JSCT_PRIVATE_USE = 18,
336 JSCT_SURROGATE = 19,
337 JSCT_DASH_PUNCTUATION = 20,
338 JSCT_START_PUNCTUATION = 21,
339 JSCT_END_PUNCTUATION = 22,
340 JSCT_CONNECTOR_PUNCTUATION = 23,
341 JSCT_OTHER_PUNCTUATION = 24,
342 JSCT_MATH_SYMBOL = 25,
343 JSCT_CURRENCY_SYMBOL = 26,
344 JSCT_MODIFIER_SYMBOL = 27,
345 JSCT_OTHER_SYMBOL = 28
346 } JSCharType;
347
348 /* Character classifying and mapping macros, based on java.lang.Character. */
349 #define JS_CCODE(c) (js_A[js_Y[(js_X[(uint16)(c)>>6]<<6)|((c)&0x3F)]])
350 #define JS_CTYPE(c) (JS_CCODE(c) & 0x1F)
351
352 #define JS_ISALPHA(c) ((((1 << JSCT_UPPERCASE_LETTER) | \
353 (1 << JSCT_LOWERCASE_LETTER) | \
354 (1 << JSCT_TITLECASE_LETTER) | \
355 (1 << JSCT_MODIFIER_LETTER) | \
356 (1 << JSCT_OTHER_LETTER)) \
357 >> JS_CTYPE(c)) & 1)
358
359 #define JS_ISALNUM(c) ((((1 << JSCT_UPPERCASE_LETTER) | \
360 (1 << JSCT_LOWERCASE_LETTER) | \
361 (1 << JSCT_TITLECASE_LETTER) | \
362 (1 << JSCT_MODIFIER_LETTER) | \
363 (1 << JSCT_OTHER_LETTER) | \
364 (1 << JSCT_DECIMAL_DIGIT_NUMBER)) \
365 >> JS_CTYPE(c)) & 1)
366
367 /* A unicode letter, suitable for use in an identifier. */
368 #define JS_ISLETTER(c) ((((1 << JSCT_UPPERCASE_LETTER) | \
369 (1 << JSCT_LOWERCASE_LETTER) | \
370 (1 << JSCT_TITLECASE_LETTER) | \
371 (1 << JSCT_MODIFIER_LETTER) | \
372 (1 << JSCT_OTHER_LETTER) | \
373 (1 << JSCT_LETTER_NUMBER)) \
374 >> JS_CTYPE(c)) & 1)
375
376 /*
377 * 'IdentifierPart' from ECMA grammar, is Unicode letter or combining mark or
378 * digit or connector punctuation.
379 */
380 #define JS_ISIDPART(c) ((((1 << JSCT_UPPERCASE_LETTER) | \
381 (1 << JSCT_LOWERCASE_LETTER) | \
382 (1 << JSCT_TITLECASE_LETTER) | \
383 (1 << JSCT_MODIFIER_LETTER) | \
384 (1 << JSCT_OTHER_LETTER) | \
385 (1 << JSCT_LETTER_NUMBER) | \
386 (1 << JSCT_NON_SPACING_MARK) | \
387 (1 << JSCT_COMBINING_SPACING_MARK) | \
388 (1 << JSCT_DECIMAL_DIGIT_NUMBER) | \
389 (1 << JSCT_CONNECTOR_PUNCTUATION)) \
390 >> JS_CTYPE(c)) & 1)
391
392 /* Unicode control-format characters, ignored in input */
393 #define JS_ISFORMAT(c) (((1 << JSCT_FORMAT) >> JS_CTYPE(c)) & 1)
394
395 /*
396 * Per ECMA-262 15.10.2.6, these characters are the only ones that make up a
397 * "word", as far as a RegExp is concerned. If we want a Unicode-friendlier
398 * definition of "word", we should rename this macro to something regexp-y.
399 */
400 #define JS_ISWORD(c) ((c) < 128 && (isalnum(c) || (c) == '_'))
401
402 #define JS_ISIDSTART(c) (JS_ISLETTER(c) || (c) == '_' || (c) == '$')
403 #define JS_ISIDENT(c) (JS_ISIDPART(c) || (c) == '_' || (c) == '$')
404
405 #define JS_ISXMLSPACE(c) ((c) == ' ' || (c) == '\t' || (c) == '\r' || \
406 (c) == '\n')
407 #define JS_ISXMLNSSTART(c) ((JS_CCODE(c) & 0x00000100) || (c) == '_')
408 #define JS_ISXMLNS(c) ((JS_CCODE(c) & 0x00000080) || (c) == '.' || \
409 (c) == '-' || (c) == '_')
410 #define JS_ISXMLNAMESTART(c) (JS_ISXMLNSSTART(c) || (c) == ':')
411 #define JS_ISXMLNAME(c) (JS_ISXMLNS(c) || (c) == ':')
412
413 #define JS_ISDIGIT(c) (JS_CTYPE(c) == JSCT_DECIMAL_DIGIT_NUMBER)
414
415 /* XXXbe unify on A/X/Y tbls, avoid ctype.h? */
416 /* XXXbe fs, etc. ? */
417 #define JS_ISSPACE(c) ((JS_CCODE(c) & 0x00070000) == 0x00040000)
418 #define JS_ISPRINT(c) ((c) < 128 && isprint(c))
419
420 #define JS_ISUPPER(c) (JS_CTYPE(c) == JSCT_UPPERCASE_LETTER)
421 #define JS_ISLOWER(c) (JS_CTYPE(c) == JSCT_LOWERCASE_LETTER)
422
423 #define JS_TOUPPER(c) ((jschar) ((JS_CCODE(c) & 0x00100000) \
424 ? (c) - ((int32)JS_CCODE(c) >> 22) \
425 : (c)))
426 #define JS_TOLOWER(c) ((jschar) ((JS_CCODE(c) & 0x00200000) \
427 ? (c) + ((int32)JS_CCODE(c) >> 22) \
428 : (c)))
429
430 /*
431 * Shorthands for ASCII (7-bit) decimal and hex conversion.
432 * Manually inline isdigit for performance; MSVC doesn't do this for us.
433 */
434 #define JS7_ISDEC(c) ((((unsigned)(c)) - '0') <= 9)
435 #define JS7_UNDEC(c) ((c) - '0')
436 #define JS7_ISHEX(c) ((c) < 128 && isxdigit(c))
437 #define JS7_UNHEX(c) (uintN)(JS7_ISDEC(c) ? (c) - '0' : 10 + tolower(c) - 'a')
438 #define JS7_ISLET(c) ((c) < 128 && isalpha(c))
439
440 /* Initialize per-runtime string state for the first context in the runtime. */
441 extern JSBool
442 js_InitRuntimeStringState(JSContext *cx);
443
444 extern JSBool
445 js_InitDeflatedStringCache(JSRuntime *rt);
446
447 /*
448 * Maximum character code for which we will create a pinned unit string on
449 * demand -- see JSRuntime.unitStrings in jscntxt.h.
450 */
451 #define UNIT_STRING_LIMIT 256U
452
453 /*
454 * Get the independent string containing only character code at index in str
455 * (backstopped with a zero character as usual for independent strings).
456 */
457 extern JSString *
458 js_GetUnitString(JSContext *cx, JSString *str, size_t index);
459
460 /*
461 * Get the independent string containing only the character code c, which must
462 * be less than UNIT_STRING_LIMIT.
463 */
464 extern JSString *
465 js_GetUnitStringForChar(JSContext *cx, jschar c);
466
467 extern void
468 js_FinishUnitStrings(JSRuntime *rt);
469
470 extern void
471 js_FinishRuntimeStringState(JSContext *cx);
472
473 extern void
474 js_FinishDeflatedStringCache(JSRuntime *rt);
475
476 /* Initialize the String class, returning its prototype object. */
477 extern JSClass js_StringClass;
478
479 extern JSObject *
480 js_InitStringClass(JSContext *cx, JSObject *obj);
481
482 extern const char js_escape_str[];
483 extern const char js_unescape_str[];
484 extern const char js_uneval_str[];
485 extern const char js_decodeURI_str[];
486 extern const char js_encodeURI_str[];
487 extern const char js_decodeURIComponent_str[];
488 extern const char js_encodeURIComponent_str[];
489
490 /* GC-allocate a string descriptor for the given malloc-allocated chars. */
491 extern JSString *
492 js_NewString(JSContext *cx, jschar *chars, size_t length);
493
494 extern JSString *
495 js_NewDependentString(JSContext *cx, JSString *base, size_t start,
496 size_t length);
497
498 /* Copy a counted string and GC-allocate a descriptor for it. */
499 extern JSString *
500 js_NewStringCopyN(JSContext *cx, const jschar *s, size_t n);
501
502 /* Copy a C string and GC-allocate a descriptor for it. */
503 extern JSString *
504 js_NewStringCopyZ(JSContext *cx, const jschar *s);
505
506 /*
507 * Free the chars held by str when it is finalized by the GC. When type is
508 * less then zero, it denotes an internal string. Otherwise it denotes the
509 * type of the external string allocated with JS_NewExternalString.
510 *
511 * This function always needs rt but can live with null cx.
512 */
513 extern void
514 js_FinalizeStringRT(JSRuntime *rt, JSString *str, intN type, JSContext *cx);
515
516 /*
517 * Convert a value to a printable C string.
518 */
519 typedef JSString *(*JSValueToStringFun)(JSContext *cx, jsval v);
520
521 extern JS_FRIEND_API(const char *)
522 js_ValueToPrintable(JSContext *cx, jsval v, JSValueToStringFun v2sfun);
523
524 #define js_ValueToPrintableString(cx,v) \
525 js_ValueToPrintable(cx, v, js_ValueToString)
526
527 #define js_ValueToPrintableSource(cx,v) \
528 js_ValueToPrintable(cx, v, js_ValueToSource)
529
530 /*
531 * Convert a value to a string, returning null after reporting an error,
532 * otherwise returning a new string reference.
533 */
534 extern JS_FRIEND_API(JSString *)
535 js_ValueToString(JSContext *cx, jsval v);
536
537 /*
538 * Convert a value to its source expression, returning null after reporting
539 * an error, otherwise returning a new string reference.
540 */
541 extern JS_FRIEND_API(JSString *)
542 js_ValueToSource(JSContext *cx, jsval v);
543
544 /*
545 * Compute a hash function from str. The caller can call this function even if
546 * str is not a GC-allocated thing.
547 */
548 extern uint32
549 js_HashString(JSString *str);
550
551 /*
552 * Test if strings are equal. The caller can call the function even if str1
553 * or str2 are not GC-allocated things.
554 */
555 extern JSBool JS_FASTCALL
556 js_EqualStrings(JSString *str1, JSString *str2);
557
558 /*
559 * Return less than, equal to, or greater than zero depending on whether
560 * str1 is less than, equal to, or greater than str2.
561 */
562 extern int32 JS_FASTCALL
563 js_CompareStrings(JSString *str1, JSString *str2);
564
565 /*
566 * Boyer-Moore-Horspool superlinear search for pat:patlen in text:textlen.
567 * The patlen argument must be positive and no greater than BMH_PATLEN_MAX.
568 * The start argument tells where in text to begin the search.
569 *
570 * Return the index of pat in text, or -1 if not found.
571 */
572 #define BMH_CHARSET_SIZE 256 /* ISO-Latin-1 */
573 #define BMH_PATLEN_MAX 255 /* skip table element is uint8 */
574
575 #define BMH_BAD_PATTERN (-2) /* return value if pat is not ISO-Latin-1 */
576
577 extern jsint
578 js_BoyerMooreHorspool(const jschar *text, jsint textlen,
579 const jschar *pat, jsint patlen,
580 jsint start);
581
582 extern size_t
583 js_strlen(const jschar *s);
584
585 extern jschar *
586 js_strchr(const jschar *s, jschar c);
587
588 extern jschar *
589 js_strchr_limit(const jschar *s, jschar c, const jschar *limit);
590
591 #define js_strncpy(t, s, n) memcpy((t), (s), (n) * sizeof(jschar))
592
593 /*
594 * Return s advanced past any Unicode white space characters.
595 */
596 extern const jschar *
597 js_SkipWhiteSpace(const jschar *s, const jschar *end);
598
599 /*
600 * Inflate bytes to JS chars and vice versa. Report out of memory via cx
601 * and return null on error, otherwise return the jschar or byte vector that
602 * was JS_malloc'ed. length is updated with the length of the new string in jschars.
603 */
604 extern jschar *
605 js_InflateString(JSContext *cx, const char *bytes, size_t *length);
606
607 extern char *
608 js_DeflateString(JSContext *cx, const jschar *chars, size_t length);
609
610 /*
611 * Inflate bytes to JS chars into a buffer. 'chars' must be large enough for
612 * 'length' jschars. The buffer is NOT null-terminated. The destination length
613 * must be be initialized with the buffer size and will contain on return the
614 * number of copied chars.
615 */
616 extern JSBool
617 js_InflateStringToBuffer(JSContext* cx, const char *bytes, size_t length,
618 jschar *chars, size_t* charsLength);
619
620 /*
621 * Get number of bytes in the deflated sequence of characters.
622 */
623 extern size_t
624 js_GetDeflatedStringLength(JSContext *cx, const jschar *chars,
625 size_t charsLength);
626
627 /*
628 * Deflate JS chars to bytes into a buffer. 'bytes' must be large enough for
629 * 'length chars. The buffer is NOT null-terminated. The destination length
630 * must to be initialized with the buffer size and will contain on return the
631 * number of copied bytes.
632 */
633 extern JSBool
634 js_DeflateStringToBuffer(JSContext* cx, const jschar *chars,
635 size_t charsLength, char *bytes, size_t* length);
636
637 /*
638 * Associate bytes with str in the deflated string cache, returning true on
639 * successful association, false on out of memory.
640 */
641 extern JSBool
642 js_SetStringBytes(JSContext *cx, JSString *str, char *bytes, size_t length);
643
644 /*
645 * Find or create a deflated string cache entry for str that contains its
646 * characters chopped from Unicode code points into bytes.
647 */
648 extern const char *
649 js_GetStringBytes(JSContext *cx, JSString *str);
650
651 /* Remove a deflated string cache entry associated with str if any. */
652 extern void
653 js_PurgeDeflatedStringCache(JSRuntime *rt, JSString *str);
654
655 /* Export a few natives and a helper to other files in SpiderMonkey. */
656 extern JSBool
657 js_str_escape(JSContext *cx, JSObject *obj, uintN argc, jsval *argv,
658 jsval *rval);
659
660 extern JSBool
661 js_StringReplaceHelper(JSContext *cx, uintN argc, JSObject *lambda,
662 JSString *repstr, jsval *vp);
663
664 /*
665 * Convert one UCS-4 char and write it into a UTF-8 buffer, which must be at
666 * least 6 bytes long. Return the number of UTF-8 bytes of data written.
667 */
668 extern int
669 js_OneUcs4ToUtf8Char(uint8 *utf8Buffer, uint32 ucs4Char);
670
671 /*
672 * Write str into buffer escaping any non-printable or non-ASCII character.
673 * Guarantees that a NUL is at the end of the buffer. Returns the length of
674 * the written output, NOT including the NUL. If buffer is null, just returns
675 * the length of the output. If quote is not 0, it must be a single or double
676 * quote character that will quote the output.
677 *
678 * The function is only defined for debug builds.
679 */
680 #define js_PutEscapedString(buffer, bufferSize, str, quote) \
681 js_PutEscapedStringImpl(buffer, bufferSize, NULL, str, quote)
682
683 /*
684 * Write str into file escaping any non-printable or non-ASCII character.
685 * Returns the number of bytes written to file. If quote is not 0, it must
686 * be a single or double quote character that will quote the output.
687 *
688 * The function is only defined for debug builds.
689 */
690 #define js_FileEscapedString(file, str, quote) \
691 (JS_ASSERT(file), js_PutEscapedStringImpl(NULL, 0, file, str, quote))
692
693 extern JS_FRIEND_API(size_t)
694 js_PutEscapedStringImpl(char *buffer, size_t bufferSize, FILE *fp,
695 JSString *str, uint32 quote);
696
697 extern JSBool
698 js_String(JSContext *cx, JSObject *obj, uintN argc, jsval *argv, jsval *rval);
699
700 JS_END_EXTERN_C
701
702 #endif /* jsstr_h___ */

  ViewVC Help
Powered by ViewVC 1.1.24