ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/JSON-XS/XS.xs
(Generate patch)

Comparing JSON-XS/XS.xs (file contents):
Revision 1.8 by root, Fri Mar 23 16:13:59 2007 UTC vs.
Revision 1.15 by root, Sun Mar 25 02:32:40 2007 UTC

… …
10#define F_UTF8 0x00000002 10#define F_UTF8 0x00000002
11#define F_INDENT 0x00000004 11#define F_INDENT 0x00000004
12#define F_CANONICAL 0x00000008 12#define F_CANONICAL 0x00000008
13#define F_SPACE_BEFORE 0x00000010 13#define F_SPACE_BEFORE 0x00000010
14#define F_SPACE_AFTER 0x00000020 14#define F_SPACE_AFTER 0x00000020
15#define F_JSON_RPC 0x00000040
16#define F_ALLOW_NONREF 0x00000080 15#define F_ALLOW_NONREF 0x00000080
17#define F_SHRINK 0x00000100 16#define F_SHRINK 0x00000100
18 17
19#define F_PRETTY F_INDENT | F_SPACE_BEFORE | F_SPACE_AFTER 18#define F_PRETTY F_INDENT | F_SPACE_BEFORE | F_SPACE_AFTER
20#define F_DEFAULT 0 19#define F_DEFAULT 0
21 20
22#define INIT_SIZE 32 // initial scalar size to be allocated 21#define INIT_SIZE 32 // initial scalar size to be allocated
22#define INDENT_STEP 3 // spaces per indentation level
23
24#define UTF8_MAX_LEN 11 // for perls UTF-X: max. number of octets per character
25#define SHORT_STRING_LEN 512 // special-case strings of up to this size
23 26
24#define SB do { 27#define SB do {
25#define SE } while (0) 28#define SE } while (0)
26 29
27static HV *json_stash; 30static HV *json_stash; // JSON::XS::
31
32/////////////////////////////////////////////////////////////////////////////
33// utility functions
34
35static UV *
36SvJSON (SV *sv)
37{
38 if (!(SvROK (sv) && SvOBJECT (SvRV (sv)) && SvSTASH (SvRV (sv)) == json_stash))
39 croak ("object is not of type JSON::XS");
40
41 return &SvUVX (SvRV (sv));
42}
43
44static void
45shrink (SV *sv)
46{
47 sv_utf8_downgrade (sv, 1);
48 if (SvLEN (sv) > SvCUR (sv) + 1)
49 {
50#ifdef SvPV_shrink_to_cur
51 SvPV_shrink_to_cur (sv);
52#elif defined (SvPV_renew)
53 SvPV_renew (sv, SvCUR (sv) + 1);
54#endif
55 }
56}
57
58// decode an utf-8 character and return it, or (UV)-1 in
59// case of an error.
60// we special-case "safe" characters from U+80 .. U+7FF,
61// but use the very good perl function to parse anything else.
62// note that we never call this function for a ascii codepoints
63static UV
64decode_utf8 (unsigned char *s, STRLEN len, STRLEN *clen)
65{
66 if (s[0] > 0xdf || s[0] < 0xc2)
67 return utf8n_to_uvuni (s, len, clen, UTF8_CHECK_ONLY);
68 else if (len > 1 && s[1] >= 0x80 && s[1] <= 0xbf)
69 {
70 *clen = 2;
71 return ((s[0] & 0x1f) << 6) | (s[1] & 0x3f);
72 }
73 else
74 return (UV)-1;
75}
76
77/////////////////////////////////////////////////////////////////////////////
78// encoder
28 79
29// structure used for encoding JSON 80// structure used for encoding JSON
30typedef struct 81typedef struct
31{ 82{
32 char *cur; 83 char *cur; // SvPVX (sv) + current output position
33 STRLEN len; // SvLEN (sv)
34 char *end; // SvEND (sv) 84 char *end; // SvEND (sv)
35 SV *sv; 85 SV *sv; // result scalar
36 UV flags; 86 UV flags; // F_*
37 int max_recurse; 87 int indent; // indentation level
38 int indent; 88 int max_depth; // max. recursion level
39} enc_t; 89} enc_t;
40
41// structure used for decoding JSON
42typedef struct
43{
44 char *cur;
45 char *end;
46 const char *err;
47 UV flags;
48} dec_t;
49
50static UV *
51SvJSON (SV *sv)
52{
53 if (!(SvROK (sv) && SvOBJECT (SvRV (sv)) && SvSTASH (SvRV (sv)) == json_stash))
54 croak ("object is not of type JSON::XS");
55
56 return &SvUVX (SvRV (sv));
57}
58
59static void
60shrink (SV *sv)
61{
62 sv_utf8_downgrade (sv, 1);
63#ifdef SvPV_shrink_to_cur
64 SvPV_shrink_to_cur (sv);
65#endif
66}
67
68/////////////////////////////////////////////////////////////////////////////
69 90
70static void 91static void
71need (enc_t *enc, STRLEN len) 92need (enc_t *enc, STRLEN len)
72{ 93{
73 if (enc->cur + len >= enc->end) 94 if (enc->cur + len >= enc->end)
… …
131 STRLEN clen; 152 STRLEN clen;
132 UV uch; 153 UV uch;
133 154
134 if (is_utf8) 155 if (is_utf8)
135 { 156 {
136 uch = utf8n_to_uvuni (str, end - str, &clen, UTF8_CHECK_ONLY); 157 //uch = utf8n_to_uvuni (str, end - str, &clen, UTF8_CHECK_ONLY);
158 uch = decode_utf8 (str, end - str, &clen);
137 if (clen == (STRLEN)-1) 159 if (clen == (STRLEN)-1)
138 croak ("malformed UTF-8 character in string, cannot convert to JSON"); 160 croak ("malformed or illegal unicode character in string [%.11s], cannot convert to JSON", str);
139 } 161 }
140 else 162 else
141 { 163 {
142 uch = ch; 164 uch = ch;
143 clen = 1; 165 clen = 1;
144 } 166 }
145 167
168 if (uch > 0x10FFFFUL)
169 croak ("out of range codepoint (0x%lx) encountered, unrepresentable in JSON", (unsigned long)uch);
170
146 if (uch < 0x80 || enc->flags & F_ASCII) 171 if (uch < 0x80 || enc->flags & F_ASCII)
147 { 172 {
148 if (uch > 0xFFFFUL) 173 if (uch > 0xFFFFUL)
149 { 174 {
150 need (enc, len += 11); 175 need (enc, len += 11);
151 sprintf (enc->cur, "\\u%04x\\u%04x", 176 sprintf (enc->cur, "\\u%04x\\u%04x",
152 (uch - 0x10000) / 0x400 + 0xD800, 177 (int)((uch - 0x10000) / 0x400 + 0xD800),
153 (uch - 0x10000) % 0x400 + 0xDC00); 178 (int)((uch - 0x10000) % 0x400 + 0xDC00));
154 enc->cur += 12; 179 enc->cur += 12;
155 } 180 }
156 else 181 else
157 { 182 {
158 static char hexdigit [16] = "0123456789abcdef"; 183 static char hexdigit [16] = "0123456789abcdef";
… …
176 } 201 }
177 while (--clen); 202 while (--clen);
178 } 203 }
179 else 204 else
180 { 205 {
181 need (enc, len += 10); // never more than 11 bytes needed 206 need (enc, len += UTF8_MAX_LEN - 1); // never more than 11 bytes needed
182 enc->cur = uvuni_to_utf8_flags (enc->cur, uch, 0); 207 enc->cur = uvuni_to_utf8_flags (enc->cur, uch, 0);
183 ++str; 208 ++str;
184 } 209 }
185 } 210 }
186 } 211 }
… …
188 213
189 --len; 214 --len;
190 } 215 }
191} 216}
192 217
193#define INDENT SB \ 218static void
219encode_indent (enc_t *enc)
220{
194 if (enc->flags & F_INDENT) \ 221 if (enc->flags & F_INDENT)
195 { \ 222 {
196 int i_; \ 223 int spaces = enc->indent * INDENT_STEP;
197 need (enc, enc->indent); \
198 for (i_ = enc->indent * 3; i_--; )\
199 encode_ch (enc, ' '); \
200 } \
201 SE
202 224
203#define SPACE SB need (enc, 1); encode_ch (enc, ' '); SE 225 need (enc, spaces);
204#define NL SB if (enc->flags & F_INDENT) { need (enc, 1); encode_ch (enc, '\n'); } SE 226 memset (enc->cur, ' ', spaces);
205#define COMMA SB \ 227 enc->cur += spaces;
228 }
229}
230
231static void
232encode_space (enc_t *enc)
233{
234 need (enc, 1);
206 encode_ch (enc, ','); \ 235 encode_ch (enc, ' ');
236}
237
238static void
239encode_nl (enc_t *enc)
240{
207 if (enc->flags & F_INDENT) \ 241 if (enc->flags & F_INDENT)
208 NL; \ 242 {
243 need (enc, 1);
244 encode_ch (enc, '\n');
245 }
246}
247
248static void
249encode_comma (enc_t *enc)
250{
251 encode_ch (enc, ',');
252
253 if (enc->flags & F_INDENT)
254 encode_nl (enc);
209 else if (enc->flags & F_SPACE_AFTER) \ 255 else if (enc->flags & F_SPACE_AFTER)
210 SPACE; \ 256 encode_space (enc);
211 SE 257}
212 258
213static void encode_sv (enc_t *enc, SV *sv); 259static void encode_sv (enc_t *enc, SV *sv);
214 260
215static void 261static void
216encode_av (enc_t *enc, AV *av) 262encode_av (enc_t *enc, AV *av)
217{ 263{
218 int i, len = av_len (av); 264 int i, len = av_len (av);
219 265
220 encode_ch (enc, '['); NL; 266 encode_ch (enc, '['); encode_nl (enc);
221 ++enc->indent; 267 ++enc->indent;
222 268
223 for (i = 0; i <= len; ++i) 269 for (i = 0; i <= len; ++i)
224 { 270 {
225 INDENT; 271 encode_indent (enc);
226 encode_sv (enc, *av_fetch (av, i, 0)); 272 encode_sv (enc, *av_fetch (av, i, 0));
227 273
228 if (i < len) 274 if (i < len)
229 COMMA; 275 encode_comma (enc);
230 } 276 }
231 277
232 NL; 278 encode_nl (enc);
233 279
234 --enc->indent; 280 --enc->indent;
235 INDENT; encode_ch (enc, ']'); 281 encode_indent (enc); encode_ch (enc, ']');
236} 282}
237 283
238static void 284static void
239encode_he (enc_t *enc, HE *he) 285encode_he (enc_t *enc, HE *he)
240{ 286{
… …
254 else 300 else
255 encode_str (enc, HeKEY (he), HeKLEN (he), HeKUTF8 (he)); 301 encode_str (enc, HeKEY (he), HeKLEN (he), HeKUTF8 (he));
256 302
257 encode_ch (enc, '"'); 303 encode_ch (enc, '"');
258 304
259 if (enc->flags & F_SPACE_BEFORE) SPACE; 305 if (enc->flags & F_SPACE_BEFORE) encode_space (enc);
260 encode_ch (enc, ':'); 306 encode_ch (enc, ':');
261 if (enc->flags & F_SPACE_AFTER ) SPACE; 307 if (enc->flags & F_SPACE_AFTER ) encode_space (enc);
262 encode_sv (enc, HeVAL (he)); 308 encode_sv (enc, HeVAL (he));
263} 309}
264 310
265// compare hash entries, used when all keys are bytestrings 311// compare hash entries, used when all keys are bytestrings
266static int 312static int
… …
272 HE *b = *(HE **)b_; 318 HE *b = *(HE **)b_;
273 319
274 STRLEN la = HeKLEN (a); 320 STRLEN la = HeKLEN (a);
275 STRLEN lb = HeKLEN (b); 321 STRLEN lb = HeKLEN (b);
276 322
277 if (!(cmp == memcmp (HeKEY (a), HeKEY (b), la < lb ? la : lb))) 323 if (!(cmp = memcmp (HeKEY (a), HeKEY (b), la < lb ? la : lb)))
278 cmp = la < lb ? -1 : la == lb ? 0 : 1; 324 cmp = la - lb;
279 325
280 return cmp; 326 return cmp;
281} 327}
282 328
283// compare hash entries, used when some keys are sv's or utf-x 329// compare hash entries, used when some keys are sv's or utf-x
… …
290static void 336static void
291encode_hv (enc_t *enc, HV *hv) 337encode_hv (enc_t *enc, HV *hv)
292{ 338{
293 int count, i; 339 int count, i;
294 340
295 encode_ch (enc, '{'); NL; ++enc->indent; 341 encode_ch (enc, '{'); encode_nl (enc); ++enc->indent;
296 342
297 if ((count = hv_iterinit (hv))) 343 if ((count = hv_iterinit (hv)))
298 { 344 {
299 // for canonical output we have to sort by keys first 345 // for canonical output we have to sort by keys first
300 // actually, this is mostly due to the stupid so-called 346 // actually, this is mostly due to the stupid so-called
301 // security workaround added somewhere in 5.8.x. 347 // security workaround added somewhere in 5.8.x.
302 // that randomises hash orderings 348 // that randomises hash orderings
303 if (enc->flags & F_CANONICAL) 349 if (enc->flags & F_CANONICAL)
304 { 350 {
305 HE *he, *hes [count]; 351 HE *he, *hes [count]; // if your compiler dies here, you need to enable C99 mode
306 int fast = 1; 352 int fast = 1;
307 353
308 i = 0; 354 i = 0;
309 while ((he = hv_iternext (hv))) 355 while ((he = hv_iternext (hv)))
310 { 356 {
… …
335 LEAVE; 381 LEAVE;
336 } 382 }
337 383
338 for (i = 0; i < count; ++i) 384 for (i = 0; i < count; ++i)
339 { 385 {
340 INDENT; 386 encode_indent (enc);
341 encode_he (enc, hes [i]); 387 encode_he (enc, hes [i]);
342 388
343 if (i < count - 1) 389 if (i < count - 1)
344 COMMA; 390 encode_comma (enc);
345 } 391 }
346 392
347 NL; 393 encode_nl (enc);
348 } 394 }
349 else 395 else
350 { 396 {
351 SV *sv; 397 SV *sv;
352 HE *he = hv_iternext (hv); 398 HE *he = hv_iternext (hv);
353 399
354 for (;;) 400 for (;;)
355 { 401 {
356 INDENT; 402 encode_indent (enc);
357 encode_he (enc, he); 403 encode_he (enc, he);
358 404
359 if (!(he = hv_iternext (hv))) 405 if (!(he = hv_iternext (hv)))
360 break; 406 break;
361 407
362 COMMA; 408 encode_comma (enc);
363 } 409 }
364 410
365 NL; 411 encode_nl (enc);
366 } 412 }
367 } 413 }
368 414
369 --enc->indent; INDENT; encode_ch (enc, '}'); 415 --enc->indent; encode_indent (enc); encode_ch (enc, '}');
370} 416}
371 417
372static void 418static void
373encode_sv (enc_t *enc, SV *sv) 419encode_sv (enc_t *enc, SV *sv)
374{ 420{
… …
396 ? snprintf (enc->cur, 64, "%"UVuf, (UV)SvUVX (sv)) 442 ? snprintf (enc->cur, 64, "%"UVuf, (UV)SvUVX (sv))
397 : snprintf (enc->cur, 64, "%"IVdf, (IV)SvIVX (sv)); 443 : snprintf (enc->cur, 64, "%"IVdf, (IV)SvIVX (sv));
398 } 444 }
399 else if (SvROK (sv)) 445 else if (SvROK (sv))
400 { 446 {
401 if (!--enc->max_recurse) 447 SV *rv = SvRV (sv);
448
449 if (enc->indent >= enc->max_depth)
402 croak ("data structure too deep (hit recursion limit)"); 450 croak ("data structure too deep (hit recursion limit)");
403 451
404 sv = SvRV (sv);
405
406 switch (SvTYPE (sv)) 452 switch (SvTYPE (rv))
407 { 453 {
408 case SVt_PVAV: encode_av (enc, (AV *)sv); break; 454 case SVt_PVAV: encode_av (enc, (AV *)rv); break;
409 case SVt_PVHV: encode_hv (enc, (HV *)sv); break; 455 case SVt_PVHV: encode_hv (enc, (HV *)rv); break;
410 456
411 default: 457 default:
412 croak ("JSON can only represent references to arrays or hashes"); 458 croak ("encountered %s, but JSON can only represent references to arrays or hashes",
459 SvPV_nolen (sv));
413 } 460 }
414 } 461 }
415 else if (!SvOK (sv)) 462 else if (!SvOK (sv))
416 encode_str (enc, "null", 4, 0); 463 encode_str (enc, "null", 4, 0);
417 else 464 else
418 croak ("encountered perl type that JSON cannot handle"); 465 croak ("encountered perl type (%s,0x%x) that JSON cannot handle, you might want to report this",
466 SvPV_nolen (sv), SvFLAGS (sv));
419} 467}
420 468
421static SV * 469static SV *
422encode_json (SV *scalar, UV flags) 470encode_json (SV *scalar, UV flags)
423{ 471{
424 if (!(flags & F_ALLOW_NONREF) && !SvROK (scalar)) 472 if (!(flags & F_ALLOW_NONREF) && !SvROK (scalar))
425 croak ("hash- or arraref required (not a simple scalar, use allow_nonref to allow this)"); 473 croak ("hash- or arrayref expected (not a simple scalar, use allow_nonref to allow this)");
426 474
427 enc_t enc; 475 enc_t enc;
428 enc.flags = flags; 476 enc.flags = flags;
429 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 477 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
430 enc.cur = SvPVX (enc.sv); 478 enc.cur = SvPVX (enc.sv);
431 enc.end = SvEND (enc.sv); 479 enc.end = SvEND (enc.sv);
432 enc.max_recurse = 0;
433 enc.indent = 0; 480 enc.indent = 0;
481 enc.max_depth = 0x7fffffffUL;
434 482
435 SvPOK_only (enc.sv); 483 SvPOK_only (enc.sv);
436 encode_sv (&enc, scalar); 484 encode_sv (&enc, scalar);
437 485
438 if (!(flags & (F_ASCII | F_UTF8))) 486 if (!(flags & (F_ASCII | F_UTF8)))
… …
445 493
446 return enc.sv; 494 return enc.sv;
447} 495}
448 496
449///////////////////////////////////////////////////////////////////////////// 497/////////////////////////////////////////////////////////////////////////////
498// decoder
450 499
451#define WS \ 500// structure used for decoding JSON
501typedef struct
502{
503 char *cur; // current parser pointer
504 char *end; // end of input string
505 const char *err; // parse error, if != 0
506 UV flags; // F_*
507} dec_t;
508
509static void
510decode_ws (dec_t *dec)
511{
452 for (;;) \ 512 for (;;)
453 { \ 513 {
454 char ch = *dec->cur; \ 514 char ch = *dec->cur;
515
455 if (ch > 0x20 \ 516 if (ch > 0x20
456 || (ch != 0x20 && ch != 0x0a && ch != 0x0d && ch != 0x09)) \ 517 || (ch != 0x20 && ch != 0x0a && ch != 0x0d && ch != 0x09))
457 break; \ 518 break;
519
458 ++dec->cur; \ 520 ++dec->cur;
459 } 521 }
522}
460 523
461#define ERR(reason) SB dec->err = reason; goto fail; SE 524#define ERR(reason) SB dec->err = reason; goto fail; SE
462#define EXPECT_CH(ch) SB \ 525#define EXPECT_CH(ch) SB \
463 if (*dec->cur != ch) \ 526 if (*dec->cur != ch) \
464 ERR (# ch " expected"); \ 527 ERR (# ch " expected"); \
… …
471 534
472static UV 535static UV
473decode_4hex (dec_t *dec) 536decode_4hex (dec_t *dec)
474{ 537{
475 signed char d1, d2, d3, d4; 538 signed char d1, d2, d3, d4;
539 unsigned char *cur = (unsigned char *)dec->cur;
476 540
477 d1 = decode_hexdigit [((unsigned char *)dec->cur) [0]];
478 if (d1 < 0) ERR ("four hexadecimal digits expected"); 541 d1 = decode_hexdigit [cur [0]]; if (d1 < 0) ERR ("four hexadecimal digits expected");
479 d2 = decode_hexdigit [((unsigned char *)dec->cur) [1]];
480 if (d2 < 0) ERR ("four hexadecimal digits expected"); 542 d2 = decode_hexdigit [cur [1]]; if (d2 < 0) ERR ("four hexadecimal digits expected");
481 d3 = decode_hexdigit [((unsigned char *)dec->cur) [2]];
482 if (d3 < 0) ERR ("four hexadecimal digits expected"); 543 d3 = decode_hexdigit [cur [2]]; if (d3 < 0) ERR ("four hexadecimal digits expected");
483 d4 = decode_hexdigit [((unsigned char *)dec->cur) [3]];
484 if (d4 < 0) ERR ("four hexadecimal digits expected"); 544 d4 = decode_hexdigit [cur [3]]; if (d4 < 0) ERR ("four hexadecimal digits expected");
485 545
486 dec->cur += 4; 546 dec->cur += 4;
487 547
488 return ((UV)d1) << 12 548 return ((UV)d1) << 12
489 | ((UV)d2) << 8 549 | ((UV)d2) << 8
… …
492 552
493fail: 553fail:
494 return (UV)-1; 554 return (UV)-1;
495} 555}
496 556
497#define APPEND_GROW(n) SB \
498 if (cur + (n) >= end) \
499 { \
500 STRLEN ofs = cur - SvPVX (sv); \
501 SvGROW (sv, ofs + (n) + 1); \
502 cur = SvPVX (sv) + ofs; \
503 end = SvEND (sv); \
504 } \
505 SE
506
507#define APPEND_CH(ch) SB \
508 APPEND_GROW (1); \
509 *cur++ = (ch); \
510 SE
511
512static SV * 557static SV *
513decode_str (dec_t *dec) 558decode_str (dec_t *dec)
514{ 559{
515 SV *sv = NEWSV (0,2); 560 SV *sv = 0;
516 int utf8 = 0; 561 int utf8 = 0;
517 char *cur = SvPVX (sv);
518 char *end = SvEND (sv);
519 562
520 for (;;) 563 do
521 { 564 {
522 unsigned char ch = *(unsigned char *)dec->cur; 565 char buf [SHORT_STRING_LEN + UTF8_MAX_LEN];
566 char *cur = buf;
523 567
524 if (ch == '"') 568 do
525 break;
526 else if (ch == '\\')
527 { 569 {
528 switch (*++dec->cur) 570 unsigned char ch = *(unsigned char *)dec->cur++;
571
572 if (ch == '"')
529 { 573 {
530 case '\\': 574 --dec->cur;
531 case '/': 575 break;
532 case '"': APPEND_CH (*dec->cur++); break; 576 }
533 577 else if (ch == '\\')
534 case 'b': APPEND_CH ('\010'); ++dec->cur; break; 578 {
535 case 't': APPEND_CH ('\011'); ++dec->cur; break; 579 switch (*dec->cur)
536 case 'n': APPEND_CH ('\012'); ++dec->cur; break;
537 case 'f': APPEND_CH ('\014'); ++dec->cur; break;
538 case 'r': APPEND_CH ('\015'); ++dec->cur; break;
539
540 case 'u':
541 { 580 {
542 UV lo, hi; 581 case '\\':
543 ++dec->cur; 582 case '/':
583 case '"': *cur++ = *dec->cur++; break;
544 584
545 hi = decode_4hex (dec); 585 case 'b': ++dec->cur; *cur++ = '\010'; break;
546 if (hi == (UV)-1) 586 case 't': ++dec->cur; *cur++ = '\011'; break;
547 goto fail; 587 case 'n': ++dec->cur; *cur++ = '\012'; break;
588 case 'f': ++dec->cur; *cur++ = '\014'; break;
589 case 'r': ++dec->cur; *cur++ = '\015'; break;
548 590
549 // possibly a surrogate pair 591 case 'u':
550 if (hi >= 0xd800 && hi < 0xdc00)
551 { 592 {
552 if (dec->cur [0] != '\\' || dec->cur [1] != 'u') 593 UV lo, hi;
553 ERR ("missing low surrogate character in surrogate pair");
554
555 dec->cur += 2; 594 ++dec->cur;
556 595
557 lo = decode_4hex (dec); 596 hi = decode_4hex (dec);
558 if (lo == (UV)-1) 597 if (hi == (UV)-1)
559 goto fail; 598 goto fail;
560 599
600 // possibly a surrogate pair
601 if (hi >= 0xd800)
602 if (hi < 0xdc00)
603 {
604 if (dec->cur [0] != '\\' || dec->cur [1] != 'u')
605 ERR ("missing low surrogate character in surrogate pair");
606
607 dec->cur += 2;
608
609 lo = decode_4hex (dec);
610 if (lo == (UV)-1)
611 goto fail;
612
561 if (lo < 0xdc00 || lo >= 0xe000) 613 if (lo < 0xdc00 || lo >= 0xe000)
562 ERR ("surrogate pair expected"); 614 ERR ("surrogate pair expected");
563 615
564 hi = (hi - 0xD800) * 0x400 + (lo - 0xDC00) + 0x10000; 616 hi = (hi - 0xD800) * 0x400 + (lo - 0xDC00) + 0x10000;
617 }
618 else if (hi < 0xe000)
619 ERR ("missing high surrogate character in surrogate pair");
620
621 if (hi >= 0x80)
622 {
623 utf8 = 1;
624
625 cur = (char *)uvuni_to_utf8_flags (cur, hi, 0);
626 }
627 else
628 *cur++ = hi;
565 } 629 }
566 else if (hi >= 0xdc00 && hi < 0xe000)
567 ERR ("missing high surrogate character in surrogate pair");
568
569 if (hi >= 0x80)
570 { 630 break;
571 utf8 = 1;
572 631
573 APPEND_GROW (4); // at most 4 bytes for 21 bits
574 cur = (char *)uvuni_to_utf8_flags (cur, hi, 0);
575 }
576 else 632 default:
577 APPEND_CH (hi); 633 --dec->cur;
634 ERR ("illegal backslash escape sequence in string");
578 } 635 }
579 break;
580
581 default:
582 --dec->cur;
583 ERR ("illegal backslash escape sequence in string");
584 } 636 }
637 else if (ch >= 0x20 && ch <= 0x7f)
638 *cur++ = ch;
639 else if (ch >= 0x80)
640 {
641 --dec->cur;
642
643 STRLEN clen;
644 UV uch = decode_utf8 (dec->cur, dec->end - dec->cur, &clen);
645 if (clen == (STRLEN)-1)
646 ERR ("malformed UTF-8 character in JSON string");
647
648 do
649 {
650 *cur++ = *dec->cur++;
651 }
652 while (--clen);
653
654 utf8 = 1;
655 }
656 else if (!ch)
657 ERR ("unexpected end of string while parsing json string");
658 else
659 ERR ("invalid character encountered");
660
585 } 661 }
586 else if (ch >= 0x20 && ch <= 0x7f) 662 while (cur < buf + SHORT_STRING_LEN);
587 APPEND_CH (*dec->cur++); 663
588 else if (ch >= 0x80) 664 STRLEN len = cur - buf;
665
666 if (sv)
589 { 667 {
590 STRLEN clen; 668 SvGROW (sv, SvCUR (sv) + len + 1);
591 UV uch = utf8n_to_uvuni (dec->cur, dec->end - dec->cur, &clen, UTF8_CHECK_ONLY); 669 memcpy (SvPVX (sv) + SvCUR (sv), buf, len);
592 if (clen == (STRLEN)-1) 670 SvCUR_set (sv, SvCUR (sv) + len);
593 ERR ("malformed UTF-8 character in JSON string");
594
595 APPEND_GROW (clen);
596 do
597 {
598 *cur++ = *dec->cur++;
599 }
600 while (--clen);
601
602 utf8 = 1;
603 } 671 }
604 else if (dec->cur == dec->end)
605 ERR ("unexpected end of string while parsing json string");
606 else 672 else
607 ERR ("invalid character encountered"); 673 sv = newSVpvn (buf, len);
608 } 674 }
675 while (*dec->cur != '"');
609 676
610 ++dec->cur; 677 ++dec->cur;
611 678
612 SvCUR_set (sv, cur - SvPVX (sv)); 679 if (sv)
613 680 {
614 SvPOK_only (sv); 681 SvPOK_only (sv);
615 *SvEND (sv) = 0; 682 *SvEND (sv) = 0;
616 683
617 if (utf8) 684 if (utf8)
618 SvUTF8_on (sv); 685 SvUTF8_on (sv);
619 686 }
620 if (dec->flags & F_SHRINK) 687 else
621 shrink (sv); 688 sv = newSVpvn ("", 0);
622 689
623 return sv; 690 return sv;
624 691
625fail: 692fail:
626 SvREFCNT_dec (sv);
627 return 0; 693 return 0;
628} 694}
629 695
630static SV * 696static SV *
631decode_num (dec_t *dec) 697decode_num (dec_t *dec)
… …
712static SV * 778static SV *
713decode_av (dec_t *dec) 779decode_av (dec_t *dec)
714{ 780{
715 AV *av = newAV (); 781 AV *av = newAV ();
716 782
717 WS; 783 decode_ws (dec);
718 if (*dec->cur == ']') 784 if (*dec->cur == ']')
719 ++dec->cur; 785 ++dec->cur;
720 else 786 else
721 for (;;) 787 for (;;)
722 { 788 {
… …
726 if (!value) 792 if (!value)
727 goto fail; 793 goto fail;
728 794
729 av_push (av, value); 795 av_push (av, value);
730 796
731 WS; 797 decode_ws (dec);
732 798
733 if (*dec->cur == ']') 799 if (*dec->cur == ']')
734 { 800 {
735 ++dec->cur; 801 ++dec->cur;
736 break; 802 break;
… …
752static SV * 818static SV *
753decode_hv (dec_t *dec) 819decode_hv (dec_t *dec)
754{ 820{
755 HV *hv = newHV (); 821 HV *hv = newHV ();
756 822
757 WS; 823 decode_ws (dec);
758 if (*dec->cur == '}') 824 if (*dec->cur == '}')
759 ++dec->cur; 825 ++dec->cur;
760 else 826 else
761 for (;;) 827 for (;;)
762 { 828 {
763 SV *key, *value; 829 SV *key, *value;
764 830
765 WS; EXPECT_CH ('"'); 831 decode_ws (dec); EXPECT_CH ('"');
766 832
767 key = decode_str (dec); 833 key = decode_str (dec);
768 if (!key) 834 if (!key)
769 goto fail; 835 goto fail;
770 836
771 WS; EXPECT_CH (':'); 837 decode_ws (dec); EXPECT_CH (':');
772 838
773 value = decode_sv (dec); 839 value = decode_sv (dec);
774 if (!value) 840 if (!value)
775 { 841 {
776 SvREFCNT_dec (key); 842 SvREFCNT_dec (key);
… …
778 } 844 }
779 845
780 //TODO: optimise 846 //TODO: optimise
781 hv_store_ent (hv, key, value, 0); 847 hv_store_ent (hv, key, value, 0);
782 848
783 WS; 849 decode_ws (dec);
784 850
785 if (*dec->cur == '}') 851 if (*dec->cur == '}')
786 { 852 {
787 ++dec->cur; 853 ++dec->cur;
788 break; 854 break;
… …
802} 868}
803 869
804static SV * 870static SV *
805decode_sv (dec_t *dec) 871decode_sv (dec_t *dec)
806{ 872{
807 WS; 873 decode_ws (dec);
808 switch (*dec->cur) 874 switch (*dec->cur)
809 { 875 {
810 case '"': ++dec->cur; return decode_str (dec); 876 case '"': ++dec->cur; return decode_str (dec);
811 case '[': ++dec->cur; return decode_av (dec); 877 case '[': ++dec->cur; return decode_av (dec);
812 case '{': ++dec->cur; return decode_hv (dec); 878 case '{': ++dec->cur; return decode_hv (dec);
… …
901 } 967 }
902 968
903 sv = sv_2mortal (sv); 969 sv = sv_2mortal (sv);
904 970
905 if (!(dec.flags & F_ALLOW_NONREF) && !SvROK (sv)) 971 if (!(dec.flags & F_ALLOW_NONREF) && !SvROK (sv))
906 croak ("JSON object or array expected (but number, string, true, false or null found, use allow_nonref to allow this)"); 972 croak ("JSON text must be an object or array (but found number, string, true, false or null, use allow_nonref to allow this)");
907 973
908 return sv; 974 return sv;
909} 975}
976
977/////////////////////////////////////////////////////////////////////////////
978// XS interface functions
910 979
911MODULE = JSON::XS PACKAGE = JSON::XS 980MODULE = JSON::XS PACKAGE = JSON::XS
912 981
913BOOT: 982BOOT:
914{ 983{
… …
941 utf8 = F_UTF8 1010 utf8 = F_UTF8
942 indent = F_INDENT 1011 indent = F_INDENT
943 canonical = F_CANONICAL 1012 canonical = F_CANONICAL
944 space_before = F_SPACE_BEFORE 1013 space_before = F_SPACE_BEFORE
945 space_after = F_SPACE_AFTER 1014 space_after = F_SPACE_AFTER
946 json_rpc = F_JSON_RPC
947 pretty = F_PRETTY 1015 pretty = F_PRETTY
948 allow_nonref = F_ALLOW_NONREF 1016 allow_nonref = F_ALLOW_NONREF
949 shrink = F_SHRINK 1017 shrink = F_SHRINK
950 CODE: 1018 CODE:
951{ 1019{
… …
969 XPUSHs (decode_json (jsonstr, *SvJSON (self))); 1037 XPUSHs (decode_json (jsonstr, *SvJSON (self)));
970 1038
971PROTOTYPES: ENABLE 1039PROTOTYPES: ENABLE
972 1040
973void to_json (SV *scalar) 1041void to_json (SV *scalar)
1042 ALIAS:
1043 objToJson = 0
974 PPCODE: 1044 PPCODE:
975 XPUSHs (encode_json (scalar, F_UTF8)); 1045 XPUSHs (encode_json (scalar, F_UTF8));
976 1046
977void from_json (SV *jsonstr) 1047void from_json (SV *jsonstr)
1048 ALIAS:
1049 jsonToObj = 0
978 PPCODE: 1050 PPCODE:
979 XPUSHs (decode_json (jsonstr, F_UTF8)); 1051 XPUSHs (decode_json (jsonstr, F_UTF8));
980 1052

Diff Legend

– Removed lines
+ Added lines
< Changed lines
> Changed lines