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 1.10 by root, Sat Mar 24 01:15:22 2007 UTC vs.
Revision 1.15 by root, Sun Mar 25 02:32:40 2007 UTC

… …
17 17
18#define F_PRETTY F_INDENT | F_SPACE_BEFORE | F_SPACE_AFTER 18#define F_PRETTY F_INDENT | F_SPACE_BEFORE | F_SPACE_AFTER
19#define F_DEFAULT 0 19#define F_DEFAULT 0
20 20
21#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
22 26
23#define SB do { 27#define SB do {
24#define SE } while (0) 28#define SE } while (0)
25 29
26static 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
27 79
28// structure used for encoding JSON 80// structure used for encoding JSON
29typedef struct 81typedef struct
30{ 82{
31 char *cur; 83 char *cur; // SvPVX (sv) + current output position
32 STRLEN len; // SvLEN (sv)
33 char *end; // SvEND (sv) 84 char *end; // SvEND (sv)
34 SV *sv; 85 SV *sv; // result scalar
35 UV flags; 86 UV flags; // F_*
36 int max_recurse; 87 int indent; // indentation level
37 int indent; 88 int max_depth; // max. recursion level
38} enc_t; 89} enc_t;
39
40// structure used for decoding JSON
41typedef struct
42{
43 char *cur;
44 char *end;
45 const char *err;
46 UV flags;
47} dec_t;
48
49static UV *
50SvJSON (SV *sv)
51{
52 if (!(SvROK (sv) && SvOBJECT (SvRV (sv)) && SvSTASH (SvRV (sv)) == json_stash))
53 croak ("object is not of type JSON::XS");
54
55 return &SvUVX (SvRV (sv));
56}
57
58static void
59shrink (SV *sv)
60{
61 sv_utf8_downgrade (sv, 1);
62#ifdef SvPV_shrink_to_cur
63 SvPV_shrink_to_cur (sv);
64#endif
65}
66
67/////////////////////////////////////////////////////////////////////////////
68 90
69static void 91static void
70need (enc_t *enc, STRLEN len) 92need (enc_t *enc, STRLEN len)
71{ 93{
72 if (enc->cur + len >= enc->end) 94 if (enc->cur + len >= enc->end)
… …
130 STRLEN clen; 152 STRLEN clen;
131 UV uch; 153 UV uch;
132 154
133 if (is_utf8) 155 if (is_utf8)
134 { 156 {
135 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);
136 if (clen == (STRLEN)-1) 159 if (clen == (STRLEN)-1)
137 croak ("malformed or illegal unicode character in string [%.11s], cannot convert to JSON", str); 160 croak ("malformed or illegal unicode character in string [%.11s], cannot convert to JSON", str);
138 } 161 }
139 else 162 else
140 { 163 {
… …
178 } 201 }
179 while (--clen); 202 while (--clen);
180 } 203 }
181 else 204 else
182 { 205 {
183 need (enc, len += 10); // never more than 11 bytes needed 206 need (enc, len += UTF8_MAX_LEN - 1); // never more than 11 bytes needed
184 enc->cur = uvuni_to_utf8_flags (enc->cur, uch, 0); 207 enc->cur = uvuni_to_utf8_flags (enc->cur, uch, 0);
185 ++str; 208 ++str;
186 } 209 }
187 } 210 }
188 } 211 }
… …
190 213
191 --len; 214 --len;
192 } 215 }
193} 216}
194 217
195#define INDENT SB \ 218static void
219encode_indent (enc_t *enc)
220{
196 if (enc->flags & F_INDENT) \ 221 if (enc->flags & F_INDENT)
197 { \ 222 {
198 int i_; \ 223 int spaces = enc->indent * INDENT_STEP;
199 need (enc, enc->indent); \
200 for (i_ = enc->indent * 3; i_--; )\
201 encode_ch (enc, ' '); \
202 } \
203 SE
204 224
205#define SPACE SB need (enc, 1); encode_ch (enc, ' '); SE 225 need (enc, spaces);
206#define NL SB if (enc->flags & F_INDENT) { need (enc, 1); encode_ch (enc, '\n'); } SE 226 memset (enc->cur, ' ', spaces);
207#define COMMA SB \ 227 enc->cur += spaces;
228 }
229}
230
231static void
232encode_space (enc_t *enc)
233{
234 need (enc, 1);
208 encode_ch (enc, ','); \ 235 encode_ch (enc, ' ');
236}
237
238static void
239encode_nl (enc_t *enc)
240{
209 if (enc->flags & F_INDENT) \ 241 if (enc->flags & F_INDENT)
210 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);
211 else if (enc->flags & F_SPACE_AFTER) \ 255 else if (enc->flags & F_SPACE_AFTER)
212 SPACE; \ 256 encode_space (enc);
213 SE 257}
214 258
215static void encode_sv (enc_t *enc, SV *sv); 259static void encode_sv (enc_t *enc, SV *sv);
216 260
217static void 261static void
218encode_av (enc_t *enc, AV *av) 262encode_av (enc_t *enc, AV *av)
219{ 263{
220 int i, len = av_len (av); 264 int i, len = av_len (av);
221 265
222 encode_ch (enc, '['); NL; 266 encode_ch (enc, '['); encode_nl (enc);
223 ++enc->indent; 267 ++enc->indent;
224 268
225 for (i = 0; i <= len; ++i) 269 for (i = 0; i <= len; ++i)
226 { 270 {
227 INDENT; 271 encode_indent (enc);
228 encode_sv (enc, *av_fetch (av, i, 0)); 272 encode_sv (enc, *av_fetch (av, i, 0));
229 273
230 if (i < len) 274 if (i < len)
231 COMMA; 275 encode_comma (enc);
232 } 276 }
233 277
234 NL; 278 encode_nl (enc);
235 279
236 --enc->indent; 280 --enc->indent;
237 INDENT; encode_ch (enc, ']'); 281 encode_indent (enc); encode_ch (enc, ']');
238} 282}
239 283
240static void 284static void
241encode_he (enc_t *enc, HE *he) 285encode_he (enc_t *enc, HE *he)
242{ 286{
… …
256 else 300 else
257 encode_str (enc, HeKEY (he), HeKLEN (he), HeKUTF8 (he)); 301 encode_str (enc, HeKEY (he), HeKLEN (he), HeKUTF8 (he));
258 302
259 encode_ch (enc, '"'); 303 encode_ch (enc, '"');
260 304
261 if (enc->flags & F_SPACE_BEFORE) SPACE; 305 if (enc->flags & F_SPACE_BEFORE) encode_space (enc);
262 encode_ch (enc, ':'); 306 encode_ch (enc, ':');
263 if (enc->flags & F_SPACE_AFTER ) SPACE; 307 if (enc->flags & F_SPACE_AFTER ) encode_space (enc);
264 encode_sv (enc, HeVAL (he)); 308 encode_sv (enc, HeVAL (he));
265} 309}
266 310
267// compare hash entries, used when all keys are bytestrings 311// compare hash entries, used when all keys are bytestrings
268static int 312static int
… …
274 HE *b = *(HE **)b_; 318 HE *b = *(HE **)b_;
275 319
276 STRLEN la = HeKLEN (a); 320 STRLEN la = HeKLEN (a);
277 STRLEN lb = HeKLEN (b); 321 STRLEN lb = HeKLEN (b);
278 322
279 if (!(cmp == memcmp (HeKEY (a), HeKEY (b), la < lb ? la : lb))) 323 if (!(cmp = memcmp (HeKEY (a), HeKEY (b), la < lb ? la : lb)))
280 cmp = la < lb ? -1 : la == lb ? 0 : 1; 324 cmp = la - lb;
281 325
282 return cmp; 326 return cmp;
283} 327}
284 328
285// 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
… …
292static void 336static void
293encode_hv (enc_t *enc, HV *hv) 337encode_hv (enc_t *enc, HV *hv)
294{ 338{
295 int count, i; 339 int count, i;
296 340
297 encode_ch (enc, '{'); NL; ++enc->indent; 341 encode_ch (enc, '{'); encode_nl (enc); ++enc->indent;
298 342
299 if ((count = hv_iterinit (hv))) 343 if ((count = hv_iterinit (hv)))
300 { 344 {
301 // for canonical output we have to sort by keys first 345 // for canonical output we have to sort by keys first
302 // actually, this is mostly due to the stupid so-called 346 // actually, this is mostly due to the stupid so-called
303 // security workaround added somewhere in 5.8.x. 347 // security workaround added somewhere in 5.8.x.
304 // that randomises hash orderings 348 // that randomises hash orderings
305 if (enc->flags & F_CANONICAL) 349 if (enc->flags & F_CANONICAL)
306 { 350 {
307 HE *he, *hes [count]; 351 HE *he, *hes [count]; // if your compiler dies here, you need to enable C99 mode
308 int fast = 1; 352 int fast = 1;
309 353
310 i = 0; 354 i = 0;
311 while ((he = hv_iternext (hv))) 355 while ((he = hv_iternext (hv)))
312 { 356 {
… …
337 LEAVE; 381 LEAVE;
338 } 382 }
339 383
340 for (i = 0; i < count; ++i) 384 for (i = 0; i < count; ++i)
341 { 385 {
342 INDENT; 386 encode_indent (enc);
343 encode_he (enc, hes [i]); 387 encode_he (enc, hes [i]);
344 388
345 if (i < count - 1) 389 if (i < count - 1)
346 COMMA; 390 encode_comma (enc);
347 } 391 }
348 392
349 NL; 393 encode_nl (enc);
350 } 394 }
351 else 395 else
352 { 396 {
353 SV *sv; 397 SV *sv;
354 HE *he = hv_iternext (hv); 398 HE *he = hv_iternext (hv);
355 399
356 for (;;) 400 for (;;)
357 { 401 {
358 INDENT; 402 encode_indent (enc);
359 encode_he (enc, he); 403 encode_he (enc, he);
360 404
361 if (!(he = hv_iternext (hv))) 405 if (!(he = hv_iternext (hv)))
362 break; 406 break;
363 407
364 COMMA; 408 encode_comma (enc);
365 } 409 }
366 410
367 NL; 411 encode_nl (enc);
368 } 412 }
369 } 413 }
370 414
371 --enc->indent; INDENT; encode_ch (enc, '}'); 415 --enc->indent; encode_indent (enc); encode_ch (enc, '}');
372} 416}
373 417
374static void 418static void
375encode_sv (enc_t *enc, SV *sv) 419encode_sv (enc_t *enc, SV *sv)
376{ 420{
… …
400 } 444 }
401 else if (SvROK (sv)) 445 else if (SvROK (sv))
402 { 446 {
403 SV *rv = SvRV (sv); 447 SV *rv = SvRV (sv);
404 448
405 if (!--enc->max_recurse) 449 if (enc->indent >= enc->max_depth)
406 croak ("data structure too deep (hit recursion limit)"); 450 croak ("data structure too deep (hit recursion limit)");
407 451
408 switch (SvTYPE (rv)) 452 switch (SvTYPE (rv))
409 { 453 {
410 case SVt_PVAV: encode_av (enc, (AV *)rv); break; 454 case SVt_PVAV: encode_av (enc, (AV *)rv); break;
… …
427{ 471{
428 if (!(flags & F_ALLOW_NONREF) && !SvROK (scalar)) 472 if (!(flags & F_ALLOW_NONREF) && !SvROK (scalar))
429 croak ("hash- or arrayref expected (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)");
430 474
431 enc_t enc; 475 enc_t enc;
432 enc.flags = flags; 476 enc.flags = flags;
433 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 477 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
434 enc.cur = SvPVX (enc.sv); 478 enc.cur = SvPVX (enc.sv);
435 enc.end = SvEND (enc.sv); 479 enc.end = SvEND (enc.sv);
436 enc.max_recurse = 0;
437 enc.indent = 0; 480 enc.indent = 0;
481 enc.max_depth = 0x7fffffffUL;
438 482
439 SvPOK_only (enc.sv); 483 SvPOK_only (enc.sv);
440 encode_sv (&enc, scalar); 484 encode_sv (&enc, scalar);
441 485
442 if (!(flags & (F_ASCII | F_UTF8))) 486 if (!(flags & (F_ASCII | F_UTF8)))
… …
449 493
450 return enc.sv; 494 return enc.sv;
451} 495}
452 496
453///////////////////////////////////////////////////////////////////////////// 497/////////////////////////////////////////////////////////////////////////////
498// decoder
454 499
455#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{
456 for (;;) \ 512 for (;;)
457 { \ 513 {
458 char ch = *dec->cur; \ 514 char ch = *dec->cur;
515
459 if (ch > 0x20 \ 516 if (ch > 0x20
460 || (ch != 0x20 && ch != 0x0a && ch != 0x0d && ch != 0x09)) \ 517 || (ch != 0x20 && ch != 0x0a && ch != 0x0d && ch != 0x09))
461 break; \ 518 break;
519
462 ++dec->cur; \ 520 ++dec->cur;
463 } 521 }
522}
464 523
465#define ERR(reason) SB dec->err = reason; goto fail; SE 524#define ERR(reason) SB dec->err = reason; goto fail; SE
466#define EXPECT_CH(ch) SB \ 525#define EXPECT_CH(ch) SB \
467 if (*dec->cur != ch) \ 526 if (*dec->cur != ch) \
468 ERR (# ch " expected"); \ 527 ERR (# ch " expected"); \
… …
475 534
476static UV 535static UV
477decode_4hex (dec_t *dec) 536decode_4hex (dec_t *dec)
478{ 537{
479 signed char d1, d2, d3, d4; 538 signed char d1, d2, d3, d4;
539 unsigned char *cur = (unsigned char *)dec->cur;
480 540
481 d1 = decode_hexdigit [((unsigned char *)dec->cur) [0]];
482 if (d1 < 0) ERR ("four hexadecimal digits expected"); 541 d1 = decode_hexdigit [cur [0]]; if (d1 < 0) ERR ("four hexadecimal digits expected");
483 d2 = decode_hexdigit [((unsigned char *)dec->cur) [1]];
484 if (d2 < 0) ERR ("four hexadecimal digits expected"); 542 d2 = decode_hexdigit [cur [1]]; if (d2 < 0) ERR ("four hexadecimal digits expected");
485 d3 = decode_hexdigit [((unsigned char *)dec->cur) [2]];
486 if (d3 < 0) ERR ("four hexadecimal digits expected"); 543 d3 = decode_hexdigit [cur [2]]; if (d3 < 0) ERR ("four hexadecimal digits expected");
487 d4 = decode_hexdigit [((unsigned char *)dec->cur) [3]];
488 if (d4 < 0) ERR ("four hexadecimal digits expected"); 544 d4 = decode_hexdigit [cur [3]]; if (d4 < 0) ERR ("four hexadecimal digits expected");
489 545
490 dec->cur += 4; 546 dec->cur += 4;
491 547
492 return ((UV)d1) << 12 548 return ((UV)d1) << 12
493 | ((UV)d2) << 8 549 | ((UV)d2) << 8
… …
496 552
497fail: 553fail:
498 return (UV)-1; 554 return (UV)-1;
499} 555}
500 556
501#define APPEND_GROW(n) SB \
502 if (cur + (n) >= end) \
503 { \
504 STRLEN ofs = cur - SvPVX (sv); \
505 SvGROW (sv, ofs + (n) + 1); \
506 cur = SvPVX (sv) + ofs; \
507 end = SvEND (sv); \
508 } \
509 SE
510
511#define APPEND_CH(ch) SB \
512 APPEND_GROW (1); \
513 *cur++ = (ch); \
514 SE
515
516static SV * 557static SV *
517decode_str (dec_t *dec) 558decode_str (dec_t *dec)
518{ 559{
519 SV *sv = NEWSV (0,2); 560 SV *sv = 0;
520 int utf8 = 0; 561 int utf8 = 0;
521 char *cur = SvPVX (sv);
522 char *end = SvEND (sv);
523 562
524 for (;;) 563 do
525 { 564 {
526 unsigned char ch = *(unsigned char *)dec->cur; 565 char buf [SHORT_STRING_LEN + UTF8_MAX_LEN];
566 char *cur = buf;
527 567
528 if (ch == '"') 568 do
529 break;
530 else if (ch == '\\')
531 { 569 {
532 switch (*++dec->cur) 570 unsigned char ch = *(unsigned char *)dec->cur++;
571
572 if (ch == '"')
533 { 573 {
534 case '\\': 574 --dec->cur;
535 case '/': 575 break;
536 case '"': APPEND_CH (*dec->cur++); break; 576 }
537 577 else if (ch == '\\')
538 case 'b': APPEND_CH ('\010'); ++dec->cur; break; 578 {
539 case 't': APPEND_CH ('\011'); ++dec->cur; break; 579 switch (*dec->cur)
540 case 'n': APPEND_CH ('\012'); ++dec->cur; break;
541 case 'f': APPEND_CH ('\014'); ++dec->cur; break;
542 case 'r': APPEND_CH ('\015'); ++dec->cur; break;
543
544 case 'u':
545 { 580 {
546 UV lo, hi; 581 case '\\':
547 ++dec->cur; 582 case '/':
583 case '"': *cur++ = *dec->cur++; break;
548 584
549 hi = decode_4hex (dec); 585 case 'b': ++dec->cur; *cur++ = '\010'; break;
550 if (hi == (UV)-1) 586 case 't': ++dec->cur; *cur++ = '\011'; break;
551 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;
552 590
553 // possibly a surrogate pair 591 case 'u':
554 if (hi >= 0xd800 && hi < 0xdc00)
555 { 592 {
556 if (dec->cur [0] != '\\' || dec->cur [1] != 'u') 593 UV lo, hi;
557 ERR ("missing low surrogate character in surrogate pair");
558
559 dec->cur += 2; 594 ++dec->cur;
560 595
561 lo = decode_4hex (dec); 596 hi = decode_4hex (dec);
562 if (lo == (UV)-1) 597 if (hi == (UV)-1)
563 goto fail; 598 goto fail;
564 599
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
565 if (lo < 0xdc00 || lo >= 0xe000) 613 if (lo < 0xdc00 || lo >= 0xe000)
566 ERR ("surrogate pair expected"); 614 ERR ("surrogate pair expected");
567 615
568 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;
569 } 629 }
570 else if (hi >= 0xdc00 && hi < 0xe000)
571 ERR ("missing high surrogate character in surrogate pair");
572
573 if (hi >= 0x80)
574 { 630 break;
575 utf8 = 1;
576 631
577 APPEND_GROW (4); // at most 4 bytes for 21 bits
578 cur = (char *)uvuni_to_utf8_flags (cur, hi, 0);
579 }
580 else 632 default:
581 APPEND_CH (hi); 633 --dec->cur;
634 ERR ("illegal backslash escape sequence in string");
582 } 635 }
583 break;
584
585 default:
586 --dec->cur;
587 ERR ("illegal backslash escape sequence in string");
588 } 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
589 } 661 }
590 else if (ch >= 0x20 && ch <= 0x7f) 662 while (cur < buf + SHORT_STRING_LEN);
591 APPEND_CH (*dec->cur++); 663
592 else if (ch >= 0x80) 664 STRLEN len = cur - buf;
665
666 if (sv)
593 { 667 {
594 STRLEN clen; 668 SvGROW (sv, SvCUR (sv) + len + 1);
595 UV uch = utf8n_to_uvuni (dec->cur, dec->end - dec->cur, &clen, UTF8_CHECK_ONLY); 669 memcpy (SvPVX (sv) + SvCUR (sv), buf, len);
596 if (clen == (STRLEN)-1) 670 SvCUR_set (sv, SvCUR (sv) + len);
597 ERR ("malformed UTF-8 character in JSON string");
598
599 APPEND_GROW (clen);
600 do
601 {
602 *cur++ = *dec->cur++;
603 }
604 while (--clen);
605
606 utf8 = 1;
607 } 671 }
608 else if (dec->cur == dec->end)
609 ERR ("unexpected end of string while parsing json string");
610 else 672 else
611 ERR ("invalid character encountered"); 673 sv = newSVpvn (buf, len);
612 } 674 }
675 while (*dec->cur != '"');
613 676
614 ++dec->cur; 677 ++dec->cur;
615 678
616 SvCUR_set (sv, cur - SvPVX (sv)); 679 if (sv)
617 680 {
618 SvPOK_only (sv); 681 SvPOK_only (sv);
619 *SvEND (sv) = 0; 682 *SvEND (sv) = 0;
620 683
621 if (utf8) 684 if (utf8)
622 SvUTF8_on (sv); 685 SvUTF8_on (sv);
623 686 }
624 if (dec->flags & F_SHRINK) 687 else
625 shrink (sv); 688 sv = newSVpvn ("", 0);
626 689
627 return sv; 690 return sv;
628 691
629fail: 692fail:
630 SvREFCNT_dec (sv);
631 return 0; 693 return 0;
632} 694}
633 695
634static SV * 696static SV *
635decode_num (dec_t *dec) 697decode_num (dec_t *dec)
… …
716static SV * 778static SV *
717decode_av (dec_t *dec) 779decode_av (dec_t *dec)
718{ 780{
719 AV *av = newAV (); 781 AV *av = newAV ();
720 782
721 WS; 783 decode_ws (dec);
722 if (*dec->cur == ']') 784 if (*dec->cur == ']')
723 ++dec->cur; 785 ++dec->cur;
724 else 786 else
725 for (;;) 787 for (;;)
726 { 788 {
… …
730 if (!value) 792 if (!value)
731 goto fail; 793 goto fail;
732 794
733 av_push (av, value); 795 av_push (av, value);
734 796
735 WS; 797 decode_ws (dec);
736 798
737 if (*dec->cur == ']') 799 if (*dec->cur == ']')
738 { 800 {
739 ++dec->cur; 801 ++dec->cur;
740 break; 802 break;
… …
756static SV * 818static SV *
757decode_hv (dec_t *dec) 819decode_hv (dec_t *dec)
758{ 820{
759 HV *hv = newHV (); 821 HV *hv = newHV ();
760 822
761 WS; 823 decode_ws (dec);
762 if (*dec->cur == '}') 824 if (*dec->cur == '}')
763 ++dec->cur; 825 ++dec->cur;
764 else 826 else
765 for (;;) 827 for (;;)
766 { 828 {
767 SV *key, *value; 829 SV *key, *value;
768 830
769 WS; EXPECT_CH ('"'); 831 decode_ws (dec); EXPECT_CH ('"');
770 832
771 key = decode_str (dec); 833 key = decode_str (dec);
772 if (!key) 834 if (!key)
773 goto fail; 835 goto fail;
774 836
775 WS; EXPECT_CH (':'); 837 decode_ws (dec); EXPECT_CH (':');
776 838
777 value = decode_sv (dec); 839 value = decode_sv (dec);
778 if (!value) 840 if (!value)
779 { 841 {
780 SvREFCNT_dec (key); 842 SvREFCNT_dec (key);
… …
782 } 844 }
783 845
784 //TODO: optimise 846 //TODO: optimise
785 hv_store_ent (hv, key, value, 0); 847 hv_store_ent (hv, key, value, 0);
786 848
787 WS; 849 decode_ws (dec);
788 850
789 if (*dec->cur == '}') 851 if (*dec->cur == '}')
790 { 852 {
791 ++dec->cur; 853 ++dec->cur;
792 break; 854 break;
… …
806} 868}
807 869
808static SV * 870static SV *
809decode_sv (dec_t *dec) 871decode_sv (dec_t *dec)
810{ 872{
811 WS; 873 decode_ws (dec);
812 switch (*dec->cur) 874 switch (*dec->cur)
813 { 875 {
814 case '"': ++dec->cur; return decode_str (dec); 876 case '"': ++dec->cur; return decode_str (dec);
815 case '[': ++dec->cur; return decode_av (dec); 877 case '[': ++dec->cur; return decode_av (dec);
816 case '{': ++dec->cur; return decode_hv (dec); 878 case '{': ++dec->cur; return decode_hv (dec);
… …
909 if (!(dec.flags & F_ALLOW_NONREF) && !SvROK (sv)) 971 if (!(dec.flags & F_ALLOW_NONREF) && !SvROK (sv))
910 croak ("JSON text must be an object or array (but found number, string, true, false or null, 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)");
911 973
912 return sv; 974 return sv;
913} 975}
976
977/////////////////////////////////////////////////////////////////////////////
978// XS interface functions
914 979
915MODULE = JSON::XS PACKAGE = JSON::XS 980MODULE = JSON::XS PACKAGE = JSON::XS
916 981
917BOOT: 982BOOT:
918{ 983{
… …
972 XPUSHs (decode_json (jsonstr, *SvJSON (self))); 1037 XPUSHs (decode_json (jsonstr, *SvJSON (self)));
973 1038
974PROTOTYPES: ENABLE 1039PROTOTYPES: ENABLE
975 1040
976void to_json (SV *scalar) 1041void to_json (SV *scalar)
1042 ALIAS:
1043 objToJson = 0
977 PPCODE: 1044 PPCODE:
978 XPUSHs (encode_json (scalar, F_UTF8)); 1045 XPUSHs (encode_json (scalar, F_UTF8));
979 1046
980void from_json (SV *jsonstr) 1047void from_json (SV *jsonstr)
1048 ALIAS:
1049 jsonToObj = 0
981 PPCODE: 1050 PPCODE:
982 XPUSHs (decode_json (jsonstr, F_UTF8)); 1051 XPUSHs (decode_json (jsonstr, F_UTF8));
983 1052

Diff Legend

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