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 1.16 by root, Sun Mar 25 02:37:00 2007 UTC

… …
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_ALLOW_NONREF 0x00000080 15#define F_ALLOW_NONREF 0x00000080
16#define F_SHRINK 0x00000100 16#define F_SHRINK 0x00000100
17 17
18// F_SKIPINVALID?
19// F_EXECCODEREF?
20// F_SELFCONVERT?
21
18#define F_PRETTY F_INDENT | F_SPACE_BEFORE | F_SPACE_AFTER 22#define F_PRETTY F_INDENT | F_SPACE_BEFORE | F_SPACE_AFTER
19#define F_DEFAULT 0 23#define F_DEFAULT 0
20 24
21#define INIT_SIZE 32 // initial scalar size to be allocated 25#define INIT_SIZE 32 // initial scalar size to be allocated
26#define INDENT_STEP 3 // spaces per indentation level
27
28#define UTF8_MAX_LEN 11 // for perls UTF-X: max. number of octets per character
29#define SHORT_STRING_LEN 512 // special-case strings of up to this size
22 30
23#define SB do { 31#define SB do {
24#define SE } while (0) 32#define SE } while (0)
25 33
26static HV *json_stash; 34static HV *json_stash; // JSON::XS::
35
36/////////////////////////////////////////////////////////////////////////////
37// utility functions
38
39static UV *
40SvJSON (SV *sv)
41{
42 if (!(SvROK (sv) && SvOBJECT (SvRV (sv)) && SvSTASH (SvRV (sv)) == json_stash))
43 croak ("object is not of type JSON::XS");
44
45 return &SvUVX (SvRV (sv));
46}
47
48static void
49shrink (SV *sv)
50{
51 sv_utf8_downgrade (sv, 1);
52 if (SvLEN (sv) > SvCUR (sv) + 1)
53 {
54#ifdef SvPV_shrink_to_cur
55 SvPV_shrink_to_cur (sv);
56#elif defined (SvPV_renew)
57 SvPV_renew (sv, SvCUR (sv) + 1);
58#endif
59 }
60}
61
62// decode an utf-8 character and return it, or (UV)-1 in
63// case of an error.
64// we special-case "safe" characters from U+80 .. U+7FF,
65// but use the very good perl function to parse anything else.
66// note that we never call this function for a ascii codepoints
67static UV
68decode_utf8 (unsigned char *s, STRLEN len, STRLEN *clen)
69{
70 if (s[0] > 0xdf || s[0] < 0xc2)
71 return utf8n_to_uvuni (s, len, clen, UTF8_CHECK_ONLY);
72 else if (len > 1 && s[1] >= 0x80 && s[1] <= 0xbf)
73 {
74 *clen = 2;
75 return ((s[0] & 0x1f) << 6) | (s[1] & 0x3f);
76 }
77 else
78 return (UV)-1;
79}
80
81/////////////////////////////////////////////////////////////////////////////
82// encoder
27 83
28// structure used for encoding JSON 84// structure used for encoding JSON
29typedef struct 85typedef struct
30{ 86{
31 char *cur; 87 char *cur; // SvPVX (sv) + current output position
32 STRLEN len; // SvLEN (sv)
33 char *end; // SvEND (sv) 88 char *end; // SvEND (sv)
34 SV *sv; 89 SV *sv; // result scalar
35 UV flags; 90 UV flags; // F_*
36 int max_recurse; 91 int indent; // indentation level
37 int indent; 92 int max_depth; // max. recursion level
38} enc_t; 93} 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 94
69static void 95static void
70need (enc_t *enc, STRLEN len) 96need (enc_t *enc, STRLEN len)
71{ 97{
72 if (enc->cur + len >= enc->end) 98 if (enc->cur + len >= enc->end)
… …
130 STRLEN clen; 156 STRLEN clen;
131 UV uch; 157 UV uch;
132 158
133 if (is_utf8) 159 if (is_utf8)
134 { 160 {
135 uch = utf8n_to_uvuni (str, end - str, &clen, UTF8_CHECK_ONLY); 161 //uch = utf8n_to_uvuni (str, end - str, &clen, UTF8_CHECK_ONLY);
162 uch = decode_utf8 (str, end - str, &clen);
136 if (clen == (STRLEN)-1) 163 if (clen == (STRLEN)-1)
137 croak ("malformed or illegal unicode character in string [%.11s], cannot convert to JSON", str); 164 croak ("malformed or illegal unicode character in string [%.11s], cannot convert to JSON", str);
138 } 165 }
139 else 166 else
140 { 167 {
… …
178 } 205 }
179 while (--clen); 206 while (--clen);
180 } 207 }
181 else 208 else
182 { 209 {
183 need (enc, len += 10); // never more than 11 bytes needed 210 need (enc, len += UTF8_MAX_LEN - 1); // never more than 11 bytes needed
184 enc->cur = uvuni_to_utf8_flags (enc->cur, uch, 0); 211 enc->cur = uvuni_to_utf8_flags (enc->cur, uch, 0);
185 ++str; 212 ++str;
186 } 213 }
187 } 214 }
188 } 215 }
… …
190 217
191 --len; 218 --len;
192 } 219 }
193} 220}
194 221
195#define INDENT SB \ 222static void
223encode_indent (enc_t *enc)
224{
196 if (enc->flags & F_INDENT) \ 225 if (enc->flags & F_INDENT)
197 { \ 226 {
198 int i_; \ 227 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
204 228
205#define SPACE SB need (enc, 1); encode_ch (enc, ' '); SE 229 need (enc, spaces);
206#define NL SB if (enc->flags & F_INDENT) { need (enc, 1); encode_ch (enc, '\n'); } SE 230 memset (enc->cur, ' ', spaces);
207#define COMMA SB \ 231 enc->cur += spaces;
232 }
233}
234
235static void
236encode_space (enc_t *enc)
237{
238 need (enc, 1);
208 encode_ch (enc, ','); \ 239 encode_ch (enc, ' ');
240}
241
242static void
243encode_nl (enc_t *enc)
244{
209 if (enc->flags & F_INDENT) \ 245 if (enc->flags & F_INDENT)
210 NL; \ 246 {
247 need (enc, 1);
248 encode_ch (enc, '\n');
249 }
250}
251
252static void
253encode_comma (enc_t *enc)
254{
255 encode_ch (enc, ',');
256
257 if (enc->flags & F_INDENT)
258 encode_nl (enc);
211 else if (enc->flags & F_SPACE_AFTER) \ 259 else if (enc->flags & F_SPACE_AFTER)
212 SPACE; \ 260 encode_space (enc);
213 SE 261}
214 262
215static void encode_sv (enc_t *enc, SV *sv); 263static void encode_sv (enc_t *enc, SV *sv);
216 264
217static void 265static void
218encode_av (enc_t *enc, AV *av) 266encode_av (enc_t *enc, AV *av)
219{ 267{
220 int i, len = av_len (av); 268 int i, len = av_len (av);
221 269
222 encode_ch (enc, '['); NL; 270 encode_ch (enc, '['); encode_nl (enc);
223 ++enc->indent; 271 ++enc->indent;
224 272
225 for (i = 0; i <= len; ++i) 273 for (i = 0; i <= len; ++i)
226 { 274 {
227 INDENT; 275 encode_indent (enc);
228 encode_sv (enc, *av_fetch (av, i, 0)); 276 encode_sv (enc, *av_fetch (av, i, 0));
229 277
230 if (i < len) 278 if (i < len)
231 COMMA; 279 encode_comma (enc);
232 } 280 }
233 281
234 NL; 282 encode_nl (enc);
235 283
236 --enc->indent; 284 --enc->indent;
237 INDENT; encode_ch (enc, ']'); 285 encode_indent (enc); encode_ch (enc, ']');
238} 286}
239 287
240static void 288static void
241encode_he (enc_t *enc, HE *he) 289encode_he (enc_t *enc, HE *he)
242{ 290{
… …
256 else 304 else
257 encode_str (enc, HeKEY (he), HeKLEN (he), HeKUTF8 (he)); 305 encode_str (enc, HeKEY (he), HeKLEN (he), HeKUTF8 (he));
258 306
259 encode_ch (enc, '"'); 307 encode_ch (enc, '"');
260 308
261 if (enc->flags & F_SPACE_BEFORE) SPACE; 309 if (enc->flags & F_SPACE_BEFORE) encode_space (enc);
262 encode_ch (enc, ':'); 310 encode_ch (enc, ':');
263 if (enc->flags & F_SPACE_AFTER ) SPACE; 311 if (enc->flags & F_SPACE_AFTER ) encode_space (enc);
264 encode_sv (enc, HeVAL (he)); 312 encode_sv (enc, HeVAL (he));
265} 313}
266 314
267// compare hash entries, used when all keys are bytestrings 315// compare hash entries, used when all keys are bytestrings
268static int 316static int
… …
274 HE *b = *(HE **)b_; 322 HE *b = *(HE **)b_;
275 323
276 STRLEN la = HeKLEN (a); 324 STRLEN la = HeKLEN (a);
277 STRLEN lb = HeKLEN (b); 325 STRLEN lb = HeKLEN (b);
278 326
279 if (!(cmp == memcmp (HeKEY (a), HeKEY (b), la < lb ? la : lb))) 327 if (!(cmp = memcmp (HeKEY (a), HeKEY (b), la < lb ? la : lb)))
280 cmp = la < lb ? -1 : la == lb ? 0 : 1; 328 cmp = la - lb;
281 329
282 return cmp; 330 return cmp;
283} 331}
284 332
285// compare hash entries, used when some keys are sv's or utf-x 333// compare hash entries, used when some keys are sv's or utf-x
… …
292static void 340static void
293encode_hv (enc_t *enc, HV *hv) 341encode_hv (enc_t *enc, HV *hv)
294{ 342{
295 int count, i; 343 int count, i;
296 344
297 encode_ch (enc, '{'); NL; ++enc->indent; 345 encode_ch (enc, '{'); encode_nl (enc); ++enc->indent;
298 346
299 if ((count = hv_iterinit (hv))) 347 if ((count = hv_iterinit (hv)))
300 { 348 {
301 // for canonical output we have to sort by keys first 349 // for canonical output we have to sort by keys first
302 // actually, this is mostly due to the stupid so-called 350 // actually, this is mostly due to the stupid so-called
303 // security workaround added somewhere in 5.8.x. 351 // security workaround added somewhere in 5.8.x.
304 // that randomises hash orderings 352 // that randomises hash orderings
305 if (enc->flags & F_CANONICAL) 353 if (enc->flags & F_CANONICAL)
306 { 354 {
307 HE *he, *hes [count]; 355 HE *he, *hes [count]; // if your compiler dies here, you need to enable C99 mode
308 int fast = 1; 356 int fast = 1;
309 357
310 i = 0; 358 i = 0;
311 while ((he = hv_iternext (hv))) 359 while ((he = hv_iternext (hv)))
312 { 360 {
… …
337 LEAVE; 385 LEAVE;
338 } 386 }
339 387
340 for (i = 0; i < count; ++i) 388 for (i = 0; i < count; ++i)
341 { 389 {
342 INDENT; 390 encode_indent (enc);
343 encode_he (enc, hes [i]); 391 encode_he (enc, hes [i]);
344 392
345 if (i < count - 1) 393 if (i < count - 1)
346 COMMA; 394 encode_comma (enc);
347 } 395 }
348 396
349 NL; 397 encode_nl (enc);
350 } 398 }
351 else 399 else
352 { 400 {
353 SV *sv; 401 SV *sv;
354 HE *he = hv_iternext (hv); 402 HE *he = hv_iternext (hv);
355 403
356 for (;;) 404 for (;;)
357 { 405 {
358 INDENT; 406 encode_indent (enc);
359 encode_he (enc, he); 407 encode_he (enc, he);
360 408
361 if (!(he = hv_iternext (hv))) 409 if (!(he = hv_iternext (hv)))
362 break; 410 break;
363 411
364 COMMA; 412 encode_comma (enc);
365 } 413 }
366 414
367 NL; 415 encode_nl (enc);
368 } 416 }
369 } 417 }
370 418
371 --enc->indent; INDENT; encode_ch (enc, '}'); 419 --enc->indent; encode_indent (enc); encode_ch (enc, '}');
372} 420}
373 421
374static void 422static void
375encode_sv (enc_t *enc, SV *sv) 423encode_sv (enc_t *enc, SV *sv)
376{ 424{
… …
400 } 448 }
401 else if (SvROK (sv)) 449 else if (SvROK (sv))
402 { 450 {
403 SV *rv = SvRV (sv); 451 SV *rv = SvRV (sv);
404 452
405 if (!--enc->max_recurse) 453 if (enc->indent >= enc->max_depth)
406 croak ("data structure too deep (hit recursion limit)"); 454 croak ("data structure too deep (hit recursion limit)");
407 455
408 switch (SvTYPE (rv)) 456 switch (SvTYPE (rv))
409 { 457 {
410 case SVt_PVAV: encode_av (enc, (AV *)rv); break; 458 case SVt_PVAV: encode_av (enc, (AV *)rv); break;
… …
427{ 475{
428 if (!(flags & F_ALLOW_NONREF) && !SvROK (scalar)) 476 if (!(flags & F_ALLOW_NONREF) && !SvROK (scalar))
429 croak ("hash- or arrayref expected (not a simple scalar, use allow_nonref to allow this)"); 477 croak ("hash- or arrayref expected (not a simple scalar, use allow_nonref to allow this)");
430 478
431 enc_t enc; 479 enc_t enc;
432 enc.flags = flags; 480 enc.flags = flags;
433 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 481 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
434 enc.cur = SvPVX (enc.sv); 482 enc.cur = SvPVX (enc.sv);
435 enc.end = SvEND (enc.sv); 483 enc.end = SvEND (enc.sv);
436 enc.max_recurse = 0;
437 enc.indent = 0; 484 enc.indent = 0;
485 enc.max_depth = 0x7fffffffUL;
438 486
439 SvPOK_only (enc.sv); 487 SvPOK_only (enc.sv);
440 encode_sv (&enc, scalar); 488 encode_sv (&enc, scalar);
441 489
442 if (!(flags & (F_ASCII | F_UTF8))) 490 if (!(flags & (F_ASCII | F_UTF8)))
… …
449 497
450 return enc.sv; 498 return enc.sv;
451} 499}
452 500
453///////////////////////////////////////////////////////////////////////////// 501/////////////////////////////////////////////////////////////////////////////
502// decoder
454 503
455#define WS \ 504// structure used for decoding JSON
505typedef struct
506{
507 char *cur; // current parser pointer
508 char *end; // end of input string
509 const char *err; // parse error, if != 0
510 UV flags; // F_*
511} dec_t;
512
513static void
514decode_ws (dec_t *dec)
515{
456 for (;;) \ 516 for (;;)
457 { \ 517 {
458 char ch = *dec->cur; \ 518 char ch = *dec->cur;
519
459 if (ch > 0x20 \ 520 if (ch > 0x20
460 || (ch != 0x20 && ch != 0x0a && ch != 0x0d && ch != 0x09)) \ 521 || (ch != 0x20 && ch != 0x0a && ch != 0x0d && ch != 0x09))
461 break; \ 522 break;
523
462 ++dec->cur; \ 524 ++dec->cur;
463 } 525 }
526}
464 527
465#define ERR(reason) SB dec->err = reason; goto fail; SE 528#define ERR(reason) SB dec->err = reason; goto fail; SE
466#define EXPECT_CH(ch) SB \ 529#define EXPECT_CH(ch) SB \
467 if (*dec->cur != ch) \ 530 if (*dec->cur != ch) \
468 ERR (# ch " expected"); \ 531 ERR (# ch " expected"); \
… …
475 538
476static UV 539static UV
477decode_4hex (dec_t *dec) 540decode_4hex (dec_t *dec)
478{ 541{
479 signed char d1, d2, d3, d4; 542 signed char d1, d2, d3, d4;
543 unsigned char *cur = (unsigned char *)dec->cur;
480 544
481 d1 = decode_hexdigit [((unsigned char *)dec->cur) [0]];
482 if (d1 < 0) ERR ("four hexadecimal digits expected"); 545 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"); 546 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"); 547 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"); 548 d4 = decode_hexdigit [cur [3]]; if (d4 < 0) ERR ("four hexadecimal digits expected");
489 549
490 dec->cur += 4; 550 dec->cur += 4;
491 551
492 return ((UV)d1) << 12 552 return ((UV)d1) << 12
493 | ((UV)d2) << 8 553 | ((UV)d2) << 8
… …
496 556
497fail: 557fail:
498 return (UV)-1; 558 return (UV)-1;
499} 559}
500 560
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 * 561static SV *
517decode_str (dec_t *dec) 562decode_str (dec_t *dec)
518{ 563{
519 SV *sv = NEWSV (0,2); 564 SV *sv = 0;
520 int utf8 = 0; 565 int utf8 = 0;
521 char *cur = SvPVX (sv);
522 char *end = SvEND (sv);
523 566
524 for (;;) 567 do
525 { 568 {
526 unsigned char ch = *(unsigned char *)dec->cur; 569 char buf [SHORT_STRING_LEN + UTF8_MAX_LEN];
570 char *cur = buf;
527 571
528 if (ch == '"') 572 do
529 break;
530 else if (ch == '\\')
531 { 573 {
532 switch (*++dec->cur) 574 unsigned char ch = *(unsigned char *)dec->cur++;
575
576 if (ch == '"')
533 { 577 {
534 case '\\': 578 --dec->cur;
535 case '/': 579 break;
536 case '"': APPEND_CH (*dec->cur++); break; 580 }
537 581 else if (ch == '\\')
538 case 'b': APPEND_CH ('\010'); ++dec->cur; break; 582 {
539 case 't': APPEND_CH ('\011'); ++dec->cur; break; 583 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 { 584 {
546 UV lo, hi; 585 case '\\':
547 ++dec->cur; 586 case '/':
587 case '"': *cur++ = *dec->cur++; break;
548 588
549 hi = decode_4hex (dec); 589 case 'b': ++dec->cur; *cur++ = '\010'; break;
550 if (hi == (UV)-1) 590 case 't': ++dec->cur; *cur++ = '\011'; break;
551 goto fail; 591 case 'n': ++dec->cur; *cur++ = '\012'; break;
592 case 'f': ++dec->cur; *cur++ = '\014'; break;
593 case 'r': ++dec->cur; *cur++ = '\015'; break;
552 594
553 // possibly a surrogate pair 595 case 'u':
554 if (hi >= 0xd800 && hi < 0xdc00)
555 { 596 {
556 if (dec->cur [0] != '\\' || dec->cur [1] != 'u') 597 UV lo, hi;
557 ERR ("missing low surrogate character in surrogate pair");
558
559 dec->cur += 2; 598 ++dec->cur;
560 599
561 lo = decode_4hex (dec); 600 hi = decode_4hex (dec);
562 if (lo == (UV)-1) 601 if (hi == (UV)-1)
563 goto fail; 602 goto fail;
564 603
604 // possibly a surrogate pair
605 if (hi >= 0xd800)
606 if (hi < 0xdc00)
607 {
608 if (dec->cur [0] != '\\' || dec->cur [1] != 'u')
609 ERR ("missing low surrogate character in surrogate pair");
610
611 dec->cur += 2;
612
613 lo = decode_4hex (dec);
614 if (lo == (UV)-1)
615 goto fail;
616
565 if (lo < 0xdc00 || lo >= 0xe000) 617 if (lo < 0xdc00 || lo >= 0xe000)
566 ERR ("surrogate pair expected"); 618 ERR ("surrogate pair expected");
567 619
568 hi = (hi - 0xD800) * 0x400 + (lo - 0xDC00) + 0x10000; 620 hi = (hi - 0xD800) * 0x400 + (lo - 0xDC00) + 0x10000;
621 }
622 else if (hi < 0xe000)
623 ERR ("missing high surrogate character in surrogate pair");
624
625 if (hi >= 0x80)
626 {
627 utf8 = 1;
628
629 cur = (char *)uvuni_to_utf8_flags (cur, hi, 0);
630 }
631 else
632 *cur++ = hi;
569 } 633 }
570 else if (hi >= 0xdc00 && hi < 0xe000)
571 ERR ("missing high surrogate character in surrogate pair");
572
573 if (hi >= 0x80)
574 { 634 break;
575 utf8 = 1;
576 635
577 APPEND_GROW (4); // at most 4 bytes for 21 bits
578 cur = (char *)uvuni_to_utf8_flags (cur, hi, 0);
579 }
580 else 636 default:
581 APPEND_CH (hi); 637 --dec->cur;
638 ERR ("illegal backslash escape sequence in string");
582 } 639 }
583 break;
584
585 default:
586 --dec->cur;
587 ERR ("illegal backslash escape sequence in string");
588 } 640 }
641 else if (ch >= 0x20 && ch <= 0x7f)
642 *cur++ = ch;
643 else if (ch >= 0x80)
644 {
645 --dec->cur;
646
647 STRLEN clen;
648 UV uch = decode_utf8 (dec->cur, dec->end - dec->cur, &clen);
649 if (clen == (STRLEN)-1)
650 ERR ("malformed UTF-8 character in JSON string");
651
652 do
653 {
654 *cur++ = *dec->cur++;
655 }
656 while (--clen);
657
658 utf8 = 1;
659 }
660 else if (!ch)
661 ERR ("unexpected end of string while parsing json string");
662 else
663 ERR ("invalid character encountered");
664
589 } 665 }
590 else if (ch >= 0x20 && ch <= 0x7f) 666 while (cur < buf + SHORT_STRING_LEN);
591 APPEND_CH (*dec->cur++); 667
592 else if (ch >= 0x80) 668 STRLEN len = cur - buf;
669
670 if (sv)
593 { 671 {
594 STRLEN clen; 672 SvGROW (sv, SvCUR (sv) + len + 1);
595 UV uch = utf8n_to_uvuni (dec->cur, dec->end - dec->cur, &clen, UTF8_CHECK_ONLY); 673 memcpy (SvPVX (sv) + SvCUR (sv), buf, len);
596 if (clen == (STRLEN)-1) 674 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 } 675 }
608 else if (dec->cur == dec->end)
609 ERR ("unexpected end of string while parsing json string");
610 else 676 else
611 ERR ("invalid character encountered"); 677 sv = newSVpvn (buf, len);
612 } 678 }
679 while (*dec->cur != '"');
613 680
614 ++dec->cur; 681 ++dec->cur;
615 682
616 SvCUR_set (sv, cur - SvPVX (sv)); 683 if (sv)
617 684 {
618 SvPOK_only (sv); 685 SvPOK_only (sv);
619 *SvEND (sv) = 0; 686 *SvEND (sv) = 0;
620 687
621 if (utf8) 688 if (utf8)
622 SvUTF8_on (sv); 689 SvUTF8_on (sv);
623 690 }
624 if (dec->flags & F_SHRINK) 691 else
625 shrink (sv); 692 sv = newSVpvn ("", 0);
626 693
627 return sv; 694 return sv;
628 695
629fail: 696fail:
630 SvREFCNT_dec (sv);
631 return 0; 697 return 0;
632} 698}
633 699
634static SV * 700static SV *
635decode_num (dec_t *dec) 701decode_num (dec_t *dec)
… …
716static SV * 782static SV *
717decode_av (dec_t *dec) 783decode_av (dec_t *dec)
718{ 784{
719 AV *av = newAV (); 785 AV *av = newAV ();
720 786
721 WS; 787 decode_ws (dec);
722 if (*dec->cur == ']') 788 if (*dec->cur == ']')
723 ++dec->cur; 789 ++dec->cur;
724 else 790 else
725 for (;;) 791 for (;;)
726 { 792 {
… …
730 if (!value) 796 if (!value)
731 goto fail; 797 goto fail;
732 798
733 av_push (av, value); 799 av_push (av, value);
734 800
735 WS; 801 decode_ws (dec);
736 802
737 if (*dec->cur == ']') 803 if (*dec->cur == ']')
738 { 804 {
739 ++dec->cur; 805 ++dec->cur;
740 break; 806 break;
… …
756static SV * 822static SV *
757decode_hv (dec_t *dec) 823decode_hv (dec_t *dec)
758{ 824{
759 HV *hv = newHV (); 825 HV *hv = newHV ();
760 826
761 WS; 827 decode_ws (dec);
762 if (*dec->cur == '}') 828 if (*dec->cur == '}')
763 ++dec->cur; 829 ++dec->cur;
764 else 830 else
765 for (;;) 831 for (;;)
766 { 832 {
767 SV *key, *value; 833 SV *key, *value;
768 834
769 WS; EXPECT_CH ('"'); 835 decode_ws (dec); EXPECT_CH ('"');
770 836
771 key = decode_str (dec); 837 key = decode_str (dec);
772 if (!key) 838 if (!key)
773 goto fail; 839 goto fail;
774 840
775 WS; EXPECT_CH (':'); 841 decode_ws (dec); EXPECT_CH (':');
776 842
777 value = decode_sv (dec); 843 value = decode_sv (dec);
778 if (!value) 844 if (!value)
779 { 845 {
780 SvREFCNT_dec (key); 846 SvREFCNT_dec (key);
… …
782 } 848 }
783 849
784 //TODO: optimise 850 //TODO: optimise
785 hv_store_ent (hv, key, value, 0); 851 hv_store_ent (hv, key, value, 0);
786 852
787 WS; 853 decode_ws (dec);
788 854
789 if (*dec->cur == '}') 855 if (*dec->cur == '}')
790 { 856 {
791 ++dec->cur; 857 ++dec->cur;
792 break; 858 break;
… …
806} 872}
807 873
808static SV * 874static SV *
809decode_sv (dec_t *dec) 875decode_sv (dec_t *dec)
810{ 876{
811 WS; 877 decode_ws (dec);
812 switch (*dec->cur) 878 switch (*dec->cur)
813 { 879 {
814 case '"': ++dec->cur; return decode_str (dec); 880 case '"': ++dec->cur; return decode_str (dec);
815 case '[': ++dec->cur; return decode_av (dec); 881 case '[': ++dec->cur; return decode_av (dec);
816 case '{': ++dec->cur; return decode_hv (dec); 882 case '{': ++dec->cur; return decode_hv (dec);
… …
909 if (!(dec.flags & F_ALLOW_NONREF) && !SvROK (sv)) 975 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)"); 976 croak ("JSON text must be an object or array (but found number, string, true, false or null, use allow_nonref to allow this)");
911 977
912 return sv; 978 return sv;
913} 979}
980
981/////////////////////////////////////////////////////////////////////////////
982// XS interface functions
914 983
915MODULE = JSON::XS PACKAGE = JSON::XS 984MODULE = JSON::XS PACKAGE = JSON::XS
916 985
917BOOT: 986BOOT:
918{ 987{
… …
972 XPUSHs (decode_json (jsonstr, *SvJSON (self))); 1041 XPUSHs (decode_json (jsonstr, *SvJSON (self)));
973 1042
974PROTOTYPES: ENABLE 1043PROTOTYPES: ENABLE
975 1044
976void to_json (SV *scalar) 1045void to_json (SV *scalar)
1046 ALIAS:
1047 objToJson = 0
977 PPCODE: 1048 PPCODE:
978 XPUSHs (encode_json (scalar, F_UTF8)); 1049 XPUSHs (encode_json (scalar, F_UTF8));
979 1050
980void from_json (SV *jsonstr) 1051void from_json (SV *jsonstr)
1052 ALIAS:
1053 jsonToObj = 0
981 PPCODE: 1054 PPCODE:
982 XPUSHs (decode_json (jsonstr, F_UTF8)); 1055 XPUSHs (decode_json (jsonstr, F_UTF8));
983 1056

Diff Legend

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