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 1.12 by root, Sat Mar 24 22:10:08 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 256 // 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/////////////////////////////////////////////////////////////////////////////
59// encoder
28 60
29// structure used for encoding JSON 61// structure used for encoding JSON
30typedef struct 62typedef struct
31{ 63{
32 char *cur; 64 char *cur; // SvPVX (sv) + current output position
33 STRLEN len; // SvLEN (sv)
34 char *end; // SvEND (sv) 65 char *end; // SvEND (sv)
35 SV *sv; 66 SV *sv; // result scalar
36 UV flags; 67 UV flags; // F_*
37 int max_recurse; 68 int indent; // indentation level
38 int indent; 69 int max_depth; // max. recursion level
39} enc_t; 70} 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 71
70static void 72static void
71need (enc_t *enc, STRLEN len) 73need (enc_t *enc, STRLEN len)
72{ 74{
73 if (enc->cur + len >= enc->end) 75 if (enc->cur + len >= enc->end)
… …
133 135
134 if (is_utf8) 136 if (is_utf8)
135 { 137 {
136 uch = utf8n_to_uvuni (str, end - str, &clen, UTF8_CHECK_ONLY); 138 uch = utf8n_to_uvuni (str, end - str, &clen, UTF8_CHECK_ONLY);
137 if (clen == (STRLEN)-1) 139 if (clen == (STRLEN)-1)
138 croak ("malformed UTF-8 character in string, cannot convert to JSON"); 140 croak ("malformed or illegal unicode character in string [%.11s], cannot convert to JSON", str);
139 } 141 }
140 else 142 else
141 { 143 {
142 uch = ch; 144 uch = ch;
143 clen = 1; 145 clen = 1;
144 } 146 }
145 147
148 if (uch > 0x10FFFFUL)
149 croak ("out of range codepoint (0x%lx) encountered, unrepresentable in JSON", (unsigned long)uch);
150
146 if (uch < 0x80 || enc->flags & F_ASCII) 151 if (uch < 0x80 || enc->flags & F_ASCII)
147 { 152 {
148 if (uch > 0xFFFFUL) 153 if (uch > 0xFFFFUL)
149 { 154 {
150 need (enc, len += 11); 155 need (enc, len += 11);
151 sprintf (enc->cur, "\\u%04x\\u%04x", 156 sprintf (enc->cur, "\\u%04x\\u%04x",
152 (uch - 0x10000) / 0x400 + 0xD800, 157 (int)((uch - 0x10000) / 0x400 + 0xD800),
153 (uch - 0x10000) % 0x400 + 0xDC00); 158 (int)((uch - 0x10000) % 0x400 + 0xDC00));
154 enc->cur += 12; 159 enc->cur += 12;
155 } 160 }
156 else 161 else
157 { 162 {
158 static char hexdigit [16] = "0123456789abcdef"; 163 static char hexdigit [16] = "0123456789abcdef";
… …
176 } 181 }
177 while (--clen); 182 while (--clen);
178 } 183 }
179 else 184 else
180 { 185 {
181 need (enc, len += 10); // never more than 11 bytes needed 186 need (enc, len += UTF8_MAX_LEN - 1); // never more than 11 bytes needed
182 enc->cur = uvuni_to_utf8_flags (enc->cur, uch, 0); 187 enc->cur = uvuni_to_utf8_flags (enc->cur, uch, 0);
183 ++str; 188 ++str;
184 } 189 }
185 } 190 }
186 } 191 }
… …
188 193
189 --len; 194 --len;
190 } 195 }
191} 196}
192 197
193#define INDENT SB \ 198static void
199encode_indent (enc_t *enc)
200{
194 if (enc->flags & F_INDENT) \ 201 if (enc->flags & F_INDENT)
195 { \ 202 {
196 int i_; \ 203 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
202 204
203#define SPACE SB need (enc, 1); encode_ch (enc, ' '); SE 205 need (enc, spaces);
204#define NL SB if (enc->flags & F_INDENT) { need (enc, 1); encode_ch (enc, '\n'); } SE 206 memset (enc->cur, ' ', spaces);
205#define COMMA SB \ 207 enc->cur += spaces;
208 }
209}
210
211static void
212encode_space (enc_t *enc)
213{
214 need (enc, 1);
206 encode_ch (enc, ','); \ 215 encode_ch (enc, ' ');
216}
217
218static void
219encode_nl (enc_t *enc)
220{
207 if (enc->flags & F_INDENT) \ 221 if (enc->flags & F_INDENT)
208 NL; \ 222 {
223 need (enc, 1);
224 encode_ch (enc, '\n');
225 }
226}
227
228static void
229encode_comma (enc_t *enc)
230{
231 encode_ch (enc, ',');
232
233 if (enc->flags & F_INDENT)
234 encode_nl (enc);
209 else if (enc->flags & F_SPACE_AFTER) \ 235 else if (enc->flags & F_SPACE_AFTER)
210 SPACE; \ 236 encode_space (enc);
211 SE 237}
212 238
213static void encode_sv (enc_t *enc, SV *sv); 239static void encode_sv (enc_t *enc, SV *sv);
214 240
215static void 241static void
216encode_av (enc_t *enc, AV *av) 242encode_av (enc_t *enc, AV *av)
217{ 243{
218 int i, len = av_len (av); 244 int i, len = av_len (av);
219 245
220 encode_ch (enc, '['); NL; 246 encode_ch (enc, '['); encode_nl (enc);
221 ++enc->indent; 247 ++enc->indent;
222 248
223 for (i = 0; i <= len; ++i) 249 for (i = 0; i <= len; ++i)
224 { 250 {
225 INDENT; 251 encode_indent (enc);
226 encode_sv (enc, *av_fetch (av, i, 0)); 252 encode_sv (enc, *av_fetch (av, i, 0));
227 253
228 if (i < len) 254 if (i < len)
229 COMMA; 255 encode_comma (enc);
230 } 256 }
231 257
232 NL; 258 encode_nl (enc);
233 259
234 --enc->indent; 260 --enc->indent;
235 INDENT; encode_ch (enc, ']'); 261 encode_indent (enc); encode_ch (enc, ']');
236} 262}
237 263
238static void 264static void
239encode_he (enc_t *enc, HE *he) 265encode_he (enc_t *enc, HE *he)
240{ 266{
… …
254 else 280 else
255 encode_str (enc, HeKEY (he), HeKLEN (he), HeKUTF8 (he)); 281 encode_str (enc, HeKEY (he), HeKLEN (he), HeKUTF8 (he));
256 282
257 encode_ch (enc, '"'); 283 encode_ch (enc, '"');
258 284
259 if (enc->flags & F_SPACE_BEFORE) SPACE; 285 if (enc->flags & F_SPACE_BEFORE) encode_space (enc);
260 encode_ch (enc, ':'); 286 encode_ch (enc, ':');
261 if (enc->flags & F_SPACE_AFTER ) SPACE; 287 if (enc->flags & F_SPACE_AFTER ) encode_space (enc);
262 encode_sv (enc, HeVAL (he)); 288 encode_sv (enc, HeVAL (he));
263} 289}
264 290
265// compare hash entries, used when all keys are bytestrings 291// compare hash entries, used when all keys are bytestrings
266static int 292static int
… …
272 HE *b = *(HE **)b_; 298 HE *b = *(HE **)b_;
273 299
274 STRLEN la = HeKLEN (a); 300 STRLEN la = HeKLEN (a);
275 STRLEN lb = HeKLEN (b); 301 STRLEN lb = HeKLEN (b);
276 302
277 if (!(cmp == memcmp (HeKEY (a), HeKEY (b), la < lb ? la : lb))) 303 if (!(cmp = memcmp (HeKEY (a), HeKEY (b), la < lb ? la : lb)))
278 cmp = la < lb ? -1 : la == lb ? 0 : 1; 304 cmp = la - lb;
279 305
280 return cmp; 306 return cmp;
281} 307}
282 308
283// compare hash entries, used when some keys are sv's or utf-x 309// compare hash entries, used when some keys are sv's or utf-x
… …
290static void 316static void
291encode_hv (enc_t *enc, HV *hv) 317encode_hv (enc_t *enc, HV *hv)
292{ 318{
293 int count, i; 319 int count, i;
294 320
295 encode_ch (enc, '{'); NL; ++enc->indent; 321 encode_ch (enc, '{'); encode_nl (enc); ++enc->indent;
296 322
297 if ((count = hv_iterinit (hv))) 323 if ((count = hv_iterinit (hv)))
298 { 324 {
299 // for canonical output we have to sort by keys first 325 // for canonical output we have to sort by keys first
300 // actually, this is mostly due to the stupid so-called 326 // actually, this is mostly due to the stupid so-called
301 // security workaround added somewhere in 5.8.x. 327 // security workaround added somewhere in 5.8.x.
302 // that randomises hash orderings 328 // that randomises hash orderings
303 if (enc->flags & F_CANONICAL) 329 if (enc->flags & F_CANONICAL)
304 { 330 {
305 HE *he, *hes [count]; 331 HE *he, *hes [count]; // if your compiler dies here, you need to enable C99 mode
306 int fast = 1; 332 int fast = 1;
307 333
308 i = 0; 334 i = 0;
309 while ((he = hv_iternext (hv))) 335 while ((he = hv_iternext (hv)))
310 { 336 {
… …
335 LEAVE; 361 LEAVE;
336 } 362 }
337 363
338 for (i = 0; i < count; ++i) 364 for (i = 0; i < count; ++i)
339 { 365 {
340 INDENT; 366 encode_indent (enc);
341 encode_he (enc, hes [i]); 367 encode_he (enc, hes [i]);
342 368
343 if (i < count - 1) 369 if (i < count - 1)
344 COMMA; 370 encode_comma (enc);
345 } 371 }
346 372
347 NL; 373 encode_nl (enc);
348 } 374 }
349 else 375 else
350 { 376 {
351 SV *sv; 377 SV *sv;
352 HE *he = hv_iternext (hv); 378 HE *he = hv_iternext (hv);
353 379
354 for (;;) 380 for (;;)
355 { 381 {
356 INDENT; 382 encode_indent (enc);
357 encode_he (enc, he); 383 encode_he (enc, he);
358 384
359 if (!(he = hv_iternext (hv))) 385 if (!(he = hv_iternext (hv)))
360 break; 386 break;
361 387
362 COMMA; 388 encode_comma (enc);
363 } 389 }
364 390
365 NL; 391 encode_nl (enc);
366 } 392 }
367 } 393 }
368 394
369 --enc->indent; INDENT; encode_ch (enc, '}'); 395 --enc->indent; encode_indent (enc); encode_ch (enc, '}');
370} 396}
371 397
372static void 398static void
373encode_sv (enc_t *enc, SV *sv) 399encode_sv (enc_t *enc, SV *sv)
374{ 400{
… …
396 ? snprintf (enc->cur, 64, "%"UVuf, (UV)SvUVX (sv)) 422 ? snprintf (enc->cur, 64, "%"UVuf, (UV)SvUVX (sv))
397 : snprintf (enc->cur, 64, "%"IVdf, (IV)SvIVX (sv)); 423 : snprintf (enc->cur, 64, "%"IVdf, (IV)SvIVX (sv));
398 } 424 }
399 else if (SvROK (sv)) 425 else if (SvROK (sv))
400 { 426 {
401 if (!--enc->max_recurse) 427 SV *rv = SvRV (sv);
428
429 if (enc->indent >= enc->max_depth)
402 croak ("data structure too deep (hit recursion limit)"); 430 croak ("data structure too deep (hit recursion limit)");
403 431
404 sv = SvRV (sv);
405
406 switch (SvTYPE (sv)) 432 switch (SvTYPE (rv))
407 { 433 {
408 case SVt_PVAV: encode_av (enc, (AV *)sv); break; 434 case SVt_PVAV: encode_av (enc, (AV *)rv); break;
409 case SVt_PVHV: encode_hv (enc, (HV *)sv); break; 435 case SVt_PVHV: encode_hv (enc, (HV *)rv); break;
410 436
411 default: 437 default:
412 croak ("JSON can only represent references to arrays or hashes"); 438 croak ("encountered %s, but JSON can only represent references to arrays or hashes",
439 SvPV_nolen (sv));
413 } 440 }
414 } 441 }
415 else if (!SvOK (sv)) 442 else if (!SvOK (sv))
416 encode_str (enc, "null", 4, 0); 443 encode_str (enc, "null", 4, 0);
417 else 444 else
418 croak ("encountered perl type that JSON cannot handle"); 445 croak ("encountered perl type (%s,0x%x) that JSON cannot handle, you might want to report this",
446 SvPV_nolen (sv), SvFLAGS (sv));
419} 447}
420 448
421static SV * 449static SV *
422encode_json (SV *scalar, UV flags) 450encode_json (SV *scalar, UV flags)
423{ 451{
424 if (!(flags & F_ALLOW_NONREF) && !SvROK (scalar)) 452 if (!(flags & F_ALLOW_NONREF) && !SvROK (scalar))
425 croak ("hash- or arraref required (not a simple scalar, use allow_nonref to allow this)"); 453 croak ("hash- or arrayref expected (not a simple scalar, use allow_nonref to allow this)");
426 454
427 enc_t enc; 455 enc_t enc;
428 enc.flags = flags; 456 enc.flags = flags;
429 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 457 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
430 enc.cur = SvPVX (enc.sv); 458 enc.cur = SvPVX (enc.sv);
431 enc.end = SvEND (enc.sv); 459 enc.end = SvEND (enc.sv);
432 enc.max_recurse = 0;
433 enc.indent = 0; 460 enc.indent = 0;
461 enc.max_depth = 0x7fffffffUL;
434 462
435 SvPOK_only (enc.sv); 463 SvPOK_only (enc.sv);
436 encode_sv (&enc, scalar); 464 encode_sv (&enc, scalar);
437 465
438 if (!(flags & (F_ASCII | F_UTF8))) 466 if (!(flags & (F_ASCII | F_UTF8)))
… …
445 473
446 return enc.sv; 474 return enc.sv;
447} 475}
448 476
449///////////////////////////////////////////////////////////////////////////// 477/////////////////////////////////////////////////////////////////////////////
478// decoder
450 479
451#define WS \ 480// structure used for decoding JSON
481typedef struct
482{
483 char *cur; // current parser pointer
484 char *end; // end of input string
485 const char *err; // parse error, if != 0
486 UV flags; // F_*
487} dec_t;
488
489static void
490decode_ws (dec_t *dec)
491{
452 for (;;) \ 492 for (;;)
453 { \ 493 {
454 char ch = *dec->cur; \ 494 char ch = *dec->cur;
495
455 if (ch > 0x20 \ 496 if (ch > 0x20
456 || (ch != 0x20 && ch != 0x0a && ch != 0x0d && ch != 0x09)) \ 497 || (ch != 0x20 && ch != 0x0a && ch != 0x0d && ch != 0x09))
457 break; \ 498 break;
499
458 ++dec->cur; \ 500 ++dec->cur;
459 } 501 }
502}
460 503
461#define ERR(reason) SB dec->err = reason; goto fail; SE 504#define ERR(reason) SB dec->err = reason; goto fail; SE
462#define EXPECT_CH(ch) SB \ 505#define EXPECT_CH(ch) SB \
463 if (*dec->cur != ch) \ 506 if (*dec->cur != ch) \
464 ERR (# ch " expected"); \ 507 ERR (# ch " expected"); \
… …
471 514
472static UV 515static UV
473decode_4hex (dec_t *dec) 516decode_4hex (dec_t *dec)
474{ 517{
475 signed char d1, d2, d3, d4; 518 signed char d1, d2, d3, d4;
519 unsigned char *cur = (unsigned char *)dec->cur;
476 520
477 d1 = decode_hexdigit [((unsigned char *)dec->cur) [0]];
478 if (d1 < 0) ERR ("four hexadecimal digits expected"); 521 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"); 522 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"); 523 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"); 524 d4 = decode_hexdigit [cur [3]]; if (d4 < 0) ERR ("four hexadecimal digits expected");
485 525
486 dec->cur += 4; 526 dec->cur += 4;
487 527
488 return ((UV)d1) << 12 528 return ((UV)d1) << 12
489 | ((UV)d2) << 8 529 | ((UV)d2) << 8
… …
492 532
493fail: 533fail:
494 return (UV)-1; 534 return (UV)-1;
495} 535}
496 536
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 * 537static SV *
513decode_str (dec_t *dec) 538decode_str (dec_t *dec)
514{ 539{
515 SV *sv = NEWSV (0,2); 540 SV *sv = 0;
516 int utf8 = 0; 541 int utf8 = 0;
517 char *cur = SvPVX (sv);
518 char *end = SvEND (sv);
519 542
520 for (;;) 543 do
521 { 544 {
522 unsigned char ch = *(unsigned char *)dec->cur; 545 char buf [SHORT_STRING_LEN + UTF8_MAX_LEN];
546 char *cur = buf;
523 547
524 if (ch == '"') 548 do
525 break;
526 else if (ch == '\\')
527 { 549 {
528 switch (*++dec->cur) 550 unsigned char ch = *(unsigned char *)dec->cur++;
551
552 if (ch == '"')
529 { 553 {
530 case '\\': 554 --dec->cur;
531 case '/': 555 break;
532 case '"': APPEND_CH (*dec->cur++); break; 556 }
533 557 else if (ch == '\\')
534 case 'b': APPEND_CH ('\010'); ++dec->cur; break; 558 {
535 case 't': APPEND_CH ('\011'); ++dec->cur; break; 559 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 { 560 {
542 UV lo, hi; 561 case '\\':
543 ++dec->cur; 562 case '/':
563 case '"': *cur++ = *dec->cur++; break;
544 564
545 hi = decode_4hex (dec); 565 case 'b': ++dec->cur; *cur++ = '\010'; break;
546 if (hi == (UV)-1) 566 case 't': ++dec->cur; *cur++ = '\011'; break;
547 goto fail; 567 case 'n': ++dec->cur; *cur++ = '\012'; break;
568 case 'f': ++dec->cur; *cur++ = '\014'; break;
569 case 'r': ++dec->cur; *cur++ = '\015'; break;
548 570
549 // possibly a surrogate pair 571 case 'u':
550 if (hi >= 0xd800 && hi < 0xdc00)
551 { 572 {
552 if (dec->cur [0] != '\\' || dec->cur [1] != 'u') 573 UV lo, hi;
553 ERR ("missing low surrogate character in surrogate pair");
554
555 dec->cur += 2; 574 ++dec->cur;
556 575
557 lo = decode_4hex (dec); 576 hi = decode_4hex (dec);
558 if (lo == (UV)-1) 577 if (hi == (UV)-1)
559 goto fail; 578 goto fail;
560 579
580 // possibly a surrogate pair
581 if (hi >= 0xd800)
582 if (hi < 0xdc00)
583 {
584 if (dec->cur [0] != '\\' || dec->cur [1] != 'u')
585 ERR ("missing low surrogate character in surrogate pair");
586
587 dec->cur += 2;
588
589 lo = decode_4hex (dec);
590 if (lo == (UV)-1)
591 goto fail;
592
561 if (lo < 0xdc00 || lo >= 0xe000) 593 if (lo < 0xdc00 || lo >= 0xe000)
562 ERR ("surrogate pair expected"); 594 ERR ("surrogate pair expected");
563 595
564 hi = (hi - 0xD800) * 0x400 + (lo - 0xDC00) + 0x10000; 596 hi = (hi - 0xD800) * 0x400 + (lo - 0xDC00) + 0x10000;
597 }
598 else if (hi < 0xe000)
599 ERR ("missing high surrogate character in surrogate pair");
600
601 if (hi >= 0x80)
602 {
603 utf8 = 1;
604
605 cur = (char *)uvuni_to_utf8_flags (cur, hi, 0);
606 }
607 else
608 *cur++ = hi;
565 } 609 }
566 else if (hi >= 0xdc00 && hi < 0xe000)
567 ERR ("missing high surrogate character in surrogate pair");
568
569 if (hi >= 0x80)
570 { 610 break;
571 utf8 = 1;
572 611
573 APPEND_GROW (4); // at most 4 bytes for 21 bits
574 cur = (char *)uvuni_to_utf8_flags (cur, hi, 0);
575 }
576 else 612 default:
577 APPEND_CH (hi); 613 --dec->cur;
614 ERR ("illegal backslash escape sequence in string");
578 } 615 }
579 break;
580
581 default:
582 --dec->cur;
583 ERR ("illegal backslash escape sequence in string");
584 } 616 }
617 else if (ch >= 0x20 && ch <= 0x7f)
618 *cur++ = ch;
619 else if (ch >= 0x80)
620 {
621 --dec->cur;
622
623 STRLEN clen;
624 UV uch = utf8n_to_uvuni (dec->cur, dec->end - dec->cur, &clen, UTF8_CHECK_ONLY);
625 if (clen == (STRLEN)-1)
626 ERR ("malformed UTF-8 character in JSON string");
627
628 do
629 {
630 *cur++ = *dec->cur++;
631 }
632 while (--clen);
633
634 utf8 = 1;
635 }
636 else if (!ch)
637 ERR ("unexpected end of string while parsing json string");
638 else
639 ERR ("invalid character encountered");
640
585 } 641 }
586 else if (ch >= 0x20 && ch <= 0x7f) 642 while (cur < buf + SHORT_STRING_LEN);
587 APPEND_CH (*dec->cur++); 643
588 else if (ch >= 0x80) 644 STRLEN len = cur - buf;
645
646 if (sv)
589 { 647 {
590 STRLEN clen; 648 SvGROW (sv, SvCUR (sv) + len + 1);
591 UV uch = utf8n_to_uvuni (dec->cur, dec->end - dec->cur, &clen, UTF8_CHECK_ONLY); 649 memcpy (SvPVX (sv) + SvCUR (sv), buf, len);
592 if (clen == (STRLEN)-1) 650 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 } 651 }
604 else if (dec->cur == dec->end)
605 ERR ("unexpected end of string while parsing json string");
606 else 652 else
607 ERR ("invalid character encountered"); 653 sv = newSVpvn (buf, len);
608 } 654 }
655 while (*dec->cur != '"');
609 656
610 ++dec->cur; 657 ++dec->cur;
611 658
612 SvCUR_set (sv, cur - SvPVX (sv)); 659 if (sv)
613 660 {
614 SvPOK_only (sv); 661 SvPOK_only (sv);
615 *SvEND (sv) = 0; 662 *SvEND (sv) = 0;
616 663
617 if (utf8) 664 if (utf8)
618 SvUTF8_on (sv); 665 SvUTF8_on (sv);
619 666 }
620 if (dec->flags & F_SHRINK) 667 else
621 shrink (sv); 668 sv = newSVpvn ("", 0);
622 669
623 return sv; 670 return sv;
624 671
625fail: 672fail:
626 SvREFCNT_dec (sv);
627 return 0; 673 return 0;
628} 674}
629 675
630static SV * 676static SV *
631decode_num (dec_t *dec) 677decode_num (dec_t *dec)
… …
712static SV * 758static SV *
713decode_av (dec_t *dec) 759decode_av (dec_t *dec)
714{ 760{
715 AV *av = newAV (); 761 AV *av = newAV ();
716 762
717 WS; 763 decode_ws (dec);
718 if (*dec->cur == ']') 764 if (*dec->cur == ']')
719 ++dec->cur; 765 ++dec->cur;
720 else 766 else
721 for (;;) 767 for (;;)
722 { 768 {
… …
726 if (!value) 772 if (!value)
727 goto fail; 773 goto fail;
728 774
729 av_push (av, value); 775 av_push (av, value);
730 776
731 WS; 777 decode_ws (dec);
732 778
733 if (*dec->cur == ']') 779 if (*dec->cur == ']')
734 { 780 {
735 ++dec->cur; 781 ++dec->cur;
736 break; 782 break;
… …
752static SV * 798static SV *
753decode_hv (dec_t *dec) 799decode_hv (dec_t *dec)
754{ 800{
755 HV *hv = newHV (); 801 HV *hv = newHV ();
756 802
757 WS; 803 decode_ws (dec);
758 if (*dec->cur == '}') 804 if (*dec->cur == '}')
759 ++dec->cur; 805 ++dec->cur;
760 else 806 else
761 for (;;) 807 for (;;)
762 { 808 {
763 SV *key, *value; 809 SV *key, *value;
764 810
765 WS; EXPECT_CH ('"'); 811 decode_ws (dec); EXPECT_CH ('"');
766 812
767 key = decode_str (dec); 813 key = decode_str (dec);
768 if (!key) 814 if (!key)
769 goto fail; 815 goto fail;
770 816
771 WS; EXPECT_CH (':'); 817 decode_ws (dec); EXPECT_CH (':');
772 818
773 value = decode_sv (dec); 819 value = decode_sv (dec);
774 if (!value) 820 if (!value)
775 { 821 {
776 SvREFCNT_dec (key); 822 SvREFCNT_dec (key);
… …
778 } 824 }
779 825
780 //TODO: optimise 826 //TODO: optimise
781 hv_store_ent (hv, key, value, 0); 827 hv_store_ent (hv, key, value, 0);
782 828
783 WS; 829 decode_ws (dec);
784 830
785 if (*dec->cur == '}') 831 if (*dec->cur == '}')
786 { 832 {
787 ++dec->cur; 833 ++dec->cur;
788 break; 834 break;
… …
802} 848}
803 849
804static SV * 850static SV *
805decode_sv (dec_t *dec) 851decode_sv (dec_t *dec)
806{ 852{
807 WS; 853 decode_ws (dec);
808 switch (*dec->cur) 854 switch (*dec->cur)
809 { 855 {
810 case '"': ++dec->cur; return decode_str (dec); 856 case '"': ++dec->cur; return decode_str (dec);
811 case '[': ++dec->cur; return decode_av (dec); 857 case '[': ++dec->cur; return decode_av (dec);
812 case '{': ++dec->cur; return decode_hv (dec); 858 case '{': ++dec->cur; return decode_hv (dec);
… …
901 } 947 }
902 948
903 sv = sv_2mortal (sv); 949 sv = sv_2mortal (sv);
904 950
905 if (!(dec.flags & F_ALLOW_NONREF) && !SvROK (sv)) 951 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)"); 952 croak ("JSON text must be an object or array (but found number, string, true, false or null, use allow_nonref to allow this)");
907 953
908 return sv; 954 return sv;
909} 955}
956
957/////////////////////////////////////////////////////////////////////////////
958// XS interface functions
910 959
911MODULE = JSON::XS PACKAGE = JSON::XS 960MODULE = JSON::XS PACKAGE = JSON::XS
912 961
913BOOT: 962BOOT:
914{ 963{
… …
941 utf8 = F_UTF8 990 utf8 = F_UTF8
942 indent = F_INDENT 991 indent = F_INDENT
943 canonical = F_CANONICAL 992 canonical = F_CANONICAL
944 space_before = F_SPACE_BEFORE 993 space_before = F_SPACE_BEFORE
945 space_after = F_SPACE_AFTER 994 space_after = F_SPACE_AFTER
946 json_rpc = F_JSON_RPC
947 pretty = F_PRETTY 995 pretty = F_PRETTY
948 allow_nonref = F_ALLOW_NONREF 996 allow_nonref = F_ALLOW_NONREF
949 shrink = F_SHRINK 997 shrink = F_SHRINK
950 CODE: 998 CODE:
951{ 999{

Diff Legend

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