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 1.17 by root, Sun Mar 25 02:46:41 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
18// F_SKIPINVALID?
19// F_EXECCODEREF?
20// F_SELFCONVERT? <=> { &__class__ => }
21
19#define F_PRETTY F_INDENT | F_SPACE_BEFORE | F_SPACE_AFTER 22#define F_PRETTY F_INDENT | F_SPACE_BEFORE | F_SPACE_AFTER
20#define F_DEFAULT 0 23#define F_DEFAULT 0
21 24
22#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
23 30
24#define SB do { 31#define SB do {
25#define SE } while (0) 32#define SE } while (0)
26 33
27static 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
28 83
29// structure used for encoding JSON 84// structure used for encoding JSON
30typedef struct 85typedef struct
31{ 86{
32 char *cur; 87 char *cur; // SvPVX (sv) + current output position
33 STRLEN len; // SvLEN (sv)
34 char *end; // SvEND (sv) 88 char *end; // SvEND (sv)
35 SV *sv; 89 SV *sv; // result scalar
36 UV flags; 90 UV flags; // F_*
37 int max_recurse; 91 int indent; // indentation level
38 int indent; 92 int max_depth; // max. recursion level
39} enc_t; 93} 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 94
70static void 95static void
71need (enc_t *enc, STRLEN len) 96need (enc_t *enc, STRLEN len)
72{ 97{
73 if (enc->cur + len >= enc->end) 98 if (enc->cur + len >= enc->end)
… …
131 STRLEN clen; 156 STRLEN clen;
132 UV uch; 157 UV uch;
133 158
134 if (is_utf8) 159 if (is_utf8)
135 { 160 {
136 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);
137 if (clen == (STRLEN)-1) 163 if (clen == (STRLEN)-1)
138 croak ("malformed UTF-8 character in string, cannot convert to JSON"); 164 croak ("malformed or illegal unicode character in string [%.11s], cannot convert to JSON", str);
139 } 165 }
140 else 166 else
141 { 167 {
142 uch = ch; 168 uch = ch;
143 clen = 1; 169 clen = 1;
144 } 170 }
145 171
172 if (uch > 0x10FFFFUL)
173 croak ("out of range codepoint (0x%lx) encountered, unrepresentable in JSON", (unsigned long)uch);
174
146 if (uch < 0x80 || enc->flags & F_ASCII) 175 if (uch < 0x80 || enc->flags & F_ASCII)
147 { 176 {
148 if (uch > 0xFFFFUL) 177 if (uch > 0xFFFFUL)
149 { 178 {
150 need (enc, len += 11); 179 need (enc, len += 11);
151 sprintf (enc->cur, "\\u%04x\\u%04x", 180 sprintf (enc->cur, "\\u%04x\\u%04x",
152 (uch - 0x10000) / 0x400 + 0xD800, 181 (int)((uch - 0x10000) / 0x400 + 0xD800),
153 (uch - 0x10000) % 0x400 + 0xDC00); 182 (int)((uch - 0x10000) % 0x400 + 0xDC00));
154 enc->cur += 12; 183 enc->cur += 12;
155 } 184 }
156 else 185 else
157 { 186 {
158 static char hexdigit [16] = "0123456789abcdef"; 187 static char hexdigit [16] = "0123456789abcdef";
… …
176 } 205 }
177 while (--clen); 206 while (--clen);
178 } 207 }
179 else 208 else
180 { 209 {
181 need (enc, len += 10); // never more than 11 bytes needed 210 need (enc, len += UTF8_MAX_LEN - 1); // never more than 11 bytes needed
182 enc->cur = uvuni_to_utf8_flags (enc->cur, uch, 0); 211 enc->cur = uvuni_to_utf8_flags (enc->cur, uch, 0);
183 ++str; 212 ++str;
184 } 213 }
185 } 214 }
186 } 215 }
… …
188 217
189 --len; 218 --len;
190 } 219 }
191} 220}
192 221
193#define INDENT SB \ 222static void
223encode_indent (enc_t *enc)
224{
194 if (enc->flags & F_INDENT) \ 225 if (enc->flags & F_INDENT)
195 { \ 226 {
196 int i_; \ 227 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
202 228
203#define SPACE SB need (enc, 1); encode_ch (enc, ' '); SE 229 need (enc, spaces);
204#define NL SB if (enc->flags & F_INDENT) { need (enc, 1); encode_ch (enc, '\n'); } SE 230 memset (enc->cur, ' ', spaces);
205#define COMMA SB \ 231 enc->cur += spaces;
232 }
233}
234
235static void
236encode_space (enc_t *enc)
237{
238 need (enc, 1);
206 encode_ch (enc, ','); \ 239 encode_ch (enc, ' ');
240}
241
242static void
243encode_nl (enc_t *enc)
244{
207 if (enc->flags & F_INDENT) \ 245 if (enc->flags & F_INDENT)
208 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);
209 else if (enc->flags & F_SPACE_AFTER) \ 259 else if (enc->flags & F_SPACE_AFTER)
210 SPACE; \ 260 encode_space (enc);
211 SE 261}
212 262
213static void encode_sv (enc_t *enc, SV *sv); 263static void encode_sv (enc_t *enc, SV *sv);
214 264
215static void 265static void
216encode_av (enc_t *enc, AV *av) 266encode_av (enc_t *enc, AV *av)
217{ 267{
218 int i, len = av_len (av); 268 int i, len = av_len (av);
219 269
220 encode_ch (enc, '['); NL; 270 encode_ch (enc, '['); encode_nl (enc);
221 ++enc->indent; 271 ++enc->indent;
222 272
223 for (i = 0; i <= len; ++i) 273 for (i = 0; i <= len; ++i)
224 { 274 {
225 INDENT; 275 encode_indent (enc);
226 encode_sv (enc, *av_fetch (av, i, 0)); 276 encode_sv (enc, *av_fetch (av, i, 0));
227 277
228 if (i < len) 278 if (i < len)
229 COMMA; 279 encode_comma (enc);
230 } 280 }
231 281
232 NL; 282 encode_nl (enc);
233 283
234 --enc->indent; 284 --enc->indent;
235 INDENT; encode_ch (enc, ']'); 285 encode_indent (enc); encode_ch (enc, ']');
236} 286}
237 287
238static void 288static void
239encode_he (enc_t *enc, HE *he) 289encode_he (enc_t *enc, HE *he)
240{ 290{
… …
254 else 304 else
255 encode_str (enc, HeKEY (he), HeKLEN (he), HeKUTF8 (he)); 305 encode_str (enc, HeKEY (he), HeKLEN (he), HeKUTF8 (he));
256 306
257 encode_ch (enc, '"'); 307 encode_ch (enc, '"');
258 308
259 if (enc->flags & F_SPACE_BEFORE) SPACE; 309 if (enc->flags & F_SPACE_BEFORE) encode_space (enc);
260 encode_ch (enc, ':'); 310 encode_ch (enc, ':');
261 if (enc->flags & F_SPACE_AFTER ) SPACE; 311 if (enc->flags & F_SPACE_AFTER ) encode_space (enc);
262 encode_sv (enc, HeVAL (he)); 312 encode_sv (enc, HeVAL (he));
263} 313}
264 314
265// compare hash entries, used when all keys are bytestrings 315// compare hash entries, used when all keys are bytestrings
266static int 316static int
… …
272 HE *b = *(HE **)b_; 322 HE *b = *(HE **)b_;
273 323
274 STRLEN la = HeKLEN (a); 324 STRLEN la = HeKLEN (a);
275 STRLEN lb = HeKLEN (b); 325 STRLEN lb = HeKLEN (b);
276 326
277 if (!(cmp == memcmp (HeKEY (a), HeKEY (b), la < lb ? la : lb))) 327 if (!(cmp = memcmp (HeKEY (a), HeKEY (b), la < lb ? la : lb)))
278 cmp = la < lb ? -1 : la == lb ? 0 : 1; 328 cmp = la - lb;
279 329
280 return cmp; 330 return cmp;
281} 331}
282 332
283// 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
… …
290static void 340static void
291encode_hv (enc_t *enc, HV *hv) 341encode_hv (enc_t *enc, HV *hv)
292{ 342{
293 int count, i; 343 int count, i;
294 344
295 encode_ch (enc, '{'); NL; ++enc->indent; 345 encode_ch (enc, '{'); encode_nl (enc); ++enc->indent;
296 346
297 if ((count = hv_iterinit (hv))) 347 if ((count = hv_iterinit (hv)))
298 { 348 {
299 // for canonical output we have to sort by keys first 349 // for canonical output we have to sort by keys first
300 // actually, this is mostly due to the stupid so-called 350 // actually, this is mostly due to the stupid so-called
301 // security workaround added somewhere in 5.8.x. 351 // security workaround added somewhere in 5.8.x.
302 // that randomises hash orderings 352 // that randomises hash orderings
303 if (enc->flags & F_CANONICAL) 353 if (enc->flags & F_CANONICAL)
304 { 354 {
305 HE *he, *hes [count]; 355 HE *he, *hes [count]; // if your compiler dies here, you need to enable C99 mode
306 int fast = 1; 356 int fast = 1;
307 357
308 i = 0; 358 i = 0;
309 while ((he = hv_iternext (hv))) 359 while ((he = hv_iternext (hv)))
310 { 360 {
… …
335 LEAVE; 385 LEAVE;
336 } 386 }
337 387
338 for (i = 0; i < count; ++i) 388 for (i = 0; i < count; ++i)
339 { 389 {
340 INDENT; 390 encode_indent (enc);
341 encode_he (enc, hes [i]); 391 encode_he (enc, hes [i]);
342 392
343 if (i < count - 1) 393 if (i < count - 1)
344 COMMA; 394 encode_comma (enc);
345 } 395 }
346 396
347 NL; 397 encode_nl (enc);
348 } 398 }
349 else 399 else
350 { 400 {
351 SV *sv; 401 SV *sv;
352 HE *he = hv_iternext (hv); 402 HE *he = hv_iternext (hv);
353 403
354 for (;;) 404 for (;;)
355 { 405 {
356 INDENT; 406 encode_indent (enc);
357 encode_he (enc, he); 407 encode_he (enc, he);
358 408
359 if (!(he = hv_iternext (hv))) 409 if (!(he = hv_iternext (hv)))
360 break; 410 break;
361 411
362 COMMA; 412 encode_comma (enc);
363 } 413 }
364 414
365 NL; 415 encode_nl (enc);
366 } 416 }
367 } 417 }
368 418
369 --enc->indent; INDENT; encode_ch (enc, '}'); 419 --enc->indent; encode_indent (enc); encode_ch (enc, '}');
370} 420}
371 421
372static void 422static void
373encode_sv (enc_t *enc, SV *sv) 423encode_sv (enc_t *enc, SV *sv)
374{ 424{
… …
396 ? snprintf (enc->cur, 64, "%"UVuf, (UV)SvUVX (sv)) 446 ? snprintf (enc->cur, 64, "%"UVuf, (UV)SvUVX (sv))
397 : snprintf (enc->cur, 64, "%"IVdf, (IV)SvIVX (sv)); 447 : snprintf (enc->cur, 64, "%"IVdf, (IV)SvIVX (sv));
398 } 448 }
399 else if (SvROK (sv)) 449 else if (SvROK (sv))
400 { 450 {
401 if (!--enc->max_recurse) 451 SV *rv = SvRV (sv);
452
453 if (enc->indent >= enc->max_depth)
402 croak ("data structure too deep (hit recursion limit)"); 454 croak ("data structure too deep (hit recursion limit)");
403 455
404 sv = SvRV (sv);
405
406 switch (SvTYPE (sv)) 456 switch (SvTYPE (rv))
407 { 457 {
408 case SVt_PVAV: encode_av (enc, (AV *)sv); break; 458 case SVt_PVAV: encode_av (enc, (AV *)rv); break;
409 case SVt_PVHV: encode_hv (enc, (HV *)sv); break; 459 case SVt_PVHV: encode_hv (enc, (HV *)rv); break;
410 460
411 default: 461 default:
412 croak ("JSON can only represent references to arrays or hashes"); 462 croak ("encountered %s, but JSON can only represent references to arrays or hashes",
463 SvPV_nolen (sv));
413 } 464 }
414 } 465 }
415 else if (!SvOK (sv)) 466 else if (!SvOK (sv))
416 encode_str (enc, "null", 4, 0); 467 encode_str (enc, "null", 4, 0);
417 else 468 else
418 croak ("encountered perl type that JSON cannot handle"); 469 croak ("encountered perl type (%s,0x%x) that JSON cannot handle, you might want to report this",
470 SvPV_nolen (sv), SvFLAGS (sv));
419} 471}
420 472
421static SV * 473static SV *
422encode_json (SV *scalar, UV flags) 474encode_json (SV *scalar, UV flags)
423{ 475{
424 if (!(flags & F_ALLOW_NONREF) && !SvROK (scalar)) 476 if (!(flags & F_ALLOW_NONREF) && !SvROK (scalar))
425 croak ("hash- or arraref required (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)");
426 478
427 enc_t enc; 479 enc_t enc;
428 enc.flags = flags; 480 enc.flags = flags;
429 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 481 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
430 enc.cur = SvPVX (enc.sv); 482 enc.cur = SvPVX (enc.sv);
431 enc.end = SvEND (enc.sv); 483 enc.end = SvEND (enc.sv);
432 enc.max_recurse = 0;
433 enc.indent = 0; 484 enc.indent = 0;
485 enc.max_depth = 0x7fffffffUL;
434 486
435 SvPOK_only (enc.sv); 487 SvPOK_only (enc.sv);
436 encode_sv (&enc, scalar); 488 encode_sv (&enc, scalar);
437 489
438 if (!(flags & (F_ASCII | F_UTF8))) 490 if (!(flags & (F_ASCII | F_UTF8)))
… …
445 497
446 return enc.sv; 498 return enc.sv;
447} 499}
448 500
449///////////////////////////////////////////////////////////////////////////// 501/////////////////////////////////////////////////////////////////////////////
502// decoder
450 503
451#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{
452 for (;;) \ 516 for (;;)
453 { \ 517 {
454 char ch = *dec->cur; \ 518 char ch = *dec->cur;
519
455 if (ch > 0x20 \ 520 if (ch > 0x20
456 || (ch != 0x20 && ch != 0x0a && ch != 0x0d && ch != 0x09)) \ 521 || (ch != 0x20 && ch != 0x0a && ch != 0x0d && ch != 0x09))
457 break; \ 522 break;
523
458 ++dec->cur; \ 524 ++dec->cur;
459 } 525 }
526}
460 527
461#define ERR(reason) SB dec->err = reason; goto fail; SE 528#define ERR(reason) SB dec->err = reason; goto fail; SE
462#define EXPECT_CH(ch) SB \ 529#define EXPECT_CH(ch) SB \
463 if (*dec->cur != ch) \ 530 if (*dec->cur != ch) \
464 ERR (# ch " expected"); \ 531 ERR (# ch " expected"); \
… …
471 538
472static UV 539static UV
473decode_4hex (dec_t *dec) 540decode_4hex (dec_t *dec)
474{ 541{
475 signed char d1, d2, d3, d4; 542 signed char d1, d2, d3, d4;
543 unsigned char *cur = (unsigned char *)dec->cur;
476 544
477 d1 = decode_hexdigit [((unsigned char *)dec->cur) [0]];
478 if (d1 < 0) ERR ("four hexadecimal digits expected"); 545 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"); 546 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"); 547 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"); 548 d4 = decode_hexdigit [cur [3]]; if (d4 < 0) ERR ("four hexadecimal digits expected");
485 549
486 dec->cur += 4; 550 dec->cur += 4;
487 551
488 return ((UV)d1) << 12 552 return ((UV)d1) << 12
489 | ((UV)d2) << 8 553 | ((UV)d2) << 8
… …
492 556
493fail: 557fail:
494 return (UV)-1; 558 return (UV)-1;
495} 559}
496 560
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 * 561static SV *
513decode_str (dec_t *dec) 562decode_str (dec_t *dec)
514{ 563{
515 SV *sv = NEWSV (0,2); 564 SV *sv = 0;
516 int utf8 = 0; 565 int utf8 = 0;
517 char *cur = SvPVX (sv);
518 char *end = SvEND (sv);
519 566
520 for (;;) 567 do
521 { 568 {
522 unsigned char ch = *(unsigned char *)dec->cur; 569 char buf [SHORT_STRING_LEN + UTF8_MAX_LEN];
570 char *cur = buf;
523 571
524 if (ch == '"') 572 do
525 break;
526 else if (ch == '\\')
527 { 573 {
528 switch (*++dec->cur) 574 unsigned char ch = *(unsigned char *)dec->cur++;
575
576 if (ch == '"')
529 { 577 {
530 case '\\': 578 --dec->cur;
531 case '/': 579 break;
532 case '"': APPEND_CH (*dec->cur++); break; 580 }
533 581 else if (ch == '\\')
534 case 'b': APPEND_CH ('\010'); ++dec->cur; break; 582 {
535 case 't': APPEND_CH ('\011'); ++dec->cur; break; 583 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 { 584 {
542 UV lo, hi; 585 case '\\':
543 ++dec->cur; 586 case '/':
587 case '"': *cur++ = *dec->cur++; break;
544 588
545 hi = decode_4hex (dec); 589 case 'b': ++dec->cur; *cur++ = '\010'; break;
546 if (hi == (UV)-1) 590 case 't': ++dec->cur; *cur++ = '\011'; break;
547 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;
548 594
549 // possibly a surrogate pair 595 case 'u':
550 if (hi >= 0xd800 && hi < 0xdc00)
551 { 596 {
552 if (dec->cur [0] != '\\' || dec->cur [1] != 'u') 597 UV lo, hi;
553 ERR ("missing low surrogate character in surrogate pair");
554
555 dec->cur += 2; 598 ++dec->cur;
556 599
557 lo = decode_4hex (dec); 600 hi = decode_4hex (dec);
558 if (lo == (UV)-1) 601 if (hi == (UV)-1)
559 goto fail; 602 goto fail;
560 603
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
561 if (lo < 0xdc00 || lo >= 0xe000) 617 if (lo < 0xdc00 || lo >= 0xe000)
562 ERR ("surrogate pair expected"); 618 ERR ("surrogate pair expected");
563 619
564 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;
565 } 633 }
566 else if (hi >= 0xdc00 && hi < 0xe000)
567 ERR ("missing high surrogate character in surrogate pair");
568
569 if (hi >= 0x80)
570 { 634 break;
571 utf8 = 1;
572 635
573 APPEND_GROW (4); // at most 4 bytes for 21 bits
574 cur = (char *)uvuni_to_utf8_flags (cur, hi, 0);
575 }
576 else 636 default:
577 APPEND_CH (hi); 637 --dec->cur;
638 ERR ("illegal backslash escape sequence in string");
578 } 639 }
579 break;
580
581 default:
582 --dec->cur;
583 ERR ("illegal backslash escape sequence in string");
584 } 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
585 } 665 }
586 else if (ch >= 0x20 && ch <= 0x7f) 666 while (cur < buf + SHORT_STRING_LEN);
587 APPEND_CH (*dec->cur++); 667
588 else if (ch >= 0x80) 668 STRLEN len = cur - buf;
669
670 if (sv)
589 { 671 {
590 STRLEN clen; 672 SvGROW (sv, SvCUR (sv) + len + 1);
591 UV uch = utf8n_to_uvuni (dec->cur, dec->end - dec->cur, &clen, UTF8_CHECK_ONLY); 673 memcpy (SvPVX (sv) + SvCUR (sv), buf, len);
592 if (clen == (STRLEN)-1) 674 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 } 675 }
604 else if (dec->cur == dec->end)
605 ERR ("unexpected end of string while parsing json string");
606 else 676 else
607 ERR ("invalid character encountered"); 677 sv = newSVpvn (buf, len);
608 } 678 }
679 while (*dec->cur != '"');
609 680
610 ++dec->cur; 681 ++dec->cur;
611 682
612 SvCUR_set (sv, cur - SvPVX (sv)); 683 if (sv)
613 684 {
614 SvPOK_only (sv); 685 SvPOK_only (sv);
615 *SvEND (sv) = 0; 686 *SvEND (sv) = 0;
616 687
617 if (utf8) 688 if (utf8)
618 SvUTF8_on (sv); 689 SvUTF8_on (sv);
619 690 }
620 if (dec->flags & F_SHRINK) 691 else
621 shrink (sv); 692 sv = newSVpvn ("", 0);
622 693
623 return sv; 694 return sv;
624 695
625fail: 696fail:
626 SvREFCNT_dec (sv);
627 return 0; 697 return 0;
628} 698}
629 699
630static SV * 700static SV *
631decode_num (dec_t *dec) 701decode_num (dec_t *dec)
… …
712static SV * 782static SV *
713decode_av (dec_t *dec) 783decode_av (dec_t *dec)
714{ 784{
715 AV *av = newAV (); 785 AV *av = newAV ();
716 786
717 WS; 787 decode_ws (dec);
718 if (*dec->cur == ']') 788 if (*dec->cur == ']')
719 ++dec->cur; 789 ++dec->cur;
720 else 790 else
721 for (;;) 791 for (;;)
722 { 792 {
… …
726 if (!value) 796 if (!value)
727 goto fail; 797 goto fail;
728 798
729 av_push (av, value); 799 av_push (av, value);
730 800
731 WS; 801 decode_ws (dec);
732 802
733 if (*dec->cur == ']') 803 if (*dec->cur == ']')
734 { 804 {
735 ++dec->cur; 805 ++dec->cur;
736 break; 806 break;
… …
752static SV * 822static SV *
753decode_hv (dec_t *dec) 823decode_hv (dec_t *dec)
754{ 824{
755 HV *hv = newHV (); 825 HV *hv = newHV ();
756 826
757 WS; 827 decode_ws (dec);
758 if (*dec->cur == '}') 828 if (*dec->cur == '}')
759 ++dec->cur; 829 ++dec->cur;
760 else 830 else
761 for (;;) 831 for (;;)
762 { 832 {
763 SV *key, *value; 833 SV *key, *value;
764 834
765 WS; EXPECT_CH ('"'); 835 decode_ws (dec); EXPECT_CH ('"');
766 836
767 key = decode_str (dec); 837 key = decode_str (dec);
768 if (!key) 838 if (!key)
769 goto fail; 839 goto fail;
770 840
771 WS; EXPECT_CH (':'); 841 decode_ws (dec); EXPECT_CH (':');
772 842
773 value = decode_sv (dec); 843 value = decode_sv (dec);
774 if (!value) 844 if (!value)
775 { 845 {
776 SvREFCNT_dec (key); 846 SvREFCNT_dec (key);
… …
778 } 848 }
779 849
780 //TODO: optimise 850 //TODO: optimise
781 hv_store_ent (hv, key, value, 0); 851 hv_store_ent (hv, key, value, 0);
782 852
783 WS; 853 decode_ws (dec);
784 854
785 if (*dec->cur == '}') 855 if (*dec->cur == '}')
786 { 856 {
787 ++dec->cur; 857 ++dec->cur;
788 break; 858 break;
… …
802} 872}
803 873
804static SV * 874static SV *
805decode_sv (dec_t *dec) 875decode_sv (dec_t *dec)
806{ 876{
807 WS; 877 decode_ws (dec);
808 switch (*dec->cur) 878 switch (*dec->cur)
809 { 879 {
810 case '"': ++dec->cur; return decode_str (dec); 880 case '"': ++dec->cur; return decode_str (dec);
811 case '[': ++dec->cur; return decode_av (dec); 881 case '[': ++dec->cur; return decode_av (dec);
812 case '{': ++dec->cur; return decode_hv (dec); 882 case '{': ++dec->cur; return decode_hv (dec);
… …
901 } 971 }
902 972
903 sv = sv_2mortal (sv); 973 sv = sv_2mortal (sv);
904 974
905 if (!(dec.flags & F_ALLOW_NONREF) && !SvROK (sv)) 975 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)"); 976 croak ("JSON text must be an object or array (but found number, string, true, false or null, use allow_nonref to allow this)");
907 977
908 return sv; 978 return sv;
909} 979}
980
981/////////////////////////////////////////////////////////////////////////////
982// XS interface functions
910 983
911MODULE = JSON::XS PACKAGE = JSON::XS 984MODULE = JSON::XS PACKAGE = JSON::XS
912 985
913BOOT: 986BOOT:
914{ 987{
… …
941 utf8 = F_UTF8 1014 utf8 = F_UTF8
942 indent = F_INDENT 1015 indent = F_INDENT
943 canonical = F_CANONICAL 1016 canonical = F_CANONICAL
944 space_before = F_SPACE_BEFORE 1017 space_before = F_SPACE_BEFORE
945 space_after = F_SPACE_AFTER 1018 space_after = F_SPACE_AFTER
946 json_rpc = F_JSON_RPC
947 pretty = F_PRETTY 1019 pretty = F_PRETTY
948 allow_nonref = F_ALLOW_NONREF 1020 allow_nonref = F_ALLOW_NONREF
949 shrink = F_SHRINK 1021 shrink = F_SHRINK
950 CODE: 1022 CODE:
951{ 1023{
… …
969 XPUSHs (decode_json (jsonstr, *SvJSON (self))); 1041 XPUSHs (decode_json (jsonstr, *SvJSON (self)));
970 1042
971PROTOTYPES: ENABLE 1043PROTOTYPES: ENABLE
972 1044
973void to_json (SV *scalar) 1045void to_json (SV *scalar)
1046 ALIAS:
1047 objToJson = 0
974 PPCODE: 1048 PPCODE:
975 XPUSHs (encode_json (scalar, F_UTF8)); 1049 XPUSHs (encode_json (scalar, F_UTF8));
976 1050
977void from_json (SV *jsonstr) 1051void from_json (SV *jsonstr)
1052 ALIAS:
1053 jsonToObj = 0
978 PPCODE: 1054 PPCODE:
979 XPUSHs (decode_json (jsonstr, F_UTF8)); 1055 XPUSHs (decode_json (jsonstr, F_UTF8));
980 1056

Diff Legend

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