… | |
… | |
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 256 // 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 | |
26 | static HV *json_stash; |
30 | static HV *json_stash; // JSON::XS:: |
|
|
31 | |
|
|
32 | ///////////////////////////////////////////////////////////////////////////// |
|
|
33 | // utility functions |
|
|
34 | |
|
|
35 | static UV * |
|
|
36 | SvJSON (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 | |
|
|
44 | static void |
|
|
45 | shrink (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 |
27 | |
60 | |
28 | // structure used for encoding JSON |
61 | // structure used for encoding JSON |
29 | typedef struct |
62 | typedef struct |
30 | { |
63 | { |
31 | char *cur; |
64 | char *cur; // SvPVX (sv) + current output position |
32 | STRLEN len; // SvLEN (sv) |
|
|
33 | char *end; // SvEND (sv) |
65 | char *end; // SvEND (sv) |
34 | SV *sv; |
66 | SV *sv; // result scalar |
35 | UV flags; |
67 | UV flags; // F_* |
36 | int max_recurse; |
68 | int indent; // indentation level |
37 | int indent; |
69 | int max_depth; // max. recursion level |
38 | } enc_t; |
70 | } enc_t; |
39 | |
|
|
40 | // structure used for decoding JSON |
|
|
41 | typedef struct |
|
|
42 | { |
|
|
43 | char *cur; |
|
|
44 | char *end; |
|
|
45 | const char *err; |
|
|
46 | UV flags; |
|
|
47 | } dec_t; |
|
|
48 | |
|
|
49 | static UV * |
|
|
50 | SvJSON (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 | |
|
|
58 | static void |
|
|
59 | shrink (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 | |
71 | |
69 | static void |
72 | static void |
70 | need (enc_t *enc, STRLEN len) |
73 | need (enc_t *enc, STRLEN len) |
71 | { |
74 | { |
72 | if (enc->cur + len >= enc->end) |
75 | if (enc->cur + len >= enc->end) |
… | |
… | |
178 | } |
181 | } |
179 | while (--clen); |
182 | while (--clen); |
180 | } |
183 | } |
181 | else |
184 | else |
182 | { |
185 | { |
183 | need (enc, len += 10); // never more than 11 bytes needed |
186 | need (enc, len += UTF8_MAX_LEN - 1); // never more than 11 bytes needed |
184 | enc->cur = uvuni_to_utf8_flags (enc->cur, uch, 0); |
187 | enc->cur = uvuni_to_utf8_flags (enc->cur, uch, 0); |
185 | ++str; |
188 | ++str; |
186 | } |
189 | } |
187 | } |
190 | } |
188 | } |
191 | } |
… | |
… | |
190 | |
193 | |
191 | --len; |
194 | --len; |
192 | } |
195 | } |
193 | } |
196 | } |
194 | |
197 | |
195 | #define INDENT SB \ |
198 | static void |
|
|
199 | encode_indent (enc_t *enc) |
|
|
200 | { |
196 | if (enc->flags & F_INDENT) \ |
201 | if (enc->flags & F_INDENT) |
197 | { \ |
202 | { |
198 | int i_; \ |
203 | 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 | |
204 | |
205 | #define SPACE SB need (enc, 1); encode_ch (enc, ' '); SE |
205 | need (enc, spaces); |
206 | #define NL SB if (enc->flags & F_INDENT) { need (enc, 1); encode_ch (enc, '\n'); } SE |
206 | memset (enc->cur, ' ', spaces); |
207 | #define COMMA SB \ |
207 | enc->cur += spaces; |
|
|
208 | } |
|
|
209 | } |
|
|
210 | |
|
|
211 | static void |
|
|
212 | encode_space (enc_t *enc) |
|
|
213 | { |
|
|
214 | need (enc, 1); |
208 | encode_ch (enc, ','); \ |
215 | encode_ch (enc, ' '); |
|
|
216 | } |
|
|
217 | |
|
|
218 | static void |
|
|
219 | encode_nl (enc_t *enc) |
|
|
220 | { |
209 | if (enc->flags & F_INDENT) \ |
221 | if (enc->flags & F_INDENT) |
210 | NL; \ |
222 | { |
|
|
223 | need (enc, 1); |
|
|
224 | encode_ch (enc, '\n'); |
|
|
225 | } |
|
|
226 | } |
|
|
227 | |
|
|
228 | static void |
|
|
229 | encode_comma (enc_t *enc) |
|
|
230 | { |
|
|
231 | encode_ch (enc, ','); |
|
|
232 | |
|
|
233 | if (enc->flags & F_INDENT) |
|
|
234 | encode_nl (enc); |
211 | else if (enc->flags & F_SPACE_AFTER) \ |
235 | else if (enc->flags & F_SPACE_AFTER) |
212 | SPACE; \ |
236 | encode_space (enc); |
213 | SE |
237 | } |
214 | |
238 | |
215 | static void encode_sv (enc_t *enc, SV *sv); |
239 | static void encode_sv (enc_t *enc, SV *sv); |
216 | |
240 | |
217 | static void |
241 | static void |
218 | encode_av (enc_t *enc, AV *av) |
242 | encode_av (enc_t *enc, AV *av) |
219 | { |
243 | { |
220 | int i, len = av_len (av); |
244 | int i, len = av_len (av); |
221 | |
245 | |
222 | encode_ch (enc, '['); NL; |
246 | encode_ch (enc, '['); encode_nl (enc); |
223 | ++enc->indent; |
247 | ++enc->indent; |
224 | |
248 | |
225 | for (i = 0; i <= len; ++i) |
249 | for (i = 0; i <= len; ++i) |
226 | { |
250 | { |
227 | INDENT; |
251 | encode_indent (enc); |
228 | encode_sv (enc, *av_fetch (av, i, 0)); |
252 | encode_sv (enc, *av_fetch (av, i, 0)); |
229 | |
253 | |
230 | if (i < len) |
254 | if (i < len) |
231 | COMMA; |
255 | encode_comma (enc); |
232 | } |
256 | } |
233 | |
257 | |
234 | NL; |
258 | encode_nl (enc); |
235 | |
259 | |
236 | --enc->indent; |
260 | --enc->indent; |
237 | INDENT; encode_ch (enc, ']'); |
261 | encode_indent (enc); encode_ch (enc, ']'); |
238 | } |
262 | } |
239 | |
263 | |
240 | static void |
264 | static void |
241 | encode_he (enc_t *enc, HE *he) |
265 | encode_he (enc_t *enc, HE *he) |
242 | { |
266 | { |
… | |
… | |
256 | else |
280 | else |
257 | encode_str (enc, HeKEY (he), HeKLEN (he), HeKUTF8 (he)); |
281 | encode_str (enc, HeKEY (he), HeKLEN (he), HeKUTF8 (he)); |
258 | |
282 | |
259 | encode_ch (enc, '"'); |
283 | encode_ch (enc, '"'); |
260 | |
284 | |
261 | if (enc->flags & F_SPACE_BEFORE) SPACE; |
285 | if (enc->flags & F_SPACE_BEFORE) encode_space (enc); |
262 | encode_ch (enc, ':'); |
286 | encode_ch (enc, ':'); |
263 | if (enc->flags & F_SPACE_AFTER ) SPACE; |
287 | if (enc->flags & F_SPACE_AFTER ) encode_space (enc); |
264 | encode_sv (enc, HeVAL (he)); |
288 | encode_sv (enc, HeVAL (he)); |
265 | } |
289 | } |
266 | |
290 | |
267 | // compare hash entries, used when all keys are bytestrings |
291 | // compare hash entries, used when all keys are bytestrings |
268 | static int |
292 | static int |
… | |
… | |
292 | static void |
316 | static void |
293 | encode_hv (enc_t *enc, HV *hv) |
317 | encode_hv (enc_t *enc, HV *hv) |
294 | { |
318 | { |
295 | int count, i; |
319 | int count, i; |
296 | |
320 | |
297 | encode_ch (enc, '{'); NL; ++enc->indent; |
321 | encode_ch (enc, '{'); encode_nl (enc); ++enc->indent; |
298 | |
322 | |
299 | if ((count = hv_iterinit (hv))) |
323 | if ((count = hv_iterinit (hv))) |
300 | { |
324 | { |
301 | // for canonical output we have to sort by keys first |
325 | // for canonical output we have to sort by keys first |
302 | // actually, this is mostly due to the stupid so-called |
326 | // actually, this is mostly due to the stupid so-called |
303 | // security workaround added somewhere in 5.8.x. |
327 | // security workaround added somewhere in 5.8.x. |
304 | // that randomises hash orderings |
328 | // that randomises hash orderings |
305 | if (enc->flags & F_CANONICAL) |
329 | if (enc->flags & F_CANONICAL) |
306 | { |
330 | { |
307 | HE *he, *hes [count]; |
331 | HE *he, *hes [count]; // if your compiler dies here, you need to enable C99 mode |
308 | int fast = 1; |
332 | int fast = 1; |
309 | |
333 | |
310 | i = 0; |
334 | i = 0; |
311 | while ((he = hv_iternext (hv))) |
335 | while ((he = hv_iternext (hv))) |
312 | { |
336 | { |
… | |
… | |
337 | LEAVE; |
361 | LEAVE; |
338 | } |
362 | } |
339 | |
363 | |
340 | for (i = 0; i < count; ++i) |
364 | for (i = 0; i < count; ++i) |
341 | { |
365 | { |
342 | INDENT; |
366 | encode_indent (enc); |
343 | encode_he (enc, hes [i]); |
367 | encode_he (enc, hes [i]); |
344 | |
368 | |
345 | if (i < count - 1) |
369 | if (i < count - 1) |
346 | COMMA; |
370 | encode_comma (enc); |
347 | } |
371 | } |
348 | |
372 | |
349 | NL; |
373 | encode_nl (enc); |
350 | } |
374 | } |
351 | else |
375 | else |
352 | { |
376 | { |
353 | SV *sv; |
377 | SV *sv; |
354 | HE *he = hv_iternext (hv); |
378 | HE *he = hv_iternext (hv); |
355 | |
379 | |
356 | for (;;) |
380 | for (;;) |
357 | { |
381 | { |
358 | INDENT; |
382 | encode_indent (enc); |
359 | encode_he (enc, he); |
383 | encode_he (enc, he); |
360 | |
384 | |
361 | if (!(he = hv_iternext (hv))) |
385 | if (!(he = hv_iternext (hv))) |
362 | break; |
386 | break; |
363 | |
387 | |
364 | COMMA; |
388 | encode_comma (enc); |
365 | } |
389 | } |
366 | |
390 | |
367 | NL; |
391 | encode_nl (enc); |
368 | } |
392 | } |
369 | } |
393 | } |
370 | |
394 | |
371 | --enc->indent; INDENT; encode_ch (enc, '}'); |
395 | --enc->indent; encode_indent (enc); encode_ch (enc, '}'); |
372 | } |
396 | } |
373 | |
397 | |
374 | static void |
398 | static void |
375 | encode_sv (enc_t *enc, SV *sv) |
399 | encode_sv (enc_t *enc, SV *sv) |
376 | { |
400 | { |
… | |
… | |
400 | } |
424 | } |
401 | else if (SvROK (sv)) |
425 | else if (SvROK (sv)) |
402 | { |
426 | { |
403 | SV *rv = SvRV (sv); |
427 | SV *rv = SvRV (sv); |
404 | |
428 | |
405 | if (!--enc->max_recurse) |
429 | if (enc->indent >= enc->max_depth) |
406 | croak ("data structure too deep (hit recursion limit)"); |
430 | croak ("data structure too deep (hit recursion limit)"); |
407 | |
431 | |
408 | switch (SvTYPE (rv)) |
432 | switch (SvTYPE (rv)) |
409 | { |
433 | { |
410 | case SVt_PVAV: encode_av (enc, (AV *)rv); break; |
434 | case SVt_PVAV: encode_av (enc, (AV *)rv); break; |
… | |
… | |
427 | { |
451 | { |
428 | if (!(flags & F_ALLOW_NONREF) && !SvROK (scalar)) |
452 | if (!(flags & F_ALLOW_NONREF) && !SvROK (scalar)) |
429 | croak ("hash- or arrayref expected (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)"); |
430 | |
454 | |
431 | enc_t enc; |
455 | enc_t enc; |
432 | enc.flags = flags; |
456 | enc.flags = flags; |
433 | enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); |
457 | enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); |
434 | enc.cur = SvPVX (enc.sv); |
458 | enc.cur = SvPVX (enc.sv); |
435 | enc.end = SvEND (enc.sv); |
459 | enc.end = SvEND (enc.sv); |
436 | enc.max_recurse = 0; |
|
|
437 | enc.indent = 0; |
460 | enc.indent = 0; |
|
|
461 | enc.max_depth = 0x7fffffffUL; |
438 | |
462 | |
439 | SvPOK_only (enc.sv); |
463 | SvPOK_only (enc.sv); |
440 | encode_sv (&enc, scalar); |
464 | encode_sv (&enc, scalar); |
441 | |
465 | |
442 | if (!(flags & (F_ASCII | F_UTF8))) |
466 | if (!(flags & (F_ASCII | F_UTF8))) |
… | |
… | |
449 | |
473 | |
450 | return enc.sv; |
474 | return enc.sv; |
451 | } |
475 | } |
452 | |
476 | |
453 | ///////////////////////////////////////////////////////////////////////////// |
477 | ///////////////////////////////////////////////////////////////////////////// |
|
|
478 | // decoder |
454 | |
479 | |
455 | #define WS \ |
480 | // structure used for decoding JSON |
|
|
481 | typedef 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 | |
|
|
489 | static void |
|
|
490 | decode_ws (dec_t *dec) |
|
|
491 | { |
456 | for (;;) \ |
492 | for (;;) |
457 | { \ |
493 | { |
458 | char ch = *dec->cur; \ |
494 | char ch = *dec->cur; |
|
|
495 | |
459 | if (ch > 0x20 \ |
496 | if (ch > 0x20 |
460 | || (ch != 0x20 && ch != 0x0a && ch != 0x0d && ch != 0x09)) \ |
497 | || (ch != 0x20 && ch != 0x0a && ch != 0x0d && ch != 0x09)) |
461 | break; \ |
498 | break; |
|
|
499 | |
462 | ++dec->cur; \ |
500 | ++dec->cur; |
463 | } |
501 | } |
|
|
502 | } |
464 | |
503 | |
465 | #define ERR(reason) SB dec->err = reason; goto fail; SE |
504 | #define ERR(reason) SB dec->err = reason; goto fail; SE |
466 | #define EXPECT_CH(ch) SB \ |
505 | #define EXPECT_CH(ch) SB \ |
467 | if (*dec->cur != ch) \ |
506 | if (*dec->cur != ch) \ |
468 | ERR (# ch " expected"); \ |
507 | ERR (# ch " expected"); \ |
… | |
… | |
475 | |
514 | |
476 | static UV |
515 | static UV |
477 | decode_4hex (dec_t *dec) |
516 | decode_4hex (dec_t *dec) |
478 | { |
517 | { |
479 | signed char d1, d2, d3, d4; |
518 | signed char d1, d2, d3, d4; |
|
|
519 | unsigned char *cur = (unsigned char *)dec->cur; |
480 | |
520 | |
481 | d1 = decode_hexdigit [((unsigned char *)dec->cur) [0]]; |
|
|
482 | if (d1 < 0) ERR ("four hexadecimal digits expected"); |
521 | 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"); |
522 | 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"); |
523 | 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"); |
524 | d4 = decode_hexdigit [cur [3]]; if (d4 < 0) ERR ("four hexadecimal digits expected"); |
489 | |
525 | |
490 | dec->cur += 4; |
526 | dec->cur += 4; |
491 | |
527 | |
492 | return ((UV)d1) << 12 |
528 | return ((UV)d1) << 12 |
493 | | ((UV)d2) << 8 |
529 | | ((UV)d2) << 8 |
… | |
… | |
496 | |
532 | |
497 | fail: |
533 | fail: |
498 | return (UV)-1; |
534 | return (UV)-1; |
499 | } |
535 | } |
500 | |
536 | |
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 | |
|
|
516 | static SV * |
537 | static SV * |
517 | decode_str (dec_t *dec) |
538 | decode_str (dec_t *dec) |
518 | { |
539 | { |
519 | SV *sv = NEWSV (0,2); |
540 | SV *sv = 0; |
520 | int utf8 = 0; |
541 | int utf8 = 0; |
521 | char *cur = SvPVX (sv); |
|
|
522 | char *end = SvEND (sv); |
|
|
523 | |
542 | |
524 | for (;;) |
543 | do |
525 | { |
544 | { |
526 | unsigned char ch = *(unsigned char *)dec->cur; |
545 | char buf [SHORT_STRING_LEN + UTF8_MAX_LEN]; |
|
|
546 | char *cur = buf; |
527 | |
547 | |
528 | if (ch == '"') |
548 | do |
529 | break; |
|
|
530 | else if (ch == '\\') |
|
|
531 | { |
549 | { |
532 | switch (*++dec->cur) |
550 | unsigned char ch = *(unsigned char *)dec->cur++; |
|
|
551 | |
|
|
552 | if (ch == '"') |
533 | { |
553 | { |
534 | case '\\': |
554 | --dec->cur; |
535 | case '/': |
555 | break; |
536 | case '"': APPEND_CH (*dec->cur++); break; |
556 | } |
537 | |
557 | else if (ch == '\\') |
538 | case 'b': APPEND_CH ('\010'); ++dec->cur; break; |
558 | { |
539 | case 't': APPEND_CH ('\011'); ++dec->cur; break; |
559 | 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 | { |
560 | { |
546 | UV lo, hi; |
561 | case '\\': |
547 | ++dec->cur; |
562 | case '/': |
|
|
563 | case '"': *cur++ = *dec->cur++; break; |
548 | |
564 | |
549 | hi = decode_4hex (dec); |
565 | case 'b': ++dec->cur; *cur++ = '\010'; break; |
550 | if (hi == (UV)-1) |
566 | case 't': ++dec->cur; *cur++ = '\011'; break; |
551 | 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; |
552 | |
570 | |
553 | // possibly a surrogate pair |
571 | case 'u': |
554 | if (hi >= 0xd800 && hi < 0xdc00) |
|
|
555 | { |
572 | { |
556 | if (dec->cur [0] != '\\' || dec->cur [1] != 'u') |
573 | UV lo, hi; |
557 | ERR ("missing low surrogate character in surrogate pair"); |
|
|
558 | |
|
|
559 | dec->cur += 2; |
574 | ++dec->cur; |
560 | |
575 | |
561 | lo = decode_4hex (dec); |
576 | hi = decode_4hex (dec); |
562 | if (lo == (UV)-1) |
577 | if (hi == (UV)-1) |
563 | goto fail; |
578 | goto fail; |
564 | |
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 | |
565 | if (lo < 0xdc00 || lo >= 0xe000) |
593 | if (lo < 0xdc00 || lo >= 0xe000) |
566 | ERR ("surrogate pair expected"); |
594 | ERR ("surrogate pair expected"); |
567 | |
595 | |
568 | 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; |
569 | } |
609 | } |
570 | else if (hi >= 0xdc00 && hi < 0xe000) |
|
|
571 | ERR ("missing high surrogate character in surrogate pair"); |
|
|
572 | |
|
|
573 | if (hi >= 0x80) |
|
|
574 | { |
610 | break; |
575 | utf8 = 1; |
|
|
576 | |
611 | |
577 | APPEND_GROW (4); // at most 4 bytes for 21 bits |
|
|
578 | cur = (char *)uvuni_to_utf8_flags (cur, hi, 0); |
|
|
579 | } |
|
|
580 | else |
612 | default: |
581 | APPEND_CH (hi); |
613 | --dec->cur; |
|
|
614 | ERR ("illegal backslash escape sequence in string"); |
582 | } |
615 | } |
583 | break; |
|
|
584 | |
|
|
585 | default: |
|
|
586 | --dec->cur; |
|
|
587 | ERR ("illegal backslash escape sequence in string"); |
|
|
588 | } |
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 | |
589 | } |
641 | } |
590 | else if (ch >= 0x20 && ch <= 0x7f) |
642 | while (cur < buf + SHORT_STRING_LEN); |
591 | APPEND_CH (*dec->cur++); |
643 | |
592 | else if (ch >= 0x80) |
644 | STRLEN len = cur - buf; |
|
|
645 | |
|
|
646 | if (sv) |
593 | { |
647 | { |
594 | STRLEN clen; |
648 | SvGROW (sv, SvCUR (sv) + len + 1); |
595 | UV uch = utf8n_to_uvuni (dec->cur, dec->end - dec->cur, &clen, UTF8_CHECK_ONLY); |
649 | memcpy (SvPVX (sv) + SvCUR (sv), buf, len); |
596 | if (clen == (STRLEN)-1) |
650 | 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 | } |
651 | } |
608 | else if (dec->cur == dec->end) |
|
|
609 | ERR ("unexpected end of string while parsing json string"); |
|
|
610 | else |
652 | else |
611 | ERR ("invalid character encountered"); |
653 | sv = newSVpvn (buf, len); |
612 | } |
654 | } |
|
|
655 | while (*dec->cur != '"'); |
613 | |
656 | |
614 | ++dec->cur; |
657 | ++dec->cur; |
615 | |
658 | |
616 | SvCUR_set (sv, cur - SvPVX (sv)); |
659 | if (sv) |
617 | |
660 | { |
618 | SvPOK_only (sv); |
661 | SvPOK_only (sv); |
619 | *SvEND (sv) = 0; |
662 | *SvEND (sv) = 0; |
620 | |
663 | |
621 | if (utf8) |
664 | if (utf8) |
622 | SvUTF8_on (sv); |
665 | SvUTF8_on (sv); |
623 | |
666 | } |
624 | if (dec->flags & F_SHRINK) |
667 | else |
625 | shrink (sv); |
668 | sv = newSVpvn ("", 0); |
626 | |
669 | |
627 | return sv; |
670 | return sv; |
628 | |
671 | |
629 | fail: |
672 | fail: |
630 | SvREFCNT_dec (sv); |
|
|
631 | return 0; |
673 | return 0; |
632 | } |
674 | } |
633 | |
675 | |
634 | static SV * |
676 | static SV * |
635 | decode_num (dec_t *dec) |
677 | decode_num (dec_t *dec) |
… | |
… | |
716 | static SV * |
758 | static SV * |
717 | decode_av (dec_t *dec) |
759 | decode_av (dec_t *dec) |
718 | { |
760 | { |
719 | AV *av = newAV (); |
761 | AV *av = newAV (); |
720 | |
762 | |
721 | WS; |
763 | decode_ws (dec); |
722 | if (*dec->cur == ']') |
764 | if (*dec->cur == ']') |
723 | ++dec->cur; |
765 | ++dec->cur; |
724 | else |
766 | else |
725 | for (;;) |
767 | for (;;) |
726 | { |
768 | { |
… | |
… | |
730 | if (!value) |
772 | if (!value) |
731 | goto fail; |
773 | goto fail; |
732 | |
774 | |
733 | av_push (av, value); |
775 | av_push (av, value); |
734 | |
776 | |
735 | WS; |
777 | decode_ws (dec); |
736 | |
778 | |
737 | if (*dec->cur == ']') |
779 | if (*dec->cur == ']') |
738 | { |
780 | { |
739 | ++dec->cur; |
781 | ++dec->cur; |
740 | break; |
782 | break; |
… | |
… | |
756 | static SV * |
798 | static SV * |
757 | decode_hv (dec_t *dec) |
799 | decode_hv (dec_t *dec) |
758 | { |
800 | { |
759 | HV *hv = newHV (); |
801 | HV *hv = newHV (); |
760 | |
802 | |
761 | WS; |
803 | decode_ws (dec); |
762 | if (*dec->cur == '}') |
804 | if (*dec->cur == '}') |
763 | ++dec->cur; |
805 | ++dec->cur; |
764 | else |
806 | else |
765 | for (;;) |
807 | for (;;) |
766 | { |
808 | { |
767 | SV *key, *value; |
809 | SV *key, *value; |
768 | |
810 | |
769 | WS; EXPECT_CH ('"'); |
811 | decode_ws (dec); EXPECT_CH ('"'); |
770 | |
812 | |
771 | key = decode_str (dec); |
813 | key = decode_str (dec); |
772 | if (!key) |
814 | if (!key) |
773 | goto fail; |
815 | goto fail; |
774 | |
816 | |
775 | WS; EXPECT_CH (':'); |
817 | decode_ws (dec); EXPECT_CH (':'); |
776 | |
818 | |
777 | value = decode_sv (dec); |
819 | value = decode_sv (dec); |
778 | if (!value) |
820 | if (!value) |
779 | { |
821 | { |
780 | SvREFCNT_dec (key); |
822 | SvREFCNT_dec (key); |
… | |
… | |
782 | } |
824 | } |
783 | |
825 | |
784 | //TODO: optimise |
826 | //TODO: optimise |
785 | hv_store_ent (hv, key, value, 0); |
827 | hv_store_ent (hv, key, value, 0); |
786 | |
828 | |
787 | WS; |
829 | decode_ws (dec); |
788 | |
830 | |
789 | if (*dec->cur == '}') |
831 | if (*dec->cur == '}') |
790 | { |
832 | { |
791 | ++dec->cur; |
833 | ++dec->cur; |
792 | break; |
834 | break; |
… | |
… | |
806 | } |
848 | } |
807 | |
849 | |
808 | static SV * |
850 | static SV * |
809 | decode_sv (dec_t *dec) |
851 | decode_sv (dec_t *dec) |
810 | { |
852 | { |
811 | WS; |
853 | decode_ws (dec); |
812 | switch (*dec->cur) |
854 | switch (*dec->cur) |
813 | { |
855 | { |
814 | case '"': ++dec->cur; return decode_str (dec); |
856 | case '"': ++dec->cur; return decode_str (dec); |
815 | case '[': ++dec->cur; return decode_av (dec); |
857 | case '[': ++dec->cur; return decode_av (dec); |
816 | case '{': ++dec->cur; return decode_hv (dec); |
858 | case '{': ++dec->cur; return decode_hv (dec); |
… | |
… | |
909 | if (!(dec.flags & F_ALLOW_NONREF) && !SvROK (sv)) |
951 | 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)"); |
952 | croak ("JSON text must be an object or array (but found number, string, true, false or null, use allow_nonref to allow this)"); |
911 | |
953 | |
912 | return sv; |
954 | return sv; |
913 | } |
955 | } |
|
|
956 | |
|
|
957 | ///////////////////////////////////////////////////////////////////////////// |
|
|
958 | // XS interface functions |
914 | |
959 | |
915 | MODULE = JSON::XS PACKAGE = JSON::XS |
960 | MODULE = JSON::XS PACKAGE = JSON::XS |
916 | |
961 | |
917 | BOOT: |
962 | BOOT: |
918 | { |
963 | { |