… | |
… | |
6 | #include <string.h> |
6 | #include <string.h> |
7 | #include <stdlib.h> |
7 | #include <stdlib.h> |
8 | #include <stdio.h> |
8 | #include <stdio.h> |
9 | #include <limits.h> |
9 | #include <limits.h> |
10 | #include <float.h> |
10 | #include <float.h> |
|
|
11 | #include <inttypes.h> |
11 | |
12 | |
|
|
13 | #define ECB_NO_THREADS 1 |
12 | #include "ecb.h" |
14 | #include "ecb.h" |
13 | |
15 | |
14 | // compatibility with perl <5.18 |
16 | // compatibility with perl <5.18 |
15 | #ifndef HvNAMELEN_get |
17 | #ifndef HvNAMELEN_get |
16 | # define HvNAMELEN_get(hv) strlen (HvNAME (hv)) |
18 | # define HvNAMELEN_get(hv) strlen (HvNAME (hv)) |
… | |
… | |
19 | # define HvNAMELEN(hv) HvNAMELEN_get (hv) |
21 | # define HvNAMELEN(hv) HvNAMELEN_get (hv) |
20 | #endif |
22 | #endif |
21 | #ifndef HvNAMEUTF8 |
23 | #ifndef HvNAMEUTF8 |
22 | # define HvNAMEUTF8(hv) 0 |
24 | # define HvNAMEUTF8(hv) 0 |
23 | #endif |
25 | #endif |
|
|
26 | #ifndef SvREFCNT_inc_NN |
|
|
27 | # define SvREFCNT_inc_NN(sv) SvREFCNT_inc (sv) |
|
|
28 | #endif |
24 | #ifndef SvREFCNT_dec_NN |
29 | #ifndef SvREFCNT_dec_NN |
25 | # define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv) |
30 | # define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv) |
26 | #endif |
31 | #endif |
27 | |
32 | |
28 | // known major and minor types |
33 | // known major and minor types |
… | |
… | |
94 | CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8 |
99 | CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8 |
95 | |
100 | |
96 | CBOR_TAG_MAGIC = 55799, // self-describe cbor |
101 | CBOR_TAG_MAGIC = 55799, // self-describe cbor |
97 | }; |
102 | }; |
98 | |
103 | |
|
|
104 | // known forced types, also hardcoded in CBOR.pm |
|
|
105 | enum |
|
|
106 | { |
|
|
107 | AS_CBOR = 0, |
|
|
108 | AS_INT = 1, |
|
|
109 | AS_BYTES = 2, |
|
|
110 | AS_TEXT = 3, |
|
|
111 | AS_FLOAT16 = 4, |
|
|
112 | AS_FLOAT32 = 5, |
|
|
113 | AS_FLOAT64 = 6, |
|
|
114 | AS_MAP = 7, |
|
|
115 | // possibly future enhancements: (generic) float, (generic) string |
|
|
116 | }; |
|
|
117 | |
99 | #define F_SHRINK 0x00000001UL |
118 | #define F_SHRINK 0x00000001UL |
100 | #define F_ALLOW_UNKNOWN 0x00000002UL |
119 | #define F_ALLOW_UNKNOWN 0x00000002UL |
101 | #define F_ALLOW_SHARING 0x00000004UL |
120 | #define F_ALLOW_SHARING 0x00000004UL |
102 | #define F_ALLOW_CYCLES 0x00000008UL |
121 | #define F_ALLOW_CYCLES 0x00000008UL |
|
|
122 | #define F_FORBID_OBJECTS 0x00000010UL |
103 | #define F_PACK_STRINGS 0x00000010UL |
123 | #define F_PACK_STRINGS 0x00000020UL |
|
|
124 | #define F_TEXT_KEYS 0x00000040UL |
|
|
125 | #define F_TEXT_STRINGS 0x00000080UL |
104 | #define F_VALIDATE_UTF8 0x00000020UL |
126 | #define F_VALIDATE_UTF8 0x00000100UL |
105 | |
127 | |
106 | #define INIT_SIZE 32 // initial scalar size to be allocated |
128 | #define INIT_SIZE 32 // initial scalar size to be allocated |
107 | |
129 | |
108 | #define SB do { |
130 | #define SB do { |
109 | #define SE } while (0) |
131 | #define SE } while (0) |
… | |
… | |
182 | |
204 | |
183 | // minimum length of a string to be registered for stringref |
205 | // minimum length of a string to be registered for stringref |
184 | ecb_inline int |
206 | ecb_inline int |
185 | minimum_string_length (UV idx) |
207 | minimum_string_length (UV idx) |
186 | { |
208 | { |
187 | return idx > 23 |
209 | return idx <= 23 ? 3 |
188 | ? idx > 0xffU |
210 | : idx <= 0xffU ? 4 |
189 | ? idx > 0xffffU |
211 | : idx <= 0xffffU ? 5 |
190 | ? idx > 0xffffffffU |
212 | : idx <= 0xffffffffU ? 7 |
191 | ? 11 |
213 | : 11; |
192 | : 7 |
|
|
193 | : 5 |
|
|
194 | : 4 |
|
|
195 | : 3; |
|
|
196 | } |
214 | } |
197 | |
215 | |
198 | ///////////////////////////////////////////////////////////////////////////// |
216 | ///////////////////////////////////////////////////////////////////////////// |
199 | // encoder |
217 | // encoder |
200 | |
218 | |
… | |
… | |
213 | } enc_t; |
231 | } enc_t; |
214 | |
232 | |
215 | ecb_inline void |
233 | ecb_inline void |
216 | need (enc_t *enc, STRLEN len) |
234 | need (enc_t *enc, STRLEN len) |
217 | { |
235 | { |
218 | if (ecb_expect_false (enc->cur + len >= enc->end)) |
236 | if (ecb_expect_false ((uintptr_t)(enc->end - enc->cur) < len)) |
219 | { |
237 | { |
220 | STRLEN cur = enc->cur - (char *)SvPVX (enc->sv); |
238 | STRLEN cur = enc->cur - (char *)SvPVX (enc->sv); |
221 | SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1); |
239 | SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1); |
222 | enc->cur = SvPVX (enc->sv) + cur; |
240 | enc->cur = SvPVX (enc->sv) + cur; |
223 | enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1; |
241 | enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1; |
224 | } |
242 | } |
225 | } |
243 | } |
226 | |
244 | |
|
|
245 | static void encode_sv (enc_t *enc, SV *sv); |
|
|
246 | |
227 | ecb_inline void |
247 | ecb_inline void |
228 | encode_ch (enc_t *enc, char ch) |
248 | encode_ch (enc_t *enc, char ch) |
229 | { |
249 | { |
230 | need (enc, 1); |
250 | need (enc, 1); |
231 | *enc->cur++ = ch; |
251 | *enc->cur++ = ch; |
232 | } |
252 | } |
233 | |
253 | |
|
|
254 | // used for tags, intregers, element counts and so on |
234 | static void |
255 | static void |
235 | encode_uint (enc_t *enc, int major, UV len) |
256 | encode_uint (enc_t *enc, int major, UV len) |
236 | { |
257 | { |
237 | need (enc, 9); |
258 | need (enc, 9); |
238 | |
259 | |
… | |
… | |
269 | *enc->cur++ = len >> 8; |
290 | *enc->cur++ = len >> 8; |
270 | *enc->cur++ = len; |
291 | *enc->cur++ = len; |
271 | } |
292 | } |
272 | } |
293 | } |
273 | |
294 | |
|
|
295 | // encodes a perl value into a CBOR integer |
|
|
296 | ecb_inline void |
|
|
297 | encode_int (enc_t *enc, SV *sv) |
|
|
298 | { |
|
|
299 | if (SvIsUV (sv)) |
|
|
300 | encode_uint (enc, MAJOR_POS_INT, SvUVX (sv)); |
|
|
301 | else if (SvIVX (sv) >= 0) |
|
|
302 | encode_uint (enc, MAJOR_POS_INT, SvIVX (sv)); |
|
|
303 | else |
|
|
304 | encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1)); |
|
|
305 | } |
|
|
306 | |
274 | ecb_inline void |
307 | ecb_inline void |
275 | encode_tag (enc_t *enc, UV tag) |
308 | encode_tag (enc_t *enc, UV tag) |
276 | { |
309 | { |
277 | encode_uint (enc, MAJOR_TAG, tag); |
310 | encode_uint (enc, MAJOR_TAG, tag); |
278 | } |
311 | } |
279 | |
312 | |
|
|
313 | // exceptional (hopefully) slow path for byte strings that need to be utf8-encoded |
|
|
314 | ecb_noinline static void |
|
|
315 | encode_str_utf8 (enc_t *enc, int utf8, char *str, STRLEN len) |
|
|
316 | { |
|
|
317 | STRLEN ulen = len; |
|
|
318 | U8 *p, *pend = (U8 *)str + len; |
|
|
319 | |
|
|
320 | for (p = (U8 *)str; p < pend; ++p) |
|
|
321 | ulen += *p >> 7; // count set high bits |
|
|
322 | |
|
|
323 | encode_uint (enc, MAJOR_TEXT, ulen); |
|
|
324 | |
|
|
325 | need (enc, ulen); |
|
|
326 | for (p = (U8 *)str; p < pend; ++p) |
|
|
327 | if (*p < 0x80) |
|
|
328 | *enc->cur++ = *p; |
|
|
329 | else |
|
|
330 | { |
|
|
331 | *enc->cur++ = 0xc0 + (*p >> 6); |
|
|
332 | *enc->cur++ = 0x80 + (*p & 63); |
|
|
333 | } |
|
|
334 | } |
|
|
335 | |
280 | ecb_inline void |
336 | ecb_inline void |
281 | encode_str (enc_t *enc, int utf8, char *str, STRLEN len) |
337 | encode_str (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len) |
282 | { |
338 | { |
|
|
339 | if (ecb_expect_false (upgrade_utf8)) |
|
|
340 | if (!utf8) |
|
|
341 | { |
|
|
342 | encode_str_utf8 (enc, utf8, str, len); |
|
|
343 | return; |
|
|
344 | } |
|
|
345 | |
283 | encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len); |
346 | encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len); |
284 | need (enc, len); |
347 | need (enc, len); |
285 | memcpy (enc->cur, str, len); |
348 | memcpy (enc->cur, str, len); |
286 | enc->cur += len; |
349 | enc->cur += len; |
287 | } |
350 | } |
288 | |
351 | |
289 | static void |
352 | ecb_inline void |
290 | encode_strref (enc_t *enc, int utf8, char *str, STRLEN len) |
353 | encode_strref (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len) |
291 | { |
354 | { |
292 | if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS)) |
355 | if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS)) |
293 | { |
356 | { |
294 | SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1); |
357 | SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1); |
295 | |
358 | |
… | |
… | |
306 | sv_setuv (*svp, enc->stringref_idx); |
369 | sv_setuv (*svp, enc->stringref_idx); |
307 | ++enc->stringref_idx; |
370 | ++enc->stringref_idx; |
308 | } |
371 | } |
309 | } |
372 | } |
310 | |
373 | |
311 | encode_str (enc, utf8, str, len); |
374 | encode_str (enc, upgrade_utf8, utf8, str, len); |
312 | } |
375 | } |
313 | |
376 | |
314 | static void encode_sv (enc_t *enc, SV *sv); |
377 | ecb_inline void |
|
|
378 | encode_float16 (enc_t *enc, NV nv) |
|
|
379 | { |
|
|
380 | need (enc, 1+2); |
|
|
381 | |
|
|
382 | *enc->cur++ = MAJOR_MISC | MISC_FLOAT16; |
|
|
383 | |
|
|
384 | uint16_t fp = ecb_float_to_binary16 (nv); |
|
|
385 | |
|
|
386 | if (!ecb_big_endian ()) |
|
|
387 | fp = ecb_bswap16 (fp); |
|
|
388 | |
|
|
389 | memcpy (enc->cur, &fp, 2); |
|
|
390 | enc->cur += 2; |
|
|
391 | } |
|
|
392 | |
|
|
393 | ecb_inline void |
|
|
394 | encode_float32 (enc_t *enc, NV nv) |
|
|
395 | { |
|
|
396 | need (enc, 1+4); |
|
|
397 | |
|
|
398 | *enc->cur++ = MAJOR_MISC | MISC_FLOAT32; |
|
|
399 | |
|
|
400 | uint32_t fp = ecb_float_to_binary32 (nv); |
|
|
401 | |
|
|
402 | if (!ecb_big_endian ()) |
|
|
403 | fp = ecb_bswap32 (fp); |
|
|
404 | |
|
|
405 | memcpy (enc->cur, &fp, 4); |
|
|
406 | enc->cur += 4; |
|
|
407 | } |
|
|
408 | |
|
|
409 | ecb_inline void |
|
|
410 | encode_float64 (enc_t *enc, NV nv) |
|
|
411 | { |
|
|
412 | need (enc, 1+8); |
|
|
413 | |
|
|
414 | *enc->cur++ = MAJOR_MISC | MISC_FLOAT64; |
|
|
415 | |
|
|
416 | uint64_t fp = ecb_double_to_binary64 (nv); |
|
|
417 | |
|
|
418 | if (!ecb_big_endian ()) |
|
|
419 | fp = ecb_bswap64 (fp); |
|
|
420 | |
|
|
421 | memcpy (enc->cur, &fp, 8); |
|
|
422 | enc->cur += 8; |
|
|
423 | } |
|
|
424 | |
|
|
425 | ecb_inline void |
|
|
426 | encode_bool (enc_t *enc, int istrue) |
|
|
427 | { |
|
|
428 | encode_ch (enc, istrue ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE); |
|
|
429 | } |
|
|
430 | |
|
|
431 | // encodes an arrayref containing key-value pairs as CBOR map |
|
|
432 | ecb_inline void |
|
|
433 | encode_array_as_map (enc_t *enc, SV *sv) |
|
|
434 | { |
|
|
435 | if (enc->depth >= enc->cbor.max_depth) |
|
|
436 | croak (ERR_NESTING_EXCEEDED); |
|
|
437 | |
|
|
438 | ++enc->depth; |
|
|
439 | |
|
|
440 | // as_map does error checking for us, but we re-check in case |
|
|
441 | // things have changed. |
|
|
442 | |
|
|
443 | if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) |
|
|
444 | croak ("CBOR::XS::as_map requires an array reference (did you change the array after calling as_map?)"); |
|
|
445 | |
|
|
446 | AV *av = (AV *)SvRV (sv); |
|
|
447 | int i, len = av_len (av); |
|
|
448 | |
|
|
449 | if (!(len & 1)) |
|
|
450 | croak ("CBOR::XS::as_map requires an even number of elements (did you change the array after calling as_map?)"); |
|
|
451 | |
|
|
452 | encode_uint (enc, MAJOR_MAP, (len + 1) >> 1); |
|
|
453 | |
|
|
454 | for (i = 0; i <= len; ++i) |
|
|
455 | { |
|
|
456 | SV **svp = av_fetch (av, i, 0); |
|
|
457 | encode_sv (enc, svp ? *svp : &PL_sv_undef); |
|
|
458 | } |
|
|
459 | |
|
|
460 | --enc->depth; |
|
|
461 | } |
|
|
462 | |
|
|
463 | ecb_inline void |
|
|
464 | encode_forced (enc_t *enc, UV type, SV *sv) |
|
|
465 | { |
|
|
466 | switch (type) |
|
|
467 | { |
|
|
468 | case AS_CBOR: |
|
|
469 | { |
|
|
470 | STRLEN len; |
|
|
471 | char *str = SvPVbyte (sv, len); |
|
|
472 | |
|
|
473 | need (enc, len); |
|
|
474 | memcpy (enc->cur, str, len); |
|
|
475 | enc->cur += len; |
|
|
476 | } |
|
|
477 | break; |
|
|
478 | |
|
|
479 | case AS_BYTES: |
|
|
480 | { |
|
|
481 | STRLEN len; |
|
|
482 | char *str = SvPVbyte (sv, len); |
|
|
483 | encode_strref (enc, 0, 0, str, len); |
|
|
484 | } |
|
|
485 | break; |
|
|
486 | |
|
|
487 | case AS_TEXT: |
|
|
488 | { |
|
|
489 | STRLEN len; |
|
|
490 | char *str = SvPVutf8 (sv, len); |
|
|
491 | encode_strref (enc, 1, 1, str, len); |
|
|
492 | } |
|
|
493 | break; |
|
|
494 | |
|
|
495 | case AS_INT: encode_int (enc, sv); break; |
|
|
496 | |
|
|
497 | case AS_FLOAT16: encode_float16 (enc, SvNV (sv)); break; |
|
|
498 | case AS_FLOAT32: encode_float32 (enc, SvNV (sv)); break; |
|
|
499 | case AS_FLOAT64: encode_float64 (enc, SvNV (sv)); break; |
|
|
500 | |
|
|
501 | case AS_MAP: encode_array_as_map (enc, sv); break; |
|
|
502 | |
|
|
503 | default: |
|
|
504 | croak ("encountered malformed CBOR::XS::Tagged object"); |
|
|
505 | } |
|
|
506 | } |
315 | |
507 | |
316 | static void |
508 | static void |
317 | encode_av (enc_t *enc, AV *av) |
509 | encode_av (enc_t *enc, AV *av) |
318 | { |
510 | { |
319 | int i, len = av_len (av); |
511 | int i, len = av_len (av); |
… | |
… | |
323 | |
515 | |
324 | ++enc->depth; |
516 | ++enc->depth; |
325 | |
517 | |
326 | encode_uint (enc, MAJOR_ARRAY, len + 1); |
518 | encode_uint (enc, MAJOR_ARRAY, len + 1); |
327 | |
519 | |
|
|
520 | if (ecb_expect_false (SvMAGICAL (av))) |
328 | for (i = 0; i <= len; ++i) |
521 | for (i = 0; i <= len; ++i) |
329 | { |
522 | { |
330 | SV **svp = av_fetch (av, i, 0); |
523 | SV **svp = av_fetch (av, i, 0); |
331 | encode_sv (enc, svp ? *svp : &PL_sv_undef); |
524 | encode_sv (enc, svp ? *svp : &PL_sv_undef); |
332 | } |
525 | } |
|
|
526 | else |
|
|
527 | for (i = 0; i <= len; ++i) |
|
|
528 | { |
|
|
529 | SV *sv = AvARRAY (av)[i]; |
|
|
530 | encode_sv (enc, sv ? sv : &PL_sv_undef); |
|
|
531 | } |
333 | |
532 | |
334 | --enc->depth; |
533 | --enc->depth; |
335 | } |
534 | } |
336 | |
535 | |
337 | static void |
536 | static void |
… | |
… | |
345 | ++enc->depth; |
544 | ++enc->depth; |
346 | |
545 | |
347 | int pairs = hv_iterinit (hv); |
546 | int pairs = hv_iterinit (hv); |
348 | int mg = SvMAGICAL (hv); |
547 | int mg = SvMAGICAL (hv); |
349 | |
548 | |
350 | if (mg) |
549 | if (ecb_expect_false (mg)) |
351 | encode_ch (enc, MAJOR_MAP | MINOR_INDEF); |
550 | encode_ch (enc, MAJOR_MAP | MINOR_INDEF); |
352 | else |
551 | else |
353 | encode_uint (enc, MAJOR_MAP, pairs); |
552 | encode_uint (enc, MAJOR_MAP, pairs); |
354 | |
553 | |
355 | while ((he = hv_iternext (hv))) |
554 | while ((he = hv_iternext (hv))) |
356 | { |
555 | { |
357 | if (HeKLEN (he) == HEf_SVKEY) |
556 | if (HeKLEN (he) == HEf_SVKEY) |
358 | encode_sv (enc, HeSVKEY (he)); |
557 | encode_sv (enc, HeSVKEY (he)); |
359 | else |
558 | else |
360 | encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); |
559 | encode_strref (enc, enc->cbor.flags & (F_TEXT_KEYS | F_TEXT_STRINGS), HeKUTF8 (he), HeKEY (he), HeKLEN (he)); |
361 | |
560 | |
362 | encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); |
561 | encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); |
363 | } |
562 | } |
364 | |
563 | |
365 | if (mg) |
564 | if (ecb_expect_false (mg)) |
366 | encode_ch (enc, MAJOR_MISC | MINOR_INDEF); |
565 | encode_ch (enc, MAJOR_MISC | MINOR_INDEF); |
367 | |
566 | |
368 | --enc->depth; |
567 | --enc->depth; |
369 | } |
568 | } |
370 | |
569 | |
… | |
… | |
390 | |
589 | |
391 | HV *stash = SvSTASH (sv); |
590 | HV *stash = SvSTASH (sv); |
392 | |
591 | |
393 | if (stash == boolean_stash) |
592 | if (stash == boolean_stash) |
394 | { |
593 | { |
395 | encode_ch (enc, SvIV (sv) ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE); |
594 | encode_bool (enc, SvIV (sv)); |
396 | return; |
595 | return; |
397 | } |
596 | } |
398 | else if (stash == error_stash) |
597 | else if (stash == error_stash) |
399 | { |
598 | { |
400 | encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF); |
599 | encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF); |
… | |
… | |
403 | else if (stash == tagged_stash) |
602 | else if (stash == tagged_stash) |
404 | { |
603 | { |
405 | if (svt != SVt_PVAV) |
604 | if (svt != SVt_PVAV) |
406 | croak ("encountered CBOR::XS::Tagged object that isn't an array"); |
605 | croak ("encountered CBOR::XS::Tagged object that isn't an array"); |
407 | |
606 | |
|
|
607 | switch (av_len ((AV *)sv)) |
|
|
608 | { |
|
|
609 | case 2-1: |
|
|
610 | // actually a tagged value |
408 | encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1))); |
611 | encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1))); |
409 | encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); |
612 | encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); |
|
|
613 | break; |
|
|
614 | |
|
|
615 | case 3-1: |
|
|
616 | // a forced type [value, type, undef] |
|
|
617 | encode_forced (enc, SvUV (*av_fetch ((AV *)sv, 1, 1)), *av_fetch ((AV *)sv, 0, 1)); |
|
|
618 | break; |
|
|
619 | |
|
|
620 | default: |
|
|
621 | croak ("encountered malformed CBOR::XS::Tagged object"); |
|
|
622 | } |
410 | |
623 | |
411 | return; |
624 | return; |
412 | } |
625 | } |
413 | } |
626 | } |
414 | |
627 | |
415 | if (ecb_expect_false (SvREFCNT (sv) > 1) |
628 | if (ecb_expect_false (SvREFCNT (sv) > 1) |
416 | && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)) |
629 | && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)) |
417 | { |
630 | { |
418 | if (!enc->shareable) |
631 | if (ecb_expect_false (!enc->shareable)) |
419 | enc->shareable = (HV *)sv_2mortal ((SV *)newHV ()); |
632 | enc->shareable = (HV *)sv_2mortal ((SV *)newHV ()); |
420 | |
633 | |
421 | SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1); |
634 | SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1); |
422 | |
635 | |
423 | if (SvOK (*svp)) |
636 | if (SvOK (*svp)) |
… | |
… | |
437 | if (ecb_expect_false (SvOBJECT (sv))) |
650 | if (ecb_expect_false (SvOBJECT (sv))) |
438 | { |
651 | { |
439 | HV *stash = SvSTASH (sv); |
652 | HV *stash = SvSTASH (sv); |
440 | GV *method; |
653 | GV *method; |
441 | |
654 | |
|
|
655 | if (enc->cbor.flags & F_FORBID_OBJECTS) |
|
|
656 | croak ("encountered object '%s', but forbid_objects is enabled", |
|
|
657 | SvPV_nolen (sv_2mortal (newRV_inc (sv)))); |
442 | if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) |
658 | else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) |
443 | { |
659 | { |
444 | dSP; |
660 | dSP; |
445 | |
661 | |
446 | ENTER; SAVETMPS; PUSHMARK (SP); |
662 | ENTER; SAVETMPS; |
|
|
663 | PUSHMARK (SP); |
447 | // we re-bless the reference to get overload and other niceties right |
664 | // we re-bless the reference to get overload and other niceties right |
448 | XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); |
665 | XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); |
449 | |
666 | |
450 | PUTBACK; |
667 | PUTBACK; |
451 | // G_SCALAR ensures that return value is 1 |
668 | // G_SCALAR ensures that return value is 1 |
… | |
… | |
464 | } |
681 | } |
465 | else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0) |
682 | else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0) |
466 | { |
683 | { |
467 | dSP; |
684 | dSP; |
468 | |
685 | |
469 | ENTER; SAVETMPS; PUSHMARK (SP); |
686 | ENTER; SAVETMPS; |
|
|
687 | PUSHMARK (SP); |
470 | EXTEND (SP, 2); |
688 | EXTEND (SP, 2); |
471 | // we re-bless the reference to get overload and other niceties right |
689 | // we re-bless the reference to get overload and other niceties right |
472 | PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); |
690 | PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); |
473 | PUSHs (sv_cbor); |
691 | PUSHs (sv_cbor); |
474 | |
692 | |
… | |
… | |
480 | if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) |
698 | if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) |
481 | croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash)); |
699 | croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash)); |
482 | |
700 | |
483 | encode_tag (enc, CBOR_TAG_PERL_OBJECT); |
701 | encode_tag (enc, CBOR_TAG_PERL_OBJECT); |
484 | encode_uint (enc, MAJOR_ARRAY, count + 1); |
702 | encode_uint (enc, MAJOR_ARRAY, count + 1); |
485 | encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); |
703 | encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); |
486 | |
704 | |
487 | while (count) |
705 | { |
|
|
706 | int i; |
|
|
707 | |
|
|
708 | for (i = 0; i < count; ++i) |
488 | encode_sv (enc, SP[1 - count--]); |
709 | encode_sv (enc, SP[i + 1 - count]); |
|
|
710 | |
|
|
711 | SP -= count; |
|
|
712 | } |
489 | |
713 | |
490 | PUTBACK; |
714 | PUTBACK; |
491 | |
715 | |
492 | FREETMPS; LEAVE; |
716 | FREETMPS; LEAVE; |
493 | } |
717 | } |
… | |
… | |
515 | |
739 | |
516 | if (ecb_expect_false (nv == (NV)(U32)nv)) |
740 | if (ecb_expect_false (nv == (NV)(U32)nv)) |
517 | encode_uint (enc, MAJOR_POS_INT, (U32)nv); |
741 | encode_uint (enc, MAJOR_POS_INT, (U32)nv); |
518 | //TODO: maybe I32? |
742 | //TODO: maybe I32? |
519 | else if (ecb_expect_false (nv == (float)nv)) |
743 | else if (ecb_expect_false (nv == (float)nv)) |
520 | { |
744 | encode_float32 (enc, nv); |
521 | uint32_t fp = ecb_float_to_binary32 (nv); |
|
|
522 | |
|
|
523 | *enc->cur++ = MAJOR_MISC | MISC_FLOAT32; |
|
|
524 | |
|
|
525 | if (!ecb_big_endian ()) |
|
|
526 | fp = ecb_bswap32 (fp); |
|
|
527 | |
|
|
528 | memcpy (enc->cur, &fp, 4); |
|
|
529 | enc->cur += 4; |
|
|
530 | } |
|
|
531 | else |
745 | else |
532 | { |
746 | encode_float64 (enc, nv); |
533 | uint64_t fp = ecb_double_to_binary64 (nv); |
|
|
534 | |
|
|
535 | *enc->cur++ = MAJOR_MISC | MISC_FLOAT64; |
|
|
536 | |
|
|
537 | if (!ecb_big_endian ()) |
|
|
538 | fp = ecb_bswap64 (fp); |
|
|
539 | |
|
|
540 | memcpy (enc->cur, &fp, 8); |
|
|
541 | enc->cur += 8; |
|
|
542 | } |
|
|
543 | } |
747 | } |
544 | |
748 | |
545 | static void |
749 | static void |
546 | encode_sv (enc_t *enc, SV *sv) |
750 | encode_sv (enc_t *enc, SV *sv) |
547 | { |
751 | { |
… | |
… | |
549 | |
753 | |
550 | if (SvPOKp (sv)) |
754 | if (SvPOKp (sv)) |
551 | { |
755 | { |
552 | STRLEN len; |
756 | STRLEN len; |
553 | char *str = SvPV (sv, len); |
757 | char *str = SvPV (sv, len); |
554 | encode_strref (enc, SvUTF8 (sv), str, len); |
758 | encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len); |
555 | } |
759 | } |
556 | else if (SvNOKp (sv)) |
760 | else if (SvNOKp (sv)) |
557 | encode_nv (enc, sv); |
761 | encode_nv (enc, sv); |
558 | else if (SvIOKp (sv)) |
762 | else if (SvIOKp (sv)) |
559 | { |
763 | encode_int (enc, sv); |
560 | if (SvIsUV (sv)) |
|
|
561 | encode_uint (enc, MAJOR_POS_INT, SvUVX (sv)); |
|
|
562 | else if (SvIVX (sv) >= 0) |
|
|
563 | encode_uint (enc, MAJOR_POS_INT, SvIVX (sv)); |
|
|
564 | else |
|
|
565 | encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1)); |
|
|
566 | } |
|
|
567 | else if (SvROK (sv)) |
764 | else if (SvROK (sv)) |
568 | encode_rv (enc, SvRV (sv)); |
765 | encode_rv (enc, SvRV (sv)); |
569 | else if (!SvOK (sv)) |
766 | else if (!SvOK (sv)) |
570 | encode_ch (enc, MAJOR_MISC | SIMPLE_NULL); |
767 | encode_ch (enc, MAJOR_MISC | SIMPLE_NULL); |
571 | else if (enc->cbor.flags & F_ALLOW_UNKNOWN) |
768 | else if (enc->cbor.flags & F_ALLOW_UNKNOWN) |
… | |
… | |
576 | } |
773 | } |
577 | |
774 | |
578 | static SV * |
775 | static SV * |
579 | encode_cbor (SV *scalar, CBOR *cbor) |
776 | encode_cbor (SV *scalar, CBOR *cbor) |
580 | { |
777 | { |
581 | enc_t enc = { }; |
778 | enc_t enc = { 0 }; |
582 | |
779 | |
583 | enc.cbor = *cbor; |
780 | enc.cbor = *cbor; |
584 | enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); |
781 | enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); |
585 | enc.cur = SvPVX (enc.sv); |
782 | enc.cur = SvPVX (enc.sv); |
586 | enc.end = SvEND (enc.sv); |
783 | enc.end = SvEND (enc.sv); |
587 | |
784 | |
588 | SvPOK_only (enc.sv); |
785 | SvPOK_only (enc.sv); |
589 | |
786 | |
590 | if (cbor->flags & F_PACK_STRINGS) |
787 | if (cbor->flags & F_PACK_STRINGS) |
591 | { |
788 | { |
… | |
… | |
618 | U32 depth; // recursion depth |
815 | U32 depth; // recursion depth |
619 | U32 maxdepth; // recursion depth limit |
816 | U32 maxdepth; // recursion depth limit |
620 | AV *shareable; |
817 | AV *shareable; |
621 | AV *stringref; |
818 | AV *stringref; |
622 | SV *decode_tagged; |
819 | SV *decode_tagged; |
|
|
820 | SV *err_sv; // optional sv for error, needs to be freed |
623 | } dec_t; |
821 | } dec_t; |
624 | |
822 | |
625 | #define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE |
823 | // set dec->err to ERRSV |
|
|
824 | ecb_cold static void |
|
|
825 | err_errsv (dec_t *dec) |
|
|
826 | { |
|
|
827 | if (!dec->err) |
|
|
828 | { |
|
|
829 | dec->err_sv = newSVsv (ERRSV); |
626 | |
830 | |
627 | #define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data") |
831 | // chop off the trailing \n |
|
|
832 | SvCUR_set (dec->err_sv, SvCUR (dec->err_sv) - 1); |
|
|
833 | *SvEND (dec->err_sv) = 0; |
628 | |
834 | |
|
|
835 | dec->err = SvPVutf8_nolen (dec->err_sv); |
|
|
836 | } |
|
|
837 | } |
|
|
838 | |
|
|
839 | // the following functions are used to reduce code size and help the compiler to optimise |
|
|
840 | ecb_cold static void |
|
|
841 | err_set (dec_t *dec, const char *reason) |
|
|
842 | { |
|
|
843 | if (!dec->err) |
|
|
844 | dec->err = reason; |
|
|
845 | } |
|
|
846 | |
|
|
847 | ecb_cold static void |
|
|
848 | err_unexpected_end (dec_t *dec) |
|
|
849 | { |
|
|
850 | err_set (dec, "unexpected end of CBOR data"); |
|
|
851 | } |
|
|
852 | |
|
|
853 | #define ERR_DO(do) SB do; goto fail; SE |
|
|
854 | #define ERR(reason) ERR_DO (err_set (dec, reason)) |
|
|
855 | #define ERR_ERRSV ERR_DO (err_errsv (dec)) |
|
|
856 | |
|
|
857 | #define WANT(len) if (ecb_expect_false ((uintptr_t)(dec->end - dec->cur) < (STRLEN)len)) ERR_DO (err_unexpected_end (dec)) |
|
|
858 | |
629 | #define DEC_INC_DEPTH if (++dec->depth > dec->cbor.max_depth) ERR (ERR_NESTING_EXCEEDED) |
859 | #define DEC_INC_DEPTH if (ecb_expect_false (++dec->depth > dec->cbor.max_depth)) ERR (ERR_NESTING_EXCEEDED) |
630 | #define DEC_DEC_DEPTH --dec->depth |
860 | #define DEC_DEC_DEPTH --dec->depth |
631 | |
861 | |
632 | static UV |
862 | static UV |
633 | decode_uint (dec_t *dec) |
863 | decode_uint (dec_t *dec) |
634 | { |
864 | { |
… | |
… | |
700 | |
930 | |
701 | for (;;) |
931 | for (;;) |
702 | { |
932 | { |
703 | WANT (1); |
933 | WANT (1); |
704 | |
934 | |
705 | if (*dec->cur == (MAJOR_MISC | MINOR_INDEF)) |
935 | if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err) |
706 | { |
936 | { |
707 | ++dec->cur; |
937 | ++dec->cur; |
708 | break; |
938 | break; |
709 | } |
939 | } |
710 | |
940 | |
711 | av_push (av, decode_sv (dec)); |
941 | av_push (av, decode_sv (dec)); |
712 | } |
942 | } |
713 | } |
943 | } |
714 | else |
944 | else |
715 | { |
945 | { |
716 | int i, len = decode_uint (dec); |
946 | UV i, len = decode_uint (dec); |
717 | |
947 | |
718 | WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays |
948 | WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays |
719 | av_fill (av, len - 1); |
949 | av_fill (av, len - 1); |
720 | |
950 | |
721 | for (i = 0; i < len; ++i) |
951 | for (i = 0; i < len; ++i) |
… | |
… | |
724 | |
954 | |
725 | DEC_DEC_DEPTH; |
955 | DEC_DEC_DEPTH; |
726 | return newRV_noinc ((SV *)av); |
956 | return newRV_noinc ((SV *)av); |
727 | |
957 | |
728 | fail: |
958 | fail: |
729 | SvREFCNT_dec (av); |
959 | SvREFCNT_dec_NN (av); |
730 | DEC_DEC_DEPTH; |
960 | DEC_DEC_DEPTH; |
731 | return &PL_sv_undef; |
961 | return &PL_sv_undef; |
732 | } |
962 | } |
733 | |
963 | |
734 | static void |
964 | static void |
… | |
… | |
736 | { |
966 | { |
737 | // for speed reasons, we specialcase single-string |
967 | // for speed reasons, we specialcase single-string |
738 | // byte or utf-8 strings as keys, but only when !stringref |
968 | // byte or utf-8 strings as keys, but only when !stringref |
739 | |
969 | |
740 | if (ecb_expect_true (!dec->stringref)) |
970 | if (ecb_expect_true (!dec->stringref)) |
741 | if (ecb_expect_true ((*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8)) |
971 | if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8)) |
742 | { |
972 | { |
743 | I32 len = decode_uint (dec); |
973 | STRLEN len = decode_uint (dec); |
744 | char *key = (char *)dec->cur; |
974 | char *key = (char *)dec->cur; |
745 | |
975 | |
|
|
976 | WANT (len); |
746 | dec->cur += len; |
977 | dec->cur += len; |
747 | |
978 | |
748 | hv_store (hv, key, len, decode_sv (dec), 0); |
979 | hv_store (hv, key, len, decode_sv (dec), 0); |
749 | |
980 | |
750 | return; |
981 | return; |
751 | } |
982 | } |
752 | else if (ecb_expect_true ((*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8)) |
983 | else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8)) |
753 | { |
984 | { |
754 | I32 len = decode_uint (dec); |
985 | STRLEN len = decode_uint (dec); |
755 | char *key = (char *)dec->cur; |
986 | char *key = (char *)dec->cur; |
756 | |
987 | |
|
|
988 | WANT (len); |
757 | dec->cur += len; |
989 | dec->cur += len; |
758 | |
990 | |
759 | if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8)) |
991 | if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8)) |
760 | if (!is_utf8_string (key, len)) |
992 | if (!is_utf8_string (key, len)) |
761 | ERR ("corrupted CBOR data (invalid UTF-8 in map key)"); |
993 | ERR ("corrupted CBOR data (invalid UTF-8 in map key)"); |
… | |
… | |
766 | } |
998 | } |
767 | |
999 | |
768 | SV *k = decode_sv (dec); |
1000 | SV *k = decode_sv (dec); |
769 | SV *v = decode_sv (dec); |
1001 | SV *v = decode_sv (dec); |
770 | |
1002 | |
|
|
1003 | // we leak memory if uncaught exceptions are thrown by random magical |
|
|
1004 | // methods, and this is hopefully the only place where it can happen, |
|
|
1005 | // so if there is a chance of an exception, take the very slow path. |
|
|
1006 | // since catching exceptions is "undocumented/internal/forbidden" by |
|
|
1007 | // the new p5p powers, we need to call out to a perl function :/ |
|
|
1008 | if (ecb_expect_false (SvAMAGIC (k))) |
|
|
1009 | { |
|
|
1010 | dSP; |
|
|
1011 | |
|
|
1012 | ENTER; SAVETMPS; |
|
|
1013 | PUSHMARK (SP); |
|
|
1014 | EXTEND (SP, 3); |
|
|
1015 | PUSHs (sv_2mortal (newRV_inc ((SV *)hv))); |
|
|
1016 | PUSHs (sv_2mortal (k)); |
|
|
1017 | PUSHs (sv_2mortal (v)); |
|
|
1018 | |
|
|
1019 | PUTBACK; |
|
|
1020 | call_pv ("CBOR::XS::_hv_store", G_VOID | G_DISCARD | G_EVAL); |
|
|
1021 | SPAGAIN; |
|
|
1022 | |
|
|
1023 | FREETMPS; LEAVE; |
|
|
1024 | |
|
|
1025 | if (SvTRUE (ERRSV)) |
|
|
1026 | ERR_ERRSV; |
|
|
1027 | |
|
|
1028 | return; |
|
|
1029 | } |
|
|
1030 | |
771 | hv_store_ent (hv, k, v, 0); |
1031 | hv_store_ent (hv, k, v, 0); |
772 | SvREFCNT_dec (k); |
1032 | SvREFCNT_dec_NN (k); |
773 | |
1033 | |
774 | fail: |
1034 | fail: |
775 | ; |
1035 | ; |
776 | } |
1036 | } |
777 | |
1037 | |
… | |
… | |
788 | |
1048 | |
789 | for (;;) |
1049 | for (;;) |
790 | { |
1050 | { |
791 | WANT (1); |
1051 | WANT (1); |
792 | |
1052 | |
793 | if (*dec->cur == (MAJOR_MISC | MINOR_INDEF)) |
1053 | if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err) |
794 | { |
1054 | { |
795 | ++dec->cur; |
1055 | ++dec->cur; |
796 | break; |
1056 | break; |
797 | } |
1057 | } |
798 | |
1058 | |
799 | decode_he (dec, hv); |
1059 | decode_he (dec, hv); |
800 | } |
1060 | } |
801 | } |
1061 | } |
802 | else |
1062 | else |
803 | { |
1063 | { |
804 | int pairs = decode_uint (dec); |
1064 | UV pairs = decode_uint (dec); |
|
|
1065 | |
|
|
1066 | WANT (pairs); // complexity check - need at least one byte per value, do not allow supersize hashes |
805 | |
1067 | |
806 | while (pairs--) |
1068 | while (pairs--) |
807 | decode_he (dec, hv); |
1069 | decode_he (dec, hv); |
808 | } |
1070 | } |
809 | |
1071 | |
810 | DEC_DEC_DEPTH; |
1072 | DEC_DEC_DEPTH; |
811 | return newRV_noinc ((SV *)hv); |
1073 | return newRV_noinc ((SV *)hv); |
812 | |
1074 | |
813 | fail: |
1075 | fail: |
814 | SvREFCNT_dec (hv); |
1076 | SvREFCNT_dec_NN (hv); |
815 | DEC_DEC_DEPTH; |
1077 | DEC_DEC_DEPTH; |
816 | return &PL_sv_undef; |
1078 | return &PL_sv_undef; |
817 | } |
1079 | } |
818 | |
1080 | |
819 | static SV * |
1081 | static SV * |
820 | decode_str (dec_t *dec, int utf8) |
1082 | decode_str (dec_t *dec, int utf8) |
821 | { |
1083 | { |
822 | SV *sv = 0; |
1084 | SV *sv = 0; |
823 | |
1085 | |
824 | if ((*dec->cur & MINOR_MASK) == MINOR_INDEF) |
1086 | if (ecb_expect_false ((*dec->cur & MINOR_MASK) == MINOR_INDEF)) |
825 | { |
1087 | { |
826 | // indefinite length strings |
1088 | // indefinite length strings |
827 | ++dec->cur; |
1089 | ++dec->cur; |
828 | |
1090 | |
829 | U8 major = *dec->cur & MAJOR_MISC; |
1091 | U8 major = *dec->cur & MAJOR_MISC; |
… | |
… | |
897 | sv = newRV_noinc (decode_sv (dec)); |
1159 | sv = newRV_noinc (decode_sv (dec)); |
898 | break; |
1160 | break; |
899 | |
1161 | |
900 | case CBOR_TAG_STRINGREF_NAMESPACE: |
1162 | case CBOR_TAG_STRINGREF_NAMESPACE: |
901 | { |
1163 | { |
|
|
1164 | // do not use SAVETMPS/FREETMPS, as these will |
|
|
1165 | // erase mortalised caches, e.g. "shareable" |
902 | ENTER; SAVETMPS; |
1166 | ENTER; |
903 | |
1167 | |
904 | SAVESPTR (dec->stringref); |
1168 | SAVESPTR (dec->stringref); |
905 | dec->stringref = (AV *)sv_2mortal ((SV *)newAV ()); |
1169 | dec->stringref = (AV *)sv_2mortal ((SV *)newAV ()); |
906 | |
1170 | |
907 | sv = decode_sv (dec); |
1171 | sv = decode_sv (dec); |
908 | |
1172 | |
909 | FREETMPS; LEAVE; |
1173 | LEAVE; |
910 | } |
1174 | } |
911 | break; |
1175 | break; |
912 | |
1176 | |
913 | case CBOR_TAG_STRINGREF: |
1177 | case CBOR_TAG_STRINGREF: |
914 | { |
1178 | { |
915 | if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT)) |
1179 | if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT)) |
916 | ERR ("corrupted CBOR data (stringref index not an unsigned integer)"); |
1180 | ERR ("corrupted CBOR data (stringref index not an unsigned integer)"); |
917 | |
1181 | |
918 | UV idx = decode_uint (dec); |
1182 | UV idx = decode_uint (dec); |
919 | |
1183 | |
920 | if (!dec->stringref || (int)idx > AvFILLp (dec->stringref)) |
1184 | if (!dec->stringref || idx >= (UV)(1 + AvFILLp (dec->stringref))) |
921 | ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)"); |
1185 | ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)"); |
922 | |
1186 | |
923 | sv = newSVsv (AvARRAY (dec->stringref)[idx]); |
1187 | sv = newSVsv (AvARRAY (dec->stringref)[idx]); |
924 | } |
1188 | } |
925 | break; |
1189 | break; |
… | |
… | |
953 | if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT)) |
1217 | if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT)) |
954 | ERR ("corrupted CBOR data (sharedref index not an unsigned integer)"); |
1218 | ERR ("corrupted CBOR data (sharedref index not an unsigned integer)"); |
955 | |
1219 | |
956 | UV idx = decode_uint (dec); |
1220 | UV idx = decode_uint (dec); |
957 | |
1221 | |
958 | if (!dec->shareable || (int)idx > AvFILLp (dec->shareable)) |
1222 | if (!dec->shareable || idx >= (UV)(1 + AvFILLp (dec->shareable))) |
959 | ERR ("corrupted CBOR data (sharedref index out of bounds)"); |
1223 | ERR ("corrupted CBOR data (sharedref index out of bounds)"); |
960 | |
1224 | |
961 | sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]); |
1225 | sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]); |
962 | |
1226 | |
963 | if (sv == &PL_sv_undef) |
1227 | if (sv == &PL_sv_undef) |
… | |
… | |
965 | } |
1229 | } |
966 | break; |
1230 | break; |
967 | |
1231 | |
968 | case CBOR_TAG_PERL_OBJECT: |
1232 | case CBOR_TAG_PERL_OBJECT: |
969 | { |
1233 | { |
|
|
1234 | if (dec->cbor.flags & F_FORBID_OBJECTS) |
|
|
1235 | goto filter; |
|
|
1236 | |
970 | sv = decode_sv (dec); |
1237 | sv = decode_sv (dec); |
971 | |
1238 | |
972 | if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) |
1239 | if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) |
973 | ERR ("corrupted CBOR data (non-array perl object)"); |
1240 | ERR ("corrupted CBOR data (non-array perl object)"); |
974 | |
1241 | |
… | |
… | |
984 | if (!method) |
1251 | if (!method) |
985 | ERR ("cannot decode perl-object (package does not have a THAW method)"); |
1252 | ERR ("cannot decode perl-object (package does not have a THAW method)"); |
986 | |
1253 | |
987 | dSP; |
1254 | dSP; |
988 | |
1255 | |
989 | ENTER; SAVETMPS; PUSHMARK (SP); |
1256 | ENTER; SAVETMPS; |
|
|
1257 | PUSHMARK (SP); |
990 | EXTEND (SP, len + 1); |
1258 | EXTEND (SP, len + 1); |
991 | // we re-bless the reference to get overload and other niceties right |
1259 | // we re-bless the reference to get overload and other niceties right |
992 | PUSHs (*av_fetch (av, 0, 1)); |
1260 | PUSHs (*av_fetch (av, 0, 1)); |
993 | PUSHs (sv_cbor); |
1261 | PUSHs (sv_cbor); |
994 | |
1262 | |
… | |
… | |
1002 | SPAGAIN; |
1270 | SPAGAIN; |
1003 | |
1271 | |
1004 | if (SvTRUE (ERRSV)) |
1272 | if (SvTRUE (ERRSV)) |
1005 | { |
1273 | { |
1006 | FREETMPS; LEAVE; |
1274 | FREETMPS; LEAVE; |
1007 | ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV)))); |
1275 | ERR_ERRSV; |
1008 | } |
1276 | } |
1009 | |
1277 | |
1010 | SvREFCNT_dec (sv); |
1278 | SvREFCNT_dec_NN (sv); |
1011 | sv = SvREFCNT_inc (POPs); |
1279 | sv = SvREFCNT_inc (POPs); |
1012 | |
1280 | |
1013 | PUTBACK; |
1281 | PUTBACK; |
1014 | |
1282 | |
1015 | FREETMPS; LEAVE; |
1283 | FREETMPS; LEAVE; |
1016 | } |
1284 | } |
1017 | break; |
1285 | break; |
1018 | |
1286 | |
1019 | default: |
1287 | default: |
|
|
1288 | filter: |
1020 | { |
1289 | { |
|
|
1290 | SV *tag_sv = newSVuv (tag); |
|
|
1291 | |
1021 | sv = decode_sv (dec); |
1292 | sv = decode_sv (dec); |
1022 | |
1293 | |
1023 | dSP; |
1294 | dSP; |
1024 | ENTER; SAVETMPS; PUSHMARK (SP); |
1295 | ENTER; SAVETMPS; |
|
|
1296 | PUSHMARK (SP); |
1025 | EXTEND (SP, 2); |
1297 | EXTEND (SP, 2); |
1026 | PUSHs (newSVuv (tag)); |
1298 | PUSHs (tag_sv); |
1027 | PUSHs (sv); |
1299 | PUSHs (sv); |
1028 | |
1300 | |
1029 | PUTBACK; |
1301 | PUTBACK; |
1030 | int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL); |
1302 | int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL); |
1031 | SPAGAIN; |
1303 | SPAGAIN; |
1032 | |
1304 | |
1033 | if (SvTRUE (ERRSV)) |
1305 | if (SvTRUE (ERRSV)) |
1034 | { |
1306 | { |
|
|
1307 | SvREFCNT_dec_NN (tag_sv); |
1035 | FREETMPS; LEAVE; |
1308 | FREETMPS; LEAVE; |
1036 | ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV)))); |
1309 | ERR_ERRSV; |
1037 | } |
1310 | } |
1038 | |
1311 | |
1039 | if (count) |
1312 | if (count) |
1040 | { |
1313 | { |
|
|
1314 | SvREFCNT_dec_NN (tag_sv); |
1041 | SvREFCNT_dec (sv); |
1315 | SvREFCNT_dec_NN (sv); |
1042 | sv = SvREFCNT_inc (POPs); |
1316 | sv = SvREFCNT_inc_NN (TOPs); |
|
|
1317 | SP -= count; |
1043 | } |
1318 | } |
1044 | else |
1319 | else |
1045 | { |
1320 | { |
1046 | AV *av = newAV (); |
1321 | AV *av = newAV (); |
1047 | av_push (av, newSVuv (tag)); |
1322 | av_push (av, tag_sv); |
1048 | av_push (av, sv); |
1323 | av_push (av, sv); |
1049 | |
1324 | |
1050 | HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash |
1325 | HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash |
1051 | ? cbor_tagged_stash |
1326 | ? cbor_tagged_stash |
1052 | : gv_stashpv ("CBOR::XS::Tagged" , 1); |
1327 | : gv_stashpv ("CBOR::XS::Tagged" , 1); |
… | |
… | |
1155 | } |
1430 | } |
1156 | |
1431 | |
1157 | static SV * |
1432 | static SV * |
1158 | decode_cbor (SV *string, CBOR *cbor, char **offset_return) |
1433 | decode_cbor (SV *string, CBOR *cbor, char **offset_return) |
1159 | { |
1434 | { |
1160 | dec_t dec = { }; |
1435 | dec_t dec = { 0 }; |
1161 | SV *sv; |
1436 | SV *sv; |
1162 | STRLEN len; |
1437 | STRLEN len; |
1163 | char *data = SvPVbyte (string, len); |
1438 | char *data = SvPVbyte (string, len); |
1164 | |
1439 | |
1165 | if (len > cbor->max_size && cbor->max_size) |
1440 | if (len > cbor->max_size && cbor->max_size) |
… | |
… | |
1181 | |
1456 | |
1182 | if (dec.err) |
1457 | if (dec.err) |
1183 | { |
1458 | { |
1184 | if (dec.shareable) |
1459 | if (dec.shareable) |
1185 | { |
1460 | { |
1186 | // need to break cyclic links, which whould all be in shareable |
1461 | // need to break cyclic links, which would all be in shareable |
1187 | int i; |
1462 | int i; |
1188 | SV **svp; |
1463 | SV **svp; |
1189 | |
1464 | |
1190 | for (i = av_len (dec.shareable) + 1; i--; ) |
1465 | for (i = av_len (dec.shareable) + 1; i--; ) |
1191 | if ((svp = av_fetch (dec.shareable, i, 0))) |
1466 | if ((svp = av_fetch (dec.shareable, i, 0))) |
1192 | sv_setsv (*svp, &PL_sv_undef); |
1467 | sv_setsv (*svp, &PL_sv_undef); |
1193 | } |
1468 | } |
1194 | |
1469 | |
1195 | SvREFCNT_dec (sv); |
1470 | SvREFCNT_dec_NN (sv); |
|
|
1471 | |
|
|
1472 | if (dec.err_sv) |
|
|
1473 | sv_2mortal (dec.err_sv); |
|
|
1474 | |
1196 | croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur); |
1475 | croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur); |
1197 | } |
1476 | } |
1198 | |
1477 | |
1199 | sv = sv_2mortal (sv); |
1478 | sv = sv_2mortal (sv); |
1200 | |
1479 | |
… | |
… | |
1256 | |
1535 | |
1257 | int major = *p >> MAJOR_SHIFT; |
1536 | int major = *p >> MAJOR_SHIFT; |
1258 | |
1537 | |
1259 | switch (major) |
1538 | switch (major) |
1260 | { |
1539 | { |
|
|
1540 | case MAJOR_TAG >> MAJOR_SHIFT: |
|
|
1541 | ++count; // tags merely prefix another value |
|
|
1542 | break; |
|
|
1543 | |
1261 | case MAJOR_BYTES >> MAJOR_SHIFT: |
1544 | case MAJOR_BYTES >> MAJOR_SHIFT: |
1262 | case MAJOR_TEXT >> MAJOR_SHIFT: |
1545 | case MAJOR_TEXT >> MAJOR_SHIFT: |
1263 | case MAJOR_ARRAY >> MAJOR_SHIFT: |
1546 | case MAJOR_ARRAY >> MAJOR_SHIFT: |
1264 | case MAJOR_MAP >> MAJOR_SHIFT: |
1547 | case MAJOR_MAP >> MAJOR_SHIFT: |
1265 | { |
1548 | { |
… | |
… | |
1340 | |
1623 | |
1341 | default_filter = newSVpv ("CBOR::XS::default_filter", 0); |
1624 | default_filter = newSVpv ("CBOR::XS::default_filter", 0); |
1342 | |
1625 | |
1343 | sv_cbor = newSVpv ("CBOR", 0); |
1626 | sv_cbor = newSVpv ("CBOR", 0); |
1344 | SvREADONLY_on (sv_cbor); |
1627 | SvREADONLY_on (sv_cbor); |
|
|
1628 | |
|
|
1629 | assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1)); |
1345 | } |
1630 | } |
1346 | |
1631 | |
1347 | PROTOTYPES: DISABLE |
1632 | PROTOTYPES: DISABLE |
1348 | |
1633 | |
1349 | void CLONE (...) |
1634 | void CLONE (...) |
… | |
… | |
1369 | ALIAS: |
1654 | ALIAS: |
1370 | shrink = F_SHRINK |
1655 | shrink = F_SHRINK |
1371 | allow_unknown = F_ALLOW_UNKNOWN |
1656 | allow_unknown = F_ALLOW_UNKNOWN |
1372 | allow_sharing = F_ALLOW_SHARING |
1657 | allow_sharing = F_ALLOW_SHARING |
1373 | allow_cycles = F_ALLOW_CYCLES |
1658 | allow_cycles = F_ALLOW_CYCLES |
|
|
1659 | forbid_objects = F_FORBID_OBJECTS |
1374 | pack_strings = F_PACK_STRINGS |
1660 | pack_strings = F_PACK_STRINGS |
|
|
1661 | text_keys = F_TEXT_KEYS |
|
|
1662 | text_strings = F_TEXT_STRINGS |
1375 | validate_utf8 = F_VALIDATE_UTF8 |
1663 | validate_utf8 = F_VALIDATE_UTF8 |
1376 | PPCODE: |
1664 | PPCODE: |
1377 | { |
1665 | { |
1378 | if (enable) |
1666 | if (enable) |
1379 | self->flags |= ix; |
1667 | self->flags |= ix; |
… | |
… | |
1387 | ALIAS: |
1675 | ALIAS: |
1388 | get_shrink = F_SHRINK |
1676 | get_shrink = F_SHRINK |
1389 | get_allow_unknown = F_ALLOW_UNKNOWN |
1677 | get_allow_unknown = F_ALLOW_UNKNOWN |
1390 | get_allow_sharing = F_ALLOW_SHARING |
1678 | get_allow_sharing = F_ALLOW_SHARING |
1391 | get_allow_cycles = F_ALLOW_CYCLES |
1679 | get_allow_cycles = F_ALLOW_CYCLES |
|
|
1680 | get_forbid_objects = F_FORBID_OBJECTS |
1392 | get_pack_strings = F_PACK_STRINGS |
1681 | get_pack_strings = F_PACK_STRINGS |
|
|
1682 | get_text_keys = F_TEXT_KEYS |
|
|
1683 | get_text_strings = F_TEXT_STRINGS |
1393 | get_validate_utf8 = F_VALIDATE_UTF8 |
1684 | get_validate_utf8 = F_VALIDATE_UTF8 |
1394 | PPCODE: |
1685 | PPCODE: |
1395 | XPUSHs (boolSV (self->flags & ix)); |
1686 | XPUSHs (boolSV (self->flags & ix)); |
1396 | |
1687 | |
1397 | void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) |
1688 | void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) |
… | |
… | |
1448 | PUSHs (sv); |
1739 | PUSHs (sv); |
1449 | PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); |
1740 | PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); |
1450 | } |
1741 | } |
1451 | |
1742 | |
1452 | void incr_parse (CBOR *self, SV *cborstr) |
1743 | void incr_parse (CBOR *self, SV *cborstr) |
|
|
1744 | ALIAS: |
|
|
1745 | incr_parse_multiple = 1 |
1453 | PPCODE: |
1746 | PPCODE: |
1454 | { |
1747 | { |
1455 | if (SvUTF8 (cborstr)) |
1748 | if (SvUTF8 (cborstr)) |
1456 | sv_utf8_downgrade (cborstr, 0); |
1749 | sv_utf8_downgrade (cborstr, 0); |
1457 | |
1750 | |
… | |
… | |
1487 | av_push (self->incr_count, newSViv (1)); |
1780 | av_push (self->incr_count, newSViv (1)); |
1488 | |
1781 | |
1489 | self->incr_pos = 0; |
1782 | self->incr_pos = 0; |
1490 | self->incr_need = self->incr_pos + 1; |
1783 | self->incr_need = self->incr_pos + 1; |
1491 | } |
1784 | } |
1492 | while (GIMME_V == G_ARRAY); |
1785 | while (ix); |
1493 | } |
1786 | } |
1494 | |
1787 | |
1495 | void incr_reset (CBOR *self) |
1788 | void incr_reset (CBOR *self) |
1496 | CODE: |
1789 | CODE: |
1497 | { |
1790 | { |
… | |
… | |
1525 | cbor_init (&cbor); |
1818 | cbor_init (&cbor); |
1526 | PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN; |
1819 | PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN; |
1527 | XPUSHs (cborstr); |
1820 | XPUSHs (cborstr); |
1528 | } |
1821 | } |
1529 | |
1822 | |
|
|
1823 | #ifdef __AFL_COMPILER |
|
|
1824 | |
|
|
1825 | void |
|
|
1826 | afl_init () |
|
|
1827 | CODE: |
|
|
1828 | __AFL_INIT (); |
|
|
1829 | |
|
|
1830 | int |
|
|
1831 | afl_loop (unsigned int count = 10000) |
|
|
1832 | CODE: |
|
|
1833 | RETVAL = __AFL_LOOP (count); |
|
|
1834 | OUTPUT: |
|
|
1835 | RETVAL |
|
|
1836 | |
|
|
1837 | #endif |
|
|
1838 | |