… | |
… | |
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 | |
12 | #define ECB_NO_THREADS 1 |
13 | #define ECB_NO_THREADS 1 |
13 | #include "ecb.h" |
14 | #include "ecb.h" |
14 | |
15 | |
15 | // compatibility with perl <5.18 |
16 | // compatibility with perl <5.18 |
… | |
… | |
20 | # define HvNAMELEN(hv) HvNAMELEN_get (hv) |
21 | # define HvNAMELEN(hv) HvNAMELEN_get (hv) |
21 | #endif |
22 | #endif |
22 | #ifndef HvNAMEUTF8 |
23 | #ifndef HvNAMEUTF8 |
23 | # define HvNAMEUTF8(hv) 0 |
24 | # define HvNAMEUTF8(hv) 0 |
24 | #endif |
25 | #endif |
|
|
26 | #ifndef SvREFCNT_inc_NN |
|
|
27 | # define SvREFCNT_inc_NN(sv) SvREFCNT_inc (sv) |
|
|
28 | #endif |
25 | #ifndef SvREFCNT_dec_NN |
29 | #ifndef SvREFCNT_dec_NN |
26 | # define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv) |
30 | # define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv) |
27 | #endif |
31 | #endif |
|
|
32 | |
|
|
33 | // perl's is_utf8_string interprets len=0 as "calculate len", but we want it to mean 0 |
|
|
34 | #define cbor_is_utf8_string(str,len) (!(len) || is_utf8_string ((str), (len))) |
28 | |
35 | |
29 | // known major and minor types |
36 | // known major and minor types |
30 | enum cbor_type |
37 | enum cbor_type |
31 | { |
38 | { |
32 | MAJOR_SHIFT = 5, |
39 | MAJOR_SHIFT = 5, |
… | |
… | |
95 | CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8 |
102 | CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8 |
96 | |
103 | |
97 | CBOR_TAG_MAGIC = 55799, // self-describe cbor |
104 | CBOR_TAG_MAGIC = 55799, // self-describe cbor |
98 | }; |
105 | }; |
99 | |
106 | |
|
|
107 | // known forced types, also hardcoded in CBOR.pm |
|
|
108 | enum |
|
|
109 | { |
|
|
110 | AS_CBOR = 0, |
|
|
111 | AS_INT = 1, |
|
|
112 | AS_BYTES = 2, |
|
|
113 | AS_TEXT = 3, |
|
|
114 | AS_FLOAT16 = 4, |
|
|
115 | AS_FLOAT32 = 5, |
|
|
116 | AS_FLOAT64 = 6, |
|
|
117 | AS_MAP = 7, |
|
|
118 | // possibly future enhancements: (generic) float, (generic) string |
|
|
119 | }; |
|
|
120 | |
100 | #define F_SHRINK 0x00000001UL |
121 | #define F_SHRINK 0x00000001UL |
101 | #define F_ALLOW_UNKNOWN 0x00000002UL |
122 | #define F_ALLOW_UNKNOWN 0x00000002UL |
102 | #define F_ALLOW_SHARING 0x00000004UL |
123 | #define F_ALLOW_SHARING 0x00000004UL |
103 | #define F_ALLOW_CYCLES 0x00000008UL |
124 | #define F_ALLOW_CYCLES 0x00000008UL |
|
|
125 | #define F_FORBID_OBJECTS 0x00000010UL |
104 | #define F_PACK_STRINGS 0x00000010UL |
126 | #define F_PACK_STRINGS 0x00000020UL |
|
|
127 | #define F_TEXT_KEYS 0x00000040UL |
105 | #define F_UTF8_STRINGS 0x00000020UL |
128 | #define F_TEXT_STRINGS 0x00000080UL |
106 | #define F_VALIDATE_UTF8 0x00000040UL |
129 | #define F_VALIDATE_UTF8 0x00000100UL |
107 | |
130 | |
108 | #define INIT_SIZE 32 // initial scalar size to be allocated |
131 | #define INIT_SIZE 32 // initial scalar size to be allocated |
109 | |
132 | |
110 | #define SB do { |
133 | #define SB do { |
111 | #define SE } while (0) |
134 | #define SE } while (0) |
… | |
… | |
181 | #endif |
204 | #endif |
182 | } |
205 | } |
183 | } |
206 | } |
184 | |
207 | |
185 | // minimum length of a string to be registered for stringref |
208 | // minimum length of a string to be registered for stringref |
186 | ecb_inline int |
209 | ecb_inline STRLEN |
187 | minimum_string_length (UV idx) |
210 | minimum_string_length (UV idx) |
188 | { |
211 | { |
189 | return idx > 23 |
212 | return idx <= 23 ? 3 |
190 | ? idx > 0xffU |
213 | : idx <= 0xffU ? 4 |
191 | ? idx > 0xffffU |
214 | : idx <= 0xffffU ? 5 |
192 | ? idx > 0xffffffffU |
215 | : idx <= 0xffffffffU ? 7 |
193 | ? 11 |
216 | : 11; |
194 | : 7 |
|
|
195 | : 5 |
|
|
196 | : 4 |
|
|
197 | : 3; |
|
|
198 | } |
217 | } |
199 | |
218 | |
200 | ///////////////////////////////////////////////////////////////////////////// |
219 | ///////////////////////////////////////////////////////////////////////////// |
201 | // encoder |
220 | // encoder |
202 | |
221 | |
… | |
… | |
215 | } enc_t; |
234 | } enc_t; |
216 | |
235 | |
217 | ecb_inline void |
236 | ecb_inline void |
218 | need (enc_t *enc, STRLEN len) |
237 | need (enc_t *enc, STRLEN len) |
219 | { |
238 | { |
220 | if (ecb_expect_false (enc->cur + len >= enc->end)) |
239 | if (ecb_expect_false ((uintptr_t)(enc->end - enc->cur) < len)) |
221 | { |
240 | { |
222 | STRLEN cur = enc->cur - (char *)SvPVX (enc->sv); |
241 | STRLEN cur = enc->cur - (char *)SvPVX (enc->sv); |
223 | SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1); |
242 | SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1); |
224 | enc->cur = SvPVX (enc->sv) + cur; |
243 | enc->cur = SvPVX (enc->sv) + cur; |
225 | enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1; |
244 | enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1; |
226 | } |
245 | } |
227 | } |
246 | } |
228 | |
247 | |
|
|
248 | static void encode_sv (enc_t *enc, SV *sv); |
|
|
249 | |
229 | ecb_inline void |
250 | ecb_inline void |
230 | encode_ch (enc_t *enc, char ch) |
251 | encode_ch (enc_t *enc, char ch) |
231 | { |
252 | { |
232 | need (enc, 1); |
253 | need (enc, 1); |
233 | *enc->cur++ = ch; |
254 | *enc->cur++ = ch; |
234 | } |
255 | } |
235 | |
256 | |
|
|
257 | // used for tags, intregers, element counts and so on |
236 | static void |
258 | static void |
237 | encode_uint (enc_t *enc, int major, UV len) |
259 | encode_uint (enc_t *enc, int major, UV len) |
238 | { |
260 | { |
239 | need (enc, 9); |
261 | need (enc, 9); |
240 | |
262 | |
… | |
… | |
271 | *enc->cur++ = len >> 8; |
293 | *enc->cur++ = len >> 8; |
272 | *enc->cur++ = len; |
294 | *enc->cur++ = len; |
273 | } |
295 | } |
274 | } |
296 | } |
275 | |
297 | |
|
|
298 | // encodes a perl value into a CBOR integer |
|
|
299 | ecb_inline void |
|
|
300 | encode_int (enc_t *enc, SV *sv) |
|
|
301 | { |
|
|
302 | if (SvIsUV (sv)) |
|
|
303 | encode_uint (enc, MAJOR_POS_INT, SvUVX (sv)); |
|
|
304 | else if (SvIVX (sv) >= 0) |
|
|
305 | encode_uint (enc, MAJOR_POS_INT, SvIVX (sv)); |
|
|
306 | else |
|
|
307 | encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1)); |
|
|
308 | } |
|
|
309 | |
276 | ecb_inline void |
310 | ecb_inline void |
277 | encode_tag (enc_t *enc, UV tag) |
311 | encode_tag (enc_t *enc, UV tag) |
278 | { |
312 | { |
279 | encode_uint (enc, MAJOR_TAG, tag); |
313 | encode_uint (enc, MAJOR_TAG, tag); |
280 | } |
314 | } |
281 | |
315 | |
|
|
316 | // exceptional (hopefully) slow path for byte strings that need to be utf8-encoded |
|
|
317 | ecb_noinline static void |
|
|
318 | encode_str_utf8 (enc_t *enc, int utf8, char *str, STRLEN len) |
|
|
319 | { |
|
|
320 | STRLEN ulen = len; |
|
|
321 | U8 *p, *pend = (U8 *)str + len; |
|
|
322 | |
|
|
323 | for (p = (U8 *)str; p < pend; ++p) |
|
|
324 | ulen += *p >> 7; // count set high bits |
|
|
325 | |
|
|
326 | encode_uint (enc, MAJOR_TEXT, ulen); |
|
|
327 | |
|
|
328 | need (enc, ulen); |
|
|
329 | for (p = (U8 *)str; p < pend; ++p) |
|
|
330 | if (*p < 0x80) |
|
|
331 | *enc->cur++ = *p; |
|
|
332 | else |
|
|
333 | { |
|
|
334 | *enc->cur++ = 0xc0 + (*p >> 6); |
|
|
335 | *enc->cur++ = 0x80 + (*p & 63); |
|
|
336 | } |
|
|
337 | } |
|
|
338 | |
282 | ecb_inline void |
339 | ecb_inline void |
283 | encode_str (enc_t *enc, int utf8, char *str, STRLEN len) |
340 | encode_str (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len) |
284 | { |
341 | { |
285 | if (ecb_expect_false (enc->cbor.flags & F_UTF8_STRINGS)) |
342 | if (ecb_expect_false (upgrade_utf8)) |
286 | if (!utf8) |
343 | if (!utf8) |
287 | { |
344 | { |
288 | SV *sv = sv_newmortal (); |
|
|
289 | char *s; STRLEN l; |
|
|
290 | |
|
|
291 | sv_setpvn (sv, str, len); |
|
|
292 | |
|
|
293 | s = SvPVutf8 (sv, l); |
|
|
294 | encode_str (enc, 1, s, l); |
345 | encode_str_utf8 (enc, utf8, str, len); |
295 | return; |
346 | return; |
296 | } |
347 | } |
297 | |
348 | |
298 | encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len); |
349 | encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len); |
299 | need (enc, len); |
350 | need (enc, len); |
300 | memcpy (enc->cur, str, len); |
351 | memcpy (enc->cur, str, len); |
301 | enc->cur += len; |
352 | enc->cur += len; |
302 | } |
353 | } |
303 | |
354 | |
304 | static void |
355 | ecb_inline void |
305 | encode_strref (enc_t *enc, int utf8, char *str, STRLEN len) |
356 | encode_strref (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len) |
306 | { |
357 | { |
307 | if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS)) |
358 | if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS)) |
308 | { |
359 | { |
309 | SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1); |
360 | SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1); |
310 | |
361 | |
… | |
… | |
321 | sv_setuv (*svp, enc->stringref_idx); |
372 | sv_setuv (*svp, enc->stringref_idx); |
322 | ++enc->stringref_idx; |
373 | ++enc->stringref_idx; |
323 | } |
374 | } |
324 | } |
375 | } |
325 | |
376 | |
326 | encode_str (enc, utf8, str, len); |
377 | encode_str (enc, upgrade_utf8, utf8, str, len); |
327 | } |
378 | } |
328 | |
379 | |
329 | static void encode_sv (enc_t *enc, SV *sv); |
380 | ecb_inline void |
|
|
381 | encode_float16 (enc_t *enc, NV nv) |
|
|
382 | { |
|
|
383 | need (enc, 1+2); |
|
|
384 | |
|
|
385 | *enc->cur++ = MAJOR_MISC | MISC_FLOAT16; |
|
|
386 | |
|
|
387 | uint16_t fp = ecb_float_to_binary16 (nv); |
|
|
388 | |
|
|
389 | if (!ecb_big_endian ()) |
|
|
390 | fp = ecb_bswap16 (fp); |
|
|
391 | |
|
|
392 | memcpy (enc->cur, &fp, 2); |
|
|
393 | enc->cur += 2; |
|
|
394 | } |
|
|
395 | |
|
|
396 | ecb_inline void |
|
|
397 | encode_float32 (enc_t *enc, NV nv) |
|
|
398 | { |
|
|
399 | need (enc, 1+4); |
|
|
400 | |
|
|
401 | *enc->cur++ = MAJOR_MISC | MISC_FLOAT32; |
|
|
402 | |
|
|
403 | uint32_t fp = ecb_float_to_binary32 (nv); |
|
|
404 | |
|
|
405 | if (!ecb_big_endian ()) |
|
|
406 | fp = ecb_bswap32 (fp); |
|
|
407 | |
|
|
408 | memcpy (enc->cur, &fp, 4); |
|
|
409 | enc->cur += 4; |
|
|
410 | } |
|
|
411 | |
|
|
412 | ecb_inline void |
|
|
413 | encode_float64 (enc_t *enc, NV nv) |
|
|
414 | { |
|
|
415 | need (enc, 1+8); |
|
|
416 | |
|
|
417 | *enc->cur++ = MAJOR_MISC | MISC_FLOAT64; |
|
|
418 | |
|
|
419 | uint64_t fp = ecb_double_to_binary64 (nv); |
|
|
420 | |
|
|
421 | if (!ecb_big_endian ()) |
|
|
422 | fp = ecb_bswap64 (fp); |
|
|
423 | |
|
|
424 | memcpy (enc->cur, &fp, 8); |
|
|
425 | enc->cur += 8; |
|
|
426 | } |
|
|
427 | |
|
|
428 | ecb_inline void |
|
|
429 | encode_bool (enc_t *enc, int istrue) |
|
|
430 | { |
|
|
431 | encode_ch (enc, istrue ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE); |
|
|
432 | } |
|
|
433 | |
|
|
434 | // encodes an arrayref containing key-value pairs as CBOR map |
|
|
435 | ecb_inline void |
|
|
436 | encode_array_as_map (enc_t *enc, SV *sv) |
|
|
437 | { |
|
|
438 | if (enc->depth >= enc->cbor.max_depth) |
|
|
439 | croak (ERR_NESTING_EXCEEDED); |
|
|
440 | |
|
|
441 | ++enc->depth; |
|
|
442 | |
|
|
443 | // as_map does error checking for us, but we re-check in case |
|
|
444 | // things have changed. |
|
|
445 | |
|
|
446 | if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) |
|
|
447 | croak ("CBOR::XS::as_map requires an array reference (did you change the array after calling as_map?)"); |
|
|
448 | |
|
|
449 | AV *av = (AV *)SvRV (sv); |
|
|
450 | int i, len = av_len (av); |
|
|
451 | |
|
|
452 | if (!(len & 1)) |
|
|
453 | croak ("CBOR::XS::as_map requires an even number of elements (did you change the array after calling as_map?)"); |
|
|
454 | |
|
|
455 | encode_uint (enc, MAJOR_MAP, (len + 1) >> 1); |
|
|
456 | |
|
|
457 | for (i = 0; i <= len; ++i) |
|
|
458 | { |
|
|
459 | SV **svp = av_fetch (av, i, 0); |
|
|
460 | encode_sv (enc, svp ? *svp : &PL_sv_undef); |
|
|
461 | } |
|
|
462 | |
|
|
463 | --enc->depth; |
|
|
464 | } |
|
|
465 | |
|
|
466 | ecb_inline void |
|
|
467 | encode_forced (enc_t *enc, UV type, SV *sv) |
|
|
468 | { |
|
|
469 | switch (type) |
|
|
470 | { |
|
|
471 | case AS_CBOR: |
|
|
472 | { |
|
|
473 | STRLEN len; |
|
|
474 | char *str = SvPVbyte (sv, len); |
|
|
475 | |
|
|
476 | need (enc, len); |
|
|
477 | memcpy (enc->cur, str, len); |
|
|
478 | enc->cur += len; |
|
|
479 | } |
|
|
480 | break; |
|
|
481 | |
|
|
482 | case AS_BYTES: |
|
|
483 | { |
|
|
484 | STRLEN len; |
|
|
485 | char *str = SvPVbyte (sv, len); |
|
|
486 | encode_strref (enc, 0, 0, str, len); |
|
|
487 | } |
|
|
488 | break; |
|
|
489 | |
|
|
490 | case AS_TEXT: |
|
|
491 | { |
|
|
492 | STRLEN len; |
|
|
493 | char *str = SvPVutf8 (sv, len); |
|
|
494 | encode_strref (enc, 1, 1, str, len); |
|
|
495 | } |
|
|
496 | break; |
|
|
497 | |
|
|
498 | case AS_INT: encode_int (enc, sv); break; |
|
|
499 | |
|
|
500 | case AS_FLOAT16: encode_float16 (enc, SvNV (sv)); break; |
|
|
501 | case AS_FLOAT32: encode_float32 (enc, SvNV (sv)); break; |
|
|
502 | case AS_FLOAT64: encode_float64 (enc, SvNV (sv)); break; |
|
|
503 | |
|
|
504 | case AS_MAP: encode_array_as_map (enc, sv); break; |
|
|
505 | |
|
|
506 | default: |
|
|
507 | croak ("encountered malformed CBOR::XS::Tagged object"); |
|
|
508 | } |
|
|
509 | } |
330 | |
510 | |
331 | static void |
511 | static void |
332 | encode_av (enc_t *enc, AV *av) |
512 | encode_av (enc_t *enc, AV *av) |
333 | { |
513 | { |
334 | int i, len = av_len (av); |
514 | int i, len = av_len (av); |
… | |
… | |
338 | |
518 | |
339 | ++enc->depth; |
519 | ++enc->depth; |
340 | |
520 | |
341 | encode_uint (enc, MAJOR_ARRAY, len + 1); |
521 | encode_uint (enc, MAJOR_ARRAY, len + 1); |
342 | |
522 | |
343 | if (SvMAGICAL (av)) |
523 | if (ecb_expect_false (SvMAGICAL (av))) |
344 | for (i = 0; i <= len; ++i) |
524 | for (i = 0; i <= len; ++i) |
345 | { |
525 | { |
346 | SV **svp = av_fetch (av, i, 0); |
526 | SV **svp = av_fetch (av, i, 0); |
347 | encode_sv (enc, svp ? *svp : &PL_sv_undef); |
527 | encode_sv (enc, svp ? *svp : &PL_sv_undef); |
348 | } |
528 | } |
… | |
… | |
367 | ++enc->depth; |
547 | ++enc->depth; |
368 | |
548 | |
369 | int pairs = hv_iterinit (hv); |
549 | int pairs = hv_iterinit (hv); |
370 | int mg = SvMAGICAL (hv); |
550 | int mg = SvMAGICAL (hv); |
371 | |
551 | |
372 | if (mg) |
552 | if (ecb_expect_false (mg)) |
373 | encode_ch (enc, MAJOR_MAP | MINOR_INDEF); |
553 | encode_ch (enc, MAJOR_MAP | MINOR_INDEF); |
374 | else |
554 | else |
375 | encode_uint (enc, MAJOR_MAP, pairs); |
555 | encode_uint (enc, MAJOR_MAP, pairs); |
376 | |
556 | |
377 | while ((he = hv_iternext (hv))) |
557 | while ((he = hv_iternext (hv))) |
378 | { |
558 | { |
379 | if (HeKLEN (he) == HEf_SVKEY) |
559 | if (HeKLEN (he) == HEf_SVKEY) |
380 | encode_sv (enc, HeSVKEY (he)); |
560 | encode_sv (enc, HeSVKEY (he)); |
381 | else |
561 | else |
382 | encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); |
562 | encode_strref (enc, enc->cbor.flags & (F_TEXT_KEYS | F_TEXT_STRINGS), HeKUTF8 (he), HeKEY (he), HeKLEN (he)); |
383 | |
563 | |
384 | encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); |
564 | encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); |
385 | } |
565 | } |
386 | |
566 | |
387 | if (mg) |
567 | if (ecb_expect_false (mg)) |
388 | encode_ch (enc, MAJOR_MISC | MINOR_INDEF); |
568 | encode_ch (enc, MAJOR_MISC | MINOR_INDEF); |
389 | |
569 | |
390 | --enc->depth; |
570 | --enc->depth; |
391 | } |
571 | } |
392 | |
572 | |
… | |
… | |
412 | |
592 | |
413 | HV *stash = SvSTASH (sv); |
593 | HV *stash = SvSTASH (sv); |
414 | |
594 | |
415 | if (stash == boolean_stash) |
595 | if (stash == boolean_stash) |
416 | { |
596 | { |
417 | encode_ch (enc, SvIV (sv) ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE); |
597 | encode_bool (enc, SvIV (sv)); |
418 | return; |
598 | return; |
419 | } |
599 | } |
420 | else if (stash == error_stash) |
600 | else if (stash == error_stash) |
421 | { |
601 | { |
422 | encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF); |
602 | encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF); |
… | |
… | |
425 | else if (stash == tagged_stash) |
605 | else if (stash == tagged_stash) |
426 | { |
606 | { |
427 | if (svt != SVt_PVAV) |
607 | if (svt != SVt_PVAV) |
428 | croak ("encountered CBOR::XS::Tagged object that isn't an array"); |
608 | croak ("encountered CBOR::XS::Tagged object that isn't an array"); |
429 | |
609 | |
|
|
610 | switch (av_len ((AV *)sv)) |
|
|
611 | { |
|
|
612 | case 2-1: |
|
|
613 | // actually a tagged value |
430 | encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1))); |
614 | encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1))); |
431 | encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); |
615 | encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); |
|
|
616 | break; |
|
|
617 | |
|
|
618 | case 3-1: |
|
|
619 | // a forced type [value, type, undef] |
|
|
620 | encode_forced (enc, SvUV (*av_fetch ((AV *)sv, 1, 1)), *av_fetch ((AV *)sv, 0, 1)); |
|
|
621 | break; |
|
|
622 | |
|
|
623 | default: |
|
|
624 | croak ("encountered malformed CBOR::XS::Tagged object"); |
|
|
625 | } |
432 | |
626 | |
433 | return; |
627 | return; |
434 | } |
628 | } |
435 | } |
629 | } |
436 | |
630 | |
437 | if (ecb_expect_false (SvREFCNT (sv) > 1) |
631 | if (ecb_expect_false (SvREFCNT (sv) > 1) |
438 | && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)) |
632 | && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)) |
439 | { |
633 | { |
440 | if (!enc->shareable) |
634 | if (ecb_expect_false (!enc->shareable)) |
441 | enc->shareable = (HV *)sv_2mortal ((SV *)newHV ()); |
635 | enc->shareable = (HV *)sv_2mortal ((SV *)newHV ()); |
442 | |
636 | |
443 | SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1); |
637 | SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1); |
444 | |
638 | |
445 | if (SvOK (*svp)) |
639 | if (SvOK (*svp)) |
… | |
… | |
459 | if (ecb_expect_false (SvOBJECT (sv))) |
653 | if (ecb_expect_false (SvOBJECT (sv))) |
460 | { |
654 | { |
461 | HV *stash = SvSTASH (sv); |
655 | HV *stash = SvSTASH (sv); |
462 | GV *method; |
656 | GV *method; |
463 | |
657 | |
|
|
658 | if (enc->cbor.flags & F_FORBID_OBJECTS) |
|
|
659 | croak ("encountered object '%s', but forbid_objects is enabled", |
|
|
660 | SvPV_nolen (sv_2mortal (newRV_inc (sv)))); |
464 | if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) |
661 | else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) |
465 | { |
662 | { |
466 | dSP; |
663 | dSP; |
467 | |
664 | |
468 | ENTER; SAVETMPS; |
665 | ENTER; SAVETMPS; |
469 | PUSHMARK (SP); |
666 | PUSHMARK (SP); |
… | |
… | |
488 | else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0) |
685 | else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0) |
489 | { |
686 | { |
490 | dSP; |
687 | dSP; |
491 | |
688 | |
492 | ENTER; SAVETMPS; |
689 | ENTER; SAVETMPS; |
493 | SAVESTACK_POS (); |
|
|
494 | PUSHMARK (SP); |
690 | PUSHMARK (SP); |
495 | EXTEND (SP, 2); |
691 | EXTEND (SP, 2); |
496 | // we re-bless the reference to get overload and other niceties right |
692 | // we re-bless the reference to get overload and other niceties right |
497 | PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); |
693 | PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); |
498 | PUSHs (sv_cbor); |
694 | PUSHs (sv_cbor); |
… | |
… | |
505 | if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) |
701 | if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) |
506 | croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash)); |
702 | croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash)); |
507 | |
703 | |
508 | encode_tag (enc, CBOR_TAG_PERL_OBJECT); |
704 | encode_tag (enc, CBOR_TAG_PERL_OBJECT); |
509 | encode_uint (enc, MAJOR_ARRAY, count + 1); |
705 | encode_uint (enc, MAJOR_ARRAY, count + 1); |
510 | encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); |
706 | encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); |
511 | |
707 | |
512 | while (count) |
708 | { |
|
|
709 | int i; |
|
|
710 | |
|
|
711 | for (i = 0; i < count; ++i) |
513 | encode_sv (enc, SP[1 - count--]); |
712 | encode_sv (enc, SP[i + 1 - count]); |
|
|
713 | |
|
|
714 | SP -= count; |
|
|
715 | } |
514 | |
716 | |
515 | PUTBACK; |
717 | PUTBACK; |
516 | |
718 | |
517 | FREETMPS; LEAVE; |
719 | FREETMPS; LEAVE; |
518 | } |
720 | } |
… | |
… | |
540 | |
742 | |
541 | if (ecb_expect_false (nv == (NV)(U32)nv)) |
743 | if (ecb_expect_false (nv == (NV)(U32)nv)) |
542 | encode_uint (enc, MAJOR_POS_INT, (U32)nv); |
744 | encode_uint (enc, MAJOR_POS_INT, (U32)nv); |
543 | //TODO: maybe I32? |
745 | //TODO: maybe I32? |
544 | else if (ecb_expect_false (nv == (float)nv)) |
746 | else if (ecb_expect_false (nv == (float)nv)) |
545 | { |
747 | encode_float32 (enc, nv); |
546 | uint32_t fp = ecb_float_to_binary32 (nv); |
|
|
547 | |
|
|
548 | *enc->cur++ = MAJOR_MISC | MISC_FLOAT32; |
|
|
549 | |
|
|
550 | if (!ecb_big_endian ()) |
|
|
551 | fp = ecb_bswap32 (fp); |
|
|
552 | |
|
|
553 | memcpy (enc->cur, &fp, 4); |
|
|
554 | enc->cur += 4; |
|
|
555 | } |
|
|
556 | else |
748 | else |
557 | { |
749 | encode_float64 (enc, nv); |
558 | uint64_t fp = ecb_double_to_binary64 (nv); |
|
|
559 | |
|
|
560 | *enc->cur++ = MAJOR_MISC | MISC_FLOAT64; |
|
|
561 | |
|
|
562 | if (!ecb_big_endian ()) |
|
|
563 | fp = ecb_bswap64 (fp); |
|
|
564 | |
|
|
565 | memcpy (enc->cur, &fp, 8); |
|
|
566 | enc->cur += 8; |
|
|
567 | } |
|
|
568 | } |
750 | } |
569 | |
751 | |
570 | static void |
752 | static void |
571 | encode_sv (enc_t *enc, SV *sv) |
753 | encode_sv (enc_t *enc, SV *sv) |
572 | { |
754 | { |
… | |
… | |
574 | |
756 | |
575 | if (SvPOKp (sv)) |
757 | if (SvPOKp (sv)) |
576 | { |
758 | { |
577 | STRLEN len; |
759 | STRLEN len; |
578 | char *str = SvPV (sv, len); |
760 | char *str = SvPV (sv, len); |
579 | encode_strref (enc, SvUTF8 (sv), str, len); |
761 | encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len); |
580 | } |
762 | } |
581 | else if (SvNOKp (sv)) |
763 | else if (SvNOKp (sv)) |
582 | encode_nv (enc, sv); |
764 | encode_nv (enc, sv); |
583 | else if (SvIOKp (sv)) |
765 | else if (SvIOKp (sv)) |
584 | { |
766 | encode_int (enc, sv); |
585 | if (SvIsUV (sv)) |
|
|
586 | encode_uint (enc, MAJOR_POS_INT, SvUVX (sv)); |
|
|
587 | else if (SvIVX (sv) >= 0) |
|
|
588 | encode_uint (enc, MAJOR_POS_INT, SvIVX (sv)); |
|
|
589 | else |
|
|
590 | encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1)); |
|
|
591 | } |
|
|
592 | else if (SvROK (sv)) |
767 | else if (SvROK (sv)) |
593 | encode_rv (enc, SvRV (sv)); |
768 | encode_rv (enc, SvRV (sv)); |
594 | else if (!SvOK (sv)) |
769 | else if (!SvOK (sv)) |
595 | encode_ch (enc, MAJOR_MISC | SIMPLE_NULL); |
770 | encode_ch (enc, MAJOR_MISC | SIMPLE_NULL); |
596 | else if (enc->cbor.flags & F_ALLOW_UNKNOWN) |
771 | else if (enc->cbor.flags & F_ALLOW_UNKNOWN) |
… | |
… | |
603 | static SV * |
778 | static SV * |
604 | encode_cbor (SV *scalar, CBOR *cbor) |
779 | encode_cbor (SV *scalar, CBOR *cbor) |
605 | { |
780 | { |
606 | enc_t enc = { 0 }; |
781 | enc_t enc = { 0 }; |
607 | |
782 | |
608 | enc.cbor = *cbor; |
783 | enc.cbor = *cbor; |
609 | enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); |
784 | enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); |
610 | enc.cur = SvPVX (enc.sv); |
785 | enc.cur = SvPVX (enc.sv); |
611 | enc.end = SvEND (enc.sv); |
786 | enc.end = SvEND (enc.sv); |
612 | |
787 | |
613 | SvPOK_only (enc.sv); |
788 | SvPOK_only (enc.sv); |
614 | |
789 | |
615 | if (cbor->flags & F_PACK_STRINGS) |
790 | if (cbor->flags & F_PACK_STRINGS) |
616 | { |
791 | { |
… | |
… | |
643 | U32 depth; // recursion depth |
818 | U32 depth; // recursion depth |
644 | U32 maxdepth; // recursion depth limit |
819 | U32 maxdepth; // recursion depth limit |
645 | AV *shareable; |
820 | AV *shareable; |
646 | AV *stringref; |
821 | AV *stringref; |
647 | SV *decode_tagged; |
822 | SV *decode_tagged; |
|
|
823 | SV *err_sv; // optional sv for error, needs to be freed |
648 | } dec_t; |
824 | } dec_t; |
649 | |
825 | |
650 | #define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE |
826 | // set dec->err to ERRSV |
|
|
827 | ecb_cold static void |
|
|
828 | err_errsv (dec_t *dec) |
|
|
829 | { |
|
|
830 | if (!dec->err) |
|
|
831 | { |
|
|
832 | dec->err_sv = newSVsv (ERRSV); |
651 | |
833 | |
652 | #define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data") |
834 | // chop off the trailing \n |
|
|
835 | SvCUR_set (dec->err_sv, SvCUR (dec->err_sv) - 1); |
|
|
836 | *SvEND (dec->err_sv) = 0; |
653 | |
837 | |
|
|
838 | dec->err = SvPVutf8_nolen (dec->err_sv); |
|
|
839 | } |
|
|
840 | } |
|
|
841 | |
|
|
842 | // the following functions are used to reduce code size and help the compiler to optimise |
|
|
843 | ecb_cold static void |
|
|
844 | err_set (dec_t *dec, const char *reason) |
|
|
845 | { |
|
|
846 | if (!dec->err) |
|
|
847 | dec->err = reason; |
|
|
848 | } |
|
|
849 | |
|
|
850 | ecb_cold static void |
|
|
851 | err_unexpected_end (dec_t *dec) |
|
|
852 | { |
|
|
853 | err_set (dec, "unexpected end of CBOR data"); |
|
|
854 | } |
|
|
855 | |
|
|
856 | #define ERR_DO(do) SB do; goto fail; SE |
|
|
857 | #define ERR(reason) ERR_DO (err_set (dec, reason)) |
|
|
858 | #define ERR_ERRSV ERR_DO (err_errsv (dec)) |
|
|
859 | |
|
|
860 | #define WANT(len) if (ecb_expect_false ((uintptr_t)(dec->end - dec->cur) < (STRLEN)len)) ERR_DO (err_unexpected_end (dec)) |
|
|
861 | |
654 | #define DEC_INC_DEPTH if (++dec->depth > dec->cbor.max_depth) ERR (ERR_NESTING_EXCEEDED) |
862 | #define DEC_INC_DEPTH if (ecb_expect_false (++dec->depth > dec->cbor.max_depth)) ERR (ERR_NESTING_EXCEEDED) |
655 | #define DEC_DEC_DEPTH --dec->depth |
863 | #define DEC_DEC_DEPTH --dec->depth |
656 | |
864 | |
657 | static UV |
865 | static UV |
658 | decode_uint (dec_t *dec) |
866 | decode_uint (dec_t *dec) |
659 | { |
867 | { |
… | |
… | |
725 | |
933 | |
726 | for (;;) |
934 | for (;;) |
727 | { |
935 | { |
728 | WANT (1); |
936 | WANT (1); |
729 | |
937 | |
730 | if (*dec->cur == (MAJOR_MISC | MINOR_INDEF)) |
938 | if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err) |
731 | { |
939 | { |
732 | ++dec->cur; |
940 | ++dec->cur; |
733 | break; |
941 | break; |
734 | } |
942 | } |
735 | |
943 | |
736 | av_push (av, decode_sv (dec)); |
944 | av_push (av, decode_sv (dec)); |
737 | } |
945 | } |
738 | } |
946 | } |
739 | else |
947 | else |
740 | { |
948 | { |
741 | int i, len = decode_uint (dec); |
949 | UV i, len = decode_uint (dec); |
742 | |
950 | |
743 | WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays |
951 | WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays |
744 | av_fill (av, len - 1); |
952 | av_fill (av, len - 1); |
745 | |
953 | |
746 | for (i = 0; i < len; ++i) |
954 | for (i = 0; i < len; ++i) |
… | |
… | |
749 | |
957 | |
750 | DEC_DEC_DEPTH; |
958 | DEC_DEC_DEPTH; |
751 | return newRV_noinc ((SV *)av); |
959 | return newRV_noinc ((SV *)av); |
752 | |
960 | |
753 | fail: |
961 | fail: |
754 | SvREFCNT_dec (av); |
962 | SvREFCNT_dec_NN (av); |
755 | DEC_DEC_DEPTH; |
963 | DEC_DEC_DEPTH; |
756 | return &PL_sv_undef; |
964 | return &PL_sv_undef; |
757 | } |
965 | } |
758 | |
966 | |
759 | static void |
967 | static void |
… | |
… | |
763 | // byte or utf-8 strings as keys, but only when !stringref |
971 | // byte or utf-8 strings as keys, but only when !stringref |
764 | |
972 | |
765 | if (ecb_expect_true (!dec->stringref)) |
973 | if (ecb_expect_true (!dec->stringref)) |
766 | if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8)) |
974 | if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8)) |
767 | { |
975 | { |
768 | I32 len = decode_uint (dec); |
976 | STRLEN len = decode_uint (dec); |
769 | char *key = (char *)dec->cur; |
977 | char *key = (char *)dec->cur; |
770 | |
978 | |
771 | WANT (len); |
979 | WANT (len); |
772 | dec->cur += len; |
980 | dec->cur += len; |
773 | |
981 | |
… | |
… | |
775 | |
983 | |
776 | return; |
984 | return; |
777 | } |
985 | } |
778 | else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8)) |
986 | else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8)) |
779 | { |
987 | { |
780 | I32 len = decode_uint (dec); |
988 | STRLEN len = decode_uint (dec); |
781 | char *key = (char *)dec->cur; |
989 | char *key = (char *)dec->cur; |
782 | |
990 | |
783 | WANT (len); |
991 | WANT (len); |
784 | dec->cur += len; |
992 | dec->cur += len; |
785 | |
993 | |
786 | if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8)) |
994 | if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8)) |
787 | if (!is_utf8_string (key, len)) |
995 | if (!cbor_is_utf8_string ((U8 *)key, len)) |
788 | ERR ("corrupted CBOR data (invalid UTF-8 in map key)"); |
996 | ERR ("corrupted CBOR data (invalid UTF-8 in map key)"); |
789 | |
997 | |
790 | hv_store (hv, key, -len, decode_sv (dec), 0); |
998 | hv_store (hv, key, -len, decode_sv (dec), 0); |
791 | |
999 | |
792 | return; |
1000 | return; |
793 | } |
1001 | } |
794 | |
1002 | |
795 | SV *k = decode_sv (dec); |
1003 | SV *k = decode_sv (dec); |
796 | SV *v = decode_sv (dec); |
1004 | SV *v = decode_sv (dec); |
797 | |
1005 | |
|
|
1006 | // we leak memory if uncaught exceptions are thrown by random magical |
|
|
1007 | // methods, and this is hopefully the only place where it can happen, |
|
|
1008 | // so if there is a chance of an exception, take the very slow path. |
|
|
1009 | // since catching exceptions is "undocumented/internal/forbidden" by |
|
|
1010 | // the new p5p powers, we need to call out to a perl function :/ |
|
|
1011 | if (ecb_expect_false (SvAMAGIC (k))) |
|
|
1012 | { |
|
|
1013 | dSP; |
|
|
1014 | |
|
|
1015 | ENTER; SAVETMPS; |
|
|
1016 | PUSHMARK (SP); |
|
|
1017 | EXTEND (SP, 3); |
|
|
1018 | PUSHs (sv_2mortal (newRV_inc ((SV *)hv))); |
|
|
1019 | PUSHs (sv_2mortal (k)); |
|
|
1020 | PUSHs (sv_2mortal (v)); |
|
|
1021 | |
|
|
1022 | PUTBACK; |
|
|
1023 | call_pv ("CBOR::XS::_hv_store", G_VOID | G_DISCARD | G_EVAL); |
|
|
1024 | SPAGAIN; |
|
|
1025 | |
|
|
1026 | FREETMPS; LEAVE; |
|
|
1027 | |
|
|
1028 | if (SvTRUE (ERRSV)) |
|
|
1029 | ERR_ERRSV; |
|
|
1030 | |
|
|
1031 | return; |
|
|
1032 | } |
|
|
1033 | |
798 | hv_store_ent (hv, k, v, 0); |
1034 | hv_store_ent (hv, k, v, 0); |
799 | SvREFCNT_dec (k); |
1035 | SvREFCNT_dec_NN (k); |
800 | |
1036 | |
801 | fail: |
1037 | fail: |
802 | ; |
1038 | ; |
803 | } |
1039 | } |
804 | |
1040 | |
… | |
… | |
815 | |
1051 | |
816 | for (;;) |
1052 | for (;;) |
817 | { |
1053 | { |
818 | WANT (1); |
1054 | WANT (1); |
819 | |
1055 | |
820 | if (*dec->cur == (MAJOR_MISC | MINOR_INDEF)) |
1056 | if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err) |
821 | { |
1057 | { |
822 | ++dec->cur; |
1058 | ++dec->cur; |
823 | break; |
1059 | break; |
824 | } |
1060 | } |
825 | |
1061 | |
826 | decode_he (dec, hv); |
1062 | decode_he (dec, hv); |
827 | } |
1063 | } |
828 | } |
1064 | } |
829 | else |
1065 | else |
830 | { |
1066 | { |
831 | int pairs = decode_uint (dec); |
1067 | UV pairs = decode_uint (dec); |
|
|
1068 | |
|
|
1069 | WANT (pairs); // complexity check - need at least one byte per value, do not allow supersize hashes |
832 | |
1070 | |
833 | while (pairs--) |
1071 | while (pairs--) |
834 | decode_he (dec, hv); |
1072 | decode_he (dec, hv); |
835 | } |
1073 | } |
836 | |
1074 | |
837 | DEC_DEC_DEPTH; |
1075 | DEC_DEC_DEPTH; |
838 | return newRV_noinc ((SV *)hv); |
1076 | return newRV_noinc ((SV *)hv); |
839 | |
1077 | |
840 | fail: |
1078 | fail: |
841 | SvREFCNT_dec (hv); |
1079 | SvREFCNT_dec_NN (hv); |
842 | DEC_DEC_DEPTH; |
1080 | DEC_DEC_DEPTH; |
843 | return &PL_sv_undef; |
1081 | return &PL_sv_undef; |
844 | } |
1082 | } |
845 | |
1083 | |
846 | static SV * |
1084 | static SV * |
847 | decode_str (dec_t *dec, int utf8) |
1085 | decode_str (dec_t *dec, int utf8) |
848 | { |
1086 | { |
849 | SV *sv = 0; |
1087 | SV *sv = 0; |
850 | |
1088 | |
851 | if ((*dec->cur & MINOR_MASK) == MINOR_INDEF) |
1089 | if (ecb_expect_false ((*dec->cur & MINOR_MASK) == MINOR_INDEF)) |
852 | { |
1090 | { |
853 | // indefinite length strings |
1091 | // indefinite length strings |
854 | ++dec->cur; |
1092 | ++dec->cur; |
855 | |
1093 | |
856 | U8 major = *dec->cur & MAJOR_MISC; |
1094 | U8 major = *dec->cur & MAJOR_MISC; |
… | |
… | |
891 | } |
1129 | } |
892 | |
1130 | |
893 | if (utf8) |
1131 | if (utf8) |
894 | { |
1132 | { |
895 | if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8)) |
1133 | if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8)) |
896 | if (!is_utf8_string (SvPVX (sv), SvCUR (sv))) |
1134 | if (!cbor_is_utf8_string (SvPVX (sv), SvCUR (sv))) |
897 | ERR ("corrupted CBOR data (invalid UTF-8 in text string)"); |
1135 | ERR ("corrupted CBOR data (invalid UTF-8 in text string)"); |
898 | |
1136 | |
899 | SvUTF8_on (sv); |
1137 | SvUTF8_on (sv); |
900 | } |
1138 | } |
901 | |
1139 | |
… | |
… | |
924 | sv = newRV_noinc (decode_sv (dec)); |
1162 | sv = newRV_noinc (decode_sv (dec)); |
925 | break; |
1163 | break; |
926 | |
1164 | |
927 | case CBOR_TAG_STRINGREF_NAMESPACE: |
1165 | case CBOR_TAG_STRINGREF_NAMESPACE: |
928 | { |
1166 | { |
|
|
1167 | // do not use SAVETMPS/FREETMPS, as these will |
|
|
1168 | // erase mortalised caches, e.g. "shareable" |
929 | ENTER; SAVETMPS; |
1169 | ENTER; |
930 | |
1170 | |
931 | SAVESPTR (dec->stringref); |
1171 | SAVESPTR (dec->stringref); |
932 | dec->stringref = (AV *)sv_2mortal ((SV *)newAV ()); |
1172 | dec->stringref = (AV *)sv_2mortal ((SV *)newAV ()); |
933 | |
1173 | |
934 | sv = decode_sv (dec); |
1174 | sv = decode_sv (dec); |
935 | |
1175 | |
936 | FREETMPS; LEAVE; |
1176 | LEAVE; |
937 | } |
1177 | } |
938 | break; |
1178 | break; |
939 | |
1179 | |
940 | case CBOR_TAG_STRINGREF: |
1180 | case CBOR_TAG_STRINGREF: |
941 | { |
1181 | { |
942 | if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT)) |
1182 | if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT)) |
943 | ERR ("corrupted CBOR data (stringref index not an unsigned integer)"); |
1183 | ERR ("corrupted CBOR data (stringref index not an unsigned integer)"); |
944 | |
1184 | |
945 | UV idx = decode_uint (dec); |
1185 | UV idx = decode_uint (dec); |
946 | |
1186 | |
947 | if (!dec->stringref || (int)idx > AvFILLp (dec->stringref)) |
1187 | if (!dec->stringref || idx >= (UV)(1 + AvFILLp (dec->stringref))) |
948 | ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)"); |
1188 | ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)"); |
949 | |
1189 | |
950 | sv = newSVsv (AvARRAY (dec->stringref)[idx]); |
1190 | sv = newSVsv (AvARRAY (dec->stringref)[idx]); |
951 | } |
1191 | } |
952 | break; |
1192 | break; |
… | |
… | |
956 | if (ecb_expect_false (!dec->shareable)) |
1196 | if (ecb_expect_false (!dec->shareable)) |
957 | dec->shareable = (AV *)sv_2mortal ((SV *)newAV ()); |
1197 | dec->shareable = (AV *)sv_2mortal ((SV *)newAV ()); |
958 | |
1198 | |
959 | if (dec->cbor.flags & F_ALLOW_CYCLES) |
1199 | if (dec->cbor.flags & F_ALLOW_CYCLES) |
960 | { |
1200 | { |
|
|
1201 | // if cycles are allowed, then we store an AV as value |
|
|
1202 | // while it is being decoded, and gather unresolved |
|
|
1203 | // references in it, to be re4solved after decoding. |
|
|
1204 | int idx, i; |
961 | sv = newSV (0); |
1205 | AV *av = newAV (); |
962 | av_push (dec->shareable, SvREFCNT_inc_NN (sv)); |
1206 | av_push (dec->shareable, (SV *)av); |
|
|
1207 | idx = AvFILLp (dec->shareable); |
963 | |
1208 | |
964 | SV *osv = decode_sv (dec); |
1209 | sv = decode_sv (dec); |
965 | sv_setsv (sv, osv); |
1210 | |
|
|
1211 | // the AV now contains \undef for all unresolved references, |
|
|
1212 | // so we fix them up here. |
|
|
1213 | for (i = 0; i <= AvFILLp (av); ++i) |
|
|
1214 | SvRV_set (AvARRAY (av)[i], SvREFCNT_inc_NN (SvRV (sv))); |
|
|
1215 | |
|
|
1216 | // now replace the AV by a reference to the completed value |
966 | SvREFCNT_dec_NN (osv); |
1217 | SvREFCNT_dec_NN ((SV *)av); |
|
|
1218 | AvARRAY (dec->shareable)[idx] = SvREFCNT_inc_NN (sv); |
967 | } |
1219 | } |
968 | else |
1220 | else |
969 | { |
1221 | { |
970 | av_push (dec->shareable, &PL_sv_undef); |
1222 | av_push (dec->shareable, &PL_sv_undef); |
971 | int idx = AvFILLp (dec->shareable); |
1223 | int idx = AvFILLp (dec->shareable); |
972 | sv = decode_sv (dec); |
1224 | sv = decode_sv (dec); |
973 | av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv)); |
1225 | AvARRAY (dec->shareable)[idx] = SvREFCNT_inc_NN (sv); |
974 | } |
1226 | } |
975 | } |
1227 | } |
976 | break; |
1228 | break; |
977 | |
1229 | |
978 | case CBOR_TAG_VALUE_SHAREDREF: |
1230 | case CBOR_TAG_VALUE_SHAREDREF: |
… | |
… | |
980 | if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT)) |
1232 | if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT)) |
981 | ERR ("corrupted CBOR data (sharedref index not an unsigned integer)"); |
1233 | ERR ("corrupted CBOR data (sharedref index not an unsigned integer)"); |
982 | |
1234 | |
983 | UV idx = decode_uint (dec); |
1235 | UV idx = decode_uint (dec); |
984 | |
1236 | |
985 | if (!dec->shareable || (int)idx > AvFILLp (dec->shareable)) |
1237 | if (!dec->shareable || idx >= (UV)(1 + AvFILLp (dec->shareable))) |
986 | ERR ("corrupted CBOR data (sharedref index out of bounds)"); |
1238 | ERR ("corrupted CBOR data (sharedref index out of bounds)"); |
987 | |
1239 | |
988 | sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]); |
1240 | sv = AvARRAY (dec->shareable)[idx]; |
989 | |
1241 | |
990 | if (sv == &PL_sv_undef) |
1242 | // reference to cycle, we create a new \undef and use that, and also |
|
|
1243 | // registerr it in the AV for later fixing |
|
|
1244 | if (SvTYPE (sv) == SVt_PVAV) |
|
|
1245 | { |
|
|
1246 | AV *av = (AV *)sv; |
|
|
1247 | sv = newRV_noinc (&PL_sv_undef); |
|
|
1248 | av_push (av, SvREFCNT_inc_NN (sv)); |
|
|
1249 | } |
|
|
1250 | else if (sv == &PL_sv_undef) // not yet decoded, but cycles not allowed |
991 | ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled"); |
1251 | ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled"); |
|
|
1252 | else // we decoded the object earlier, no cycle |
|
|
1253 | sv = newSVsv (sv); |
992 | } |
1254 | } |
993 | break; |
1255 | break; |
994 | |
1256 | |
995 | case CBOR_TAG_PERL_OBJECT: |
1257 | case CBOR_TAG_PERL_OBJECT: |
996 | { |
1258 | { |
|
|
1259 | if (dec->cbor.flags & F_FORBID_OBJECTS) |
|
|
1260 | goto filter; |
|
|
1261 | |
997 | sv = decode_sv (dec); |
1262 | sv = decode_sv (dec); |
998 | |
1263 | |
999 | if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) |
1264 | if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) |
1000 | ERR ("corrupted CBOR data (non-array perl object)"); |
1265 | ERR ("corrupted CBOR data (non-array perl object)"); |
1001 | |
1266 | |
… | |
… | |
1030 | SPAGAIN; |
1295 | SPAGAIN; |
1031 | |
1296 | |
1032 | if (SvTRUE (ERRSV)) |
1297 | if (SvTRUE (ERRSV)) |
1033 | { |
1298 | { |
1034 | FREETMPS; LEAVE; |
1299 | FREETMPS; LEAVE; |
1035 | ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV)))); |
1300 | ERR_ERRSV; |
1036 | } |
1301 | } |
1037 | |
1302 | |
1038 | SvREFCNT_dec (sv); |
1303 | SvREFCNT_dec_NN (sv); |
1039 | sv = SvREFCNT_inc (POPs); |
1304 | sv = SvREFCNT_inc (POPs); |
1040 | |
1305 | |
1041 | PUTBACK; |
1306 | PUTBACK; |
1042 | |
1307 | |
1043 | FREETMPS; LEAVE; |
1308 | FREETMPS; LEAVE; |
1044 | } |
1309 | } |
1045 | break; |
1310 | break; |
1046 | |
1311 | |
1047 | default: |
1312 | default: |
|
|
1313 | filter: |
1048 | { |
1314 | { |
|
|
1315 | SV *tag_sv = newSVuv (tag); |
|
|
1316 | |
1049 | sv = decode_sv (dec); |
1317 | sv = decode_sv (dec); |
1050 | |
1318 | |
1051 | dSP; |
1319 | dSP; |
1052 | ENTER; SAVETMPS; |
1320 | ENTER; SAVETMPS; |
1053 | SAVESTACK_POS (); |
|
|
1054 | PUSHMARK (SP); |
1321 | PUSHMARK (SP); |
1055 | EXTEND (SP, 2); |
1322 | EXTEND (SP, 2); |
1056 | PUSHs (newSVuv (tag)); |
1323 | PUSHs (tag_sv); |
1057 | PUSHs (sv); |
1324 | PUSHs (sv); |
1058 | |
1325 | |
1059 | PUTBACK; |
1326 | PUTBACK; |
1060 | int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL); |
1327 | int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL); |
1061 | SPAGAIN; |
1328 | SPAGAIN; |
1062 | |
1329 | |
1063 | if (SvTRUE (ERRSV)) |
1330 | if (SvTRUE (ERRSV)) |
1064 | { |
1331 | { |
|
|
1332 | SvREFCNT_dec_NN (tag_sv); |
1065 | FREETMPS; LEAVE; |
1333 | FREETMPS; LEAVE; |
1066 | ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV)))); |
1334 | ERR_ERRSV; |
1067 | } |
1335 | } |
1068 | |
1336 | |
1069 | if (count) |
1337 | if (count) |
1070 | { |
1338 | { |
|
|
1339 | SvREFCNT_dec_NN (tag_sv); |
1071 | SvREFCNT_dec (sv); |
1340 | SvREFCNT_dec_NN (sv); |
1072 | sv = SvREFCNT_inc (POPs); |
1341 | sv = SvREFCNT_inc_NN (TOPs); |
|
|
1342 | SP -= count; |
1073 | } |
1343 | } |
1074 | else |
1344 | else |
1075 | { |
1345 | { |
1076 | AV *av = newAV (); |
1346 | AV *av = newAV (); |
1077 | av_push (av, newSVuv (tag)); |
1347 | av_push (av, tag_sv); |
1078 | av_push (av, sv); |
1348 | av_push (av, sv); |
1079 | |
1349 | |
1080 | HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash |
1350 | HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash |
1081 | ? cbor_tagged_stash |
1351 | ? cbor_tagged_stash |
1082 | : gv_stashpv ("CBOR::XS::Tagged" , 1); |
1352 | : gv_stashpv ("CBOR::XS::Tagged" , 1); |
… | |
… | |
1211 | |
1481 | |
1212 | if (dec.err) |
1482 | if (dec.err) |
1213 | { |
1483 | { |
1214 | if (dec.shareable) |
1484 | if (dec.shareable) |
1215 | { |
1485 | { |
1216 | // need to break cyclic links, which whould all be in shareable |
1486 | // need to break cyclic links, which would all be in shareable |
1217 | int i; |
1487 | int i; |
1218 | SV **svp; |
1488 | SV **svp; |
1219 | |
1489 | |
1220 | for (i = av_len (dec.shareable) + 1; i--; ) |
1490 | for (i = av_len (dec.shareable) + 1; i--; ) |
1221 | if ((svp = av_fetch (dec.shareable, i, 0))) |
1491 | if ((svp = av_fetch (dec.shareable, i, 0))) |
1222 | sv_setsv (*svp, &PL_sv_undef); |
1492 | sv_setsv (*svp, &PL_sv_undef); |
1223 | } |
1493 | } |
1224 | |
1494 | |
1225 | SvREFCNT_dec (sv); |
1495 | SvREFCNT_dec_NN (sv); |
|
|
1496 | |
|
|
1497 | if (dec.err_sv) |
|
|
1498 | sv_2mortal (dec.err_sv); |
|
|
1499 | |
1226 | croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur); |
1500 | croak ("%s, at offset %ld (octet 0x%02x)", dec.err, (long)(dec.cur - (U8 *)data), (int)(uint8_t)*dec.cur); |
1227 | } |
1501 | } |
1228 | |
1502 | |
1229 | sv = sv_2mortal (sv); |
1503 | sv = sv_2mortal (sv); |
1230 | |
1504 | |
1231 | return sv; |
1505 | return sv; |
… | |
… | |
1320 | |
1594 | |
1321 | break; |
1595 | break; |
1322 | |
1596 | |
1323 | case MAJOR_MAP >> MAJOR_SHIFT: |
1597 | case MAJOR_MAP >> MAJOR_SHIFT: |
1324 | len <<= 1; |
1598 | len <<= 1; |
|
|
1599 | /* FALLTHROUGH */ |
1325 | case MAJOR_ARRAY >> MAJOR_SHIFT: |
1600 | case MAJOR_ARRAY >> MAJOR_SHIFT: |
1326 | if (len) |
1601 | if (len) |
1327 | { |
1602 | { |
1328 | av_push (self->incr_count, newSViv (len + 1)); //TODO: nest |
1603 | av_push (self->incr_count, newSViv (len + 1)); //TODO: nest |
1329 | count = len + 1; |
1604 | count = len + 1; |
… | |
… | |
1374 | |
1649 | |
1375 | default_filter = newSVpv ("CBOR::XS::default_filter", 0); |
1650 | default_filter = newSVpv ("CBOR::XS::default_filter", 0); |
1376 | |
1651 | |
1377 | sv_cbor = newSVpv ("CBOR", 0); |
1652 | sv_cbor = newSVpv ("CBOR", 0); |
1378 | SvREADONLY_on (sv_cbor); |
1653 | SvREADONLY_on (sv_cbor); |
|
|
1654 | |
|
|
1655 | assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1)); |
1379 | } |
1656 | } |
1380 | |
1657 | |
1381 | PROTOTYPES: DISABLE |
1658 | PROTOTYPES: DISABLE |
1382 | |
1659 | |
1383 | void CLONE (...) |
1660 | void CLONE (...) |
… | |
… | |
1403 | ALIAS: |
1680 | ALIAS: |
1404 | shrink = F_SHRINK |
1681 | shrink = F_SHRINK |
1405 | allow_unknown = F_ALLOW_UNKNOWN |
1682 | allow_unknown = F_ALLOW_UNKNOWN |
1406 | allow_sharing = F_ALLOW_SHARING |
1683 | allow_sharing = F_ALLOW_SHARING |
1407 | allow_cycles = F_ALLOW_CYCLES |
1684 | allow_cycles = F_ALLOW_CYCLES |
|
|
1685 | forbid_objects = F_FORBID_OBJECTS |
1408 | pack_strings = F_PACK_STRINGS |
1686 | pack_strings = F_PACK_STRINGS |
|
|
1687 | text_keys = F_TEXT_KEYS |
1409 | utf8_strings = F_UTF8_STRINGS |
1688 | text_strings = F_TEXT_STRINGS |
1410 | validate_utf8 = F_VALIDATE_UTF8 |
1689 | validate_utf8 = F_VALIDATE_UTF8 |
1411 | PPCODE: |
1690 | PPCODE: |
1412 | { |
1691 | { |
1413 | if (enable) |
1692 | if (enable) |
1414 | self->flags |= ix; |
1693 | self->flags |= ix; |
… | |
… | |
1422 | ALIAS: |
1701 | ALIAS: |
1423 | get_shrink = F_SHRINK |
1702 | get_shrink = F_SHRINK |
1424 | get_allow_unknown = F_ALLOW_UNKNOWN |
1703 | get_allow_unknown = F_ALLOW_UNKNOWN |
1425 | get_allow_sharing = F_ALLOW_SHARING |
1704 | get_allow_sharing = F_ALLOW_SHARING |
1426 | get_allow_cycles = F_ALLOW_CYCLES |
1705 | get_allow_cycles = F_ALLOW_CYCLES |
|
|
1706 | get_forbid_objects = F_FORBID_OBJECTS |
1427 | get_pack_strings = F_PACK_STRINGS |
1707 | get_pack_strings = F_PACK_STRINGS |
|
|
1708 | get_text_keys = F_TEXT_KEYS |
|
|
1709 | get_text_strings = F_TEXT_STRINGS |
1428 | get_validate_utf8 = F_VALIDATE_UTF8 |
1710 | get_validate_utf8 = F_VALIDATE_UTF8 |
1429 | PPCODE: |
1711 | PPCODE: |
1430 | XPUSHs (boolSV (self->flags & ix)); |
1712 | XPUSHs (boolSV (self->flags & ix)); |
1431 | |
1713 | |
1432 | void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) |
1714 | void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) |
… | |
… | |
1562 | cbor_init (&cbor); |
1844 | cbor_init (&cbor); |
1563 | PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN; |
1845 | PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN; |
1564 | XPUSHs (cborstr); |
1846 | XPUSHs (cborstr); |
1565 | } |
1847 | } |
1566 | |
1848 | |
|
|
1849 | #ifdef __AFL_COMPILER |
|
|
1850 | |
|
|
1851 | void |
|
|
1852 | afl_init () |
|
|
1853 | CODE: |
|
|
1854 | __AFL_INIT (); |
|
|
1855 | |
|
|
1856 | int |
|
|
1857 | afl_loop (unsigned int count = 10000) |
|
|
1858 | CODE: |
|
|
1859 | RETVAL = __AFL_LOOP (count); |
|
|
1860 | OUTPUT: |
|
|
1861 | RETVAL |
|
|
1862 | |
|
|
1863 | #endif |
|
|
1864 | |