… | |
… | |
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 | // possibly future enhancements: (generic) float, (generic) string |
|
|
115 | }; |
|
|
116 | |
99 | #define F_SHRINK 0x00000001UL |
117 | #define F_SHRINK 0x00000001UL |
100 | #define F_ALLOW_UNKNOWN 0x00000002UL |
118 | #define F_ALLOW_UNKNOWN 0x00000002UL |
101 | #define F_ALLOW_SHARING 0x00000004UL |
119 | #define F_ALLOW_SHARING 0x00000004UL |
102 | #define F_ALLOW_CYCLES 0x00000008UL |
120 | #define F_ALLOW_CYCLES 0x00000008UL |
|
|
121 | #define F_FORBID_OBJECTS 0x00000010UL |
103 | #define F_PACK_STRINGS 0x00000010UL |
122 | #define F_PACK_STRINGS 0x00000020UL |
|
|
123 | #define F_TEXT_KEYS 0x00000040UL |
|
|
124 | #define F_TEXT_STRINGS 0x00000080UL |
|
|
125 | #define F_VALIDATE_UTF8 0x00000100UL |
104 | |
126 | |
105 | #define INIT_SIZE 32 // initial scalar size to be allocated |
127 | #define INIT_SIZE 32 // initial scalar size to be allocated |
106 | |
128 | |
107 | #define SB do { |
129 | #define SB do { |
108 | #define SE } while (0) |
130 | #define SE } while (0) |
… | |
… | |
127 | typedef struct { |
149 | typedef struct { |
128 | U32 flags; |
150 | U32 flags; |
129 | U32 max_depth; |
151 | U32 max_depth; |
130 | STRLEN max_size; |
152 | STRLEN max_size; |
131 | SV *filter; |
153 | SV *filter; |
|
|
154 | |
|
|
155 | // for the incremental parser |
|
|
156 | STRLEN incr_pos; // the current offset into the text |
|
|
157 | STRLEN incr_need; // minimum bytes needed to decode |
|
|
158 | AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef. |
132 | } CBOR; |
159 | } CBOR; |
133 | |
160 | |
134 | ecb_inline void |
161 | ecb_inline void |
135 | cbor_init (CBOR *cbor) |
162 | cbor_init (CBOR *cbor) |
136 | { |
163 | { |
… | |
… | |
140 | |
167 | |
141 | ecb_inline void |
168 | ecb_inline void |
142 | cbor_free (CBOR *cbor) |
169 | cbor_free (CBOR *cbor) |
143 | { |
170 | { |
144 | SvREFCNT_dec (cbor->filter); |
171 | SvREFCNT_dec (cbor->filter); |
|
|
172 | SvREFCNT_dec (cbor->incr_count); |
145 | } |
173 | } |
146 | |
174 | |
147 | ///////////////////////////////////////////////////////////////////////////// |
175 | ///////////////////////////////////////////////////////////////////////////// |
148 | // utility functions |
176 | // utility functions |
149 | |
177 | |
… | |
… | |
175 | |
203 | |
176 | // minimum length of a string to be registered for stringref |
204 | // minimum length of a string to be registered for stringref |
177 | ecb_inline int |
205 | ecb_inline int |
178 | minimum_string_length (UV idx) |
206 | minimum_string_length (UV idx) |
179 | { |
207 | { |
180 | return idx > 23 |
208 | return idx <= 23 ? 3 |
181 | ? idx > 0xffU |
209 | : idx <= 0xffU ? 4 |
182 | ? idx > 0xffffU |
210 | : idx <= 0xffffU ? 5 |
183 | ? idx > 0xffffffffU |
211 | : idx <= 0xffffffffU ? 7 |
184 | ? 11 |
212 | : 11; |
185 | : 7 |
|
|
186 | : 5 |
|
|
187 | : 4 |
|
|
188 | : 3; |
|
|
189 | } |
213 | } |
190 | |
214 | |
191 | ///////////////////////////////////////////////////////////////////////////// |
215 | ///////////////////////////////////////////////////////////////////////////// |
192 | // encoder |
216 | // encoder |
193 | |
217 | |
… | |
… | |
206 | } enc_t; |
230 | } enc_t; |
207 | |
231 | |
208 | ecb_inline void |
232 | ecb_inline void |
209 | need (enc_t *enc, STRLEN len) |
233 | need (enc_t *enc, STRLEN len) |
210 | { |
234 | { |
211 | if (ecb_expect_false (enc->cur + len >= enc->end)) |
235 | if (ecb_expect_false ((uintptr_t)(enc->end - enc->cur) < len)) |
212 | { |
236 | { |
213 | STRLEN cur = enc->cur - (char *)SvPVX (enc->sv); |
237 | STRLEN cur = enc->cur - (char *)SvPVX (enc->sv); |
214 | SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1); |
238 | SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1); |
215 | enc->cur = SvPVX (enc->sv) + cur; |
239 | enc->cur = SvPVX (enc->sv) + cur; |
216 | enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1; |
240 | enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1; |
… | |
… | |
222 | { |
246 | { |
223 | need (enc, 1); |
247 | need (enc, 1); |
224 | *enc->cur++ = ch; |
248 | *enc->cur++ = ch; |
225 | } |
249 | } |
226 | |
250 | |
|
|
251 | // used for tags, intregers, element counts and so on |
227 | static void |
252 | static void |
228 | encode_uint (enc_t *enc, int major, UV len) |
253 | encode_uint (enc_t *enc, int major, UV len) |
229 | { |
254 | { |
230 | need (enc, 9); |
255 | need (enc, 9); |
231 | |
256 | |
… | |
… | |
262 | *enc->cur++ = len >> 8; |
287 | *enc->cur++ = len >> 8; |
263 | *enc->cur++ = len; |
288 | *enc->cur++ = len; |
264 | } |
289 | } |
265 | } |
290 | } |
266 | |
291 | |
|
|
292 | // encodes a perl value into a CBOR integer |
|
|
293 | ecb_inline void |
|
|
294 | encode_int (enc_t *enc, SV *sv) |
|
|
295 | { |
|
|
296 | if (SvIsUV (sv)) |
|
|
297 | encode_uint (enc, MAJOR_POS_INT, SvUVX (sv)); |
|
|
298 | else if (SvIVX (sv) >= 0) |
|
|
299 | encode_uint (enc, MAJOR_POS_INT, SvIVX (sv)); |
|
|
300 | else |
|
|
301 | encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1)); |
|
|
302 | } |
|
|
303 | |
267 | ecb_inline void |
304 | ecb_inline void |
268 | encode_tag (enc_t *enc, UV tag) |
305 | encode_tag (enc_t *enc, UV tag) |
269 | { |
306 | { |
270 | encode_uint (enc, MAJOR_TAG, tag); |
307 | encode_uint (enc, MAJOR_TAG, tag); |
271 | } |
308 | } |
272 | |
309 | |
|
|
310 | // exceptional (hopefully) slow path for byte strings that need to be utf8-encoded |
|
|
311 | ecb_noinline static void |
|
|
312 | encode_str_utf8 (enc_t *enc, int utf8, char *str, STRLEN len) |
|
|
313 | { |
|
|
314 | STRLEN ulen = len; |
|
|
315 | U8 *p, *pend = (U8 *)str + len; |
|
|
316 | |
|
|
317 | for (p = (U8 *)str; p < pend; ++p) |
|
|
318 | ulen += *p >> 7; // count set high bits |
|
|
319 | |
|
|
320 | encode_uint (enc, MAJOR_TEXT, ulen); |
|
|
321 | |
|
|
322 | need (enc, ulen); |
|
|
323 | for (p = (U8 *)str; p < pend; ++p) |
|
|
324 | if (*p < 0x80) |
|
|
325 | *enc->cur++ = *p; |
|
|
326 | else |
|
|
327 | { |
|
|
328 | *enc->cur++ = 0xc0 + (*p >> 6); |
|
|
329 | *enc->cur++ = 0x80 + (*p & 63); |
|
|
330 | } |
|
|
331 | } |
|
|
332 | |
273 | ecb_inline void |
333 | ecb_inline void |
274 | encode_str (enc_t *enc, int utf8, char *str, STRLEN len) |
334 | encode_str (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len) |
275 | { |
335 | { |
|
|
336 | if (ecb_expect_false (upgrade_utf8)) |
|
|
337 | if (!utf8) |
|
|
338 | { |
|
|
339 | encode_str_utf8 (enc, utf8, str, len); |
|
|
340 | return; |
|
|
341 | } |
|
|
342 | |
276 | encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len); |
343 | encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len); |
277 | need (enc, len); |
344 | need (enc, len); |
278 | memcpy (enc->cur, str, len); |
345 | memcpy (enc->cur, str, len); |
279 | enc->cur += len; |
346 | enc->cur += len; |
280 | } |
347 | } |
281 | |
348 | |
282 | static void |
349 | ecb_inline void |
283 | encode_strref (enc_t *enc, int utf8, char *str, STRLEN len) |
350 | encode_strref (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len) |
284 | { |
351 | { |
285 | if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS)) |
352 | if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS)) |
286 | { |
353 | { |
287 | SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1); |
354 | SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1); |
288 | |
355 | |
… | |
… | |
299 | sv_setuv (*svp, enc->stringref_idx); |
366 | sv_setuv (*svp, enc->stringref_idx); |
300 | ++enc->stringref_idx; |
367 | ++enc->stringref_idx; |
301 | } |
368 | } |
302 | } |
369 | } |
303 | |
370 | |
304 | encode_str (enc, utf8, str, len); |
371 | encode_str (enc, upgrade_utf8, utf8, str, len); |
|
|
372 | } |
|
|
373 | |
|
|
374 | ecb_inline void |
|
|
375 | encode_float16 (enc_t *enc, NV nv) |
|
|
376 | { |
|
|
377 | need (enc, 1+2); |
|
|
378 | |
|
|
379 | *enc->cur++ = MAJOR_MISC | MISC_FLOAT16; |
|
|
380 | |
|
|
381 | uint16_t fp = ecb_float_to_binary16 (nv); |
|
|
382 | |
|
|
383 | if (!ecb_big_endian ()) |
|
|
384 | fp = ecb_bswap16 (fp); |
|
|
385 | |
|
|
386 | memcpy (enc->cur, &fp, 2); |
|
|
387 | enc->cur += 2; |
|
|
388 | } |
|
|
389 | |
|
|
390 | ecb_inline void |
|
|
391 | encode_float32 (enc_t *enc, NV nv) |
|
|
392 | { |
|
|
393 | need (enc, 1+4); |
|
|
394 | |
|
|
395 | *enc->cur++ = MAJOR_MISC | MISC_FLOAT32; |
|
|
396 | |
|
|
397 | uint32_t fp = ecb_float_to_binary32 (nv); |
|
|
398 | |
|
|
399 | if (!ecb_big_endian ()) |
|
|
400 | fp = ecb_bswap32 (fp); |
|
|
401 | |
|
|
402 | memcpy (enc->cur, &fp, 4); |
|
|
403 | enc->cur += 4; |
|
|
404 | } |
|
|
405 | |
|
|
406 | ecb_inline void |
|
|
407 | encode_float64 (enc_t *enc, NV nv) |
|
|
408 | { |
|
|
409 | need (enc, 1+8); |
|
|
410 | |
|
|
411 | *enc->cur++ = MAJOR_MISC | MISC_FLOAT64; |
|
|
412 | |
|
|
413 | uint64_t fp = ecb_double_to_binary64 (nv); |
|
|
414 | |
|
|
415 | if (!ecb_big_endian ()) |
|
|
416 | fp = ecb_bswap64 (fp); |
|
|
417 | |
|
|
418 | memcpy (enc->cur, &fp, 8); |
|
|
419 | enc->cur += 8; |
|
|
420 | } |
|
|
421 | |
|
|
422 | ecb_inline void |
|
|
423 | encode_forced (enc_t *enc, UV type, SV *sv) |
|
|
424 | { |
|
|
425 | switch (type) |
|
|
426 | { |
|
|
427 | case AS_CBOR: |
|
|
428 | { |
|
|
429 | STRLEN len; |
|
|
430 | char *str = SvPVbyte (sv, len); |
|
|
431 | |
|
|
432 | need (enc, len); |
|
|
433 | memcpy (enc->cur, str, len); |
|
|
434 | enc->cur += len; |
|
|
435 | } |
|
|
436 | break; |
|
|
437 | |
|
|
438 | case AS_BYTES: |
|
|
439 | { |
|
|
440 | STRLEN len; |
|
|
441 | char *str = SvPVbyte (sv, len); |
|
|
442 | encode_strref (enc, 0, 0, str, len); |
|
|
443 | } |
|
|
444 | break; |
|
|
445 | |
|
|
446 | case AS_TEXT: |
|
|
447 | { |
|
|
448 | STRLEN len; |
|
|
449 | char *str = SvPVutf8 (sv, len); |
|
|
450 | encode_strref (enc, 1, 1, str, len); |
|
|
451 | } |
|
|
452 | break; |
|
|
453 | |
|
|
454 | case AS_INT: encode_int (enc, sv); break; |
|
|
455 | case AS_FLOAT16: encode_float16 (enc, SvNV (sv)); break; |
|
|
456 | case AS_FLOAT32: encode_float32 (enc, SvNV (sv)); break; |
|
|
457 | case AS_FLOAT64: encode_float64 (enc, SvNV (sv)); break; |
|
|
458 | |
|
|
459 | default: |
|
|
460 | croak ("encountered malformed CBOR::XS::Tagged object"); |
|
|
461 | } |
305 | } |
462 | } |
306 | |
463 | |
307 | static void encode_sv (enc_t *enc, SV *sv); |
464 | static void encode_sv (enc_t *enc, SV *sv); |
308 | |
465 | |
309 | static void |
466 | static void |
… | |
… | |
316 | |
473 | |
317 | ++enc->depth; |
474 | ++enc->depth; |
318 | |
475 | |
319 | encode_uint (enc, MAJOR_ARRAY, len + 1); |
476 | encode_uint (enc, MAJOR_ARRAY, len + 1); |
320 | |
477 | |
|
|
478 | if (ecb_expect_false (SvMAGICAL (av))) |
321 | for (i = 0; i <= len; ++i) |
479 | for (i = 0; i <= len; ++i) |
322 | { |
480 | { |
323 | SV **svp = av_fetch (av, i, 0); |
481 | SV **svp = av_fetch (av, i, 0); |
324 | encode_sv (enc, svp ? *svp : &PL_sv_undef); |
482 | encode_sv (enc, svp ? *svp : &PL_sv_undef); |
325 | } |
483 | } |
|
|
484 | else |
|
|
485 | for (i = 0; i <= len; ++i) |
|
|
486 | { |
|
|
487 | SV *sv = AvARRAY (av)[i]; |
|
|
488 | encode_sv (enc, sv ? sv : &PL_sv_undef); |
|
|
489 | } |
326 | |
490 | |
327 | --enc->depth; |
491 | --enc->depth; |
328 | } |
492 | } |
329 | |
493 | |
330 | static void |
494 | static void |
… | |
… | |
338 | ++enc->depth; |
502 | ++enc->depth; |
339 | |
503 | |
340 | int pairs = hv_iterinit (hv); |
504 | int pairs = hv_iterinit (hv); |
341 | int mg = SvMAGICAL (hv); |
505 | int mg = SvMAGICAL (hv); |
342 | |
506 | |
343 | if (mg) |
507 | if (ecb_expect_false (mg)) |
344 | encode_ch (enc, MAJOR_MAP | MINOR_INDEF); |
508 | encode_ch (enc, MAJOR_MAP | MINOR_INDEF); |
345 | else |
509 | else |
346 | encode_uint (enc, MAJOR_MAP, pairs); |
510 | encode_uint (enc, MAJOR_MAP, pairs); |
347 | |
511 | |
348 | while ((he = hv_iternext (hv))) |
512 | while ((he = hv_iternext (hv))) |
349 | { |
513 | { |
350 | if (HeKLEN (he) == HEf_SVKEY) |
514 | if (HeKLEN (he) == HEf_SVKEY) |
351 | encode_sv (enc, HeSVKEY (he)); |
515 | encode_sv (enc, HeSVKEY (he)); |
352 | else |
516 | else |
353 | encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); |
517 | encode_strref (enc, enc->cbor.flags & (F_TEXT_KEYS | F_TEXT_STRINGS), HeKUTF8 (he), HeKEY (he), HeKLEN (he)); |
354 | |
518 | |
355 | encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); |
519 | encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); |
356 | } |
520 | } |
357 | |
521 | |
358 | if (mg) |
522 | if (ecb_expect_false (mg)) |
359 | encode_ch (enc, MAJOR_MISC | MINOR_INDEF); |
523 | encode_ch (enc, MAJOR_MISC | MINOR_INDEF); |
360 | |
524 | |
361 | --enc->depth; |
525 | --enc->depth; |
362 | } |
526 | } |
363 | |
527 | |
… | |
… | |
396 | else if (stash == tagged_stash) |
560 | else if (stash == tagged_stash) |
397 | { |
561 | { |
398 | if (svt != SVt_PVAV) |
562 | if (svt != SVt_PVAV) |
399 | croak ("encountered CBOR::XS::Tagged object that isn't an array"); |
563 | croak ("encountered CBOR::XS::Tagged object that isn't an array"); |
400 | |
564 | |
|
|
565 | switch (av_len ((AV *)sv)) |
|
|
566 | { |
|
|
567 | case 2-1: |
|
|
568 | // actually a tagged value |
401 | encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1))); |
569 | encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1))); |
402 | encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); |
570 | encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); |
|
|
571 | break; |
|
|
572 | |
|
|
573 | case 3-1: |
|
|
574 | // a forced type [value, type, undef] |
|
|
575 | encode_forced (enc, SvUV (*av_fetch ((AV *)sv, 1, 1)), *av_fetch ((AV *)sv, 0, 1)); |
|
|
576 | break; |
|
|
577 | |
|
|
578 | default: |
|
|
579 | croak ("encountered malformed CBOR::XS::Tagged object"); |
|
|
580 | } |
403 | |
581 | |
404 | return; |
582 | return; |
405 | } |
583 | } |
406 | } |
584 | } |
407 | |
585 | |
408 | if (ecb_expect_false (SvREFCNT (sv) > 1) |
586 | if (ecb_expect_false (SvREFCNT (sv) > 1) |
409 | && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)) |
587 | && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)) |
410 | { |
588 | { |
411 | if (!enc->shareable) |
589 | if (ecb_expect_false (!enc->shareable)) |
412 | enc->shareable = (HV *)sv_2mortal ((SV *)newHV ()); |
590 | enc->shareable = (HV *)sv_2mortal ((SV *)newHV ()); |
413 | |
591 | |
414 | SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1); |
592 | SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1); |
415 | |
593 | |
416 | if (SvOK (*svp)) |
594 | if (SvOK (*svp)) |
… | |
… | |
430 | if (ecb_expect_false (SvOBJECT (sv))) |
608 | if (ecb_expect_false (SvOBJECT (sv))) |
431 | { |
609 | { |
432 | HV *stash = SvSTASH (sv); |
610 | HV *stash = SvSTASH (sv); |
433 | GV *method; |
611 | GV *method; |
434 | |
612 | |
|
|
613 | if (enc->cbor.flags & F_FORBID_OBJECTS) |
|
|
614 | croak ("encountered object '%s', but forbid_objects is enabled", |
|
|
615 | SvPV_nolen (sv_2mortal (newRV_inc (sv)))); |
435 | if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) |
616 | else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) |
436 | { |
617 | { |
437 | dSP; |
618 | dSP; |
438 | |
619 | |
439 | ENTER; SAVETMPS; PUSHMARK (SP); |
620 | ENTER; SAVETMPS; |
|
|
621 | PUSHMARK (SP); |
440 | // we re-bless the reference to get overload and other niceties right |
622 | // we re-bless the reference to get overload and other niceties right |
441 | XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); |
623 | XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); |
442 | |
624 | |
443 | PUTBACK; |
625 | PUTBACK; |
444 | // G_SCALAR ensures that return value is 1 |
626 | // G_SCALAR ensures that return value is 1 |
… | |
… | |
457 | } |
639 | } |
458 | else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0) |
640 | else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0) |
459 | { |
641 | { |
460 | dSP; |
642 | dSP; |
461 | |
643 | |
462 | ENTER; SAVETMPS; PUSHMARK (SP); |
644 | ENTER; SAVETMPS; |
|
|
645 | PUSHMARK (SP); |
463 | EXTEND (SP, 2); |
646 | EXTEND (SP, 2); |
464 | // we re-bless the reference to get overload and other niceties right |
647 | // we re-bless the reference to get overload and other niceties right |
465 | PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); |
648 | PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); |
466 | PUSHs (sv_cbor); |
649 | PUSHs (sv_cbor); |
467 | |
650 | |
… | |
… | |
473 | if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) |
656 | if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) |
474 | croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash)); |
657 | croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash)); |
475 | |
658 | |
476 | encode_tag (enc, CBOR_TAG_PERL_OBJECT); |
659 | encode_tag (enc, CBOR_TAG_PERL_OBJECT); |
477 | encode_uint (enc, MAJOR_ARRAY, count + 1); |
660 | encode_uint (enc, MAJOR_ARRAY, count + 1); |
478 | encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); |
661 | encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); |
479 | |
662 | |
480 | while (count) |
663 | { |
|
|
664 | int i; |
|
|
665 | |
|
|
666 | for (i = 0; i < count; ++i) |
481 | encode_sv (enc, SP[1 - count--]); |
667 | encode_sv (enc, SP[i + 1 - count]); |
|
|
668 | |
|
|
669 | SP -= count; |
|
|
670 | } |
482 | |
671 | |
483 | PUTBACK; |
672 | PUTBACK; |
484 | |
673 | |
485 | FREETMPS; LEAVE; |
674 | FREETMPS; LEAVE; |
486 | } |
675 | } |
… | |
… | |
508 | |
697 | |
509 | if (ecb_expect_false (nv == (NV)(U32)nv)) |
698 | if (ecb_expect_false (nv == (NV)(U32)nv)) |
510 | encode_uint (enc, MAJOR_POS_INT, (U32)nv); |
699 | encode_uint (enc, MAJOR_POS_INT, (U32)nv); |
511 | //TODO: maybe I32? |
700 | //TODO: maybe I32? |
512 | else if (ecb_expect_false (nv == (float)nv)) |
701 | else if (ecb_expect_false (nv == (float)nv)) |
513 | { |
702 | encode_float32 (enc, nv); |
514 | uint32_t fp = ecb_float_to_binary32 (nv); |
|
|
515 | |
|
|
516 | *enc->cur++ = MAJOR_MISC | MISC_FLOAT32; |
|
|
517 | |
|
|
518 | if (!ecb_big_endian ()) |
|
|
519 | fp = ecb_bswap32 (fp); |
|
|
520 | |
|
|
521 | memcpy (enc->cur, &fp, 4); |
|
|
522 | enc->cur += 4; |
|
|
523 | } |
|
|
524 | else |
703 | else |
525 | { |
704 | encode_float64 (enc, nv); |
526 | uint64_t fp = ecb_double_to_binary64 (nv); |
|
|
527 | |
|
|
528 | *enc->cur++ = MAJOR_MISC | MISC_FLOAT64; |
|
|
529 | |
|
|
530 | if (!ecb_big_endian ()) |
|
|
531 | fp = ecb_bswap64 (fp); |
|
|
532 | |
|
|
533 | memcpy (enc->cur, &fp, 8); |
|
|
534 | enc->cur += 8; |
|
|
535 | } |
|
|
536 | } |
705 | } |
537 | |
706 | |
538 | static void |
707 | static void |
539 | encode_sv (enc_t *enc, SV *sv) |
708 | encode_sv (enc_t *enc, SV *sv) |
540 | { |
709 | { |
… | |
… | |
542 | |
711 | |
543 | if (SvPOKp (sv)) |
712 | if (SvPOKp (sv)) |
544 | { |
713 | { |
545 | STRLEN len; |
714 | STRLEN len; |
546 | char *str = SvPV (sv, len); |
715 | char *str = SvPV (sv, len); |
547 | encode_strref (enc, SvUTF8 (sv), str, len); |
716 | encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len); |
548 | } |
717 | } |
549 | else if (SvNOKp (sv)) |
718 | else if (SvNOKp (sv)) |
550 | encode_nv (enc, sv); |
719 | encode_nv (enc, sv); |
551 | else if (SvIOKp (sv)) |
720 | else if (SvIOKp (sv)) |
552 | { |
721 | encode_int (enc, sv); |
553 | if (SvIsUV (sv)) |
|
|
554 | encode_uint (enc, MAJOR_POS_INT, SvUVX (sv)); |
|
|
555 | else if (SvIVX (sv) >= 0) |
|
|
556 | encode_uint (enc, MAJOR_POS_INT, SvIVX (sv)); |
|
|
557 | else |
|
|
558 | encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1)); |
|
|
559 | } |
|
|
560 | else if (SvROK (sv)) |
722 | else if (SvROK (sv)) |
561 | encode_rv (enc, SvRV (sv)); |
723 | encode_rv (enc, SvRV (sv)); |
562 | else if (!SvOK (sv)) |
724 | else if (!SvOK (sv)) |
563 | encode_ch (enc, MAJOR_MISC | SIMPLE_NULL); |
725 | encode_ch (enc, MAJOR_MISC | SIMPLE_NULL); |
564 | else if (enc->cbor.flags & F_ALLOW_UNKNOWN) |
726 | else if (enc->cbor.flags & F_ALLOW_UNKNOWN) |
… | |
… | |
569 | } |
731 | } |
570 | |
732 | |
571 | static SV * |
733 | static SV * |
572 | encode_cbor (SV *scalar, CBOR *cbor) |
734 | encode_cbor (SV *scalar, CBOR *cbor) |
573 | { |
735 | { |
574 | enc_t enc = { }; |
736 | enc_t enc = { 0 }; |
575 | |
737 | |
576 | enc.cbor = *cbor; |
738 | enc.cbor = *cbor; |
577 | enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); |
739 | enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); |
578 | enc.cur = SvPVX (enc.sv); |
740 | enc.cur = SvPVX (enc.sv); |
579 | enc.end = SvEND (enc.sv); |
741 | enc.end = SvEND (enc.sv); |
580 | |
742 | |
581 | SvPOK_only (enc.sv); |
743 | SvPOK_only (enc.sv); |
582 | |
744 | |
583 | if (cbor->flags & F_PACK_STRINGS) |
745 | if (cbor->flags & F_PACK_STRINGS) |
584 | { |
746 | { |
… | |
… | |
611 | U32 depth; // recursion depth |
773 | U32 depth; // recursion depth |
612 | U32 maxdepth; // recursion depth limit |
774 | U32 maxdepth; // recursion depth limit |
613 | AV *shareable; |
775 | AV *shareable; |
614 | AV *stringref; |
776 | AV *stringref; |
615 | SV *decode_tagged; |
777 | SV *decode_tagged; |
|
|
778 | SV *err_sv; // optional sv for error, needs to be freed |
616 | } dec_t; |
779 | } dec_t; |
617 | |
780 | |
618 | #define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE |
781 | // set dec->err to ERRSV |
|
|
782 | ecb_cold static void |
|
|
783 | err_errsv (dec_t *dec) |
|
|
784 | { |
|
|
785 | if (!dec->err) |
|
|
786 | { |
|
|
787 | dec->err_sv = newSVsv (ERRSV); |
619 | |
788 | |
620 | #define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data") |
789 | // chop off the trailing \n |
|
|
790 | SvCUR_set (dec->err_sv, SvCUR (dec->err_sv) - 1); |
|
|
791 | *SvEND (dec->err_sv) = 0; |
621 | |
792 | |
|
|
793 | dec->err = SvPVutf8_nolen (dec->err_sv); |
|
|
794 | } |
|
|
795 | } |
|
|
796 | |
|
|
797 | // the following functions are used to reduce code size and help the compiler to optimise |
|
|
798 | ecb_cold static void |
|
|
799 | err_set (dec_t *dec, const char *reason) |
|
|
800 | { |
|
|
801 | if (!dec->err) |
|
|
802 | dec->err = reason; |
|
|
803 | } |
|
|
804 | |
|
|
805 | ecb_cold static void |
|
|
806 | err_unexpected_end (dec_t *dec) |
|
|
807 | { |
|
|
808 | err_set (dec, "unexpected end of CBOR data"); |
|
|
809 | } |
|
|
810 | |
|
|
811 | #define ERR_DO(do) SB do; goto fail; SE |
|
|
812 | #define ERR(reason) ERR_DO (err_set (dec, reason)) |
|
|
813 | #define ERR_ERRSV ERR_DO (err_errsv (dec)) |
|
|
814 | |
|
|
815 | #define WANT(len) if (ecb_expect_false ((uintptr_t)(dec->end - dec->cur) < (STRLEN)len)) ERR_DO (err_unexpected_end (dec)) |
|
|
816 | |
622 | #define DEC_INC_DEPTH if (++dec->depth > dec->cbor.max_depth) ERR (ERR_NESTING_EXCEEDED) |
817 | #define DEC_INC_DEPTH if (ecb_expect_false (++dec->depth > dec->cbor.max_depth)) ERR (ERR_NESTING_EXCEEDED) |
623 | #define DEC_DEC_DEPTH --dec->depth |
818 | #define DEC_DEC_DEPTH --dec->depth |
624 | |
819 | |
625 | static UV |
820 | static UV |
626 | decode_uint (dec_t *dec) |
821 | decode_uint (dec_t *dec) |
627 | { |
822 | { |
… | |
… | |
693 | |
888 | |
694 | for (;;) |
889 | for (;;) |
695 | { |
890 | { |
696 | WANT (1); |
891 | WANT (1); |
697 | |
892 | |
698 | if (*dec->cur == (MAJOR_MISC | MINOR_INDEF)) |
893 | if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err) |
699 | { |
894 | { |
700 | ++dec->cur; |
895 | ++dec->cur; |
701 | break; |
896 | break; |
702 | } |
897 | } |
703 | |
898 | |
704 | av_push (av, decode_sv (dec)); |
899 | av_push (av, decode_sv (dec)); |
705 | } |
900 | } |
706 | } |
901 | } |
707 | else |
902 | else |
708 | { |
903 | { |
709 | int i, len = decode_uint (dec); |
904 | UV i, len = decode_uint (dec); |
710 | |
905 | |
711 | WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays |
906 | WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays |
712 | av_fill (av, len - 1); |
907 | av_fill (av, len - 1); |
713 | |
908 | |
714 | for (i = 0; i < len; ++i) |
909 | for (i = 0; i < len; ++i) |
… | |
… | |
717 | |
912 | |
718 | DEC_DEC_DEPTH; |
913 | DEC_DEC_DEPTH; |
719 | return newRV_noinc ((SV *)av); |
914 | return newRV_noinc ((SV *)av); |
720 | |
915 | |
721 | fail: |
916 | fail: |
722 | SvREFCNT_dec (av); |
917 | SvREFCNT_dec_NN (av); |
723 | DEC_DEC_DEPTH; |
918 | DEC_DEC_DEPTH; |
724 | return &PL_sv_undef; |
919 | return &PL_sv_undef; |
725 | } |
920 | } |
726 | |
921 | |
727 | static void |
922 | static void |
… | |
… | |
729 | { |
924 | { |
730 | // for speed reasons, we specialcase single-string |
925 | // for speed reasons, we specialcase single-string |
731 | // byte or utf-8 strings as keys, but only when !stringref |
926 | // byte or utf-8 strings as keys, but only when !stringref |
732 | |
927 | |
733 | if (ecb_expect_true (!dec->stringref)) |
928 | if (ecb_expect_true (!dec->stringref)) |
734 | if (ecb_expect_true ((*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8)) |
929 | if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8)) |
735 | { |
930 | { |
736 | I32 len = decode_uint (dec); |
931 | STRLEN len = decode_uint (dec); |
737 | char *key = (char *)dec->cur; |
932 | char *key = (char *)dec->cur; |
738 | |
933 | |
|
|
934 | WANT (len); |
739 | dec->cur += len; |
935 | dec->cur += len; |
740 | |
936 | |
741 | if (ecb_expect_false (dec->stringref)) |
|
|
742 | av_push (dec->stringref, newSVpvn (key, len)); |
|
|
743 | |
|
|
744 | hv_store (hv, key, len, decode_sv (dec), 0); |
937 | hv_store (hv, key, len, decode_sv (dec), 0); |
745 | |
938 | |
746 | return; |
939 | return; |
747 | } |
940 | } |
748 | else if (ecb_expect_true ((*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8)) |
941 | else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8)) |
749 | { |
942 | { |
750 | I32 len = decode_uint (dec); |
943 | STRLEN len = decode_uint (dec); |
751 | char *key = (char *)dec->cur; |
944 | char *key = (char *)dec->cur; |
752 | |
945 | |
|
|
946 | WANT (len); |
753 | dec->cur += len; |
947 | dec->cur += len; |
754 | |
948 | |
755 | if (ecb_expect_false (dec->stringref)) |
949 | if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8)) |
756 | av_push (dec->stringref, newSVpvn_utf8 (key, len, 1)); |
950 | if (!is_utf8_string (key, len)) |
|
|
951 | ERR ("corrupted CBOR data (invalid UTF-8 in map key)"); |
757 | |
952 | |
758 | hv_store (hv, key, -len, decode_sv (dec), 0); |
953 | hv_store (hv, key, -len, decode_sv (dec), 0); |
759 | |
954 | |
760 | return; |
955 | return; |
761 | } |
956 | } |
762 | |
957 | |
763 | SV *k = decode_sv (dec); |
958 | SV *k = decode_sv (dec); |
764 | SV *v = decode_sv (dec); |
959 | SV *v = decode_sv (dec); |
765 | |
960 | |
|
|
961 | // we leak memory if uncaught exceptions are thrown by random magical |
|
|
962 | // methods, and this is hopefully the only place where it can happen, |
|
|
963 | // so if there is a chance of an exception, take the very slow path. |
|
|
964 | // since catching exceptions is "undocumented/internal/forbidden" by |
|
|
965 | // the new p5p powers, we need to call out to a perl function :/ |
|
|
966 | if (ecb_expect_false (SvAMAGIC (k))) |
|
|
967 | { |
|
|
968 | dSP; |
|
|
969 | |
|
|
970 | ENTER; SAVETMPS; |
|
|
971 | PUSHMARK (SP); |
|
|
972 | EXTEND (SP, 3); |
|
|
973 | PUSHs (sv_2mortal (newRV_inc ((SV *)hv))); |
|
|
974 | PUSHs (sv_2mortal (k)); |
|
|
975 | PUSHs (sv_2mortal (v)); |
|
|
976 | |
|
|
977 | PUTBACK; |
|
|
978 | call_pv ("CBOR::XS::_hv_store", G_VOID | G_DISCARD | G_EVAL); |
|
|
979 | SPAGAIN; |
|
|
980 | |
|
|
981 | FREETMPS; LEAVE; |
|
|
982 | |
|
|
983 | if (SvTRUE (ERRSV)) |
|
|
984 | ERR_ERRSV; |
|
|
985 | |
|
|
986 | return; |
|
|
987 | } |
|
|
988 | |
766 | hv_store_ent (hv, k, v, 0); |
989 | hv_store_ent (hv, k, v, 0); |
767 | SvREFCNT_dec (k); |
990 | SvREFCNT_dec_NN (k); |
|
|
991 | |
|
|
992 | fail: |
|
|
993 | ; |
768 | } |
994 | } |
769 | |
995 | |
770 | static SV * |
996 | static SV * |
771 | decode_hv (dec_t *dec) |
997 | decode_hv (dec_t *dec) |
772 | { |
998 | { |
… | |
… | |
780 | |
1006 | |
781 | for (;;) |
1007 | for (;;) |
782 | { |
1008 | { |
783 | WANT (1); |
1009 | WANT (1); |
784 | |
1010 | |
785 | if (*dec->cur == (MAJOR_MISC | MINOR_INDEF)) |
1011 | if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err) |
786 | { |
1012 | { |
787 | ++dec->cur; |
1013 | ++dec->cur; |
788 | break; |
1014 | break; |
789 | } |
1015 | } |
790 | |
1016 | |
791 | decode_he (dec, hv); |
1017 | decode_he (dec, hv); |
792 | } |
1018 | } |
793 | } |
1019 | } |
794 | else |
1020 | else |
795 | { |
1021 | { |
796 | int pairs = decode_uint (dec); |
1022 | UV pairs = decode_uint (dec); |
|
|
1023 | |
|
|
1024 | WANT (pairs); // complexity check - need at least one byte per value, do not allow supersize hashes |
797 | |
1025 | |
798 | while (pairs--) |
1026 | while (pairs--) |
799 | decode_he (dec, hv); |
1027 | decode_he (dec, hv); |
800 | } |
1028 | } |
801 | |
1029 | |
802 | DEC_DEC_DEPTH; |
1030 | DEC_DEC_DEPTH; |
803 | return newRV_noinc ((SV *)hv); |
1031 | return newRV_noinc ((SV *)hv); |
804 | |
1032 | |
805 | fail: |
1033 | fail: |
806 | SvREFCNT_dec (hv); |
1034 | SvREFCNT_dec_NN (hv); |
807 | DEC_DEC_DEPTH; |
1035 | DEC_DEC_DEPTH; |
808 | return &PL_sv_undef; |
1036 | return &PL_sv_undef; |
809 | } |
1037 | } |
810 | |
1038 | |
811 | static SV * |
1039 | static SV * |
812 | decode_str (dec_t *dec, int utf8) |
1040 | decode_str (dec_t *dec, int utf8) |
813 | { |
1041 | { |
814 | SV *sv = 0; |
1042 | SV *sv = 0; |
815 | |
1043 | |
816 | if ((*dec->cur & MINOR_MASK) == MINOR_INDEF) |
1044 | if (ecb_expect_false ((*dec->cur & MINOR_MASK) == MINOR_INDEF)) |
817 | { |
1045 | { |
818 | // indefinite length strings |
1046 | // indefinite length strings |
819 | ++dec->cur; |
1047 | ++dec->cur; |
820 | |
1048 | |
821 | U8 major = *dec->cur & MAJOR_MISC; |
1049 | U8 major = *dec->cur & MAJOR_MISC; |
… | |
… | |
854 | && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1)) |
1082 | && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1)) |
855 | av_push (dec->stringref, SvREFCNT_inc_NN (sv)); |
1083 | av_push (dec->stringref, SvREFCNT_inc_NN (sv)); |
856 | } |
1084 | } |
857 | |
1085 | |
858 | if (utf8) |
1086 | if (utf8) |
|
|
1087 | { |
|
|
1088 | if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8)) |
|
|
1089 | if (!is_utf8_string (SvPVX (sv), SvCUR (sv))) |
|
|
1090 | ERR ("corrupted CBOR data (invalid UTF-8 in text string)"); |
|
|
1091 | |
859 | SvUTF8_on (sv); |
1092 | SvUTF8_on (sv); |
|
|
1093 | } |
860 | |
1094 | |
861 | return sv; |
1095 | return sv; |
862 | |
1096 | |
863 | fail: |
1097 | fail: |
864 | SvREFCNT_dec (sv); |
1098 | SvREFCNT_dec (sv); |
… | |
… | |
883 | sv = newRV_noinc (decode_sv (dec)); |
1117 | sv = newRV_noinc (decode_sv (dec)); |
884 | break; |
1118 | break; |
885 | |
1119 | |
886 | case CBOR_TAG_STRINGREF_NAMESPACE: |
1120 | case CBOR_TAG_STRINGREF_NAMESPACE: |
887 | { |
1121 | { |
|
|
1122 | // do not use SAVETMPS/FREETMPS, as these will |
|
|
1123 | // erase mortalised caches, e.g. "shareable" |
888 | ENTER; SAVETMPS; |
1124 | ENTER; |
889 | |
1125 | |
890 | SAVESPTR (dec->stringref); |
1126 | SAVESPTR (dec->stringref); |
891 | dec->stringref = (AV *)sv_2mortal ((SV *)newAV ()); |
1127 | dec->stringref = (AV *)sv_2mortal ((SV *)newAV ()); |
892 | |
1128 | |
893 | sv = decode_sv (dec); |
1129 | sv = decode_sv (dec); |
894 | |
1130 | |
895 | FREETMPS; LEAVE; |
1131 | LEAVE; |
896 | } |
1132 | } |
897 | break; |
1133 | break; |
898 | |
1134 | |
899 | case CBOR_TAG_STRINGREF: |
1135 | case CBOR_TAG_STRINGREF: |
900 | { |
1136 | { |
901 | if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT)) |
1137 | if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT)) |
902 | ERR ("corrupted CBOR data (stringref index not an unsigned integer)"); |
1138 | ERR ("corrupted CBOR data (stringref index not an unsigned integer)"); |
903 | |
1139 | |
904 | UV idx = decode_uint (dec); |
1140 | UV idx = decode_uint (dec); |
905 | |
1141 | |
906 | if (!dec->stringref || (int)idx > AvFILLp (dec->stringref)) |
1142 | if (!dec->stringref || idx >= (UV)(1 + AvFILLp (dec->stringref))) |
907 | ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)"); |
1143 | ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)"); |
908 | |
1144 | |
909 | sv = newSVsv (AvARRAY (dec->stringref)[idx]); |
1145 | sv = newSVsv (AvARRAY (dec->stringref)[idx]); |
910 | } |
1146 | } |
911 | break; |
1147 | break; |
… | |
… | |
939 | if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT)) |
1175 | if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT)) |
940 | ERR ("corrupted CBOR data (sharedref index not an unsigned integer)"); |
1176 | ERR ("corrupted CBOR data (sharedref index not an unsigned integer)"); |
941 | |
1177 | |
942 | UV idx = decode_uint (dec); |
1178 | UV idx = decode_uint (dec); |
943 | |
1179 | |
944 | if (!dec->shareable || (int)idx > AvFILLp (dec->shareable)) |
1180 | if (!dec->shareable || idx >= (UV)(1 + AvFILLp (dec->shareable))) |
945 | ERR ("corrupted CBOR data (sharedref index out of bounds)"); |
1181 | ERR ("corrupted CBOR data (sharedref index out of bounds)"); |
946 | |
1182 | |
947 | sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]); |
1183 | sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]); |
948 | |
1184 | |
949 | if (sv == &PL_sv_undef) |
1185 | if (sv == &PL_sv_undef) |
… | |
… | |
951 | } |
1187 | } |
952 | break; |
1188 | break; |
953 | |
1189 | |
954 | case CBOR_TAG_PERL_OBJECT: |
1190 | case CBOR_TAG_PERL_OBJECT: |
955 | { |
1191 | { |
|
|
1192 | if (dec->cbor.flags & F_FORBID_OBJECTS) |
|
|
1193 | goto filter; |
|
|
1194 | |
956 | sv = decode_sv (dec); |
1195 | sv = decode_sv (dec); |
957 | |
1196 | |
958 | if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) |
1197 | if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) |
959 | ERR ("corrupted CBOR data (non-array perl object)"); |
1198 | ERR ("corrupted CBOR data (non-array perl object)"); |
960 | |
1199 | |
… | |
… | |
970 | if (!method) |
1209 | if (!method) |
971 | ERR ("cannot decode perl-object (package does not have a THAW method)"); |
1210 | ERR ("cannot decode perl-object (package does not have a THAW method)"); |
972 | |
1211 | |
973 | dSP; |
1212 | dSP; |
974 | |
1213 | |
975 | ENTER; SAVETMPS; PUSHMARK (SP); |
1214 | ENTER; SAVETMPS; |
|
|
1215 | PUSHMARK (SP); |
976 | EXTEND (SP, len + 1); |
1216 | EXTEND (SP, len + 1); |
977 | // we re-bless the reference to get overload and other niceties right |
1217 | // we re-bless the reference to get overload and other niceties right |
978 | PUSHs (*av_fetch (av, 0, 1)); |
1218 | PUSHs (*av_fetch (av, 0, 1)); |
979 | PUSHs (sv_cbor); |
1219 | PUSHs (sv_cbor); |
980 | |
1220 | |
… | |
… | |
988 | SPAGAIN; |
1228 | SPAGAIN; |
989 | |
1229 | |
990 | if (SvTRUE (ERRSV)) |
1230 | if (SvTRUE (ERRSV)) |
991 | { |
1231 | { |
992 | FREETMPS; LEAVE; |
1232 | FREETMPS; LEAVE; |
993 | ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV)))); |
1233 | ERR_ERRSV; |
994 | } |
1234 | } |
995 | |
1235 | |
996 | SvREFCNT_dec (sv); |
1236 | SvREFCNT_dec_NN (sv); |
997 | sv = SvREFCNT_inc (POPs); |
1237 | sv = SvREFCNT_inc (POPs); |
998 | |
1238 | |
999 | PUTBACK; |
1239 | PUTBACK; |
1000 | |
1240 | |
1001 | FREETMPS; LEAVE; |
1241 | FREETMPS; LEAVE; |
1002 | } |
1242 | } |
1003 | break; |
1243 | break; |
1004 | |
1244 | |
1005 | default: |
1245 | default: |
|
|
1246 | filter: |
1006 | { |
1247 | { |
|
|
1248 | SV *tag_sv = newSVuv (tag); |
|
|
1249 | |
1007 | sv = decode_sv (dec); |
1250 | sv = decode_sv (dec); |
1008 | |
1251 | |
1009 | dSP; |
1252 | dSP; |
1010 | ENTER; SAVETMPS; PUSHMARK (SP); |
1253 | ENTER; SAVETMPS; |
|
|
1254 | PUSHMARK (SP); |
1011 | EXTEND (SP, 2); |
1255 | EXTEND (SP, 2); |
1012 | PUSHs (newSVuv (tag)); |
1256 | PUSHs (tag_sv); |
1013 | PUSHs (sv); |
1257 | PUSHs (sv); |
1014 | |
1258 | |
1015 | PUTBACK; |
1259 | PUTBACK; |
1016 | int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL); |
1260 | int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL); |
1017 | SPAGAIN; |
1261 | SPAGAIN; |
1018 | |
1262 | |
1019 | if (SvTRUE (ERRSV)) |
1263 | if (SvTRUE (ERRSV)) |
1020 | { |
1264 | { |
|
|
1265 | SvREFCNT_dec_NN (tag_sv); |
1021 | FREETMPS; LEAVE; |
1266 | FREETMPS; LEAVE; |
1022 | ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV)))); |
1267 | ERR_ERRSV; |
1023 | } |
1268 | } |
1024 | |
1269 | |
1025 | if (count) |
1270 | if (count) |
1026 | { |
1271 | { |
|
|
1272 | SvREFCNT_dec_NN (tag_sv); |
1027 | SvREFCNT_dec (sv); |
1273 | SvREFCNT_dec_NN (sv); |
1028 | sv = SvREFCNT_inc (POPs); |
1274 | sv = SvREFCNT_inc_NN (TOPs); |
|
|
1275 | SP -= count; |
1029 | } |
1276 | } |
1030 | else |
1277 | else |
1031 | { |
1278 | { |
1032 | AV *av = newAV (); |
1279 | AV *av = newAV (); |
1033 | av_push (av, newSVuv (tag)); |
1280 | av_push (av, tag_sv); |
1034 | av_push (av, sv); |
1281 | av_push (av, sv); |
1035 | |
1282 | |
1036 | HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash |
1283 | HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash |
1037 | ? cbor_tagged_stash |
1284 | ? cbor_tagged_stash |
1038 | : gv_stashpv ("CBOR::XS::Tagged" , 1); |
1285 | : gv_stashpv ("CBOR::XS::Tagged" , 1); |
… | |
… | |
1124 | |
1371 | |
1125 | return newSVnv (ecb_binary64_to_double (fp)); |
1372 | return newSVnv (ecb_binary64_to_double (fp)); |
1126 | } |
1373 | } |
1127 | |
1374 | |
1128 | // 0..19 unassigned simple |
1375 | // 0..19 unassigned simple |
1129 | // 24 reserved + unassigned (reserved values are not encodable) |
1376 | // 24 reserved + unassigned simple (reserved values are not encodable) |
|
|
1377 | // 28-30 unassigned misc |
|
|
1378 | // 31 break code |
1130 | default: |
1379 | default: |
1131 | ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); |
1380 | ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)"); |
1132 | } |
1381 | } |
1133 | |
1382 | |
1134 | break; |
1383 | break; |
1135 | } |
1384 | } |
1136 | |
1385 | |
… | |
… | |
1139 | } |
1388 | } |
1140 | |
1389 | |
1141 | static SV * |
1390 | static SV * |
1142 | decode_cbor (SV *string, CBOR *cbor, char **offset_return) |
1391 | decode_cbor (SV *string, CBOR *cbor, char **offset_return) |
1143 | { |
1392 | { |
1144 | dec_t dec = { }; |
1393 | dec_t dec = { 0 }; |
1145 | SV *sv; |
1394 | SV *sv; |
1146 | STRLEN len; |
1395 | STRLEN len; |
1147 | char *data = SvPVbyte (string, len); |
1396 | char *data = SvPVbyte (string, len); |
1148 | |
1397 | |
1149 | if (len > cbor->max_size && cbor->max_size) |
1398 | if (len > cbor->max_size && cbor->max_size) |
… | |
… | |
1163 | if (dec.cur != dec.end && !dec.err) |
1412 | if (dec.cur != dec.end && !dec.err) |
1164 | dec.err = "garbage after CBOR object"; |
1413 | dec.err = "garbage after CBOR object"; |
1165 | |
1414 | |
1166 | if (dec.err) |
1415 | if (dec.err) |
1167 | { |
1416 | { |
|
|
1417 | if (dec.shareable) |
|
|
1418 | { |
|
|
1419 | // need to break cyclic links, which would all be in shareable |
|
|
1420 | int i; |
|
|
1421 | SV **svp; |
|
|
1422 | |
|
|
1423 | for (i = av_len (dec.shareable) + 1; i--; ) |
|
|
1424 | if ((svp = av_fetch (dec.shareable, i, 0))) |
|
|
1425 | sv_setsv (*svp, &PL_sv_undef); |
|
|
1426 | } |
|
|
1427 | |
1168 | SvREFCNT_dec (sv); |
1428 | SvREFCNT_dec_NN (sv); |
|
|
1429 | |
|
|
1430 | if (dec.err_sv) |
|
|
1431 | sv_2mortal (dec.err_sv); |
|
|
1432 | |
1169 | croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur); |
1433 | croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur); |
1170 | } |
1434 | } |
1171 | |
1435 | |
1172 | sv = sv_2mortal (sv); |
1436 | sv = sv_2mortal (sv); |
1173 | |
1437 | |
1174 | return sv; |
1438 | return sv; |
1175 | } |
1439 | } |
1176 | |
1440 | |
|
|
1441 | ///////////////////////////////////////////////////////////////////////////// |
|
|
1442 | // incremental parser |
|
|
1443 | |
|
|
1444 | #define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0) |
|
|
1445 | |
|
|
1446 | // returns 0 for notyet, 1 for success or error |
|
|
1447 | static int |
|
|
1448 | incr_parse (CBOR *self, SV *cborstr) |
|
|
1449 | { |
|
|
1450 | STRLEN cur; |
|
|
1451 | SvPV (cborstr, cur); |
|
|
1452 | |
|
|
1453 | while (ecb_expect_true (self->incr_need <= cur)) |
|
|
1454 | { |
|
|
1455 | // table of integer count bytes |
|
|
1456 | static I8 incr_len[MINOR_MASK + 1] = { |
|
|
1457 | 0, 0, 0, 0, 0, 0, 0, 0, |
|
|
1458 | 0, 0, 0, 0, 0, 0, 0, 0, |
|
|
1459 | 0, 0, 0, 0, 0, 0, 0, 0, |
|
|
1460 | 1, 2, 4, 8,-1,-1,-1,-2 |
|
|
1461 | }; |
|
|
1462 | |
|
|
1463 | const U8 *p = SvPVX (cborstr) + self->incr_pos; |
|
|
1464 | U8 m = *p & MINOR_MASK; |
|
|
1465 | IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]); |
|
|
1466 | I8 ilen = incr_len[m]; |
|
|
1467 | |
|
|
1468 | self->incr_need = self->incr_pos + 1; |
|
|
1469 | |
|
|
1470 | if (ecb_expect_false (ilen < 0)) |
|
|
1471 | { |
|
|
1472 | if (m != MINOR_INDEF) |
|
|
1473 | return 1; // error |
|
|
1474 | |
|
|
1475 | if (*p == (MAJOR_MISC | MINOR_INDEF)) |
|
|
1476 | { |
|
|
1477 | if (count >= 0) |
|
|
1478 | return 1; // error |
|
|
1479 | |
|
|
1480 | count = 1; |
|
|
1481 | } |
|
|
1482 | else |
|
|
1483 | { |
|
|
1484 | av_push (self->incr_count, newSViv (-1)); //TODO: nest |
|
|
1485 | count = -1; |
|
|
1486 | } |
|
|
1487 | } |
|
|
1488 | else |
|
|
1489 | { |
|
|
1490 | self->incr_need += ilen; |
|
|
1491 | if (ecb_expect_false (self->incr_need > cur)) |
|
|
1492 | return 0; |
|
|
1493 | |
|
|
1494 | int major = *p >> MAJOR_SHIFT; |
|
|
1495 | |
|
|
1496 | switch (major) |
|
|
1497 | { |
|
|
1498 | case MAJOR_TAG >> MAJOR_SHIFT: |
|
|
1499 | ++count; // tags merely prefix another value |
|
|
1500 | break; |
|
|
1501 | |
|
|
1502 | case MAJOR_BYTES >> MAJOR_SHIFT: |
|
|
1503 | case MAJOR_TEXT >> MAJOR_SHIFT: |
|
|
1504 | case MAJOR_ARRAY >> MAJOR_SHIFT: |
|
|
1505 | case MAJOR_MAP >> MAJOR_SHIFT: |
|
|
1506 | { |
|
|
1507 | UV len; |
|
|
1508 | |
|
|
1509 | if (ecb_expect_false (ilen)) |
|
|
1510 | { |
|
|
1511 | len = 0; |
|
|
1512 | |
|
|
1513 | do { |
|
|
1514 | len = (len << 8) | *++p; |
|
|
1515 | } while (--ilen); |
|
|
1516 | } |
|
|
1517 | else |
|
|
1518 | len = m; |
|
|
1519 | |
|
|
1520 | switch (major) |
|
|
1521 | { |
|
|
1522 | case MAJOR_BYTES >> MAJOR_SHIFT: |
|
|
1523 | case MAJOR_TEXT >> MAJOR_SHIFT: |
|
|
1524 | self->incr_need += len; |
|
|
1525 | if (ecb_expect_false (self->incr_need > cur)) |
|
|
1526 | return 0; |
|
|
1527 | |
|
|
1528 | break; |
|
|
1529 | |
|
|
1530 | case MAJOR_MAP >> MAJOR_SHIFT: |
|
|
1531 | len <<= 1; |
|
|
1532 | case MAJOR_ARRAY >> MAJOR_SHIFT: |
|
|
1533 | if (len) |
|
|
1534 | { |
|
|
1535 | av_push (self->incr_count, newSViv (len + 1)); //TODO: nest |
|
|
1536 | count = len + 1; |
|
|
1537 | } |
|
|
1538 | break; |
|
|
1539 | } |
|
|
1540 | } |
|
|
1541 | } |
|
|
1542 | } |
|
|
1543 | |
|
|
1544 | self->incr_pos = self->incr_need; |
|
|
1545 | |
|
|
1546 | if (count > 0) |
|
|
1547 | { |
|
|
1548 | while (!--count) |
|
|
1549 | { |
|
|
1550 | if (!AvFILLp (self->incr_count)) |
|
|
1551 | return 1; // done |
|
|
1552 | |
|
|
1553 | SvREFCNT_dec_NN (av_pop (self->incr_count)); |
|
|
1554 | count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]); |
|
|
1555 | } |
|
|
1556 | |
|
|
1557 | SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count; |
|
|
1558 | } |
|
|
1559 | } |
|
|
1560 | |
|
|
1561 | return 0; |
|
|
1562 | } |
|
|
1563 | |
|
|
1564 | |
1177 | ///////////////////////////////////////////////////////////////////////////// |
1565 | ///////////////////////////////////////////////////////////////////////////// |
1178 | // XS interface functions |
1566 | // XS interface functions |
1179 | |
1567 | |
1180 | MODULE = CBOR::XS PACKAGE = CBOR::XS |
1568 | MODULE = CBOR::XS PACKAGE = CBOR::XS |
1181 | |
1569 | |
… | |
… | |
1193 | |
1581 | |
1194 | default_filter = newSVpv ("CBOR::XS::default_filter", 0); |
1582 | default_filter = newSVpv ("CBOR::XS::default_filter", 0); |
1195 | |
1583 | |
1196 | sv_cbor = newSVpv ("CBOR", 0); |
1584 | sv_cbor = newSVpv ("CBOR", 0); |
1197 | SvREADONLY_on (sv_cbor); |
1585 | SvREADONLY_on (sv_cbor); |
|
|
1586 | |
|
|
1587 | assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1)); |
1198 | } |
1588 | } |
1199 | |
1589 | |
1200 | PROTOTYPES: DISABLE |
1590 | PROTOTYPES: DISABLE |
1201 | |
1591 | |
1202 | void CLONE (...) |
1592 | void CLONE (...) |
… | |
… | |
1222 | ALIAS: |
1612 | ALIAS: |
1223 | shrink = F_SHRINK |
1613 | shrink = F_SHRINK |
1224 | allow_unknown = F_ALLOW_UNKNOWN |
1614 | allow_unknown = F_ALLOW_UNKNOWN |
1225 | allow_sharing = F_ALLOW_SHARING |
1615 | allow_sharing = F_ALLOW_SHARING |
1226 | allow_cycles = F_ALLOW_CYCLES |
1616 | allow_cycles = F_ALLOW_CYCLES |
|
|
1617 | forbid_objects = F_FORBID_OBJECTS |
1227 | pack_strings = F_PACK_STRINGS |
1618 | pack_strings = F_PACK_STRINGS |
|
|
1619 | text_keys = F_TEXT_KEYS |
|
|
1620 | text_strings = F_TEXT_STRINGS |
|
|
1621 | validate_utf8 = F_VALIDATE_UTF8 |
1228 | PPCODE: |
1622 | PPCODE: |
1229 | { |
1623 | { |
1230 | if (enable) |
1624 | if (enable) |
1231 | self->flags |= ix; |
1625 | self->flags |= ix; |
1232 | else |
1626 | else |
… | |
… | |
1239 | ALIAS: |
1633 | ALIAS: |
1240 | get_shrink = F_SHRINK |
1634 | get_shrink = F_SHRINK |
1241 | get_allow_unknown = F_ALLOW_UNKNOWN |
1635 | get_allow_unknown = F_ALLOW_UNKNOWN |
1242 | get_allow_sharing = F_ALLOW_SHARING |
1636 | get_allow_sharing = F_ALLOW_SHARING |
1243 | get_allow_cycles = F_ALLOW_CYCLES |
1637 | get_allow_cycles = F_ALLOW_CYCLES |
|
|
1638 | get_forbid_objects = F_FORBID_OBJECTS |
1244 | get_pack_strings = F_PACK_STRINGS |
1639 | get_pack_strings = F_PACK_STRINGS |
|
|
1640 | get_text_keys = F_TEXT_KEYS |
|
|
1641 | get_text_strings = F_TEXT_STRINGS |
|
|
1642 | get_validate_utf8 = F_VALIDATE_UTF8 |
1245 | PPCODE: |
1643 | PPCODE: |
1246 | XPUSHs (boolSV (self->flags & ix)); |
1644 | XPUSHs (boolSV (self->flags & ix)); |
1247 | |
1645 | |
1248 | void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) |
1646 | void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) |
1249 | PPCODE: |
1647 | PPCODE: |
… | |
… | |
1298 | EXTEND (SP, 2); |
1696 | EXTEND (SP, 2); |
1299 | PUSHs (sv); |
1697 | PUSHs (sv); |
1300 | PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); |
1698 | PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); |
1301 | } |
1699 | } |
1302 | |
1700 | |
|
|
1701 | void incr_parse (CBOR *self, SV *cborstr) |
|
|
1702 | ALIAS: |
|
|
1703 | incr_parse_multiple = 1 |
|
|
1704 | PPCODE: |
|
|
1705 | { |
|
|
1706 | if (SvUTF8 (cborstr)) |
|
|
1707 | sv_utf8_downgrade (cborstr, 0); |
|
|
1708 | |
|
|
1709 | if (!self->incr_count) |
|
|
1710 | { |
|
|
1711 | self->incr_count = newAV (); |
|
|
1712 | self->incr_pos = 0; |
|
|
1713 | self->incr_need = 1; |
|
|
1714 | |
|
|
1715 | av_push (self->incr_count, newSViv (1)); |
|
|
1716 | } |
|
|
1717 | |
|
|
1718 | do |
|
|
1719 | { |
|
|
1720 | if (!incr_parse (self, cborstr)) |
|
|
1721 | { |
|
|
1722 | if (self->incr_need > self->max_size && self->max_size) |
|
|
1723 | croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu", |
|
|
1724 | (unsigned long)self->incr_need, (unsigned long)self->max_size); |
|
|
1725 | |
|
|
1726 | break; |
|
|
1727 | } |
|
|
1728 | |
|
|
1729 | SV *sv; |
|
|
1730 | char *offset; |
|
|
1731 | |
|
|
1732 | PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN; |
|
|
1733 | XPUSHs (sv); |
|
|
1734 | |
|
|
1735 | sv_chop (cborstr, offset); |
|
|
1736 | |
|
|
1737 | av_clear (self->incr_count); |
|
|
1738 | av_push (self->incr_count, newSViv (1)); |
|
|
1739 | |
|
|
1740 | self->incr_pos = 0; |
|
|
1741 | self->incr_need = self->incr_pos + 1; |
|
|
1742 | } |
|
|
1743 | while (ix); |
|
|
1744 | } |
|
|
1745 | |
|
|
1746 | void incr_reset (CBOR *self) |
|
|
1747 | CODE: |
|
|
1748 | { |
|
|
1749 | SvREFCNT_dec (self->incr_count); |
|
|
1750 | self->incr_count = 0; |
|
|
1751 | } |
|
|
1752 | |
1303 | void DESTROY (CBOR *self) |
1753 | void DESTROY (CBOR *self) |
1304 | PPCODE: |
1754 | PPCODE: |
1305 | cbor_free (self); |
1755 | cbor_free (self); |
1306 | |
1756 | |
1307 | PROTOTYPES: ENABLE |
1757 | PROTOTYPES: ENABLE |
… | |
… | |
1326 | cbor_init (&cbor); |
1776 | cbor_init (&cbor); |
1327 | PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN; |
1777 | PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN; |
1328 | XPUSHs (cborstr); |
1778 | XPUSHs (cborstr); |
1329 | } |
1779 | } |
1330 | |
1780 | |
|
|
1781 | #ifdef __AFL_COMPILER |
|
|
1782 | |
|
|
1783 | void |
|
|
1784 | afl_init () |
|
|
1785 | CODE: |
|
|
1786 | __AFL_INIT (); |
|
|
1787 | |
|
|
1788 | int |
|
|
1789 | afl_loop (unsigned int count = 10000) |
|
|
1790 | CODE: |
|
|
1791 | RETVAL = __AFL_LOOP (count); |
|
|
1792 | OUTPUT: |
|
|
1793 | RETVAL |
|
|
1794 | |
|
|
1795 | #endif |
|
|
1796 | |