ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/CBOR-XS/XS.xs
(Generate patch)

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.29 by root, Thu Nov 28 09:13:12 2013 UTC vs.
Revision 1.38 by root, Sun Dec 1 14:30:52 2013 UTC

23#endif 23#endif
24#ifndef SvREFCNT_dec_NN 24#ifndef SvREFCNT_dec_NN
25# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv) 25# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
26#endif 26#endif
27 27
28// known major and minor types
29enum cbor_type
30{
31 MAJOR_SHIFT = 5,
32 MINOR_MASK = 0x1f,
33
34 MAJOR_POS_INT = 0 << MAJOR_SHIFT,
35 MAJOR_NEG_INT = 1 << MAJOR_SHIFT,
36 MAJOR_BYTES = 2 << MAJOR_SHIFT,
37 MAJOR_TEXT = 3 << MAJOR_SHIFT,
38 MAJOR_ARRAY = 4 << MAJOR_SHIFT,
39 MAJOR_MAP = 5 << MAJOR_SHIFT,
40 MAJOR_TAG = 6 << MAJOR_SHIFT,
41 MAJOR_MISC = 7 << MAJOR_SHIFT,
42
43 // INT/STRING/ARRAY/MAP subtypes
44 LENGTH_EXT1 = 24,
45 LENGTH_EXT2 = 25,
46 LENGTH_EXT4 = 26,
47 LENGTH_EXT8 = 27,
48
49 // SIMPLE types (effectively MISC subtypes)
50 SIMPLE_FALSE = 20,
51 SIMPLE_TRUE = 21,
52 SIMPLE_NULL = 22,
53 SIMPLE_UNDEF = 23,
54
55 // MISC subtype (unused)
56 MISC_EXT1 = 24,
57 MISC_FLOAT16 = 25,
58 MISC_FLOAT32 = 26,
59 MISC_FLOAT64 = 27,
60
61 // BYTES/TEXT/ARRAY/MAP
62 MINOR_INDEF = 31,
63};
64
28// known tags 65// known tags
29enum cbor_tag 66enum cbor_tag
30{ 67{
31 // extensions 68 // extensions
32 CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref 69 CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
33 CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object 70 CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
34 CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object 71 CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
35 CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing 72 CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
36 CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing 73 CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
37 CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref 74 CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref
38 CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection 75 CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
39 76
40 // rfc7049 77 // rfc7049
41 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8 78 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
42 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any 79 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
43 CBOR_TAG_POS_BIGNUM = 2, // byte string 80 CBOR_TAG_POS_BIGNUM = 2, // byte string
44 CBOR_TAG_NEG_BIGNUM = 3, // byte string 81 CBOR_TAG_NEG_BIGNUM = 3, // byte string
45 CBOR_TAG_DECIMAL = 4, // decimal fraction, array 82 CBOR_TAG_DECIMAL = 4, // decimal fraction, array
46 CBOR_TAG_BIGFLOAT = 5, // array 83 CBOR_TAG_BIGFLOAT = 5, // array
47 84
48 CBOR_TAG_CONV_B64U = 21, // base64url, any 85 CBOR_TAG_CONV_B64U = 21, // base64url, any
49 CBOR_TAG_CONV_B64 = 22, // base64, any 86 CBOR_TAG_CONV_B64 = 22, // base64, any
50 CBOR_TAG_CONV_HEX = 23, // base16, any 87 CBOR_TAG_CONV_HEX = 23, // base16, any
51 CBOR_TAG_CBOR = 24, // embedded cbor, byte string 88 CBOR_TAG_CBOR = 24, // embedded cbor, byte string
52 89
53 CBOR_TAG_URI = 32, // URI rfc3986, utf-8 90 CBOR_TAG_URI = 32, // URI rfc3986, utf-8
54 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8 91 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
55 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8 92 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
56 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8 93 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
57 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8 94 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
58 95
59 CBOR_TAG_MAGIC = 55799 // self-describe cbor 96 CBOR_TAG_MAGIC = 55799, // self-describe cbor
60}; 97};
61 98
62#define F_SHRINK 0x00000001UL 99#define F_SHRINK 0x00000001UL
63#define F_ALLOW_UNKNOWN 0x00000002UL 100#define F_ALLOW_UNKNOWN 0x00000002UL
64#define F_ALLOW_SHARING 0x00000004UL //TODO 101#define F_ALLOW_SHARING 0x00000004UL
65#define F_ALLOW_STRINGREF 0x00000008UL //TODO 102#define F_ALLOW_CYCLES 0x00000008UL
103#define F_PACK_STRINGS 0x00000010UL
104#define F_VALIDATE_UTF8 0x00000020UL
66 105
67#define INIT_SIZE 32 // initial scalar size to be allocated 106#define INIT_SIZE 32 // initial scalar size to be allocated
68 107
69#define SB do { 108#define SB do {
70#define SE } while (0) 109#define SE } while (0)
189static void 228static void
190encode_uint (enc_t *enc, int major, UV len) 229encode_uint (enc_t *enc, int major, UV len)
191{ 230{
192 need (enc, 9); 231 need (enc, 9);
193 232
194 if (len < 24) 233 if (ecb_expect_true (len < LENGTH_EXT1))
195 *enc->cur++ = major | len; 234 *enc->cur++ = major | len;
196 else if (len <= 0xff) 235 else if (ecb_expect_true (len <= 0xffU))
197 { 236 {
198 *enc->cur++ = major | 24; 237 *enc->cur++ = major | LENGTH_EXT1;
199 *enc->cur++ = len; 238 *enc->cur++ = len;
200 } 239 }
201 else if (len <= 0xffff) 240 else if (len <= 0xffffU)
202 { 241 {
203 *enc->cur++ = major | 25; 242 *enc->cur++ = major | LENGTH_EXT2;
204 *enc->cur++ = len >> 8; 243 *enc->cur++ = len >> 8;
205 *enc->cur++ = len; 244 *enc->cur++ = len;
206 } 245 }
207 else if (len <= 0xffffffff) 246 else if (len <= 0xffffffffU)
208 { 247 {
209 *enc->cur++ = major | 26; 248 *enc->cur++ = major | LENGTH_EXT4;
210 *enc->cur++ = len >> 24; 249 *enc->cur++ = len >> 24;
211 *enc->cur++ = len >> 16; 250 *enc->cur++ = len >> 16;
212 *enc->cur++ = len >> 8; 251 *enc->cur++ = len >> 8;
213 *enc->cur++ = len; 252 *enc->cur++ = len;
214 } 253 }
215 else 254 else
216 { 255 {
217 *enc->cur++ = major | 27; 256 *enc->cur++ = major | LENGTH_EXT8;
218 *enc->cur++ = len >> 56; 257 *enc->cur++ = len >> 56;
219 *enc->cur++ = len >> 48; 258 *enc->cur++ = len >> 48;
220 *enc->cur++ = len >> 40; 259 *enc->cur++ = len >> 40;
221 *enc->cur++ = len >> 32; 260 *enc->cur++ = len >> 32;
222 *enc->cur++ = len >> 24; 261 *enc->cur++ = len >> 24;
227} 266}
228 267
229ecb_inline void 268ecb_inline void
230encode_tag (enc_t *enc, UV tag) 269encode_tag (enc_t *enc, UV tag)
231{ 270{
232 encode_uint (enc, 0xc0, tag); 271 encode_uint (enc, MAJOR_TAG, tag);
272}
273
274ecb_inline void
275encode_str (enc_t *enc, int utf8, char *str, STRLEN len)
276{
277 encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
278 need (enc, len);
279 memcpy (enc->cur, str, len);
280 enc->cur += len;
233} 281}
234 282
235static void 283static void
236encode_str (enc_t *enc, int utf8, char *str, STRLEN len) 284encode_strref (enc_t *enc, int utf8, char *str, STRLEN len)
237{ 285{
238 if (ecb_expect_false (enc->cbor.flags & F_ALLOW_STRINGREF)) 286 if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
239 { 287 {
240 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1); 288 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
241 289
242 if (SvOK (*svp)) 290 if (SvOK (*svp))
243 { 291 {
244 // already registered, use stringref 292 // already registered, use stringref
245 encode_tag (enc, CBOR_TAG_STRINGREF); 293 encode_tag (enc, CBOR_TAG_STRINGREF);
246 encode_uint (enc, 0x00, SvUV (*svp)); 294 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
247 return; 295 return;
248 } 296 }
249 else if (len >= minimum_string_length (enc->stringref_idx)) 297 else if (len >= minimum_string_length (enc->stringref_idx))
250 { 298 {
251 // register only 299 // register only
252 sv_setuv (*svp, enc->stringref_idx); 300 sv_setuv (*svp, enc->stringref_idx);
253 ++enc->stringref_idx; 301 ++enc->stringref_idx;
254 } 302 }
255 } 303 }
256 304
257 encode_uint (enc, utf8 ? 0x60 : 0x40, len); 305 encode_str (enc, utf8, str, len);
258 need (enc, len);
259 memcpy (enc->cur, str, len);
260 enc->cur += len;
261} 306}
262 307
263static void encode_sv (enc_t *enc, SV *sv); 308static void encode_sv (enc_t *enc, SV *sv);
264 309
265static void 310static void
270 if (enc->depth >= enc->cbor.max_depth) 315 if (enc->depth >= enc->cbor.max_depth)
271 croak (ERR_NESTING_EXCEEDED); 316 croak (ERR_NESTING_EXCEEDED);
272 317
273 ++enc->depth; 318 ++enc->depth;
274 319
275 encode_uint (enc, 0x80, len + 1); 320 encode_uint (enc, MAJOR_ARRAY, len + 1);
276 321
277 for (i = 0; i <= len; ++i) 322 for (i = 0; i <= len; ++i)
278 { 323 {
279 SV **svp = av_fetch (av, i, 0); 324 SV **svp = av_fetch (av, i, 0);
280 encode_sv (enc, svp ? *svp : &PL_sv_undef); 325 encode_sv (enc, svp ? *svp : &PL_sv_undef);
295 340
296 int pairs = hv_iterinit (hv); 341 int pairs = hv_iterinit (hv);
297 int mg = SvMAGICAL (hv); 342 int mg = SvMAGICAL (hv);
298 343
299 if (mg) 344 if (mg)
300 encode_ch (enc, 0xa0 | 31); 345 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
301 else 346 else
302 encode_uint (enc, 0xa0, pairs); 347 encode_uint (enc, MAJOR_MAP, pairs);
303 348
304 while ((he = hv_iternext (hv))) 349 while ((he = hv_iternext (hv)))
305 { 350 {
306 if (HeKLEN (he) == HEf_SVKEY) 351 if (HeKLEN (he) == HEf_SVKEY)
307 encode_sv (enc, HeSVKEY (he)); 352 encode_sv (enc, HeSVKEY (he));
308 else 353 else
309 encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); 354 encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
310 355
311 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); 356 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
312 } 357 }
313 358
314 if (mg) 359 if (mg)
315 encode_ch (enc, 0xe0 | 31); 360 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
316 361
317 --enc->depth; 362 --enc->depth;
318} 363}
319 364
320// encode objects, arrays and special \0=false and \1=true values. 365// encode objects, arrays and special \0=false and \1=true values.
321static void 366static void
322encode_rv (enc_t *enc, SV *sv) 367encode_rv (enc_t *enc, SV *sv)
323{ 368{
324 SvGETMAGIC (sv); 369 SvGETMAGIC (sv);
325
326 if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)
327 && ecb_expect_false (SvREFCNT (sv) > 1))
328 {
329 if (!enc->shareable)
330 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
331
332 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
333
334 if (SvOK (*svp))
335 {
336 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
337 encode_uint (enc, 0x00, SvUV (*svp));
338 return;
339 }
340 else
341 {
342 sv_setuv (*svp, enc->shareable_idx);
343 ++enc->shareable_idx;
344 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
345 }
346 }
347 370
348 svtype svt = SvTYPE (sv); 371 svtype svt = SvTYPE (sv);
349 372
350 if (ecb_expect_false (SvOBJECT (sv))) 373 if (ecb_expect_false (SvOBJECT (sv)))
351 { 374 {
358 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 381 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
359 ? cbor_tagged_stash 382 ? cbor_tagged_stash
360 : gv_stashpv ("CBOR::XS::Tagged" , 1); 383 : gv_stashpv ("CBOR::XS::Tagged" , 1);
361 384
362 HV *stash = SvSTASH (sv); 385 HV *stash = SvSTASH (sv);
363 GV *method;
364 386
365 if (stash == boolean_stash) 387 if (stash == boolean_stash)
366 encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20); 388 {
389 encode_ch (enc, SvIV (sv) ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE);
390 return;
391 }
367 else if (stash == error_stash) 392 else if (stash == error_stash)
368 encode_ch (enc, 0xe0 | 23); 393 {
394 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
395 return;
396 }
369 else if (stash == tagged_stash) 397 else if (stash == tagged_stash)
370 { 398 {
371 if (svt != SVt_PVAV) 399 if (svt != SVt_PVAV)
372 croak ("encountered CBOR::XS::Tagged object that isn't an array"); 400 croak ("encountered CBOR::XS::Tagged object that isn't an array");
373 401
374 encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); 402 encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
375 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); 403 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1));
404
405 return;
406 }
407 }
408
409 if (ecb_expect_false (SvREFCNT (sv) > 1)
410 && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
411 {
412 if (!enc->shareable)
413 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
414
415 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
416
417 if (SvOK (*svp))
376 } 418 {
419 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
420 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
421 return;
422 }
423 else
424 {
425 sv_setuv (*svp, enc->shareable_idx);
426 ++enc->shareable_idx;
427 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
428 }
429 }
430
431 if (ecb_expect_false (SvOBJECT (sv)))
432 {
433 HV *stash = SvSTASH (sv);
434 GV *method;
435
377 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) 436 if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
378 { 437 {
379 dSP; 438 dSP;
380 439
381 ENTER; SAVETMPS; PUSHMARK (SP); 440 ENTER; SAVETMPS; PUSHMARK (SP);
382 // we re-bless the reference to get overload and other niceties right 441 // we re-bless the reference to get overload and other niceties right
414 // catch this surprisingly common error 473 // catch this surprisingly common error
415 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) 474 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
416 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash)); 475 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
417 476
418 encode_tag (enc, CBOR_TAG_PERL_OBJECT); 477 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
419 encode_uint (enc, 0x80, count + 1); 478 encode_uint (enc, MAJOR_ARRAY, count + 1);
420 encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); 479 encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
421 480
422 while (count) 481 while (count)
423 encode_sv (enc, SP[1 - count--]); 482 encode_sv (enc, SP[1 - count--]);
424 483
425 PUTBACK; 484 PUTBACK;
446{ 505{
447 double nv = SvNVX (sv); 506 double nv = SvNVX (sv);
448 507
449 need (enc, 9); 508 need (enc, 9);
450 509
451 if (ecb_expect_false (nv == (U32)nv)) 510 if (ecb_expect_false (nv == (NV)(U32)nv))
452 encode_uint (enc, 0x00, (U32)nv); 511 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
453 //TODO: maybe I32? 512 //TODO: maybe I32?
454 else if (ecb_expect_false (nv == (float)nv)) 513 else if (ecb_expect_false (nv == (float)nv))
455 { 514 {
456 uint32_t fp = ecb_float_to_binary32 (nv); 515 uint32_t fp = ecb_float_to_binary32 (nv);
457 516
458 *enc->cur++ = 0xe0 | 26; 517 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
459 518
460 if (!ecb_big_endian ()) 519 if (!ecb_big_endian ())
461 fp = ecb_bswap32 (fp); 520 fp = ecb_bswap32 (fp);
462 521
463 memcpy (enc->cur, &fp, 4); 522 memcpy (enc->cur, &fp, 4);
465 } 524 }
466 else 525 else
467 { 526 {
468 uint64_t fp = ecb_double_to_binary64 (nv); 527 uint64_t fp = ecb_double_to_binary64 (nv);
469 528
470 *enc->cur++ = 0xe0 | 27; 529 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
471 530
472 if (!ecb_big_endian ()) 531 if (!ecb_big_endian ())
473 fp = ecb_bswap64 (fp); 532 fp = ecb_bswap64 (fp);
474 533
475 memcpy (enc->cur, &fp, 8); 534 memcpy (enc->cur, &fp, 8);
484 543
485 if (SvPOKp (sv)) 544 if (SvPOKp (sv))
486 { 545 {
487 STRLEN len; 546 STRLEN len;
488 char *str = SvPV (sv, len); 547 char *str = SvPV (sv, len);
489 encode_str (enc, SvUTF8 (sv), str, len); 548 encode_strref (enc, SvUTF8 (sv), str, len);
490 } 549 }
491 else if (SvNOKp (sv)) 550 else if (SvNOKp (sv))
492 encode_nv (enc, sv); 551 encode_nv (enc, sv);
493 else if (SvIOKp (sv)) 552 else if (SvIOKp (sv))
494 { 553 {
495 if (SvIsUV (sv)) 554 if (SvIsUV (sv))
496 encode_uint (enc, 0x00, SvUVX (sv)); 555 encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
497 else if (SvIVX (sv) >= 0) 556 else if (SvIVX (sv) >= 0)
498 encode_uint (enc, 0x00, SvIVX (sv)); 557 encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
499 else 558 else
500 encode_uint (enc, 0x20, -(SvIVX (sv) + 1)); 559 encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
501 } 560 }
502 else if (SvROK (sv)) 561 else if (SvROK (sv))
503 encode_rv (enc, SvRV (sv)); 562 encode_rv (enc, SvRV (sv));
504 else if (!SvOK (sv)) 563 else if (!SvOK (sv))
505 encode_ch (enc, 0xe0 | 22); 564 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
506 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 565 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
507 encode_ch (enc, 0xe0 | 23); 566 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
508 else 567 else
509 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data", 568 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
510 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 569 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
511} 570}
512 571
520 enc.cur = SvPVX (enc.sv); 579 enc.cur = SvPVX (enc.sv);
521 enc.end = SvEND (enc.sv); 580 enc.end = SvEND (enc.sv);
522 581
523 SvPOK_only (enc.sv); 582 SvPOK_only (enc.sv);
524 583
525 if (cbor->flags & F_ALLOW_STRINGREF) 584 if (cbor->flags & F_PACK_STRINGS)
526 { 585 {
527 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE); 586 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
528 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ()); 587 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
529 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ()); 588 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
530 } 589 }
565#define DEC_DEC_DEPTH --dec->depth 624#define DEC_DEC_DEPTH --dec->depth
566 625
567static UV 626static UV
568decode_uint (dec_t *dec) 627decode_uint (dec_t *dec)
569{ 628{
570 switch (*dec->cur & 31) 629 U8 m = *dec->cur & MINOR_MASK;
571 { 630 ++dec->cur;
572 case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
573 case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
574 case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
575 return *dec->cur++ & 31;
576 631
577 case 24: 632 if (ecb_expect_true (m < LENGTH_EXT1))
633 return m;
634 else if (ecb_expect_true (m == LENGTH_EXT1))
635 {
578 WANT (2); 636 WANT (1);
579 dec->cur += 2; 637 dec->cur += 1;
580 return dec->cur[-1]; 638 return dec->cur[-1];
581 639 }
582 case 25: 640 else if (ecb_expect_true (m == LENGTH_EXT2))
641 {
583 WANT (3); 642 WANT (2);
584 dec->cur += 3; 643 dec->cur += 2;
585 return (((UV)dec->cur[-2]) << 8) 644 return (((UV)dec->cur[-2]) << 8)
586 | ((UV)dec->cur[-1]); 645 | ((UV)dec->cur[-1]);
587 646 }
588 case 26: 647 else if (ecb_expect_true (m == LENGTH_EXT4))
648 {
589 WANT (5); 649 WANT (4);
590 dec->cur += 5; 650 dec->cur += 4;
591 return (((UV)dec->cur[-4]) << 24) 651 return (((UV)dec->cur[-4]) << 24)
592 | (((UV)dec->cur[-3]) << 16) 652 | (((UV)dec->cur[-3]) << 16)
593 | (((UV)dec->cur[-2]) << 8) 653 | (((UV)dec->cur[-2]) << 8)
594 | ((UV)dec->cur[-1]); 654 | ((UV)dec->cur[-1]);
595 655 }
596 case 27: 656 else if (ecb_expect_true (m == LENGTH_EXT8))
657 {
597 WANT (9); 658 WANT (8);
598 dec->cur += 9; 659 dec->cur += 8;
660
661 return
662#if UVSIZE < 8
663 0
664#else
599 return (((UV)dec->cur[-8]) << 56) 665 (((UV)dec->cur[-8]) << 56)
600 | (((UV)dec->cur[-7]) << 48) 666 | (((UV)dec->cur[-7]) << 48)
601 | (((UV)dec->cur[-6]) << 40) 667 | (((UV)dec->cur[-6]) << 40)
602 | (((UV)dec->cur[-5]) << 32) 668 | (((UV)dec->cur[-5]) << 32)
669#endif
603 | (((UV)dec->cur[-4]) << 24) 670 | (((UV)dec->cur[-4]) << 24)
604 | (((UV)dec->cur[-3]) << 16) 671 | (((UV)dec->cur[-3]) << 16)
605 | (((UV)dec->cur[-2]) << 8) 672 | (((UV)dec->cur[-2]) << 8)
606 | ((UV)dec->cur[-1]); 673 | ((UV)dec->cur[-1]);
607 674 }
608 default: 675 else
609 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 676 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
610 }
611 677
612fail: 678fail:
613 return 0; 679 return 0;
614} 680}
615 681
620{ 686{
621 AV *av = newAV (); 687 AV *av = newAV ();
622 688
623 DEC_INC_DEPTH; 689 DEC_INC_DEPTH;
624 690
625 if ((*dec->cur & 31) == 31) 691 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
626 { 692 {
627 ++dec->cur; 693 ++dec->cur;
628 694
629 for (;;) 695 for (;;)
630 { 696 {
631 WANT (1); 697 WANT (1);
632 698
633 if (*dec->cur == (0xe0 | 31)) 699 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
634 { 700 {
635 ++dec->cur; 701 ++dec->cur;
636 break; 702 break;
637 } 703 }
638 704
641 } 707 }
642 else 708 else
643 { 709 {
644 int i, len = decode_uint (dec); 710 int i, len = decode_uint (dec);
645 711
712 WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
646 av_fill (av, len - 1); 713 av_fill (av, len - 1);
647 714
648 for (i = 0; i < len; ++i) 715 for (i = 0; i < len; ++i)
649 AvARRAY (av)[i] = decode_sv (dec); 716 AvARRAY (av)[i] = decode_sv (dec);
650 } 717 }
663{ 730{
664 // for speed reasons, we specialcase single-string 731 // for speed reasons, we specialcase single-string
665 // byte or utf-8 strings as keys, but only when !stringref 732 // byte or utf-8 strings as keys, but only when !stringref
666 733
667 if (ecb_expect_true (!dec->stringref)) 734 if (ecb_expect_true (!dec->stringref))
668 if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27) 735 if (ecb_expect_true ((*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
669 { 736 {
670 I32 len = decode_uint (dec); 737 I32 len = decode_uint (dec);
671 char *key = (char *)dec->cur; 738 char *key = (char *)dec->cur;
672 739
673 dec->cur += len; 740 dec->cur += len;
674 741
675 if (ecb_expect_false (dec->stringref))
676 av_push (dec->stringref, newSVpvn (key, len));
677
678 hv_store (hv, key, len, decode_sv (dec), 0); 742 hv_store (hv, key, len, decode_sv (dec), 0);
679 743
680 return; 744 return;
681 } 745 }
682 else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27) 746 else if (ecb_expect_true ((*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
683 { 747 {
684 I32 len = decode_uint (dec); 748 I32 len = decode_uint (dec);
685 char *key = (char *)dec->cur; 749 char *key = (char *)dec->cur;
686 750
687 dec->cur += len; 751 dec->cur += len;
688 752
689 if (ecb_expect_false (dec->stringref)) 753 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
690 av_push (dec->stringref, newSVpvn_utf8 (key, len, 1)); 754 if (!is_utf8_string (key, len))
755 ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
691 756
692 hv_store (hv, key, -len, decode_sv (dec), 0); 757 hv_store (hv, key, -len, decode_sv (dec), 0);
693 758
694 return; 759 return;
695 } 760 }
697 SV *k = decode_sv (dec); 762 SV *k = decode_sv (dec);
698 SV *v = decode_sv (dec); 763 SV *v = decode_sv (dec);
699 764
700 hv_store_ent (hv, k, v, 0); 765 hv_store_ent (hv, k, v, 0);
701 SvREFCNT_dec (k); 766 SvREFCNT_dec (k);
767
768fail:
769 ;
702} 770}
703 771
704static SV * 772static SV *
705decode_hv (dec_t *dec) 773decode_hv (dec_t *dec)
706{ 774{
707 HV *hv = newHV (); 775 HV *hv = newHV ();
708 776
709 DEC_INC_DEPTH; 777 DEC_INC_DEPTH;
710 778
711 if ((*dec->cur & 31) == 31) 779 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
712 { 780 {
713 ++dec->cur; 781 ++dec->cur;
714 782
715 for (;;) 783 for (;;)
716 { 784 {
717 WANT (1); 785 WANT (1);
718 786
719 if (*dec->cur == (0xe0 | 31)) 787 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
720 { 788 {
721 ++dec->cur; 789 ++dec->cur;
722 break; 790 break;
723 } 791 }
724 792
745static SV * 813static SV *
746decode_str (dec_t *dec, int utf8) 814decode_str (dec_t *dec, int utf8)
747{ 815{
748 SV *sv = 0; 816 SV *sv = 0;
749 817
750 if ((*dec->cur & 31) == 31) 818 if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
751 { 819 {
820 // indefinite length strings
752 ++dec->cur; 821 ++dec->cur;
753 822
823 U8 major = *dec->cur & MAJOR_MISC;
824
754 sv = newSVpvn ("", 0); 825 sv = newSVpvn ("", 0);
755 826
756 // not very fast, and certainly not robust against illegal input
757 for (;;) 827 for (;;)
758 { 828 {
759 WANT (1); 829 WANT (1);
760 830
761 if (*dec->cur == (0xe0 | 31)) 831 if ((*dec->cur - major) > LENGTH_EXT8)
832 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
762 { 833 {
763 ++dec->cur; 834 ++dec->cur;
764 break; 835 break;
765 } 836 }
837 else
838 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
766 839
767 sv_catsv (sv, decode_sv (dec)); 840 STRLEN len = decode_uint (dec);
841
842 WANT (len);
843 sv_catpvn (sv, dec->cur, len);
844 dec->cur += len;
768 } 845 }
769 } 846 }
770 else 847 else
771 { 848 {
772 STRLEN len = decode_uint (dec); 849 STRLEN len = decode_uint (dec);
779 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1)) 856 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
780 av_push (dec->stringref, SvREFCNT_inc_NN (sv)); 857 av_push (dec->stringref, SvREFCNT_inc_NN (sv));
781 } 858 }
782 859
783 if (utf8) 860 if (utf8)
861 {
862 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
863 if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
864 ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
865
784 SvUTF8_on (sv); 866 SvUTF8_on (sv);
867 }
785 868
786 return sv; 869 return sv;
787 870
788fail: 871fail:
789 SvREFCNT_dec (sv); 872 SvREFCNT_dec (sv);
821 } 904 }
822 break; 905 break;
823 906
824 case CBOR_TAG_STRINGREF: 907 case CBOR_TAG_STRINGREF:
825 { 908 {
826 if ((*dec->cur >> 5) != 0) 909 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
827 ERR ("corrupted CBOR data (stringref index not an unsigned integer)"); 910 ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
828 911
829 UV idx = decode_uint (dec); 912 UV idx = decode_uint (dec);
830 913
831 if (!dec->stringref || (int)idx > AvFILLp (dec->stringref)) 914 if (!dec->stringref || (int)idx > AvFILLp (dec->stringref))
838 case CBOR_TAG_VALUE_SHAREABLE: 921 case CBOR_TAG_VALUE_SHAREABLE:
839 { 922 {
840 if (ecb_expect_false (!dec->shareable)) 923 if (ecb_expect_false (!dec->shareable))
841 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ()); 924 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
842 925
926 if (dec->cbor.flags & F_ALLOW_CYCLES)
927 {
843 sv = newSV (0); 928 sv = newSV (0);
844 av_push (dec->shareable, SvREFCNT_inc_NN (sv)); 929 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
845 930
846 SV *osv = decode_sv (dec); 931 SV *osv = decode_sv (dec);
847 sv_setsv (sv, osv); 932 sv_setsv (sv, osv);
848 SvREFCNT_dec_NN (osv); 933 SvREFCNT_dec_NN (osv);
934 }
935 else
936 {
937 av_push (dec->shareable, &PL_sv_undef);
938 int idx = AvFILLp (dec->shareable);
939 sv = decode_sv (dec);
940 av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
941 }
849 } 942 }
850 break; 943 break;
851 944
852 case CBOR_TAG_VALUE_SHAREDREF: 945 case CBOR_TAG_VALUE_SHAREDREF:
853 { 946 {
854 if ((*dec->cur >> 5) != 0) 947 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
855 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)"); 948 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
856 949
857 UV idx = decode_uint (dec); 950 UV idx = decode_uint (dec);
858 951
859 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable)) 952 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable))
860 ERR ("corrupted CBOR data (sharedref index out of bounds)"); 953 ERR ("corrupted CBOR data (sharedref index out of bounds)");
861 954
862 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]); 955 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
956
957 if (sv == &PL_sv_undef)
958 ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
863 } 959 }
864 break; 960 break;
865 961
866 case CBOR_TAG_PERL_OBJECT: 962 case CBOR_TAG_PERL_OBJECT:
867 { 963 {
968static SV * 1064static SV *
969decode_sv (dec_t *dec) 1065decode_sv (dec_t *dec)
970{ 1066{
971 WANT (1); 1067 WANT (1);
972 1068
973 switch (*dec->cur >> 5) 1069 switch (*dec->cur >> MAJOR_SHIFT)
974 { 1070 {
975 case 0: // unsigned int 1071 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
976 return newSVuv (decode_uint (dec)); 1072 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
977 case 1: // negative int 1073 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
978 return newSViv (-1 - (IV)decode_uint (dec)); 1074 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
979 case 2: // octet string 1075 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
980 return decode_str (dec, 0); 1076 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
981 case 3: // utf-8 string 1077 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
982 return decode_str (dec, 1); 1078
983 case 4: // array 1079 case MAJOR_MISC >> MAJOR_SHIFT:
984 return decode_av (dec);
985 case 5: // map
986 return decode_hv (dec);
987 case 6: // tag
988 return decode_tagged (dec);
989 case 7: // misc
990 switch (*dec->cur++ & 31) 1080 switch (*dec->cur++ & MINOR_MASK)
991 { 1081 {
992 case 20: 1082 case SIMPLE_FALSE:
993#if CBOR_SLOW 1083#if CBOR_SLOW
994 types_false = get_bool ("Types::Serialiser::false"); 1084 types_false = get_bool ("Types::Serialiser::false");
995#endif 1085#endif
996 return newSVsv (types_false); 1086 return newSVsv (types_false);
997 case 21: 1087 case SIMPLE_TRUE:
998#if CBOR_SLOW 1088#if CBOR_SLOW
999 types_true = get_bool ("Types::Serialiser::true"); 1089 types_true = get_bool ("Types::Serialiser::true");
1000#endif 1090#endif
1001 return newSVsv (types_true); 1091 return newSVsv (types_true);
1002 case 22: 1092 case SIMPLE_NULL:
1003 return newSVsv (&PL_sv_undef); 1093 return newSVsv (&PL_sv_undef);
1004 case 23: 1094 case SIMPLE_UNDEF:
1005#if CBOR_SLOW 1095#if CBOR_SLOW
1006 types_error = get_bool ("Types::Serialiser::error"); 1096 types_error = get_bool ("Types::Serialiser::error");
1007#endif 1097#endif
1008 return newSVsv (types_error); 1098 return newSVsv (types_error);
1009 1099
1010 case 25: 1100 case MISC_FLOAT16:
1011 { 1101 {
1012 WANT (2); 1102 WANT (2);
1013 1103
1014 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1104 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
1015 dec->cur += 2; 1105 dec->cur += 2;
1016 1106
1017 return newSVnv (ecb_binary16_to_float (fp)); 1107 return newSVnv (ecb_binary16_to_float (fp));
1018 } 1108 }
1019 1109
1020 case 26: 1110 case MISC_FLOAT32:
1021 { 1111 {
1022 uint32_t fp; 1112 uint32_t fp;
1023 WANT (4); 1113 WANT (4);
1024 memcpy (&fp, dec->cur, 4); 1114 memcpy (&fp, dec->cur, 4);
1025 dec->cur += 4; 1115 dec->cur += 4;
1028 fp = ecb_bswap32 (fp); 1118 fp = ecb_bswap32 (fp);
1029 1119
1030 return newSVnv (ecb_binary32_to_float (fp)); 1120 return newSVnv (ecb_binary32_to_float (fp));
1031 } 1121 }
1032 1122
1033 case 27: 1123 case MISC_FLOAT64:
1034 { 1124 {
1035 uint64_t fp; 1125 uint64_t fp;
1036 WANT (8); 1126 WANT (8);
1037 memcpy (&fp, dec->cur, 8); 1127 memcpy (&fp, dec->cur, 8);
1038 dec->cur += 8; 1128 dec->cur += 8;
1041 fp = ecb_bswap64 (fp); 1131 fp = ecb_bswap64 (fp);
1042 1132
1043 return newSVnv (ecb_binary64_to_double (fp)); 1133 return newSVnv (ecb_binary64_to_double (fp));
1044 } 1134 }
1045 1135
1046 // 0..19 unassigned 1136 // 0..19 unassigned simple
1047 // 24 reserved + unassigned (reserved values are not encodable) 1137 // 24 reserved + unassigned (reserved values are not encodable)
1048 default: 1138 default:
1049 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1139 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");
1050 } 1140 }
1051 1141
1139void shrink (CBOR *self, int enable = 1) 1229void shrink (CBOR *self, int enable = 1)
1140 ALIAS: 1230 ALIAS:
1141 shrink = F_SHRINK 1231 shrink = F_SHRINK
1142 allow_unknown = F_ALLOW_UNKNOWN 1232 allow_unknown = F_ALLOW_UNKNOWN
1143 allow_sharing = F_ALLOW_SHARING 1233 allow_sharing = F_ALLOW_SHARING
1144 allow_stringref = F_ALLOW_STRINGREF 1234 allow_cycles = F_ALLOW_CYCLES
1235 pack_strings = F_PACK_STRINGS
1236 validate_utf8 = F_VALIDATE_UTF8
1145 PPCODE: 1237 PPCODE:
1146{ 1238{
1147 if (enable) 1239 if (enable)
1148 self->flags |= ix; 1240 self->flags |= ix;
1149 else 1241 else
1155void get_shrink (CBOR *self) 1247void get_shrink (CBOR *self)
1156 ALIAS: 1248 ALIAS:
1157 get_shrink = F_SHRINK 1249 get_shrink = F_SHRINK
1158 get_allow_unknown = F_ALLOW_UNKNOWN 1250 get_allow_unknown = F_ALLOW_UNKNOWN
1159 get_allow_sharing = F_ALLOW_SHARING 1251 get_allow_sharing = F_ALLOW_SHARING
1160 get_allow_stringref = F_ALLOW_STRINGREF 1252 get_allow_cycles = F_ALLOW_CYCLES
1253 get_pack_strings = F_PACK_STRINGS
1254 get_validate_utf8 = F_VALIDATE_UTF8
1161 PPCODE: 1255 PPCODE:
1162 XPUSHs (boolSV (self->flags & ix)); 1256 XPUSHs (boolSV (self->flags & ix));
1163 1257
1164void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1258void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1165 PPCODE: 1259 PPCODE:
1221 cbor_free (self); 1315 cbor_free (self);
1222 1316
1223PROTOTYPES: ENABLE 1317PROTOTYPES: ENABLE
1224 1318
1225void encode_cbor (SV *scalar) 1319void encode_cbor (SV *scalar)
1320 ALIAS:
1321 encode_cbor = 0
1322 encode_cbor_sharing = F_ALLOW_SHARING
1226 PPCODE: 1323 PPCODE:
1227{ 1324{
1228 CBOR cbor; 1325 CBOR cbor;
1229 cbor_init (&cbor); 1326 cbor_init (&cbor);
1327 cbor.flags |= ix;
1230 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1328 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1231 XPUSHs (scalar); 1329 XPUSHs (scalar);
1232} 1330}
1233 1331
1234void decode_cbor (SV *cborstr) 1332void decode_cbor (SV *cborstr)

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines