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 1.27 by root, Fri Nov 22 15:28:38 2013 UTC vs.
Revision 1.38 by root, Sun Dec 1 14:30:52 2013 UTC

19# define HvNAMELEN(hv) HvNAMELEN_get (hv) 19# define HvNAMELEN(hv) HvNAMELEN_get (hv)
20#endif 20#endif
21#ifndef HvNAMEUTF8 21#ifndef HvNAMEUTF8
22# define HvNAMEUTF8(hv) 0 22# define HvNAMEUTF8(hv) 0
23#endif 23#endif
24#ifndef SvREFCNT_dec_NN
25# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
26#endif
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};
24 64
25// known tags 65// known tags
26enum cbor_tag 66enum cbor_tag
27{ 67{
28 // inofficial extensions (pending iana registration) 68 // extensions
69 CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
29 CBOR_TAG_PERL_OBJECT = 24, // http://cbor.schmorp.de/perl-object 70 CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
30 CBOR_TAG_GENERIC_OBJECT = 25, // http://cbor.schmorp.de/generic-object 71 CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
31 CBOR_TAG_VALUE_SHAREABLE = 26, // http://cbor.schmorp.de/value-sharing 72 CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
32 CBOR_TAG_VALUE_SHAREDREF = 27, // http://cbor.schmorp.de/value-sharing 73 CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
33 CBOR_TAG_STRINGREF_NAMESPACE = 65537, // http://cbor.schmorp.de/stringref 74 CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref
34 CBOR_TAG_STRINGREF = 28, // http://cbor.schmorp.de/stringref
35 CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection 75 CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
36 76
37 // rfc7049 77 // rfc7049
38 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8 78 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
39 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any 79 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
40 CBOR_TAG_POS_BIGNUM = 2, // byte string 80 CBOR_TAG_POS_BIGNUM = 2, // byte string
41 CBOR_TAG_NEG_BIGNUM = 3, // byte string 81 CBOR_TAG_NEG_BIGNUM = 3, // byte string
42 CBOR_TAG_DECIMAL = 4, // decimal fraction, array 82 CBOR_TAG_DECIMAL = 4, // decimal fraction, array
43 CBOR_TAG_BIGFLOAT = 5, // array 83 CBOR_TAG_BIGFLOAT = 5, // array
44 84
45 CBOR_TAG_CONV_B64U = 21, // base64url, any 85 CBOR_TAG_CONV_B64U = 21, // base64url, any
46 CBOR_TAG_CONV_B64 = 22, // base64, any 86 CBOR_TAG_CONV_B64 = 22, // base64, any
47 CBOR_TAG_CONV_HEX = 23, // base16, any 87 CBOR_TAG_CONV_HEX = 23, // base16, any
48 CBOR_TAG_CBOR = 24, // embedded cbor, byte string 88 CBOR_TAG_CBOR = 24, // embedded cbor, byte string
49 89
50 CBOR_TAG_URI = 32, // URI rfc3986, utf-8 90 CBOR_TAG_URI = 32, // URI rfc3986, utf-8
51 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8 91 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
52 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8 92 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
53 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8 93 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
54 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8 94 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
55 95
56 CBOR_TAG_MAGIC = 55799 // self-describe cbor 96 CBOR_TAG_MAGIC = 55799, // self-describe cbor
57}; 97};
58 98
59#define F_SHRINK 0x00000001UL 99#define F_SHRINK 0x00000001UL
60#define F_ALLOW_UNKNOWN 0x00000002UL 100#define F_ALLOW_UNKNOWN 0x00000002UL
61#define F_ALLOW_SHARING 0x00000004UL //TODO 101#define F_ALLOW_SHARING 0x00000004UL
62#define F_ALLOW_STRINGREF 0x00000008UL //TODO 102#define F_ALLOW_CYCLES 0x00000008UL
103#define F_PACK_STRINGS 0x00000010UL
104#define F_VALIDATE_UTF8 0x00000020UL
63 105
64#define INIT_SIZE 32 // initial scalar size to be allocated 106#define INIT_SIZE 32 // initial scalar size to be allocated
65 107
66#define SB do { 108#define SB do {
67#define SE } while (0) 109#define SE } while (0)
138{ 180{
139 return idx > 23 181 return idx > 23
140 ? idx > 0xffU 182 ? idx > 0xffU
141 ? idx > 0xffffU 183 ? idx > 0xffffU
142 ? idx > 0xffffffffU 184 ? idx > 0xffffffffU
185 ? 11
143 ? 7 186 : 7
144 : 6
145 : 5 187 : 5
146 : 4 188 : 4
147 : 3; 189 : 3;
148} 190}
149 191
186static void 228static void
187encode_uint (enc_t *enc, int major, UV len) 229encode_uint (enc_t *enc, int major, UV len)
188{ 230{
189 need (enc, 9); 231 need (enc, 9);
190 232
191 if (len < 24) 233 if (ecb_expect_true (len < LENGTH_EXT1))
192 *enc->cur++ = major | len; 234 *enc->cur++ = major | len;
193 else if (len <= 0xff) 235 else if (ecb_expect_true (len <= 0xffU))
194 { 236 {
195 *enc->cur++ = major | 24; 237 *enc->cur++ = major | LENGTH_EXT1;
196 *enc->cur++ = len; 238 *enc->cur++ = len;
197 } 239 }
198 else if (len <= 0xffff) 240 else if (len <= 0xffffU)
199 { 241 {
200 *enc->cur++ = major | 25; 242 *enc->cur++ = major | LENGTH_EXT2;
201 *enc->cur++ = len >> 8; 243 *enc->cur++ = len >> 8;
202 *enc->cur++ = len; 244 *enc->cur++ = len;
203 } 245 }
204 else if (len <= 0xffffffff) 246 else if (len <= 0xffffffffU)
205 { 247 {
206 *enc->cur++ = major | 26; 248 *enc->cur++ = major | LENGTH_EXT4;
207 *enc->cur++ = len >> 24; 249 *enc->cur++ = len >> 24;
208 *enc->cur++ = len >> 16; 250 *enc->cur++ = len >> 16;
209 *enc->cur++ = len >> 8; 251 *enc->cur++ = len >> 8;
210 *enc->cur++ = len; 252 *enc->cur++ = len;
211 } 253 }
212 else 254 else
213 { 255 {
214 *enc->cur++ = major | 27; 256 *enc->cur++ = major | LENGTH_EXT8;
215 *enc->cur++ = len >> 56; 257 *enc->cur++ = len >> 56;
216 *enc->cur++ = len >> 48; 258 *enc->cur++ = len >> 48;
217 *enc->cur++ = len >> 40; 259 *enc->cur++ = len >> 40;
218 *enc->cur++ = len >> 32; 260 *enc->cur++ = len >> 32;
219 *enc->cur++ = len >> 24; 261 *enc->cur++ = len >> 24;
224} 266}
225 267
226ecb_inline void 268ecb_inline void
227encode_tag (enc_t *enc, UV tag) 269encode_tag (enc_t *enc, UV tag)
228{ 270{
229 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;
230} 281}
231 282
232static void 283static void
233encode_str (enc_t *enc, int utf8, char *str, STRLEN len) 284encode_strref (enc_t *enc, int utf8, char *str, STRLEN len)
234{ 285{
235 if (ecb_expect_false (enc->cbor.flags & F_ALLOW_STRINGREF)) 286 if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
236 { 287 {
237 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1); 288 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
238 289
239 if (SvOK (*svp)) 290 if (SvOK (*svp))
240 { 291 {
241 // already registered, use stringref 292 // already registered, use stringref
242 encode_tag (enc, CBOR_TAG_STRINGREF); 293 encode_tag (enc, CBOR_TAG_STRINGREF);
243 encode_uint (enc, 0x00, SvUV (*svp)); 294 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
244 return; 295 return;
245 } 296 }
246 else if (len >= minimum_string_length (enc->stringref_idx)) 297 else if (len >= minimum_string_length (enc->stringref_idx))
247 { 298 {
248 // register only 299 // register only
249 sv_setuv (*svp, enc->stringref_idx); 300 sv_setuv (*svp, enc->stringref_idx);
250 ++enc->stringref_idx; 301 ++enc->stringref_idx;
251 } 302 }
252 } 303 }
253 304
254 encode_uint (enc, utf8 ? 0x60 : 0x40, len); 305 encode_str (enc, utf8, str, len);
255 need (enc, len);
256 memcpy (enc->cur, str, len);
257 enc->cur += len;
258} 306}
259 307
260static void encode_sv (enc_t *enc, SV *sv); 308static void encode_sv (enc_t *enc, SV *sv);
261 309
262static void 310static void
267 if (enc->depth >= enc->cbor.max_depth) 315 if (enc->depth >= enc->cbor.max_depth)
268 croak (ERR_NESTING_EXCEEDED); 316 croak (ERR_NESTING_EXCEEDED);
269 317
270 ++enc->depth; 318 ++enc->depth;
271 319
272 encode_uint (enc, 0x80, len + 1); 320 encode_uint (enc, MAJOR_ARRAY, len + 1);
273 321
274 for (i = 0; i <= len; ++i) 322 for (i = 0; i <= len; ++i)
275 { 323 {
276 SV **svp = av_fetch (av, i, 0); 324 SV **svp = av_fetch (av, i, 0);
277 encode_sv (enc, svp ? *svp : &PL_sv_undef); 325 encode_sv (enc, svp ? *svp : &PL_sv_undef);
292 340
293 int pairs = hv_iterinit (hv); 341 int pairs = hv_iterinit (hv);
294 int mg = SvMAGICAL (hv); 342 int mg = SvMAGICAL (hv);
295 343
296 if (mg) 344 if (mg)
297 encode_ch (enc, 0xa0 | 31); 345 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
298 else 346 else
299 encode_uint (enc, 0xa0, pairs); 347 encode_uint (enc, MAJOR_MAP, pairs);
300 348
301 while ((he = hv_iternext (hv))) 349 while ((he = hv_iternext (hv)))
302 { 350 {
303 if (HeKLEN (he) == HEf_SVKEY) 351 if (HeKLEN (he) == HEf_SVKEY)
304 encode_sv (enc, HeSVKEY (he)); 352 encode_sv (enc, HeSVKEY (he));
305 else 353 else
306 encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); 354 encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
307 355
308 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));
309 } 357 }
310 358
311 if (mg) 359 if (mg)
312 encode_ch (enc, 0xe0 | 31); 360 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
313 361
314 --enc->depth; 362 --enc->depth;
315} 363}
316 364
317// encode objects, arrays and special \0=false and \1=true values. 365// encode objects, arrays and special \0=false and \1=true values.
318static void 366static void
319encode_rv (enc_t *enc, SV *sv) 367encode_rv (enc_t *enc, SV *sv)
320{ 368{
321 SvGETMAGIC (sv); 369 SvGETMAGIC (sv);
322
323 if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)
324 && ecb_expect_false (SvREFCNT (sv) > 1))
325 {
326 if (!enc->shareable)
327 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
328
329 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
330
331 if (SvOK (*svp))
332 {
333 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
334 encode_uint (enc, 0x00, SvUV (*svp));
335 return;
336 }
337 else
338 {
339 sv_setuv (*svp, enc->shareable_idx);
340 ++enc->shareable_idx;
341 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
342 }
343 }
344 370
345 svtype svt = SvTYPE (sv); 371 svtype svt = SvTYPE (sv);
346 372
347 if (ecb_expect_false (SvOBJECT (sv))) 373 if (ecb_expect_false (SvOBJECT (sv)))
348 { 374 {
355 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 381 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
356 ? cbor_tagged_stash 382 ? cbor_tagged_stash
357 : gv_stashpv ("CBOR::XS::Tagged" , 1); 383 : gv_stashpv ("CBOR::XS::Tagged" , 1);
358 384
359 HV *stash = SvSTASH (sv); 385 HV *stash = SvSTASH (sv);
360 GV *method;
361 386
362 if (stash == boolean_stash) 387 if (stash == boolean_stash)
363 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 }
364 else if (stash == error_stash) 392 else if (stash == error_stash)
365 encode_ch (enc, 0xe0 | 23); 393 {
394 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
395 return;
396 }
366 else if (stash == tagged_stash) 397 else if (stash == tagged_stash)
367 { 398 {
368 if (svt != SVt_PVAV) 399 if (svt != SVt_PVAV)
369 croak ("encountered CBOR::XS::Tagged object that isn't an array"); 400 croak ("encountered CBOR::XS::Tagged object that isn't an array");
370 401
371 encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); 402 encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
372 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))
373 } 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
374 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) 436 if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
375 { 437 {
376 dSP; 438 dSP;
377 439
378 ENTER; SAVETMPS; PUSHMARK (SP); 440 ENTER; SAVETMPS; PUSHMARK (SP);
379 // 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
411 // catch this surprisingly common error 473 // catch this surprisingly common error
412 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) 474 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
413 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));
414 476
415 encode_tag (enc, CBOR_TAG_PERL_OBJECT); 477 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
416 encode_uint (enc, 0x80, count + 1); 478 encode_uint (enc, MAJOR_ARRAY, count + 1);
417 encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); 479 encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
418 480
419 while (count) 481 while (count)
420 encode_sv (enc, SP[1 - count--]); 482 encode_sv (enc, SP[1 - count--]);
421 483
422 PUTBACK; 484 PUTBACK;
443{ 505{
444 double nv = SvNVX (sv); 506 double nv = SvNVX (sv);
445 507
446 need (enc, 9); 508 need (enc, 9);
447 509
448 if (ecb_expect_false (nv == (U32)nv)) 510 if (ecb_expect_false (nv == (NV)(U32)nv))
449 encode_uint (enc, 0x00, (U32)nv); 511 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
450 //TODO: maybe I32? 512 //TODO: maybe I32?
451 else if (ecb_expect_false (nv == (float)nv)) 513 else if (ecb_expect_false (nv == (float)nv))
452 { 514 {
453 uint32_t fp = ecb_float_to_binary32 (nv); 515 uint32_t fp = ecb_float_to_binary32 (nv);
454 516
455 *enc->cur++ = 0xe0 | 26; 517 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
456 518
457 if (!ecb_big_endian ()) 519 if (!ecb_big_endian ())
458 fp = ecb_bswap32 (fp); 520 fp = ecb_bswap32 (fp);
459 521
460 memcpy (enc->cur, &fp, 4); 522 memcpy (enc->cur, &fp, 4);
462 } 524 }
463 else 525 else
464 { 526 {
465 uint64_t fp = ecb_double_to_binary64 (nv); 527 uint64_t fp = ecb_double_to_binary64 (nv);
466 528
467 *enc->cur++ = 0xe0 | 27; 529 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
468 530
469 if (!ecb_big_endian ()) 531 if (!ecb_big_endian ())
470 fp = ecb_bswap64 (fp); 532 fp = ecb_bswap64 (fp);
471 533
472 memcpy (enc->cur, &fp, 8); 534 memcpy (enc->cur, &fp, 8);
481 543
482 if (SvPOKp (sv)) 544 if (SvPOKp (sv))
483 { 545 {
484 STRLEN len; 546 STRLEN len;
485 char *str = SvPV (sv, len); 547 char *str = SvPV (sv, len);
486 encode_str (enc, SvUTF8 (sv), str, len); 548 encode_strref (enc, SvUTF8 (sv), str, len);
487 } 549 }
488 else if (SvNOKp (sv)) 550 else if (SvNOKp (sv))
489 encode_nv (enc, sv); 551 encode_nv (enc, sv);
490 else if (SvIOKp (sv)) 552 else if (SvIOKp (sv))
491 { 553 {
492 if (SvIsUV (sv)) 554 if (SvIsUV (sv))
493 encode_uint (enc, 0x00, SvUVX (sv)); 555 encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
494 else if (SvIVX (sv) >= 0) 556 else if (SvIVX (sv) >= 0)
495 encode_uint (enc, 0x00, SvIVX (sv)); 557 encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
496 else 558 else
497 encode_uint (enc, 0x20, -(SvIVX (sv) + 1)); 559 encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
498 } 560 }
499 else if (SvROK (sv)) 561 else if (SvROK (sv))
500 encode_rv (enc, SvRV (sv)); 562 encode_rv (enc, SvRV (sv));
501 else if (!SvOK (sv)) 563 else if (!SvOK (sv))
502 encode_ch (enc, 0xe0 | 22); 564 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
503 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 565 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
504 encode_ch (enc, 0xe0 | 23); 566 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
505 else 567 else
506 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",
507 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 569 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
508} 570}
509 571
517 enc.cur = SvPVX (enc.sv); 579 enc.cur = SvPVX (enc.sv);
518 enc.end = SvEND (enc.sv); 580 enc.end = SvEND (enc.sv);
519 581
520 SvPOK_only (enc.sv); 582 SvPOK_only (enc.sv);
521 583
522 if (cbor->flags & F_ALLOW_STRINGREF) 584 if (cbor->flags & F_PACK_STRINGS)
523 { 585 {
524 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE); 586 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
525 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ()); 587 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
526 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ()); 588 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
527 } 589 }
562#define DEC_DEC_DEPTH --dec->depth 624#define DEC_DEC_DEPTH --dec->depth
563 625
564static UV 626static UV
565decode_uint (dec_t *dec) 627decode_uint (dec_t *dec)
566{ 628{
567 switch (*dec->cur & 31) 629 U8 m = *dec->cur & MINOR_MASK;
568 { 630 ++dec->cur;
569 case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
570 case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
571 case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
572 return *dec->cur++ & 31;
573 631
574 case 24: 632 if (ecb_expect_true (m < LENGTH_EXT1))
633 return m;
634 else if (ecb_expect_true (m == LENGTH_EXT1))
635 {
575 WANT (2); 636 WANT (1);
576 dec->cur += 2; 637 dec->cur += 1;
577 return dec->cur[-1]; 638 return dec->cur[-1];
578 639 }
579 case 25: 640 else if (ecb_expect_true (m == LENGTH_EXT2))
641 {
580 WANT (3); 642 WANT (2);
581 dec->cur += 3; 643 dec->cur += 2;
582 return (((UV)dec->cur[-2]) << 8) 644 return (((UV)dec->cur[-2]) << 8)
583 | ((UV)dec->cur[-1]); 645 | ((UV)dec->cur[-1]);
584 646 }
585 case 26: 647 else if (ecb_expect_true (m == LENGTH_EXT4))
648 {
586 WANT (5); 649 WANT (4);
587 dec->cur += 5; 650 dec->cur += 4;
588 return (((UV)dec->cur[-4]) << 24) 651 return (((UV)dec->cur[-4]) << 24)
589 | (((UV)dec->cur[-3]) << 16) 652 | (((UV)dec->cur[-3]) << 16)
590 | (((UV)dec->cur[-2]) << 8) 653 | (((UV)dec->cur[-2]) << 8)
591 | ((UV)dec->cur[-1]); 654 | ((UV)dec->cur[-1]);
592 655 }
593 case 27: 656 else if (ecb_expect_true (m == LENGTH_EXT8))
657 {
594 WANT (9); 658 WANT (8);
595 dec->cur += 9; 659 dec->cur += 8;
660
661 return
662#if UVSIZE < 8
663 0
664#else
596 return (((UV)dec->cur[-8]) << 56) 665 (((UV)dec->cur[-8]) << 56)
597 | (((UV)dec->cur[-7]) << 48) 666 | (((UV)dec->cur[-7]) << 48)
598 | (((UV)dec->cur[-6]) << 40) 667 | (((UV)dec->cur[-6]) << 40)
599 | (((UV)dec->cur[-5]) << 32) 668 | (((UV)dec->cur[-5]) << 32)
669#endif
600 | (((UV)dec->cur[-4]) << 24) 670 | (((UV)dec->cur[-4]) << 24)
601 | (((UV)dec->cur[-3]) << 16) 671 | (((UV)dec->cur[-3]) << 16)
602 | (((UV)dec->cur[-2]) << 8) 672 | (((UV)dec->cur[-2]) << 8)
603 | ((UV)dec->cur[-1]); 673 | ((UV)dec->cur[-1]);
604 674 }
605 default: 675 else
606 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 676 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
607 }
608 677
609fail: 678fail:
610 return 0; 679 return 0;
611} 680}
612 681
617{ 686{
618 AV *av = newAV (); 687 AV *av = newAV ();
619 688
620 DEC_INC_DEPTH; 689 DEC_INC_DEPTH;
621 690
622 if ((*dec->cur & 31) == 31) 691 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
623 { 692 {
624 ++dec->cur; 693 ++dec->cur;
625 694
626 for (;;) 695 for (;;)
627 { 696 {
628 WANT (1); 697 WANT (1);
629 698
630 if (*dec->cur == (0xe0 | 31)) 699 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
631 { 700 {
632 ++dec->cur; 701 ++dec->cur;
633 break; 702 break;
634 } 703 }
635 704
638 } 707 }
639 else 708 else
640 { 709 {
641 int i, len = decode_uint (dec); 710 int i, len = decode_uint (dec);
642 711
712 WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
643 av_fill (av, len - 1); 713 av_fill (av, len - 1);
644 714
645 for (i = 0; i < len; ++i) 715 for (i = 0; i < len; ++i)
646 AvARRAY (av)[i] = decode_sv (dec); 716 AvARRAY (av)[i] = decode_sv (dec);
647 } 717 }
660{ 730{
661 // for speed reasons, we specialcase single-string 731 // for speed reasons, we specialcase single-string
662 // byte or utf-8 strings as keys, but only when !stringref 732 // byte or utf-8 strings as keys, but only when !stringref
663 733
664 if (ecb_expect_true (!dec->stringref)) 734 if (ecb_expect_true (!dec->stringref))
665 if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27) 735 if (ecb_expect_true ((*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
666 { 736 {
667 I32 len = decode_uint (dec); 737 I32 len = decode_uint (dec);
668 char *key = (char *)dec->cur; 738 char *key = (char *)dec->cur;
669 739
670 dec->cur += len; 740 dec->cur += len;
671 741
672 if (ecb_expect_false (dec->stringref))
673 av_push (dec->stringref, newSVpvn (key, len));
674
675 hv_store (hv, key, len, decode_sv (dec), 0); 742 hv_store (hv, key, len, decode_sv (dec), 0);
676 743
677 return; 744 return;
678 } 745 }
679 else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27) 746 else if (ecb_expect_true ((*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
680 { 747 {
681 I32 len = decode_uint (dec); 748 I32 len = decode_uint (dec);
682 char *key = (char *)dec->cur; 749 char *key = (char *)dec->cur;
683 750
684 dec->cur += len; 751 dec->cur += len;
685 752
686 if (ecb_expect_false (dec->stringref)) 753 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
687 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)");
688 756
689 hv_store (hv, key, -len, decode_sv (dec), 0); 757 hv_store (hv, key, -len, decode_sv (dec), 0);
690 758
691 return; 759 return;
692 } 760 }
694 SV *k = decode_sv (dec); 762 SV *k = decode_sv (dec);
695 SV *v = decode_sv (dec); 763 SV *v = decode_sv (dec);
696 764
697 hv_store_ent (hv, k, v, 0); 765 hv_store_ent (hv, k, v, 0);
698 SvREFCNT_dec (k); 766 SvREFCNT_dec (k);
767
768fail:
769 ;
699} 770}
700 771
701static SV * 772static SV *
702decode_hv (dec_t *dec) 773decode_hv (dec_t *dec)
703{ 774{
704 HV *hv = newHV (); 775 HV *hv = newHV ();
705 776
706 DEC_INC_DEPTH; 777 DEC_INC_DEPTH;
707 778
708 if ((*dec->cur & 31) == 31) 779 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
709 { 780 {
710 ++dec->cur; 781 ++dec->cur;
711 782
712 for (;;) 783 for (;;)
713 { 784 {
714 WANT (1); 785 WANT (1);
715 786
716 if (*dec->cur == (0xe0 | 31)) 787 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
717 { 788 {
718 ++dec->cur; 789 ++dec->cur;
719 break; 790 break;
720 } 791 }
721 792
742static SV * 813static SV *
743decode_str (dec_t *dec, int utf8) 814decode_str (dec_t *dec, int utf8)
744{ 815{
745 SV *sv = 0; 816 SV *sv = 0;
746 817
747 if ((*dec->cur & 31) == 31) 818 if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
748 { 819 {
820 // indefinite length strings
749 ++dec->cur; 821 ++dec->cur;
750 822
823 U8 major = *dec->cur & MAJOR_MISC;
824
751 sv = newSVpvn ("", 0); 825 sv = newSVpvn ("", 0);
752 826
753 // not very fast, and certainly not robust against illegal input
754 for (;;) 827 for (;;)
755 { 828 {
756 WANT (1); 829 WANT (1);
757 830
758 if (*dec->cur == (0xe0 | 31)) 831 if ((*dec->cur - major) > LENGTH_EXT8)
832 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
759 { 833 {
760 ++dec->cur; 834 ++dec->cur;
761 break; 835 break;
762 } 836 }
837 else
838 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
763 839
764 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;
765 } 845 }
766 } 846 }
767 else 847 else
768 { 848 {
769 STRLEN len = decode_uint (dec); 849 STRLEN len = decode_uint (dec);
776 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1)) 856 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
777 av_push (dec->stringref, SvREFCNT_inc_NN (sv)); 857 av_push (dec->stringref, SvREFCNT_inc_NN (sv));
778 } 858 }
779 859
780 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
781 SvUTF8_on (sv); 866 SvUTF8_on (sv);
867 }
782 868
783 return sv; 869 return sv;
784 870
785fail: 871fail:
786 SvREFCNT_dec (sv); 872 SvREFCNT_dec (sv);
818 } 904 }
819 break; 905 break;
820 906
821 case CBOR_TAG_STRINGREF: 907 case CBOR_TAG_STRINGREF:
822 { 908 {
823 if ((*dec->cur >> 5) != 0) 909 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
824 ERR ("corrupted CBOR data (stringref index not an unsigned integer)"); 910 ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
825 911
826 UV idx = decode_uint (dec); 912 UV idx = decode_uint (dec);
827 913
828 if (!dec->stringref || (int)idx > AvFILLp (dec->stringref)) 914 if (!dec->stringref || (int)idx > AvFILLp (dec->stringref))
835 case CBOR_TAG_VALUE_SHAREABLE: 921 case CBOR_TAG_VALUE_SHAREABLE:
836 { 922 {
837 if (ecb_expect_false (!dec->shareable)) 923 if (ecb_expect_false (!dec->shareable))
838 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ()); 924 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
839 925
926 if (dec->cbor.flags & F_ALLOW_CYCLES)
927 {
840 sv = newSV (0); 928 sv = newSV (0);
841 av_push (dec->shareable, SvREFCNT_inc_NN (sv)); 929 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
842 930
843 SV *osv = decode_sv (dec); 931 SV *osv = decode_sv (dec);
844 sv_setsv (sv, osv); 932 sv_setsv (sv, osv);
845 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 }
846 } 942 }
847 break; 943 break;
848 944
849 case CBOR_TAG_VALUE_SHAREDREF: 945 case CBOR_TAG_VALUE_SHAREDREF:
850 { 946 {
851 if ((*dec->cur >> 5) != 0) 947 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
852 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)"); 948 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
853 949
854 UV idx = decode_uint (dec); 950 UV idx = decode_uint (dec);
855 951
856 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable)) 952 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable))
857 ERR ("corrupted CBOR data (sharedref index out of bounds)"); 953 ERR ("corrupted CBOR data (sharedref index out of bounds)");
858 954
859 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");
860 } 959 }
861 break; 960 break;
862 961
863 case CBOR_TAG_PERL_OBJECT: 962 case CBOR_TAG_PERL_OBJECT:
864 { 963 {
965static SV * 1064static SV *
966decode_sv (dec_t *dec) 1065decode_sv (dec_t *dec)
967{ 1066{
968 WANT (1); 1067 WANT (1);
969 1068
970 switch (*dec->cur >> 5) 1069 switch (*dec->cur >> MAJOR_SHIFT)
971 { 1070 {
972 case 0: // unsigned int 1071 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
973 return newSVuv (decode_uint (dec)); 1072 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
974 case 1: // negative int 1073 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
975 return newSViv (-1 - (IV)decode_uint (dec)); 1074 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
976 case 2: // octet string 1075 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
977 return decode_str (dec, 0); 1076 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
978 case 3: // utf-8 string 1077 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
979 return decode_str (dec, 1); 1078
980 case 4: // array 1079 case MAJOR_MISC >> MAJOR_SHIFT:
981 return decode_av (dec);
982 case 5: // map
983 return decode_hv (dec);
984 case 6: // tag
985 return decode_tagged (dec);
986 case 7: // misc
987 switch (*dec->cur++ & 31) 1080 switch (*dec->cur++ & MINOR_MASK)
988 { 1081 {
989 case 20: 1082 case SIMPLE_FALSE:
990#if CBOR_SLOW 1083#if CBOR_SLOW
991 types_false = get_bool ("Types::Serialiser::false"); 1084 types_false = get_bool ("Types::Serialiser::false");
992#endif 1085#endif
993 return newSVsv (types_false); 1086 return newSVsv (types_false);
994 case 21: 1087 case SIMPLE_TRUE:
995#if CBOR_SLOW 1088#if CBOR_SLOW
996 types_true = get_bool ("Types::Serialiser::true"); 1089 types_true = get_bool ("Types::Serialiser::true");
997#endif 1090#endif
998 return newSVsv (types_true); 1091 return newSVsv (types_true);
999 case 22: 1092 case SIMPLE_NULL:
1000 return newSVsv (&PL_sv_undef); 1093 return newSVsv (&PL_sv_undef);
1001 case 23: 1094 case SIMPLE_UNDEF:
1002#if CBOR_SLOW 1095#if CBOR_SLOW
1003 types_error = get_bool ("Types::Serialiser::error"); 1096 types_error = get_bool ("Types::Serialiser::error");
1004#endif 1097#endif
1005 return newSVsv (types_error); 1098 return newSVsv (types_error);
1006 1099
1007 case 25: 1100 case MISC_FLOAT16:
1008 { 1101 {
1009 WANT (2); 1102 WANT (2);
1010 1103
1011 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1104 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
1012 dec->cur += 2; 1105 dec->cur += 2;
1013 1106
1014 return newSVnv (ecb_binary16_to_float (fp)); 1107 return newSVnv (ecb_binary16_to_float (fp));
1015 } 1108 }
1016 1109
1017 case 26: 1110 case MISC_FLOAT32:
1018 { 1111 {
1019 uint32_t fp; 1112 uint32_t fp;
1020 WANT (4); 1113 WANT (4);
1021 memcpy (&fp, dec->cur, 4); 1114 memcpy (&fp, dec->cur, 4);
1022 dec->cur += 4; 1115 dec->cur += 4;
1025 fp = ecb_bswap32 (fp); 1118 fp = ecb_bswap32 (fp);
1026 1119
1027 return newSVnv (ecb_binary32_to_float (fp)); 1120 return newSVnv (ecb_binary32_to_float (fp));
1028 } 1121 }
1029 1122
1030 case 27: 1123 case MISC_FLOAT64:
1031 { 1124 {
1032 uint64_t fp; 1125 uint64_t fp;
1033 WANT (8); 1126 WANT (8);
1034 memcpy (&fp, dec->cur, 8); 1127 memcpy (&fp, dec->cur, 8);
1035 dec->cur += 8; 1128 dec->cur += 8;
1038 fp = ecb_bswap64 (fp); 1131 fp = ecb_bswap64 (fp);
1039 1132
1040 return newSVnv (ecb_binary64_to_double (fp)); 1133 return newSVnv (ecb_binary64_to_double (fp));
1041 } 1134 }
1042 1135
1043 // 0..19 unassigned 1136 // 0..19 unassigned simple
1044 // 24 reserved + unassigned (reserved values are not encodable) 1137 // 24 reserved + unassigned (reserved values are not encodable)
1045 default: 1138 default:
1046 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1139 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");
1047 } 1140 }
1048 1141
1136void shrink (CBOR *self, int enable = 1) 1229void shrink (CBOR *self, int enable = 1)
1137 ALIAS: 1230 ALIAS:
1138 shrink = F_SHRINK 1231 shrink = F_SHRINK
1139 allow_unknown = F_ALLOW_UNKNOWN 1232 allow_unknown = F_ALLOW_UNKNOWN
1140 allow_sharing = F_ALLOW_SHARING 1233 allow_sharing = F_ALLOW_SHARING
1141 allow_stringref = F_ALLOW_STRINGREF 1234 allow_cycles = F_ALLOW_CYCLES
1235 pack_strings = F_PACK_STRINGS
1236 validate_utf8 = F_VALIDATE_UTF8
1142 PPCODE: 1237 PPCODE:
1143{ 1238{
1144 if (enable) 1239 if (enable)
1145 self->flags |= ix; 1240 self->flags |= ix;
1146 else 1241 else
1152void get_shrink (CBOR *self) 1247void get_shrink (CBOR *self)
1153 ALIAS: 1248 ALIAS:
1154 get_shrink = F_SHRINK 1249 get_shrink = F_SHRINK
1155 get_allow_unknown = F_ALLOW_UNKNOWN 1250 get_allow_unknown = F_ALLOW_UNKNOWN
1156 get_allow_sharing = F_ALLOW_SHARING 1251 get_allow_sharing = F_ALLOW_SHARING
1157 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
1158 PPCODE: 1255 PPCODE:
1159 XPUSHs (boolSV (self->flags & ix)); 1256 XPUSHs (boolSV (self->flags & ix));
1160 1257
1161void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1258void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1162 PPCODE: 1259 PPCODE:
1218 cbor_free (self); 1315 cbor_free (self);
1219 1316
1220PROTOTYPES: ENABLE 1317PROTOTYPES: ENABLE
1221 1318
1222void encode_cbor (SV *scalar) 1319void encode_cbor (SV *scalar)
1320 ALIAS:
1321 encode_cbor = 0
1322 encode_cbor_sharing = F_ALLOW_SHARING
1223 PPCODE: 1323 PPCODE:
1224{ 1324{
1225 CBOR cbor; 1325 CBOR cbor;
1226 cbor_init (&cbor); 1326 cbor_init (&cbor);
1327 cbor.flags |= ix;
1227 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1328 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1228 XPUSHs (scalar); 1329 XPUSHs (scalar);
1229} 1330}
1230 1331
1231void decode_cbor (SV *cborstr) 1332void decode_cbor (SV *cborstr)

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines