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 1.37 by root, Sat Nov 30 18:13:53 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
63 104
64#define INIT_SIZE 32 // initial scalar size to be allocated 105#define INIT_SIZE 32 // initial scalar size to be allocated
65 106
66#define SB do { 107#define SB do {
67#define SE } while (0) 108#define SE } while (0)
138{ 179{
139 return idx > 23 180 return idx > 23
140 ? idx > 0xffU 181 ? idx > 0xffU
141 ? idx > 0xffffU 182 ? idx > 0xffffU
142 ? idx > 0xffffffffU 183 ? idx > 0xffffffffU
184 ? 11
143 ? 7 185 : 7
144 : 6
145 : 5 186 : 5
146 : 4 187 : 4
147 : 3; 188 : 3;
148} 189}
149 190
186static void 227static void
187encode_uint (enc_t *enc, int major, UV len) 228encode_uint (enc_t *enc, int major, UV len)
188{ 229{
189 need (enc, 9); 230 need (enc, 9);
190 231
191 if (len < 24) 232 if (ecb_expect_true (len < LENGTH_EXT1))
192 *enc->cur++ = major | len; 233 *enc->cur++ = major | len;
193 else if (len <= 0xff) 234 else if (ecb_expect_true (len <= 0xffU))
194 { 235 {
195 *enc->cur++ = major | 24; 236 *enc->cur++ = major | LENGTH_EXT1;
196 *enc->cur++ = len; 237 *enc->cur++ = len;
197 } 238 }
198 else if (len <= 0xffff) 239 else if (len <= 0xffffU)
199 { 240 {
200 *enc->cur++ = major | 25; 241 *enc->cur++ = major | LENGTH_EXT2;
201 *enc->cur++ = len >> 8; 242 *enc->cur++ = len >> 8;
202 *enc->cur++ = len; 243 *enc->cur++ = len;
203 } 244 }
204 else if (len <= 0xffffffff) 245 else if (len <= 0xffffffffU)
205 { 246 {
206 *enc->cur++ = major | 26; 247 *enc->cur++ = major | LENGTH_EXT4;
207 *enc->cur++ = len >> 24; 248 *enc->cur++ = len >> 24;
208 *enc->cur++ = len >> 16; 249 *enc->cur++ = len >> 16;
209 *enc->cur++ = len >> 8; 250 *enc->cur++ = len >> 8;
210 *enc->cur++ = len; 251 *enc->cur++ = len;
211 } 252 }
212 else 253 else
213 { 254 {
214 *enc->cur++ = major | 27; 255 *enc->cur++ = major | LENGTH_EXT8;
215 *enc->cur++ = len >> 56; 256 *enc->cur++ = len >> 56;
216 *enc->cur++ = len >> 48; 257 *enc->cur++ = len >> 48;
217 *enc->cur++ = len >> 40; 258 *enc->cur++ = len >> 40;
218 *enc->cur++ = len >> 32; 259 *enc->cur++ = len >> 32;
219 *enc->cur++ = len >> 24; 260 *enc->cur++ = len >> 24;
224} 265}
225 266
226ecb_inline void 267ecb_inline void
227encode_tag (enc_t *enc, UV tag) 268encode_tag (enc_t *enc, UV tag)
228{ 269{
229 encode_uint (enc, 0xc0, tag); 270 encode_uint (enc, MAJOR_TAG, tag);
271}
272
273ecb_inline void
274encode_str (enc_t *enc, int utf8, char *str, STRLEN len)
275{
276 encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
277 need (enc, len);
278 memcpy (enc->cur, str, len);
279 enc->cur += len;
230} 280}
231 281
232static void 282static void
233encode_str (enc_t *enc, int utf8, char *str, STRLEN len) 283encode_strref (enc_t *enc, int utf8, char *str, STRLEN len)
234{ 284{
235 if (ecb_expect_false (enc->cbor.flags & F_ALLOW_STRINGREF)) 285 if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
236 { 286 {
237 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1); 287 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
238 288
239 if (SvOK (*svp)) 289 if (SvOK (*svp))
240 { 290 {
241 // already registered, use stringref 291 // already registered, use stringref
242 encode_tag (enc, CBOR_TAG_STRINGREF); 292 encode_tag (enc, CBOR_TAG_STRINGREF);
243 encode_uint (enc, 0x00, SvUV (*svp)); 293 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
244 return; 294 return;
245 } 295 }
246 else if (len >= minimum_string_length (enc->stringref_idx)) 296 else if (len >= minimum_string_length (enc->stringref_idx))
247 { 297 {
248 // register only 298 // register only
249 sv_setuv (*svp, enc->stringref_idx); 299 sv_setuv (*svp, enc->stringref_idx);
250 ++enc->stringref_idx; 300 ++enc->stringref_idx;
251 } 301 }
252 } 302 }
253 303
254 encode_uint (enc, utf8 ? 0x60 : 0x40, len); 304 encode_str (enc, utf8, str, len);
255 need (enc, len);
256 memcpy (enc->cur, str, len);
257 enc->cur += len;
258} 305}
259 306
260static void encode_sv (enc_t *enc, SV *sv); 307static void encode_sv (enc_t *enc, SV *sv);
261 308
262static void 309static void
267 if (enc->depth >= enc->cbor.max_depth) 314 if (enc->depth >= enc->cbor.max_depth)
268 croak (ERR_NESTING_EXCEEDED); 315 croak (ERR_NESTING_EXCEEDED);
269 316
270 ++enc->depth; 317 ++enc->depth;
271 318
272 encode_uint (enc, 0x80, len + 1); 319 encode_uint (enc, MAJOR_ARRAY, len + 1);
273 320
274 for (i = 0; i <= len; ++i) 321 for (i = 0; i <= len; ++i)
275 { 322 {
276 SV **svp = av_fetch (av, i, 0); 323 SV **svp = av_fetch (av, i, 0);
277 encode_sv (enc, svp ? *svp : &PL_sv_undef); 324 encode_sv (enc, svp ? *svp : &PL_sv_undef);
292 339
293 int pairs = hv_iterinit (hv); 340 int pairs = hv_iterinit (hv);
294 int mg = SvMAGICAL (hv); 341 int mg = SvMAGICAL (hv);
295 342
296 if (mg) 343 if (mg)
297 encode_ch (enc, 0xa0 | 31); 344 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
298 else 345 else
299 encode_uint (enc, 0xa0, pairs); 346 encode_uint (enc, MAJOR_MAP, pairs);
300 347
301 while ((he = hv_iternext (hv))) 348 while ((he = hv_iternext (hv)))
302 { 349 {
303 if (HeKLEN (he) == HEf_SVKEY) 350 if (HeKLEN (he) == HEf_SVKEY)
304 encode_sv (enc, HeSVKEY (he)); 351 encode_sv (enc, HeSVKEY (he));
305 else 352 else
306 encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); 353 encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
307 354
308 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); 355 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
309 } 356 }
310 357
311 if (mg) 358 if (mg)
312 encode_ch (enc, 0xe0 | 31); 359 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
313 360
314 --enc->depth; 361 --enc->depth;
315} 362}
316 363
317// encode objects, arrays and special \0=false and \1=true values. 364// encode objects, arrays and special \0=false and \1=true values.
318static void 365static void
319encode_rv (enc_t *enc, SV *sv) 366encode_rv (enc_t *enc, SV *sv)
320{ 367{
321 SvGETMAGIC (sv); 368 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 }
344 369
345 svtype svt = SvTYPE (sv); 370 svtype svt = SvTYPE (sv);
346 371
347 if (ecb_expect_false (SvOBJECT (sv))) 372 if (ecb_expect_false (SvOBJECT (sv)))
348 { 373 {
355 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 380 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
356 ? cbor_tagged_stash 381 ? cbor_tagged_stash
357 : gv_stashpv ("CBOR::XS::Tagged" , 1); 382 : gv_stashpv ("CBOR::XS::Tagged" , 1);
358 383
359 HV *stash = SvSTASH (sv); 384 HV *stash = SvSTASH (sv);
360 GV *method;
361 385
362 if (stash == boolean_stash) 386 if (stash == boolean_stash)
363 encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20); 387 {
388 encode_ch (enc, SvIV (sv) ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE);
389 return;
390 }
364 else if (stash == error_stash) 391 else if (stash == error_stash)
365 encode_ch (enc, 0xe0 | 23); 392 {
393 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
394 return;
395 }
366 else if (stash == tagged_stash) 396 else if (stash == tagged_stash)
367 { 397 {
368 if (svt != SVt_PVAV) 398 if (svt != SVt_PVAV)
369 croak ("encountered CBOR::XS::Tagged object that isn't an array"); 399 croak ("encountered CBOR::XS::Tagged object that isn't an array");
370 400
371 encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); 401 encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
372 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); 402 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1));
403
404 return;
405 }
406 }
407
408 if (ecb_expect_false (SvREFCNT (sv) > 1)
409 && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
410 {
411 if (!enc->shareable)
412 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
413
414 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
415
416 if (SvOK (*svp))
373 } 417 {
418 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
419 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
420 return;
421 }
422 else
423 {
424 sv_setuv (*svp, enc->shareable_idx);
425 ++enc->shareable_idx;
426 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
427 }
428 }
429
430 if (ecb_expect_false (SvOBJECT (sv)))
431 {
432 HV *stash = SvSTASH (sv);
433 GV *method;
434
374 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) 435 if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
375 { 436 {
376 dSP; 437 dSP;
377 438
378 ENTER; SAVETMPS; PUSHMARK (SP); 439 ENTER; SAVETMPS; PUSHMARK (SP);
379 // we re-bless the reference to get overload and other niceties right 440 // we re-bless the reference to get overload and other niceties right
411 // catch this surprisingly common error 472 // catch this surprisingly common error
412 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) 473 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)); 474 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
414 475
415 encode_tag (enc, CBOR_TAG_PERL_OBJECT); 476 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
416 encode_uint (enc, 0x80, count + 1); 477 encode_uint (enc, MAJOR_ARRAY, count + 1);
417 encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); 478 encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
418 479
419 while (count) 480 while (count)
420 encode_sv (enc, SP[1 - count--]); 481 encode_sv (enc, SP[1 - count--]);
421 482
422 PUTBACK; 483 PUTBACK;
443{ 504{
444 double nv = SvNVX (sv); 505 double nv = SvNVX (sv);
445 506
446 need (enc, 9); 507 need (enc, 9);
447 508
448 if (ecb_expect_false (nv == (U32)nv)) 509 if (ecb_expect_false (nv == (NV)(U32)nv))
449 encode_uint (enc, 0x00, (U32)nv); 510 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
450 //TODO: maybe I32? 511 //TODO: maybe I32?
451 else if (ecb_expect_false (nv == (float)nv)) 512 else if (ecb_expect_false (nv == (float)nv))
452 { 513 {
453 uint32_t fp = ecb_float_to_binary32 (nv); 514 uint32_t fp = ecb_float_to_binary32 (nv);
454 515
455 *enc->cur++ = 0xe0 | 26; 516 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
456 517
457 if (!ecb_big_endian ()) 518 if (!ecb_big_endian ())
458 fp = ecb_bswap32 (fp); 519 fp = ecb_bswap32 (fp);
459 520
460 memcpy (enc->cur, &fp, 4); 521 memcpy (enc->cur, &fp, 4);
462 } 523 }
463 else 524 else
464 { 525 {
465 uint64_t fp = ecb_double_to_binary64 (nv); 526 uint64_t fp = ecb_double_to_binary64 (nv);
466 527
467 *enc->cur++ = 0xe0 | 27; 528 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
468 529
469 if (!ecb_big_endian ()) 530 if (!ecb_big_endian ())
470 fp = ecb_bswap64 (fp); 531 fp = ecb_bswap64 (fp);
471 532
472 memcpy (enc->cur, &fp, 8); 533 memcpy (enc->cur, &fp, 8);
481 542
482 if (SvPOKp (sv)) 543 if (SvPOKp (sv))
483 { 544 {
484 STRLEN len; 545 STRLEN len;
485 char *str = SvPV (sv, len); 546 char *str = SvPV (sv, len);
486 encode_str (enc, SvUTF8 (sv), str, len); 547 encode_strref (enc, SvUTF8 (sv), str, len);
487 } 548 }
488 else if (SvNOKp (sv)) 549 else if (SvNOKp (sv))
489 encode_nv (enc, sv); 550 encode_nv (enc, sv);
490 else if (SvIOKp (sv)) 551 else if (SvIOKp (sv))
491 { 552 {
492 if (SvIsUV (sv)) 553 if (SvIsUV (sv))
493 encode_uint (enc, 0x00, SvUVX (sv)); 554 encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
494 else if (SvIVX (sv) >= 0) 555 else if (SvIVX (sv) >= 0)
495 encode_uint (enc, 0x00, SvIVX (sv)); 556 encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
496 else 557 else
497 encode_uint (enc, 0x20, -(SvIVX (sv) + 1)); 558 encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
498 } 559 }
499 else if (SvROK (sv)) 560 else if (SvROK (sv))
500 encode_rv (enc, SvRV (sv)); 561 encode_rv (enc, SvRV (sv));
501 else if (!SvOK (sv)) 562 else if (!SvOK (sv))
502 encode_ch (enc, 0xe0 | 22); 563 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
503 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 564 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
504 encode_ch (enc, 0xe0 | 23); 565 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
505 else 566 else
506 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data", 567 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
507 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 568 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
508} 569}
509 570
517 enc.cur = SvPVX (enc.sv); 578 enc.cur = SvPVX (enc.sv);
518 enc.end = SvEND (enc.sv); 579 enc.end = SvEND (enc.sv);
519 580
520 SvPOK_only (enc.sv); 581 SvPOK_only (enc.sv);
521 582
522 if (cbor->flags & F_ALLOW_STRINGREF) 583 if (cbor->flags & F_PACK_STRINGS)
523 { 584 {
524 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE); 585 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
525 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ()); 586 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
526 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ()); 587 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
527 } 588 }
562#define DEC_DEC_DEPTH --dec->depth 623#define DEC_DEC_DEPTH --dec->depth
563 624
564static UV 625static UV
565decode_uint (dec_t *dec) 626decode_uint (dec_t *dec)
566{ 627{
567 switch (*dec->cur & 31) 628 U8 m = *dec->cur & MINOR_MASK;
568 { 629 ++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;
573 630
574 case 24: 631 if (ecb_expect_true (m < LENGTH_EXT1))
632 return m;
633 else if (ecb_expect_true (m == LENGTH_EXT1))
634 {
575 WANT (2); 635 WANT (1);
576 dec->cur += 2; 636 dec->cur += 1;
577 return dec->cur[-1]; 637 return dec->cur[-1];
578 638 }
579 case 25: 639 else if (ecb_expect_true (m == LENGTH_EXT2))
640 {
580 WANT (3); 641 WANT (2);
581 dec->cur += 3; 642 dec->cur += 2;
582 return (((UV)dec->cur[-2]) << 8) 643 return (((UV)dec->cur[-2]) << 8)
583 | ((UV)dec->cur[-1]); 644 | ((UV)dec->cur[-1]);
584 645 }
585 case 26: 646 else if (ecb_expect_true (m == LENGTH_EXT4))
647 {
586 WANT (5); 648 WANT (4);
587 dec->cur += 5; 649 dec->cur += 4;
588 return (((UV)dec->cur[-4]) << 24) 650 return (((UV)dec->cur[-4]) << 24)
589 | (((UV)dec->cur[-3]) << 16) 651 | (((UV)dec->cur[-3]) << 16)
590 | (((UV)dec->cur[-2]) << 8) 652 | (((UV)dec->cur[-2]) << 8)
591 | ((UV)dec->cur[-1]); 653 | ((UV)dec->cur[-1]);
592 654 }
593 case 27: 655 else if (ecb_expect_true (m == LENGTH_EXT8))
656 {
594 WANT (9); 657 WANT (8);
595 dec->cur += 9; 658 dec->cur += 8;
659
660 return
661#if UVSIZE < 8
662 0
663#else
596 return (((UV)dec->cur[-8]) << 56) 664 (((UV)dec->cur[-8]) << 56)
597 | (((UV)dec->cur[-7]) << 48) 665 | (((UV)dec->cur[-7]) << 48)
598 | (((UV)dec->cur[-6]) << 40) 666 | (((UV)dec->cur[-6]) << 40)
599 | (((UV)dec->cur[-5]) << 32) 667 | (((UV)dec->cur[-5]) << 32)
668#endif
600 | (((UV)dec->cur[-4]) << 24) 669 | (((UV)dec->cur[-4]) << 24)
601 | (((UV)dec->cur[-3]) << 16) 670 | (((UV)dec->cur[-3]) << 16)
602 | (((UV)dec->cur[-2]) << 8) 671 | (((UV)dec->cur[-2]) << 8)
603 | ((UV)dec->cur[-1]); 672 | ((UV)dec->cur[-1]);
604 673 }
605 default: 674 else
606 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 675 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
607 }
608 676
609fail: 677fail:
610 return 0; 678 return 0;
611} 679}
612 680
617{ 685{
618 AV *av = newAV (); 686 AV *av = newAV ();
619 687
620 DEC_INC_DEPTH; 688 DEC_INC_DEPTH;
621 689
622 if ((*dec->cur & 31) == 31) 690 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
623 { 691 {
624 ++dec->cur; 692 ++dec->cur;
625 693
626 for (;;) 694 for (;;)
627 { 695 {
628 WANT (1); 696 WANT (1);
629 697
630 if (*dec->cur == (0xe0 | 31)) 698 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
631 { 699 {
632 ++dec->cur; 700 ++dec->cur;
633 break; 701 break;
634 } 702 }
635 703
638 } 706 }
639 else 707 else
640 { 708 {
641 int i, len = decode_uint (dec); 709 int i, len = decode_uint (dec);
642 710
711 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); 712 av_fill (av, len - 1);
644 713
645 for (i = 0; i < len; ++i) 714 for (i = 0; i < len; ++i)
646 AvARRAY (av)[i] = decode_sv (dec); 715 AvARRAY (av)[i] = decode_sv (dec);
647 } 716 }
660{ 729{
661 // for speed reasons, we specialcase single-string 730 // for speed reasons, we specialcase single-string
662 // byte or utf-8 strings as keys, but only when !stringref 731 // byte or utf-8 strings as keys, but only when !stringref
663 732
664 if (ecb_expect_true (!dec->stringref)) 733 if (ecb_expect_true (!dec->stringref))
665 if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27) 734 if (ecb_expect_true ((*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
666 { 735 {
667 I32 len = decode_uint (dec); 736 I32 len = decode_uint (dec);
668 char *key = (char *)dec->cur; 737 char *key = (char *)dec->cur;
669 738
670 dec->cur += len; 739 dec->cur += len;
674 743
675 hv_store (hv, key, len, decode_sv (dec), 0); 744 hv_store (hv, key, len, decode_sv (dec), 0);
676 745
677 return; 746 return;
678 } 747 }
679 else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27) 748 else if (ecb_expect_true ((*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
680 { 749 {
681 I32 len = decode_uint (dec); 750 I32 len = decode_uint (dec);
682 char *key = (char *)dec->cur; 751 char *key = (char *)dec->cur;
683 752
684 dec->cur += len; 753 dec->cur += len;
703{ 772{
704 HV *hv = newHV (); 773 HV *hv = newHV ();
705 774
706 DEC_INC_DEPTH; 775 DEC_INC_DEPTH;
707 776
708 if ((*dec->cur & 31) == 31) 777 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
709 { 778 {
710 ++dec->cur; 779 ++dec->cur;
711 780
712 for (;;) 781 for (;;)
713 { 782 {
714 WANT (1); 783 WANT (1);
715 784
716 if (*dec->cur == (0xe0 | 31)) 785 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
717 { 786 {
718 ++dec->cur; 787 ++dec->cur;
719 break; 788 break;
720 } 789 }
721 790
742static SV * 811static SV *
743decode_str (dec_t *dec, int utf8) 812decode_str (dec_t *dec, int utf8)
744{ 813{
745 SV *sv = 0; 814 SV *sv = 0;
746 815
747 if ((*dec->cur & 31) == 31) 816 if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
748 { 817 {
818 // indefinite length strings
749 ++dec->cur; 819 ++dec->cur;
750 820
821 U8 major = *dec->cur & MAJOR_MISC;
822
751 sv = newSVpvn ("", 0); 823 sv = newSVpvn ("", 0);
752 824
753 // not very fast, and certainly not robust against illegal input
754 for (;;) 825 for (;;)
755 { 826 {
756 WANT (1); 827 WANT (1);
757 828
758 if (*dec->cur == (0xe0 | 31)) 829 if ((*dec->cur - major) > LENGTH_EXT8)
830 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
759 { 831 {
760 ++dec->cur; 832 ++dec->cur;
761 break; 833 break;
762 } 834 }
835 else
836 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
763 837
764 sv_catsv (sv, decode_sv (dec)); 838 STRLEN len = decode_uint (dec);
839
840 WANT (len);
841 sv_catpvn (sv, dec->cur, len);
842 dec->cur += len;
765 } 843 }
766 } 844 }
767 else 845 else
768 { 846 {
769 STRLEN len = decode_uint (dec); 847 STRLEN len = decode_uint (dec);
818 } 896 }
819 break; 897 break;
820 898
821 case CBOR_TAG_STRINGREF: 899 case CBOR_TAG_STRINGREF:
822 { 900 {
823 if ((*dec->cur >> 5) != 0) 901 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
824 ERR ("corrupted CBOR data (stringref index not an unsigned integer)"); 902 ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
825 903
826 UV idx = decode_uint (dec); 904 UV idx = decode_uint (dec);
827 905
828 if (!dec->stringref || (int)idx > AvFILLp (dec->stringref)) 906 if (!dec->stringref || (int)idx > AvFILLp (dec->stringref))
835 case CBOR_TAG_VALUE_SHAREABLE: 913 case CBOR_TAG_VALUE_SHAREABLE:
836 { 914 {
837 if (ecb_expect_false (!dec->shareable)) 915 if (ecb_expect_false (!dec->shareable))
838 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ()); 916 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
839 917
918 if (dec->cbor.flags & F_ALLOW_CYCLES)
919 {
840 sv = newSV (0); 920 sv = newSV (0);
841 av_push (dec->shareable, SvREFCNT_inc_NN (sv)); 921 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
842 922
843 SV *osv = decode_sv (dec); 923 SV *osv = decode_sv (dec);
844 sv_setsv (sv, osv); 924 sv_setsv (sv, osv);
845 SvREFCNT_dec_NN (osv); 925 SvREFCNT_dec_NN (osv);
926 }
927 else
928 {
929 av_push (dec->shareable, &PL_sv_undef);
930 int idx = AvFILLp (dec->shareable);
931 sv = decode_sv (dec);
932 av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
933 }
846 } 934 }
847 break; 935 break;
848 936
849 case CBOR_TAG_VALUE_SHAREDREF: 937 case CBOR_TAG_VALUE_SHAREDREF:
850 { 938 {
851 if ((*dec->cur >> 5) != 0) 939 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
852 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)"); 940 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
853 941
854 UV idx = decode_uint (dec); 942 UV idx = decode_uint (dec);
855 943
856 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable)) 944 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable))
857 ERR ("corrupted CBOR data (sharedref index out of bounds)"); 945 ERR ("corrupted CBOR data (sharedref index out of bounds)");
858 946
859 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]); 947 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
948
949 if (sv == &PL_sv_undef)
950 ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
860 } 951 }
861 break; 952 break;
862 953
863 case CBOR_TAG_PERL_OBJECT: 954 case CBOR_TAG_PERL_OBJECT:
864 { 955 {
965static SV * 1056static SV *
966decode_sv (dec_t *dec) 1057decode_sv (dec_t *dec)
967{ 1058{
968 WANT (1); 1059 WANT (1);
969 1060
970 switch (*dec->cur >> 5) 1061 switch (*dec->cur >> MAJOR_SHIFT)
971 { 1062 {
972 case 0: // unsigned int 1063 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
973 return newSVuv (decode_uint (dec)); 1064 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
974 case 1: // negative int 1065 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
975 return newSViv (-1 - (IV)decode_uint (dec)); 1066 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
976 case 2: // octet string 1067 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
977 return decode_str (dec, 0); 1068 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
978 case 3: // utf-8 string 1069 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
979 return decode_str (dec, 1); 1070
980 case 4: // array 1071 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) 1072 switch (*dec->cur++ & MINOR_MASK)
988 { 1073 {
989 case 20: 1074 case SIMPLE_FALSE:
990#if CBOR_SLOW 1075#if CBOR_SLOW
991 types_false = get_bool ("Types::Serialiser::false"); 1076 types_false = get_bool ("Types::Serialiser::false");
992#endif 1077#endif
993 return newSVsv (types_false); 1078 return newSVsv (types_false);
994 case 21: 1079 case SIMPLE_TRUE:
995#if CBOR_SLOW 1080#if CBOR_SLOW
996 types_true = get_bool ("Types::Serialiser::true"); 1081 types_true = get_bool ("Types::Serialiser::true");
997#endif 1082#endif
998 return newSVsv (types_true); 1083 return newSVsv (types_true);
999 case 22: 1084 case SIMPLE_NULL:
1000 return newSVsv (&PL_sv_undef); 1085 return newSVsv (&PL_sv_undef);
1001 case 23: 1086 case SIMPLE_UNDEF:
1002#if CBOR_SLOW 1087#if CBOR_SLOW
1003 types_error = get_bool ("Types::Serialiser::error"); 1088 types_error = get_bool ("Types::Serialiser::error");
1004#endif 1089#endif
1005 return newSVsv (types_error); 1090 return newSVsv (types_error);
1006 1091
1007 case 25: 1092 case MISC_FLOAT16:
1008 { 1093 {
1009 WANT (2); 1094 WANT (2);
1010 1095
1011 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1096 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
1012 dec->cur += 2; 1097 dec->cur += 2;
1013 1098
1014 return newSVnv (ecb_binary16_to_float (fp)); 1099 return newSVnv (ecb_binary16_to_float (fp));
1015 } 1100 }
1016 1101
1017 case 26: 1102 case MISC_FLOAT32:
1018 { 1103 {
1019 uint32_t fp; 1104 uint32_t fp;
1020 WANT (4); 1105 WANT (4);
1021 memcpy (&fp, dec->cur, 4); 1106 memcpy (&fp, dec->cur, 4);
1022 dec->cur += 4; 1107 dec->cur += 4;
1025 fp = ecb_bswap32 (fp); 1110 fp = ecb_bswap32 (fp);
1026 1111
1027 return newSVnv (ecb_binary32_to_float (fp)); 1112 return newSVnv (ecb_binary32_to_float (fp));
1028 } 1113 }
1029 1114
1030 case 27: 1115 case MISC_FLOAT64:
1031 { 1116 {
1032 uint64_t fp; 1117 uint64_t fp;
1033 WANT (8); 1118 WANT (8);
1034 memcpy (&fp, dec->cur, 8); 1119 memcpy (&fp, dec->cur, 8);
1035 dec->cur += 8; 1120 dec->cur += 8;
1038 fp = ecb_bswap64 (fp); 1123 fp = ecb_bswap64 (fp);
1039 1124
1040 return newSVnv (ecb_binary64_to_double (fp)); 1125 return newSVnv (ecb_binary64_to_double (fp));
1041 } 1126 }
1042 1127
1043 // 0..19 unassigned 1128 // 0..19 unassigned simple
1044 // 24 reserved + unassigned (reserved values are not encodable) 1129 // 24 reserved + unassigned (reserved values are not encodable)
1045 default: 1130 default:
1046 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1131 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");
1047 } 1132 }
1048 1133
1136void shrink (CBOR *self, int enable = 1) 1221void shrink (CBOR *self, int enable = 1)
1137 ALIAS: 1222 ALIAS:
1138 shrink = F_SHRINK 1223 shrink = F_SHRINK
1139 allow_unknown = F_ALLOW_UNKNOWN 1224 allow_unknown = F_ALLOW_UNKNOWN
1140 allow_sharing = F_ALLOW_SHARING 1225 allow_sharing = F_ALLOW_SHARING
1141 allow_stringref = F_ALLOW_STRINGREF 1226 allow_cycles = F_ALLOW_CYCLES
1227 pack_strings = F_PACK_STRINGS
1142 PPCODE: 1228 PPCODE:
1143{ 1229{
1144 if (enable) 1230 if (enable)
1145 self->flags |= ix; 1231 self->flags |= ix;
1146 else 1232 else
1152void get_shrink (CBOR *self) 1238void get_shrink (CBOR *self)
1153 ALIAS: 1239 ALIAS:
1154 get_shrink = F_SHRINK 1240 get_shrink = F_SHRINK
1155 get_allow_unknown = F_ALLOW_UNKNOWN 1241 get_allow_unknown = F_ALLOW_UNKNOWN
1156 get_allow_sharing = F_ALLOW_SHARING 1242 get_allow_sharing = F_ALLOW_SHARING
1157 get_allow_stringref = F_ALLOW_STRINGREF 1243 get_allow_cycles = F_ALLOW_CYCLES
1244 get_pack_strings = F_PACK_STRINGS
1158 PPCODE: 1245 PPCODE:
1159 XPUSHs (boolSV (self->flags & ix)); 1246 XPUSHs (boolSV (self->flags & ix));
1160 1247
1161void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1248void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1162 PPCODE: 1249 PPCODE:
1218 cbor_free (self); 1305 cbor_free (self);
1219 1306
1220PROTOTYPES: ENABLE 1307PROTOTYPES: ENABLE
1221 1308
1222void encode_cbor (SV *scalar) 1309void encode_cbor (SV *scalar)
1310 ALIAS:
1311 encode_cbor = 0
1312 encode_cbor_sharing = F_ALLOW_SHARING
1223 PPCODE: 1313 PPCODE:
1224{ 1314{
1225 CBOR cbor; 1315 CBOR cbor;
1226 cbor_init (&cbor); 1316 cbor_init (&cbor);
1317 cbor.flags |= ix;
1227 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1318 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1228 XPUSHs (scalar); 1319 XPUSHs (scalar);
1229} 1320}
1230 1321
1231void decode_cbor (SV *cborstr) 1322void decode_cbor (SV *cborstr)

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines