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 1.26 by root, Fri Nov 22 09:40:13 2013 UTC vs.
Revision 1.39 by root, Sun Dec 1 14:45:03 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)
79# define CBOR_SLOW 0 121# define CBOR_SLOW 0
80# define CBOR_STASH cbor_stash 122# define CBOR_STASH cbor_stash
81#endif 123#endif
82 124
83static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS:: 125static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS::
84static SV *types_true, *types_false, *types_error, *sv_cbor; 126static SV *types_true, *types_false, *types_error, *sv_cbor, *default_filter;
85 127
86typedef struct { 128typedef struct {
87 U32 flags; 129 U32 flags;
88 U32 max_depth; 130 U32 max_depth;
89 STRLEN max_size; 131 STRLEN max_size;
132 SV *filter;
90} CBOR; 133} CBOR;
91 134
92ecb_inline void 135ecb_inline void
93cbor_init (CBOR *cbor) 136cbor_init (CBOR *cbor)
94{ 137{
95 Zero (cbor, 1, CBOR); 138 Zero (cbor, 1, CBOR);
96 cbor->max_depth = 512; 139 cbor->max_depth = 512;
140}
141
142ecb_inline void
143cbor_free (CBOR *cbor)
144{
145 SvREFCNT_dec (cbor->filter);
97} 146}
98 147
99///////////////////////////////////////////////////////////////////////////// 148/////////////////////////////////////////////////////////////////////////////
100// utility functions 149// utility functions
101 150
131{ 180{
132 return idx > 23 181 return idx > 23
133 ? idx > 0xffU 182 ? idx > 0xffU
134 ? idx > 0xffffU 183 ? idx > 0xffffU
135 ? idx > 0xffffffffU 184 ? idx > 0xffffffffU
185 ? 11
136 ? 7 186 : 7
137 : 6
138 : 5 187 : 5
139 : 4 188 : 4
140 : 3; 189 : 3;
141} 190}
142 191
179static void 228static void
180encode_uint (enc_t *enc, int major, UV len) 229encode_uint (enc_t *enc, int major, UV len)
181{ 230{
182 need (enc, 9); 231 need (enc, 9);
183 232
184 if (len < 24) 233 if (ecb_expect_true (len < LENGTH_EXT1))
185 *enc->cur++ = major | len; 234 *enc->cur++ = major | len;
186 else if (len <= 0xff) 235 else if (ecb_expect_true (len <= 0xffU))
187 { 236 {
188 *enc->cur++ = major | 24; 237 *enc->cur++ = major | LENGTH_EXT1;
189 *enc->cur++ = len; 238 *enc->cur++ = len;
190 } 239 }
191 else if (len <= 0xffff) 240 else if (len <= 0xffffU)
192 { 241 {
193 *enc->cur++ = major | 25; 242 *enc->cur++ = major | LENGTH_EXT2;
194 *enc->cur++ = len >> 8; 243 *enc->cur++ = len >> 8;
195 *enc->cur++ = len; 244 *enc->cur++ = len;
196 } 245 }
197 else if (len <= 0xffffffff) 246 else if (len <= 0xffffffffU)
198 { 247 {
199 *enc->cur++ = major | 26; 248 *enc->cur++ = major | LENGTH_EXT4;
200 *enc->cur++ = len >> 24; 249 *enc->cur++ = len >> 24;
201 *enc->cur++ = len >> 16; 250 *enc->cur++ = len >> 16;
202 *enc->cur++ = len >> 8; 251 *enc->cur++ = len >> 8;
203 *enc->cur++ = len; 252 *enc->cur++ = len;
204 } 253 }
205 else 254 else
206 { 255 {
207 *enc->cur++ = major | 27; 256 *enc->cur++ = major | LENGTH_EXT8;
208 *enc->cur++ = len >> 56; 257 *enc->cur++ = len >> 56;
209 *enc->cur++ = len >> 48; 258 *enc->cur++ = len >> 48;
210 *enc->cur++ = len >> 40; 259 *enc->cur++ = len >> 40;
211 *enc->cur++ = len >> 32; 260 *enc->cur++ = len >> 32;
212 *enc->cur++ = len >> 24; 261 *enc->cur++ = len >> 24;
217} 266}
218 267
219ecb_inline void 268ecb_inline void
220encode_tag (enc_t *enc, UV tag) 269encode_tag (enc_t *enc, UV tag)
221{ 270{
222 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;
223} 281}
224 282
225static void 283static void
226encode_str (enc_t *enc, int utf8, char *str, STRLEN len) 284encode_strref (enc_t *enc, int utf8, char *str, STRLEN len)
227{ 285{
228 if (ecb_expect_false (enc->cbor.flags & F_ALLOW_STRINGREF)) 286 if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
229 { 287 {
230 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1); 288 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
231 289
232 if (SvOK (*svp)) 290 if (SvOK (*svp))
233 { 291 {
234 // already registered, use stringref 292 // already registered, use stringref
235 encode_tag (enc, CBOR_TAG_STRINGREF); 293 encode_tag (enc, CBOR_TAG_STRINGREF);
236 encode_uint (enc, 0x00, SvUV (*svp)); 294 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
237 return; 295 return;
238 } 296 }
239 else if (len >= minimum_string_length (enc->stringref_idx)) 297 else if (len >= minimum_string_length (enc->stringref_idx))
240 { 298 {
241 // register only 299 // register only
242 sv_setuv (*svp, enc->stringref_idx); 300 sv_setuv (*svp, enc->stringref_idx);
243 ++enc->stringref_idx; 301 ++enc->stringref_idx;
244 } 302 }
245 } 303 }
246 304
247 encode_uint (enc, utf8 ? 0x60 : 0x40, len); 305 encode_str (enc, utf8, str, len);
248 need (enc, len);
249 memcpy (enc->cur, str, len);
250 enc->cur += len;
251} 306}
252 307
253static void encode_sv (enc_t *enc, SV *sv); 308static void encode_sv (enc_t *enc, SV *sv);
254 309
255static void 310static void
260 if (enc->depth >= enc->cbor.max_depth) 315 if (enc->depth >= enc->cbor.max_depth)
261 croak (ERR_NESTING_EXCEEDED); 316 croak (ERR_NESTING_EXCEEDED);
262 317
263 ++enc->depth; 318 ++enc->depth;
264 319
265 encode_uint (enc, 0x80, len + 1); 320 encode_uint (enc, MAJOR_ARRAY, len + 1);
266 321
267 for (i = 0; i <= len; ++i) 322 for (i = 0; i <= len; ++i)
268 { 323 {
269 SV **svp = av_fetch (av, i, 0); 324 SV **svp = av_fetch (av, i, 0);
270 encode_sv (enc, svp ? *svp : &PL_sv_undef); 325 encode_sv (enc, svp ? *svp : &PL_sv_undef);
285 340
286 int pairs = hv_iterinit (hv); 341 int pairs = hv_iterinit (hv);
287 int mg = SvMAGICAL (hv); 342 int mg = SvMAGICAL (hv);
288 343
289 if (mg) 344 if (mg)
290 encode_ch (enc, 0xa0 | 31); 345 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
291 else 346 else
292 encode_uint (enc, 0xa0, pairs); 347 encode_uint (enc, MAJOR_MAP, pairs);
293 348
294 while ((he = hv_iternext (hv))) 349 while ((he = hv_iternext (hv)))
295 { 350 {
296 if (HeKLEN (he) == HEf_SVKEY) 351 if (HeKLEN (he) == HEf_SVKEY)
297 encode_sv (enc, HeSVKEY (he)); 352 encode_sv (enc, HeSVKEY (he));
298 else 353 else
299 encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); 354 encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
300 355
301 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));
302 } 357 }
303 358
304 if (mg) 359 if (mg)
305 encode_ch (enc, 0xe0 | 31); 360 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
306 361
307 --enc->depth; 362 --enc->depth;
308} 363}
309 364
310// encode objects, arrays and special \0=false and \1=true values. 365// encode objects, arrays and special \0=false and \1=true values.
311static void 366static void
312encode_rv (enc_t *enc, SV *sv) 367encode_rv (enc_t *enc, SV *sv)
313{ 368{
314 SvGETMAGIC (sv); 369 SvGETMAGIC (sv);
315
316 if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)
317 && ecb_expect_false (SvREFCNT (sv) > 1))
318 {
319 if (!enc->shareable)
320 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
321
322 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
323
324 if (SvOK (*svp))
325 {
326 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
327 encode_uint (enc, 0x00, SvUV (*svp));
328 return;
329 }
330 else
331 {
332 sv_setuv (*svp, enc->shareable_idx);
333 ++enc->shareable_idx;
334 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
335 }
336 }
337 370
338 svtype svt = SvTYPE (sv); 371 svtype svt = SvTYPE (sv);
339 372
340 if (ecb_expect_false (SvOBJECT (sv))) 373 if (ecb_expect_false (SvOBJECT (sv)))
341 { 374 {
348 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 381 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
349 ? cbor_tagged_stash 382 ? cbor_tagged_stash
350 : gv_stashpv ("CBOR::XS::Tagged" , 1); 383 : gv_stashpv ("CBOR::XS::Tagged" , 1);
351 384
352 HV *stash = SvSTASH (sv); 385 HV *stash = SvSTASH (sv);
353 GV *method;
354 386
355 if (stash == boolean_stash) 387 if (stash == boolean_stash)
356 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 }
357 else if (stash == error_stash) 392 else if (stash == error_stash)
358 encode_ch (enc, 0xe0 | 23); 393 {
394 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
395 return;
396 }
359 else if (stash == tagged_stash) 397 else if (stash == tagged_stash)
360 { 398 {
361 if (svt != SVt_PVAV) 399 if (svt != SVt_PVAV)
362 croak ("encountered CBOR::XS::Tagged object that isn't an array"); 400 croak ("encountered CBOR::XS::Tagged object that isn't an array");
363 401
364 encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); 402 encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
365 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))
366 } 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
367 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) 436 if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
368 { 437 {
369 dSP; 438 dSP;
370 439
371 ENTER; SAVETMPS; PUSHMARK (SP); 440 ENTER; SAVETMPS; PUSHMARK (SP);
372 // 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
404 // catch this surprisingly common error 473 // catch this surprisingly common error
405 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) 474 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
406 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));
407 476
408 encode_tag (enc, CBOR_TAG_PERL_OBJECT); 477 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
409 encode_uint (enc, 0x80, count + 1); 478 encode_uint (enc, MAJOR_ARRAY, count + 1);
410 encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); 479 encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
411 480
412 while (count) 481 while (count)
413 encode_sv (enc, SP[1 - count--]); 482 encode_sv (enc, SP[1 - count--]);
414 483
415 PUTBACK; 484 PUTBACK;
436{ 505{
437 double nv = SvNVX (sv); 506 double nv = SvNVX (sv);
438 507
439 need (enc, 9); 508 need (enc, 9);
440 509
441 if (ecb_expect_false (nv == (U32)nv)) 510 if (ecb_expect_false (nv == (NV)(U32)nv))
442 encode_uint (enc, 0x00, (U32)nv); 511 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
443 //TODO: maybe I32? 512 //TODO: maybe I32?
444 else if (ecb_expect_false (nv == (float)nv)) 513 else if (ecb_expect_false (nv == (float)nv))
445 { 514 {
446 uint32_t fp = ecb_float_to_binary32 (nv); 515 uint32_t fp = ecb_float_to_binary32 (nv);
447 516
448 *enc->cur++ = 0xe0 | 26; 517 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
449 518
450 if (!ecb_big_endian ()) 519 if (!ecb_big_endian ())
451 fp = ecb_bswap32 (fp); 520 fp = ecb_bswap32 (fp);
452 521
453 memcpy (enc->cur, &fp, 4); 522 memcpy (enc->cur, &fp, 4);
455 } 524 }
456 else 525 else
457 { 526 {
458 uint64_t fp = ecb_double_to_binary64 (nv); 527 uint64_t fp = ecb_double_to_binary64 (nv);
459 528
460 *enc->cur++ = 0xe0 | 27; 529 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
461 530
462 if (!ecb_big_endian ()) 531 if (!ecb_big_endian ())
463 fp = ecb_bswap64 (fp); 532 fp = ecb_bswap64 (fp);
464 533
465 memcpy (enc->cur, &fp, 8); 534 memcpy (enc->cur, &fp, 8);
474 543
475 if (SvPOKp (sv)) 544 if (SvPOKp (sv))
476 { 545 {
477 STRLEN len; 546 STRLEN len;
478 char *str = SvPV (sv, len); 547 char *str = SvPV (sv, len);
479 encode_str (enc, SvUTF8 (sv), str, len); 548 encode_strref (enc, SvUTF8 (sv), str, len);
480 } 549 }
481 else if (SvNOKp (sv)) 550 else if (SvNOKp (sv))
482 encode_nv (enc, sv); 551 encode_nv (enc, sv);
483 else if (SvIOKp (sv)) 552 else if (SvIOKp (sv))
484 { 553 {
485 if (SvIsUV (sv)) 554 if (SvIsUV (sv))
486 encode_uint (enc, 0x00, SvUVX (sv)); 555 encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
487 else if (SvIVX (sv) >= 0) 556 else if (SvIVX (sv) >= 0)
488 encode_uint (enc, 0x00, SvIVX (sv)); 557 encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
489 else 558 else
490 encode_uint (enc, 0x20, -(SvIVX (sv) + 1)); 559 encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
491 } 560 }
492 else if (SvROK (sv)) 561 else if (SvROK (sv))
493 encode_rv (enc, SvRV (sv)); 562 encode_rv (enc, SvRV (sv));
494 else if (!SvOK (sv)) 563 else if (!SvOK (sv))
495 encode_ch (enc, 0xe0 | 22); 564 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
496 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 565 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
497 encode_ch (enc, 0xe0 | 23); 566 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
498 else 567 else
499 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",
500 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 569 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
501} 570}
502 571
510 enc.cur = SvPVX (enc.sv); 579 enc.cur = SvPVX (enc.sv);
511 enc.end = SvEND (enc.sv); 580 enc.end = SvEND (enc.sv);
512 581
513 SvPOK_only (enc.sv); 582 SvPOK_only (enc.sv);
514 583
515 if (cbor->flags & F_ALLOW_STRINGREF) 584 if (cbor->flags & F_PACK_STRINGS)
516 { 585 {
517 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE); 586 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
518 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ()); 587 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
519 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ()); 588 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
520 } 589 }
542 CBOR cbor; 611 CBOR cbor;
543 U32 depth; // recursion depth 612 U32 depth; // recursion depth
544 U32 maxdepth; // recursion depth limit 613 U32 maxdepth; // recursion depth limit
545 AV *shareable; 614 AV *shareable;
546 AV *stringref; 615 AV *stringref;
616 SV *decode_tagged;
547} dec_t; 617} dec_t;
548 618
549#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE 619#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE
550 620
551#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data") 621#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data")
554#define DEC_DEC_DEPTH --dec->depth 624#define DEC_DEC_DEPTH --dec->depth
555 625
556static UV 626static UV
557decode_uint (dec_t *dec) 627decode_uint (dec_t *dec)
558{ 628{
559 switch (*dec->cur & 31) 629 U8 m = *dec->cur & MINOR_MASK;
560 { 630 ++dec->cur;
561 case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
562 case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
563 case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
564 return *dec->cur++ & 31;
565 631
566 case 24: 632 if (ecb_expect_true (m < LENGTH_EXT1))
633 return m;
634 else if (ecb_expect_true (m == LENGTH_EXT1))
635 {
567 WANT (2); 636 WANT (1);
568 dec->cur += 2; 637 dec->cur += 1;
569 return dec->cur[-1]; 638 return dec->cur[-1];
570 639 }
571 case 25: 640 else if (ecb_expect_true (m == LENGTH_EXT2))
641 {
572 WANT (3); 642 WANT (2);
573 dec->cur += 3; 643 dec->cur += 2;
574 return (((UV)dec->cur[-2]) << 8) 644 return (((UV)dec->cur[-2]) << 8)
575 | ((UV)dec->cur[-1]); 645 | ((UV)dec->cur[-1]);
576 646 }
577 case 26: 647 else if (ecb_expect_true (m == LENGTH_EXT4))
648 {
578 WANT (5); 649 WANT (4);
579 dec->cur += 5; 650 dec->cur += 4;
580 return (((UV)dec->cur[-4]) << 24) 651 return (((UV)dec->cur[-4]) << 24)
581 | (((UV)dec->cur[-3]) << 16) 652 | (((UV)dec->cur[-3]) << 16)
582 | (((UV)dec->cur[-2]) << 8) 653 | (((UV)dec->cur[-2]) << 8)
583 | ((UV)dec->cur[-1]); 654 | ((UV)dec->cur[-1]);
584 655 }
585 case 27: 656 else if (ecb_expect_true (m == LENGTH_EXT8))
657 {
586 WANT (9); 658 WANT (8);
587 dec->cur += 9; 659 dec->cur += 8;
660
661 return
662#if UVSIZE < 8
663 0
664#else
588 return (((UV)dec->cur[-8]) << 56) 665 (((UV)dec->cur[-8]) << 56)
589 | (((UV)dec->cur[-7]) << 48) 666 | (((UV)dec->cur[-7]) << 48)
590 | (((UV)dec->cur[-6]) << 40) 667 | (((UV)dec->cur[-6]) << 40)
591 | (((UV)dec->cur[-5]) << 32) 668 | (((UV)dec->cur[-5]) << 32)
669#endif
592 | (((UV)dec->cur[-4]) << 24) 670 | (((UV)dec->cur[-4]) << 24)
593 | (((UV)dec->cur[-3]) << 16) 671 | (((UV)dec->cur[-3]) << 16)
594 | (((UV)dec->cur[-2]) << 8) 672 | (((UV)dec->cur[-2]) << 8)
595 | ((UV)dec->cur[-1]); 673 | ((UV)dec->cur[-1]);
596 674 }
597 default: 675 else
598 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 676 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
599 }
600 677
601fail: 678fail:
602 return 0; 679 return 0;
603} 680}
604 681
609{ 686{
610 AV *av = newAV (); 687 AV *av = newAV ();
611 688
612 DEC_INC_DEPTH; 689 DEC_INC_DEPTH;
613 690
614 if ((*dec->cur & 31) == 31) 691 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
615 { 692 {
616 ++dec->cur; 693 ++dec->cur;
617 694
618 for (;;) 695 for (;;)
619 { 696 {
620 WANT (1); 697 WANT (1);
621 698
622 if (*dec->cur == (0xe0 | 31)) 699 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
623 { 700 {
624 ++dec->cur; 701 ++dec->cur;
625 break; 702 break;
626 } 703 }
627 704
630 } 707 }
631 else 708 else
632 { 709 {
633 int i, len = decode_uint (dec); 710 int i, len = decode_uint (dec);
634 711
712 WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
635 av_fill (av, len - 1); 713 av_fill (av, len - 1);
636 714
637 for (i = 0; i < len; ++i) 715 for (i = 0; i < len; ++i)
638 AvARRAY (av)[i] = decode_sv (dec); 716 AvARRAY (av)[i] = decode_sv (dec);
639 } 717 }
652{ 730{
653 // for speed reasons, we specialcase single-string 731 // for speed reasons, we specialcase single-string
654 // byte or utf-8 strings as keys, but only when !stringref 732 // byte or utf-8 strings as keys, but only when !stringref
655 733
656 if (ecb_expect_true (!dec->stringref)) 734 if (ecb_expect_true (!dec->stringref))
657 if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27) 735 if (ecb_expect_true ((*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
658 { 736 {
659 I32 len = decode_uint (dec); 737 I32 len = decode_uint (dec);
660 char *key = (char *)dec->cur; 738 char *key = (char *)dec->cur;
661 739
662 dec->cur += len; 740 dec->cur += len;
663 741
664 if (ecb_expect_false (dec->stringref))
665 av_push (dec->stringref, newSVpvn (key, len));
666
667 hv_store (hv, key, len, decode_sv (dec), 0); 742 hv_store (hv, key, len, decode_sv (dec), 0);
668 743
669 return; 744 return;
670 } 745 }
671 else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27) 746 else if (ecb_expect_true ((*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
672 { 747 {
673 I32 len = decode_uint (dec); 748 I32 len = decode_uint (dec);
674 char *key = (char *)dec->cur; 749 char *key = (char *)dec->cur;
675 750
676 dec->cur += len; 751 dec->cur += len;
677 752
678 if (ecb_expect_false (dec->stringref)) 753 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
679 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)");
680 756
681 hv_store (hv, key, -len, decode_sv (dec), 0); 757 hv_store (hv, key, -len, decode_sv (dec), 0);
682 758
683 return; 759 return;
684 } 760 }
686 SV *k = decode_sv (dec); 762 SV *k = decode_sv (dec);
687 SV *v = decode_sv (dec); 763 SV *v = decode_sv (dec);
688 764
689 hv_store_ent (hv, k, v, 0); 765 hv_store_ent (hv, k, v, 0);
690 SvREFCNT_dec (k); 766 SvREFCNT_dec (k);
767
768fail:
769 ;
691} 770}
692 771
693static SV * 772static SV *
694decode_hv (dec_t *dec) 773decode_hv (dec_t *dec)
695{ 774{
696 HV *hv = newHV (); 775 HV *hv = newHV ();
697 776
698 DEC_INC_DEPTH; 777 DEC_INC_DEPTH;
699 778
700 if ((*dec->cur & 31) == 31) 779 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
701 { 780 {
702 ++dec->cur; 781 ++dec->cur;
703 782
704 for (;;) 783 for (;;)
705 { 784 {
706 WANT (1); 785 WANT (1);
707 786
708 if (*dec->cur == (0xe0 | 31)) 787 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
709 { 788 {
710 ++dec->cur; 789 ++dec->cur;
711 break; 790 break;
712 } 791 }
713 792
734static SV * 813static SV *
735decode_str (dec_t *dec, int utf8) 814decode_str (dec_t *dec, int utf8)
736{ 815{
737 SV *sv = 0; 816 SV *sv = 0;
738 817
739 if ((*dec->cur & 31) == 31) 818 if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
740 { 819 {
820 // indefinite length strings
741 ++dec->cur; 821 ++dec->cur;
742 822
823 U8 major = *dec->cur & MAJOR_MISC;
824
743 sv = newSVpvn ("", 0); 825 sv = newSVpvn ("", 0);
744 826
745 // not very fast, and certainly not robust against illegal input
746 for (;;) 827 for (;;)
747 { 828 {
748 WANT (1); 829 WANT (1);
749 830
750 if (*dec->cur == (0xe0 | 31)) 831 if ((*dec->cur - major) > LENGTH_EXT8)
832 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
751 { 833 {
752 ++dec->cur; 834 ++dec->cur;
753 break; 835 break;
754 } 836 }
837 else
838 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
755 839
756 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;
757 } 845 }
758 } 846 }
759 else 847 else
760 { 848 {
761 STRLEN len = decode_uint (dec); 849 STRLEN len = decode_uint (dec);
768 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1)) 856 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
769 av_push (dec->stringref, SvREFCNT_inc_NN (sv)); 857 av_push (dec->stringref, SvREFCNT_inc_NN (sv));
770 } 858 }
771 859
772 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
773 SvUTF8_on (sv); 866 SvUTF8_on (sv);
867 }
774 868
775 return sv; 869 return sv;
776 870
777fail: 871fail:
778 SvREFCNT_dec (sv); 872 SvREFCNT_dec (sv);
810 } 904 }
811 break; 905 break;
812 906
813 case CBOR_TAG_STRINGREF: 907 case CBOR_TAG_STRINGREF:
814 { 908 {
815 if ((*dec->cur >> 5) != 0) 909 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
816 ERR ("corrupted CBOR data (stringref index not an unsigned integer)"); 910 ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
817 911
818 UV idx = decode_uint (dec); 912 UV idx = decode_uint (dec);
819 913
820 if (!dec->stringref || (int)idx > AvFILLp (dec->stringref)) 914 if (!dec->stringref || (int)idx > AvFILLp (dec->stringref))
827 case CBOR_TAG_VALUE_SHAREABLE: 921 case CBOR_TAG_VALUE_SHAREABLE:
828 { 922 {
829 if (ecb_expect_false (!dec->shareable)) 923 if (ecb_expect_false (!dec->shareable))
830 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ()); 924 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
831 925
926 if (dec->cbor.flags & F_ALLOW_CYCLES)
927 {
832 sv = newSV (0); 928 sv = newSV (0);
833 av_push (dec->shareable, SvREFCNT_inc_NN (sv)); 929 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
834 930
835 SV *osv = decode_sv (dec); 931 SV *osv = decode_sv (dec);
836 sv_setsv (sv, osv); 932 sv_setsv (sv, osv);
837 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 }
838 } 942 }
839 break; 943 break;
840 944
841 case CBOR_TAG_VALUE_SHAREDREF: 945 case CBOR_TAG_VALUE_SHAREDREF:
842 { 946 {
843 if ((*dec->cur >> 5) != 0) 947 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
844 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)"); 948 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
845 949
846 UV idx = decode_uint (dec); 950 UV idx = decode_uint (dec);
847 951
848 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable)) 952 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable))
849 ERR ("corrupted CBOR data (sharedref index out of bounds)"); 953 ERR ("corrupted CBOR data (sharedref index out of bounds)");
850 954
851 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");
852 } 959 }
853 break; 960 break;
854 961
855 case CBOR_TAG_PERL_OBJECT: 962 case CBOR_TAG_PERL_OBJECT:
856 { 963 {
905 1012
906 default: 1013 default:
907 { 1014 {
908 sv = decode_sv (dec); 1015 sv = decode_sv (dec);
909 1016
1017 dSP;
1018 ENTER; SAVETMPS; PUSHMARK (SP);
1019 EXTEND (SP, 2);
1020 PUSHs (newSVuv (tag));
1021 PUSHs (sv);
1022
1023 PUTBACK;
1024 int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL);
1025 SPAGAIN;
1026
1027 if (SvTRUE (ERRSV))
1028 {
1029 FREETMPS; LEAVE;
1030 ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));
1031 }
1032
1033 if (count)
1034 {
1035 SvREFCNT_dec (sv);
1036 sv = SvREFCNT_inc (POPs);
1037 }
1038 else
1039 {
910 AV *av = newAV (); 1040 AV *av = newAV ();
911 av_push (av, newSVuv (tag)); 1041 av_push (av, newSVuv (tag));
912 av_push (av, sv); 1042 av_push (av, sv);
913 1043
914 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 1044 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
915 ? cbor_tagged_stash 1045 ? cbor_tagged_stash
916 : gv_stashpv ("CBOR::XS::Tagged" , 1); 1046 : gv_stashpv ("CBOR::XS::Tagged" , 1);
917
918 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash); 1047 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
1048 }
1049
1050 PUTBACK;
1051
1052 FREETMPS; LEAVE;
919 } 1053 }
920 break; 1054 break;
921 } 1055 }
922 1056
923 return sv; 1057 return sv;
930static SV * 1064static SV *
931decode_sv (dec_t *dec) 1065decode_sv (dec_t *dec)
932{ 1066{
933 WANT (1); 1067 WANT (1);
934 1068
935 switch (*dec->cur >> 5) 1069 switch (*dec->cur >> MAJOR_SHIFT)
936 { 1070 {
937 case 0: // unsigned int 1071 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
938 return newSVuv (decode_uint (dec)); 1072 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
939 case 1: // negative int 1073 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
940 return newSViv (-1 - (IV)decode_uint (dec)); 1074 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
941 case 2: // octet string 1075 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
942 return decode_str (dec, 0); 1076 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
943 case 3: // utf-8 string 1077 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
944 return decode_str (dec, 1); 1078
945 case 4: // array 1079 case MAJOR_MISC >> MAJOR_SHIFT:
946 return decode_av (dec);
947 case 5: // map
948 return decode_hv (dec);
949 case 6: // tag
950 return decode_tagged (dec);
951 case 7: // misc
952 switch (*dec->cur++ & 31) 1080 switch (*dec->cur++ & MINOR_MASK)
953 { 1081 {
954 case 20: 1082 case SIMPLE_FALSE:
955#if CBOR_SLOW 1083#if CBOR_SLOW
956 types_false = get_bool ("Types::Serialiser::false"); 1084 types_false = get_bool ("Types::Serialiser::false");
957#endif 1085#endif
958 return newSVsv (types_false); 1086 return newSVsv (types_false);
959 case 21: 1087 case SIMPLE_TRUE:
960#if CBOR_SLOW 1088#if CBOR_SLOW
961 types_true = get_bool ("Types::Serialiser::true"); 1089 types_true = get_bool ("Types::Serialiser::true");
962#endif 1090#endif
963 return newSVsv (types_true); 1091 return newSVsv (types_true);
964 case 22: 1092 case SIMPLE_NULL:
965 return newSVsv (&PL_sv_undef); 1093 return newSVsv (&PL_sv_undef);
966 case 23: 1094 case SIMPLE_UNDEF:
967#if CBOR_SLOW 1095#if CBOR_SLOW
968 types_error = get_bool ("Types::Serialiser::error"); 1096 types_error = get_bool ("Types::Serialiser::error");
969#endif 1097#endif
970 return newSVsv (types_error); 1098 return newSVsv (types_error);
971 1099
972 case 25: 1100 case MISC_FLOAT16:
973 { 1101 {
974 WANT (2); 1102 WANT (2);
975 1103
976 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1104 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
977 dec->cur += 2; 1105 dec->cur += 2;
978 1106
979 return newSVnv (ecb_binary16_to_float (fp)); 1107 return newSVnv (ecb_binary16_to_float (fp));
980 } 1108 }
981 1109
982 case 26: 1110 case MISC_FLOAT32:
983 { 1111 {
984 uint32_t fp; 1112 uint32_t fp;
985 WANT (4); 1113 WANT (4);
986 memcpy (&fp, dec->cur, 4); 1114 memcpy (&fp, dec->cur, 4);
987 dec->cur += 4; 1115 dec->cur += 4;
990 fp = ecb_bswap32 (fp); 1118 fp = ecb_bswap32 (fp);
991 1119
992 return newSVnv (ecb_binary32_to_float (fp)); 1120 return newSVnv (ecb_binary32_to_float (fp));
993 } 1121 }
994 1122
995 case 27: 1123 case MISC_FLOAT64:
996 { 1124 {
997 uint64_t fp; 1125 uint64_t fp;
998 WANT (8); 1126 WANT (8);
999 memcpy (&fp, dec->cur, 8); 1127 memcpy (&fp, dec->cur, 8);
1000 dec->cur += 8; 1128 dec->cur += 8;
1003 fp = ecb_bswap64 (fp); 1131 fp = ecb_bswap64 (fp);
1004 1132
1005 return newSVnv (ecb_binary64_to_double (fp)); 1133 return newSVnv (ecb_binary64_to_double (fp));
1006 } 1134 }
1007 1135
1008 // 0..19 unassigned 1136 // 0..19 unassigned simple
1009 // 24 reserved + unassigned (reserved values are not encodable) 1137 // 24 reserved + unassigned (reserved values are not encodable)
1010 default: 1138 default:
1011 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1139 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");
1012 } 1140 }
1013 1141
1043 if (dec.cur != dec.end && !dec.err) 1171 if (dec.cur != dec.end && !dec.err)
1044 dec.err = "garbage after CBOR object"; 1172 dec.err = "garbage after CBOR object";
1045 1173
1046 if (dec.err) 1174 if (dec.err)
1047 { 1175 {
1176 if (dec.shareable)
1177 {
1178 // need to break cyclic links, which whould all be in shareable
1179 int i;
1180 SV **svp;
1181
1182 for (i = av_len (dec.shareable) + 1; i--; )
1183 if ((svp = av_fetch (dec.shareable, i, 0)))
1184 sv_setsv (*svp, &PL_sv_undef);
1185 }
1186
1048 SvREFCNT_dec (sv); 1187 SvREFCNT_dec (sv);
1049 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur); 1188 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur);
1050 } 1189 }
1051 1190
1052 sv = sv_2mortal (sv); 1191 sv = sv_2mortal (sv);
1068 types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1); 1207 types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1);
1069 1208
1070 types_true = get_bool ("Types::Serialiser::true" ); 1209 types_true = get_bool ("Types::Serialiser::true" );
1071 types_false = get_bool ("Types::Serialiser::false"); 1210 types_false = get_bool ("Types::Serialiser::false");
1072 types_error = get_bool ("Types::Serialiser::error"); 1211 types_error = get_bool ("Types::Serialiser::error");
1212
1213 default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1073 1214
1074 sv_cbor = newSVpv ("CBOR", 0); 1215 sv_cbor = newSVpv ("CBOR", 0);
1075 SvREADONLY_on (sv_cbor); 1216 SvREADONLY_on (sv_cbor);
1076} 1217}
1077 1218
1099void shrink (CBOR *self, int enable = 1) 1240void shrink (CBOR *self, int enable = 1)
1100 ALIAS: 1241 ALIAS:
1101 shrink = F_SHRINK 1242 shrink = F_SHRINK
1102 allow_unknown = F_ALLOW_UNKNOWN 1243 allow_unknown = F_ALLOW_UNKNOWN
1103 allow_sharing = F_ALLOW_SHARING 1244 allow_sharing = F_ALLOW_SHARING
1104 allow_stringref = F_ALLOW_STRINGREF 1245 allow_cycles = F_ALLOW_CYCLES
1246 pack_strings = F_PACK_STRINGS
1247 validate_utf8 = F_VALIDATE_UTF8
1105 PPCODE: 1248 PPCODE:
1106{ 1249{
1107 if (enable) 1250 if (enable)
1108 self->flags |= ix; 1251 self->flags |= ix;
1109 else 1252 else
1115void get_shrink (CBOR *self) 1258void get_shrink (CBOR *self)
1116 ALIAS: 1259 ALIAS:
1117 get_shrink = F_SHRINK 1260 get_shrink = F_SHRINK
1118 get_allow_unknown = F_ALLOW_UNKNOWN 1261 get_allow_unknown = F_ALLOW_UNKNOWN
1119 get_allow_sharing = F_ALLOW_SHARING 1262 get_allow_sharing = F_ALLOW_SHARING
1120 get_allow_stringref = F_ALLOW_STRINGREF 1263 get_allow_cycles = F_ALLOW_CYCLES
1264 get_pack_strings = F_PACK_STRINGS
1265 get_validate_utf8 = F_VALIDATE_UTF8
1121 PPCODE: 1266 PPCODE:
1122 XPUSHs (boolSV (self->flags & ix)); 1267 XPUSHs (boolSV (self->flags & ix));
1123 1268
1124void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1269void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1125 PPCODE: 1270 PPCODE:
1138 XPUSHs (ST (0)); 1283 XPUSHs (ST (0));
1139 1284
1140int get_max_size (CBOR *self) 1285int get_max_size (CBOR *self)
1141 CODE: 1286 CODE:
1142 RETVAL = self->max_size; 1287 RETVAL = self->max_size;
1288 OUTPUT:
1289 RETVAL
1290
1291void filter (CBOR *self, SV *filter = 0)
1292 PPCODE:
1293 SvREFCNT_dec (self->filter);
1294 self->filter = filter ? newSVsv (filter) : filter;
1295 XPUSHs (ST (0));
1296
1297SV *get_filter (CBOR *self)
1298 CODE:
1299 RETVAL = self->filter ? self->filter : NEWSV (0, 0);
1143 OUTPUT: 1300 OUTPUT:
1144 RETVAL 1301 RETVAL
1145 1302
1146void encode (CBOR *self, SV *scalar) 1303void encode (CBOR *self, SV *scalar)
1147 PPCODE: 1304 PPCODE:
1162 EXTEND (SP, 2); 1319 EXTEND (SP, 2);
1163 PUSHs (sv); 1320 PUSHs (sv);
1164 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1321 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1165} 1322}
1166 1323
1324void DESTROY (CBOR *self)
1325 PPCODE:
1326 cbor_free (self);
1327
1167PROTOTYPES: ENABLE 1328PROTOTYPES: ENABLE
1168 1329
1169void encode_cbor (SV *scalar) 1330void encode_cbor (SV *scalar)
1331 ALIAS:
1332 encode_cbor = 0
1333 encode_cbor_sharing = F_ALLOW_SHARING
1170 PPCODE: 1334 PPCODE:
1171{ 1335{
1172 CBOR cbor; 1336 CBOR cbor;
1173 cbor_init (&cbor); 1337 cbor_init (&cbor);
1338 cbor.flags |= ix;
1174 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1339 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1175 XPUSHs (scalar); 1340 XPUSHs (scalar);
1176} 1341}
1177 1342
1178void decode_cbor (SV *cborstr) 1343void decode_cbor (SV *cborstr)

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines