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 1.41 by root, Tue Dec 10 14:01:52 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)
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;
133
134 // for the incremental parser
135 STRLEN incr_pos; // the current offset into the text
136 STRLEN incr_need; // minimum bytes needed to decode
137 AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef.
90} CBOR; 138} CBOR;
91 139
92ecb_inline void 140ecb_inline void
93cbor_init (CBOR *cbor) 141cbor_init (CBOR *cbor)
94{ 142{
95 Zero (cbor, 1, CBOR); 143 Zero (cbor, 1, CBOR);
96 cbor->max_depth = 512; 144 cbor->max_depth = 512;
145}
146
147ecb_inline void
148cbor_free (CBOR *cbor)
149{
150 SvREFCNT_dec (cbor->filter);
151 SvREFCNT_dec (cbor->incr_count);
97} 152}
98 153
99///////////////////////////////////////////////////////////////////////////// 154/////////////////////////////////////////////////////////////////////////////
100// utility functions 155// utility functions
101 156
131{ 186{
132 return idx > 23 187 return idx > 23
133 ? idx > 0xffU 188 ? idx > 0xffU
134 ? idx > 0xffffU 189 ? idx > 0xffffU
135 ? idx > 0xffffffffU 190 ? idx > 0xffffffffU
191 ? 11
136 ? 7 192 : 7
137 : 6
138 : 5 193 : 5
139 : 4 194 : 4
140 : 3; 195 : 3;
141} 196}
142 197
179static void 234static void
180encode_uint (enc_t *enc, int major, UV len) 235encode_uint (enc_t *enc, int major, UV len)
181{ 236{
182 need (enc, 9); 237 need (enc, 9);
183 238
184 if (len < 24) 239 if (ecb_expect_true (len < LENGTH_EXT1))
185 *enc->cur++ = major | len; 240 *enc->cur++ = major | len;
186 else if (len <= 0xff) 241 else if (ecb_expect_true (len <= 0xffU))
187 { 242 {
188 *enc->cur++ = major | 24; 243 *enc->cur++ = major | LENGTH_EXT1;
189 *enc->cur++ = len; 244 *enc->cur++ = len;
190 } 245 }
191 else if (len <= 0xffff) 246 else if (len <= 0xffffU)
192 { 247 {
193 *enc->cur++ = major | 25; 248 *enc->cur++ = major | LENGTH_EXT2;
194 *enc->cur++ = len >> 8; 249 *enc->cur++ = len >> 8;
195 *enc->cur++ = len; 250 *enc->cur++ = len;
196 } 251 }
197 else if (len <= 0xffffffff) 252 else if (len <= 0xffffffffU)
198 { 253 {
199 *enc->cur++ = major | 26; 254 *enc->cur++ = major | LENGTH_EXT4;
200 *enc->cur++ = len >> 24; 255 *enc->cur++ = len >> 24;
201 *enc->cur++ = len >> 16; 256 *enc->cur++ = len >> 16;
202 *enc->cur++ = len >> 8; 257 *enc->cur++ = len >> 8;
203 *enc->cur++ = len; 258 *enc->cur++ = len;
204 } 259 }
205 else 260 else
206 { 261 {
207 *enc->cur++ = major | 27; 262 *enc->cur++ = major | LENGTH_EXT8;
208 *enc->cur++ = len >> 56; 263 *enc->cur++ = len >> 56;
209 *enc->cur++ = len >> 48; 264 *enc->cur++ = len >> 48;
210 *enc->cur++ = len >> 40; 265 *enc->cur++ = len >> 40;
211 *enc->cur++ = len >> 32; 266 *enc->cur++ = len >> 32;
212 *enc->cur++ = len >> 24; 267 *enc->cur++ = len >> 24;
217} 272}
218 273
219ecb_inline void 274ecb_inline void
220encode_tag (enc_t *enc, UV tag) 275encode_tag (enc_t *enc, UV tag)
221{ 276{
222 encode_uint (enc, 0xc0, tag); 277 encode_uint (enc, MAJOR_TAG, tag);
278}
279
280ecb_inline void
281encode_str (enc_t *enc, int utf8, char *str, STRLEN len)
282{
283 encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
284 need (enc, len);
285 memcpy (enc->cur, str, len);
286 enc->cur += len;
223} 287}
224 288
225static void 289static void
226encode_str (enc_t *enc, int utf8, char *str, STRLEN len) 290encode_strref (enc_t *enc, int utf8, char *str, STRLEN len)
227{ 291{
228 if (ecb_expect_false (enc->cbor.flags & F_ALLOW_STRINGREF)) 292 if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
229 { 293 {
230 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1); 294 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
231 295
232 if (SvOK (*svp)) 296 if (SvOK (*svp))
233 { 297 {
234 // already registered, use stringref 298 // already registered, use stringref
235 encode_tag (enc, CBOR_TAG_STRINGREF); 299 encode_tag (enc, CBOR_TAG_STRINGREF);
236 encode_uint (enc, 0x00, SvUV (*svp)); 300 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
237 return; 301 return;
238 } 302 }
239 else if (len >= minimum_string_length (enc->stringref_idx)) 303 else if (len >= minimum_string_length (enc->stringref_idx))
240 { 304 {
241 // register only 305 // register only
242 sv_setuv (*svp, enc->stringref_idx); 306 sv_setuv (*svp, enc->stringref_idx);
243 ++enc->stringref_idx; 307 ++enc->stringref_idx;
244 } 308 }
245 } 309 }
246 310
247 encode_uint (enc, utf8 ? 0x60 : 0x40, len); 311 encode_str (enc, utf8, str, len);
248 need (enc, len);
249 memcpy (enc->cur, str, len);
250 enc->cur += len;
251} 312}
252 313
253static void encode_sv (enc_t *enc, SV *sv); 314static void encode_sv (enc_t *enc, SV *sv);
254 315
255static void 316static void
260 if (enc->depth >= enc->cbor.max_depth) 321 if (enc->depth >= enc->cbor.max_depth)
261 croak (ERR_NESTING_EXCEEDED); 322 croak (ERR_NESTING_EXCEEDED);
262 323
263 ++enc->depth; 324 ++enc->depth;
264 325
265 encode_uint (enc, 0x80, len + 1); 326 encode_uint (enc, MAJOR_ARRAY, len + 1);
266 327
267 for (i = 0; i <= len; ++i) 328 for (i = 0; i <= len; ++i)
268 { 329 {
269 SV **svp = av_fetch (av, i, 0); 330 SV **svp = av_fetch (av, i, 0);
270 encode_sv (enc, svp ? *svp : &PL_sv_undef); 331 encode_sv (enc, svp ? *svp : &PL_sv_undef);
285 346
286 int pairs = hv_iterinit (hv); 347 int pairs = hv_iterinit (hv);
287 int mg = SvMAGICAL (hv); 348 int mg = SvMAGICAL (hv);
288 349
289 if (mg) 350 if (mg)
290 encode_ch (enc, 0xa0 | 31); 351 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
291 else 352 else
292 encode_uint (enc, 0xa0, pairs); 353 encode_uint (enc, MAJOR_MAP, pairs);
293 354
294 while ((he = hv_iternext (hv))) 355 while ((he = hv_iternext (hv)))
295 { 356 {
296 if (HeKLEN (he) == HEf_SVKEY) 357 if (HeKLEN (he) == HEf_SVKEY)
297 encode_sv (enc, HeSVKEY (he)); 358 encode_sv (enc, HeSVKEY (he));
298 else 359 else
299 encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); 360 encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
300 361
301 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); 362 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
302 } 363 }
303 364
304 if (mg) 365 if (mg)
305 encode_ch (enc, 0xe0 | 31); 366 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
306 367
307 --enc->depth; 368 --enc->depth;
308} 369}
309 370
310// encode objects, arrays and special \0=false and \1=true values. 371// encode objects, arrays and special \0=false and \1=true values.
311static void 372static void
312encode_rv (enc_t *enc, SV *sv) 373encode_rv (enc_t *enc, SV *sv)
313{ 374{
314 SvGETMAGIC (sv); 375 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 }
337 376
338 svtype svt = SvTYPE (sv); 377 svtype svt = SvTYPE (sv);
339 378
340 if (ecb_expect_false (SvOBJECT (sv))) 379 if (ecb_expect_false (SvOBJECT (sv)))
341 { 380 {
348 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 387 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
349 ? cbor_tagged_stash 388 ? cbor_tagged_stash
350 : gv_stashpv ("CBOR::XS::Tagged" , 1); 389 : gv_stashpv ("CBOR::XS::Tagged" , 1);
351 390
352 HV *stash = SvSTASH (sv); 391 HV *stash = SvSTASH (sv);
353 GV *method;
354 392
355 if (stash == boolean_stash) 393 if (stash == boolean_stash)
356 encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20); 394 {
395 encode_ch (enc, SvIV (sv) ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE);
396 return;
397 }
357 else if (stash == error_stash) 398 else if (stash == error_stash)
358 encode_ch (enc, 0xe0 | 23); 399 {
400 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
401 return;
402 }
359 else if (stash == tagged_stash) 403 else if (stash == tagged_stash)
360 { 404 {
361 if (svt != SVt_PVAV) 405 if (svt != SVt_PVAV)
362 croak ("encountered CBOR::XS::Tagged object that isn't an array"); 406 croak ("encountered CBOR::XS::Tagged object that isn't an array");
363 407
364 encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); 408 encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
365 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); 409 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1));
410
411 return;
412 }
413 }
414
415 if (ecb_expect_false (SvREFCNT (sv) > 1)
416 && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
417 {
418 if (!enc->shareable)
419 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
420
421 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
422
423 if (SvOK (*svp))
366 } 424 {
425 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
426 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
427 return;
428 }
429 else
430 {
431 sv_setuv (*svp, enc->shareable_idx);
432 ++enc->shareable_idx;
433 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
434 }
435 }
436
437 if (ecb_expect_false (SvOBJECT (sv)))
438 {
439 HV *stash = SvSTASH (sv);
440 GV *method;
441
367 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) 442 if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
368 { 443 {
369 dSP; 444 dSP;
370 445
371 ENTER; SAVETMPS; PUSHMARK (SP); 446 ENTER; SAVETMPS; PUSHMARK (SP);
372 // we re-bless the reference to get overload and other niceties right 447 // we re-bless the reference to get overload and other niceties right
404 // catch this surprisingly common error 479 // catch this surprisingly common error
405 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) 480 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)); 481 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
407 482
408 encode_tag (enc, CBOR_TAG_PERL_OBJECT); 483 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
409 encode_uint (enc, 0x80, count + 1); 484 encode_uint (enc, MAJOR_ARRAY, count + 1);
410 encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); 485 encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
411 486
412 while (count) 487 while (count)
413 encode_sv (enc, SP[1 - count--]); 488 encode_sv (enc, SP[1 - count--]);
414 489
415 PUTBACK; 490 PUTBACK;
436{ 511{
437 double nv = SvNVX (sv); 512 double nv = SvNVX (sv);
438 513
439 need (enc, 9); 514 need (enc, 9);
440 515
441 if (ecb_expect_false (nv == (U32)nv)) 516 if (ecb_expect_false (nv == (NV)(U32)nv))
442 encode_uint (enc, 0x00, (U32)nv); 517 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
443 //TODO: maybe I32? 518 //TODO: maybe I32?
444 else if (ecb_expect_false (nv == (float)nv)) 519 else if (ecb_expect_false (nv == (float)nv))
445 { 520 {
446 uint32_t fp = ecb_float_to_binary32 (nv); 521 uint32_t fp = ecb_float_to_binary32 (nv);
447 522
448 *enc->cur++ = 0xe0 | 26; 523 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
449 524
450 if (!ecb_big_endian ()) 525 if (!ecb_big_endian ())
451 fp = ecb_bswap32 (fp); 526 fp = ecb_bswap32 (fp);
452 527
453 memcpy (enc->cur, &fp, 4); 528 memcpy (enc->cur, &fp, 4);
455 } 530 }
456 else 531 else
457 { 532 {
458 uint64_t fp = ecb_double_to_binary64 (nv); 533 uint64_t fp = ecb_double_to_binary64 (nv);
459 534
460 *enc->cur++ = 0xe0 | 27; 535 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
461 536
462 if (!ecb_big_endian ()) 537 if (!ecb_big_endian ())
463 fp = ecb_bswap64 (fp); 538 fp = ecb_bswap64 (fp);
464 539
465 memcpy (enc->cur, &fp, 8); 540 memcpy (enc->cur, &fp, 8);
474 549
475 if (SvPOKp (sv)) 550 if (SvPOKp (sv))
476 { 551 {
477 STRLEN len; 552 STRLEN len;
478 char *str = SvPV (sv, len); 553 char *str = SvPV (sv, len);
479 encode_str (enc, SvUTF8 (sv), str, len); 554 encode_strref (enc, SvUTF8 (sv), str, len);
480 } 555 }
481 else if (SvNOKp (sv)) 556 else if (SvNOKp (sv))
482 encode_nv (enc, sv); 557 encode_nv (enc, sv);
483 else if (SvIOKp (sv)) 558 else if (SvIOKp (sv))
484 { 559 {
485 if (SvIsUV (sv)) 560 if (SvIsUV (sv))
486 encode_uint (enc, 0x00, SvUVX (sv)); 561 encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
487 else if (SvIVX (sv) >= 0) 562 else if (SvIVX (sv) >= 0)
488 encode_uint (enc, 0x00, SvIVX (sv)); 563 encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
489 else 564 else
490 encode_uint (enc, 0x20, -(SvIVX (sv) + 1)); 565 encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
491 } 566 }
492 else if (SvROK (sv)) 567 else if (SvROK (sv))
493 encode_rv (enc, SvRV (sv)); 568 encode_rv (enc, SvRV (sv));
494 else if (!SvOK (sv)) 569 else if (!SvOK (sv))
495 encode_ch (enc, 0xe0 | 22); 570 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
496 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 571 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
497 encode_ch (enc, 0xe0 | 23); 572 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
498 else 573 else
499 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data", 574 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
500 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 575 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
501} 576}
502 577
510 enc.cur = SvPVX (enc.sv); 585 enc.cur = SvPVX (enc.sv);
511 enc.end = SvEND (enc.sv); 586 enc.end = SvEND (enc.sv);
512 587
513 SvPOK_only (enc.sv); 588 SvPOK_only (enc.sv);
514 589
515 if (cbor->flags & F_ALLOW_STRINGREF) 590 if (cbor->flags & F_PACK_STRINGS)
516 { 591 {
517 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE); 592 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
518 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ()); 593 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
519 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ()); 594 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
520 } 595 }
542 CBOR cbor; 617 CBOR cbor;
543 U32 depth; // recursion depth 618 U32 depth; // recursion depth
544 U32 maxdepth; // recursion depth limit 619 U32 maxdepth; // recursion depth limit
545 AV *shareable; 620 AV *shareable;
546 AV *stringref; 621 AV *stringref;
622 SV *decode_tagged;
547} dec_t; 623} dec_t;
548 624
549#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE 625#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE
550 626
551#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data") 627#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 630#define DEC_DEC_DEPTH --dec->depth
555 631
556static UV 632static UV
557decode_uint (dec_t *dec) 633decode_uint (dec_t *dec)
558{ 634{
559 switch (*dec->cur & 31) 635 U8 m = *dec->cur & MINOR_MASK;
560 { 636 ++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;
565 637
566 case 24: 638 if (ecb_expect_true (m < LENGTH_EXT1))
639 return m;
640 else if (ecb_expect_true (m == LENGTH_EXT1))
641 {
567 WANT (2); 642 WANT (1);
568 dec->cur += 2; 643 dec->cur += 1;
569 return dec->cur[-1]; 644 return dec->cur[-1];
570 645 }
571 case 25: 646 else if (ecb_expect_true (m == LENGTH_EXT2))
647 {
572 WANT (3); 648 WANT (2);
573 dec->cur += 3; 649 dec->cur += 2;
574 return (((UV)dec->cur[-2]) << 8) 650 return (((UV)dec->cur[-2]) << 8)
575 | ((UV)dec->cur[-1]); 651 | ((UV)dec->cur[-1]);
576 652 }
577 case 26: 653 else if (ecb_expect_true (m == LENGTH_EXT4))
654 {
578 WANT (5); 655 WANT (4);
579 dec->cur += 5; 656 dec->cur += 4;
580 return (((UV)dec->cur[-4]) << 24) 657 return (((UV)dec->cur[-4]) << 24)
581 | (((UV)dec->cur[-3]) << 16) 658 | (((UV)dec->cur[-3]) << 16)
582 | (((UV)dec->cur[-2]) << 8) 659 | (((UV)dec->cur[-2]) << 8)
583 | ((UV)dec->cur[-1]); 660 | ((UV)dec->cur[-1]);
584 661 }
585 case 27: 662 else if (ecb_expect_true (m == LENGTH_EXT8))
663 {
586 WANT (9); 664 WANT (8);
587 dec->cur += 9; 665 dec->cur += 8;
666
667 return
668#if UVSIZE < 8
669 0
670#else
588 return (((UV)dec->cur[-8]) << 56) 671 (((UV)dec->cur[-8]) << 56)
589 | (((UV)dec->cur[-7]) << 48) 672 | (((UV)dec->cur[-7]) << 48)
590 | (((UV)dec->cur[-6]) << 40) 673 | (((UV)dec->cur[-6]) << 40)
591 | (((UV)dec->cur[-5]) << 32) 674 | (((UV)dec->cur[-5]) << 32)
675#endif
592 | (((UV)dec->cur[-4]) << 24) 676 | (((UV)dec->cur[-4]) << 24)
593 | (((UV)dec->cur[-3]) << 16) 677 | (((UV)dec->cur[-3]) << 16)
594 | (((UV)dec->cur[-2]) << 8) 678 | (((UV)dec->cur[-2]) << 8)
595 | ((UV)dec->cur[-1]); 679 | ((UV)dec->cur[-1]);
596 680 }
597 default: 681 else
598 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 682 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
599 }
600 683
601fail: 684fail:
602 return 0; 685 return 0;
603} 686}
604 687
609{ 692{
610 AV *av = newAV (); 693 AV *av = newAV ();
611 694
612 DEC_INC_DEPTH; 695 DEC_INC_DEPTH;
613 696
614 if ((*dec->cur & 31) == 31) 697 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
615 { 698 {
616 ++dec->cur; 699 ++dec->cur;
617 700
618 for (;;) 701 for (;;)
619 { 702 {
620 WANT (1); 703 WANT (1);
621 704
622 if (*dec->cur == (0xe0 | 31)) 705 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
623 { 706 {
624 ++dec->cur; 707 ++dec->cur;
625 break; 708 break;
626 } 709 }
627 710
630 } 713 }
631 else 714 else
632 { 715 {
633 int i, len = decode_uint (dec); 716 int i, len = decode_uint (dec);
634 717
718 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); 719 av_fill (av, len - 1);
636 720
637 for (i = 0; i < len; ++i) 721 for (i = 0; i < len; ++i)
638 AvARRAY (av)[i] = decode_sv (dec); 722 AvARRAY (av)[i] = decode_sv (dec);
639 } 723 }
652{ 736{
653 // for speed reasons, we specialcase single-string 737 // for speed reasons, we specialcase single-string
654 // byte or utf-8 strings as keys, but only when !stringref 738 // byte or utf-8 strings as keys, but only when !stringref
655 739
656 if (ecb_expect_true (!dec->stringref)) 740 if (ecb_expect_true (!dec->stringref))
657 if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27) 741 if (ecb_expect_true ((*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
658 { 742 {
659 I32 len = decode_uint (dec); 743 I32 len = decode_uint (dec);
660 char *key = (char *)dec->cur; 744 char *key = (char *)dec->cur;
661 745
662 dec->cur += len; 746 dec->cur += len;
663 747
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); 748 hv_store (hv, key, len, decode_sv (dec), 0);
668 749
669 return; 750 return;
670 } 751 }
671 else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27) 752 else if (ecb_expect_true ((*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
672 { 753 {
673 I32 len = decode_uint (dec); 754 I32 len = decode_uint (dec);
674 char *key = (char *)dec->cur; 755 char *key = (char *)dec->cur;
675 756
676 dec->cur += len; 757 dec->cur += len;
677 758
678 if (ecb_expect_false (dec->stringref)) 759 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
679 av_push (dec->stringref, newSVpvn_utf8 (key, len, 1)); 760 if (!is_utf8_string (key, len))
761 ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
680 762
681 hv_store (hv, key, -len, decode_sv (dec), 0); 763 hv_store (hv, key, -len, decode_sv (dec), 0);
682 764
683 return; 765 return;
684 } 766 }
686 SV *k = decode_sv (dec); 768 SV *k = decode_sv (dec);
687 SV *v = decode_sv (dec); 769 SV *v = decode_sv (dec);
688 770
689 hv_store_ent (hv, k, v, 0); 771 hv_store_ent (hv, k, v, 0);
690 SvREFCNT_dec (k); 772 SvREFCNT_dec (k);
773
774fail:
775 ;
691} 776}
692 777
693static SV * 778static SV *
694decode_hv (dec_t *dec) 779decode_hv (dec_t *dec)
695{ 780{
696 HV *hv = newHV (); 781 HV *hv = newHV ();
697 782
698 DEC_INC_DEPTH; 783 DEC_INC_DEPTH;
699 784
700 if ((*dec->cur & 31) == 31) 785 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
701 { 786 {
702 ++dec->cur; 787 ++dec->cur;
703 788
704 for (;;) 789 for (;;)
705 { 790 {
706 WANT (1); 791 WANT (1);
707 792
708 if (*dec->cur == (0xe0 | 31)) 793 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
709 { 794 {
710 ++dec->cur; 795 ++dec->cur;
711 break; 796 break;
712 } 797 }
713 798
734static SV * 819static SV *
735decode_str (dec_t *dec, int utf8) 820decode_str (dec_t *dec, int utf8)
736{ 821{
737 SV *sv = 0; 822 SV *sv = 0;
738 823
739 if ((*dec->cur & 31) == 31) 824 if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
740 { 825 {
826 // indefinite length strings
741 ++dec->cur; 827 ++dec->cur;
742 828
829 U8 major = *dec->cur & MAJOR_MISC;
830
743 sv = newSVpvn ("", 0); 831 sv = newSVpvn ("", 0);
744 832
745 // not very fast, and certainly not robust against illegal input
746 for (;;) 833 for (;;)
747 { 834 {
748 WANT (1); 835 WANT (1);
749 836
750 if (*dec->cur == (0xe0 | 31)) 837 if ((*dec->cur - major) > LENGTH_EXT8)
838 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
751 { 839 {
752 ++dec->cur; 840 ++dec->cur;
753 break; 841 break;
754 } 842 }
843 else
844 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
755 845
756 sv_catsv (sv, decode_sv (dec)); 846 STRLEN len = decode_uint (dec);
847
848 WANT (len);
849 sv_catpvn (sv, dec->cur, len);
850 dec->cur += len;
757 } 851 }
758 } 852 }
759 else 853 else
760 { 854 {
761 STRLEN len = decode_uint (dec); 855 STRLEN len = decode_uint (dec);
768 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1)) 862 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
769 av_push (dec->stringref, SvREFCNT_inc_NN (sv)); 863 av_push (dec->stringref, SvREFCNT_inc_NN (sv));
770 } 864 }
771 865
772 if (utf8) 866 if (utf8)
867 {
868 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
869 if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
870 ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
871
773 SvUTF8_on (sv); 872 SvUTF8_on (sv);
873 }
774 874
775 return sv; 875 return sv;
776 876
777fail: 877fail:
778 SvREFCNT_dec (sv); 878 SvREFCNT_dec (sv);
810 } 910 }
811 break; 911 break;
812 912
813 case CBOR_TAG_STRINGREF: 913 case CBOR_TAG_STRINGREF:
814 { 914 {
815 if ((*dec->cur >> 5) != 0) 915 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
816 ERR ("corrupted CBOR data (stringref index not an unsigned integer)"); 916 ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
817 917
818 UV idx = decode_uint (dec); 918 UV idx = decode_uint (dec);
819 919
820 if (!dec->stringref || (int)idx > AvFILLp (dec->stringref)) 920 if (!dec->stringref || (int)idx > AvFILLp (dec->stringref))
827 case CBOR_TAG_VALUE_SHAREABLE: 927 case CBOR_TAG_VALUE_SHAREABLE:
828 { 928 {
829 if (ecb_expect_false (!dec->shareable)) 929 if (ecb_expect_false (!dec->shareable))
830 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ()); 930 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
831 931
932 if (dec->cbor.flags & F_ALLOW_CYCLES)
933 {
832 sv = newSV (0); 934 sv = newSV (0);
833 av_push (dec->shareable, SvREFCNT_inc_NN (sv)); 935 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
834 936
835 SV *osv = decode_sv (dec); 937 SV *osv = decode_sv (dec);
836 sv_setsv (sv, osv); 938 sv_setsv (sv, osv);
837 SvREFCNT_dec_NN (osv); 939 SvREFCNT_dec_NN (osv);
940 }
941 else
942 {
943 av_push (dec->shareable, &PL_sv_undef);
944 int idx = AvFILLp (dec->shareable);
945 sv = decode_sv (dec);
946 av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
947 }
838 } 948 }
839 break; 949 break;
840 950
841 case CBOR_TAG_VALUE_SHAREDREF: 951 case CBOR_TAG_VALUE_SHAREDREF:
842 { 952 {
843 if ((*dec->cur >> 5) != 0) 953 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
844 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)"); 954 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
845 955
846 UV idx = decode_uint (dec); 956 UV idx = decode_uint (dec);
847 957
848 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable)) 958 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable))
849 ERR ("corrupted CBOR data (sharedref index out of bounds)"); 959 ERR ("corrupted CBOR data (sharedref index out of bounds)");
850 960
851 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]); 961 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
962
963 if (sv == &PL_sv_undef)
964 ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
852 } 965 }
853 break; 966 break;
854 967
855 case CBOR_TAG_PERL_OBJECT: 968 case CBOR_TAG_PERL_OBJECT:
856 { 969 {
905 1018
906 default: 1019 default:
907 { 1020 {
908 sv = decode_sv (dec); 1021 sv = decode_sv (dec);
909 1022
1023 dSP;
1024 ENTER; SAVETMPS; PUSHMARK (SP);
1025 EXTEND (SP, 2);
1026 PUSHs (newSVuv (tag));
1027 PUSHs (sv);
1028
1029 PUTBACK;
1030 int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL);
1031 SPAGAIN;
1032
1033 if (SvTRUE (ERRSV))
1034 {
1035 FREETMPS; LEAVE;
1036 ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));
1037 }
1038
1039 if (count)
1040 {
1041 SvREFCNT_dec (sv);
1042 sv = SvREFCNT_inc (POPs);
1043 }
1044 else
1045 {
910 AV *av = newAV (); 1046 AV *av = newAV ();
911 av_push (av, newSVuv (tag)); 1047 av_push (av, newSVuv (tag));
912 av_push (av, sv); 1048 av_push (av, sv);
913 1049
914 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 1050 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
915 ? cbor_tagged_stash 1051 ? cbor_tagged_stash
916 : gv_stashpv ("CBOR::XS::Tagged" , 1); 1052 : gv_stashpv ("CBOR::XS::Tagged" , 1);
917
918 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash); 1053 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
1054 }
1055
1056 PUTBACK;
1057
1058 FREETMPS; LEAVE;
919 } 1059 }
920 break; 1060 break;
921 } 1061 }
922 1062
923 return sv; 1063 return sv;
930static SV * 1070static SV *
931decode_sv (dec_t *dec) 1071decode_sv (dec_t *dec)
932{ 1072{
933 WANT (1); 1073 WANT (1);
934 1074
935 switch (*dec->cur >> 5) 1075 switch (*dec->cur >> MAJOR_SHIFT)
936 { 1076 {
937 case 0: // unsigned int 1077 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
938 return newSVuv (decode_uint (dec)); 1078 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
939 case 1: // negative int 1079 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
940 return newSViv (-1 - (IV)decode_uint (dec)); 1080 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
941 case 2: // octet string 1081 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
942 return decode_str (dec, 0); 1082 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
943 case 3: // utf-8 string 1083 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
944 return decode_str (dec, 1); 1084
945 case 4: // array 1085 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) 1086 switch (*dec->cur++ & MINOR_MASK)
953 { 1087 {
954 case 20: 1088 case SIMPLE_FALSE:
955#if CBOR_SLOW 1089#if CBOR_SLOW
956 types_false = get_bool ("Types::Serialiser::false"); 1090 types_false = get_bool ("Types::Serialiser::false");
957#endif 1091#endif
958 return newSVsv (types_false); 1092 return newSVsv (types_false);
959 case 21: 1093 case SIMPLE_TRUE:
960#if CBOR_SLOW 1094#if CBOR_SLOW
961 types_true = get_bool ("Types::Serialiser::true"); 1095 types_true = get_bool ("Types::Serialiser::true");
962#endif 1096#endif
963 return newSVsv (types_true); 1097 return newSVsv (types_true);
964 case 22: 1098 case SIMPLE_NULL:
965 return newSVsv (&PL_sv_undef); 1099 return newSVsv (&PL_sv_undef);
966 case 23: 1100 case SIMPLE_UNDEF:
967#if CBOR_SLOW 1101#if CBOR_SLOW
968 types_error = get_bool ("Types::Serialiser::error"); 1102 types_error = get_bool ("Types::Serialiser::error");
969#endif 1103#endif
970 return newSVsv (types_error); 1104 return newSVsv (types_error);
971 1105
972 case 25: 1106 case MISC_FLOAT16:
973 { 1107 {
974 WANT (2); 1108 WANT (2);
975 1109
976 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1110 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
977 dec->cur += 2; 1111 dec->cur += 2;
978 1112
979 return newSVnv (ecb_binary16_to_float (fp)); 1113 return newSVnv (ecb_binary16_to_float (fp));
980 } 1114 }
981 1115
982 case 26: 1116 case MISC_FLOAT32:
983 { 1117 {
984 uint32_t fp; 1118 uint32_t fp;
985 WANT (4); 1119 WANT (4);
986 memcpy (&fp, dec->cur, 4); 1120 memcpy (&fp, dec->cur, 4);
987 dec->cur += 4; 1121 dec->cur += 4;
990 fp = ecb_bswap32 (fp); 1124 fp = ecb_bswap32 (fp);
991 1125
992 return newSVnv (ecb_binary32_to_float (fp)); 1126 return newSVnv (ecb_binary32_to_float (fp));
993 } 1127 }
994 1128
995 case 27: 1129 case MISC_FLOAT64:
996 { 1130 {
997 uint64_t fp; 1131 uint64_t fp;
998 WANT (8); 1132 WANT (8);
999 memcpy (&fp, dec->cur, 8); 1133 memcpy (&fp, dec->cur, 8);
1000 dec->cur += 8; 1134 dec->cur += 8;
1003 fp = ecb_bswap64 (fp); 1137 fp = ecb_bswap64 (fp);
1004 1138
1005 return newSVnv (ecb_binary64_to_double (fp)); 1139 return newSVnv (ecb_binary64_to_double (fp));
1006 } 1140 }
1007 1141
1008 // 0..19 unassigned 1142 // 0..19 unassigned simple
1009 // 24 reserved + unassigned (reserved values are not encodable) 1143 // 24 reserved + unassigned simple (reserved values are not encodable)
1144 // 28-30 unassigned misc
1145 // 31 break code
1010 default: 1146 default:
1011 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1147 ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1012 } 1148 }
1013 1149
1014 break; 1150 break;
1015 } 1151 }
1016 1152
1043 if (dec.cur != dec.end && !dec.err) 1179 if (dec.cur != dec.end && !dec.err)
1044 dec.err = "garbage after CBOR object"; 1180 dec.err = "garbage after CBOR object";
1045 1181
1046 if (dec.err) 1182 if (dec.err)
1047 { 1183 {
1184 if (dec.shareable)
1185 {
1186 // need to break cyclic links, which whould all be in shareable
1187 int i;
1188 SV **svp;
1189
1190 for (i = av_len (dec.shareable) + 1; i--; )
1191 if ((svp = av_fetch (dec.shareable, i, 0)))
1192 sv_setsv (*svp, &PL_sv_undef);
1193 }
1194
1048 SvREFCNT_dec (sv); 1195 SvREFCNT_dec (sv);
1049 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur); 1196 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur);
1050 } 1197 }
1051 1198
1052 sv = sv_2mortal (sv); 1199 sv = sv_2mortal (sv);
1053 1200
1054 return sv; 1201 return sv;
1055} 1202}
1056 1203
1204/////////////////////////////////////////////////////////////////////////////
1205// incremental parser
1206
1207#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
1208
1209// returns 0 for notyet, 1 for success or error
1210static int
1211incr_parse (CBOR *self, SV *cborstr)
1212{
1213 STRLEN cur;
1214 SvPV (cborstr, cur);
1215
1216 while (ecb_expect_true (self->incr_need <= cur))
1217 {
1218 // table of integer count bytes
1219 static I8 incr_len[MINOR_MASK + 1] = {
1220 0, 0, 0, 0, 0, 0, 0, 0,
1221 0, 0, 0, 0, 0, 0, 0, 0,
1222 0, 0, 0, 0, 0, 0, 0, 0,
1223 1, 2, 4, 8,-1,-1,-1,-2
1224 };
1225
1226 const U8 *p = SvPVX (cborstr) + self->incr_pos;
1227 U8 m = *p & MINOR_MASK;
1228 IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1229 I8 ilen = incr_len[m];
1230
1231 self->incr_need = self->incr_pos + 1;
1232
1233 if (ecb_expect_false (ilen < 0))
1234 {
1235 if (m != MINOR_INDEF)
1236 return 1; // error
1237
1238 if (*p == (MAJOR_MISC | MINOR_INDEF))
1239 {
1240 if (count >= 0)
1241 return 1; // error
1242
1243 count = 1;
1244 }
1245 else
1246 {
1247 av_push (self->incr_count, newSViv (-1)); //TODO: nest
1248 count = -1;
1249 }
1250 }
1251 else
1252 {
1253 self->incr_need += ilen;
1254 if (ecb_expect_false (self->incr_need > cur))
1255 return 0;
1256
1257 int major = *p >> MAJOR_SHIFT;
1258
1259 switch (major)
1260 {
1261 case MAJOR_BYTES >> MAJOR_SHIFT:
1262 case MAJOR_TEXT >> MAJOR_SHIFT:
1263 case MAJOR_ARRAY >> MAJOR_SHIFT:
1264 case MAJOR_MAP >> MAJOR_SHIFT:
1265 {
1266 UV len;
1267
1268 if (ecb_expect_false (ilen))
1269 {
1270 len = 0;
1271
1272 do {
1273 len = (len << 8) | *++p;
1274 } while (--ilen);
1275 }
1276 else
1277 len = m;
1278
1279 switch (major)
1280 {
1281 case MAJOR_BYTES >> MAJOR_SHIFT:
1282 case MAJOR_TEXT >> MAJOR_SHIFT:
1283 self->incr_need += len;
1284 if (ecb_expect_false (self->incr_need > cur))
1285 return 0;
1286
1287 break;
1288
1289 case MAJOR_MAP >> MAJOR_SHIFT:
1290 len <<= 1;
1291 case MAJOR_ARRAY >> MAJOR_SHIFT:
1292 if (len)
1293 {
1294 av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
1295 count = len + 1;
1296 }
1297 break;
1298 }
1299 }
1300 }
1301 }
1302
1303 self->incr_pos = self->incr_need;
1304
1305 if (count > 0)
1306 {
1307 while (!--count)
1308 {
1309 if (!AvFILLp (self->incr_count))
1310 return 1; // done
1311
1312 SvREFCNT_dec_NN (av_pop (self->incr_count));
1313 count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1314 }
1315
1316 SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
1317 }
1318 }
1319
1320 return 0;
1321}
1322
1323
1057///////////////////////////////////////////////////////////////////////////// 1324/////////////////////////////////////////////////////////////////////////////
1058// XS interface functions 1325// XS interface functions
1059 1326
1060MODULE = CBOR::XS PACKAGE = CBOR::XS 1327MODULE = CBOR::XS PACKAGE = CBOR::XS
1061 1328
1068 types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1); 1335 types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1);
1069 1336
1070 types_true = get_bool ("Types::Serialiser::true" ); 1337 types_true = get_bool ("Types::Serialiser::true" );
1071 types_false = get_bool ("Types::Serialiser::false"); 1338 types_false = get_bool ("Types::Serialiser::false");
1072 types_error = get_bool ("Types::Serialiser::error"); 1339 types_error = get_bool ("Types::Serialiser::error");
1340
1341 default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1073 1342
1074 sv_cbor = newSVpv ("CBOR", 0); 1343 sv_cbor = newSVpv ("CBOR", 0);
1075 SvREADONLY_on (sv_cbor); 1344 SvREADONLY_on (sv_cbor);
1076} 1345}
1077 1346
1099void shrink (CBOR *self, int enable = 1) 1368void shrink (CBOR *self, int enable = 1)
1100 ALIAS: 1369 ALIAS:
1101 shrink = F_SHRINK 1370 shrink = F_SHRINK
1102 allow_unknown = F_ALLOW_UNKNOWN 1371 allow_unknown = F_ALLOW_UNKNOWN
1103 allow_sharing = F_ALLOW_SHARING 1372 allow_sharing = F_ALLOW_SHARING
1104 allow_stringref = F_ALLOW_STRINGREF 1373 allow_cycles = F_ALLOW_CYCLES
1374 pack_strings = F_PACK_STRINGS
1375 validate_utf8 = F_VALIDATE_UTF8
1105 PPCODE: 1376 PPCODE:
1106{ 1377{
1107 if (enable) 1378 if (enable)
1108 self->flags |= ix; 1379 self->flags |= ix;
1109 else 1380 else
1115void get_shrink (CBOR *self) 1386void get_shrink (CBOR *self)
1116 ALIAS: 1387 ALIAS:
1117 get_shrink = F_SHRINK 1388 get_shrink = F_SHRINK
1118 get_allow_unknown = F_ALLOW_UNKNOWN 1389 get_allow_unknown = F_ALLOW_UNKNOWN
1119 get_allow_sharing = F_ALLOW_SHARING 1390 get_allow_sharing = F_ALLOW_SHARING
1120 get_allow_stringref = F_ALLOW_STRINGREF 1391 get_allow_cycles = F_ALLOW_CYCLES
1392 get_pack_strings = F_PACK_STRINGS
1393 get_validate_utf8 = F_VALIDATE_UTF8
1121 PPCODE: 1394 PPCODE:
1122 XPUSHs (boolSV (self->flags & ix)); 1395 XPUSHs (boolSV (self->flags & ix));
1123 1396
1124void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1397void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1125 PPCODE: 1398 PPCODE:
1138 XPUSHs (ST (0)); 1411 XPUSHs (ST (0));
1139 1412
1140int get_max_size (CBOR *self) 1413int get_max_size (CBOR *self)
1141 CODE: 1414 CODE:
1142 RETVAL = self->max_size; 1415 RETVAL = self->max_size;
1416 OUTPUT:
1417 RETVAL
1418
1419void filter (CBOR *self, SV *filter = 0)
1420 PPCODE:
1421 SvREFCNT_dec (self->filter);
1422 self->filter = filter ? newSVsv (filter) : filter;
1423 XPUSHs (ST (0));
1424
1425SV *get_filter (CBOR *self)
1426 CODE:
1427 RETVAL = self->filter ? self->filter : NEWSV (0, 0);
1143 OUTPUT: 1428 OUTPUT:
1144 RETVAL 1429 RETVAL
1145 1430
1146void encode (CBOR *self, SV *scalar) 1431void encode (CBOR *self, SV *scalar)
1147 PPCODE: 1432 PPCODE:
1162 EXTEND (SP, 2); 1447 EXTEND (SP, 2);
1163 PUSHs (sv); 1448 PUSHs (sv);
1164 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1449 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1165} 1450}
1166 1451
1452void incr_parse (CBOR *self, SV *cborstr)
1453 PPCODE:
1454{
1455 if (SvUTF8 (cborstr))
1456 sv_utf8_downgrade (cborstr, 0);
1457
1458 if (!self->incr_count)
1459 {
1460 self->incr_count = newAV ();
1461 self->incr_pos = 0;
1462 self->incr_need = 1;
1463
1464 av_push (self->incr_count, newSViv (1));
1465 }
1466
1467 do
1468 {
1469 if (!incr_parse (self, cborstr))
1470 {
1471 if (self->incr_need > self->max_size && self->max_size)
1472 croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
1473 (unsigned long)self->incr_need, (unsigned long)self->max_size);
1474
1475 break;
1476 }
1477
1478 SV *sv;
1479 char *offset;
1480
1481 PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
1482 XPUSHs (sv);
1483
1484 sv_chop (cborstr, offset);
1485
1486 av_clear (self->incr_count);
1487 av_push (self->incr_count, newSViv (1));
1488
1489 self->incr_pos = 0;
1490 self->incr_need = self->incr_pos + 1;
1491 }
1492 while (GIMME_V == G_ARRAY);
1493}
1494
1495void incr_reset (CBOR *self)
1496 CODE:
1497{
1498 SvREFCNT_dec (self->incr_count);
1499 self->incr_count = 0;
1500}
1501
1502void DESTROY (CBOR *self)
1503 PPCODE:
1504 cbor_free (self);
1505
1167PROTOTYPES: ENABLE 1506PROTOTYPES: ENABLE
1168 1507
1169void encode_cbor (SV *scalar) 1508void encode_cbor (SV *scalar)
1509 ALIAS:
1510 encode_cbor = 0
1511 encode_cbor_sharing = F_ALLOW_SHARING
1170 PPCODE: 1512 PPCODE:
1171{ 1513{
1172 CBOR cbor; 1514 CBOR cbor;
1173 cbor_init (&cbor); 1515 cbor_init (&cbor);
1516 cbor.flags |= ix;
1174 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1517 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1175 XPUSHs (scalar); 1518 XPUSHs (scalar);
1176} 1519}
1177 1520
1178void decode_cbor (SV *cborstr) 1521void decode_cbor (SV *cborstr)

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines