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 1.19 by root, Wed Nov 20 02:03:09 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
102#define F_ALLOW_CYCLES 0x00000008UL
62#define F_DEDUP_STRINGS 0x00000008UL //TODO 103#define F_PACK_STRINGS 0x00000010UL
63#define F_DEDUP_KEYS 0x00000010UL //TODO 104#define F_VALIDATE_UTF8 0x00000020UL
64 105
65#define INIT_SIZE 32 // initial scalar size to be allocated 106#define INIT_SIZE 32 // initial scalar size to be allocated
66 107
67#define SB do { 108#define SB do {
68#define SE } while (0) 109#define SE } while (0)
80# define CBOR_SLOW 0 121# define CBOR_SLOW 0
81# define CBOR_STASH cbor_stash 122# define CBOR_STASH cbor_stash
82#endif 123#endif
83 124
84static 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::
85static SV *types_true, *types_false, *types_error, *sv_cbor; 126static SV *types_true, *types_false, *types_error, *sv_cbor, *default_filter;
86 127
87typedef struct { 128typedef struct {
88 U32 flags; 129 U32 flags;
89 U32 max_depth; 130 U32 max_depth;
90 STRLEN max_size; 131 STRLEN max_size;
132 SV *filter;
91} CBOR; 133} CBOR;
92 134
93ecb_inline void 135ecb_inline void
94cbor_init (CBOR *cbor) 136cbor_init (CBOR *cbor)
95{ 137{
96 Zero (cbor, 1, CBOR); 138 Zero (cbor, 1, CBOR);
97 cbor->max_depth = 512; 139 cbor->max_depth = 512;
140}
141
142ecb_inline void
143cbor_free (CBOR *cbor)
144{
145 SvREFCNT_dec (cbor->filter);
98} 146}
99 147
100///////////////////////////////////////////////////////////////////////////// 148/////////////////////////////////////////////////////////////////////////////
101// utility functions 149// utility functions
102 150
124 SvPV_renew (sv, SvCUR (sv) + 1); 172 SvPV_renew (sv, SvCUR (sv) + 1);
125#endif 173#endif
126 } 174 }
127} 175}
128 176
177// minimum length of a string to be registered for stringref
178ecb_inline int
179minimum_string_length (UV idx)
180{
181 return idx > 23
182 ? idx > 0xffU
183 ? idx > 0xffffU
184 ? idx > 0xffffffffU
185 ? 11
186 : 7
187 : 5
188 : 4
189 : 3;
190}
191
129///////////////////////////////////////////////////////////////////////////// 192/////////////////////////////////////////////////////////////////////////////
130// encoder 193// encoder
131 194
132// structure used for encoding CBOR 195// structure used for encoding CBOR
133typedef struct 196typedef struct
135 char *cur; // SvPVX (sv) + current output position 198 char *cur; // SvPVX (sv) + current output position
136 char *end; // SvEND (sv) 199 char *end; // SvEND (sv)
137 SV *sv; // result scalar 200 SV *sv; // result scalar
138 CBOR cbor; 201 CBOR cbor;
139 U32 depth; // recursion level 202 U32 depth; // recursion level
140 HV *stringref; // string => index, or 0 203 HV *stringref[2]; // string => index, or 0 ([0] = bytes, [1] = utf-8)
204 UV stringref_idx;
141 HV *shareable; // ptr => index, or 0 205 HV *shareable; // ptr => index, or 0
142 UV shareable_idx; 206 UV shareable_idx;
143} enc_t; 207} enc_t;
144 208
145ecb_inline void 209ecb_inline void
164static void 228static void
165encode_uint (enc_t *enc, int major, UV len) 229encode_uint (enc_t *enc, int major, UV len)
166{ 230{
167 need (enc, 9); 231 need (enc, 9);
168 232
169 if (len < 24) 233 if (ecb_expect_true (len < LENGTH_EXT1))
170 *enc->cur++ = major | len; 234 *enc->cur++ = major | len;
171 else if (len <= 0xff) 235 else if (ecb_expect_true (len <= 0xffU))
172 { 236 {
173 *enc->cur++ = major | 24; 237 *enc->cur++ = major | LENGTH_EXT1;
174 *enc->cur++ = len; 238 *enc->cur++ = len;
175 } 239 }
176 else if (len <= 0xffff) 240 else if (len <= 0xffffU)
177 { 241 {
178 *enc->cur++ = major | 25; 242 *enc->cur++ = major | LENGTH_EXT2;
179 *enc->cur++ = len >> 8; 243 *enc->cur++ = len >> 8;
180 *enc->cur++ = len; 244 *enc->cur++ = len;
181 } 245 }
182 else if (len <= 0xffffffff) 246 else if (len <= 0xffffffffU)
183 { 247 {
184 *enc->cur++ = major | 26; 248 *enc->cur++ = major | LENGTH_EXT4;
185 *enc->cur++ = len >> 24; 249 *enc->cur++ = len >> 24;
186 *enc->cur++ = len >> 16; 250 *enc->cur++ = len >> 16;
187 *enc->cur++ = len >> 8; 251 *enc->cur++ = len >> 8;
188 *enc->cur++ = len; 252 *enc->cur++ = len;
189 } 253 }
190 else 254 else
191 { 255 {
192 *enc->cur++ = major | 27; 256 *enc->cur++ = major | LENGTH_EXT8;
193 *enc->cur++ = len >> 56; 257 *enc->cur++ = len >> 56;
194 *enc->cur++ = len >> 48; 258 *enc->cur++ = len >> 48;
195 *enc->cur++ = len >> 40; 259 *enc->cur++ = len >> 40;
196 *enc->cur++ = len >> 32; 260 *enc->cur++ = len >> 32;
197 *enc->cur++ = len >> 24; 261 *enc->cur++ = len >> 24;
199 *enc->cur++ = len >> 8; 263 *enc->cur++ = len >> 8;
200 *enc->cur++ = len; 264 *enc->cur++ = len;
201 } 265 }
202} 266}
203 267
204static void 268ecb_inline void
269encode_tag (enc_t *enc, UV tag)
270{
271 encode_uint (enc, MAJOR_TAG, tag);
272}
273
274ecb_inline void
205encode_str (enc_t *enc, int utf8, char *str, STRLEN len) 275encode_str (enc_t *enc, int utf8, char *str, STRLEN len)
206{ 276{
207 encode_uint (enc, utf8 ? 0x60 : 0x40, len); 277 encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
208 need (enc, len); 278 need (enc, len);
209 memcpy (enc->cur, str, len); 279 memcpy (enc->cur, str, len);
210 enc->cur += len; 280 enc->cur += len;
211} 281}
212 282
213ecb_inline void 283static void
214encode_tag (enc_t *enc, UV tag) 284encode_strref (enc_t *enc, int utf8, char *str, STRLEN len)
215{ 285{
216 encode_uint (enc, 0xc0, tag); 286 if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
287 {
288 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
289
290 if (SvOK (*svp))
291 {
292 // already registered, use stringref
293 encode_tag (enc, CBOR_TAG_STRINGREF);
294 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
295 return;
296 }
297 else if (len >= minimum_string_length (enc->stringref_idx))
298 {
299 // register only
300 sv_setuv (*svp, enc->stringref_idx);
301 ++enc->stringref_idx;
302 }
303 }
304
305 encode_str (enc, utf8, str, len);
217} 306}
218 307
219static void encode_sv (enc_t *enc, SV *sv); 308static void encode_sv (enc_t *enc, SV *sv);
220 309
221static void 310static void
226 if (enc->depth >= enc->cbor.max_depth) 315 if (enc->depth >= enc->cbor.max_depth)
227 croak (ERR_NESTING_EXCEEDED); 316 croak (ERR_NESTING_EXCEEDED);
228 317
229 ++enc->depth; 318 ++enc->depth;
230 319
231 encode_uint (enc, 0x80, len + 1); 320 encode_uint (enc, MAJOR_ARRAY, len + 1);
232 321
233 for (i = 0; i <= len; ++i) 322 for (i = 0; i <= len; ++i)
234 { 323 {
235 SV **svp = av_fetch (av, i, 0); 324 SV **svp = av_fetch (av, i, 0);
236 encode_sv (enc, svp ? *svp : &PL_sv_undef); 325 encode_sv (enc, svp ? *svp : &PL_sv_undef);
251 340
252 int pairs = hv_iterinit (hv); 341 int pairs = hv_iterinit (hv);
253 int mg = SvMAGICAL (hv); 342 int mg = SvMAGICAL (hv);
254 343
255 if (mg) 344 if (mg)
256 encode_ch (enc, 0xa0 | 31); 345 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
257 else 346 else
258 encode_uint (enc, 0xa0, pairs); 347 encode_uint (enc, MAJOR_MAP, pairs);
259 348
260 while ((he = hv_iternext (hv))) 349 while ((he = hv_iternext (hv)))
261 { 350 {
262 if (HeKLEN (he) == HEf_SVKEY) 351 if (HeKLEN (he) == HEf_SVKEY)
263 encode_sv (enc, HeSVKEY (he)); 352 encode_sv (enc, HeSVKEY (he));
264 else 353 else
265 encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); 354 encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
266 355
267 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));
268 } 357 }
269 358
270 if (mg) 359 if (mg)
271 encode_ch (enc, 0xe0 | 31); 360 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
272 361
273 --enc->depth; 362 --enc->depth;
274} 363}
275 364
276// encode objects, arrays and special \0=false and \1=true values. 365// encode objects, arrays and special \0=false and \1=true values.
277static void 366static void
278encode_rv (enc_t *enc, SV *sv) 367encode_rv (enc_t *enc, SV *sv)
279{ 368{
280 SvGETMAGIC (sv); 369 SvGETMAGIC (sv);
281
282 if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)
283 && ecb_expect_false (SvREFCNT (sv) > 1))
284 {
285 if (!enc->shareable)
286 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
287
288 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
289
290 if (SvOK (*svp))
291 {
292 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
293 encode_uint (enc, 0x00, SvUV (*svp));
294 return;
295 }
296 else
297 {
298 sv_setuv (*svp, enc->shareable_idx);
299 ++enc->shareable_idx;
300 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
301 }
302 }
303 370
304 svtype svt = SvTYPE (sv); 371 svtype svt = SvTYPE (sv);
305 372
306 if (ecb_expect_false (SvOBJECT (sv))) 373 if (ecb_expect_false (SvOBJECT (sv)))
307 { 374 {
314 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 381 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
315 ? cbor_tagged_stash 382 ? cbor_tagged_stash
316 : gv_stashpv ("CBOR::XS::Tagged" , 1); 383 : gv_stashpv ("CBOR::XS::Tagged" , 1);
317 384
318 HV *stash = SvSTASH (sv); 385 HV *stash = SvSTASH (sv);
319 GV *method;
320 386
321 if (stash == boolean_stash) 387 if (stash == boolean_stash)
322 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 }
323 else if (stash == error_stash) 392 else if (stash == error_stash)
324 encode_ch (enc, 0xe0 | 23); 393 {
394 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
395 return;
396 }
325 else if (stash == tagged_stash) 397 else if (stash == tagged_stash)
326 { 398 {
327 if (svt != SVt_PVAV) 399 if (svt != SVt_PVAV)
328 croak ("encountered CBOR::XS::Tagged object that isn't an array"); 400 croak ("encountered CBOR::XS::Tagged object that isn't an array");
329 401
330 encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); 402 encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
331 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))
332 } 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
333 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) 436 if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
334 { 437 {
335 dSP; 438 dSP;
336 439
337 ENTER; SAVETMPS; PUSHMARK (SP); 440 ENTER; SAVETMPS; PUSHMARK (SP);
338 // 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
370 // catch this surprisingly common error 473 // catch this surprisingly common error
371 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) 474 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
372 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));
373 476
374 encode_tag (enc, CBOR_TAG_PERL_OBJECT); 477 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
375 encode_uint (enc, 0x80, count + 1); 478 encode_uint (enc, MAJOR_ARRAY, count + 1);
376 encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); 479 encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
377 480
378 while (count) 481 while (count)
379 encode_sv (enc, SP[1 - count--]); 482 encode_sv (enc, SP[1 - count--]);
380 483
381 PUTBACK; 484 PUTBACK;
402{ 505{
403 double nv = SvNVX (sv); 506 double nv = SvNVX (sv);
404 507
405 need (enc, 9); 508 need (enc, 9);
406 509
407 if (ecb_expect_false (nv == (U32)nv)) 510 if (ecb_expect_false (nv == (NV)(U32)nv))
408 encode_uint (enc, 0x00, (U32)nv); 511 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
409 //TODO: maybe I32? 512 //TODO: maybe I32?
410 else if (ecb_expect_false (nv == (float)nv)) 513 else if (ecb_expect_false (nv == (float)nv))
411 { 514 {
412 uint32_t fp = ecb_float_to_binary32 (nv); 515 uint32_t fp = ecb_float_to_binary32 (nv);
413 516
414 *enc->cur++ = 0xe0 | 26; 517 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
415 518
416 if (!ecb_big_endian ()) 519 if (!ecb_big_endian ())
417 fp = ecb_bswap32 (fp); 520 fp = ecb_bswap32 (fp);
418 521
419 memcpy (enc->cur, &fp, 4); 522 memcpy (enc->cur, &fp, 4);
421 } 524 }
422 else 525 else
423 { 526 {
424 uint64_t fp = ecb_double_to_binary64 (nv); 527 uint64_t fp = ecb_double_to_binary64 (nv);
425 528
426 *enc->cur++ = 0xe0 | 27; 529 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
427 530
428 if (!ecb_big_endian ()) 531 if (!ecb_big_endian ())
429 fp = ecb_bswap64 (fp); 532 fp = ecb_bswap64 (fp);
430 533
431 memcpy (enc->cur, &fp, 8); 534 memcpy (enc->cur, &fp, 8);
440 543
441 if (SvPOKp (sv)) 544 if (SvPOKp (sv))
442 { 545 {
443 STRLEN len; 546 STRLEN len;
444 char *str = SvPV (sv, len); 547 char *str = SvPV (sv, len);
445 encode_str (enc, SvUTF8 (sv), str, len); 548 encode_strref (enc, SvUTF8 (sv), str, len);
446 } 549 }
447 else if (SvNOKp (sv)) 550 else if (SvNOKp (sv))
448 encode_nv (enc, sv); 551 encode_nv (enc, sv);
449 else if (SvIOKp (sv)) 552 else if (SvIOKp (sv))
450 { 553 {
451 if (SvIsUV (sv)) 554 if (SvIsUV (sv))
452 encode_uint (enc, 0x00, SvUVX (sv)); 555 encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
453 else if (SvIVX (sv) >= 0) 556 else if (SvIVX (sv) >= 0)
454 encode_uint (enc, 0x00, SvIVX (sv)); 557 encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
455 else 558 else
456 encode_uint (enc, 0x20, -(SvIVX (sv) + 1)); 559 encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
457 } 560 }
458 else if (SvROK (sv)) 561 else if (SvROK (sv))
459 encode_rv (enc, SvRV (sv)); 562 encode_rv (enc, SvRV (sv));
460 else if (!SvOK (sv)) 563 else if (!SvOK (sv))
461 encode_ch (enc, 0xe0 | 22); 564 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
462 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 565 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
463 encode_ch (enc, 0xe0 | 23); 566 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
464 else 567 else
465 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",
466 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 569 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
467} 570}
468 571
475 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 578 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
476 enc.cur = SvPVX (enc.sv); 579 enc.cur = SvPVX (enc.sv);
477 enc.end = SvEND (enc.sv); 580 enc.end = SvEND (enc.sv);
478 581
479 SvPOK_only (enc.sv); 582 SvPOK_only (enc.sv);
583
584 if (cbor->flags & F_PACK_STRINGS)
585 {
586 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
587 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
588 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
589 }
590
480 encode_sv (&enc, scalar); 591 encode_sv (&enc, scalar);
481 592
482 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv)); 593 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv));
483 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings 594 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings
484 595
499 const char *err; // parse error, if != 0 610 const char *err; // parse error, if != 0
500 CBOR cbor; 611 CBOR cbor;
501 U32 depth; // recursion depth 612 U32 depth; // recursion depth
502 U32 maxdepth; // recursion depth limit 613 U32 maxdepth; // recursion depth limit
503 AV *shareable; 614 AV *shareable;
615 AV *stringref;
616 SV *decode_tagged;
504} dec_t; 617} dec_t;
505 618
506#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
507 620
508#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")
511#define DEC_DEC_DEPTH --dec->depth 624#define DEC_DEC_DEPTH --dec->depth
512 625
513static UV 626static UV
514decode_uint (dec_t *dec) 627decode_uint (dec_t *dec)
515{ 628{
516 switch (*dec->cur & 31) 629 U8 m = *dec->cur & MINOR_MASK;
517 { 630 ++dec->cur;
518 case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
519 case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
520 case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
521 return *dec->cur++ & 31;
522 631
523 case 24: 632 if (ecb_expect_true (m < LENGTH_EXT1))
633 return m;
634 else if (ecb_expect_true (m == LENGTH_EXT1))
635 {
524 WANT (2); 636 WANT (1);
525 dec->cur += 2; 637 dec->cur += 1;
526 return dec->cur[-1]; 638 return dec->cur[-1];
527 639 }
528 case 25: 640 else if (ecb_expect_true (m == LENGTH_EXT2))
641 {
529 WANT (3); 642 WANT (2);
530 dec->cur += 3; 643 dec->cur += 2;
531 return (((UV)dec->cur[-2]) << 8) 644 return (((UV)dec->cur[-2]) << 8)
532 | ((UV)dec->cur[-1]); 645 | ((UV)dec->cur[-1]);
533 646 }
534 case 26: 647 else if (ecb_expect_true (m == LENGTH_EXT4))
648 {
535 WANT (5); 649 WANT (4);
536 dec->cur += 5; 650 dec->cur += 4;
537 return (((UV)dec->cur[-4]) << 24) 651 return (((UV)dec->cur[-4]) << 24)
538 | (((UV)dec->cur[-3]) << 16) 652 | (((UV)dec->cur[-3]) << 16)
539 | (((UV)dec->cur[-2]) << 8) 653 | (((UV)dec->cur[-2]) << 8)
540 | ((UV)dec->cur[-1]); 654 | ((UV)dec->cur[-1]);
541 655 }
542 case 27: 656 else if (ecb_expect_true (m == LENGTH_EXT8))
657 {
543 WANT (9); 658 WANT (8);
544 dec->cur += 9; 659 dec->cur += 8;
660
661 return
662#if UVSIZE < 8
663 0
664#else
545 return (((UV)dec->cur[-8]) << 56) 665 (((UV)dec->cur[-8]) << 56)
546 | (((UV)dec->cur[-7]) << 48) 666 | (((UV)dec->cur[-7]) << 48)
547 | (((UV)dec->cur[-6]) << 40) 667 | (((UV)dec->cur[-6]) << 40)
548 | (((UV)dec->cur[-5]) << 32) 668 | (((UV)dec->cur[-5]) << 32)
669#endif
549 | (((UV)dec->cur[-4]) << 24) 670 | (((UV)dec->cur[-4]) << 24)
550 | (((UV)dec->cur[-3]) << 16) 671 | (((UV)dec->cur[-3]) << 16)
551 | (((UV)dec->cur[-2]) << 8) 672 | (((UV)dec->cur[-2]) << 8)
552 | ((UV)dec->cur[-1]); 673 | ((UV)dec->cur[-1]);
553 674 }
554 default: 675 else
555 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 676 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
556 }
557 677
558fail: 678fail:
559 return 0; 679 return 0;
560} 680}
561 681
566{ 686{
567 AV *av = newAV (); 687 AV *av = newAV ();
568 688
569 DEC_INC_DEPTH; 689 DEC_INC_DEPTH;
570 690
571 if ((*dec->cur & 31) == 31) 691 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
572 { 692 {
573 ++dec->cur; 693 ++dec->cur;
574 694
575 for (;;) 695 for (;;)
576 { 696 {
577 WANT (1); 697 WANT (1);
578 698
579 if (*dec->cur == (0xe0 | 31)) 699 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
580 { 700 {
581 ++dec->cur; 701 ++dec->cur;
582 break; 702 break;
583 } 703 }
584 704
587 } 707 }
588 else 708 else
589 { 709 {
590 int i, len = decode_uint (dec); 710 int i, len = decode_uint (dec);
591 711
712 WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
592 av_fill (av, len - 1); 713 av_fill (av, len - 1);
593 714
594 for (i = 0; i < len; ++i) 715 for (i = 0; i < len; ++i)
595 AvARRAY (av)[i] = decode_sv (dec); 716 AvARRAY (av)[i] = decode_sv (dec);
596 } 717 }
606 727
607static void 728static void
608decode_he (dec_t *dec, HV *hv) 729decode_he (dec_t *dec, HV *hv)
609{ 730{
610 // for speed reasons, we specialcase single-string 731 // for speed reasons, we specialcase single-string
611 // byte or utf-8 strings as keys. 732 // byte or utf-8 strings as keys, but only when !stringref
612 733
613 if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27) 734 if (ecb_expect_true (!dec->stringref))
735 if (ecb_expect_true ((*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
614 { 736 {
615 I32 len = decode_uint (dec); 737 I32 len = decode_uint (dec);
616 char *key = (char *)dec->cur; 738 char *key = (char *)dec->cur;
617 739
618 dec->cur += len; 740 dec->cur += len;
619 741
620 hv_store (hv, key, len, decode_sv (dec), 0); 742 hv_store (hv, key, len, decode_sv (dec), 0);
743
744 return;
621 } 745 }
622 else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27) 746 else if (ecb_expect_true ((*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
623 { 747 {
624 I32 len = decode_uint (dec); 748 I32 len = decode_uint (dec);
625 char *key = (char *)dec->cur; 749 char *key = (char *)dec->cur;
626 750
627 dec->cur += len; 751 dec->cur += len;
628 752
753 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
754 if (!is_utf8_string (key, len))
755 ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
756
629 hv_store (hv, key, -len, decode_sv (dec), 0); 757 hv_store (hv, key, -len, decode_sv (dec), 0);
758
759 return;
630 } 760 }
631 else 761
632 {
633 SV *k = decode_sv (dec); 762 SV *k = decode_sv (dec);
634 SV *v = decode_sv (dec); 763 SV *v = decode_sv (dec);
635 764
636 hv_store_ent (hv, k, v, 0); 765 hv_store_ent (hv, k, v, 0);
637 SvREFCNT_dec (k); 766 SvREFCNT_dec (k);
638 } 767
768fail:
769 ;
639} 770}
640 771
641static SV * 772static SV *
642decode_hv (dec_t *dec) 773decode_hv (dec_t *dec)
643{ 774{
644 HV *hv = newHV (); 775 HV *hv = newHV ();
645 776
646 DEC_INC_DEPTH; 777 DEC_INC_DEPTH;
647 778
648 if ((*dec->cur & 31) == 31) 779 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
649 { 780 {
650 ++dec->cur; 781 ++dec->cur;
651 782
652 for (;;) 783 for (;;)
653 { 784 {
654 WANT (1); 785 WANT (1);
655 786
656 if (*dec->cur == (0xe0 | 31)) 787 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
657 { 788 {
658 ++dec->cur; 789 ++dec->cur;
659 break; 790 break;
660 } 791 }
661 792
682static SV * 813static SV *
683decode_str (dec_t *dec, int utf8) 814decode_str (dec_t *dec, int utf8)
684{ 815{
685 SV *sv = 0; 816 SV *sv = 0;
686 817
687 if ((*dec->cur & 31) == 31) 818 if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
688 { 819 {
820 // indefinite length strings
689 ++dec->cur; 821 ++dec->cur;
690 822
823 U8 major = *dec->cur & MAJOR_MISC;
824
691 sv = newSVpvn ("", 0); 825 sv = newSVpvn ("", 0);
692 826
693 // not very fast, and certainly not robust against illegal input
694 for (;;) 827 for (;;)
695 { 828 {
696 WANT (1); 829 WANT (1);
697 830
698 if (*dec->cur == (0xe0 | 31)) 831 if ((*dec->cur - major) > LENGTH_EXT8)
832 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
699 { 833 {
700 ++dec->cur; 834 ++dec->cur;
701 break; 835 break;
702 } 836 }
837 else
838 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
703 839
704 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;
705 } 845 }
706 } 846 }
707 else 847 else
708 { 848 {
709 STRLEN len = decode_uint (dec); 849 STRLEN len = decode_uint (dec);
710 850
711 WANT (len); 851 WANT (len);
712 sv = newSVpvn (dec->cur, len); 852 sv = newSVpvn (dec->cur, len);
713 dec->cur += len; 853 dec->cur += len;
854
855 if (ecb_expect_false (dec->stringref)
856 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
857 av_push (dec->stringref, SvREFCNT_inc_NN (sv));
714 } 858 }
715 859
716 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
717 SvUTF8_on (sv); 866 SvUTF8_on (sv);
867 }
718 868
719 return sv; 869 return sv;
720 870
721fail: 871fail:
722 SvREFCNT_dec (sv); 872 SvREFCNT_dec (sv);
732 WANT (1); 882 WANT (1);
733 883
734 switch (tag) 884 switch (tag)
735 { 885 {
736 case CBOR_TAG_MAGIC: 886 case CBOR_TAG_MAGIC:
737 return decode_sv (dec); 887 sv = decode_sv (dec);
888 break;
738 889
739 case CBOR_TAG_INDIRECTION: 890 case CBOR_TAG_INDIRECTION:
740 return newRV_noinc (decode_sv (dec)); 891 sv = newRV_noinc (decode_sv (dec));
892 break;
893
894 case CBOR_TAG_STRINGREF_NAMESPACE:
895 {
896 ENTER; SAVETMPS;
897
898 SAVESPTR (dec->stringref);
899 dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());
900
901 sv = decode_sv (dec);
902
903 FREETMPS; LEAVE;
904 }
905 break;
906
907 case CBOR_TAG_STRINGREF:
908 {
909 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
910 ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
911
912 UV idx = decode_uint (dec);
913
914 if (!dec->stringref || (int)idx > AvFILLp (dec->stringref))
915 ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)");
916
917 sv = newSVsv (AvARRAY (dec->stringref)[idx]);
918 }
919 break;
741 920
742 case CBOR_TAG_VALUE_SHAREABLE: 921 case CBOR_TAG_VALUE_SHAREABLE:
743 { 922 {
744 if (ecb_expect_false (!dec->shareable)) 923 if (ecb_expect_false (!dec->shareable))
745 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ()); 924 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
746 925
926 if (dec->cbor.flags & F_ALLOW_CYCLES)
927 {
747 sv = newSV (0); 928 sv = newSV (0);
748 av_push (dec->shareable, SvREFCNT_inc_NN (sv)); 929 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
749 930
750 SV *osv = decode_sv (dec); 931 SV *osv = decode_sv (dec);
751 sv_setsv (sv, osv); 932 sv_setsv (sv, osv);
752 SvREFCNT_dec_NN (osv); 933 SvREFCNT_dec_NN (osv);
753 } 934 }
754 935 else
755 return sv; 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 }
942 }
943 break;
756 944
757 case CBOR_TAG_VALUE_SHAREDREF: 945 case CBOR_TAG_VALUE_SHAREDREF:
758 { 946 {
759 if ((*dec->cur >> 5) != 0) 947 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
760 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)"); 948 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
761 949
762 UV idx = decode_uint (dec); 950 UV idx = decode_uint (dec);
763 951
764 if (!dec->shareable || idx > AvFILLp (dec->shareable)) 952 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable))
765 ERR ("corrupted CBOR data (sharedref index out of bounds)"); 953 ERR ("corrupted CBOR data (sharedref index out of bounds)");
766 954
767 return 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");
768 } 959 }
960 break;
769 961
770 case CBOR_TAG_PERL_OBJECT: 962 case CBOR_TAG_PERL_OBJECT:
771 { 963 {
772 sv = decode_sv (dec); 964 sv = decode_sv (dec);
773 965
813 sv = SvREFCNT_inc (POPs); 1005 sv = SvREFCNT_inc (POPs);
814 1006
815 PUTBACK; 1007 PUTBACK;
816 1008
817 FREETMPS; LEAVE; 1009 FREETMPS; LEAVE;
818
819 return sv;
820 } 1010 }
1011 break;
821 1012
822 default: 1013 default:
823 { 1014 {
824 sv = decode_sv (dec); 1015 sv = decode_sv (dec);
825 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 {
826 AV *av = newAV (); 1040 AV *av = newAV ();
827 av_push (av, newSVuv (tag)); 1041 av_push (av, newSVuv (tag));
828 av_push (av, sv); 1042 av_push (av, sv);
829 1043
830 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 1044 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
831 ? cbor_tagged_stash 1045 ? cbor_tagged_stash
832 : gv_stashpv ("CBOR::XS::Tagged" , 1); 1046 : gv_stashpv ("CBOR::XS::Tagged" , 1);
833
834 return sv_bless (newRV_noinc ((SV *)av), tagged_stash); 1047 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
835 } 1048 }
1049
1050 PUTBACK;
1051
1052 FREETMPS; LEAVE;
1053 }
1054 break;
836 } 1055 }
1056
1057 return sv;
837 1058
838fail: 1059fail:
839 SvREFCNT_dec (sv); 1060 SvREFCNT_dec (sv);
840 return &PL_sv_undef; 1061 return &PL_sv_undef;
841} 1062}
843static SV * 1064static SV *
844decode_sv (dec_t *dec) 1065decode_sv (dec_t *dec)
845{ 1066{
846 WANT (1); 1067 WANT (1);
847 1068
848 switch (*dec->cur >> 5) 1069 switch (*dec->cur >> MAJOR_SHIFT)
849 { 1070 {
850 case 0: // unsigned int 1071 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
851 return newSVuv (decode_uint (dec)); 1072 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
852 case 1: // negative int 1073 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
853 return newSViv (-1 - (IV)decode_uint (dec)); 1074 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
854 case 2: // octet string 1075 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
855 return decode_str (dec, 0); 1076 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
856 case 3: // utf-8 string 1077 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
857 return decode_str (dec, 1); 1078
858 case 4: // array 1079 case MAJOR_MISC >> MAJOR_SHIFT:
859 return decode_av (dec);
860 case 5: // map
861 return decode_hv (dec);
862 case 6: // tag
863 return decode_tagged (dec);
864 case 7: // misc
865 switch (*dec->cur++ & 31) 1080 switch (*dec->cur++ & MINOR_MASK)
866 { 1081 {
867 case 20: 1082 case SIMPLE_FALSE:
868#if CBOR_SLOW 1083#if CBOR_SLOW
869 types_false = get_bool ("Types::Serialiser::false"); 1084 types_false = get_bool ("Types::Serialiser::false");
870#endif 1085#endif
871 return newSVsv (types_false); 1086 return newSVsv (types_false);
872 case 21: 1087 case SIMPLE_TRUE:
873#if CBOR_SLOW 1088#if CBOR_SLOW
874 types_true = get_bool ("Types::Serialiser::true"); 1089 types_true = get_bool ("Types::Serialiser::true");
875#endif 1090#endif
876 return newSVsv (types_true); 1091 return newSVsv (types_true);
877 case 22: 1092 case SIMPLE_NULL:
878 return newSVsv (&PL_sv_undef); 1093 return newSVsv (&PL_sv_undef);
879 case 23: 1094 case SIMPLE_UNDEF:
880#if CBOR_SLOW 1095#if CBOR_SLOW
881 types_error = get_bool ("Types::Serialiser::error"); 1096 types_error = get_bool ("Types::Serialiser::error");
882#endif 1097#endif
883 return newSVsv (types_error); 1098 return newSVsv (types_error);
884 1099
885 case 25: 1100 case MISC_FLOAT16:
886 { 1101 {
887 WANT (2); 1102 WANT (2);
888 1103
889 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1104 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
890 dec->cur += 2; 1105 dec->cur += 2;
891 1106
892 return newSVnv (ecb_binary16_to_float (fp)); 1107 return newSVnv (ecb_binary16_to_float (fp));
893 } 1108 }
894 1109
895 case 26: 1110 case MISC_FLOAT32:
896 { 1111 {
897 uint32_t fp; 1112 uint32_t fp;
898 WANT (4); 1113 WANT (4);
899 memcpy (&fp, dec->cur, 4); 1114 memcpy (&fp, dec->cur, 4);
900 dec->cur += 4; 1115 dec->cur += 4;
903 fp = ecb_bswap32 (fp); 1118 fp = ecb_bswap32 (fp);
904 1119
905 return newSVnv (ecb_binary32_to_float (fp)); 1120 return newSVnv (ecb_binary32_to_float (fp));
906 } 1121 }
907 1122
908 case 27: 1123 case MISC_FLOAT64:
909 { 1124 {
910 uint64_t fp; 1125 uint64_t fp;
911 WANT (8); 1126 WANT (8);
912 memcpy (&fp, dec->cur, 8); 1127 memcpy (&fp, dec->cur, 8);
913 dec->cur += 8; 1128 dec->cur += 8;
916 fp = ecb_bswap64 (fp); 1131 fp = ecb_bswap64 (fp);
917 1132
918 return newSVnv (ecb_binary64_to_double (fp)); 1133 return newSVnv (ecb_binary64_to_double (fp));
919 } 1134 }
920 1135
921 // 0..19 unassigned 1136 // 0..19 unassigned simple
922 // 24 reserved + unassigned (reserved values are not encodable) 1137 // 24 reserved + unassigned (reserved values are not encodable)
923 default: 1138 default:
924 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1139 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");
925 } 1140 }
926 1141
956 if (dec.cur != dec.end && !dec.err) 1171 if (dec.cur != dec.end && !dec.err)
957 dec.err = "garbage after CBOR object"; 1172 dec.err = "garbage after CBOR object";
958 1173
959 if (dec.err) 1174 if (dec.err)
960 { 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
961 SvREFCNT_dec (sv); 1187 SvREFCNT_dec (sv);
962 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);
963 } 1189 }
964 1190
965 sv = sv_2mortal (sv); 1191 sv = sv_2mortal (sv);
981 types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1); 1207 types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1);
982 1208
983 types_true = get_bool ("Types::Serialiser::true" ); 1209 types_true = get_bool ("Types::Serialiser::true" );
984 types_false = get_bool ("Types::Serialiser::false"); 1210 types_false = get_bool ("Types::Serialiser::false");
985 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);
986 1214
987 sv_cbor = newSVpv ("CBOR", 0); 1215 sv_cbor = newSVpv ("CBOR", 0);
988 SvREADONLY_on (sv_cbor); 1216 SvREADONLY_on (sv_cbor);
989} 1217}
990 1218
1012void shrink (CBOR *self, int enable = 1) 1240void shrink (CBOR *self, int enable = 1)
1013 ALIAS: 1241 ALIAS:
1014 shrink = F_SHRINK 1242 shrink = F_SHRINK
1015 allow_unknown = F_ALLOW_UNKNOWN 1243 allow_unknown = F_ALLOW_UNKNOWN
1016 allow_sharing = F_ALLOW_SHARING 1244 allow_sharing = F_ALLOW_SHARING
1017 dedup_keys = F_DEDUP_KEYS 1245 allow_cycles = F_ALLOW_CYCLES
1018 dedup_strings = F_DEDUP_STRINGS 1246 pack_strings = F_PACK_STRINGS
1247 validate_utf8 = F_VALIDATE_UTF8
1019 PPCODE: 1248 PPCODE:
1020{ 1249{
1021 if (enable) 1250 if (enable)
1022 self->flags |= ix; 1251 self->flags |= ix;
1023 else 1252 else
1029void get_shrink (CBOR *self) 1258void get_shrink (CBOR *self)
1030 ALIAS: 1259 ALIAS:
1031 get_shrink = F_SHRINK 1260 get_shrink = F_SHRINK
1032 get_allow_unknown = F_ALLOW_UNKNOWN 1261 get_allow_unknown = F_ALLOW_UNKNOWN
1033 get_allow_sharing = F_ALLOW_SHARING 1262 get_allow_sharing = F_ALLOW_SHARING
1034 get_dedup_keys = F_DEDUP_KEYS 1263 get_allow_cycles = F_ALLOW_CYCLES
1035 get_dedup_strings = F_DEDUP_STRINGS 1264 get_pack_strings = F_PACK_STRINGS
1265 get_validate_utf8 = F_VALIDATE_UTF8
1036 PPCODE: 1266 PPCODE:
1037 XPUSHs (boolSV (self->flags & ix)); 1267 XPUSHs (boolSV (self->flags & ix));
1038 1268
1039void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1269void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1040 PPCODE: 1270 PPCODE:
1053 XPUSHs (ST (0)); 1283 XPUSHs (ST (0));
1054 1284
1055int get_max_size (CBOR *self) 1285int get_max_size (CBOR *self)
1056 CODE: 1286 CODE:
1057 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);
1058 OUTPUT: 1300 OUTPUT:
1059 RETVAL 1301 RETVAL
1060 1302
1061void encode (CBOR *self, SV *scalar) 1303void encode (CBOR *self, SV *scalar)
1062 PPCODE: 1304 PPCODE:
1077 EXTEND (SP, 2); 1319 EXTEND (SP, 2);
1078 PUSHs (sv); 1320 PUSHs (sv);
1079 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1321 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1080} 1322}
1081 1323
1324void DESTROY (CBOR *self)
1325 PPCODE:
1326 cbor_free (self);
1327
1082PROTOTYPES: ENABLE 1328PROTOTYPES: ENABLE
1083 1329
1084void encode_cbor (SV *scalar) 1330void encode_cbor (SV *scalar)
1331 ALIAS:
1332 encode_cbor = 0
1333 encode_cbor_sharing = F_ALLOW_SHARING
1085 PPCODE: 1334 PPCODE:
1086{ 1335{
1087 CBOR cbor; 1336 CBOR cbor;
1088 cbor_init (&cbor); 1337 cbor_init (&cbor);
1338 cbor.flags |= ix;
1089 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1339 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1090 XPUSHs (scalar); 1340 XPUSHs (scalar);
1091} 1341}
1092 1342
1093void decode_cbor (SV *cborstr) 1343void decode_cbor (SV *cborstr)

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines