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 1.16 by root, Tue Oct 29 20:59:16 2013 UTC vs.
Revision 1.37 by root, Sat Nov 30 18:13:53 2013 UTC

19# define HvNAMELEN(hv) HvNAMELEN_get (hv) 19# define HvNAMELEN(hv) HvNAMELEN_get (hv)
20#endif 20#endif
21#ifndef HvNAMEUTF8 21#ifndef HvNAMEUTF8
22# define HvNAMEUTF8(hv) 0 22# define HvNAMEUTF8(hv) 0
23#endif 23#endif
24#ifndef SvREFCNT_dec_NN
25# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
26#endif
27
28// known major and minor types
29enum cbor_type
30{
31 MAJOR_SHIFT = 5,
32 MINOR_MASK = 0x1f,
33
34 MAJOR_POS_INT = 0 << MAJOR_SHIFT,
35 MAJOR_NEG_INT = 1 << MAJOR_SHIFT,
36 MAJOR_BYTES = 2 << MAJOR_SHIFT,
37 MAJOR_TEXT = 3 << MAJOR_SHIFT,
38 MAJOR_ARRAY = 4 << MAJOR_SHIFT,
39 MAJOR_MAP = 5 << MAJOR_SHIFT,
40 MAJOR_TAG = 6 << MAJOR_SHIFT,
41 MAJOR_MISC = 7 << MAJOR_SHIFT,
42
43 // INT/STRING/ARRAY/MAP subtypes
44 LENGTH_EXT1 = 24,
45 LENGTH_EXT2 = 25,
46 LENGTH_EXT4 = 26,
47 LENGTH_EXT8 = 27,
48
49 // SIMPLE types (effectively MISC subtypes)
50 SIMPLE_FALSE = 20,
51 SIMPLE_TRUE = 21,
52 SIMPLE_NULL = 22,
53 SIMPLE_UNDEF = 23,
54
55 // MISC subtype (unused)
56 MISC_EXT1 = 24,
57 MISC_FLOAT16 = 25,
58 MISC_FLOAT32 = 26,
59 MISC_FLOAT64 = 27,
60
61 // BYTES/TEXT/ARRAY/MAP
62 MINOR_INDEF = 31,
63};
24 64
25// known tags 65// known tags
26enum cbor_tag 66enum cbor_tag
27{ 67{
28 // inofficial extensions (pending iana registration) 68 // extensions
29 CBOR_TAG_PERL_OBJECT = 256, 69 CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
30 CBOR_TAG_GENERIC_OBJECT = 257, 70 CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
71 CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
72 CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
73 CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
74 CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref
75 CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
31 76
32 // rfc7049 77 // rfc7049
33 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8 78 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
34 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any 79 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
35 CBOR_TAG_POS_BIGNUM = 2, // byte string 80 CBOR_TAG_POS_BIGNUM = 2, // byte string
36 CBOR_TAG_NEG_BIGNUM = 3, // byte string 81 CBOR_TAG_NEG_BIGNUM = 3, // byte string
37 CBOR_TAG_DECIMAL = 4, // decimal fraction, array 82 CBOR_TAG_DECIMAL = 4, // decimal fraction, array
38 CBOR_TAG_BIGFLOAT = 5, // array 83 CBOR_TAG_BIGFLOAT = 5, // array
39 84
40 CBOR_TAG_CONV_B64U = 21, // base64url, any 85 CBOR_TAG_CONV_B64U = 21, // base64url, any
41 CBOR_TAG_CONV_B64 = 22, // base64, any 86 CBOR_TAG_CONV_B64 = 22, // base64, any
42 CBOR_TAG_CONV_HEX = 23, // base16, any 87 CBOR_TAG_CONV_HEX = 23, // base16, any
43 CBOR_TAG_CBOR = 24, // embedded cbor, byte string 88 CBOR_TAG_CBOR = 24, // embedded cbor, byte string
44 89
45 CBOR_TAG_URI = 32, // URI rfc3986, utf-8 90 CBOR_TAG_URI = 32, // URI rfc3986, utf-8
46 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8 91 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
47 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8 92 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
48 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8 93 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
49 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8 94 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
50 95
51 CBOR_TAG_MAGIC = 55799 // self-describe cbor 96 CBOR_TAG_MAGIC = 55799, // self-describe cbor
52}; 97};
53 98
54#define F_SHRINK 0x00000200UL 99#define F_SHRINK 0x00000001UL
55#define F_ALLOW_UNKNOWN 0x00002000UL 100#define F_ALLOW_UNKNOWN 0x00000002UL
101#define F_ALLOW_SHARING 0x00000004UL
102#define F_ALLOW_CYCLES 0x00000008UL
103#define F_PACK_STRINGS 0x00000010UL
56 104
57#define INIT_SIZE 32 // initial scalar size to be allocated 105#define INIT_SIZE 32 // initial scalar size to be allocated
58 106
59#define SB do { 107#define SB do {
60#define SE } while (0) 108#define SE } while (0)
72# define CBOR_SLOW 0 120# define CBOR_SLOW 0
73# define CBOR_STASH cbor_stash 121# define CBOR_STASH cbor_stash
74#endif 122#endif
75 123
76static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS:: 124static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS::
77static SV *types_true, *types_false, *types_error, *sv_cbor; 125static SV *types_true, *types_false, *types_error, *sv_cbor, *default_filter;
78 126
79typedef struct { 127typedef struct {
80 U32 flags; 128 U32 flags;
81 U32 max_depth; 129 U32 max_depth;
82 STRLEN max_size; 130 STRLEN max_size;
131 SV *filter;
83} CBOR; 132} CBOR;
84 133
85ecb_inline void 134ecb_inline void
86cbor_init (CBOR *cbor) 135cbor_init (CBOR *cbor)
87{ 136{
88 Zero (cbor, 1, CBOR); 137 Zero (cbor, 1, CBOR);
89 cbor->max_depth = 512; 138 cbor->max_depth = 512;
139}
140
141ecb_inline void
142cbor_free (CBOR *cbor)
143{
144 SvREFCNT_dec (cbor->filter);
90} 145}
91 146
92///////////////////////////////////////////////////////////////////////////// 147/////////////////////////////////////////////////////////////////////////////
93// utility functions 148// utility functions
94 149
116 SvPV_renew (sv, SvCUR (sv) + 1); 171 SvPV_renew (sv, SvCUR (sv) + 1);
117#endif 172#endif
118 } 173 }
119} 174}
120 175
121///////////////////////////////////////////////////////////////////////////// 176// minimum length of a string to be registered for stringref
122// fp hell 177ecb_inline int
123 178minimum_string_length (UV idx)
124//TODO 179{
180 return idx > 23
181 ? idx > 0xffU
182 ? idx > 0xffffU
183 ? idx > 0xffffffffU
184 ? 11
185 : 7
186 : 5
187 : 4
188 : 3;
189}
125 190
126///////////////////////////////////////////////////////////////////////////// 191/////////////////////////////////////////////////////////////////////////////
127// encoder 192// encoder
128 193
129// structure used for encoding CBOR 194// structure used for encoding CBOR
132 char *cur; // SvPVX (sv) + current output position 197 char *cur; // SvPVX (sv) + current output position
133 char *end; // SvEND (sv) 198 char *end; // SvEND (sv)
134 SV *sv; // result scalar 199 SV *sv; // result scalar
135 CBOR cbor; 200 CBOR cbor;
136 U32 depth; // recursion level 201 U32 depth; // recursion level
202 HV *stringref[2]; // string => index, or 0 ([0] = bytes, [1] = utf-8)
203 UV stringref_idx;
204 HV *shareable; // ptr => index, or 0
205 UV shareable_idx;
137} enc_t; 206} enc_t;
138 207
139ecb_inline void 208ecb_inline void
140need (enc_t *enc, STRLEN len) 209need (enc_t *enc, STRLEN len)
141{ 210{
158static void 227static void
159encode_uint (enc_t *enc, int major, UV len) 228encode_uint (enc_t *enc, int major, UV len)
160{ 229{
161 need (enc, 9); 230 need (enc, 9);
162 231
163 if (len < 24) 232 if (ecb_expect_true (len < LENGTH_EXT1))
164 *enc->cur++ = major | len; 233 *enc->cur++ = major | len;
165 else if (len <= 0xff) 234 else if (ecb_expect_true (len <= 0xffU))
166 { 235 {
167 *enc->cur++ = major | 24; 236 *enc->cur++ = major | LENGTH_EXT1;
168 *enc->cur++ = len; 237 *enc->cur++ = len;
169 } 238 }
170 else if (len <= 0xffff) 239 else if (len <= 0xffffU)
171 { 240 {
172 *enc->cur++ = major | 25; 241 *enc->cur++ = major | LENGTH_EXT2;
173 *enc->cur++ = len >> 8; 242 *enc->cur++ = len >> 8;
174 *enc->cur++ = len; 243 *enc->cur++ = len;
175 } 244 }
176 else if (len <= 0xffffffff) 245 else if (len <= 0xffffffffU)
177 { 246 {
178 *enc->cur++ = major | 26; 247 *enc->cur++ = major | LENGTH_EXT4;
179 *enc->cur++ = len >> 24; 248 *enc->cur++ = len >> 24;
180 *enc->cur++ = len >> 16; 249 *enc->cur++ = len >> 16;
181 *enc->cur++ = len >> 8; 250 *enc->cur++ = len >> 8;
182 *enc->cur++ = len; 251 *enc->cur++ = len;
183 } 252 }
184 else 253 else
185 { 254 {
186 *enc->cur++ = major | 27; 255 *enc->cur++ = major | LENGTH_EXT8;
187 *enc->cur++ = len >> 56; 256 *enc->cur++ = len >> 56;
188 *enc->cur++ = len >> 48; 257 *enc->cur++ = len >> 48;
189 *enc->cur++ = len >> 40; 258 *enc->cur++ = len >> 40;
190 *enc->cur++ = len >> 32; 259 *enc->cur++ = len >> 32;
191 *enc->cur++ = len >> 24; 260 *enc->cur++ = len >> 24;
193 *enc->cur++ = len >> 8; 262 *enc->cur++ = len >> 8;
194 *enc->cur++ = len; 263 *enc->cur++ = len;
195 } 264 }
196} 265}
197 266
198static void 267ecb_inline void
268encode_tag (enc_t *enc, UV tag)
269{
270 encode_uint (enc, MAJOR_TAG, tag);
271}
272
273ecb_inline void
199encode_str (enc_t *enc, int utf8, char *str, STRLEN len) 274encode_str (enc_t *enc, int utf8, char *str, STRLEN len)
200{ 275{
201 encode_uint (enc, utf8 ? 0x60 : 0x40, len); 276 encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
202 need (enc, len); 277 need (enc, len);
203 memcpy (enc->cur, str, len); 278 memcpy (enc->cur, str, len);
204 enc->cur += len; 279 enc->cur += len;
205} 280}
206 281
282static void
283encode_strref (enc_t *enc, int utf8, char *str, STRLEN len)
284{
285 if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
286 {
287 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
288
289 if (SvOK (*svp))
290 {
291 // already registered, use stringref
292 encode_tag (enc, CBOR_TAG_STRINGREF);
293 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
294 return;
295 }
296 else if (len >= minimum_string_length (enc->stringref_idx))
297 {
298 // register only
299 sv_setuv (*svp, enc->stringref_idx);
300 ++enc->stringref_idx;
301 }
302 }
303
304 encode_str (enc, utf8, str, len);
305}
306
207static void encode_sv (enc_t *enc, SV *sv); 307static void encode_sv (enc_t *enc, SV *sv);
208 308
209static void 309static void
210encode_av (enc_t *enc, AV *av) 310encode_av (enc_t *enc, AV *av)
211{ 311{
214 if (enc->depth >= enc->cbor.max_depth) 314 if (enc->depth >= enc->cbor.max_depth)
215 croak (ERR_NESTING_EXCEEDED); 315 croak (ERR_NESTING_EXCEEDED);
216 316
217 ++enc->depth; 317 ++enc->depth;
218 318
219 encode_uint (enc, 0x80, len + 1); 319 encode_uint (enc, MAJOR_ARRAY, len + 1);
220 320
221 for (i = 0; i <= len; ++i) 321 for (i = 0; i <= len; ++i)
222 { 322 {
223 SV **svp = av_fetch (av, i, 0); 323 SV **svp = av_fetch (av, i, 0);
224 encode_sv (enc, svp ? *svp : &PL_sv_undef); 324 encode_sv (enc, svp ? *svp : &PL_sv_undef);
239 339
240 int pairs = hv_iterinit (hv); 340 int pairs = hv_iterinit (hv);
241 int mg = SvMAGICAL (hv); 341 int mg = SvMAGICAL (hv);
242 342
243 if (mg) 343 if (mg)
244 encode_ch (enc, 0xa0 | 31); 344 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
245 else 345 else
246 encode_uint (enc, 0xa0, pairs); 346 encode_uint (enc, MAJOR_MAP, pairs);
247 347
248 while ((he = hv_iternext (hv))) 348 while ((he = hv_iternext (hv)))
249 { 349 {
250 if (HeKLEN (he) == HEf_SVKEY) 350 if (HeKLEN (he) == HEf_SVKEY)
251 encode_sv (enc, HeSVKEY (he)); 351 encode_sv (enc, HeSVKEY (he));
252 else 352 else
253 encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); 353 encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
254 354
255 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); 355 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
256 } 356 }
257 357
258 if (mg) 358 if (mg)
259 encode_ch (enc, 0xe0 | 31); 359 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
260 360
261 --enc->depth; 361 --enc->depth;
262} 362}
263 363
264// encode objects, arrays and special \0=false and \1=true values. 364// encode objects, arrays and special \0=false and \1=true values.
265static void 365static void
266encode_rv (enc_t *enc, SV *sv) 366encode_rv (enc_t *enc, SV *sv)
267{ 367{
268 svtype svt;
269
270 SvGETMAGIC (sv); 368 SvGETMAGIC (sv);
369
271 svt = SvTYPE (sv); 370 svtype svt = SvTYPE (sv);
272 371
273 if (ecb_expect_false (SvOBJECT (sv))) 372 if (ecb_expect_false (SvOBJECT (sv)))
274 { 373 {
275 HV *boolean_stash = !CBOR_SLOW || types_boolean_stash 374 HV *boolean_stash = !CBOR_SLOW || types_boolean_stash
276 ? types_boolean_stash 375 ? types_boolean_stash
281 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 380 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
282 ? cbor_tagged_stash 381 ? cbor_tagged_stash
283 : gv_stashpv ("CBOR::XS::Tagged" , 1); 382 : gv_stashpv ("CBOR::XS::Tagged" , 1);
284 383
285 HV *stash = SvSTASH (sv); 384 HV *stash = SvSTASH (sv);
286 GV *method;
287 385
288 if (stash == boolean_stash) 386 if (stash == boolean_stash)
289 encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20); 387 {
388 encode_ch (enc, SvIV (sv) ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE);
389 return;
390 }
290 else if (stash == error_stash) 391 else if (stash == error_stash)
291 encode_ch (enc, 0xe0 | 23); 392 {
393 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
394 return;
395 }
292 else if (stash == tagged_stash) 396 else if (stash == tagged_stash)
293 { 397 {
294 if (svt != SVt_PVAV) 398 if (svt != SVt_PVAV)
295 croak ("encountered CBOR::XS::Tagged object that isn't an array"); 399 croak ("encountered CBOR::XS::Tagged object that isn't an array");
296 400
297 encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); 401 encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
298 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); 402 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1));
403
404 return;
405 }
406 }
407
408 if (ecb_expect_false (SvREFCNT (sv) > 1)
409 && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
410 {
411 if (!enc->shareable)
412 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
413
414 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
415
416 if (SvOK (*svp))
299 } 417 {
418 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
419 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
420 return;
421 }
422 else
423 {
424 sv_setuv (*svp, enc->shareable_idx);
425 ++enc->shareable_idx;
426 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
427 }
428 }
429
430 if (ecb_expect_false (SvOBJECT (sv)))
431 {
432 HV *stash = SvSTASH (sv);
433 GV *method;
434
300 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) 435 if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
301 { 436 {
302 dSP; 437 dSP;
303 438
304 ENTER; SAVETMPS; PUSHMARK (SP); 439 ENTER; SAVETMPS; PUSHMARK (SP);
305 // we re-bless the reference to get overload and other niceties right 440 // we re-bless the reference to get overload and other niceties right
336 471
337 // catch this surprisingly common error 472 // catch this surprisingly common error
338 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) 473 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
339 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash)); 474 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
340 475
341 encode_uint (enc, 0xc0, CBOR_TAG_PERL_OBJECT); 476 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
342 encode_uint (enc, 0x80, count + 1); 477 encode_uint (enc, MAJOR_ARRAY, count + 1);
343 encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); 478 encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
344 479
345 while (count) 480 while (count)
346 encode_sv (enc, SP[1 - count--]); 481 encode_sv (enc, SP[1 - count--]);
347 482
348 PUTBACK; 483 PUTBACK;
355 } 490 }
356 else if (svt == SVt_PVHV) 491 else if (svt == SVt_PVHV)
357 encode_hv (enc, (HV *)sv); 492 encode_hv (enc, (HV *)sv);
358 else if (svt == SVt_PVAV) 493 else if (svt == SVt_PVAV)
359 encode_av (enc, (AV *)sv); 494 encode_av (enc, (AV *)sv);
360 else if (svt < SVt_PVAV)
361 {
362 STRLEN len = 0;
363 char *pv = svt ? SvPV (sv, len) : 0;
364
365 if (len == 1 && *pv == '1')
366 encode_ch (enc, 0xe0 | 21);
367 else if (len == 1 && *pv == '0')
368 encode_ch (enc, 0xe0 | 20);
369 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
370 encode_ch (enc, 0xe0 | 23);
371 else
372 croak ("cannot encode reference to scalar '%s' unless the scalar is 0 or 1",
373 SvPV_nolen (sv_2mortal (newRV_inc (sv))));
374 }
375 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
376 encode_ch (enc, 0xe0 | 23);
377 else 495 else
378 croak ("encountered %s, but CBOR can only represent references to arrays or hashes", 496 {
379 SvPV_nolen (sv_2mortal (newRV_inc (sv)))); 497 encode_tag (enc, CBOR_TAG_INDIRECTION);
498 encode_sv (enc, sv);
499 }
380} 500}
381 501
382static void 502static void
383encode_nv (enc_t *enc, SV *sv) 503encode_nv (enc_t *enc, SV *sv)
384{ 504{
385 double nv = SvNVX (sv); 505 double nv = SvNVX (sv);
386 506
387 need (enc, 9); 507 need (enc, 9);
388 508
389 if (ecb_expect_false (nv == (U32)nv)) 509 if (ecb_expect_false (nv == (NV)(U32)nv))
390 encode_uint (enc, 0x00, (U32)nv); 510 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
391 //TODO: maybe I32? 511 //TODO: maybe I32?
392 else if (ecb_expect_false (nv == (float)nv)) 512 else if (ecb_expect_false (nv == (float)nv))
393 { 513 {
394 uint32_t fp = ecb_float_to_binary32 (nv); 514 uint32_t fp = ecb_float_to_binary32 (nv);
395 515
396 *enc->cur++ = 0xe0 | 26; 516 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
397 517
398 if (!ecb_big_endian ()) 518 if (!ecb_big_endian ())
399 fp = ecb_bswap32 (fp); 519 fp = ecb_bswap32 (fp);
400 520
401 memcpy (enc->cur, &fp, 4); 521 memcpy (enc->cur, &fp, 4);
403 } 523 }
404 else 524 else
405 { 525 {
406 uint64_t fp = ecb_double_to_binary64 (nv); 526 uint64_t fp = ecb_double_to_binary64 (nv);
407 527
408 *enc->cur++ = 0xe0 | 27; 528 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
409 529
410 if (!ecb_big_endian ()) 530 if (!ecb_big_endian ())
411 fp = ecb_bswap64 (fp); 531 fp = ecb_bswap64 (fp);
412 532
413 memcpy (enc->cur, &fp, 8); 533 memcpy (enc->cur, &fp, 8);
422 542
423 if (SvPOKp (sv)) 543 if (SvPOKp (sv))
424 { 544 {
425 STRLEN len; 545 STRLEN len;
426 char *str = SvPV (sv, len); 546 char *str = SvPV (sv, len);
427 encode_str (enc, SvUTF8 (sv), str, len); 547 encode_strref (enc, SvUTF8 (sv), str, len);
428 } 548 }
429 else if (SvNOKp (sv)) 549 else if (SvNOKp (sv))
430 encode_nv (enc, sv); 550 encode_nv (enc, sv);
431 else if (SvIOKp (sv)) 551 else if (SvIOKp (sv))
432 { 552 {
433 if (SvIsUV (sv)) 553 if (SvIsUV (sv))
434 encode_uint (enc, 0x00, SvUVX (sv)); 554 encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
435 else if (SvIVX (sv) >= 0) 555 else if (SvIVX (sv) >= 0)
436 encode_uint (enc, 0x00, SvIVX (sv)); 556 encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
437 else 557 else
438 encode_uint (enc, 0x20, -(SvIVX (sv) + 1)); 558 encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
439 } 559 }
440 else if (SvROK (sv)) 560 else if (SvROK (sv))
441 encode_rv (enc, SvRV (sv)); 561 encode_rv (enc, SvRV (sv));
442 else if (!SvOK (sv)) 562 else if (!SvOK (sv))
443 encode_ch (enc, 0xe0 | 22); 563 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
444 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 564 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
445 encode_ch (enc, 0xe0 | 23); 565 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
446 else 566 else
447 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data", 567 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
448 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 568 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
449} 569}
450 570
451static SV * 571static SV *
452encode_cbor (SV *scalar, CBOR *cbor) 572encode_cbor (SV *scalar, CBOR *cbor)
453{ 573{
454 enc_t enc; 574 enc_t enc = { };
455 575
456 enc.cbor = *cbor; 576 enc.cbor = *cbor;
457 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 577 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
458 enc.cur = SvPVX (enc.sv); 578 enc.cur = SvPVX (enc.sv);
459 enc.end = SvEND (enc.sv); 579 enc.end = SvEND (enc.sv);
460 enc.depth = 0;
461 580
462 SvPOK_only (enc.sv); 581 SvPOK_only (enc.sv);
582
583 if (cbor->flags & F_PACK_STRINGS)
584 {
585 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
586 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
587 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
588 }
589
463 encode_sv (&enc, scalar); 590 encode_sv (&enc, scalar);
464 591
465 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv)); 592 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv));
466 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings 593 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings
467 594
481 U8 *end; // end of input string 608 U8 *end; // end of input string
482 const char *err; // parse error, if != 0 609 const char *err; // parse error, if != 0
483 CBOR cbor; 610 CBOR cbor;
484 U32 depth; // recursion depth 611 U32 depth; // recursion depth
485 U32 maxdepth; // recursion depth limit 612 U32 maxdepth; // recursion depth limit
613 AV *shareable;
614 AV *stringref;
615 SV *decode_tagged;
486} dec_t; 616} dec_t;
487 617
488#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE 618#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE
489 619
490#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data") 620#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data")
493#define DEC_DEC_DEPTH --dec->depth 623#define DEC_DEC_DEPTH --dec->depth
494 624
495static UV 625static UV
496decode_uint (dec_t *dec) 626decode_uint (dec_t *dec)
497{ 627{
498 switch (*dec->cur & 31) 628 U8 m = *dec->cur & MINOR_MASK;
499 { 629 ++dec->cur;
500 case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
501 case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
502 case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
503 return *dec->cur++ & 31;
504 630
505 case 24: 631 if (ecb_expect_true (m < LENGTH_EXT1))
632 return m;
633 else if (ecb_expect_true (m == LENGTH_EXT1))
634 {
506 WANT (2); 635 WANT (1);
507 dec->cur += 2; 636 dec->cur += 1;
508 return dec->cur[-1]; 637 return dec->cur[-1];
509 638 }
510 case 25: 639 else if (ecb_expect_true (m == LENGTH_EXT2))
640 {
511 WANT (3); 641 WANT (2);
512 dec->cur += 3; 642 dec->cur += 2;
513 return (((UV)dec->cur[-2]) << 8) 643 return (((UV)dec->cur[-2]) << 8)
514 | ((UV)dec->cur[-1]); 644 | ((UV)dec->cur[-1]);
515 645 }
516 case 26: 646 else if (ecb_expect_true (m == LENGTH_EXT4))
647 {
517 WANT (5); 648 WANT (4);
518 dec->cur += 5; 649 dec->cur += 4;
519 return (((UV)dec->cur[-4]) << 24) 650 return (((UV)dec->cur[-4]) << 24)
520 | (((UV)dec->cur[-3]) << 16) 651 | (((UV)dec->cur[-3]) << 16)
521 | (((UV)dec->cur[-2]) << 8) 652 | (((UV)dec->cur[-2]) << 8)
522 | ((UV)dec->cur[-1]); 653 | ((UV)dec->cur[-1]);
523 654 }
524 case 27: 655 else if (ecb_expect_true (m == LENGTH_EXT8))
656 {
525 WANT (9); 657 WANT (8);
526 dec->cur += 9; 658 dec->cur += 8;
659
660 return
661#if UVSIZE < 8
662 0
663#else
527 return (((UV)dec->cur[-8]) << 56) 664 (((UV)dec->cur[-8]) << 56)
528 | (((UV)dec->cur[-7]) << 48) 665 | (((UV)dec->cur[-7]) << 48)
529 | (((UV)dec->cur[-6]) << 40) 666 | (((UV)dec->cur[-6]) << 40)
530 | (((UV)dec->cur[-5]) << 32) 667 | (((UV)dec->cur[-5]) << 32)
668#endif
531 | (((UV)dec->cur[-4]) << 24) 669 | (((UV)dec->cur[-4]) << 24)
532 | (((UV)dec->cur[-3]) << 16) 670 | (((UV)dec->cur[-3]) << 16)
533 | (((UV)dec->cur[-2]) << 8) 671 | (((UV)dec->cur[-2]) << 8)
534 | ((UV)dec->cur[-1]); 672 | ((UV)dec->cur[-1]);
535 673 }
536 default: 674 else
537 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 675 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
538 }
539 676
540fail: 677fail:
541 return 0; 678 return 0;
542} 679}
543 680
548{ 685{
549 AV *av = newAV (); 686 AV *av = newAV ();
550 687
551 DEC_INC_DEPTH; 688 DEC_INC_DEPTH;
552 689
553 if ((*dec->cur & 31) == 31) 690 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
554 { 691 {
555 ++dec->cur; 692 ++dec->cur;
556 693
557 for (;;) 694 for (;;)
558 { 695 {
559 WANT (1); 696 WANT (1);
560 697
561 if (*dec->cur == (0xe0 | 31)) 698 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
562 { 699 {
563 ++dec->cur; 700 ++dec->cur;
564 break; 701 break;
565 } 702 }
566 703
569 } 706 }
570 else 707 else
571 { 708 {
572 int i, len = decode_uint (dec); 709 int i, len = decode_uint (dec);
573 710
711 WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
574 av_fill (av, len - 1); 712 av_fill (av, len - 1);
575 713
576 for (i = 0; i < len; ++i) 714 for (i = 0; i < len; ++i)
577 AvARRAY (av)[i] = decode_sv (dec); 715 AvARRAY (av)[i] = decode_sv (dec);
578 } 716 }
588 726
589static void 727static void
590decode_he (dec_t *dec, HV *hv) 728decode_he (dec_t *dec, HV *hv)
591{ 729{
592 // for speed reasons, we specialcase single-string 730 // for speed reasons, we specialcase single-string
593 // byte or utf-8 strings as keys. 731 // byte or utf-8 strings as keys, but only when !stringref
594 732
595 if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27) 733 if (ecb_expect_true (!dec->stringref))
734 if (ecb_expect_true ((*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
596 { 735 {
597 I32 len = decode_uint (dec); 736 I32 len = decode_uint (dec);
598 char *key = (char *)dec->cur; 737 char *key = (char *)dec->cur;
599 738
600 dec->cur += len; 739 dec->cur += len;
601 740
741 if (ecb_expect_false (dec->stringref))
742 av_push (dec->stringref, newSVpvn (key, len));
743
602 hv_store (hv, key, len, decode_sv (dec), 0); 744 hv_store (hv, key, len, decode_sv (dec), 0);
745
746 return;
603 } 747 }
604 else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27) 748 else if (ecb_expect_true ((*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
605 { 749 {
606 I32 len = decode_uint (dec); 750 I32 len = decode_uint (dec);
607 char *key = (char *)dec->cur; 751 char *key = (char *)dec->cur;
608 752
609 dec->cur += len; 753 dec->cur += len;
610 754
755 if (ecb_expect_false (dec->stringref))
756 av_push (dec->stringref, newSVpvn_utf8 (key, len, 1));
757
611 hv_store (hv, key, -len, decode_sv (dec), 0); 758 hv_store (hv, key, -len, decode_sv (dec), 0);
759
760 return;
612 } 761 }
613 else 762
614 {
615 SV *k = decode_sv (dec); 763 SV *k = decode_sv (dec);
616 SV *v = decode_sv (dec); 764 SV *v = decode_sv (dec);
617 765
618 hv_store_ent (hv, k, v, 0); 766 hv_store_ent (hv, k, v, 0);
619 SvREFCNT_dec (k); 767 SvREFCNT_dec (k);
620 }
621} 768}
622 769
623static SV * 770static SV *
624decode_hv (dec_t *dec) 771decode_hv (dec_t *dec)
625{ 772{
626 HV *hv = newHV (); 773 HV *hv = newHV ();
627 774
628 DEC_INC_DEPTH; 775 DEC_INC_DEPTH;
629 776
630 if ((*dec->cur & 31) == 31) 777 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
631 { 778 {
632 ++dec->cur; 779 ++dec->cur;
633 780
634 for (;;) 781 for (;;)
635 { 782 {
636 WANT (1); 783 WANT (1);
637 784
638 if (*dec->cur == (0xe0 | 31)) 785 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
639 { 786 {
640 ++dec->cur; 787 ++dec->cur;
641 break; 788 break;
642 } 789 }
643 790
664static SV * 811static SV *
665decode_str (dec_t *dec, int utf8) 812decode_str (dec_t *dec, int utf8)
666{ 813{
667 SV *sv = 0; 814 SV *sv = 0;
668 815
669 if ((*dec->cur & 31) == 31) 816 if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
670 { 817 {
818 // indefinite length strings
671 ++dec->cur; 819 ++dec->cur;
672 820
821 U8 major = *dec->cur & MAJOR_MISC;
822
673 sv = newSVpvn ("", 0); 823 sv = newSVpvn ("", 0);
674 824
675 // not very fast, and certainly not robust against illegal input
676 for (;;) 825 for (;;)
677 { 826 {
678 WANT (1); 827 WANT (1);
679 828
680 if (*dec->cur == (0xe0 | 31)) 829 if ((*dec->cur - major) > LENGTH_EXT8)
830 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
681 { 831 {
682 ++dec->cur; 832 ++dec->cur;
683 break; 833 break;
684 } 834 }
835 else
836 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
685 837
686 sv_catsv (sv, decode_sv (dec)); 838 STRLEN len = decode_uint (dec);
839
840 WANT (len);
841 sv_catpvn (sv, dec->cur, len);
842 dec->cur += len;
687 } 843 }
688 } 844 }
689 else 845 else
690 { 846 {
691 STRLEN len = decode_uint (dec); 847 STRLEN len = decode_uint (dec);
692 848
693 WANT (len); 849 WANT (len);
694 sv = newSVpvn (dec->cur, len); 850 sv = newSVpvn (dec->cur, len);
695 dec->cur += len; 851 dec->cur += len;
852
853 if (ecb_expect_false (dec->stringref)
854 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
855 av_push (dec->stringref, SvREFCNT_inc_NN (sv));
696 } 856 }
697 857
698 if (utf8) 858 if (utf8)
699 SvUTF8_on (sv); 859 SvUTF8_on (sv);
700 860
706} 866}
707 867
708static SV * 868static SV *
709decode_tagged (dec_t *dec) 869decode_tagged (dec_t *dec)
710{ 870{
871 SV *sv = 0;
711 UV tag = decode_uint (dec); 872 UV tag = decode_uint (dec);
873
874 WANT (1);
875
876 switch (tag)
877 {
878 case CBOR_TAG_MAGIC:
712 SV *sv = decode_sv (dec); 879 sv = decode_sv (dec);
880 break;
713 881
714 if (tag == CBOR_TAG_MAGIC) 882 case CBOR_TAG_INDIRECTION:
715 return sv; 883 sv = newRV_noinc (decode_sv (dec));
716 else if (tag == CBOR_TAG_PERL_OBJECT) 884 break;
717 { 885
886 case CBOR_TAG_STRINGREF_NAMESPACE:
887 {
888 ENTER; SAVETMPS;
889
890 SAVESPTR (dec->stringref);
891 dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());
892
893 sv = decode_sv (dec);
894
895 FREETMPS; LEAVE;
896 }
897 break;
898
899 case CBOR_TAG_STRINGREF:
900 {
901 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
902 ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
903
904 UV idx = decode_uint (dec);
905
906 if (!dec->stringref || (int)idx > AvFILLp (dec->stringref))
907 ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)");
908
909 sv = newSVsv (AvARRAY (dec->stringref)[idx]);
910 }
911 break;
912
913 case CBOR_TAG_VALUE_SHAREABLE:
914 {
915 if (ecb_expect_false (!dec->shareable))
916 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
917
918 if (dec->cbor.flags & F_ALLOW_CYCLES)
919 {
920 sv = newSV (0);
921 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
922
923 SV *osv = decode_sv (dec);
924 sv_setsv (sv, osv);
925 SvREFCNT_dec_NN (osv);
926 }
927 else
928 {
929 av_push (dec->shareable, &PL_sv_undef);
930 int idx = AvFILLp (dec->shareable);
931 sv = decode_sv (dec);
932 av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
933 }
934 }
935 break;
936
937 case CBOR_TAG_VALUE_SHAREDREF:
938 {
939 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
940 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
941
942 UV idx = decode_uint (dec);
943
944 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable))
945 ERR ("corrupted CBOR data (sharedref index out of bounds)");
946
947 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
948
949 if (sv == &PL_sv_undef)
950 ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
951 }
952 break;
953
954 case CBOR_TAG_PERL_OBJECT:
955 {
956 sv = decode_sv (dec);
957
718 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) 958 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
719 ERR ("corrupted CBOR data (non-array perl object)"); 959 ERR ("corrupted CBOR data (non-array perl object)");
720 960
721 AV *av = (AV *)SvRV (sv); 961 AV *av = (AV *)SvRV (sv);
722 int len = av_len (av) + 1; 962 int len = av_len (av) + 1;
723 HV *stash = gv_stashsv (*av_fetch (av, 0, 1), 0); 963 HV *stash = gv_stashsv (*av_fetch (av, 0, 1), 0);
724 964
725 if (!stash) 965 if (!stash)
726 ERR ("cannot decode perl-object (package does not exist)"); 966 ERR ("cannot decode perl-object (package does not exist)");
727 967
728 GV *method = gv_fetchmethod_autoload (stash, "THAW", 0); 968 GV *method = gv_fetchmethod_autoload (stash, "THAW", 0);
729 969
730 if (!method) 970 if (!method)
731 ERR ("cannot decode perl-object (package does not have a THAW method)"); 971 ERR ("cannot decode perl-object (package does not have a THAW method)");
732 972
733 dSP; 973 dSP;
734 974
735 ENTER; SAVETMPS; PUSHMARK (SP); 975 ENTER; SAVETMPS; PUSHMARK (SP);
736 EXTEND (SP, len + 1); 976 EXTEND (SP, len + 1);
737 // we re-bless the reference to get overload and other niceties right 977 // we re-bless the reference to get overload and other niceties right
738 PUSHs (*av_fetch (av, 0, 1)); 978 PUSHs (*av_fetch (av, 0, 1));
739 PUSHs (sv_cbor); 979 PUSHs (sv_cbor);
740 980
741 int i; 981 int i;
742 982
743 for (i = 1; i < len; ++i) 983 for (i = 1; i < len; ++i)
744 PUSHs (*av_fetch (av, i, 1)); 984 PUSHs (*av_fetch (av, i, 1));
745 985
746 PUTBACK; 986 PUTBACK;
747 call_sv ((SV *)GvCV (method), G_SCALAR); 987 call_sv ((SV *)GvCV (method), G_SCALAR | G_EVAL);
748 SPAGAIN; 988 SPAGAIN;
749 989
990 if (SvTRUE (ERRSV))
991 {
992 FREETMPS; LEAVE;
993 ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));
994 }
995
750 SvREFCNT_dec (sv); 996 SvREFCNT_dec (sv);
751 sv = SvREFCNT_inc (POPs); 997 sv = SvREFCNT_inc (POPs);
752 998
753 PUTBACK; 999 PUTBACK;
754 1000
755 FREETMPS; LEAVE; 1001 FREETMPS; LEAVE;
1002 }
1003 break;
756 1004
757 return sv; 1005 default:
758 } 1006 {
759 else 1007 sv = decode_sv (dec);
760 { 1008
1009 dSP;
1010 ENTER; SAVETMPS; PUSHMARK (SP);
1011 EXTEND (SP, 2);
1012 PUSHs (newSVuv (tag));
1013 PUSHs (sv);
1014
1015 PUTBACK;
1016 int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL);
1017 SPAGAIN;
1018
1019 if (SvTRUE (ERRSV))
1020 {
1021 FREETMPS; LEAVE;
1022 ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));
1023 }
1024
1025 if (count)
1026 {
1027 SvREFCNT_dec (sv);
1028 sv = SvREFCNT_inc (POPs);
1029 }
1030 else
1031 {
761 AV *av = newAV (); 1032 AV *av = newAV ();
762 av_push (av, newSVuv (tag)); 1033 av_push (av, newSVuv (tag));
763 av_push (av, sv); 1034 av_push (av, sv);
764 1035
765 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 1036 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
766 ? cbor_tagged_stash 1037 ? cbor_tagged_stash
767 : gv_stashpv ("CBOR::XS::Tagged" , 1); 1038 : gv_stashpv ("CBOR::XS::Tagged" , 1);
768
769 return sv_bless (newRV_noinc ((SV *)av), tagged_stash); 1039 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
1040 }
1041
1042 PUTBACK;
1043
1044 FREETMPS; LEAVE;
1045 }
1046 break;
770 } 1047 }
1048
1049 return sv;
771 1050
772fail: 1051fail:
773 SvREFCNT_dec (sv); 1052 SvREFCNT_dec (sv);
774 return &PL_sv_undef; 1053 return &PL_sv_undef;
775} 1054}
777static SV * 1056static SV *
778decode_sv (dec_t *dec) 1057decode_sv (dec_t *dec)
779{ 1058{
780 WANT (1); 1059 WANT (1);
781 1060
782 switch (*dec->cur >> 5) 1061 switch (*dec->cur >> MAJOR_SHIFT)
783 { 1062 {
784 case 0: // unsigned int 1063 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
785 return newSVuv (decode_uint (dec)); 1064 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
786 case 1: // negative int 1065 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
787 return newSViv (-1 - (IV)decode_uint (dec)); 1066 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
788 case 2: // octet string 1067 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
789 return decode_str (dec, 0); 1068 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
790 case 3: // utf-8 string 1069 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
791 return decode_str (dec, 1); 1070
792 case 4: // array 1071 case MAJOR_MISC >> MAJOR_SHIFT:
793 return decode_av (dec);
794 case 5: // map
795 return decode_hv (dec);
796 case 6: // tag
797 return decode_tagged (dec);
798 case 7: // misc
799 switch (*dec->cur++ & 31) 1072 switch (*dec->cur++ & MINOR_MASK)
800 { 1073 {
801 case 20: 1074 case SIMPLE_FALSE:
802#if CBOR_SLOW 1075#if CBOR_SLOW
803 types_false = get_bool ("Types::Serialiser::false"); 1076 types_false = get_bool ("Types::Serialiser::false");
804#endif 1077#endif
805 return newSVsv (types_false); 1078 return newSVsv (types_false);
806 case 21: 1079 case SIMPLE_TRUE:
807#if CBOR_SLOW 1080#if CBOR_SLOW
808 types_true = get_bool ("Types::Serialiser::true"); 1081 types_true = get_bool ("Types::Serialiser::true");
809#endif 1082#endif
810 return newSVsv (types_true); 1083 return newSVsv (types_true);
811 case 22: 1084 case SIMPLE_NULL:
812 return newSVsv (&PL_sv_undef); 1085 return newSVsv (&PL_sv_undef);
813 case 23: 1086 case SIMPLE_UNDEF:
814#if CBOR_SLOW 1087#if CBOR_SLOW
815 types_error = get_bool ("Types::Serialiser::error"); 1088 types_error = get_bool ("Types::Serialiser::error");
816#endif 1089#endif
817 return newSVsv (types_error); 1090 return newSVsv (types_error);
818 1091
819 case 25: 1092 case MISC_FLOAT16:
820 { 1093 {
821 WANT (2); 1094 WANT (2);
822 1095
823 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1096 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
824 dec->cur += 2; 1097 dec->cur += 2;
825 1098
826 return newSVnv (ecb_binary16_to_float (fp)); 1099 return newSVnv (ecb_binary16_to_float (fp));
827 } 1100 }
828 1101
829 case 26: 1102 case MISC_FLOAT32:
830 { 1103 {
831 uint32_t fp; 1104 uint32_t fp;
832 WANT (4); 1105 WANT (4);
833 memcpy (&fp, dec->cur, 4); 1106 memcpy (&fp, dec->cur, 4);
834 dec->cur += 4; 1107 dec->cur += 4;
837 fp = ecb_bswap32 (fp); 1110 fp = ecb_bswap32 (fp);
838 1111
839 return newSVnv (ecb_binary32_to_float (fp)); 1112 return newSVnv (ecb_binary32_to_float (fp));
840 } 1113 }
841 1114
842 case 27: 1115 case MISC_FLOAT64:
843 { 1116 {
844 uint64_t fp; 1117 uint64_t fp;
845 WANT (8); 1118 WANT (8);
846 memcpy (&fp, dec->cur, 8); 1119 memcpy (&fp, dec->cur, 8);
847 dec->cur += 8; 1120 dec->cur += 8;
850 fp = ecb_bswap64 (fp); 1123 fp = ecb_bswap64 (fp);
851 1124
852 return newSVnv (ecb_binary64_to_double (fp)); 1125 return newSVnv (ecb_binary64_to_double (fp));
853 } 1126 }
854 1127
855 // 0..19 unassigned 1128 // 0..19 unassigned simple
856 // 24 reserved + unassigned (reserved values are not encodable) 1129 // 24 reserved + unassigned (reserved values are not encodable)
857 default: 1130 default:
858 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1131 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");
859 } 1132 }
860 1133
866} 1139}
867 1140
868static SV * 1141static SV *
869decode_cbor (SV *string, CBOR *cbor, char **offset_return) 1142decode_cbor (SV *string, CBOR *cbor, char **offset_return)
870{ 1143{
871 dec_t dec; 1144 dec_t dec = { };
872 SV *sv; 1145 SV *sv;
873 STRLEN len; 1146 STRLEN len;
874 char *data = SvPVbyte (string, len); 1147 char *data = SvPVbyte (string, len);
875 1148
876 if (len > cbor->max_size && cbor->max_size) 1149 if (len > cbor->max_size && cbor->max_size)
878 (unsigned long)len, (unsigned long)cbor->max_size); 1151 (unsigned long)len, (unsigned long)cbor->max_size);
879 1152
880 dec.cbor = *cbor; 1153 dec.cbor = *cbor;
881 dec.cur = (U8 *)data; 1154 dec.cur = (U8 *)data;
882 dec.end = (U8 *)data + len; 1155 dec.end = (U8 *)data + len;
883 dec.err = 0;
884 dec.depth = 0;
885 1156
886 sv = decode_sv (&dec); 1157 sv = decode_sv (&dec);
887 1158
888 if (offset_return) 1159 if (offset_return)
889 *offset_return = dec.cur; 1160 *offset_return = dec.cur;
917 types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1); 1188 types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1);
918 1189
919 types_true = get_bool ("Types::Serialiser::true" ); 1190 types_true = get_bool ("Types::Serialiser::true" );
920 types_false = get_bool ("Types::Serialiser::false"); 1191 types_false = get_bool ("Types::Serialiser::false");
921 types_error = get_bool ("Types::Serialiser::error"); 1192 types_error = get_bool ("Types::Serialiser::error");
1193
1194 default_filter = newSVpv ("CBOR::XS::default_filter", 0);
922 1195
923 sv_cbor = newSVpv ("CBOR", 0); 1196 sv_cbor = newSVpv ("CBOR", 0);
924 SvREADONLY_on (sv_cbor); 1197 SvREADONLY_on (sv_cbor);
925} 1198}
926 1199
947 1220
948void shrink (CBOR *self, int enable = 1) 1221void shrink (CBOR *self, int enable = 1)
949 ALIAS: 1222 ALIAS:
950 shrink = F_SHRINK 1223 shrink = F_SHRINK
951 allow_unknown = F_ALLOW_UNKNOWN 1224 allow_unknown = F_ALLOW_UNKNOWN
1225 allow_sharing = F_ALLOW_SHARING
1226 allow_cycles = F_ALLOW_CYCLES
1227 pack_strings = F_PACK_STRINGS
952 PPCODE: 1228 PPCODE:
953{ 1229{
954 if (enable) 1230 if (enable)
955 self->flags |= ix; 1231 self->flags |= ix;
956 else 1232 else
961 1237
962void get_shrink (CBOR *self) 1238void get_shrink (CBOR *self)
963 ALIAS: 1239 ALIAS:
964 get_shrink = F_SHRINK 1240 get_shrink = F_SHRINK
965 get_allow_unknown = F_ALLOW_UNKNOWN 1241 get_allow_unknown = F_ALLOW_UNKNOWN
1242 get_allow_sharing = F_ALLOW_SHARING
1243 get_allow_cycles = F_ALLOW_CYCLES
1244 get_pack_strings = F_PACK_STRINGS
966 PPCODE: 1245 PPCODE:
967 XPUSHs (boolSV (self->flags & ix)); 1246 XPUSHs (boolSV (self->flags & ix));
968 1247
969void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1248void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
970 PPCODE: 1249 PPCODE:
983 XPUSHs (ST (0)); 1262 XPUSHs (ST (0));
984 1263
985int get_max_size (CBOR *self) 1264int get_max_size (CBOR *self)
986 CODE: 1265 CODE:
987 RETVAL = self->max_size; 1266 RETVAL = self->max_size;
1267 OUTPUT:
1268 RETVAL
1269
1270void filter (CBOR *self, SV *filter = 0)
1271 PPCODE:
1272 SvREFCNT_dec (self->filter);
1273 self->filter = filter ? newSVsv (filter) : filter;
1274 XPUSHs (ST (0));
1275
1276SV *get_filter (CBOR *self)
1277 CODE:
1278 RETVAL = self->filter ? self->filter : NEWSV (0, 0);
988 OUTPUT: 1279 OUTPUT:
989 RETVAL 1280 RETVAL
990 1281
991void encode (CBOR *self, SV *scalar) 1282void encode (CBOR *self, SV *scalar)
992 PPCODE: 1283 PPCODE:
1007 EXTEND (SP, 2); 1298 EXTEND (SP, 2);
1008 PUSHs (sv); 1299 PUSHs (sv);
1009 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1300 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1010} 1301}
1011 1302
1303void DESTROY (CBOR *self)
1304 PPCODE:
1305 cbor_free (self);
1306
1012PROTOTYPES: ENABLE 1307PROTOTYPES: ENABLE
1013 1308
1014void encode_cbor (SV *scalar) 1309void encode_cbor (SV *scalar)
1310 ALIAS:
1311 encode_cbor = 0
1312 encode_cbor_sharing = F_ALLOW_SHARING
1015 PPCODE: 1313 PPCODE:
1016{ 1314{
1017 CBOR cbor; 1315 CBOR cbor;
1018 cbor_init (&cbor); 1316 cbor_init (&cbor);
1317 cbor.flags |= ix;
1019 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1318 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1020 XPUSHs (scalar); 1319 XPUSHs (scalar);
1021} 1320}
1022 1321
1023void decode_cbor (SV *cborstr) 1322void decode_cbor (SV *cborstr)

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines