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 1.18 by root, Wed Nov 20 01:09:46 2013 UTC vs.
Revision 1.37 by root, Sat Nov 30 18:13:53 2013 UTC

19# define HvNAMELEN(hv) HvNAMELEN_get (hv) 19# define HvNAMELEN(hv) HvNAMELEN_get (hv)
20#endif 20#endif
21#ifndef HvNAMEUTF8 21#ifndef HvNAMEUTF8
22# define HvNAMEUTF8(hv) 0 22# define HvNAMEUTF8(hv) 0
23#endif 23#endif
24#ifndef SvREFCNT_dec_NN
25# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
26#endif
27
28// known major and minor types
29enum cbor_type
30{
31 MAJOR_SHIFT = 5,
32 MINOR_MASK = 0x1f,
33
34 MAJOR_POS_INT = 0 << MAJOR_SHIFT,
35 MAJOR_NEG_INT = 1 << MAJOR_SHIFT,
36 MAJOR_BYTES = 2 << MAJOR_SHIFT,
37 MAJOR_TEXT = 3 << MAJOR_SHIFT,
38 MAJOR_ARRAY = 4 << MAJOR_SHIFT,
39 MAJOR_MAP = 5 << MAJOR_SHIFT,
40 MAJOR_TAG = 6 << MAJOR_SHIFT,
41 MAJOR_MISC = 7 << MAJOR_SHIFT,
42
43 // INT/STRING/ARRAY/MAP subtypes
44 LENGTH_EXT1 = 24,
45 LENGTH_EXT2 = 25,
46 LENGTH_EXT4 = 26,
47 LENGTH_EXT8 = 27,
48
49 // SIMPLE types (effectively MISC subtypes)
50 SIMPLE_FALSE = 20,
51 SIMPLE_TRUE = 21,
52 SIMPLE_NULL = 22,
53 SIMPLE_UNDEF = 23,
54
55 // MISC subtype (unused)
56 MISC_EXT1 = 24,
57 MISC_FLOAT16 = 25,
58 MISC_FLOAT32 = 26,
59 MISC_FLOAT64 = 27,
60
61 // BYTES/TEXT/ARRAY/MAP
62 MINOR_INDEF = 31,
63};
24 64
25// known tags 65// known tags
26enum cbor_tag 66enum cbor_tag
27{ 67{
28 // inofficial extensions (pending iana registration) 68 // extensions
69 CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
29 CBOR_TAG_PERL_OBJECT = 24, // http://cbor.schmorp.de/perl-object 70 CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
30 CBOR_TAG_GENERIC_OBJECT = 25, // http://cbor.schmorp.de/generic-object 71 CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
31 CBOR_TAG_VALUE_SHARABLE = 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
64 104
65#define INIT_SIZE 32 // initial scalar size to be allocated 105#define INIT_SIZE 32 // initial scalar size to be allocated
66 106
67#define SB do { 107#define SB do {
68#define SE } while (0) 108#define SE } while (0)
80# define CBOR_SLOW 0 120# define CBOR_SLOW 0
81# define CBOR_STASH cbor_stash 121# define CBOR_STASH cbor_stash
82#endif 122#endif
83 123
84static 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::
85static SV *types_true, *types_false, *types_error, *sv_cbor; 125static SV *types_true, *types_false, *types_error, *sv_cbor, *default_filter;
86 126
87typedef struct { 127typedef struct {
88 U32 flags; 128 U32 flags;
89 U32 max_depth; 129 U32 max_depth;
90 STRLEN max_size; 130 STRLEN max_size;
131 SV *filter;
91} CBOR; 132} CBOR;
92 133
93ecb_inline void 134ecb_inline void
94cbor_init (CBOR *cbor) 135cbor_init (CBOR *cbor)
95{ 136{
96 Zero (cbor, 1, CBOR); 137 Zero (cbor, 1, CBOR);
97 cbor->max_depth = 512; 138 cbor->max_depth = 512;
139}
140
141ecb_inline void
142cbor_free (CBOR *cbor)
143{
144 SvREFCNT_dec (cbor->filter);
98} 145}
99 146
100///////////////////////////////////////////////////////////////////////////// 147/////////////////////////////////////////////////////////////////////////////
101// utility functions 148// utility functions
102 149
124 SvPV_renew (sv, SvCUR (sv) + 1); 171 SvPV_renew (sv, SvCUR (sv) + 1);
125#endif 172#endif
126 } 173 }
127} 174}
128 175
176// minimum length of a string to be registered for stringref
177ecb_inline int
178minimum_string_length (UV idx)
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}
190
129///////////////////////////////////////////////////////////////////////////// 191/////////////////////////////////////////////////////////////////////////////
130// encoder 192// encoder
131 193
132// structure used for encoding CBOR 194// structure used for encoding CBOR
133typedef struct 195typedef struct
135 char *cur; // SvPVX (sv) + current output position 197 char *cur; // SvPVX (sv) + current output position
136 char *end; // SvEND (sv) 198 char *end; // SvEND (sv)
137 SV *sv; // result scalar 199 SV *sv; // result scalar
138 CBOR cbor; 200 CBOR cbor;
139 U32 depth; // recursion level 201 U32 depth; // recursion level
140 HV *stringref; // string => index, or 0 202 HV *stringref[2]; // string => index, or 0 ([0] = bytes, [1] = utf-8)
203 UV stringref_idx;
141 HV *sharable; // ptr => index, or 0 204 HV *shareable; // ptr => index, or 0
142 HV *sharable_idx; 205 UV shareable_idx;
143} enc_t; 206} enc_t;
144 207
145ecb_inline void 208ecb_inline void
146need (enc_t *enc, STRLEN len) 209need (enc_t *enc, STRLEN len)
147{ 210{
164static void 227static void
165encode_uint (enc_t *enc, int major, UV len) 228encode_uint (enc_t *enc, int major, UV len)
166{ 229{
167 need (enc, 9); 230 need (enc, 9);
168 231
169 if (len < 24) 232 if (ecb_expect_true (len < LENGTH_EXT1))
170 *enc->cur++ = major | len; 233 *enc->cur++ = major | len;
171 else if (len <= 0xff) 234 else if (ecb_expect_true (len <= 0xffU))
172 { 235 {
173 *enc->cur++ = major | 24; 236 *enc->cur++ = major | LENGTH_EXT1;
174 *enc->cur++ = len; 237 *enc->cur++ = len;
175 } 238 }
176 else if (len <= 0xffff) 239 else if (len <= 0xffffU)
177 { 240 {
178 *enc->cur++ = major | 25; 241 *enc->cur++ = major | LENGTH_EXT2;
179 *enc->cur++ = len >> 8; 242 *enc->cur++ = len >> 8;
180 *enc->cur++ = len; 243 *enc->cur++ = len;
181 } 244 }
182 else if (len <= 0xffffffff) 245 else if (len <= 0xffffffffU)
183 { 246 {
184 *enc->cur++ = major | 26; 247 *enc->cur++ = major | LENGTH_EXT4;
185 *enc->cur++ = len >> 24; 248 *enc->cur++ = len >> 24;
186 *enc->cur++ = len >> 16; 249 *enc->cur++ = len >> 16;
187 *enc->cur++ = len >> 8; 250 *enc->cur++ = len >> 8;
188 *enc->cur++ = len; 251 *enc->cur++ = len;
189 } 252 }
190 else 253 else
191 { 254 {
192 *enc->cur++ = major | 27; 255 *enc->cur++ = major | LENGTH_EXT8;
193 *enc->cur++ = len >> 56; 256 *enc->cur++ = len >> 56;
194 *enc->cur++ = len >> 48; 257 *enc->cur++ = len >> 48;
195 *enc->cur++ = len >> 40; 258 *enc->cur++ = len >> 40;
196 *enc->cur++ = len >> 32; 259 *enc->cur++ = len >> 32;
197 *enc->cur++ = len >> 24; 260 *enc->cur++ = len >> 24;
199 *enc->cur++ = len >> 8; 262 *enc->cur++ = len >> 8;
200 *enc->cur++ = len; 263 *enc->cur++ = len;
201 } 264 }
202} 265}
203 266
204static void 267ecb_inline void
268encode_tag (enc_t *enc, UV tag)
269{
270 encode_uint (enc, MAJOR_TAG, tag);
271}
272
273ecb_inline void
205encode_str (enc_t *enc, int utf8, char *str, STRLEN len) 274encode_str (enc_t *enc, int utf8, char *str, STRLEN len)
206{ 275{
207 encode_uint (enc, utf8 ? 0x60 : 0x40, len); 276 encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
208 need (enc, len); 277 need (enc, len);
209 memcpy (enc->cur, str, len); 278 memcpy (enc->cur, str, len);
210 enc->cur += len; 279 enc->cur += len;
211} 280}
212 281
213ecb_inline void
214encode_tag (enc_t *enc, UV tag)
215{
216 encode_uint (enc, 0xc0, tag);
217}
218
219static int 282static void
220encode_sharable2 (enc_t *enc, SV *sv) 283encode_strref (enc_t *enc, int utf8, char *str, STRLEN len)
221{ 284{
222 if (!enc->sharable) 285 if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
223 enc->sharable = (HV *)sv_2mortal ((SV *)newHV ()); 286 {
287 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
224 288
225 SV **svp = hv_fetch (enc->sharable, &sv, sizeof (sv), 1);
226
227 if (SvOK (*svp)) 289 if (SvOK (*svp))
228 { 290 {
291 // already registered, use stringref
229 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF); 292 encode_tag (enc, CBOR_TAG_STRINGREF);
230 encode_uint (enc, 0x00, SvUV (*svp)); 293 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
231
232 return 1; 294 return;
233 } 295 }
234 else 296 else if (len >= minimum_string_length (enc->stringref_idx))
235 { 297 {
298 // register only
236 sv_setuv (*svp, enc->sharable_idx++); 299 sv_setuv (*svp, enc->stringref_idx);
237 encode_tag (enc, CBOR_TAG_VALUE_SHARABLE); 300 ++enc->stringref_idx;
238 301 }
239 return 0;
240 } 302 }
241}
242 303
243ecb_inline int 304 encode_str (enc, utf8, str, len);
244encode_sharable (enc_t *enc, SV *sv)
245{
246 if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)
247 && ecb_expect_false (SvREFCNT (sv) > 1))
248 return encode_sharable2 (enc, sv);
249
250 return 0;
251} 305}
252 306
253static void encode_sv (enc_t *enc, SV *sv); 307static void encode_sv (enc_t *enc, SV *sv);
254 308
255static void 309static void
260 if (enc->depth >= enc->cbor.max_depth) 314 if (enc->depth >= enc->cbor.max_depth)
261 croak (ERR_NESTING_EXCEEDED); 315 croak (ERR_NESTING_EXCEEDED);
262 316
263 ++enc->depth; 317 ++enc->depth;
264 318
265 encode_uint (enc, 0x80, len + 1); 319 encode_uint (enc, MAJOR_ARRAY, len + 1);
266 320
267 for (i = 0; i <= len; ++i) 321 for (i = 0; i <= len; ++i)
268 { 322 {
269 SV **svp = av_fetch (av, i, 0); 323 SV **svp = av_fetch (av, i, 0);
270 encode_sv (enc, svp ? *svp : &PL_sv_undef); 324 encode_sv (enc, svp ? *svp : &PL_sv_undef);
285 339
286 int pairs = hv_iterinit (hv); 340 int pairs = hv_iterinit (hv);
287 int mg = SvMAGICAL (hv); 341 int mg = SvMAGICAL (hv);
288 342
289 if (mg) 343 if (mg)
290 encode_ch (enc, 0xa0 | 31); 344 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
291 else 345 else
292 encode_uint (enc, 0xa0, pairs); 346 encode_uint (enc, MAJOR_MAP, pairs);
293 347
294 while ((he = hv_iternext (hv))) 348 while ((he = hv_iternext (hv)))
295 { 349 {
296 if (HeKLEN (he) == HEf_SVKEY) 350 if (HeKLEN (he) == HEf_SVKEY)
297 encode_sv (enc, HeSVKEY (he)); 351 encode_sv (enc, HeSVKEY (he));
298 else 352 else
299 encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); 353 encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
300 354
301 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));
302 } 356 }
303 357
304 if (mg) 358 if (mg)
305 encode_ch (enc, 0xe0 | 31); 359 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
306 360
307 --enc->depth; 361 --enc->depth;
308} 362}
309 363
310// encode objects, arrays and special \0=false and \1=true values. 364// encode objects, arrays and special \0=false and \1=true values.
311static void 365static void
312encode_rv (enc_t *enc, SV *sv) 366encode_rv (enc_t *enc, SV *sv)
313{ 367{
314 svtype svt;
315
316 SvGETMAGIC (sv); 368 SvGETMAGIC (sv);
369
317 svt = SvTYPE (sv); 370 svtype svt = SvTYPE (sv);
318
319 if (encode_sharable (enc, sv))
320 return;
321 371
322 if (ecb_expect_false (SvOBJECT (sv))) 372 if (ecb_expect_false (SvOBJECT (sv)))
323 { 373 {
324 HV *boolean_stash = !CBOR_SLOW || types_boolean_stash 374 HV *boolean_stash = !CBOR_SLOW || types_boolean_stash
325 ? types_boolean_stash 375 ? types_boolean_stash
330 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 380 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
331 ? cbor_tagged_stash 381 ? cbor_tagged_stash
332 : gv_stashpv ("CBOR::XS::Tagged" , 1); 382 : gv_stashpv ("CBOR::XS::Tagged" , 1);
333 383
334 HV *stash = SvSTASH (sv); 384 HV *stash = SvSTASH (sv);
335 GV *method;
336 385
337 if (stash == boolean_stash) 386 if (stash == boolean_stash)
338 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 }
339 else if (stash == error_stash) 391 else if (stash == error_stash)
340 encode_ch (enc, 0xe0 | 23); 392 {
393 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
394 return;
395 }
341 else if (stash == tagged_stash) 396 else if (stash == tagged_stash)
342 { 397 {
343 if (svt != SVt_PVAV) 398 if (svt != SVt_PVAV)
344 croak ("encountered CBOR::XS::Tagged object that isn't an array"); 399 croak ("encountered CBOR::XS::Tagged object that isn't an array");
345 400
346 encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); 401 encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
347 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))
348 } 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
349 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) 435 if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
350 { 436 {
351 dSP; 437 dSP;
352 438
353 ENTER; SAVETMPS; PUSHMARK (SP); 439 ENTER; SAVETMPS; PUSHMARK (SP);
354 // 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
386 // catch this surprisingly common error 472 // catch this surprisingly common error
387 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) 473 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
388 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));
389 475
390 encode_tag (enc, CBOR_TAG_PERL_OBJECT); 476 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
391 encode_uint (enc, 0x80, count + 1); 477 encode_uint (enc, MAJOR_ARRAY, count + 1);
392 encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); 478 encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
393 479
394 while (count) 480 while (count)
395 encode_sv (enc, SP[1 - count--]); 481 encode_sv (enc, SP[1 - count--]);
396 482
397 PUTBACK; 483 PUTBACK;
418{ 504{
419 double nv = SvNVX (sv); 505 double nv = SvNVX (sv);
420 506
421 need (enc, 9); 507 need (enc, 9);
422 508
423 if (ecb_expect_false (nv == (U32)nv)) 509 if (ecb_expect_false (nv == (NV)(U32)nv))
424 encode_uint (enc, 0x00, (U32)nv); 510 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
425 //TODO: maybe I32? 511 //TODO: maybe I32?
426 else if (ecb_expect_false (nv == (float)nv)) 512 else if (ecb_expect_false (nv == (float)nv))
427 { 513 {
428 uint32_t fp = ecb_float_to_binary32 (nv); 514 uint32_t fp = ecb_float_to_binary32 (nv);
429 515
430 *enc->cur++ = 0xe0 | 26; 516 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
431 517
432 if (!ecb_big_endian ()) 518 if (!ecb_big_endian ())
433 fp = ecb_bswap32 (fp); 519 fp = ecb_bswap32 (fp);
434 520
435 memcpy (enc->cur, &fp, 4); 521 memcpy (enc->cur, &fp, 4);
437 } 523 }
438 else 524 else
439 { 525 {
440 uint64_t fp = ecb_double_to_binary64 (nv); 526 uint64_t fp = ecb_double_to_binary64 (nv);
441 527
442 *enc->cur++ = 0xe0 | 27; 528 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
443 529
444 if (!ecb_big_endian ()) 530 if (!ecb_big_endian ())
445 fp = ecb_bswap64 (fp); 531 fp = ecb_bswap64 (fp);
446 532
447 memcpy (enc->cur, &fp, 8); 533 memcpy (enc->cur, &fp, 8);
452static void 538static void
453encode_sv (enc_t *enc, SV *sv) 539encode_sv (enc_t *enc, SV *sv)
454{ 540{
455 SvGETMAGIC (sv); 541 SvGETMAGIC (sv);
456 542
457 if (encode_sharable (enc, sv))
458 return;
459
460 if (SvPOKp (sv)) 543 if (SvPOKp (sv))
461 { 544 {
462 STRLEN len; 545 STRLEN len;
463 char *str = SvPV (sv, len); 546 char *str = SvPV (sv, len);
464 encode_str (enc, SvUTF8 (sv), str, len); 547 encode_strref (enc, SvUTF8 (sv), str, len);
465 } 548 }
466 else if (SvNOKp (sv)) 549 else if (SvNOKp (sv))
467 encode_nv (enc, sv); 550 encode_nv (enc, sv);
468 else if (SvIOKp (sv)) 551 else if (SvIOKp (sv))
469 { 552 {
470 if (SvIsUV (sv)) 553 if (SvIsUV (sv))
471 encode_uint (enc, 0x00, SvUVX (sv)); 554 encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
472 else if (SvIVX (sv) >= 0) 555 else if (SvIVX (sv) >= 0)
473 encode_uint (enc, 0x00, SvIVX (sv)); 556 encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
474 else 557 else
475 encode_uint (enc, 0x20, -(SvIVX (sv) + 1)); 558 encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
476 } 559 }
477 else if (SvROK (sv)) 560 else if (SvROK (sv))
478 encode_rv (enc, SvRV (sv)); 561 encode_rv (enc, SvRV (sv));
479 else if (!SvOK (sv)) 562 else if (!SvOK (sv))
480 encode_ch (enc, 0xe0 | 22); 563 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
481 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 564 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
482 encode_ch (enc, 0xe0 | 23); 565 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
483 else 566 else
484 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",
485 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 568 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
486} 569}
487 570
494 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 577 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
495 enc.cur = SvPVX (enc.sv); 578 enc.cur = SvPVX (enc.sv);
496 enc.end = SvEND (enc.sv); 579 enc.end = SvEND (enc.sv);
497 580
498 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
499 encode_sv (&enc, scalar); 590 encode_sv (&enc, scalar);
500 591
501 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv)); 592 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv));
502 *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
503 594
517 U8 *end; // end of input string 608 U8 *end; // end of input string
518 const char *err; // parse error, if != 0 609 const char *err; // parse error, if != 0
519 CBOR cbor; 610 CBOR cbor;
520 U32 depth; // recursion depth 611 U32 depth; // recursion depth
521 U32 maxdepth; // recursion depth limit 612 U32 maxdepth; // recursion depth limit
522 AV *sharable; 613 AV *shareable;
614 AV *stringref;
615 SV *decode_tagged;
523} dec_t; 616} dec_t;
524 617
525#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
526 619
527#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")
530#define DEC_DEC_DEPTH --dec->depth 623#define DEC_DEC_DEPTH --dec->depth
531 624
532static UV 625static UV
533decode_uint (dec_t *dec) 626decode_uint (dec_t *dec)
534{ 627{
535 switch (*dec->cur & 31) 628 U8 m = *dec->cur & MINOR_MASK;
536 { 629 ++dec->cur;
537 case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
538 case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
539 case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
540 return *dec->cur++ & 31;
541 630
542 case 24: 631 if (ecb_expect_true (m < LENGTH_EXT1))
632 return m;
633 else if (ecb_expect_true (m == LENGTH_EXT1))
634 {
543 WANT (2); 635 WANT (1);
544 dec->cur += 2; 636 dec->cur += 1;
545 return dec->cur[-1]; 637 return dec->cur[-1];
546 638 }
547 case 25: 639 else if (ecb_expect_true (m == LENGTH_EXT2))
640 {
548 WANT (3); 641 WANT (2);
549 dec->cur += 3; 642 dec->cur += 2;
550 return (((UV)dec->cur[-2]) << 8) 643 return (((UV)dec->cur[-2]) << 8)
551 | ((UV)dec->cur[-1]); 644 | ((UV)dec->cur[-1]);
552 645 }
553 case 26: 646 else if (ecb_expect_true (m == LENGTH_EXT4))
647 {
554 WANT (5); 648 WANT (4);
555 dec->cur += 5; 649 dec->cur += 4;
556 return (((UV)dec->cur[-4]) << 24) 650 return (((UV)dec->cur[-4]) << 24)
557 | (((UV)dec->cur[-3]) << 16) 651 | (((UV)dec->cur[-3]) << 16)
558 | (((UV)dec->cur[-2]) << 8) 652 | (((UV)dec->cur[-2]) << 8)
559 | ((UV)dec->cur[-1]); 653 | ((UV)dec->cur[-1]);
560 654 }
561 case 27: 655 else if (ecb_expect_true (m == LENGTH_EXT8))
656 {
562 WANT (9); 657 WANT (8);
563 dec->cur += 9; 658 dec->cur += 8;
659
660 return
661#if UVSIZE < 8
662 0
663#else
564 return (((UV)dec->cur[-8]) << 56) 664 (((UV)dec->cur[-8]) << 56)
565 | (((UV)dec->cur[-7]) << 48) 665 | (((UV)dec->cur[-7]) << 48)
566 | (((UV)dec->cur[-6]) << 40) 666 | (((UV)dec->cur[-6]) << 40)
567 | (((UV)dec->cur[-5]) << 32) 667 | (((UV)dec->cur[-5]) << 32)
668#endif
568 | (((UV)dec->cur[-4]) << 24) 669 | (((UV)dec->cur[-4]) << 24)
569 | (((UV)dec->cur[-3]) << 16) 670 | (((UV)dec->cur[-3]) << 16)
570 | (((UV)dec->cur[-2]) << 8) 671 | (((UV)dec->cur[-2]) << 8)
571 | ((UV)dec->cur[-1]); 672 | ((UV)dec->cur[-1]);
572 673 }
573 default: 674 else
574 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 675 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
575 }
576 676
577fail: 677fail:
578 return 0; 678 return 0;
579} 679}
580 680
585{ 685{
586 AV *av = newAV (); 686 AV *av = newAV ();
587 687
588 DEC_INC_DEPTH; 688 DEC_INC_DEPTH;
589 689
590 if ((*dec->cur & 31) == 31) 690 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
591 { 691 {
592 ++dec->cur; 692 ++dec->cur;
593 693
594 for (;;) 694 for (;;)
595 { 695 {
596 WANT (1); 696 WANT (1);
597 697
598 if (*dec->cur == (0xe0 | 31)) 698 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
599 { 699 {
600 ++dec->cur; 700 ++dec->cur;
601 break; 701 break;
602 } 702 }
603 703
606 } 706 }
607 else 707 else
608 { 708 {
609 int i, len = decode_uint (dec); 709 int i, len = decode_uint (dec);
610 710
711 WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
611 av_fill (av, len - 1); 712 av_fill (av, len - 1);
612 713
613 for (i = 0; i < len; ++i) 714 for (i = 0; i < len; ++i)
614 AvARRAY (av)[i] = decode_sv (dec); 715 AvARRAY (av)[i] = decode_sv (dec);
615 } 716 }
625 726
626static void 727static void
627decode_he (dec_t *dec, HV *hv) 728decode_he (dec_t *dec, HV *hv)
628{ 729{
629 // for speed reasons, we specialcase single-string 730 // for speed reasons, we specialcase single-string
630 // byte or utf-8 strings as keys. 731 // byte or utf-8 strings as keys, but only when !stringref
631 732
632 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))
633 { 735 {
634 I32 len = decode_uint (dec); 736 I32 len = decode_uint (dec);
635 char *key = (char *)dec->cur; 737 char *key = (char *)dec->cur;
636 738
637 dec->cur += len; 739 dec->cur += len;
638 740
741 if (ecb_expect_false (dec->stringref))
742 av_push (dec->stringref, newSVpvn (key, len));
743
639 hv_store (hv, key, len, decode_sv (dec), 0); 744 hv_store (hv, key, len, decode_sv (dec), 0);
745
746 return;
640 } 747 }
641 else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27) 748 else if (ecb_expect_true ((*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
642 { 749 {
643 I32 len = decode_uint (dec); 750 I32 len = decode_uint (dec);
644 char *key = (char *)dec->cur; 751 char *key = (char *)dec->cur;
645 752
646 dec->cur += len; 753 dec->cur += len;
647 754
755 if (ecb_expect_false (dec->stringref))
756 av_push (dec->stringref, newSVpvn_utf8 (key, len, 1));
757
648 hv_store (hv, key, -len, decode_sv (dec), 0); 758 hv_store (hv, key, -len, decode_sv (dec), 0);
759
760 return;
649 } 761 }
650 else 762
651 {
652 SV *k = decode_sv (dec); 763 SV *k = decode_sv (dec);
653 SV *v = decode_sv (dec); 764 SV *v = decode_sv (dec);
654 765
655 hv_store_ent (hv, k, v, 0); 766 hv_store_ent (hv, k, v, 0);
656 SvREFCNT_dec (k); 767 SvREFCNT_dec (k);
657 }
658} 768}
659 769
660static SV * 770static SV *
661decode_hv (dec_t *dec) 771decode_hv (dec_t *dec)
662{ 772{
663 HV *hv = newHV (); 773 HV *hv = newHV ();
664 774
665 DEC_INC_DEPTH; 775 DEC_INC_DEPTH;
666 776
667 if ((*dec->cur & 31) == 31) 777 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
668 { 778 {
669 ++dec->cur; 779 ++dec->cur;
670 780
671 for (;;) 781 for (;;)
672 { 782 {
673 WANT (1); 783 WANT (1);
674 784
675 if (*dec->cur == (0xe0 | 31)) 785 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
676 { 786 {
677 ++dec->cur; 787 ++dec->cur;
678 break; 788 break;
679 } 789 }
680 790
701static SV * 811static SV *
702decode_str (dec_t *dec, int utf8) 812decode_str (dec_t *dec, int utf8)
703{ 813{
704 SV *sv = 0; 814 SV *sv = 0;
705 815
706 if ((*dec->cur & 31) == 31) 816 if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
707 { 817 {
818 // indefinite length strings
708 ++dec->cur; 819 ++dec->cur;
709 820
821 U8 major = *dec->cur & MAJOR_MISC;
822
710 sv = newSVpvn ("", 0); 823 sv = newSVpvn ("", 0);
711 824
712 // not very fast, and certainly not robust against illegal input
713 for (;;) 825 for (;;)
714 { 826 {
715 WANT (1); 827 WANT (1);
716 828
717 if (*dec->cur == (0xe0 | 31)) 829 if ((*dec->cur - major) > LENGTH_EXT8)
830 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
718 { 831 {
719 ++dec->cur; 832 ++dec->cur;
720 break; 833 break;
721 } 834 }
835 else
836 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
722 837
723 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;
724 } 843 }
725 } 844 }
726 else 845 else
727 { 846 {
728 STRLEN len = decode_uint (dec); 847 STRLEN len = decode_uint (dec);
729 848
730 WANT (len); 849 WANT (len);
731 sv = newSVpvn (dec->cur, len); 850 sv = newSVpvn (dec->cur, len);
732 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));
733 } 856 }
734 857
735 if (utf8) 858 if (utf8)
736 SvUTF8_on (sv); 859 SvUTF8_on (sv);
737 860
743} 866}
744 867
745static SV * 868static SV *
746decode_tagged (dec_t *dec) 869decode_tagged (dec_t *dec)
747{ 870{
871 SV *sv = 0;
748 UV tag = decode_uint (dec); 872 UV tag = decode_uint (dec);
749 SV *sv = decode_sv (dec); 873
874 WANT (1);
750 875
751 switch (tag) 876 switch (tag)
752 { 877 {
753 case CBOR_TAG_MAGIC: 878 case CBOR_TAG_MAGIC:
754 return sv; 879 sv = decode_sv (dec);
880 break;
755 881
756 case CBOR_TAG_INDIRECTION: 882 case CBOR_TAG_INDIRECTION:
757 return newRV_noinc (sv); 883 sv = newRV_noinc (decode_sv (dec));
884 break;
758 885
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
759 case CBOR_TAG_VALUE_SHARABLE: 913 case CBOR_TAG_VALUE_SHAREABLE:
914 {
760 if (ecb_expect_false (!dec->sharable)) 915 if (ecb_expect_false (!dec->shareable))
761 dec->sharable = (AV *)sv_2mortal ((SV *)newAV ()); 916 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
762 917
918 if (dec->cbor.flags & F_ALLOW_CYCLES)
919 {
920 sv = newSV (0);
763 av_push (dec->sharable, SvREFCNT_inc_NN (sv)); 921 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
764 922
765 return sv; 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;
766 936
767 case CBOR_TAG_VALUE_SHAREDREF: 937 case CBOR_TAG_VALUE_SHAREDREF:
768 { 938 {
769 // TODO: should verify that the sv atcually was a CBOR unsigned integer 939 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
770 UV idx = SvUV (sv); 940 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
771 941
942 UV idx = decode_uint (dec);
943
772 if (!dec->sharable || idx > AvFILLp (dec->sharable)) 944 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable))
773 ERR ("corrupted CBOR data (sharedref index out of bounds)"); 945 ERR ("corrupted CBOR data (sharedref index out of bounds)");
774 946
775 SvREFCNT_dec (sv);
776
777 return SvREFCNT_inc_NN (AvARRAY (dec->sharable)[idx]); 947 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
948
949 if (sv == &PL_sv_undef)
950 ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
778 } 951 }
952 break;
779 953
780 case CBOR_TAG_PERL_OBJECT: 954 case CBOR_TAG_PERL_OBJECT:
781 { 955 {
956 sv = decode_sv (dec);
957
782 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) 958 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
783 ERR ("corrupted CBOR data (non-array perl object)"); 959 ERR ("corrupted CBOR data (non-array perl object)");
784 960
785 AV *av = (AV *)SvRV (sv); 961 AV *av = (AV *)SvRV (sv);
786 int len = av_len (av) + 1; 962 int len = av_len (av) + 1;
821 sv = SvREFCNT_inc (POPs); 997 sv = SvREFCNT_inc (POPs);
822 998
823 PUTBACK; 999 PUTBACK;
824 1000
825 FREETMPS; LEAVE; 1001 FREETMPS; LEAVE;
826
827 return sv;
828 } 1002 }
1003 break;
829 1004
830 default: 1005 default:
831 { 1006 {
1007 sv = decode_sv (dec);
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 {
832 AV *av = newAV (); 1032 AV *av = newAV ();
833 av_push (av, newSVuv (tag)); 1033 av_push (av, newSVuv (tag));
834 av_push (av, sv); 1034 av_push (av, sv);
835 1035
836 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 1036 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
837 ? cbor_tagged_stash 1037 ? cbor_tagged_stash
838 : gv_stashpv ("CBOR::XS::Tagged" , 1); 1038 : gv_stashpv ("CBOR::XS::Tagged" , 1);
839
840 return sv_bless (newRV_noinc ((SV *)av), tagged_stash); 1039 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
841 } 1040 }
1041
1042 PUTBACK;
1043
1044 FREETMPS; LEAVE;
1045 }
1046 break;
842 } 1047 }
1048
1049 return sv;
843 1050
844fail: 1051fail:
845 SvREFCNT_dec (sv); 1052 SvREFCNT_dec (sv);
846 return &PL_sv_undef; 1053 return &PL_sv_undef;
847} 1054}
849static SV * 1056static SV *
850decode_sv (dec_t *dec) 1057decode_sv (dec_t *dec)
851{ 1058{
852 WANT (1); 1059 WANT (1);
853 1060
854 switch (*dec->cur >> 5) 1061 switch (*dec->cur >> MAJOR_SHIFT)
855 { 1062 {
856 case 0: // unsigned int 1063 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
857 return newSVuv (decode_uint (dec)); 1064 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
858 case 1: // negative int 1065 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
859 return newSViv (-1 - (IV)decode_uint (dec)); 1066 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
860 case 2: // octet string 1067 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
861 return decode_str (dec, 0); 1068 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
862 case 3: // utf-8 string 1069 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
863 return decode_str (dec, 1); 1070
864 case 4: // array 1071 case MAJOR_MISC >> MAJOR_SHIFT:
865 return decode_av (dec);
866 case 5: // map
867 return decode_hv (dec);
868 case 6: // tag
869 return decode_tagged (dec);
870 case 7: // misc
871 switch (*dec->cur++ & 31) 1072 switch (*dec->cur++ & MINOR_MASK)
872 { 1073 {
873 case 20: 1074 case SIMPLE_FALSE:
874#if CBOR_SLOW 1075#if CBOR_SLOW
875 types_false = get_bool ("Types::Serialiser::false"); 1076 types_false = get_bool ("Types::Serialiser::false");
876#endif 1077#endif
877 return newSVsv (types_false); 1078 return newSVsv (types_false);
878 case 21: 1079 case SIMPLE_TRUE:
879#if CBOR_SLOW 1080#if CBOR_SLOW
880 types_true = get_bool ("Types::Serialiser::true"); 1081 types_true = get_bool ("Types::Serialiser::true");
881#endif 1082#endif
882 return newSVsv (types_true); 1083 return newSVsv (types_true);
883 case 22: 1084 case SIMPLE_NULL:
884 return newSVsv (&PL_sv_undef); 1085 return newSVsv (&PL_sv_undef);
885 case 23: 1086 case SIMPLE_UNDEF:
886#if CBOR_SLOW 1087#if CBOR_SLOW
887 types_error = get_bool ("Types::Serialiser::error"); 1088 types_error = get_bool ("Types::Serialiser::error");
888#endif 1089#endif
889 return newSVsv (types_error); 1090 return newSVsv (types_error);
890 1091
891 case 25: 1092 case MISC_FLOAT16:
892 { 1093 {
893 WANT (2); 1094 WANT (2);
894 1095
895 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1096 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
896 dec->cur += 2; 1097 dec->cur += 2;
897 1098
898 return newSVnv (ecb_binary16_to_float (fp)); 1099 return newSVnv (ecb_binary16_to_float (fp));
899 } 1100 }
900 1101
901 case 26: 1102 case MISC_FLOAT32:
902 { 1103 {
903 uint32_t fp; 1104 uint32_t fp;
904 WANT (4); 1105 WANT (4);
905 memcpy (&fp, dec->cur, 4); 1106 memcpy (&fp, dec->cur, 4);
906 dec->cur += 4; 1107 dec->cur += 4;
909 fp = ecb_bswap32 (fp); 1110 fp = ecb_bswap32 (fp);
910 1111
911 return newSVnv (ecb_binary32_to_float (fp)); 1112 return newSVnv (ecb_binary32_to_float (fp));
912 } 1113 }
913 1114
914 case 27: 1115 case MISC_FLOAT64:
915 { 1116 {
916 uint64_t fp; 1117 uint64_t fp;
917 WANT (8); 1118 WANT (8);
918 memcpy (&fp, dec->cur, 8); 1119 memcpy (&fp, dec->cur, 8);
919 dec->cur += 8; 1120 dec->cur += 8;
922 fp = ecb_bswap64 (fp); 1123 fp = ecb_bswap64 (fp);
923 1124
924 return newSVnv (ecb_binary64_to_double (fp)); 1125 return newSVnv (ecb_binary64_to_double (fp));
925 } 1126 }
926 1127
927 // 0..19 unassigned 1128 // 0..19 unassigned simple
928 // 24 reserved + unassigned (reserved values are not encodable) 1129 // 24 reserved + unassigned (reserved values are not encodable)
929 default: 1130 default:
930 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1131 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");
931 } 1132 }
932 1133
987 types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1); 1188 types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1);
988 1189
989 types_true = get_bool ("Types::Serialiser::true" ); 1190 types_true = get_bool ("Types::Serialiser::true" );
990 types_false = get_bool ("Types::Serialiser::false"); 1191 types_false = get_bool ("Types::Serialiser::false");
991 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);
992 1195
993 sv_cbor = newSVpv ("CBOR", 0); 1196 sv_cbor = newSVpv ("CBOR", 0);
994 SvREADONLY_on (sv_cbor); 1197 SvREADONLY_on (sv_cbor);
995} 1198}
996 1199
1018void shrink (CBOR *self, int enable = 1) 1221void shrink (CBOR *self, int enable = 1)
1019 ALIAS: 1222 ALIAS:
1020 shrink = F_SHRINK 1223 shrink = F_SHRINK
1021 allow_unknown = F_ALLOW_UNKNOWN 1224 allow_unknown = F_ALLOW_UNKNOWN
1022 allow_sharing = F_ALLOW_SHARING 1225 allow_sharing = F_ALLOW_SHARING
1023 dedup_keys = F_DEDUP_KEYS 1226 allow_cycles = F_ALLOW_CYCLES
1024 dedup_strings = F_DEDUP_STRINGS 1227 pack_strings = F_PACK_STRINGS
1025 PPCODE: 1228 PPCODE:
1026{ 1229{
1027 if (enable) 1230 if (enable)
1028 self->flags |= ix; 1231 self->flags |= ix;
1029 else 1232 else
1035void get_shrink (CBOR *self) 1238void get_shrink (CBOR *self)
1036 ALIAS: 1239 ALIAS:
1037 get_shrink = F_SHRINK 1240 get_shrink = F_SHRINK
1038 get_allow_unknown = F_ALLOW_UNKNOWN 1241 get_allow_unknown = F_ALLOW_UNKNOWN
1039 get_allow_sharing = F_ALLOW_SHARING 1242 get_allow_sharing = F_ALLOW_SHARING
1040 get_dedup_keys = F_DEDUP_KEYS 1243 get_allow_cycles = F_ALLOW_CYCLES
1041 get_dedup_strings = F_DEDUP_STRINGS 1244 get_pack_strings = F_PACK_STRINGS
1042 PPCODE: 1245 PPCODE:
1043 XPUSHs (boolSV (self->flags & ix)); 1246 XPUSHs (boolSV (self->flags & ix));
1044 1247
1045void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1248void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1046 PPCODE: 1249 PPCODE:
1059 XPUSHs (ST (0)); 1262 XPUSHs (ST (0));
1060 1263
1061int get_max_size (CBOR *self) 1264int get_max_size (CBOR *self)
1062 CODE: 1265 CODE:
1063 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);
1064 OUTPUT: 1279 OUTPUT:
1065 RETVAL 1280 RETVAL
1066 1281
1067void encode (CBOR *self, SV *scalar) 1282void encode (CBOR *self, SV *scalar)
1068 PPCODE: 1283 PPCODE:
1083 EXTEND (SP, 2); 1298 EXTEND (SP, 2);
1084 PUSHs (sv); 1299 PUSHs (sv);
1085 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1300 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1086} 1301}
1087 1302
1303void DESTROY (CBOR *self)
1304 PPCODE:
1305 cbor_free (self);
1306
1088PROTOTYPES: ENABLE 1307PROTOTYPES: ENABLE
1089 1308
1090void encode_cbor (SV *scalar) 1309void encode_cbor (SV *scalar)
1310 ALIAS:
1311 encode_cbor = 0
1312 encode_cbor_sharing = F_ALLOW_SHARING
1091 PPCODE: 1313 PPCODE:
1092{ 1314{
1093 CBOR cbor; 1315 CBOR cbor;
1094 cbor_init (&cbor); 1316 cbor_init (&cbor);
1317 cbor.flags |= ix;
1095 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1318 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1096 XPUSHs (scalar); 1319 XPUSHs (scalar);
1097} 1320}
1098 1321
1099void decode_cbor (SV *cborstr) 1322void decode_cbor (SV *cborstr)

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines