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 1.24 by root, Wed Nov 20 16:29:02 2013 UTC vs.
Revision 1.47 by root, Mon Apr 27 20:21:53 2015 UTC

7#include <stdlib.h> 7#include <stdlib.h>
8#include <stdio.h> 8#include <stdio.h>
9#include <limits.h> 9#include <limits.h>
10#include <float.h> 10#include <float.h>
11 11
12#define ECB_NO_THREADS 1
12#include "ecb.h" 13#include "ecb.h"
13 14
14// compatibility with perl <5.18 15// compatibility with perl <5.18
15#ifndef HvNAMELEN_get 16#ifndef HvNAMELEN_get
16# define HvNAMELEN_get(hv) strlen (HvNAME (hv)) 17# define HvNAMELEN_get(hv) strlen (HvNAME (hv))
19# define HvNAMELEN(hv) HvNAMELEN_get (hv) 20# define HvNAMELEN(hv) HvNAMELEN_get (hv)
20#endif 21#endif
21#ifndef HvNAMEUTF8 22#ifndef HvNAMEUTF8
22# define HvNAMEUTF8(hv) 0 23# define HvNAMEUTF8(hv) 0
23#endif 24#endif
25#ifndef SvREFCNT_dec_NN
26# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
27#endif
28
29// known major and minor types
30enum cbor_type
31{
32 MAJOR_SHIFT = 5,
33 MINOR_MASK = 0x1f,
34
35 MAJOR_POS_INT = 0 << MAJOR_SHIFT,
36 MAJOR_NEG_INT = 1 << MAJOR_SHIFT,
37 MAJOR_BYTES = 2 << MAJOR_SHIFT,
38 MAJOR_TEXT = 3 << MAJOR_SHIFT,
39 MAJOR_ARRAY = 4 << MAJOR_SHIFT,
40 MAJOR_MAP = 5 << MAJOR_SHIFT,
41 MAJOR_TAG = 6 << MAJOR_SHIFT,
42 MAJOR_MISC = 7 << MAJOR_SHIFT,
43
44 // INT/STRING/ARRAY/MAP subtypes
45 LENGTH_EXT1 = 24,
46 LENGTH_EXT2 = 25,
47 LENGTH_EXT4 = 26,
48 LENGTH_EXT8 = 27,
49
50 // SIMPLE types (effectively MISC subtypes)
51 SIMPLE_FALSE = 20,
52 SIMPLE_TRUE = 21,
53 SIMPLE_NULL = 22,
54 SIMPLE_UNDEF = 23,
55
56 // MISC subtype (unused)
57 MISC_EXT1 = 24,
58 MISC_FLOAT16 = 25,
59 MISC_FLOAT32 = 26,
60 MISC_FLOAT64 = 27,
61
62 // BYTES/TEXT/ARRAY/MAP
63 MINOR_INDEF = 31,
64};
24 65
25// known tags 66// known tags
26enum cbor_tag 67enum cbor_tag
27{ 68{
28 // inofficial extensions (pending iana registration) 69 // extensions
70 CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
29 CBOR_TAG_PERL_OBJECT = 24, // http://cbor.schmorp.de/perl-object 71 CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
30 CBOR_TAG_GENERIC_OBJECT = 25, // http://cbor.schmorp.de/generic-object 72 CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
31 CBOR_TAG_VALUE_SHAREABLE = 26, // http://cbor.schmorp.de/value-sharing 73 CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
32 CBOR_TAG_VALUE_SHAREDREF = 27, // http://cbor.schmorp.de/value-sharing 74 CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
33 CBOR_TAG_STRINGREF_NAMESPACE = 65537, // http://cbor.schmorp.de/stringref 75 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 76 CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
36 77
37 // rfc7049 78 // rfc7049
38 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8 79 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
39 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any 80 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
40 CBOR_TAG_POS_BIGNUM = 2, // byte string 81 CBOR_TAG_POS_BIGNUM = 2, // byte string
41 CBOR_TAG_NEG_BIGNUM = 3, // byte string 82 CBOR_TAG_NEG_BIGNUM = 3, // byte string
42 CBOR_TAG_DECIMAL = 4, // decimal fraction, array 83 CBOR_TAG_DECIMAL = 4, // decimal fraction, array
43 CBOR_TAG_BIGFLOAT = 5, // array 84 CBOR_TAG_BIGFLOAT = 5, // array
44 85
45 CBOR_TAG_CONV_B64U = 21, // base64url, any 86 CBOR_TAG_CONV_B64U = 21, // base64url, any
46 CBOR_TAG_CONV_B64 = 22, // base64, any 87 CBOR_TAG_CONV_B64 = 22, // base64, any
47 CBOR_TAG_CONV_HEX = 23, // base16, any 88 CBOR_TAG_CONV_HEX = 23, // base16, any
48 CBOR_TAG_CBOR = 24, // embedded cbor, byte string 89 CBOR_TAG_CBOR = 24, // embedded cbor, byte string
49 90
50 CBOR_TAG_URI = 32, // URI rfc3986, utf-8 91 CBOR_TAG_URI = 32, // URI rfc3986, utf-8
51 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8 92 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
52 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8 93 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
53 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8 94 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
54 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8 95 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
55 96
56 CBOR_TAG_MAGIC = 55799 // self-describe cbor 97 CBOR_TAG_MAGIC = 55799, // self-describe cbor
57}; 98};
58 99
59#define F_SHRINK 0x00000001UL 100#define F_SHRINK 0x00000001UL
60#define F_ALLOW_UNKNOWN 0x00000002UL 101#define F_ALLOW_UNKNOWN 0x00000002UL
61#define F_ALLOW_SHARING 0x00000004UL //TODO 102#define F_ALLOW_SHARING 0x00000004UL
62#define F_ALLOW_STRINGREF 0x00000008UL //TODO 103#define F_ALLOW_CYCLES 0x00000008UL
104#define F_PACK_STRINGS 0x00000010UL
105#define F_VALIDATE_UTF8 0x00000020UL
63 106
64#define INIT_SIZE 32 // initial scalar size to be allocated 107#define INIT_SIZE 32 // initial scalar size to be allocated
65 108
66#define SB do { 109#define SB do {
67#define SE } while (0) 110#define SE } while (0)
79# define CBOR_SLOW 0 122# define CBOR_SLOW 0
80# define CBOR_STASH cbor_stash 123# define CBOR_STASH cbor_stash
81#endif 124#endif
82 125
83static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS:: 126static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS::
84static SV *types_true, *types_false, *types_error, *sv_cbor; 127static SV *types_true, *types_false, *types_error, *sv_cbor, *default_filter;
85 128
86typedef struct { 129typedef struct {
87 U32 flags; 130 U32 flags;
88 U32 max_depth; 131 U32 max_depth;
89 STRLEN max_size; 132 STRLEN max_size;
133 SV *filter;
134
135 // for the incremental parser
136 STRLEN incr_pos; // the current offset into the text
137 STRLEN incr_need; // minimum bytes needed to decode
138 AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef.
90} CBOR; 139} CBOR;
91 140
92ecb_inline void 141ecb_inline void
93cbor_init (CBOR *cbor) 142cbor_init (CBOR *cbor)
94{ 143{
95 Zero (cbor, 1, CBOR); 144 Zero (cbor, 1, CBOR);
96 cbor->max_depth = 512; 145 cbor->max_depth = 512;
146}
147
148ecb_inline void
149cbor_free (CBOR *cbor)
150{
151 SvREFCNT_dec (cbor->filter);
152 SvREFCNT_dec (cbor->incr_count);
97} 153}
98 154
99///////////////////////////////////////////////////////////////////////////// 155/////////////////////////////////////////////////////////////////////////////
100// utility functions 156// utility functions
101 157
131{ 187{
132 return idx > 23 188 return idx > 23
133 ? idx > 0xffU 189 ? idx > 0xffU
134 ? idx > 0xffffU 190 ? idx > 0xffffU
135 ? idx > 0xffffffffU 191 ? idx > 0xffffffffU
192 ? 11
136 ? 7 193 : 7
137 : 6
138 : 5 194 : 5
139 : 4 195 : 4
140 : 3; 196 : 3;
141} 197}
142 198
179static void 235static void
180encode_uint (enc_t *enc, int major, UV len) 236encode_uint (enc_t *enc, int major, UV len)
181{ 237{
182 need (enc, 9); 238 need (enc, 9);
183 239
184 if (len < 24) 240 if (ecb_expect_true (len < LENGTH_EXT1))
185 *enc->cur++ = major | len; 241 *enc->cur++ = major | len;
186 else if (len <= 0xff) 242 else if (ecb_expect_true (len <= 0xffU))
187 { 243 {
188 *enc->cur++ = major | 24; 244 *enc->cur++ = major | LENGTH_EXT1;
189 *enc->cur++ = len; 245 *enc->cur++ = len;
190 } 246 }
191 else if (len <= 0xffff) 247 else if (len <= 0xffffU)
192 { 248 {
193 *enc->cur++ = major | 25; 249 *enc->cur++ = major | LENGTH_EXT2;
194 *enc->cur++ = len >> 8; 250 *enc->cur++ = len >> 8;
195 *enc->cur++ = len; 251 *enc->cur++ = len;
196 } 252 }
197 else if (len <= 0xffffffff) 253 else if (len <= 0xffffffffU)
198 { 254 {
199 *enc->cur++ = major | 26; 255 *enc->cur++ = major | LENGTH_EXT4;
200 *enc->cur++ = len >> 24; 256 *enc->cur++ = len >> 24;
201 *enc->cur++ = len >> 16; 257 *enc->cur++ = len >> 16;
202 *enc->cur++ = len >> 8; 258 *enc->cur++ = len >> 8;
203 *enc->cur++ = len; 259 *enc->cur++ = len;
204 } 260 }
205 else 261 else
206 { 262 {
207 *enc->cur++ = major | 27; 263 *enc->cur++ = major | LENGTH_EXT8;
208 *enc->cur++ = len >> 56; 264 *enc->cur++ = len >> 56;
209 *enc->cur++ = len >> 48; 265 *enc->cur++ = len >> 48;
210 *enc->cur++ = len >> 40; 266 *enc->cur++ = len >> 40;
211 *enc->cur++ = len >> 32; 267 *enc->cur++ = len >> 32;
212 *enc->cur++ = len >> 24; 268 *enc->cur++ = len >> 24;
217} 273}
218 274
219ecb_inline void 275ecb_inline void
220encode_tag (enc_t *enc, UV tag) 276encode_tag (enc_t *enc, UV tag)
221{ 277{
222 encode_uint (enc, 0xc0, tag); 278 encode_uint (enc, MAJOR_TAG, tag);
279}
280
281ecb_inline void
282encode_str (enc_t *enc, int utf8, char *str, STRLEN len)
283{
284 encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
285 need (enc, len);
286 memcpy (enc->cur, str, len);
287 enc->cur += len;
223} 288}
224 289
225static void 290static void
226encode_str (enc_t *enc, int utf8, char *str, STRLEN len) 291encode_strref (enc_t *enc, int utf8, char *str, STRLEN len)
227{ 292{
228 if (ecb_expect_false (enc->cbor.flags & F_ALLOW_STRINGREF)) 293 if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
229 { 294 {
230 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1); 295 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
231 296
232 if (SvOK (*svp)) 297 if (SvOK (*svp))
233 { 298 {
234 // already registered, use stringref 299 // already registered, use stringref
235 encode_tag (enc, CBOR_TAG_STRINGREF); 300 encode_tag (enc, CBOR_TAG_STRINGREF);
236 encode_uint (enc, 0x00, SvUV (*svp)); 301 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
237 return; 302 return;
238 } 303 }
239 else if (len >= minimum_string_length (enc->stringref_idx)) 304 else if (len >= minimum_string_length (enc->stringref_idx))
240 { 305 {
241 // register only 306 // register only
242 sv_setuv (*svp, enc->stringref_idx); 307 sv_setuv (*svp, enc->stringref_idx);
243 ++enc->stringref_idx; 308 ++enc->stringref_idx;
244 } 309 }
245 } 310 }
246 311
247 encode_uint (enc, utf8 ? 0x60 : 0x40, len); 312 encode_str (enc, utf8, str, len);
248 need (enc, len);
249 memcpy (enc->cur, str, len);
250 enc->cur += len;
251} 313}
252 314
253static void encode_sv (enc_t *enc, SV *sv); 315static void encode_sv (enc_t *enc, SV *sv);
254 316
255static void 317static void
260 if (enc->depth >= enc->cbor.max_depth) 322 if (enc->depth >= enc->cbor.max_depth)
261 croak (ERR_NESTING_EXCEEDED); 323 croak (ERR_NESTING_EXCEEDED);
262 324
263 ++enc->depth; 325 ++enc->depth;
264 326
265 encode_uint (enc, 0x80, len + 1); 327 encode_uint (enc, MAJOR_ARRAY, len + 1);
266 328
329 if (SvMAGICAL (av))
267 for (i = 0; i <= len; ++i) 330 for (i = 0; i <= len; ++i)
268 { 331 {
269 SV **svp = av_fetch (av, i, 0); 332 SV **svp = av_fetch (av, i, 0);
270 encode_sv (enc, svp ? *svp : &PL_sv_undef); 333 encode_sv (enc, svp ? *svp : &PL_sv_undef);
271 } 334 }
335 else
336 for (i = 0; i <= len; ++i)
337 {
338 SV *sv = AvARRAY (av)[i];
339 encode_sv (enc, sv ? sv : &PL_sv_undef);
340 }
272 341
273 --enc->depth; 342 --enc->depth;
274}
275
276ecb_inline void
277encode_he (enc_t *enc, HE *he)
278{
279} 343}
280 344
281static void 345static void
282encode_hv (enc_t *enc, HV *hv) 346encode_hv (enc_t *enc, HV *hv)
283{ 347{
290 354
291 int pairs = hv_iterinit (hv); 355 int pairs = hv_iterinit (hv);
292 int mg = SvMAGICAL (hv); 356 int mg = SvMAGICAL (hv);
293 357
294 if (mg) 358 if (mg)
295 encode_ch (enc, 0xa0 | 31); 359 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
296 else 360 else
297 encode_uint (enc, 0xa0, pairs); 361 encode_uint (enc, MAJOR_MAP, pairs);
298 362
299 while ((he = hv_iternext (hv))) 363 while ((he = hv_iternext (hv)))
300 { 364 {
301 if (HeKLEN (he) == HEf_SVKEY) 365 if (HeKLEN (he) == HEf_SVKEY)
302 encode_sv (enc, HeSVKEY (he)); 366 encode_sv (enc, HeSVKEY (he));
303 else 367 else
304 encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); 368 encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
305 369
306 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); 370 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
307 } 371 }
308 372
309 if (mg) 373 if (mg)
310 encode_ch (enc, 0xe0 | 31); 374 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
311 375
312 --enc->depth; 376 --enc->depth;
313} 377}
314 378
315// encode objects, arrays and special \0=false and \1=true values. 379// encode objects, arrays and special \0=false and \1=true values.
316static void 380static void
317encode_rv (enc_t *enc, SV *sv) 381encode_rv (enc_t *enc, SV *sv)
318{ 382{
319 SvGETMAGIC (sv); 383 SvGETMAGIC (sv);
320
321 if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)
322 && ecb_expect_false (SvREFCNT (sv) > 1))
323 {
324 if (!enc->shareable)
325 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
326
327 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
328
329 if (SvOK (*svp))
330 {
331 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
332 encode_uint (enc, 0x00, SvUV (*svp));
333 return;
334 }
335 else
336 {
337 sv_setuv (*svp, enc->shareable_idx);
338 ++enc->shareable_idx;
339 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
340 }
341 }
342 384
343 svtype svt = SvTYPE (sv); 385 svtype svt = SvTYPE (sv);
344 386
345 if (ecb_expect_false (SvOBJECT (sv))) 387 if (ecb_expect_false (SvOBJECT (sv)))
346 { 388 {
353 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 395 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
354 ? cbor_tagged_stash 396 ? cbor_tagged_stash
355 : gv_stashpv ("CBOR::XS::Tagged" , 1); 397 : gv_stashpv ("CBOR::XS::Tagged" , 1);
356 398
357 HV *stash = SvSTASH (sv); 399 HV *stash = SvSTASH (sv);
358 GV *method;
359 400
360 if (stash == boolean_stash) 401 if (stash == boolean_stash)
361 encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20); 402 {
403 encode_ch (enc, SvIV (sv) ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE);
404 return;
405 }
362 else if (stash == error_stash) 406 else if (stash == error_stash)
363 encode_ch (enc, 0xe0 | 23); 407 {
408 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
409 return;
410 }
364 else if (stash == tagged_stash) 411 else if (stash == tagged_stash)
365 { 412 {
366 if (svt != SVt_PVAV) 413 if (svt != SVt_PVAV)
367 croak ("encountered CBOR::XS::Tagged object that isn't an array"); 414 croak ("encountered CBOR::XS::Tagged object that isn't an array");
368 415
369 encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); 416 encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
370 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); 417 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1));
418
419 return;
420 }
421 }
422
423 if (ecb_expect_false (SvREFCNT (sv) > 1)
424 && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
425 {
426 if (!enc->shareable)
427 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
428
429 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
430
431 if (SvOK (*svp))
371 } 432 {
433 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
434 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
435 return;
436 }
437 else
438 {
439 sv_setuv (*svp, enc->shareable_idx);
440 ++enc->shareable_idx;
441 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
442 }
443 }
444
445 if (ecb_expect_false (SvOBJECT (sv)))
446 {
447 HV *stash = SvSTASH (sv);
448 GV *method;
449
372 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) 450 if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
373 { 451 {
374 dSP; 452 dSP;
375 453
376 ENTER; SAVETMPS; PUSHMARK (SP); 454 ENTER; SAVETMPS; PUSHMARK (SP);
377 // we re-bless the reference to get overload and other niceties right 455 // we re-bless the reference to get overload and other niceties right
409 // catch this surprisingly common error 487 // catch this surprisingly common error
410 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) 488 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
411 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash)); 489 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
412 490
413 encode_tag (enc, CBOR_TAG_PERL_OBJECT); 491 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
414 encode_uint (enc, 0x80, count + 1); 492 encode_uint (enc, MAJOR_ARRAY, count + 1);
415 encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); 493 encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
416 494
417 while (count) 495 while (count)
418 encode_sv (enc, SP[1 - count--]); 496 encode_sv (enc, SP[1 - count--]);
419 497
420 PUTBACK; 498 PUTBACK;
441{ 519{
442 double nv = SvNVX (sv); 520 double nv = SvNVX (sv);
443 521
444 need (enc, 9); 522 need (enc, 9);
445 523
446 if (ecb_expect_false (nv == (U32)nv)) 524 if (ecb_expect_false (nv == (NV)(U32)nv))
447 encode_uint (enc, 0x00, (U32)nv); 525 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
448 //TODO: maybe I32? 526 //TODO: maybe I32?
449 else if (ecb_expect_false (nv == (float)nv)) 527 else if (ecb_expect_false (nv == (float)nv))
450 { 528 {
451 uint32_t fp = ecb_float_to_binary32 (nv); 529 uint32_t fp = ecb_float_to_binary32 (nv);
452 530
453 *enc->cur++ = 0xe0 | 26; 531 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
454 532
455 if (!ecb_big_endian ()) 533 if (!ecb_big_endian ())
456 fp = ecb_bswap32 (fp); 534 fp = ecb_bswap32 (fp);
457 535
458 memcpy (enc->cur, &fp, 4); 536 memcpy (enc->cur, &fp, 4);
460 } 538 }
461 else 539 else
462 { 540 {
463 uint64_t fp = ecb_double_to_binary64 (nv); 541 uint64_t fp = ecb_double_to_binary64 (nv);
464 542
465 *enc->cur++ = 0xe0 | 27; 543 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
466 544
467 if (!ecb_big_endian ()) 545 if (!ecb_big_endian ())
468 fp = ecb_bswap64 (fp); 546 fp = ecb_bswap64 (fp);
469 547
470 memcpy (enc->cur, &fp, 8); 548 memcpy (enc->cur, &fp, 8);
479 557
480 if (SvPOKp (sv)) 558 if (SvPOKp (sv))
481 { 559 {
482 STRLEN len; 560 STRLEN len;
483 char *str = SvPV (sv, len); 561 char *str = SvPV (sv, len);
484 encode_str (enc, SvUTF8 (sv), str, len); 562 encode_strref (enc, SvUTF8 (sv), str, len);
485 } 563 }
486 else if (SvNOKp (sv)) 564 else if (SvNOKp (sv))
487 encode_nv (enc, sv); 565 encode_nv (enc, sv);
488 else if (SvIOKp (sv)) 566 else if (SvIOKp (sv))
489 { 567 {
490 if (SvIsUV (sv)) 568 if (SvIsUV (sv))
491 encode_uint (enc, 0x00, SvUVX (sv)); 569 encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
492 else if (SvIVX (sv) >= 0) 570 else if (SvIVX (sv) >= 0)
493 encode_uint (enc, 0x00, SvIVX (sv)); 571 encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
494 else 572 else
495 encode_uint (enc, 0x20, -(SvIVX (sv) + 1)); 573 encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
496 } 574 }
497 else if (SvROK (sv)) 575 else if (SvROK (sv))
498 encode_rv (enc, SvRV (sv)); 576 encode_rv (enc, SvRV (sv));
499 else if (!SvOK (sv)) 577 else if (!SvOK (sv))
500 encode_ch (enc, 0xe0 | 22); 578 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
501 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 579 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
502 encode_ch (enc, 0xe0 | 23); 580 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
503 else 581 else
504 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data", 582 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
505 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 583 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
506} 584}
507 585
515 enc.cur = SvPVX (enc.sv); 593 enc.cur = SvPVX (enc.sv);
516 enc.end = SvEND (enc.sv); 594 enc.end = SvEND (enc.sv);
517 595
518 SvPOK_only (enc.sv); 596 SvPOK_only (enc.sv);
519 597
520 if (cbor->flags & F_ALLOW_STRINGREF) 598 if (cbor->flags & F_PACK_STRINGS)
521 { 599 {
522 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE); 600 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
523 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ()); 601 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
524 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ()); 602 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
525 } 603 }
547 CBOR cbor; 625 CBOR cbor;
548 U32 depth; // recursion depth 626 U32 depth; // recursion depth
549 U32 maxdepth; // recursion depth limit 627 U32 maxdepth; // recursion depth limit
550 AV *shareable; 628 AV *shareable;
551 AV *stringref; 629 AV *stringref;
630 SV *decode_tagged;
552} dec_t; 631} dec_t;
553 632
554#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE 633#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE
555 634
556#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data") 635#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data")
559#define DEC_DEC_DEPTH --dec->depth 638#define DEC_DEC_DEPTH --dec->depth
560 639
561static UV 640static UV
562decode_uint (dec_t *dec) 641decode_uint (dec_t *dec)
563{ 642{
564 switch (*dec->cur & 31) 643 U8 m = *dec->cur & MINOR_MASK;
565 { 644 ++dec->cur;
566 case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
567 case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
568 case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
569 return *dec->cur++ & 31;
570 645
571 case 24: 646 if (ecb_expect_true (m < LENGTH_EXT1))
647 return m;
648 else if (ecb_expect_true (m == LENGTH_EXT1))
649 {
572 WANT (2); 650 WANT (1);
573 dec->cur += 2; 651 dec->cur += 1;
574 return dec->cur[-1]; 652 return dec->cur[-1];
575 653 }
576 case 25: 654 else if (ecb_expect_true (m == LENGTH_EXT2))
655 {
577 WANT (3); 656 WANT (2);
578 dec->cur += 3; 657 dec->cur += 2;
579 return (((UV)dec->cur[-2]) << 8) 658 return (((UV)dec->cur[-2]) << 8)
580 | ((UV)dec->cur[-1]); 659 | ((UV)dec->cur[-1]);
581 660 }
582 case 26: 661 else if (ecb_expect_true (m == LENGTH_EXT4))
662 {
583 WANT (5); 663 WANT (4);
584 dec->cur += 5; 664 dec->cur += 4;
585 return (((UV)dec->cur[-4]) << 24) 665 return (((UV)dec->cur[-4]) << 24)
586 | (((UV)dec->cur[-3]) << 16) 666 | (((UV)dec->cur[-3]) << 16)
587 | (((UV)dec->cur[-2]) << 8) 667 | (((UV)dec->cur[-2]) << 8)
588 | ((UV)dec->cur[-1]); 668 | ((UV)dec->cur[-1]);
589 669 }
590 case 27: 670 else if (ecb_expect_true (m == LENGTH_EXT8))
671 {
591 WANT (9); 672 WANT (8);
592 dec->cur += 9; 673 dec->cur += 8;
674
675 return
676#if UVSIZE < 8
677 0
678#else
593 return (((UV)dec->cur[-8]) << 56) 679 (((UV)dec->cur[-8]) << 56)
594 | (((UV)dec->cur[-7]) << 48) 680 | (((UV)dec->cur[-7]) << 48)
595 | (((UV)dec->cur[-6]) << 40) 681 | (((UV)dec->cur[-6]) << 40)
596 | (((UV)dec->cur[-5]) << 32) 682 | (((UV)dec->cur[-5]) << 32)
683#endif
597 | (((UV)dec->cur[-4]) << 24) 684 | (((UV)dec->cur[-4]) << 24)
598 | (((UV)dec->cur[-3]) << 16) 685 | (((UV)dec->cur[-3]) << 16)
599 | (((UV)dec->cur[-2]) << 8) 686 | (((UV)dec->cur[-2]) << 8)
600 | ((UV)dec->cur[-1]); 687 | ((UV)dec->cur[-1]);
601 688 }
602 default: 689 else
603 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 690 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
604 }
605 691
606fail: 692fail:
607 return 0; 693 return 0;
608} 694}
609 695
614{ 700{
615 AV *av = newAV (); 701 AV *av = newAV ();
616 702
617 DEC_INC_DEPTH; 703 DEC_INC_DEPTH;
618 704
619 if ((*dec->cur & 31) == 31) 705 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
620 { 706 {
621 ++dec->cur; 707 ++dec->cur;
622 708
623 for (;;) 709 for (;;)
624 { 710 {
625 WANT (1); 711 WANT (1);
626 712
627 if (*dec->cur == (0xe0 | 31)) 713 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
628 { 714 {
629 ++dec->cur; 715 ++dec->cur;
630 break; 716 break;
631 } 717 }
632 718
635 } 721 }
636 else 722 else
637 { 723 {
638 int i, len = decode_uint (dec); 724 int i, len = decode_uint (dec);
639 725
726 WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
640 av_fill (av, len - 1); 727 av_fill (av, len - 1);
641 728
642 for (i = 0; i < len; ++i) 729 for (i = 0; i < len; ++i)
643 AvARRAY (av)[i] = decode_sv (dec); 730 AvARRAY (av)[i] = decode_sv (dec);
644 } 731 }
657{ 744{
658 // for speed reasons, we specialcase single-string 745 // for speed reasons, we specialcase single-string
659 // byte or utf-8 strings as keys, but only when !stringref 746 // byte or utf-8 strings as keys, but only when !stringref
660 747
661 if (ecb_expect_true (!dec->stringref)) 748 if (ecb_expect_true (!dec->stringref))
662 if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27) 749 if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
663 { 750 {
664 I32 len = decode_uint (dec); 751 I32 len = decode_uint (dec);
665 char *key = (char *)dec->cur; 752 char *key = (char *)dec->cur;
666 753
667 dec->cur += len; 754 dec->cur += len;
668 755
669 if (ecb_expect_false (dec->stringref))
670 av_push (dec->stringref, newSVpvn (key, len));
671
672 hv_store (hv, key, len, decode_sv (dec), 0); 756 hv_store (hv, key, len, decode_sv (dec), 0);
673 757
674 return; 758 return;
675 } 759 }
676 else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27) 760 else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
677 { 761 {
678 I32 len = decode_uint (dec); 762 I32 len = decode_uint (dec);
679 char *key = (char *)dec->cur; 763 char *key = (char *)dec->cur;
680 764
681 dec->cur += len; 765 dec->cur += len;
682 766
683 if (ecb_expect_false (dec->stringref)) 767 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
684 av_push (dec->stringref, newSVpvn_utf8 (key, len, 1)); 768 if (!is_utf8_string (key, len))
769 ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
685 770
686 hv_store (hv, key, -len, decode_sv (dec), 0); 771 hv_store (hv, key, -len, decode_sv (dec), 0);
687 772
688 return; 773 return;
689 } 774 }
691 SV *k = decode_sv (dec); 776 SV *k = decode_sv (dec);
692 SV *v = decode_sv (dec); 777 SV *v = decode_sv (dec);
693 778
694 hv_store_ent (hv, k, v, 0); 779 hv_store_ent (hv, k, v, 0);
695 SvREFCNT_dec (k); 780 SvREFCNT_dec (k);
781
782fail:
783 ;
696} 784}
697 785
698static SV * 786static SV *
699decode_hv (dec_t *dec) 787decode_hv (dec_t *dec)
700{ 788{
701 HV *hv = newHV (); 789 HV *hv = newHV ();
702 790
703 DEC_INC_DEPTH; 791 DEC_INC_DEPTH;
704 792
705 if ((*dec->cur & 31) == 31) 793 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
706 { 794 {
707 ++dec->cur; 795 ++dec->cur;
708 796
709 for (;;) 797 for (;;)
710 { 798 {
711 WANT (1); 799 WANT (1);
712 800
713 if (*dec->cur == (0xe0 | 31)) 801 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
714 { 802 {
715 ++dec->cur; 803 ++dec->cur;
716 break; 804 break;
717 } 805 }
718 806
739static SV * 827static SV *
740decode_str (dec_t *dec, int utf8) 828decode_str (dec_t *dec, int utf8)
741{ 829{
742 SV *sv = 0; 830 SV *sv = 0;
743 831
744 if ((*dec->cur & 31) == 31) 832 if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
745 { 833 {
834 // indefinite length strings
746 ++dec->cur; 835 ++dec->cur;
747 836
837 U8 major = *dec->cur & MAJOR_MISC;
838
748 sv = newSVpvn ("", 0); 839 sv = newSVpvn ("", 0);
749 840
750 // not very fast, and certainly not robust against illegal input
751 for (;;) 841 for (;;)
752 { 842 {
753 WANT (1); 843 WANT (1);
754 844
755 if (*dec->cur == (0xe0 | 31)) 845 if ((*dec->cur - major) > LENGTH_EXT8)
846 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
756 { 847 {
757 ++dec->cur; 848 ++dec->cur;
758 break; 849 break;
759 } 850 }
851 else
852 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
760 853
761 sv_catsv (sv, decode_sv (dec)); 854 STRLEN len = decode_uint (dec);
855
856 WANT (len);
857 sv_catpvn (sv, dec->cur, len);
858 dec->cur += len;
762 } 859 }
763 } 860 }
764 else 861 else
765 { 862 {
766 STRLEN len = decode_uint (dec); 863 STRLEN len = decode_uint (dec);
767 864
768 WANT (len); 865 WANT (len);
769 sv = newSVpvn (dec->cur, len); 866 sv = newSVpvn (dec->cur, len);
770 dec->cur += len; 867 dec->cur += len;
868
869 if (ecb_expect_false (dec->stringref)
870 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
871 av_push (dec->stringref, SvREFCNT_inc_NN (sv));
771 } 872 }
772 873
773 if (utf8) 874 if (utf8)
875 {
876 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
877 if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
878 ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
879
774 SvUTF8_on (sv); 880 SvUTF8_on (sv);
775 881 }
776 if (ecb_expect_false (dec->stringref)
777 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
778 av_push (dec->stringref, SvREFCNT_inc_NN (sv));
779 882
780 return sv; 883 return sv;
781 884
782fail: 885fail:
783 SvREFCNT_dec (sv); 886 SvREFCNT_dec (sv);
815 } 918 }
816 break; 919 break;
817 920
818 case CBOR_TAG_STRINGREF: 921 case CBOR_TAG_STRINGREF:
819 { 922 {
820 if ((*dec->cur >> 5) != 0) 923 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
821 ERR ("corrupted CBOR data (stringref index not an unsigned integer)"); 924 ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
822 925
823 UV idx = decode_uint (dec); 926 UV idx = decode_uint (dec);
824 927
825 if (!dec->stringref || (int)idx > AvFILLp (dec->stringref)) 928 if (!dec->stringref || (int)idx > AvFILLp (dec->stringref))
832 case CBOR_TAG_VALUE_SHAREABLE: 935 case CBOR_TAG_VALUE_SHAREABLE:
833 { 936 {
834 if (ecb_expect_false (!dec->shareable)) 937 if (ecb_expect_false (!dec->shareable))
835 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ()); 938 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
836 939
940 if (dec->cbor.flags & F_ALLOW_CYCLES)
941 {
837 sv = newSV (0); 942 sv = newSV (0);
838 av_push (dec->shareable, SvREFCNT_inc_NN (sv)); 943 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
839 944
840 SV *osv = decode_sv (dec); 945 SV *osv = decode_sv (dec);
841 sv_setsv (sv, osv); 946 sv_setsv (sv, osv);
842 SvREFCNT_dec_NN (osv); 947 SvREFCNT_dec_NN (osv);
948 }
949 else
950 {
951 av_push (dec->shareable, &PL_sv_undef);
952 int idx = AvFILLp (dec->shareable);
953 sv = decode_sv (dec);
954 av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
955 }
843 } 956 }
844 break; 957 break;
845 958
846 case CBOR_TAG_VALUE_SHAREDREF: 959 case CBOR_TAG_VALUE_SHAREDREF:
847 { 960 {
848 if ((*dec->cur >> 5) != 0) 961 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
849 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)"); 962 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
850 963
851 UV idx = decode_uint (dec); 964 UV idx = decode_uint (dec);
852 965
853 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable)) 966 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable))
854 ERR ("corrupted CBOR data (sharedref index out of bounds)"); 967 ERR ("corrupted CBOR data (sharedref index out of bounds)");
855 968
856 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]); 969 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
970
971 if (sv == &PL_sv_undef)
972 ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
857 } 973 }
858 break; 974 break;
859 975
860 case CBOR_TAG_PERL_OBJECT: 976 case CBOR_TAG_PERL_OBJECT:
861 { 977 {
910 1026
911 default: 1027 default:
912 { 1028 {
913 sv = decode_sv (dec); 1029 sv = decode_sv (dec);
914 1030
1031 dSP;
1032 ENTER; SAVETMPS; PUSHMARK (SP);
1033 EXTEND (SP, 2);
1034 PUSHs (newSVuv (tag));
1035 PUSHs (sv);
1036
1037 PUTBACK;
1038 int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL);
1039 SPAGAIN;
1040
1041 if (SvTRUE (ERRSV))
1042 {
1043 FREETMPS; LEAVE;
1044 ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));
1045 }
1046
1047 if (count)
1048 {
1049 SvREFCNT_dec (sv);
1050 sv = SvREFCNT_inc (POPs);
1051 }
1052 else
1053 {
915 AV *av = newAV (); 1054 AV *av = newAV ();
916 av_push (av, newSVuv (tag)); 1055 av_push (av, newSVuv (tag));
917 av_push (av, sv); 1056 av_push (av, sv);
918 1057
919 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 1058 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
920 ? cbor_tagged_stash 1059 ? cbor_tagged_stash
921 : gv_stashpv ("CBOR::XS::Tagged" , 1); 1060 : gv_stashpv ("CBOR::XS::Tagged" , 1);
922
923 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash); 1061 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
1062 }
1063
1064 PUTBACK;
1065
1066 FREETMPS; LEAVE;
924 } 1067 }
925 break; 1068 break;
926 } 1069 }
927 1070
928 return sv; 1071 return sv;
935static SV * 1078static SV *
936decode_sv (dec_t *dec) 1079decode_sv (dec_t *dec)
937{ 1080{
938 WANT (1); 1081 WANT (1);
939 1082
940 switch (*dec->cur >> 5) 1083 switch (*dec->cur >> MAJOR_SHIFT)
941 { 1084 {
942 case 0: // unsigned int 1085 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
943 return newSVuv (decode_uint (dec)); 1086 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
944 case 1: // negative int 1087 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
945 return newSViv (-1 - (IV)decode_uint (dec)); 1088 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
946 case 2: // octet string 1089 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
947 return decode_str (dec, 0); 1090 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
948 case 3: // utf-8 string 1091 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
949 return decode_str (dec, 1); 1092
950 case 4: // array 1093 case MAJOR_MISC >> MAJOR_SHIFT:
951 return decode_av (dec);
952 case 5: // map
953 return decode_hv (dec);
954 case 6: // tag
955 return decode_tagged (dec);
956 case 7: // misc
957 switch (*dec->cur++ & 31) 1094 switch (*dec->cur++ & MINOR_MASK)
958 { 1095 {
959 case 20: 1096 case SIMPLE_FALSE:
960#if CBOR_SLOW 1097#if CBOR_SLOW
961 types_false = get_bool ("Types::Serialiser::false"); 1098 types_false = get_bool ("Types::Serialiser::false");
962#endif 1099#endif
963 return newSVsv (types_false); 1100 return newSVsv (types_false);
964 case 21: 1101 case SIMPLE_TRUE:
965#if CBOR_SLOW 1102#if CBOR_SLOW
966 types_true = get_bool ("Types::Serialiser::true"); 1103 types_true = get_bool ("Types::Serialiser::true");
967#endif 1104#endif
968 return newSVsv (types_true); 1105 return newSVsv (types_true);
969 case 22: 1106 case SIMPLE_NULL:
970 return newSVsv (&PL_sv_undef); 1107 return newSVsv (&PL_sv_undef);
971 case 23: 1108 case SIMPLE_UNDEF:
972#if CBOR_SLOW 1109#if CBOR_SLOW
973 types_error = get_bool ("Types::Serialiser::error"); 1110 types_error = get_bool ("Types::Serialiser::error");
974#endif 1111#endif
975 return newSVsv (types_error); 1112 return newSVsv (types_error);
976 1113
977 case 25: 1114 case MISC_FLOAT16:
978 { 1115 {
979 WANT (2); 1116 WANT (2);
980 1117
981 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1118 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
982 dec->cur += 2; 1119 dec->cur += 2;
983 1120
984 return newSVnv (ecb_binary16_to_float (fp)); 1121 return newSVnv (ecb_binary16_to_float (fp));
985 } 1122 }
986 1123
987 case 26: 1124 case MISC_FLOAT32:
988 { 1125 {
989 uint32_t fp; 1126 uint32_t fp;
990 WANT (4); 1127 WANT (4);
991 memcpy (&fp, dec->cur, 4); 1128 memcpy (&fp, dec->cur, 4);
992 dec->cur += 4; 1129 dec->cur += 4;
995 fp = ecb_bswap32 (fp); 1132 fp = ecb_bswap32 (fp);
996 1133
997 return newSVnv (ecb_binary32_to_float (fp)); 1134 return newSVnv (ecb_binary32_to_float (fp));
998 } 1135 }
999 1136
1000 case 27: 1137 case MISC_FLOAT64:
1001 { 1138 {
1002 uint64_t fp; 1139 uint64_t fp;
1003 WANT (8); 1140 WANT (8);
1004 memcpy (&fp, dec->cur, 8); 1141 memcpy (&fp, dec->cur, 8);
1005 dec->cur += 8; 1142 dec->cur += 8;
1008 fp = ecb_bswap64 (fp); 1145 fp = ecb_bswap64 (fp);
1009 1146
1010 return newSVnv (ecb_binary64_to_double (fp)); 1147 return newSVnv (ecb_binary64_to_double (fp));
1011 } 1148 }
1012 1149
1013 // 0..19 unassigned 1150 // 0..19 unassigned simple
1014 // 24 reserved + unassigned (reserved values are not encodable) 1151 // 24 reserved + unassigned simple (reserved values are not encodable)
1152 // 28-30 unassigned misc
1153 // 31 break code
1015 default: 1154 default:
1016 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1155 ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1017 } 1156 }
1018 1157
1019 break; 1158 break;
1020 } 1159 }
1021 1160
1048 if (dec.cur != dec.end && !dec.err) 1187 if (dec.cur != dec.end && !dec.err)
1049 dec.err = "garbage after CBOR object"; 1188 dec.err = "garbage after CBOR object";
1050 1189
1051 if (dec.err) 1190 if (dec.err)
1052 { 1191 {
1192 if (dec.shareable)
1193 {
1194 // need to break cyclic links, which whould all be in shareable
1195 int i;
1196 SV **svp;
1197
1198 for (i = av_len (dec.shareable) + 1; i--; )
1199 if ((svp = av_fetch (dec.shareable, i, 0)))
1200 sv_setsv (*svp, &PL_sv_undef);
1201 }
1202
1053 SvREFCNT_dec (sv); 1203 SvREFCNT_dec (sv);
1054 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur); 1204 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur);
1055 } 1205 }
1056 1206
1057 sv = sv_2mortal (sv); 1207 sv = sv_2mortal (sv);
1058 1208
1059 return sv; 1209 return sv;
1060} 1210}
1061 1211
1212/////////////////////////////////////////////////////////////////////////////
1213// incremental parser
1214
1215#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
1216
1217// returns 0 for notyet, 1 for success or error
1218static int
1219incr_parse (CBOR *self, SV *cborstr)
1220{
1221 STRLEN cur;
1222 SvPV (cborstr, cur);
1223
1224 while (ecb_expect_true (self->incr_need <= cur))
1225 {
1226 // table of integer count bytes
1227 static I8 incr_len[MINOR_MASK + 1] = {
1228 0, 0, 0, 0, 0, 0, 0, 0,
1229 0, 0, 0, 0, 0, 0, 0, 0,
1230 0, 0, 0, 0, 0, 0, 0, 0,
1231 1, 2, 4, 8,-1,-1,-1,-2
1232 };
1233
1234 const U8 *p = SvPVX (cborstr) + self->incr_pos;
1235 U8 m = *p & MINOR_MASK;
1236 IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1237 I8 ilen = incr_len[m];
1238
1239 self->incr_need = self->incr_pos + 1;
1240
1241 if (ecb_expect_false (ilen < 0))
1242 {
1243 if (m != MINOR_INDEF)
1244 return 1; // error
1245
1246 if (*p == (MAJOR_MISC | MINOR_INDEF))
1247 {
1248 if (count >= 0)
1249 return 1; // error
1250
1251 count = 1;
1252 }
1253 else
1254 {
1255 av_push (self->incr_count, newSViv (-1)); //TODO: nest
1256 count = -1;
1257 }
1258 }
1259 else
1260 {
1261 self->incr_need += ilen;
1262 if (ecb_expect_false (self->incr_need > cur))
1263 return 0;
1264
1265 int major = *p >> MAJOR_SHIFT;
1266
1267 switch (major)
1268 {
1269 case MAJOR_TAG >> MAJOR_SHIFT:
1270 ++count; // tags merely prefix another value
1271 break;
1272
1273 case MAJOR_BYTES >> MAJOR_SHIFT:
1274 case MAJOR_TEXT >> MAJOR_SHIFT:
1275 case MAJOR_ARRAY >> MAJOR_SHIFT:
1276 case MAJOR_MAP >> MAJOR_SHIFT:
1277 {
1278 UV len;
1279
1280 if (ecb_expect_false (ilen))
1281 {
1282 len = 0;
1283
1284 do {
1285 len = (len << 8) | *++p;
1286 } while (--ilen);
1287 }
1288 else
1289 len = m;
1290
1291 switch (major)
1292 {
1293 case MAJOR_BYTES >> MAJOR_SHIFT:
1294 case MAJOR_TEXT >> MAJOR_SHIFT:
1295 self->incr_need += len;
1296 if (ecb_expect_false (self->incr_need > cur))
1297 return 0;
1298
1299 break;
1300
1301 case MAJOR_MAP >> MAJOR_SHIFT:
1302 len <<= 1;
1303 case MAJOR_ARRAY >> MAJOR_SHIFT:
1304 if (len)
1305 {
1306 av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
1307 count = len + 1;
1308 }
1309 break;
1310 }
1311 }
1312 }
1313 }
1314
1315 self->incr_pos = self->incr_need;
1316
1317 if (count > 0)
1318 {
1319 while (!--count)
1320 {
1321 if (!AvFILLp (self->incr_count))
1322 return 1; // done
1323
1324 SvREFCNT_dec_NN (av_pop (self->incr_count));
1325 count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1326 }
1327
1328 SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
1329 }
1330 }
1331
1332 return 0;
1333}
1334
1335
1062///////////////////////////////////////////////////////////////////////////// 1336/////////////////////////////////////////////////////////////////////////////
1063// XS interface functions 1337// XS interface functions
1064 1338
1065MODULE = CBOR::XS PACKAGE = CBOR::XS 1339MODULE = CBOR::XS PACKAGE = CBOR::XS
1066 1340
1073 types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1); 1347 types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1);
1074 1348
1075 types_true = get_bool ("Types::Serialiser::true" ); 1349 types_true = get_bool ("Types::Serialiser::true" );
1076 types_false = get_bool ("Types::Serialiser::false"); 1350 types_false = get_bool ("Types::Serialiser::false");
1077 types_error = get_bool ("Types::Serialiser::error"); 1351 types_error = get_bool ("Types::Serialiser::error");
1352
1353 default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1078 1354
1079 sv_cbor = newSVpv ("CBOR", 0); 1355 sv_cbor = newSVpv ("CBOR", 0);
1080 SvREADONLY_on (sv_cbor); 1356 SvREADONLY_on (sv_cbor);
1081} 1357}
1082 1358
1104void shrink (CBOR *self, int enable = 1) 1380void shrink (CBOR *self, int enable = 1)
1105 ALIAS: 1381 ALIAS:
1106 shrink = F_SHRINK 1382 shrink = F_SHRINK
1107 allow_unknown = F_ALLOW_UNKNOWN 1383 allow_unknown = F_ALLOW_UNKNOWN
1108 allow_sharing = F_ALLOW_SHARING 1384 allow_sharing = F_ALLOW_SHARING
1109 allow_stringref = F_ALLOW_STRINGREF 1385 allow_cycles = F_ALLOW_CYCLES
1386 pack_strings = F_PACK_STRINGS
1387 validate_utf8 = F_VALIDATE_UTF8
1110 PPCODE: 1388 PPCODE:
1111{ 1389{
1112 if (enable) 1390 if (enable)
1113 self->flags |= ix; 1391 self->flags |= ix;
1114 else 1392 else
1120void get_shrink (CBOR *self) 1398void get_shrink (CBOR *self)
1121 ALIAS: 1399 ALIAS:
1122 get_shrink = F_SHRINK 1400 get_shrink = F_SHRINK
1123 get_allow_unknown = F_ALLOW_UNKNOWN 1401 get_allow_unknown = F_ALLOW_UNKNOWN
1124 get_allow_sharing = F_ALLOW_SHARING 1402 get_allow_sharing = F_ALLOW_SHARING
1125 get_allow_stringref = F_ALLOW_STRINGREF 1403 get_allow_cycles = F_ALLOW_CYCLES
1404 get_pack_strings = F_PACK_STRINGS
1405 get_validate_utf8 = F_VALIDATE_UTF8
1126 PPCODE: 1406 PPCODE:
1127 XPUSHs (boolSV (self->flags & ix)); 1407 XPUSHs (boolSV (self->flags & ix));
1128 1408
1129void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1409void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1130 PPCODE: 1410 PPCODE:
1143 XPUSHs (ST (0)); 1423 XPUSHs (ST (0));
1144 1424
1145int get_max_size (CBOR *self) 1425int get_max_size (CBOR *self)
1146 CODE: 1426 CODE:
1147 RETVAL = self->max_size; 1427 RETVAL = self->max_size;
1428 OUTPUT:
1429 RETVAL
1430
1431void filter (CBOR *self, SV *filter = 0)
1432 PPCODE:
1433 SvREFCNT_dec (self->filter);
1434 self->filter = filter ? newSVsv (filter) : filter;
1435 XPUSHs (ST (0));
1436
1437SV *get_filter (CBOR *self)
1438 CODE:
1439 RETVAL = self->filter ? self->filter : NEWSV (0, 0);
1148 OUTPUT: 1440 OUTPUT:
1149 RETVAL 1441 RETVAL
1150 1442
1151void encode (CBOR *self, SV *scalar) 1443void encode (CBOR *self, SV *scalar)
1152 PPCODE: 1444 PPCODE:
1167 EXTEND (SP, 2); 1459 EXTEND (SP, 2);
1168 PUSHs (sv); 1460 PUSHs (sv);
1169 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1461 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1170} 1462}
1171 1463
1464void incr_parse (CBOR *self, SV *cborstr)
1465 ALIAS:
1466 incr_parse_multiple = 1
1467 PPCODE:
1468{
1469 if (SvUTF8 (cborstr))
1470 sv_utf8_downgrade (cborstr, 0);
1471
1472 if (!self->incr_count)
1473 {
1474 self->incr_count = newAV ();
1475 self->incr_pos = 0;
1476 self->incr_need = 1;
1477
1478 av_push (self->incr_count, newSViv (1));
1479 }
1480
1481 do
1482 {
1483 if (!incr_parse (self, cborstr))
1484 {
1485 if (self->incr_need > self->max_size && self->max_size)
1486 croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
1487 (unsigned long)self->incr_need, (unsigned long)self->max_size);
1488
1489 break;
1490 }
1491
1492 SV *sv;
1493 char *offset;
1494
1495 PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
1496 XPUSHs (sv);
1497
1498 sv_chop (cborstr, offset);
1499
1500 av_clear (self->incr_count);
1501 av_push (self->incr_count, newSViv (1));
1502
1503 self->incr_pos = 0;
1504 self->incr_need = self->incr_pos + 1;
1505 }
1506 while (ix);
1507}
1508
1509void incr_reset (CBOR *self)
1510 CODE:
1511{
1512 SvREFCNT_dec (self->incr_count);
1513 self->incr_count = 0;
1514}
1515
1516void DESTROY (CBOR *self)
1517 PPCODE:
1518 cbor_free (self);
1519
1172PROTOTYPES: ENABLE 1520PROTOTYPES: ENABLE
1173 1521
1174void encode_cbor (SV *scalar) 1522void encode_cbor (SV *scalar)
1523 ALIAS:
1524 encode_cbor = 0
1525 encode_cbor_sharing = F_ALLOW_SHARING
1175 PPCODE: 1526 PPCODE:
1176{ 1527{
1177 CBOR cbor; 1528 CBOR cbor;
1178 cbor_init (&cbor); 1529 cbor_init (&cbor);
1530 cbor.flags |= ix;
1179 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1531 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1180 XPUSHs (scalar); 1532 XPUSHs (scalar);
1181} 1533}
1182 1534
1183void decode_cbor (SV *cborstr) 1535void decode_cbor (SV *cborstr)

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines