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 1.25 by root, Fri Nov 22 05:54:07 2013 UTC vs.
Revision 1.63 by root, Sat Nov 26 04:50:58 2016 UTC

6#include <string.h> 6#include <string.h>
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#include <inttypes.h>
11 12
13#define ECB_NO_THREADS 1
12#include "ecb.h" 14#include "ecb.h"
13 15
14// compatibility with perl <5.18 16// compatibility with perl <5.18
15#ifndef HvNAMELEN_get 17#ifndef HvNAMELEN_get
16# define HvNAMELEN_get(hv) strlen (HvNAME (hv)) 18# define HvNAMELEN_get(hv) strlen (HvNAME (hv))
19# define HvNAMELEN(hv) HvNAMELEN_get (hv) 21# define HvNAMELEN(hv) HvNAMELEN_get (hv)
20#endif 22#endif
21#ifndef HvNAMEUTF8 23#ifndef HvNAMEUTF8
22# define HvNAMEUTF8(hv) 0 24# define HvNAMEUTF8(hv) 0
23#endif 25#endif
26#ifndef SvREFCNT_inc_NN
27# define SvREFCNT_inc_NN(sv) SvREFCNT_inc (sv)
28#endif
29#ifndef SvREFCNT_dec_NN
30# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
31#endif
32
33// known major and minor types
34enum cbor_type
35{
36 MAJOR_SHIFT = 5,
37 MINOR_MASK = 0x1f,
38
39 MAJOR_POS_INT = 0 << MAJOR_SHIFT,
40 MAJOR_NEG_INT = 1 << MAJOR_SHIFT,
41 MAJOR_BYTES = 2 << MAJOR_SHIFT,
42 MAJOR_TEXT = 3 << MAJOR_SHIFT,
43 MAJOR_ARRAY = 4 << MAJOR_SHIFT,
44 MAJOR_MAP = 5 << MAJOR_SHIFT,
45 MAJOR_TAG = 6 << MAJOR_SHIFT,
46 MAJOR_MISC = 7 << MAJOR_SHIFT,
47
48 // INT/STRING/ARRAY/MAP subtypes
49 LENGTH_EXT1 = 24,
50 LENGTH_EXT2 = 25,
51 LENGTH_EXT4 = 26,
52 LENGTH_EXT8 = 27,
53
54 // SIMPLE types (effectively MISC subtypes)
55 SIMPLE_FALSE = 20,
56 SIMPLE_TRUE = 21,
57 SIMPLE_NULL = 22,
58 SIMPLE_UNDEF = 23,
59
60 // MISC subtype (unused)
61 MISC_EXT1 = 24,
62 MISC_FLOAT16 = 25,
63 MISC_FLOAT32 = 26,
64 MISC_FLOAT64 = 27,
65
66 // BYTES/TEXT/ARRAY/MAP
67 MINOR_INDEF = 31,
68};
24 69
25// known tags 70// known tags
26enum cbor_tag 71enum cbor_tag
27{ 72{
28 // inofficial extensions (pending iana registration) 73 // extensions
74 CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
29 CBOR_TAG_PERL_OBJECT = 24, // http://cbor.schmorp.de/perl-object 75 CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
30 CBOR_TAG_GENERIC_OBJECT = 25, // http://cbor.schmorp.de/generic-object 76 CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
31 CBOR_TAG_VALUE_SHAREABLE = 26, // http://cbor.schmorp.de/value-sharing 77 CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
32 CBOR_TAG_VALUE_SHAREDREF = 27, // http://cbor.schmorp.de/value-sharing 78 CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
33 CBOR_TAG_STRINGREF_NAMESPACE = 65537, // http://cbor.schmorp.de/stringref 79 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 80 CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
36 81
37 // rfc7049 82 // rfc7049
38 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8 83 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
39 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any 84 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
40 CBOR_TAG_POS_BIGNUM = 2, // byte string 85 CBOR_TAG_POS_BIGNUM = 2, // byte string
41 CBOR_TAG_NEG_BIGNUM = 3, // byte string 86 CBOR_TAG_NEG_BIGNUM = 3, // byte string
42 CBOR_TAG_DECIMAL = 4, // decimal fraction, array 87 CBOR_TAG_DECIMAL = 4, // decimal fraction, array
43 CBOR_TAG_BIGFLOAT = 5, // array 88 CBOR_TAG_BIGFLOAT = 5, // array
44 89
45 CBOR_TAG_CONV_B64U = 21, // base64url, any 90 CBOR_TAG_CONV_B64U = 21, // base64url, any
46 CBOR_TAG_CONV_B64 = 22, // base64, any 91 CBOR_TAG_CONV_B64 = 22, // base64, any
47 CBOR_TAG_CONV_HEX = 23, // base16, any 92 CBOR_TAG_CONV_HEX = 23, // base16, any
48 CBOR_TAG_CBOR = 24, // embedded cbor, byte string 93 CBOR_TAG_CBOR = 24, // embedded cbor, byte string
49 94
50 CBOR_TAG_URI = 32, // URI rfc3986, utf-8 95 CBOR_TAG_URI = 32, // URI rfc3986, utf-8
51 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8 96 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
52 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8 97 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
53 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8 98 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
54 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8 99 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
55 100
56 CBOR_TAG_MAGIC = 55799 // self-describe cbor 101 CBOR_TAG_MAGIC = 55799, // self-describe cbor
57}; 102};
58 103
59#define F_SHRINK 0x00000001UL 104#define F_SHRINK 0x00000001UL
60#define F_ALLOW_UNKNOWN 0x00000002UL 105#define F_ALLOW_UNKNOWN 0x00000002UL
61#define F_ALLOW_SHARING 0x00000004UL //TODO 106#define F_ALLOW_SHARING 0x00000004UL
62#define F_ALLOW_STRINGREF 0x00000008UL //TODO 107#define F_ALLOW_CYCLES 0x00000008UL
108#define F_FORBID_OBJECTS 0x00000010UL
109#define F_PACK_STRINGS 0x00000020UL
110#define F_TEXT_KEYS 0x00000040UL
111#define F_TEXT_STRINGS 0x00000080UL
112#define F_VALIDATE_UTF8 0x00000100UL
63 113
64#define INIT_SIZE 32 // initial scalar size to be allocated 114#define INIT_SIZE 32 // initial scalar size to be allocated
65 115
66#define SB do { 116#define SB do {
67#define SE } while (0) 117#define SE } while (0)
79# define CBOR_SLOW 0 129# define CBOR_SLOW 0
80# define CBOR_STASH cbor_stash 130# define CBOR_STASH cbor_stash
81#endif 131#endif
82 132
83static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS:: 133static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS::
84static SV *types_true, *types_false, *types_error, *sv_cbor; 134static SV *types_true, *types_false, *types_error, *sv_cbor, *default_filter;
85 135
86typedef struct { 136typedef struct {
87 U32 flags; 137 U32 flags;
88 U32 max_depth; 138 U32 max_depth;
89 STRLEN max_size; 139 STRLEN max_size;
140 SV *filter;
141
142 // for the incremental parser
143 STRLEN incr_pos; // the current offset into the text
144 STRLEN incr_need; // minimum bytes needed to decode
145 AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef.
90} CBOR; 146} CBOR;
91 147
92ecb_inline void 148ecb_inline void
93cbor_init (CBOR *cbor) 149cbor_init (CBOR *cbor)
94{ 150{
95 Zero (cbor, 1, CBOR); 151 Zero (cbor, 1, CBOR);
96 cbor->max_depth = 512; 152 cbor->max_depth = 512;
153}
154
155ecb_inline void
156cbor_free (CBOR *cbor)
157{
158 SvREFCNT_dec (cbor->filter);
159 SvREFCNT_dec (cbor->incr_count);
97} 160}
98 161
99///////////////////////////////////////////////////////////////////////////// 162/////////////////////////////////////////////////////////////////////////////
100// utility functions 163// utility functions
101 164
127 190
128// minimum length of a string to be registered for stringref 191// minimum length of a string to be registered for stringref
129ecb_inline int 192ecb_inline int
130minimum_string_length (UV idx) 193minimum_string_length (UV idx)
131{ 194{
132 return idx > 23 195 return idx <= 23 ? 3
133 ? idx > 0xffU 196 : idx <= 0xffU ? 4
134 ? idx > 0xffffU 197 : idx <= 0xffffU ? 5
135 ? idx > 0xffffffffU 198 : idx <= 0xffffffffU ? 7
136 ? 7 199 : 11;
137 : 6
138 : 5
139 : 4
140 : 3;
141} 200}
142 201
143///////////////////////////////////////////////////////////////////////////// 202/////////////////////////////////////////////////////////////////////////////
144// encoder 203// encoder
145 204
158} enc_t; 217} enc_t;
159 218
160ecb_inline void 219ecb_inline void
161need (enc_t *enc, STRLEN len) 220need (enc_t *enc, STRLEN len)
162{ 221{
163 if (ecb_expect_false (enc->cur + len >= enc->end)) 222 if (ecb_expect_false ((uintptr_t)(enc->end - enc->cur) < len))
164 { 223 {
165 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv); 224 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv);
166 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1); 225 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1);
167 enc->cur = SvPVX (enc->sv) + cur; 226 enc->cur = SvPVX (enc->sv) + cur;
168 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1; 227 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;
179static void 238static void
180encode_uint (enc_t *enc, int major, UV len) 239encode_uint (enc_t *enc, int major, UV len)
181{ 240{
182 need (enc, 9); 241 need (enc, 9);
183 242
184 if (len < 24) 243 if (ecb_expect_true (len < LENGTH_EXT1))
185 *enc->cur++ = major | len; 244 *enc->cur++ = major | len;
186 else if (len <= 0xff) 245 else if (ecb_expect_true (len <= 0xffU))
187 { 246 {
188 *enc->cur++ = major | 24; 247 *enc->cur++ = major | LENGTH_EXT1;
189 *enc->cur++ = len; 248 *enc->cur++ = len;
190 } 249 }
191 else if (len <= 0xffff) 250 else if (len <= 0xffffU)
192 { 251 {
193 *enc->cur++ = major | 25; 252 *enc->cur++ = major | LENGTH_EXT2;
194 *enc->cur++ = len >> 8; 253 *enc->cur++ = len >> 8;
195 *enc->cur++ = len; 254 *enc->cur++ = len;
196 } 255 }
197 else if (len <= 0xffffffff) 256 else if (len <= 0xffffffffU)
198 { 257 {
199 *enc->cur++ = major | 26; 258 *enc->cur++ = major | LENGTH_EXT4;
200 *enc->cur++ = len >> 24; 259 *enc->cur++ = len >> 24;
201 *enc->cur++ = len >> 16; 260 *enc->cur++ = len >> 16;
202 *enc->cur++ = len >> 8; 261 *enc->cur++ = len >> 8;
203 *enc->cur++ = len; 262 *enc->cur++ = len;
204 } 263 }
205 else 264 else
206 { 265 {
207 *enc->cur++ = major | 27; 266 *enc->cur++ = major | LENGTH_EXT8;
208 *enc->cur++ = len >> 56; 267 *enc->cur++ = len >> 56;
209 *enc->cur++ = len >> 48; 268 *enc->cur++ = len >> 48;
210 *enc->cur++ = len >> 40; 269 *enc->cur++ = len >> 40;
211 *enc->cur++ = len >> 32; 270 *enc->cur++ = len >> 32;
212 *enc->cur++ = len >> 24; 271 *enc->cur++ = len >> 24;
217} 276}
218 277
219ecb_inline void 278ecb_inline void
220encode_tag (enc_t *enc, UV tag) 279encode_tag (enc_t *enc, UV tag)
221{ 280{
222 encode_uint (enc, 0xc0, tag); 281 encode_uint (enc, MAJOR_TAG, tag);
223} 282}
224 283
225static void 284// exceptional (hopefully) slow path for byte strings that need to be utf8-encoded
285ecb_noinline static void
226encode_str (enc_t *enc, int utf8, char *str, STRLEN len) 286encode_str_utf8 (enc_t *enc, int utf8, char *str, STRLEN len)
227{ 287{
288 STRLEN ulen = len;
289 U8 *p, *pend = (U8 *)str + len;
290
291 for (p = (U8 *)str; p < pend; ++p)
292 ulen += *p >> 7; // count set high bits
293
294 encode_uint (enc, MAJOR_TEXT, ulen);
295
296 need (enc, ulen);
297 for (p = (U8 *)str; p < pend; ++p)
298 if (*p < 0x80)
299 *enc->cur++ = *p;
300 else
301 {
302 *enc->cur++ = 0xc0 + (*p >> 6);
303 *enc->cur++ = 0x80 + (*p & 63);
304 }
305}
306
307ecb_inline void
308encode_str (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
309{
310 if (ecb_expect_false (upgrade_utf8))
311 if (!utf8)
312 {
313 encode_str_utf8 (enc, utf8, str, len);
314 return;
315 }
316
317 encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
318 need (enc, len);
319 memcpy (enc->cur, str, len);
320 enc->cur += len;
321}
322
323ecb_inline void
324encode_strref (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
325{
228 if (ecb_expect_false (enc->cbor.flags & F_ALLOW_STRINGREF)) 326 if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
229 { 327 {
230 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1); 328 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
231 329
232 if (SvOK (*svp)) 330 if (SvOK (*svp))
233 { 331 {
234 // already registered, use stringref 332 // already registered, use stringref
235 encode_tag (enc, CBOR_TAG_STRINGREF); 333 encode_tag (enc, CBOR_TAG_STRINGREF);
236 encode_uint (enc, 0x00, SvUV (*svp)); 334 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
237 return; 335 return;
238 } 336 }
239 else if (len >= minimum_string_length (enc->stringref_idx)) 337 else if (len >= minimum_string_length (enc->stringref_idx))
240 { 338 {
241 // register only 339 // register only
242 sv_setuv (*svp, enc->stringref_idx); 340 sv_setuv (*svp, enc->stringref_idx);
243 ++enc->stringref_idx; 341 ++enc->stringref_idx;
244 } 342 }
245 } 343 }
246 344
247 encode_uint (enc, utf8 ? 0x60 : 0x40, len); 345 encode_str (enc, upgrade_utf8, utf8, str, len);
248 need (enc, len);
249 memcpy (enc->cur, str, len);
250 enc->cur += len;
251} 346}
252 347
253static void encode_sv (enc_t *enc, SV *sv); 348static void encode_sv (enc_t *enc, SV *sv);
254 349
255static void 350static void
260 if (enc->depth >= enc->cbor.max_depth) 355 if (enc->depth >= enc->cbor.max_depth)
261 croak (ERR_NESTING_EXCEEDED); 356 croak (ERR_NESTING_EXCEEDED);
262 357
263 ++enc->depth; 358 ++enc->depth;
264 359
265 encode_uint (enc, 0x80, len + 1); 360 encode_uint (enc, MAJOR_ARRAY, len + 1);
266 361
362 if (ecb_expect_false (SvMAGICAL (av)))
267 for (i = 0; i <= len; ++i) 363 for (i = 0; i <= len; ++i)
268 { 364 {
269 SV **svp = av_fetch (av, i, 0); 365 SV **svp = av_fetch (av, i, 0);
270 encode_sv (enc, svp ? *svp : &PL_sv_undef); 366 encode_sv (enc, svp ? *svp : &PL_sv_undef);
271 } 367 }
368 else
369 for (i = 0; i <= len; ++i)
370 {
371 SV *sv = AvARRAY (av)[i];
372 encode_sv (enc, sv ? sv : &PL_sv_undef);
373 }
272 374
273 --enc->depth; 375 --enc->depth;
274}
275
276ecb_inline void
277encode_he (enc_t *enc, HE *he)
278{
279} 376}
280 377
281static void 378static void
282encode_hv (enc_t *enc, HV *hv) 379encode_hv (enc_t *enc, HV *hv)
283{ 380{
289 ++enc->depth; 386 ++enc->depth;
290 387
291 int pairs = hv_iterinit (hv); 388 int pairs = hv_iterinit (hv);
292 int mg = SvMAGICAL (hv); 389 int mg = SvMAGICAL (hv);
293 390
294 if (mg) 391 if (ecb_expect_false (mg))
295 encode_ch (enc, 0xa0 | 31); 392 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
296 else 393 else
297 encode_uint (enc, 0xa0, pairs); 394 encode_uint (enc, MAJOR_MAP, pairs);
298 395
299 while ((he = hv_iternext (hv))) 396 while ((he = hv_iternext (hv)))
300 { 397 {
301 if (HeKLEN (he) == HEf_SVKEY) 398 if (HeKLEN (he) == HEf_SVKEY)
302 encode_sv (enc, HeSVKEY (he)); 399 encode_sv (enc, HeSVKEY (he));
303 else 400 else
304 encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); 401 encode_strref (enc, enc->cbor.flags & (F_TEXT_KEYS | F_TEXT_STRINGS), HeKUTF8 (he), HeKEY (he), HeKLEN (he));
305 402
306 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); 403 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
307 } 404 }
308 405
309 if (mg) 406 if (ecb_expect_false (mg))
310 encode_ch (enc, 0xe0 | 31); 407 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
311 408
312 --enc->depth; 409 --enc->depth;
313} 410}
314 411
315// encode objects, arrays and special \0=false and \1=true values. 412// encode objects, arrays and special \0=false and \1=true values.
316static void 413static void
317encode_rv (enc_t *enc, SV *sv) 414encode_rv (enc_t *enc, SV *sv)
318{ 415{
319 SvGETMAGIC (sv); 416 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 }
342 417
343 svtype svt = SvTYPE (sv); 418 svtype svt = SvTYPE (sv);
344 419
345 if (ecb_expect_false (SvOBJECT (sv))) 420 if (ecb_expect_false (SvOBJECT (sv)))
346 { 421 {
353 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 428 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
354 ? cbor_tagged_stash 429 ? cbor_tagged_stash
355 : gv_stashpv ("CBOR::XS::Tagged" , 1); 430 : gv_stashpv ("CBOR::XS::Tagged" , 1);
356 431
357 HV *stash = SvSTASH (sv); 432 HV *stash = SvSTASH (sv);
358 GV *method;
359 433
360 if (stash == boolean_stash) 434 if (stash == boolean_stash)
361 encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20); 435 {
436 encode_ch (enc, SvIV (sv) ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE);
437 return;
438 }
362 else if (stash == error_stash) 439 else if (stash == error_stash)
363 encode_ch (enc, 0xe0 | 23); 440 {
441 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
442 return;
443 }
364 else if (stash == tagged_stash) 444 else if (stash == tagged_stash)
365 { 445 {
366 if (svt != SVt_PVAV) 446 if (svt != SVt_PVAV)
367 croak ("encountered CBOR::XS::Tagged object that isn't an array"); 447 croak ("encountered CBOR::XS::Tagged object that isn't an array");
368 448
369 encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); 449 encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
370 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); 450 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1));
451
452 return;
453 }
454 }
455
456 if (ecb_expect_false (SvREFCNT (sv) > 1)
457 && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
458 {
459 if (ecb_expect_false (!enc->shareable))
460 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
461
462 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
463
464 if (SvOK (*svp))
371 } 465 {
466 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
467 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
468 return;
469 }
470 else
471 {
472 sv_setuv (*svp, enc->shareable_idx);
473 ++enc->shareable_idx;
474 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
475 }
476 }
477
478 if (ecb_expect_false (SvOBJECT (sv)))
479 {
480 HV *stash = SvSTASH (sv);
481 GV *method;
482
483 if (enc->cbor.flags & F_FORBID_OBJECTS)
484 croak ("encountered object '%s', but forbid_objects is enabled",
485 SvPV_nolen (sv_2mortal (newRV_inc (sv))));
372 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) 486 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
373 { 487 {
374 dSP; 488 dSP;
375 489
376 ENTER; SAVETMPS; PUSHMARK (SP); 490 ENTER; SAVETMPS;
491 PUSHMARK (SP);
377 // we re-bless the reference to get overload and other niceties right 492 // we re-bless the reference to get overload and other niceties right
378 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 493 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
379 494
380 PUTBACK; 495 PUTBACK;
381 // G_SCALAR ensures that return value is 1 496 // G_SCALAR ensures that return value is 1
394 } 509 }
395 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0) 510 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
396 { 511 {
397 dSP; 512 dSP;
398 513
399 ENTER; SAVETMPS; PUSHMARK (SP); 514 ENTER; SAVETMPS;
515 SAVESTACK_POS ();
516 PUSHMARK (SP);
400 EXTEND (SP, 2); 517 EXTEND (SP, 2);
401 // we re-bless the reference to get overload and other niceties right 518 // we re-bless the reference to get overload and other niceties right
402 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 519 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
403 PUSHs (sv_cbor); 520 PUSHs (sv_cbor);
404 521
409 // catch this surprisingly common error 526 // catch this surprisingly common error
410 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) 527 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)); 528 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
412 529
413 encode_tag (enc, CBOR_TAG_PERL_OBJECT); 530 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
414 encode_uint (enc, 0x80, count + 1); 531 encode_uint (enc, MAJOR_ARRAY, count + 1);
415 encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); 532 encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
416 533
417 while (count) 534 while (count)
418 encode_sv (enc, SP[1 - count--]); 535 encode_sv (enc, SP[1 - count--]);
419 536
420 PUTBACK; 537 PUTBACK;
441{ 558{
442 double nv = SvNVX (sv); 559 double nv = SvNVX (sv);
443 560
444 need (enc, 9); 561 need (enc, 9);
445 562
446 if (ecb_expect_false (nv == (U32)nv)) 563 if (ecb_expect_false (nv == (NV)(U32)nv))
447 encode_uint (enc, 0x00, (U32)nv); 564 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
448 //TODO: maybe I32? 565 //TODO: maybe I32?
449 else if (ecb_expect_false (nv == (float)nv)) 566 else if (ecb_expect_false (nv == (float)nv))
450 { 567 {
568 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
569
451 uint32_t fp = ecb_float_to_binary32 (nv); 570 uint32_t fp = ecb_float_to_binary32 (nv);
452
453 *enc->cur++ = 0xe0 | 26;
454 571
455 if (!ecb_big_endian ()) 572 if (!ecb_big_endian ())
456 fp = ecb_bswap32 (fp); 573 fp = ecb_bswap32 (fp);
457 574
458 memcpy (enc->cur, &fp, 4); 575 memcpy (enc->cur, &fp, 4);
459 enc->cur += 4; 576 enc->cur += 4;
460 } 577 }
461 else 578 else
462 { 579 {
580 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
581
463 uint64_t fp = ecb_double_to_binary64 (nv); 582 uint64_t fp = ecb_double_to_binary64 (nv);
464
465 *enc->cur++ = 0xe0 | 27;
466 583
467 if (!ecb_big_endian ()) 584 if (!ecb_big_endian ())
468 fp = ecb_bswap64 (fp); 585 fp = ecb_bswap64 (fp);
469 586
470 memcpy (enc->cur, &fp, 8); 587 memcpy (enc->cur, &fp, 8);
479 596
480 if (SvPOKp (sv)) 597 if (SvPOKp (sv))
481 { 598 {
482 STRLEN len; 599 STRLEN len;
483 char *str = SvPV (sv, len); 600 char *str = SvPV (sv, len);
484 encode_str (enc, SvUTF8 (sv), str, len); 601 encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len);
485 } 602 }
486 else if (SvNOKp (sv)) 603 else if (SvNOKp (sv))
487 encode_nv (enc, sv); 604 encode_nv (enc, sv);
488 else if (SvIOKp (sv)) 605 else if (SvIOKp (sv))
489 { 606 {
490 if (SvIsUV (sv)) 607 if (SvIsUV (sv))
491 encode_uint (enc, 0x00, SvUVX (sv)); 608 encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
492 else if (SvIVX (sv) >= 0) 609 else if (SvIVX (sv) >= 0)
493 encode_uint (enc, 0x00, SvIVX (sv)); 610 encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
494 else 611 else
495 encode_uint (enc, 0x20, -(SvIVX (sv) + 1)); 612 encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
496 } 613 }
497 else if (SvROK (sv)) 614 else if (SvROK (sv))
498 encode_rv (enc, SvRV (sv)); 615 encode_rv (enc, SvRV (sv));
499 else if (!SvOK (sv)) 616 else if (!SvOK (sv))
500 encode_ch (enc, 0xe0 | 22); 617 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
501 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 618 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
502 encode_ch (enc, 0xe0 | 23); 619 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
503 else 620 else
504 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data", 621 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
505 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 622 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
506} 623}
507 624
508static SV * 625static SV *
509encode_cbor (SV *scalar, CBOR *cbor) 626encode_cbor (SV *scalar, CBOR *cbor)
510{ 627{
511 enc_t enc = { }; 628 enc_t enc = { 0 };
512 629
513 enc.cbor = *cbor; 630 enc.cbor = *cbor;
514 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 631 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
515 enc.cur = SvPVX (enc.sv); 632 enc.cur = SvPVX (enc.sv);
516 enc.end = SvEND (enc.sv); 633 enc.end = SvEND (enc.sv);
517 634
518 SvPOK_only (enc.sv); 635 SvPOK_only (enc.sv);
519 636
520 if (cbor->flags & F_ALLOW_STRINGREF) 637 if (cbor->flags & F_PACK_STRINGS)
521 { 638 {
522 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE); 639 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
523 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ()); 640 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
524 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ()); 641 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
525 } 642 }
547 CBOR cbor; 664 CBOR cbor;
548 U32 depth; // recursion depth 665 U32 depth; // recursion depth
549 U32 maxdepth; // recursion depth limit 666 U32 maxdepth; // recursion depth limit
550 AV *shareable; 667 AV *shareable;
551 AV *stringref; 668 AV *stringref;
669 SV *decode_tagged;
670 SV *err_sv; // optional sv for error, needs to be freed
552} dec_t; 671} dec_t;
553 672
554#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE 673// set dec->err to ERRSV
674ecb_cold static void
675err_errsv (dec_t *dec)
676{
677 if (!dec->err)
678 {
679 dec->err_sv = newSVsv (ERRSV);
555 680
556#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data") 681 // chop off the trailing \n
682 SvCUR_set (dec->err_sv, SvCUR (dec->err_sv) - 1);
683 *SvEND (dec->err_sv) = 0;
557 684
685 dec->err = SvPVutf8_nolen (dec->err_sv);
686 }
687}
688
689// the following functions are used to reduce code size and help the compiler to optimise
690ecb_cold static void
691err_set (dec_t *dec, const char *reason)
692{
693 if (!dec->err)
694 dec->err = reason;
695}
696
697ecb_cold static void
698err_unexpected_end (dec_t *dec)
699{
700 err_set (dec, "unexpected end of CBOR data");
701}
702
703#define ERR_DO(do) SB do; goto fail; SE
704#define ERR(reason) ERR_DO (err_set (dec, reason))
705#define ERR_ERRSV ERR_DO (err_errsv (dec))
706
707#define WANT(len) if (ecb_expect_false ((uintptr_t)(dec->end - dec->cur) < (STRLEN)len)) ERR_DO (err_unexpected_end (dec))
708
558#define DEC_INC_DEPTH if (++dec->depth > dec->cbor.max_depth) ERR (ERR_NESTING_EXCEEDED) 709#define DEC_INC_DEPTH if (ecb_expect_false (++dec->depth > dec->cbor.max_depth)) ERR (ERR_NESTING_EXCEEDED)
559#define DEC_DEC_DEPTH --dec->depth 710#define DEC_DEC_DEPTH --dec->depth
560 711
561static UV 712static UV
562decode_uint (dec_t *dec) 713decode_uint (dec_t *dec)
563{ 714{
564 switch (*dec->cur & 31) 715 U8 m = *dec->cur & MINOR_MASK;
565 { 716 ++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;
570 717
571 case 24: 718 if (ecb_expect_true (m < LENGTH_EXT1))
719 return m;
720 else if (ecb_expect_true (m == LENGTH_EXT1))
721 {
572 WANT (2); 722 WANT (1);
573 dec->cur += 2; 723 dec->cur += 1;
574 return dec->cur[-1]; 724 return dec->cur[-1];
575 725 }
576 case 25: 726 else if (ecb_expect_true (m == LENGTH_EXT2))
727 {
577 WANT (3); 728 WANT (2);
578 dec->cur += 3; 729 dec->cur += 2;
579 return (((UV)dec->cur[-2]) << 8) 730 return (((UV)dec->cur[-2]) << 8)
580 | ((UV)dec->cur[-1]); 731 | ((UV)dec->cur[-1]);
581 732 }
582 case 26: 733 else if (ecb_expect_true (m == LENGTH_EXT4))
734 {
583 WANT (5); 735 WANT (4);
584 dec->cur += 5; 736 dec->cur += 4;
585 return (((UV)dec->cur[-4]) << 24) 737 return (((UV)dec->cur[-4]) << 24)
586 | (((UV)dec->cur[-3]) << 16) 738 | (((UV)dec->cur[-3]) << 16)
587 | (((UV)dec->cur[-2]) << 8) 739 | (((UV)dec->cur[-2]) << 8)
588 | ((UV)dec->cur[-1]); 740 | ((UV)dec->cur[-1]);
589 741 }
590 case 27: 742 else if (ecb_expect_true (m == LENGTH_EXT8))
743 {
591 WANT (9); 744 WANT (8);
592 dec->cur += 9; 745 dec->cur += 8;
746
747 return
748#if UVSIZE < 8
749 0
750#else
593 return (((UV)dec->cur[-8]) << 56) 751 (((UV)dec->cur[-8]) << 56)
594 | (((UV)dec->cur[-7]) << 48) 752 | (((UV)dec->cur[-7]) << 48)
595 | (((UV)dec->cur[-6]) << 40) 753 | (((UV)dec->cur[-6]) << 40)
596 | (((UV)dec->cur[-5]) << 32) 754 | (((UV)dec->cur[-5]) << 32)
755#endif
597 | (((UV)dec->cur[-4]) << 24) 756 | (((UV)dec->cur[-4]) << 24)
598 | (((UV)dec->cur[-3]) << 16) 757 | (((UV)dec->cur[-3]) << 16)
599 | (((UV)dec->cur[-2]) << 8) 758 | (((UV)dec->cur[-2]) << 8)
600 | ((UV)dec->cur[-1]); 759 | ((UV)dec->cur[-1]);
601 760 }
602 default: 761 else
603 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 762 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
604 }
605 763
606fail: 764fail:
607 return 0; 765 return 0;
608} 766}
609 767
614{ 772{
615 AV *av = newAV (); 773 AV *av = newAV ();
616 774
617 DEC_INC_DEPTH; 775 DEC_INC_DEPTH;
618 776
619 if ((*dec->cur & 31) == 31) 777 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
620 { 778 {
621 ++dec->cur; 779 ++dec->cur;
622 780
623 for (;;) 781 for (;;)
624 { 782 {
625 WANT (1); 783 WANT (1);
626 784
627 if (*dec->cur == (0xe0 | 31)) 785 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
628 { 786 {
629 ++dec->cur; 787 ++dec->cur;
630 break; 788 break;
631 } 789 }
632 790
633 av_push (av, decode_sv (dec)); 791 av_push (av, decode_sv (dec));
634 } 792 }
635 } 793 }
636 else 794 else
637 { 795 {
638 int i, len = decode_uint (dec); 796 UV i, len = decode_uint (dec);
639 797
798 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); 799 av_fill (av, len - 1);
641 800
642 for (i = 0; i < len; ++i) 801 for (i = 0; i < len; ++i)
643 AvARRAY (av)[i] = decode_sv (dec); 802 AvARRAY (av)[i] = decode_sv (dec);
644 } 803 }
645 804
646 DEC_DEC_DEPTH; 805 DEC_DEC_DEPTH;
647 return newRV_noinc ((SV *)av); 806 return newRV_noinc ((SV *)av);
648 807
649fail: 808fail:
650 SvREFCNT_dec (av); 809 SvREFCNT_dec_NN (av);
651 DEC_DEC_DEPTH; 810 DEC_DEC_DEPTH;
652 return &PL_sv_undef; 811 return &PL_sv_undef;
653} 812}
654 813
655static void 814static void
657{ 816{
658 // for speed reasons, we specialcase single-string 817 // for speed reasons, we specialcase single-string
659 // byte or utf-8 strings as keys, but only when !stringref 818 // byte or utf-8 strings as keys, but only when !stringref
660 819
661 if (ecb_expect_true (!dec->stringref)) 820 if (ecb_expect_true (!dec->stringref))
662 if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27) 821 if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
663 { 822 {
664 I32 len = decode_uint (dec); 823 STRLEN len = decode_uint (dec);
665 char *key = (char *)dec->cur; 824 char *key = (char *)dec->cur;
666 825
826 WANT (len);
667 dec->cur += len; 827 dec->cur += len;
668 828
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); 829 hv_store (hv, key, len, decode_sv (dec), 0);
673 830
674 return; 831 return;
675 } 832 }
676 else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27) 833 else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
677 { 834 {
678 I32 len = decode_uint (dec); 835 STRLEN len = decode_uint (dec);
679 char *key = (char *)dec->cur; 836 char *key = (char *)dec->cur;
680 837
838 WANT (len);
681 dec->cur += len; 839 dec->cur += len;
682 840
683 if (ecb_expect_false (dec->stringref)) 841 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
684 av_push (dec->stringref, newSVpvn_utf8 (key, len, 1)); 842 if (!is_utf8_string (key, len))
843 ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
685 844
686 hv_store (hv, key, -len, decode_sv (dec), 0); 845 hv_store (hv, key, -len, decode_sv (dec), 0);
687 846
688 return; 847 return;
689 } 848 }
690 849
691 SV *k = decode_sv (dec); 850 SV *k = decode_sv (dec);
692 SV *v = decode_sv (dec); 851 SV *v = decode_sv (dec);
693 852
853 // we leak memory if uncaught exceptions are thrown by random magical
854 // methods, and this is hopefully the only place where it can happen,
855 // so if there is a chance of an exception, take the very slow path.
856 // since catching exceptions is "undocumented/internal/forbidden" by
857 // the new p5p powers, we need to call out to a perl function :/
858 if (ecb_expect_false (SvAMAGIC (k)))
859 {
860 dSP;
861
862 ENTER; SAVETMPS;
863 PUSHMARK (SP);
864 EXTEND (SP, 3);
865 PUSHs (sv_2mortal (newRV_inc ((SV *)hv)));
866 PUSHs (sv_2mortal (k));
867 PUSHs (sv_2mortal (v));
868
869 PUTBACK;
870 call_pv ("CBOR::XS::_hv_store", G_VOID | G_DISCARD | G_EVAL);
871 SPAGAIN;
872
873 FREETMPS; LEAVE;
874
875 if (SvTRUE (ERRSV))
876 ERR_ERRSV;
877
878 return;
879 }
880
694 hv_store_ent (hv, k, v, 0); 881 hv_store_ent (hv, k, v, 0);
695 SvREFCNT_dec (k); 882 SvREFCNT_dec_NN (k);
883
884fail:
885 ;
696} 886}
697 887
698static SV * 888static SV *
699decode_hv (dec_t *dec) 889decode_hv (dec_t *dec)
700{ 890{
701 HV *hv = newHV (); 891 HV *hv = newHV ();
702 892
703 DEC_INC_DEPTH; 893 DEC_INC_DEPTH;
704 894
705 if ((*dec->cur & 31) == 31) 895 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
706 { 896 {
707 ++dec->cur; 897 ++dec->cur;
708 898
709 for (;;) 899 for (;;)
710 { 900 {
711 WANT (1); 901 WANT (1);
712 902
713 if (*dec->cur == (0xe0 | 31)) 903 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
714 { 904 {
715 ++dec->cur; 905 ++dec->cur;
716 break; 906 break;
717 } 907 }
718 908
719 decode_he (dec, hv); 909 decode_he (dec, hv);
720 } 910 }
721 } 911 }
722 else 912 else
723 { 913 {
724 int pairs = decode_uint (dec); 914 UV pairs = decode_uint (dec);
915
916 WANT (pairs); // complexity check - need at least one byte per value, do not allow supersize hashes
725 917
726 while (pairs--) 918 while (pairs--)
727 decode_he (dec, hv); 919 decode_he (dec, hv);
728 } 920 }
729 921
730 DEC_DEC_DEPTH; 922 DEC_DEC_DEPTH;
731 return newRV_noinc ((SV *)hv); 923 return newRV_noinc ((SV *)hv);
732 924
733fail: 925fail:
734 SvREFCNT_dec (hv); 926 SvREFCNT_dec_NN (hv);
735 DEC_DEC_DEPTH; 927 DEC_DEC_DEPTH;
736 return &PL_sv_undef; 928 return &PL_sv_undef;
737} 929}
738 930
739static SV * 931static SV *
740decode_str (dec_t *dec, int utf8) 932decode_str (dec_t *dec, int utf8)
741{ 933{
742 SV *sv = 0; 934 SV *sv = 0;
743 935
744 if ((*dec->cur & 31) == 31) 936 if (ecb_expect_false ((*dec->cur & MINOR_MASK) == MINOR_INDEF))
745 { 937 {
938 // indefinite length strings
746 ++dec->cur; 939 ++dec->cur;
747 940
941 U8 major = *dec->cur & MAJOR_MISC;
942
748 sv = newSVpvn ("", 0); 943 sv = newSVpvn ("", 0);
749 944
750 // not very fast, and certainly not robust against illegal input
751 for (;;) 945 for (;;)
752 { 946 {
753 WANT (1); 947 WANT (1);
754 948
755 if (*dec->cur == (0xe0 | 31)) 949 if ((*dec->cur - major) > LENGTH_EXT8)
950 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
756 { 951 {
757 ++dec->cur; 952 ++dec->cur;
758 break; 953 break;
759 } 954 }
955 else
956 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
760 957
761 sv_catsv (sv, decode_sv (dec)); 958 STRLEN len = decode_uint (dec);
959
960 WANT (len);
961 sv_catpvn (sv, dec->cur, len);
962 dec->cur += len;
762 } 963 }
763 } 964 }
764 else 965 else
765 { 966 {
766 STRLEN len = decode_uint (dec); 967 STRLEN len = decode_uint (dec);
773 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1)) 974 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
774 av_push (dec->stringref, SvREFCNT_inc_NN (sv)); 975 av_push (dec->stringref, SvREFCNT_inc_NN (sv));
775 } 976 }
776 977
777 if (utf8) 978 if (utf8)
979 {
980 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
981 if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
982 ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
983
778 SvUTF8_on (sv); 984 SvUTF8_on (sv);
985 }
779 986
780 return sv; 987 return sv;
781 988
782fail: 989fail:
783 SvREFCNT_dec (sv); 990 SvREFCNT_dec (sv);
802 sv = newRV_noinc (decode_sv (dec)); 1009 sv = newRV_noinc (decode_sv (dec));
803 break; 1010 break;
804 1011
805 case CBOR_TAG_STRINGREF_NAMESPACE: 1012 case CBOR_TAG_STRINGREF_NAMESPACE:
806 { 1013 {
1014 // do not use SAVETMPS/FREETMPS, as these will
1015 // erase mortalised caches, e.g. "shareable"
807 ENTER; SAVETMPS; 1016 ENTER;
808 1017
809 SAVESPTR (dec->stringref); 1018 SAVESPTR (dec->stringref);
810 dec->stringref = (AV *)sv_2mortal ((SV *)newAV ()); 1019 dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());
811 1020
812 sv = decode_sv (dec); 1021 sv = decode_sv (dec);
813 1022
814 FREETMPS; LEAVE; 1023 LEAVE;
815 } 1024 }
816 break; 1025 break;
817 1026
818 case CBOR_TAG_STRINGREF: 1027 case CBOR_TAG_STRINGREF:
819 { 1028 {
820 if ((*dec->cur >> 5) != 0) 1029 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
821 ERR ("corrupted CBOR data (stringref index not an unsigned integer)"); 1030 ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
822 1031
823 UV idx = decode_uint (dec); 1032 UV idx = decode_uint (dec);
824 1033
825 if (!dec->stringref || (int)idx > AvFILLp (dec->stringref)) 1034 if (!dec->stringref || (int)idx > AvFILLp (dec->stringref))
832 case CBOR_TAG_VALUE_SHAREABLE: 1041 case CBOR_TAG_VALUE_SHAREABLE:
833 { 1042 {
834 if (ecb_expect_false (!dec->shareable)) 1043 if (ecb_expect_false (!dec->shareable))
835 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ()); 1044 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
836 1045
1046 if (dec->cbor.flags & F_ALLOW_CYCLES)
1047 {
837 sv = newSV (0); 1048 sv = newSV (0);
838 av_push (dec->shareable, SvREFCNT_inc_NN (sv)); 1049 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
839 1050
840 SV *osv = decode_sv (dec); 1051 SV *osv = decode_sv (dec);
841 sv_setsv (sv, osv); 1052 sv_setsv (sv, osv);
842 SvREFCNT_dec_NN (osv); 1053 SvREFCNT_dec_NN (osv);
1054 }
1055 else
1056 {
1057 av_push (dec->shareable, &PL_sv_undef);
1058 int idx = AvFILLp (dec->shareable);
1059 sv = decode_sv (dec);
1060 av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
1061 }
843 } 1062 }
844 break; 1063 break;
845 1064
846 case CBOR_TAG_VALUE_SHAREDREF: 1065 case CBOR_TAG_VALUE_SHAREDREF:
847 { 1066 {
848 if ((*dec->cur >> 5) != 0) 1067 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
849 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)"); 1068 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
850 1069
851 UV idx = decode_uint (dec); 1070 UV idx = decode_uint (dec);
852 1071
853 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable)) 1072 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable))
854 ERR ("corrupted CBOR data (sharedref index out of bounds)"); 1073 ERR ("corrupted CBOR data (sharedref index out of bounds)");
855 1074
856 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]); 1075 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
1076
1077 if (sv == &PL_sv_undef)
1078 ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
857 } 1079 }
858 break; 1080 break;
859 1081
860 case CBOR_TAG_PERL_OBJECT: 1082 case CBOR_TAG_PERL_OBJECT:
861 { 1083 {
1084 if (dec->cbor.flags & F_FORBID_OBJECTS)
1085 goto filter;
1086
862 sv = decode_sv (dec); 1087 sv = decode_sv (dec);
863 1088
864 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) 1089 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
865 ERR ("corrupted CBOR data (non-array perl object)"); 1090 ERR ("corrupted CBOR data (non-array perl object)");
866 1091
876 if (!method) 1101 if (!method)
877 ERR ("cannot decode perl-object (package does not have a THAW method)"); 1102 ERR ("cannot decode perl-object (package does not have a THAW method)");
878 1103
879 dSP; 1104 dSP;
880 1105
881 ENTER; SAVETMPS; PUSHMARK (SP); 1106 ENTER; SAVETMPS;
1107 PUSHMARK (SP);
882 EXTEND (SP, len + 1); 1108 EXTEND (SP, len + 1);
883 // we re-bless the reference to get overload and other niceties right 1109 // we re-bless the reference to get overload and other niceties right
884 PUSHs (*av_fetch (av, 0, 1)); 1110 PUSHs (*av_fetch (av, 0, 1));
885 PUSHs (sv_cbor); 1111 PUSHs (sv_cbor);
886 1112
894 SPAGAIN; 1120 SPAGAIN;
895 1121
896 if (SvTRUE (ERRSV)) 1122 if (SvTRUE (ERRSV))
897 { 1123 {
898 FREETMPS; LEAVE; 1124 FREETMPS; LEAVE;
899 ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV)))); 1125 ERR_ERRSV;
900 } 1126 }
901 1127
902 SvREFCNT_dec (sv); 1128 SvREFCNT_dec_NN (sv);
903 sv = SvREFCNT_inc (POPs); 1129 sv = SvREFCNT_inc (POPs);
904 1130
905 PUTBACK; 1131 PUTBACK;
906 1132
907 FREETMPS; LEAVE; 1133 FREETMPS; LEAVE;
908 } 1134 }
909 break; 1135 break;
910 1136
911 default: 1137 default:
1138 filter:
912 { 1139 {
1140 SV *tag_sv = newSVuv (tag);
1141
913 sv = decode_sv (dec); 1142 sv = decode_sv (dec);
914 1143
1144 dSP;
1145 ENTER; SAVETMPS;
1146 SAVESTACK_POS ();
1147 PUSHMARK (SP);
1148 EXTEND (SP, 2);
1149 PUSHs (tag_sv);
1150 PUSHs (sv);
1151
1152 PUTBACK;
1153 int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL);
1154 SPAGAIN;
1155
1156 if (SvTRUE (ERRSV))
1157 {
1158 SvREFCNT_dec_NN (tag_sv);
1159 FREETMPS; LEAVE;
1160 ERR_ERRSV;
1161 }
1162
1163 if (count)
1164 {
1165 SvREFCNT_dec_NN (tag_sv);
1166 SvREFCNT_dec_NN (sv);
1167 sv = SvREFCNT_inc_NN (POPs);
1168 }
1169 else
1170 {
915 AV *av = newAV (); 1171 AV *av = newAV ();
916 av_push (av, newSVuv (tag)); 1172 av_push (av, tag_sv);
917 av_push (av, sv); 1173 av_push (av, sv);
918 1174
919 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 1175 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
920 ? cbor_tagged_stash 1176 ? cbor_tagged_stash
921 : gv_stashpv ("CBOR::XS::Tagged" , 1); 1177 : gv_stashpv ("CBOR::XS::Tagged" , 1);
922
923 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash); 1178 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
1179 }
1180
1181 PUTBACK;
1182
1183 FREETMPS; LEAVE;
924 } 1184 }
925 break; 1185 break;
926 } 1186 }
927 1187
928 return sv; 1188 return sv;
935static SV * 1195static SV *
936decode_sv (dec_t *dec) 1196decode_sv (dec_t *dec)
937{ 1197{
938 WANT (1); 1198 WANT (1);
939 1199
940 switch (*dec->cur >> 5) 1200 switch (*dec->cur >> MAJOR_SHIFT)
941 { 1201 {
942 case 0: // unsigned int 1202 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
943 return newSVuv (decode_uint (dec)); 1203 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
944 case 1: // negative int 1204 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
945 return newSViv (-1 - (IV)decode_uint (dec)); 1205 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
946 case 2: // octet string 1206 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
947 return decode_str (dec, 0); 1207 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
948 case 3: // utf-8 string 1208 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
949 return decode_str (dec, 1); 1209
950 case 4: // array 1210 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) 1211 switch (*dec->cur++ & MINOR_MASK)
958 { 1212 {
959 case 20: 1213 case SIMPLE_FALSE:
960#if CBOR_SLOW 1214#if CBOR_SLOW
961 types_false = get_bool ("Types::Serialiser::false"); 1215 types_false = get_bool ("Types::Serialiser::false");
962#endif 1216#endif
963 return newSVsv (types_false); 1217 return newSVsv (types_false);
964 case 21: 1218 case SIMPLE_TRUE:
965#if CBOR_SLOW 1219#if CBOR_SLOW
966 types_true = get_bool ("Types::Serialiser::true"); 1220 types_true = get_bool ("Types::Serialiser::true");
967#endif 1221#endif
968 return newSVsv (types_true); 1222 return newSVsv (types_true);
969 case 22: 1223 case SIMPLE_NULL:
970 return newSVsv (&PL_sv_undef); 1224 return newSVsv (&PL_sv_undef);
971 case 23: 1225 case SIMPLE_UNDEF:
972#if CBOR_SLOW 1226#if CBOR_SLOW
973 types_error = get_bool ("Types::Serialiser::error"); 1227 types_error = get_bool ("Types::Serialiser::error");
974#endif 1228#endif
975 return newSVsv (types_error); 1229 return newSVsv (types_error);
976 1230
977 case 25: 1231 case MISC_FLOAT16:
978 { 1232 {
979 WANT (2); 1233 WANT (2);
980 1234
981 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1235 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
982 dec->cur += 2; 1236 dec->cur += 2;
983 1237
984 return newSVnv (ecb_binary16_to_float (fp)); 1238 return newSVnv (ecb_binary16_to_float (fp));
985 } 1239 }
986 1240
987 case 26: 1241 case MISC_FLOAT32:
988 { 1242 {
989 uint32_t fp; 1243 uint32_t fp;
990 WANT (4); 1244 WANT (4);
991 memcpy (&fp, dec->cur, 4); 1245 memcpy (&fp, dec->cur, 4);
992 dec->cur += 4; 1246 dec->cur += 4;
995 fp = ecb_bswap32 (fp); 1249 fp = ecb_bswap32 (fp);
996 1250
997 return newSVnv (ecb_binary32_to_float (fp)); 1251 return newSVnv (ecb_binary32_to_float (fp));
998 } 1252 }
999 1253
1000 case 27: 1254 case MISC_FLOAT64:
1001 { 1255 {
1002 uint64_t fp; 1256 uint64_t fp;
1003 WANT (8); 1257 WANT (8);
1004 memcpy (&fp, dec->cur, 8); 1258 memcpy (&fp, dec->cur, 8);
1005 dec->cur += 8; 1259 dec->cur += 8;
1008 fp = ecb_bswap64 (fp); 1262 fp = ecb_bswap64 (fp);
1009 1263
1010 return newSVnv (ecb_binary64_to_double (fp)); 1264 return newSVnv (ecb_binary64_to_double (fp));
1011 } 1265 }
1012 1266
1013 // 0..19 unassigned 1267 // 0..19 unassigned simple
1014 // 24 reserved + unassigned (reserved values are not encodable) 1268 // 24 reserved + unassigned simple (reserved values are not encodable)
1269 // 28-30 unassigned misc
1270 // 31 break code
1015 default: 1271 default:
1016 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1272 ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1017 } 1273 }
1018 1274
1019 break; 1275 break;
1020 } 1276 }
1021 1277
1024} 1280}
1025 1281
1026static SV * 1282static SV *
1027decode_cbor (SV *string, CBOR *cbor, char **offset_return) 1283decode_cbor (SV *string, CBOR *cbor, char **offset_return)
1028{ 1284{
1029 dec_t dec = { }; 1285 dec_t dec = { 0 };
1030 SV *sv; 1286 SV *sv;
1031 STRLEN len; 1287 STRLEN len;
1032 char *data = SvPVbyte (string, len); 1288 char *data = SvPVbyte (string, len);
1033 1289
1034 if (len > cbor->max_size && cbor->max_size) 1290 if (len > cbor->max_size && cbor->max_size)
1048 if (dec.cur != dec.end && !dec.err) 1304 if (dec.cur != dec.end && !dec.err)
1049 dec.err = "garbage after CBOR object"; 1305 dec.err = "garbage after CBOR object";
1050 1306
1051 if (dec.err) 1307 if (dec.err)
1052 { 1308 {
1309 if (dec.shareable)
1310 {
1311 // need to break cyclic links, which would all be in shareable
1312 int i;
1313 SV **svp;
1314
1315 for (i = av_len (dec.shareable) + 1; i--; )
1316 if ((svp = av_fetch (dec.shareable, i, 0)))
1317 sv_setsv (*svp, &PL_sv_undef);
1318 }
1319
1053 SvREFCNT_dec (sv); 1320 SvREFCNT_dec_NN (sv);
1321
1322 if (dec.err_sv)
1323 sv_2mortal (dec.err_sv);
1324
1054 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur); 1325 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur);
1055 } 1326 }
1056 1327
1057 sv = sv_2mortal (sv); 1328 sv = sv_2mortal (sv);
1058 1329
1059 return sv; 1330 return sv;
1060} 1331}
1061 1332
1333/////////////////////////////////////////////////////////////////////////////
1334// incremental parser
1335
1336#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
1337
1338// returns 0 for notyet, 1 for success or error
1339static int
1340incr_parse (CBOR *self, SV *cborstr)
1341{
1342 STRLEN cur;
1343 SvPV (cborstr, cur);
1344
1345 while (ecb_expect_true (self->incr_need <= cur))
1346 {
1347 // table of integer count bytes
1348 static I8 incr_len[MINOR_MASK + 1] = {
1349 0, 0, 0, 0, 0, 0, 0, 0,
1350 0, 0, 0, 0, 0, 0, 0, 0,
1351 0, 0, 0, 0, 0, 0, 0, 0,
1352 1, 2, 4, 8,-1,-1,-1,-2
1353 };
1354
1355 const U8 *p = SvPVX (cborstr) + self->incr_pos;
1356 U8 m = *p & MINOR_MASK;
1357 IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1358 I8 ilen = incr_len[m];
1359
1360 self->incr_need = self->incr_pos + 1;
1361
1362 if (ecb_expect_false (ilen < 0))
1363 {
1364 if (m != MINOR_INDEF)
1365 return 1; // error
1366
1367 if (*p == (MAJOR_MISC | MINOR_INDEF))
1368 {
1369 if (count >= 0)
1370 return 1; // error
1371
1372 count = 1;
1373 }
1374 else
1375 {
1376 av_push (self->incr_count, newSViv (-1)); //TODO: nest
1377 count = -1;
1378 }
1379 }
1380 else
1381 {
1382 self->incr_need += ilen;
1383 if (ecb_expect_false (self->incr_need > cur))
1384 return 0;
1385
1386 int major = *p >> MAJOR_SHIFT;
1387
1388 switch (major)
1389 {
1390 case MAJOR_TAG >> MAJOR_SHIFT:
1391 ++count; // tags merely prefix another value
1392 break;
1393
1394 case MAJOR_BYTES >> MAJOR_SHIFT:
1395 case MAJOR_TEXT >> MAJOR_SHIFT:
1396 case MAJOR_ARRAY >> MAJOR_SHIFT:
1397 case MAJOR_MAP >> MAJOR_SHIFT:
1398 {
1399 UV len;
1400
1401 if (ecb_expect_false (ilen))
1402 {
1403 len = 0;
1404
1405 do {
1406 len = (len << 8) | *++p;
1407 } while (--ilen);
1408 }
1409 else
1410 len = m;
1411
1412 switch (major)
1413 {
1414 case MAJOR_BYTES >> MAJOR_SHIFT:
1415 case MAJOR_TEXT >> MAJOR_SHIFT:
1416 self->incr_need += len;
1417 if (ecb_expect_false (self->incr_need > cur))
1418 return 0;
1419
1420 break;
1421
1422 case MAJOR_MAP >> MAJOR_SHIFT:
1423 len <<= 1;
1424 case MAJOR_ARRAY >> MAJOR_SHIFT:
1425 if (len)
1426 {
1427 av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
1428 count = len + 1;
1429 }
1430 break;
1431 }
1432 }
1433 }
1434 }
1435
1436 self->incr_pos = self->incr_need;
1437
1438 if (count > 0)
1439 {
1440 while (!--count)
1441 {
1442 if (!AvFILLp (self->incr_count))
1443 return 1; // done
1444
1445 SvREFCNT_dec_NN (av_pop (self->incr_count));
1446 count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1447 }
1448
1449 SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
1450 }
1451 }
1452
1453 return 0;
1454}
1455
1456
1062///////////////////////////////////////////////////////////////////////////// 1457/////////////////////////////////////////////////////////////////////////////
1063// XS interface functions 1458// XS interface functions
1064 1459
1065MODULE = CBOR::XS PACKAGE = CBOR::XS 1460MODULE = CBOR::XS PACKAGE = CBOR::XS
1066 1461
1074 1469
1075 types_true = get_bool ("Types::Serialiser::true" ); 1470 types_true = get_bool ("Types::Serialiser::true" );
1076 types_false = get_bool ("Types::Serialiser::false"); 1471 types_false = get_bool ("Types::Serialiser::false");
1077 types_error = get_bool ("Types::Serialiser::error"); 1472 types_error = get_bool ("Types::Serialiser::error");
1078 1473
1474 default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1475
1079 sv_cbor = newSVpv ("CBOR", 0); 1476 sv_cbor = newSVpv ("CBOR", 0);
1080 SvREADONLY_on (sv_cbor); 1477 SvREADONLY_on (sv_cbor);
1478
1479 assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1));
1081} 1480}
1082 1481
1083PROTOTYPES: DISABLE 1482PROTOTYPES: DISABLE
1084 1483
1085void CLONE (...) 1484void CLONE (...)
1104void shrink (CBOR *self, int enable = 1) 1503void shrink (CBOR *self, int enable = 1)
1105 ALIAS: 1504 ALIAS:
1106 shrink = F_SHRINK 1505 shrink = F_SHRINK
1107 allow_unknown = F_ALLOW_UNKNOWN 1506 allow_unknown = F_ALLOW_UNKNOWN
1108 allow_sharing = F_ALLOW_SHARING 1507 allow_sharing = F_ALLOW_SHARING
1109 allow_stringref = F_ALLOW_STRINGREF 1508 allow_cycles = F_ALLOW_CYCLES
1509 forbid_objects = F_FORBID_OBJECTS
1510 pack_strings = F_PACK_STRINGS
1511 text_keys = F_TEXT_KEYS
1512 text_strings = F_TEXT_STRINGS
1513 validate_utf8 = F_VALIDATE_UTF8
1110 PPCODE: 1514 PPCODE:
1111{ 1515{
1112 if (enable) 1516 if (enable)
1113 self->flags |= ix; 1517 self->flags |= ix;
1114 else 1518 else
1120void get_shrink (CBOR *self) 1524void get_shrink (CBOR *self)
1121 ALIAS: 1525 ALIAS:
1122 get_shrink = F_SHRINK 1526 get_shrink = F_SHRINK
1123 get_allow_unknown = F_ALLOW_UNKNOWN 1527 get_allow_unknown = F_ALLOW_UNKNOWN
1124 get_allow_sharing = F_ALLOW_SHARING 1528 get_allow_sharing = F_ALLOW_SHARING
1125 get_allow_stringref = F_ALLOW_STRINGREF 1529 get_allow_cycles = F_ALLOW_CYCLES
1530 get_forbid_objects = F_FORBID_OBJECTS
1531 get_pack_strings = F_PACK_STRINGS
1532 get_text_keys = F_TEXT_KEYS
1533 get_text_strings = F_TEXT_STRINGS
1534 get_validate_utf8 = F_VALIDATE_UTF8
1126 PPCODE: 1535 PPCODE:
1127 XPUSHs (boolSV (self->flags & ix)); 1536 XPUSHs (boolSV (self->flags & ix));
1128 1537
1129void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1538void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1130 PPCODE: 1539 PPCODE:
1143 XPUSHs (ST (0)); 1552 XPUSHs (ST (0));
1144 1553
1145int get_max_size (CBOR *self) 1554int get_max_size (CBOR *self)
1146 CODE: 1555 CODE:
1147 RETVAL = self->max_size; 1556 RETVAL = self->max_size;
1557 OUTPUT:
1558 RETVAL
1559
1560void filter (CBOR *self, SV *filter = 0)
1561 PPCODE:
1562 SvREFCNT_dec (self->filter);
1563 self->filter = filter ? newSVsv (filter) : filter;
1564 XPUSHs (ST (0));
1565
1566SV *get_filter (CBOR *self)
1567 CODE:
1568 RETVAL = self->filter ? self->filter : NEWSV (0, 0);
1148 OUTPUT: 1569 OUTPUT:
1149 RETVAL 1570 RETVAL
1150 1571
1151void encode (CBOR *self, SV *scalar) 1572void encode (CBOR *self, SV *scalar)
1152 PPCODE: 1573 PPCODE:
1167 EXTEND (SP, 2); 1588 EXTEND (SP, 2);
1168 PUSHs (sv); 1589 PUSHs (sv);
1169 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1590 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1170} 1591}
1171 1592
1593void incr_parse (CBOR *self, SV *cborstr)
1594 ALIAS:
1595 incr_parse_multiple = 1
1596 PPCODE:
1597{
1598 if (SvUTF8 (cborstr))
1599 sv_utf8_downgrade (cborstr, 0);
1600
1601 if (!self->incr_count)
1602 {
1603 self->incr_count = newAV ();
1604 self->incr_pos = 0;
1605 self->incr_need = 1;
1606
1607 av_push (self->incr_count, newSViv (1));
1608 }
1609
1610 do
1611 {
1612 if (!incr_parse (self, cborstr))
1613 {
1614 if (self->incr_need > self->max_size && self->max_size)
1615 croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
1616 (unsigned long)self->incr_need, (unsigned long)self->max_size);
1617
1618 break;
1619 }
1620
1621 SV *sv;
1622 char *offset;
1623
1624 PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
1625 XPUSHs (sv);
1626
1627 sv_chop (cborstr, offset);
1628
1629 av_clear (self->incr_count);
1630 av_push (self->incr_count, newSViv (1));
1631
1632 self->incr_pos = 0;
1633 self->incr_need = self->incr_pos + 1;
1634 }
1635 while (ix);
1636}
1637
1638void incr_reset (CBOR *self)
1639 CODE:
1640{
1641 SvREFCNT_dec (self->incr_count);
1642 self->incr_count = 0;
1643}
1644
1645void DESTROY (CBOR *self)
1646 PPCODE:
1647 cbor_free (self);
1648
1172PROTOTYPES: ENABLE 1649PROTOTYPES: ENABLE
1173 1650
1174void encode_cbor (SV *scalar) 1651void encode_cbor (SV *scalar)
1652 ALIAS:
1653 encode_cbor = 0
1654 encode_cbor_sharing = F_ALLOW_SHARING
1175 PPCODE: 1655 PPCODE:
1176{ 1656{
1177 CBOR cbor; 1657 CBOR cbor;
1178 cbor_init (&cbor); 1658 cbor_init (&cbor);
1659 cbor.flags |= ix;
1179 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1660 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1180 XPUSHs (scalar); 1661 XPUSHs (scalar);
1181} 1662}
1182 1663
1183void decode_cbor (SV *cborstr) 1664void decode_cbor (SV *cborstr)

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines