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 1.22 by root, Wed Nov 20 15:05:46 2013 UTC vs.
Revision 1.68 by root, Mon Nov 30 18:30:29 2020 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
104// known forced types, also hardcoded in CBOR.pm
105enum
106{
107 AS_CBOR = 0,
108 AS_INT = 1,
109 AS_BYTES = 2,
110 AS_TEXT = 3,
111 AS_FLOAT16 = 4,
112 AS_FLOAT32 = 5,
113 AS_FLOAT64 = 6,
114 // possibly future enhancements: (generic) float, (generic) string
115};
116
59#define F_SHRINK 0x00000001UL 117#define F_SHRINK 0x00000001UL
60#define F_ALLOW_UNKNOWN 0x00000002UL 118#define F_ALLOW_UNKNOWN 0x00000002UL
61#define F_ALLOW_SHARING 0x00000004UL //TODO 119#define F_ALLOW_SHARING 0x00000004UL
120#define F_ALLOW_CYCLES 0x00000008UL
121#define F_FORBID_OBJECTS 0x00000010UL
62#define F_DEDUP_STRINGS 0x00000008UL //TODO 122#define F_PACK_STRINGS 0x00000020UL
123#define F_TEXT_KEYS 0x00000040UL
124#define F_TEXT_STRINGS 0x00000080UL
125#define F_VALIDATE_UTF8 0x00000100UL
63 126
64#define INIT_SIZE 32 // initial scalar size to be allocated 127#define INIT_SIZE 32 // initial scalar size to be allocated
65 128
66#define SB do { 129#define SB do {
67#define SE } while (0) 130#define SE } while (0)
79# define CBOR_SLOW 0 142# define CBOR_SLOW 0
80# define CBOR_STASH cbor_stash 143# define CBOR_STASH cbor_stash
81#endif 144#endif
82 145
83static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS:: 146static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS::
84static SV *types_true, *types_false, *types_error, *sv_cbor; 147static SV *types_true, *types_false, *types_error, *sv_cbor, *default_filter;
85 148
86typedef struct { 149typedef struct {
87 U32 flags; 150 U32 flags;
88 U32 max_depth; 151 U32 max_depth;
89 STRLEN max_size; 152 STRLEN max_size;
153 SV *filter;
154
155 // for the incremental parser
156 STRLEN incr_pos; // the current offset into the text
157 STRLEN incr_need; // minimum bytes needed to decode
158 AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef.
90} CBOR; 159} CBOR;
91 160
92ecb_inline void 161ecb_inline void
93cbor_init (CBOR *cbor) 162cbor_init (CBOR *cbor)
94{ 163{
95 Zero (cbor, 1, CBOR); 164 Zero (cbor, 1, CBOR);
96 cbor->max_depth = 512; 165 cbor->max_depth = 512;
166}
167
168ecb_inline void
169cbor_free (CBOR *cbor)
170{
171 SvREFCNT_dec (cbor->filter);
172 SvREFCNT_dec (cbor->incr_count);
97} 173}
98 174
99///////////////////////////////////////////////////////////////////////////// 175/////////////////////////////////////////////////////////////////////////////
100// utility functions 176// utility functions
101 177
127 203
128// minimum length of a string to be registered for stringref 204// minimum length of a string to be registered for stringref
129ecb_inline int 205ecb_inline int
130minimum_string_length (UV idx) 206minimum_string_length (UV idx)
131{ 207{
132 return idx > 23 208 return idx <= 23 ? 3
133 ? idx > 0xffU 209 : idx <= 0xffU ? 4
134 ? idx > 0xffffU 210 : idx <= 0xffffU ? 5
135 ? idx > 0xffffffffU 211 : idx <= 0xffffffffU ? 7
136 ? 7 212 : 11;
137 : 6
138 : 5
139 : 4
140 : 3;
141} 213}
142 214
143///////////////////////////////////////////////////////////////////////////// 215/////////////////////////////////////////////////////////////////////////////
144// encoder 216// encoder
145 217
158} enc_t; 230} enc_t;
159 231
160ecb_inline void 232ecb_inline void
161need (enc_t *enc, STRLEN len) 233need (enc_t *enc, STRLEN len)
162{ 234{
163 if (ecb_expect_false (enc->cur + len >= enc->end)) 235 if (ecb_expect_false ((uintptr_t)(enc->end - enc->cur) < len))
164 { 236 {
165 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv); 237 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv);
166 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1); 238 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1);
167 enc->cur = SvPVX (enc->sv) + cur; 239 enc->cur = SvPVX (enc->sv) + cur;
168 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1; 240 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;
174{ 246{
175 need (enc, 1); 247 need (enc, 1);
176 *enc->cur++ = ch; 248 *enc->cur++ = ch;
177} 249}
178 250
251// used for tags, intregers, element counts and so on
179static void 252static void
180encode_uint (enc_t *enc, int major, UV len) 253encode_uint (enc_t *enc, int major, UV len)
181{ 254{
182 need (enc, 9); 255 need (enc, 9);
183 256
184 if (len < 24) 257 if (ecb_expect_true (len < LENGTH_EXT1))
185 *enc->cur++ = major | len; 258 *enc->cur++ = major | len;
186 else if (len <= 0xff) 259 else if (ecb_expect_true (len <= 0xffU))
187 { 260 {
188 *enc->cur++ = major | 24; 261 *enc->cur++ = major | LENGTH_EXT1;
189 *enc->cur++ = len; 262 *enc->cur++ = len;
190 } 263 }
191 else if (len <= 0xffff) 264 else if (len <= 0xffffU)
192 { 265 {
193 *enc->cur++ = major | 25; 266 *enc->cur++ = major | LENGTH_EXT2;
194 *enc->cur++ = len >> 8; 267 *enc->cur++ = len >> 8;
195 *enc->cur++ = len; 268 *enc->cur++ = len;
196 } 269 }
197 else if (len <= 0xffffffff) 270 else if (len <= 0xffffffffU)
198 { 271 {
199 *enc->cur++ = major | 26; 272 *enc->cur++ = major | LENGTH_EXT4;
200 *enc->cur++ = len >> 24; 273 *enc->cur++ = len >> 24;
201 *enc->cur++ = len >> 16; 274 *enc->cur++ = len >> 16;
202 *enc->cur++ = len >> 8; 275 *enc->cur++ = len >> 8;
203 *enc->cur++ = len; 276 *enc->cur++ = len;
204 } 277 }
205 else 278 else
206 { 279 {
207 *enc->cur++ = major | 27; 280 *enc->cur++ = major | LENGTH_EXT8;
208 *enc->cur++ = len >> 56; 281 *enc->cur++ = len >> 56;
209 *enc->cur++ = len >> 48; 282 *enc->cur++ = len >> 48;
210 *enc->cur++ = len >> 40; 283 *enc->cur++ = len >> 40;
211 *enc->cur++ = len >> 32; 284 *enc->cur++ = len >> 32;
212 *enc->cur++ = len >> 24; 285 *enc->cur++ = len >> 24;
214 *enc->cur++ = len >> 8; 287 *enc->cur++ = len >> 8;
215 *enc->cur++ = len; 288 *enc->cur++ = len;
216 } 289 }
217} 290}
218 291
292// encodes a perl value into a CBOR integer
293ecb_inline void
294encode_int (enc_t *enc, SV *sv)
295{
296 if (SvIsUV (sv))
297 encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
298 else if (SvIVX (sv) >= 0)
299 encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
300 else
301 encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
302}
303
219ecb_inline void 304ecb_inline void
220encode_tag (enc_t *enc, UV tag) 305encode_tag (enc_t *enc, UV tag)
221{ 306{
222 encode_uint (enc, 0xc0, tag); 307 encode_uint (enc, MAJOR_TAG, tag);
223} 308}
224 309
225static void 310// exceptional (hopefully) slow path for byte strings that need to be utf8-encoded
311ecb_noinline static void
226encode_str (enc_t *enc, int utf8, char *str, STRLEN len) 312encode_str_utf8 (enc_t *enc, int utf8, char *str, STRLEN len)
227{ 313{
314 STRLEN ulen = len;
315 U8 *p, *pend = (U8 *)str + len;
316
317 for (p = (U8 *)str; p < pend; ++p)
318 ulen += *p >> 7; // count set high bits
319
320 encode_uint (enc, MAJOR_TEXT, ulen);
321
322 need (enc, ulen);
323 for (p = (U8 *)str; p < pend; ++p)
324 if (*p < 0x80)
325 *enc->cur++ = *p;
326 else
327 {
328 *enc->cur++ = 0xc0 + (*p >> 6);
329 *enc->cur++ = 0x80 + (*p & 63);
330 }
331}
332
333ecb_inline void
334encode_str (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
335{
336 if (ecb_expect_false (upgrade_utf8))
337 if (!utf8)
338 {
339 encode_str_utf8 (enc, utf8, str, len);
340 return;
341 }
342
343 encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
344 need (enc, len);
345 memcpy (enc->cur, str, len);
346 enc->cur += len;
347}
348
349ecb_inline void
350encode_strref (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
351{
228 if (ecb_expect_false (enc->cbor.flags & F_DEDUP_STRINGS)) 352 if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
229 { 353 {
230 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1); 354 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
231 355
232 if (SvOK (*svp)) 356 if (SvOK (*svp))
233 { 357 {
234 // already registered, use stringref 358 // already registered, use stringref
235 encode_tag (enc, CBOR_TAG_STRINGREF); 359 encode_tag (enc, CBOR_TAG_STRINGREF);
236 encode_uint (enc, 0x00, SvUV (*svp)); 360 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
237 return; 361 return;
238 } 362 }
239 else if (len >= minimum_string_length (enc->stringref_idx)) 363 else if (len >= minimum_string_length (enc->stringref_idx))
240 { 364 {
241 // register only 365 // register only
242 sv_setuv (*svp, enc->stringref_idx); 366 sv_setuv (*svp, enc->stringref_idx);
243 ++enc->stringref_idx; 367 ++enc->stringref_idx;
244 } 368 }
245 } 369 }
246 370
247 encode_uint (enc, utf8 ? 0x60 : 0x40, len); 371 encode_str (enc, upgrade_utf8, utf8, str, len);
372}
373
374ecb_inline void
375encode_float16 (enc_t *enc, NV nv)
376{
248 need (enc, len); 377 need (enc, 1+2);
378
379 *enc->cur++ = MAJOR_MISC | MISC_FLOAT16;
380
381 uint16_t fp = ecb_float_to_binary16 (nv);
382
383 if (!ecb_big_endian ())
384 fp = ecb_bswap16 (fp);
385
249 memcpy (enc->cur, str, len); 386 memcpy (enc->cur, &fp, 2);
250 enc->cur += len; 387 enc->cur += 2;
388}
389
390ecb_inline void
391encode_float32 (enc_t *enc, NV nv)
392{
393 need (enc, 1+4);
394
395 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
396
397 uint32_t fp = ecb_float_to_binary32 (nv);
398
399 if (!ecb_big_endian ())
400 fp = ecb_bswap32 (fp);
401
402 memcpy (enc->cur, &fp, 4);
403 enc->cur += 4;
404}
405
406ecb_inline void
407encode_float64 (enc_t *enc, NV nv)
408{
409 need (enc, 1+8);
410
411 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
412
413 uint64_t fp = ecb_double_to_binary64 (nv);
414
415 if (!ecb_big_endian ())
416 fp = ecb_bswap64 (fp);
417
418 memcpy (enc->cur, &fp, 8);
419 enc->cur += 8;
420}
421
422ecb_inline void
423encode_forced (enc_t *enc, UV type, SV *sv)
424{
425 switch (type)
426 {
427 case AS_CBOR:
428 {
429 STRLEN len;
430 char *str = SvPVbyte (sv, len);
431
432 need (enc, len);
433 memcpy (enc->cur, str, len);
434 enc->cur += len;
435 }
436 break;
437
438 case AS_BYTES:
439 {
440 STRLEN len;
441 char *str = SvPVbyte (sv, len);
442 encode_strref (enc, 0, 0, str, len);
443 }
444 break;
445
446 case AS_TEXT:
447 {
448 STRLEN len;
449 char *str = SvPVutf8 (sv, len);
450 encode_strref (enc, 1, 1, str, len);
451 }
452 break;
453
454 case AS_INT: encode_int (enc, sv); break;
455 case AS_FLOAT16: encode_float16 (enc, SvNV (sv)); break;
456 case AS_FLOAT32: encode_float32 (enc, SvNV (sv)); break;
457 case AS_FLOAT64: encode_float64 (enc, SvNV (sv)); break;
458
459 default:
460 croak ("encountered malformed CBOR::XS::Tagged object");
461 }
251} 462}
252 463
253static void encode_sv (enc_t *enc, SV *sv); 464static void encode_sv (enc_t *enc, SV *sv);
254 465
255static void 466static void
260 if (enc->depth >= enc->cbor.max_depth) 471 if (enc->depth >= enc->cbor.max_depth)
261 croak (ERR_NESTING_EXCEEDED); 472 croak (ERR_NESTING_EXCEEDED);
262 473
263 ++enc->depth; 474 ++enc->depth;
264 475
265 encode_uint (enc, 0x80, len + 1); 476 encode_uint (enc, MAJOR_ARRAY, len + 1);
266 477
478 if (ecb_expect_false (SvMAGICAL (av)))
267 for (i = 0; i <= len; ++i) 479 for (i = 0; i <= len; ++i)
268 { 480 {
269 SV **svp = av_fetch (av, i, 0); 481 SV **svp = av_fetch (av, i, 0);
270 encode_sv (enc, svp ? *svp : &PL_sv_undef); 482 encode_sv (enc, svp ? *svp : &PL_sv_undef);
271 } 483 }
484 else
485 for (i = 0; i <= len; ++i)
486 {
487 SV *sv = AvARRAY (av)[i];
488 encode_sv (enc, sv ? sv : &PL_sv_undef);
489 }
272 490
273 --enc->depth; 491 --enc->depth;
274}
275
276ecb_inline void
277encode_he (enc_t *enc, HE *he)
278{
279} 492}
280 493
281static void 494static void
282encode_hv (enc_t *enc, HV *hv) 495encode_hv (enc_t *enc, HV *hv)
283{ 496{
289 ++enc->depth; 502 ++enc->depth;
290 503
291 int pairs = hv_iterinit (hv); 504 int pairs = hv_iterinit (hv);
292 int mg = SvMAGICAL (hv); 505 int mg = SvMAGICAL (hv);
293 506
294 if (mg) 507 if (ecb_expect_false (mg))
295 encode_ch (enc, 0xa0 | 31); 508 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
296 else 509 else
297 encode_uint (enc, 0xa0, pairs); 510 encode_uint (enc, MAJOR_MAP, pairs);
298 511
299 while ((he = hv_iternext (hv))) 512 while ((he = hv_iternext (hv)))
300 { 513 {
301 if (HeKLEN (he) == HEf_SVKEY) 514 if (HeKLEN (he) == HEf_SVKEY)
302 encode_sv (enc, HeSVKEY (he)); 515 encode_sv (enc, HeSVKEY (he));
303 else 516 else
304 encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); 517 encode_strref (enc, enc->cbor.flags & (F_TEXT_KEYS | F_TEXT_STRINGS), HeKUTF8 (he), HeKEY (he), HeKLEN (he));
305 518
306 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); 519 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
307 } 520 }
308 521
309 if (mg) 522 if (ecb_expect_false (mg))
310 encode_ch (enc, 0xe0 | 31); 523 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
311 524
312 --enc->depth; 525 --enc->depth;
313} 526}
314 527
315// encode objects, arrays and special \0=false and \1=true values. 528// encode objects, arrays and special \0=false and \1=true values.
316static void 529static void
317encode_rv (enc_t *enc, SV *sv) 530encode_rv (enc_t *enc, SV *sv)
318{ 531{
319 SvGETMAGIC (sv); 532 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 }
342 533
343 svtype svt = SvTYPE (sv); 534 svtype svt = SvTYPE (sv);
344 535
345 if (ecb_expect_false (SvOBJECT (sv))) 536 if (ecb_expect_false (SvOBJECT (sv)))
346 { 537 {
353 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 544 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
354 ? cbor_tagged_stash 545 ? cbor_tagged_stash
355 : gv_stashpv ("CBOR::XS::Tagged" , 1); 546 : gv_stashpv ("CBOR::XS::Tagged" , 1);
356 547
357 HV *stash = SvSTASH (sv); 548 HV *stash = SvSTASH (sv);
358 GV *method;
359 549
360 if (stash == boolean_stash) 550 if (stash == boolean_stash)
361 encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20); 551 {
552 encode_ch (enc, SvIV (sv) ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE);
553 return;
554 }
362 else if (stash == error_stash) 555 else if (stash == error_stash)
363 encode_ch (enc, 0xe0 | 23); 556 {
557 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
558 return;
559 }
364 else if (stash == tagged_stash) 560 else if (stash == tagged_stash)
365 { 561 {
366 if (svt != SVt_PVAV) 562 if (svt != SVt_PVAV)
367 croak ("encountered CBOR::XS::Tagged object that isn't an array"); 563 croak ("encountered CBOR::XS::Tagged object that isn't an array");
368 564
565 switch (av_len ((AV *)sv))
566 {
567 case 2-1:
568 // actually a tagged value
369 encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); 569 encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
370 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); 570 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1));
571 break;
572
573 case 3-1:
574 // a forced type [value, type, undef]
575 encode_forced (enc, SvUV (*av_fetch ((AV *)sv, 1, 1)), *av_fetch ((AV *)sv, 0, 1));
576 break;
577
578 default:
579 croak ("encountered malformed CBOR::XS::Tagged object");
580 }
581
582 return;
583 }
584 }
585
586 if (ecb_expect_false (SvREFCNT (sv) > 1)
587 && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
588 {
589 if (ecb_expect_false (!enc->shareable))
590 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
591
592 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
593
594 if (SvOK (*svp))
371 } 595 {
596 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
597 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
598 return;
599 }
600 else
601 {
602 sv_setuv (*svp, enc->shareable_idx);
603 ++enc->shareable_idx;
604 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
605 }
606 }
607
608 if (ecb_expect_false (SvOBJECT (sv)))
609 {
610 HV *stash = SvSTASH (sv);
611 GV *method;
612
613 if (enc->cbor.flags & F_FORBID_OBJECTS)
614 croak ("encountered object '%s', but forbid_objects is enabled",
615 SvPV_nolen (sv_2mortal (newRV_inc (sv))));
372 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) 616 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
373 { 617 {
374 dSP; 618 dSP;
375 619
376 ENTER; SAVETMPS; PUSHMARK (SP); 620 ENTER; SAVETMPS;
621 PUSHMARK (SP);
377 // we re-bless the reference to get overload and other niceties right 622 // we re-bless the reference to get overload and other niceties right
378 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 623 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
379 624
380 PUTBACK; 625 PUTBACK;
381 // G_SCALAR ensures that return value is 1 626 // G_SCALAR ensures that return value is 1
394 } 639 }
395 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0) 640 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
396 { 641 {
397 dSP; 642 dSP;
398 643
399 ENTER; SAVETMPS; PUSHMARK (SP); 644 ENTER; SAVETMPS;
645 PUSHMARK (SP);
400 EXTEND (SP, 2); 646 EXTEND (SP, 2);
401 // we re-bless the reference to get overload and other niceties right 647 // we re-bless the reference to get overload and other niceties right
402 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 648 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
403 PUSHs (sv_cbor); 649 PUSHs (sv_cbor);
404 650
409 // catch this surprisingly common error 655 // catch this surprisingly common error
410 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) 656 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)); 657 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
412 658
413 encode_tag (enc, CBOR_TAG_PERL_OBJECT); 659 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
414 encode_uint (enc, 0x80, count + 1); 660 encode_uint (enc, MAJOR_ARRAY, count + 1);
415 encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); 661 encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
416 662
417 while (count) 663 {
664 int i;
665
666 for (i = 0; i < count; ++i)
418 encode_sv (enc, SP[1 - count--]); 667 encode_sv (enc, SP[i + 1 - count]);
668
669 SP -= count;
670 }
419 671
420 PUTBACK; 672 PUTBACK;
421 673
422 FREETMPS; LEAVE; 674 FREETMPS; LEAVE;
423 } 675 }
441{ 693{
442 double nv = SvNVX (sv); 694 double nv = SvNVX (sv);
443 695
444 need (enc, 9); 696 need (enc, 9);
445 697
446 if (ecb_expect_false (nv == (U32)nv)) 698 if (ecb_expect_false (nv == (NV)(U32)nv))
447 encode_uint (enc, 0x00, (U32)nv); 699 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
448 //TODO: maybe I32? 700 //TODO: maybe I32?
449 else if (ecb_expect_false (nv == (float)nv)) 701 else if (ecb_expect_false (nv == (float)nv))
450 { 702 encode_float32 (enc, nv);
451 uint32_t fp = ecb_float_to_binary32 (nv);
452
453 *enc->cur++ = 0xe0 | 26;
454
455 if (!ecb_big_endian ())
456 fp = ecb_bswap32 (fp);
457
458 memcpy (enc->cur, &fp, 4);
459 enc->cur += 4;
460 }
461 else 703 else
462 { 704 encode_float64 (enc, nv);
463 uint64_t fp = ecb_double_to_binary64 (nv);
464
465 *enc->cur++ = 0xe0 | 27;
466
467 if (!ecb_big_endian ())
468 fp = ecb_bswap64 (fp);
469
470 memcpy (enc->cur, &fp, 8);
471 enc->cur += 8;
472 }
473} 705}
474 706
475static void 707static void
476encode_sv (enc_t *enc, SV *sv) 708encode_sv (enc_t *enc, SV *sv)
477{ 709{
479 711
480 if (SvPOKp (sv)) 712 if (SvPOKp (sv))
481 { 713 {
482 STRLEN len; 714 STRLEN len;
483 char *str = SvPV (sv, len); 715 char *str = SvPV (sv, len);
484 encode_str (enc, SvUTF8 (sv), str, len); 716 encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len);
485 } 717 }
486 else if (SvNOKp (sv)) 718 else if (SvNOKp (sv))
487 encode_nv (enc, sv); 719 encode_nv (enc, sv);
488 else if (SvIOKp (sv)) 720 else if (SvIOKp (sv))
489 { 721 encode_int (enc, sv);
490 if (SvIsUV (sv))
491 encode_uint (enc, 0x00, SvUVX (sv));
492 else if (SvIVX (sv) >= 0)
493 encode_uint (enc, 0x00, SvIVX (sv));
494 else
495 encode_uint (enc, 0x20, -(SvIVX (sv) + 1));
496 }
497 else if (SvROK (sv)) 722 else if (SvROK (sv))
498 encode_rv (enc, SvRV (sv)); 723 encode_rv (enc, SvRV (sv));
499 else if (!SvOK (sv)) 724 else if (!SvOK (sv))
500 encode_ch (enc, 0xe0 | 22); 725 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
501 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 726 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
502 encode_ch (enc, 0xe0 | 23); 727 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
503 else 728 else
504 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data", 729 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
505 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 730 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
506} 731}
507 732
508static SV * 733static SV *
509encode_cbor (SV *scalar, CBOR *cbor) 734encode_cbor (SV *scalar, CBOR *cbor)
510{ 735{
511 enc_t enc = { }; 736 enc_t enc = { 0 };
512 737
513 enc.cbor = *cbor; 738 enc.cbor = *cbor;
514 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 739 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
515 enc.cur = SvPVX (enc.sv); 740 enc.cur = SvPVX (enc.sv);
516 enc.end = SvEND (enc.sv); 741 enc.end = SvEND (enc.sv);
517 742
518 SvPOK_only (enc.sv); 743 SvPOK_only (enc.sv);
519 744
520 if (cbor->flags & F_DEDUP_STRINGS) 745 if (cbor->flags & F_PACK_STRINGS)
521 { 746 {
522 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE); 747 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
523 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ()); 748 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
524 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ()); 749 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
525 } 750 }
547 CBOR cbor; 772 CBOR cbor;
548 U32 depth; // recursion depth 773 U32 depth; // recursion depth
549 U32 maxdepth; // recursion depth limit 774 U32 maxdepth; // recursion depth limit
550 AV *shareable; 775 AV *shareable;
551 AV *stringref; 776 AV *stringref;
777 SV *decode_tagged;
778 SV *err_sv; // optional sv for error, needs to be freed
552} dec_t; 779} dec_t;
553 780
554#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE 781// set dec->err to ERRSV
782ecb_cold static void
783err_errsv (dec_t *dec)
784{
785 if (!dec->err)
786 {
787 dec->err_sv = newSVsv (ERRSV);
555 788
556#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data") 789 // chop off the trailing \n
790 SvCUR_set (dec->err_sv, SvCUR (dec->err_sv) - 1);
791 *SvEND (dec->err_sv) = 0;
557 792
793 dec->err = SvPVutf8_nolen (dec->err_sv);
794 }
795}
796
797// the following functions are used to reduce code size and help the compiler to optimise
798ecb_cold static void
799err_set (dec_t *dec, const char *reason)
800{
801 if (!dec->err)
802 dec->err = reason;
803}
804
805ecb_cold static void
806err_unexpected_end (dec_t *dec)
807{
808 err_set (dec, "unexpected end of CBOR data");
809}
810
811#define ERR_DO(do) SB do; goto fail; SE
812#define ERR(reason) ERR_DO (err_set (dec, reason))
813#define ERR_ERRSV ERR_DO (err_errsv (dec))
814
815#define WANT(len) if (ecb_expect_false ((uintptr_t)(dec->end - dec->cur) < (STRLEN)len)) ERR_DO (err_unexpected_end (dec))
816
558#define DEC_INC_DEPTH if (++dec->depth > dec->cbor.max_depth) ERR (ERR_NESTING_EXCEEDED) 817#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 818#define DEC_DEC_DEPTH --dec->depth
560 819
561static UV 820static UV
562decode_uint (dec_t *dec) 821decode_uint (dec_t *dec)
563{ 822{
564 switch (*dec->cur & 31) 823 U8 m = *dec->cur & MINOR_MASK;
565 { 824 ++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;
570 825
571 case 24: 826 if (ecb_expect_true (m < LENGTH_EXT1))
827 return m;
828 else if (ecb_expect_true (m == LENGTH_EXT1))
829 {
572 WANT (2); 830 WANT (1);
573 dec->cur += 2; 831 dec->cur += 1;
574 return dec->cur[-1]; 832 return dec->cur[-1];
575 833 }
576 case 25: 834 else if (ecb_expect_true (m == LENGTH_EXT2))
835 {
577 WANT (3); 836 WANT (2);
578 dec->cur += 3; 837 dec->cur += 2;
579 return (((UV)dec->cur[-2]) << 8) 838 return (((UV)dec->cur[-2]) << 8)
580 | ((UV)dec->cur[-1]); 839 | ((UV)dec->cur[-1]);
581 840 }
582 case 26: 841 else if (ecb_expect_true (m == LENGTH_EXT4))
842 {
583 WANT (5); 843 WANT (4);
584 dec->cur += 5; 844 dec->cur += 4;
585 return (((UV)dec->cur[-4]) << 24) 845 return (((UV)dec->cur[-4]) << 24)
586 | (((UV)dec->cur[-3]) << 16) 846 | (((UV)dec->cur[-3]) << 16)
587 | (((UV)dec->cur[-2]) << 8) 847 | (((UV)dec->cur[-2]) << 8)
588 | ((UV)dec->cur[-1]); 848 | ((UV)dec->cur[-1]);
589 849 }
590 case 27: 850 else if (ecb_expect_true (m == LENGTH_EXT8))
851 {
591 WANT (9); 852 WANT (8);
592 dec->cur += 9; 853 dec->cur += 8;
854
855 return
856#if UVSIZE < 8
857 0
858#else
593 return (((UV)dec->cur[-8]) << 56) 859 (((UV)dec->cur[-8]) << 56)
594 | (((UV)dec->cur[-7]) << 48) 860 | (((UV)dec->cur[-7]) << 48)
595 | (((UV)dec->cur[-6]) << 40) 861 | (((UV)dec->cur[-6]) << 40)
596 | (((UV)dec->cur[-5]) << 32) 862 | (((UV)dec->cur[-5]) << 32)
863#endif
597 | (((UV)dec->cur[-4]) << 24) 864 | (((UV)dec->cur[-4]) << 24)
598 | (((UV)dec->cur[-3]) << 16) 865 | (((UV)dec->cur[-3]) << 16)
599 | (((UV)dec->cur[-2]) << 8) 866 | (((UV)dec->cur[-2]) << 8)
600 | ((UV)dec->cur[-1]); 867 | ((UV)dec->cur[-1]);
601 868 }
602 default: 869 else
603 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 870 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
604 }
605 871
606fail: 872fail:
607 return 0; 873 return 0;
608} 874}
609 875
614{ 880{
615 AV *av = newAV (); 881 AV *av = newAV ();
616 882
617 DEC_INC_DEPTH; 883 DEC_INC_DEPTH;
618 884
619 if ((*dec->cur & 31) == 31) 885 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
620 { 886 {
621 ++dec->cur; 887 ++dec->cur;
622 888
623 for (;;) 889 for (;;)
624 { 890 {
625 WANT (1); 891 WANT (1);
626 892
627 if (*dec->cur == (0xe0 | 31)) 893 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
628 { 894 {
629 ++dec->cur; 895 ++dec->cur;
630 break; 896 break;
631 } 897 }
632 898
633 av_push (av, decode_sv (dec)); 899 av_push (av, decode_sv (dec));
634 } 900 }
635 } 901 }
636 else 902 else
637 { 903 {
638 int i, len = decode_uint (dec); 904 UV i, len = decode_uint (dec);
639 905
906 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); 907 av_fill (av, len - 1);
641 908
642 for (i = 0; i < len; ++i) 909 for (i = 0; i < len; ++i)
643 AvARRAY (av)[i] = decode_sv (dec); 910 AvARRAY (av)[i] = decode_sv (dec);
644 } 911 }
645 912
646 DEC_DEC_DEPTH; 913 DEC_DEC_DEPTH;
647 return newRV_noinc ((SV *)av); 914 return newRV_noinc ((SV *)av);
648 915
649fail: 916fail:
650 SvREFCNT_dec (av); 917 SvREFCNT_dec_NN (av);
651 DEC_DEC_DEPTH; 918 DEC_DEC_DEPTH;
652 return &PL_sv_undef; 919 return &PL_sv_undef;
653} 920}
654 921
655static void 922static void
656decode_he (dec_t *dec, HV *hv) 923decode_he (dec_t *dec, HV *hv)
657{ 924{
658 // for speed reasons, we specialcase single-string 925 // for speed reasons, we specialcase single-string
659 // byte or utf-8 strings as keys, but only when !stringref 926 // byte or utf-8 strings as keys, but only when !stringref
660 927
661 if (expect_true (!dec->stringref)) 928 if (ecb_expect_true (!dec->stringref))
662 if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27) 929 if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
663 { 930 {
664 I32 len = decode_uint (dec); 931 STRLEN len = decode_uint (dec);
665 char *key = (char *)dec->cur; 932 char *key = (char *)dec->cur;
666 933
934 WANT (len);
667 dec->cur += len; 935 dec->cur += len;
668 936
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); 937 hv_store (hv, key, len, decode_sv (dec), 0);
673 938
674 return; 939 return;
675 } 940 }
676 else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27) 941 else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
677 { 942 {
678 I32 len = decode_uint (dec); 943 STRLEN len = decode_uint (dec);
679 char *key = (char *)dec->cur; 944 char *key = (char *)dec->cur;
680 945
946 WANT (len);
681 dec->cur += len; 947 dec->cur += len;
682 948
683 if (ecb_expect_false (dec->stringref)) 949 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
684 av_push (dec->stringref, newSVpvn_utf8 (key, len, 1)); 950 if (!is_utf8_string (key, len))
951 ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
685 952
686 hv_store (hv, key, -len, decode_sv (dec), 0); 953 hv_store (hv, key, -len, decode_sv (dec), 0);
687 954
688 return; 955 return;
689 } 956 }
690 957
691 SV *k = decode_sv (dec); 958 SV *k = decode_sv (dec);
692 SV *v = decode_sv (dec); 959 SV *v = decode_sv (dec);
693 960
961 // we leak memory if uncaught exceptions are thrown by random magical
962 // methods, and this is hopefully the only place where it can happen,
963 // so if there is a chance of an exception, take the very slow path.
964 // since catching exceptions is "undocumented/internal/forbidden" by
965 // the new p5p powers, we need to call out to a perl function :/
966 if (ecb_expect_false (SvAMAGIC (k)))
967 {
968 dSP;
969
970 ENTER; SAVETMPS;
971 PUSHMARK (SP);
972 EXTEND (SP, 3);
973 PUSHs (sv_2mortal (newRV_inc ((SV *)hv)));
974 PUSHs (sv_2mortal (k));
975 PUSHs (sv_2mortal (v));
976
977 PUTBACK;
978 call_pv ("CBOR::XS::_hv_store", G_VOID | G_DISCARD | G_EVAL);
979 SPAGAIN;
980
981 FREETMPS; LEAVE;
982
983 if (SvTRUE (ERRSV))
984 ERR_ERRSV;
985
986 return;
987 }
988
694 hv_store_ent (hv, k, v, 0); 989 hv_store_ent (hv, k, v, 0);
695 SvREFCNT_dec (k); 990 SvREFCNT_dec_NN (k);
991
992fail:
993 ;
696} 994}
697 995
698static SV * 996static SV *
699decode_hv (dec_t *dec) 997decode_hv (dec_t *dec)
700{ 998{
701 HV *hv = newHV (); 999 HV *hv = newHV ();
702 1000
703 DEC_INC_DEPTH; 1001 DEC_INC_DEPTH;
704 1002
705 if ((*dec->cur & 31) == 31) 1003 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
706 { 1004 {
707 ++dec->cur; 1005 ++dec->cur;
708 1006
709 for (;;) 1007 for (;;)
710 { 1008 {
711 WANT (1); 1009 WANT (1);
712 1010
713 if (*dec->cur == (0xe0 | 31)) 1011 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
714 { 1012 {
715 ++dec->cur; 1013 ++dec->cur;
716 break; 1014 break;
717 } 1015 }
718 1016
719 decode_he (dec, hv); 1017 decode_he (dec, hv);
720 } 1018 }
721 } 1019 }
722 else 1020 else
723 { 1021 {
724 int pairs = decode_uint (dec); 1022 UV pairs = decode_uint (dec);
1023
1024 WANT (pairs); // complexity check - need at least one byte per value, do not allow supersize hashes
725 1025
726 while (pairs--) 1026 while (pairs--)
727 decode_he (dec, hv); 1027 decode_he (dec, hv);
728 } 1028 }
729 1029
730 DEC_DEC_DEPTH; 1030 DEC_DEC_DEPTH;
731 return newRV_noinc ((SV *)hv); 1031 return newRV_noinc ((SV *)hv);
732 1032
733fail: 1033fail:
734 SvREFCNT_dec (hv); 1034 SvREFCNT_dec_NN (hv);
735 DEC_DEC_DEPTH; 1035 DEC_DEC_DEPTH;
736 return &PL_sv_undef; 1036 return &PL_sv_undef;
737} 1037}
738 1038
739static SV * 1039static SV *
740decode_str (dec_t *dec, int utf8) 1040decode_str (dec_t *dec, int utf8)
741{ 1041{
742 SV *sv = 0; 1042 SV *sv = 0;
743 1043
744 if ((*dec->cur & 31) == 31) 1044 if (ecb_expect_false ((*dec->cur & MINOR_MASK) == MINOR_INDEF))
745 { 1045 {
1046 // indefinite length strings
746 ++dec->cur; 1047 ++dec->cur;
747 1048
1049 U8 major = *dec->cur & MAJOR_MISC;
1050
748 sv = newSVpvn ("", 0); 1051 sv = newSVpvn ("", 0);
749 1052
750 // not very fast, and certainly not robust against illegal input
751 for (;;) 1053 for (;;)
752 { 1054 {
753 WANT (1); 1055 WANT (1);
754 1056
755 if (*dec->cur == (0xe0 | 31)) 1057 if ((*dec->cur - major) > LENGTH_EXT8)
1058 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
756 { 1059 {
757 ++dec->cur; 1060 ++dec->cur;
758 break; 1061 break;
759 } 1062 }
1063 else
1064 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
760 1065
761 sv_catsv (sv, decode_sv (dec)); 1066 STRLEN len = decode_uint (dec);
1067
1068 WANT (len);
1069 sv_catpvn (sv, dec->cur, len);
1070 dec->cur += len;
762 } 1071 }
763 } 1072 }
764 else 1073 else
765 { 1074 {
766 STRLEN len = decode_uint (dec); 1075 STRLEN len = decode_uint (dec);
767 1076
768 WANT (len); 1077 WANT (len);
769 sv = newSVpvn (dec->cur, len); 1078 sv = newSVpvn (dec->cur, len);
770 dec->cur += len; 1079 dec->cur += len;
1080
1081 if (ecb_expect_false (dec->stringref)
1082 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
1083 av_push (dec->stringref, SvREFCNT_inc_NN (sv));
771 } 1084 }
772 1085
773 if (utf8) 1086 if (utf8)
1087 {
1088 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
1089 if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
1090 ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
1091
774 SvUTF8_on (sv); 1092 SvUTF8_on (sv);
775 1093 }
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 1094
780 return sv; 1095 return sv;
781 1096
782fail: 1097fail:
783 SvREFCNT_dec (sv); 1098 SvREFCNT_dec (sv);
802 sv = newRV_noinc (decode_sv (dec)); 1117 sv = newRV_noinc (decode_sv (dec));
803 break; 1118 break;
804 1119
805 case CBOR_TAG_STRINGREF_NAMESPACE: 1120 case CBOR_TAG_STRINGREF_NAMESPACE:
806 { 1121 {
1122 // do not use SAVETMPS/FREETMPS, as these will
1123 // erase mortalised caches, e.g. "shareable"
807 ENTER; SAVETMPS; 1124 ENTER;
808 1125
809 SAVESPTR (dec->stringref); 1126 SAVESPTR (dec->stringref);
810 dec->stringref = (AV *)sv_2mortal ((SV *)newAV ()); 1127 dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());
811 1128
812 sv = decode_sv (dec); 1129 sv = decode_sv (dec);
813 1130
814 FREETMPS; LEAVE; 1131 LEAVE;
815 } 1132 }
816 break; 1133 break;
817 1134
818 case CBOR_TAG_STRINGREF: 1135 case CBOR_TAG_STRINGREF:
819 { 1136 {
820 if ((*dec->cur >> 5) != 0) 1137 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
821 ERR ("corrupted CBOR data (stringref index not an unsigned integer)"); 1138 ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
822 1139
823 UV idx = decode_uint (dec); 1140 UV idx = decode_uint (dec);
824 1141
825 if (!dec->stringref || (int)idx > AvFILLp (dec->stringref)) 1142 if (!dec->stringref || idx >= (UV)(1 + AvFILLp (dec->stringref)))
826 ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)"); 1143 ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)");
827 1144
828 sv = newSVsv (AvARRAY (dec->stringref)[idx]); 1145 sv = newSVsv (AvARRAY (dec->stringref)[idx]);
829 } 1146 }
830 break; 1147 break;
832 case CBOR_TAG_VALUE_SHAREABLE: 1149 case CBOR_TAG_VALUE_SHAREABLE:
833 { 1150 {
834 if (ecb_expect_false (!dec->shareable)) 1151 if (ecb_expect_false (!dec->shareable))
835 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ()); 1152 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
836 1153
1154 if (dec->cbor.flags & F_ALLOW_CYCLES)
1155 {
837 sv = newSV (0); 1156 sv = newSV (0);
838 av_push (dec->shareable, SvREFCNT_inc_NN (sv)); 1157 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
839 1158
840 SV *osv = decode_sv (dec); 1159 SV *osv = decode_sv (dec);
841 sv_setsv (sv, osv); 1160 sv_setsv (sv, osv);
842 SvREFCNT_dec_NN (osv); 1161 SvREFCNT_dec_NN (osv);
1162 }
1163 else
1164 {
1165 av_push (dec->shareable, &PL_sv_undef);
1166 int idx = AvFILLp (dec->shareable);
1167 sv = decode_sv (dec);
1168 av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
1169 }
843 } 1170 }
844 break; 1171 break;
845 1172
846 case CBOR_TAG_VALUE_SHAREDREF: 1173 case CBOR_TAG_VALUE_SHAREDREF:
847 { 1174 {
848 if ((*dec->cur >> 5) != 0) 1175 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
849 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)"); 1176 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
850 1177
851 UV idx = decode_uint (dec); 1178 UV idx = decode_uint (dec);
852 1179
853 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable)) 1180 if (!dec->shareable || idx >= (UV)(1 + AvFILLp (dec->shareable)))
854 ERR ("corrupted CBOR data (sharedref index out of bounds)"); 1181 ERR ("corrupted CBOR data (sharedref index out of bounds)");
855 1182
856 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]); 1183 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
1184
1185 if (sv == &PL_sv_undef)
1186 ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
857 } 1187 }
858 break; 1188 break;
859 1189
860 case CBOR_TAG_PERL_OBJECT: 1190 case CBOR_TAG_PERL_OBJECT:
861 { 1191 {
1192 if (dec->cbor.flags & F_FORBID_OBJECTS)
1193 goto filter;
1194
862 sv = decode_sv (dec); 1195 sv = decode_sv (dec);
863 1196
864 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) 1197 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
865 ERR ("corrupted CBOR data (non-array perl object)"); 1198 ERR ("corrupted CBOR data (non-array perl object)");
866 1199
876 if (!method) 1209 if (!method)
877 ERR ("cannot decode perl-object (package does not have a THAW method)"); 1210 ERR ("cannot decode perl-object (package does not have a THAW method)");
878 1211
879 dSP; 1212 dSP;
880 1213
881 ENTER; SAVETMPS; PUSHMARK (SP); 1214 ENTER; SAVETMPS;
1215 PUSHMARK (SP);
882 EXTEND (SP, len + 1); 1216 EXTEND (SP, len + 1);
883 // we re-bless the reference to get overload and other niceties right 1217 // we re-bless the reference to get overload and other niceties right
884 PUSHs (*av_fetch (av, 0, 1)); 1218 PUSHs (*av_fetch (av, 0, 1));
885 PUSHs (sv_cbor); 1219 PUSHs (sv_cbor);
886 1220
894 SPAGAIN; 1228 SPAGAIN;
895 1229
896 if (SvTRUE (ERRSV)) 1230 if (SvTRUE (ERRSV))
897 { 1231 {
898 FREETMPS; LEAVE; 1232 FREETMPS; LEAVE;
899 ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV)))); 1233 ERR_ERRSV;
900 } 1234 }
901 1235
902 SvREFCNT_dec (sv); 1236 SvREFCNT_dec_NN (sv);
903 sv = SvREFCNT_inc (POPs); 1237 sv = SvREFCNT_inc (POPs);
904 1238
905 PUTBACK; 1239 PUTBACK;
906 1240
907 FREETMPS; LEAVE; 1241 FREETMPS; LEAVE;
908 } 1242 }
909 break; 1243 break;
910 1244
911 default: 1245 default:
1246 filter:
912 { 1247 {
1248 SV *tag_sv = newSVuv (tag);
1249
913 sv = decode_sv (dec); 1250 sv = decode_sv (dec);
914 1251
1252 dSP;
1253 ENTER; SAVETMPS;
1254 PUSHMARK (SP);
1255 EXTEND (SP, 2);
1256 PUSHs (tag_sv);
1257 PUSHs (sv);
1258
1259 PUTBACK;
1260 int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL);
1261 SPAGAIN;
1262
1263 if (SvTRUE (ERRSV))
1264 {
1265 SvREFCNT_dec_NN (tag_sv);
1266 FREETMPS; LEAVE;
1267 ERR_ERRSV;
1268 }
1269
1270 if (count)
1271 {
1272 SvREFCNT_dec_NN (tag_sv);
1273 SvREFCNT_dec_NN (sv);
1274 sv = SvREFCNT_inc_NN (TOPs);
1275 SP -= count;
1276 }
1277 else
1278 {
915 AV *av = newAV (); 1279 AV *av = newAV ();
916 av_push (av, newSVuv (tag)); 1280 av_push (av, tag_sv);
917 av_push (av, sv); 1281 av_push (av, sv);
918 1282
919 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 1283 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
920 ? cbor_tagged_stash 1284 ? cbor_tagged_stash
921 : gv_stashpv ("CBOR::XS::Tagged" , 1); 1285 : gv_stashpv ("CBOR::XS::Tagged" , 1);
922
923 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash); 1286 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
1287 }
1288
1289 PUTBACK;
1290
1291 FREETMPS; LEAVE;
924 } 1292 }
925 break; 1293 break;
926 } 1294 }
927 1295
928 return sv; 1296 return sv;
935static SV * 1303static SV *
936decode_sv (dec_t *dec) 1304decode_sv (dec_t *dec)
937{ 1305{
938 WANT (1); 1306 WANT (1);
939 1307
940 switch (*dec->cur >> 5) 1308 switch (*dec->cur >> MAJOR_SHIFT)
941 { 1309 {
942 case 0: // unsigned int 1310 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
943 return newSVuv (decode_uint (dec)); 1311 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
944 case 1: // negative int 1312 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
945 return newSViv (-1 - (IV)decode_uint (dec)); 1313 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
946 case 2: // octet string 1314 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
947 return decode_str (dec, 0); 1315 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
948 case 3: // utf-8 string 1316 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
949 return decode_str (dec, 1); 1317
950 case 4: // array 1318 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) 1319 switch (*dec->cur++ & MINOR_MASK)
958 { 1320 {
959 case 20: 1321 case SIMPLE_FALSE:
960#if CBOR_SLOW 1322#if CBOR_SLOW
961 types_false = get_bool ("Types::Serialiser::false"); 1323 types_false = get_bool ("Types::Serialiser::false");
962#endif 1324#endif
963 return newSVsv (types_false); 1325 return newSVsv (types_false);
964 case 21: 1326 case SIMPLE_TRUE:
965#if CBOR_SLOW 1327#if CBOR_SLOW
966 types_true = get_bool ("Types::Serialiser::true"); 1328 types_true = get_bool ("Types::Serialiser::true");
967#endif 1329#endif
968 return newSVsv (types_true); 1330 return newSVsv (types_true);
969 case 22: 1331 case SIMPLE_NULL:
970 return newSVsv (&PL_sv_undef); 1332 return newSVsv (&PL_sv_undef);
971 case 23: 1333 case SIMPLE_UNDEF:
972#if CBOR_SLOW 1334#if CBOR_SLOW
973 types_error = get_bool ("Types::Serialiser::error"); 1335 types_error = get_bool ("Types::Serialiser::error");
974#endif 1336#endif
975 return newSVsv (types_error); 1337 return newSVsv (types_error);
976 1338
977 case 25: 1339 case MISC_FLOAT16:
978 { 1340 {
979 WANT (2); 1341 WANT (2);
980 1342
981 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1343 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
982 dec->cur += 2; 1344 dec->cur += 2;
983 1345
984 return newSVnv (ecb_binary16_to_float (fp)); 1346 return newSVnv (ecb_binary16_to_float (fp));
985 } 1347 }
986 1348
987 case 26: 1349 case MISC_FLOAT32:
988 { 1350 {
989 uint32_t fp; 1351 uint32_t fp;
990 WANT (4); 1352 WANT (4);
991 memcpy (&fp, dec->cur, 4); 1353 memcpy (&fp, dec->cur, 4);
992 dec->cur += 4; 1354 dec->cur += 4;
995 fp = ecb_bswap32 (fp); 1357 fp = ecb_bswap32 (fp);
996 1358
997 return newSVnv (ecb_binary32_to_float (fp)); 1359 return newSVnv (ecb_binary32_to_float (fp));
998 } 1360 }
999 1361
1000 case 27: 1362 case MISC_FLOAT64:
1001 { 1363 {
1002 uint64_t fp; 1364 uint64_t fp;
1003 WANT (8); 1365 WANT (8);
1004 memcpy (&fp, dec->cur, 8); 1366 memcpy (&fp, dec->cur, 8);
1005 dec->cur += 8; 1367 dec->cur += 8;
1008 fp = ecb_bswap64 (fp); 1370 fp = ecb_bswap64 (fp);
1009 1371
1010 return newSVnv (ecb_binary64_to_double (fp)); 1372 return newSVnv (ecb_binary64_to_double (fp));
1011 } 1373 }
1012 1374
1013 // 0..19 unassigned 1375 // 0..19 unassigned simple
1014 // 24 reserved + unassigned (reserved values are not encodable) 1376 // 24 reserved + unassigned simple (reserved values are not encodable)
1377 // 28-30 unassigned misc
1378 // 31 break code
1015 default: 1379 default:
1016 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1380 ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1017 } 1381 }
1018 1382
1019 break; 1383 break;
1020 } 1384 }
1021 1385
1024} 1388}
1025 1389
1026static SV * 1390static SV *
1027decode_cbor (SV *string, CBOR *cbor, char **offset_return) 1391decode_cbor (SV *string, CBOR *cbor, char **offset_return)
1028{ 1392{
1029 dec_t dec = { }; 1393 dec_t dec = { 0 };
1030 SV *sv; 1394 SV *sv;
1031 STRLEN len; 1395 STRLEN len;
1032 char *data = SvPVbyte (string, len); 1396 char *data = SvPVbyte (string, len);
1033 1397
1034 if (len > cbor->max_size && cbor->max_size) 1398 if (len > cbor->max_size && cbor->max_size)
1048 if (dec.cur != dec.end && !dec.err) 1412 if (dec.cur != dec.end && !dec.err)
1049 dec.err = "garbage after CBOR object"; 1413 dec.err = "garbage after CBOR object";
1050 1414
1051 if (dec.err) 1415 if (dec.err)
1052 { 1416 {
1417 if (dec.shareable)
1418 {
1419 // need to break cyclic links, which would all be in shareable
1420 int i;
1421 SV **svp;
1422
1423 for (i = av_len (dec.shareable) + 1; i--; )
1424 if ((svp = av_fetch (dec.shareable, i, 0)))
1425 sv_setsv (*svp, &PL_sv_undef);
1426 }
1427
1053 SvREFCNT_dec (sv); 1428 SvREFCNT_dec_NN (sv);
1429
1430 if (dec.err_sv)
1431 sv_2mortal (dec.err_sv);
1432
1054 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur); 1433 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur);
1055 } 1434 }
1056 1435
1057 sv = sv_2mortal (sv); 1436 sv = sv_2mortal (sv);
1058 1437
1059 return sv; 1438 return sv;
1060} 1439}
1061 1440
1441/////////////////////////////////////////////////////////////////////////////
1442// incremental parser
1443
1444#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
1445
1446// returns 0 for notyet, 1 for success or error
1447static int
1448incr_parse (CBOR *self, SV *cborstr)
1449{
1450 STRLEN cur;
1451 SvPV (cborstr, cur);
1452
1453 while (ecb_expect_true (self->incr_need <= cur))
1454 {
1455 // table of integer count bytes
1456 static I8 incr_len[MINOR_MASK + 1] = {
1457 0, 0, 0, 0, 0, 0, 0, 0,
1458 0, 0, 0, 0, 0, 0, 0, 0,
1459 0, 0, 0, 0, 0, 0, 0, 0,
1460 1, 2, 4, 8,-1,-1,-1,-2
1461 };
1462
1463 const U8 *p = SvPVX (cborstr) + self->incr_pos;
1464 U8 m = *p & MINOR_MASK;
1465 IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1466 I8 ilen = incr_len[m];
1467
1468 self->incr_need = self->incr_pos + 1;
1469
1470 if (ecb_expect_false (ilen < 0))
1471 {
1472 if (m != MINOR_INDEF)
1473 return 1; // error
1474
1475 if (*p == (MAJOR_MISC | MINOR_INDEF))
1476 {
1477 if (count >= 0)
1478 return 1; // error
1479
1480 count = 1;
1481 }
1482 else
1483 {
1484 av_push (self->incr_count, newSViv (-1)); //TODO: nest
1485 count = -1;
1486 }
1487 }
1488 else
1489 {
1490 self->incr_need += ilen;
1491 if (ecb_expect_false (self->incr_need > cur))
1492 return 0;
1493
1494 int major = *p >> MAJOR_SHIFT;
1495
1496 switch (major)
1497 {
1498 case MAJOR_TAG >> MAJOR_SHIFT:
1499 ++count; // tags merely prefix another value
1500 break;
1501
1502 case MAJOR_BYTES >> MAJOR_SHIFT:
1503 case MAJOR_TEXT >> MAJOR_SHIFT:
1504 case MAJOR_ARRAY >> MAJOR_SHIFT:
1505 case MAJOR_MAP >> MAJOR_SHIFT:
1506 {
1507 UV len;
1508
1509 if (ecb_expect_false (ilen))
1510 {
1511 len = 0;
1512
1513 do {
1514 len = (len << 8) | *++p;
1515 } while (--ilen);
1516 }
1517 else
1518 len = m;
1519
1520 switch (major)
1521 {
1522 case MAJOR_BYTES >> MAJOR_SHIFT:
1523 case MAJOR_TEXT >> MAJOR_SHIFT:
1524 self->incr_need += len;
1525 if (ecb_expect_false (self->incr_need > cur))
1526 return 0;
1527
1528 break;
1529
1530 case MAJOR_MAP >> MAJOR_SHIFT:
1531 len <<= 1;
1532 case MAJOR_ARRAY >> MAJOR_SHIFT:
1533 if (len)
1534 {
1535 av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
1536 count = len + 1;
1537 }
1538 break;
1539 }
1540 }
1541 }
1542 }
1543
1544 self->incr_pos = self->incr_need;
1545
1546 if (count > 0)
1547 {
1548 while (!--count)
1549 {
1550 if (!AvFILLp (self->incr_count))
1551 return 1; // done
1552
1553 SvREFCNT_dec_NN (av_pop (self->incr_count));
1554 count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1555 }
1556
1557 SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
1558 }
1559 }
1560
1561 return 0;
1562}
1563
1564
1062///////////////////////////////////////////////////////////////////////////// 1565/////////////////////////////////////////////////////////////////////////////
1063// XS interface functions 1566// XS interface functions
1064 1567
1065MODULE = CBOR::XS PACKAGE = CBOR::XS 1568MODULE = CBOR::XS PACKAGE = CBOR::XS
1066 1569
1074 1577
1075 types_true = get_bool ("Types::Serialiser::true" ); 1578 types_true = get_bool ("Types::Serialiser::true" );
1076 types_false = get_bool ("Types::Serialiser::false"); 1579 types_false = get_bool ("Types::Serialiser::false");
1077 types_error = get_bool ("Types::Serialiser::error"); 1580 types_error = get_bool ("Types::Serialiser::error");
1078 1581
1582 default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1583
1079 sv_cbor = newSVpv ("CBOR", 0); 1584 sv_cbor = newSVpv ("CBOR", 0);
1080 SvREADONLY_on (sv_cbor); 1585 SvREADONLY_on (sv_cbor);
1586
1587 assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1));
1081} 1588}
1082 1589
1083PROTOTYPES: DISABLE 1590PROTOTYPES: DISABLE
1084 1591
1085void CLONE (...) 1592void CLONE (...)
1104void shrink (CBOR *self, int enable = 1) 1611void shrink (CBOR *self, int enable = 1)
1105 ALIAS: 1612 ALIAS:
1106 shrink = F_SHRINK 1613 shrink = F_SHRINK
1107 allow_unknown = F_ALLOW_UNKNOWN 1614 allow_unknown = F_ALLOW_UNKNOWN
1108 allow_sharing = F_ALLOW_SHARING 1615 allow_sharing = F_ALLOW_SHARING
1616 allow_cycles = F_ALLOW_CYCLES
1617 forbid_objects = F_FORBID_OBJECTS
1109 dedup_strings = F_DEDUP_STRINGS 1618 pack_strings = F_PACK_STRINGS
1619 text_keys = F_TEXT_KEYS
1620 text_strings = F_TEXT_STRINGS
1621 validate_utf8 = F_VALIDATE_UTF8
1110 PPCODE: 1622 PPCODE:
1111{ 1623{
1112 if (enable) 1624 if (enable)
1113 self->flags |= ix; 1625 self->flags |= ix;
1114 else 1626 else
1120void get_shrink (CBOR *self) 1632void get_shrink (CBOR *self)
1121 ALIAS: 1633 ALIAS:
1122 get_shrink = F_SHRINK 1634 get_shrink = F_SHRINK
1123 get_allow_unknown = F_ALLOW_UNKNOWN 1635 get_allow_unknown = F_ALLOW_UNKNOWN
1124 get_allow_sharing = F_ALLOW_SHARING 1636 get_allow_sharing = F_ALLOW_SHARING
1637 get_allow_cycles = F_ALLOW_CYCLES
1638 get_forbid_objects = F_FORBID_OBJECTS
1125 get_dedup_strings = F_DEDUP_STRINGS 1639 get_pack_strings = F_PACK_STRINGS
1640 get_text_keys = F_TEXT_KEYS
1641 get_text_strings = F_TEXT_STRINGS
1642 get_validate_utf8 = F_VALIDATE_UTF8
1126 PPCODE: 1643 PPCODE:
1127 XPUSHs (boolSV (self->flags & ix)); 1644 XPUSHs (boolSV (self->flags & ix));
1128 1645
1129void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1646void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1130 PPCODE: 1647 PPCODE:
1143 XPUSHs (ST (0)); 1660 XPUSHs (ST (0));
1144 1661
1145int get_max_size (CBOR *self) 1662int get_max_size (CBOR *self)
1146 CODE: 1663 CODE:
1147 RETVAL = self->max_size; 1664 RETVAL = self->max_size;
1665 OUTPUT:
1666 RETVAL
1667
1668void filter (CBOR *self, SV *filter = 0)
1669 PPCODE:
1670 SvREFCNT_dec (self->filter);
1671 self->filter = filter ? newSVsv (filter) : filter;
1672 XPUSHs (ST (0));
1673
1674SV *get_filter (CBOR *self)
1675 CODE:
1676 RETVAL = self->filter ? self->filter : NEWSV (0, 0);
1148 OUTPUT: 1677 OUTPUT:
1149 RETVAL 1678 RETVAL
1150 1679
1151void encode (CBOR *self, SV *scalar) 1680void encode (CBOR *self, SV *scalar)
1152 PPCODE: 1681 PPCODE:
1167 EXTEND (SP, 2); 1696 EXTEND (SP, 2);
1168 PUSHs (sv); 1697 PUSHs (sv);
1169 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1698 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1170} 1699}
1171 1700
1701void incr_parse (CBOR *self, SV *cborstr)
1702 ALIAS:
1703 incr_parse_multiple = 1
1704 PPCODE:
1705{
1706 if (SvUTF8 (cborstr))
1707 sv_utf8_downgrade (cborstr, 0);
1708
1709 if (!self->incr_count)
1710 {
1711 self->incr_count = newAV ();
1712 self->incr_pos = 0;
1713 self->incr_need = 1;
1714
1715 av_push (self->incr_count, newSViv (1));
1716 }
1717
1718 do
1719 {
1720 if (!incr_parse (self, cborstr))
1721 {
1722 if (self->incr_need > self->max_size && self->max_size)
1723 croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
1724 (unsigned long)self->incr_need, (unsigned long)self->max_size);
1725
1726 break;
1727 }
1728
1729 SV *sv;
1730 char *offset;
1731
1732 PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
1733 XPUSHs (sv);
1734
1735 sv_chop (cborstr, offset);
1736
1737 av_clear (self->incr_count);
1738 av_push (self->incr_count, newSViv (1));
1739
1740 self->incr_pos = 0;
1741 self->incr_need = self->incr_pos + 1;
1742 }
1743 while (ix);
1744}
1745
1746void incr_reset (CBOR *self)
1747 CODE:
1748{
1749 SvREFCNT_dec (self->incr_count);
1750 self->incr_count = 0;
1751}
1752
1753void DESTROY (CBOR *self)
1754 PPCODE:
1755 cbor_free (self);
1756
1172PROTOTYPES: ENABLE 1757PROTOTYPES: ENABLE
1173 1758
1174void encode_cbor (SV *scalar) 1759void encode_cbor (SV *scalar)
1760 ALIAS:
1761 encode_cbor = 0
1762 encode_cbor_sharing = F_ALLOW_SHARING
1175 PPCODE: 1763 PPCODE:
1176{ 1764{
1177 CBOR cbor; 1765 CBOR cbor;
1178 cbor_init (&cbor); 1766 cbor_init (&cbor);
1767 cbor.flags |= ix;
1179 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1768 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1180 XPUSHs (scalar); 1769 XPUSHs (scalar);
1181} 1770}
1182 1771
1183void decode_cbor (SV *cborstr) 1772void decode_cbor (SV *cborstr)
1187 cbor_init (&cbor); 1776 cbor_init (&cbor);
1188 PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN; 1777 PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN;
1189 XPUSHs (cborstr); 1778 XPUSHs (cborstr);
1190} 1779}
1191 1780
1781#ifdef __AFL_COMPILER
1782
1783void
1784afl_init ()
1785 CODE:
1786 __AFL_INIT ();
1787
1788int
1789afl_loop (unsigned int count = 10000)
1790 CODE:
1791 RETVAL = __AFL_LOOP (count);
1792 OUTPUT:
1793 RETVAL
1794
1795#endif
1796

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines