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 1.7 by root, Sat Oct 26 23:02:55 2013 UTC vs.
Revision 1.66 by root, Sun Nov 29 21:32:01 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
16// compatibility with perl <5.18
17#ifndef HvNAMELEN_get
18# define HvNAMELEN_get(hv) strlen (HvNAME (hv))
19#endif
20#ifndef HvNAMELEN
21# define HvNAMELEN(hv) HvNAMELEN_get (hv)
22#endif
23#ifndef HvNAMEUTF8
24# define HvNAMEUTF8(hv) 0
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};
69
70// known tags
71enum cbor_tag
72{
73 // extensions
74 CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
75 CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
76 CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
77 CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
78 CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
79 CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref
80 CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
81
82 // rfc7049
83 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
84 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
85 CBOR_TAG_POS_BIGNUM = 2, // byte string
86 CBOR_TAG_NEG_BIGNUM = 3, // byte string
87 CBOR_TAG_DECIMAL = 4, // decimal fraction, array
88 CBOR_TAG_BIGFLOAT = 5, // array
89
90 CBOR_TAG_CONV_B64U = 21, // base64url, any
91 CBOR_TAG_CONV_B64 = 22, // base64, any
92 CBOR_TAG_CONV_HEX = 23, // base16, any
93 CBOR_TAG_CBOR = 24, // embedded cbor, byte string
94
95 CBOR_TAG_URI = 32, // URI rfc3986, utf-8
96 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
97 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
98 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
99 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
100
101 CBOR_TAG_MAGIC = 55799, // self-describe cbor
102};
103
104// known forced types, also hardcoded in CBOR.pm
105enum
106{
107 AS_CBOR = 0,
108 AS_BYTES = 1,
109 AS_TEXT = 2,
110 AS_FLOAT16 = 3,
111 AS_FLOAT32 = 4,
112 AS_FLOAT64 = 5,
113 // possibly future enhancements: float, integer
114};
115
14#define F_SHRINK 0x00000200UL 116#define F_SHRINK 0x00000001UL
15#define F_ALLOW_UNKNOWN 0x00002000UL 117#define F_ALLOW_UNKNOWN 0x00000002UL
118#define F_ALLOW_SHARING 0x00000004UL
119#define F_ALLOW_CYCLES 0x00000008UL
120#define F_FORBID_OBJECTS 0x00000010UL
121#define F_PACK_STRINGS 0x00000020UL
122#define F_TEXT_KEYS 0x00000040UL
123#define F_TEXT_STRINGS 0x00000080UL
124#define F_VALIDATE_UTF8 0x00000100UL
16 125
17#define INIT_SIZE 32 // initial scalar size to be allocated 126#define INIT_SIZE 32 // initial scalar size to be allocated
18 127
19#define SB do { 128#define SB do {
20#define SE } while (0) 129#define SE } while (0)
31#else 140#else
32# define CBOR_SLOW 0 141# define CBOR_SLOW 0
33# define CBOR_STASH cbor_stash 142# define CBOR_STASH cbor_stash
34#endif 143#endif
35 144
36static HV *cbor_stash, *cbor_boolean_stash, *cbor_tagged_stash; // CBOR::XS:: 145static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS::
37static SV *cbor_true, *cbor_false; 146static SV *types_true, *types_false, *types_error, *sv_cbor, *default_filter;
38 147
39typedef struct { 148typedef struct {
40 U32 flags; 149 U32 flags;
41 U32 max_depth; 150 U32 max_depth;
42 STRLEN max_size; 151 STRLEN max_size;
152 SV *filter;
153
154 // for the incremental parser
155 STRLEN incr_pos; // the current offset into the text
156 STRLEN incr_need; // minimum bytes needed to decode
157 AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef.
43} CBOR; 158} CBOR;
44 159
45ecb_inline void 160ecb_inline void
46cbor_init (CBOR *cbor) 161cbor_init (CBOR *cbor)
47{ 162{
48 Zero (cbor, 1, CBOR); 163 Zero (cbor, 1, CBOR);
49 cbor->max_depth = 512; 164 cbor->max_depth = 512;
165}
166
167ecb_inline void
168cbor_free (CBOR *cbor)
169{
170 SvREFCNT_dec (cbor->filter);
171 SvREFCNT_dec (cbor->incr_count);
50} 172}
51 173
52///////////////////////////////////////////////////////////////////////////// 174/////////////////////////////////////////////////////////////////////////////
53// utility functions 175// utility functions
54 176
76 SvPV_renew (sv, SvCUR (sv) + 1); 198 SvPV_renew (sv, SvCUR (sv) + 1);
77#endif 199#endif
78 } 200 }
79} 201}
80 202
81///////////////////////////////////////////////////////////////////////////// 203// minimum length of a string to be registered for stringref
82// fp hell 204ecb_inline int
83 205minimum_string_length (UV idx)
84//TODO 206{
207 return idx <= 23 ? 3
208 : idx <= 0xffU ? 4
209 : idx <= 0xffffU ? 5
210 : idx <= 0xffffffffU ? 7
211 : 11;
212}
85 213
86///////////////////////////////////////////////////////////////////////////// 214/////////////////////////////////////////////////////////////////////////////
87// encoder 215// encoder
88 216
89// structure used for encoding CBOR 217// structure used for encoding CBOR
92 char *cur; // SvPVX (sv) + current output position 220 char *cur; // SvPVX (sv) + current output position
93 char *end; // SvEND (sv) 221 char *end; // SvEND (sv)
94 SV *sv; // result scalar 222 SV *sv; // result scalar
95 CBOR cbor; 223 CBOR cbor;
96 U32 depth; // recursion level 224 U32 depth; // recursion level
225 HV *stringref[2]; // string => index, or 0 ([0] = bytes, [1] = utf-8)
226 UV stringref_idx;
227 HV *shareable; // ptr => index, or 0
228 UV shareable_idx;
97} enc_t; 229} enc_t;
98 230
99ecb_inline void 231ecb_inline void
100need (enc_t *enc, STRLEN len) 232need (enc_t *enc, STRLEN len)
101{ 233{
102 if (ecb_expect_false (enc->cur + len >= enc->end)) 234 if (ecb_expect_false ((uintptr_t)(enc->end - enc->cur) < len))
103 { 235 {
104 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv); 236 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv);
105 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1); 237 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1);
106 enc->cur = SvPVX (enc->sv) + cur; 238 enc->cur = SvPVX (enc->sv) + cur;
107 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1; 239 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;
113{ 245{
114 need (enc, 1); 246 need (enc, 1);
115 *enc->cur++ = ch; 247 *enc->cur++ = ch;
116} 248}
117 249
250// used for tags, intregers, element counts and so on
118static void 251static void
119encode_uint (enc_t *enc, int major, UV len) 252encode_uint (enc_t *enc, int major, UV len)
120{ 253{
121 need (enc, 9); 254 need (enc, 9);
122 255
123 if (len < 24) 256 if (ecb_expect_true (len < LENGTH_EXT1))
124 *enc->cur++ = major | len; 257 *enc->cur++ = major | len;
125 else if (len <= 0xff) 258 else if (ecb_expect_true (len <= 0xffU))
126 { 259 {
127 *enc->cur++ = major | 24; 260 *enc->cur++ = major | LENGTH_EXT1;
128 *enc->cur++ = len; 261 *enc->cur++ = len;
129 } 262 }
130 else if (len <= 0xffff) 263 else if (len <= 0xffffU)
131 { 264 {
132 *enc->cur++ = major | 25; 265 *enc->cur++ = major | LENGTH_EXT2;
133 *enc->cur++ = len >> 8; 266 *enc->cur++ = len >> 8;
134 *enc->cur++ = len; 267 *enc->cur++ = len;
135 } 268 }
136 else if (len <= 0xffffffff) 269 else if (len <= 0xffffffffU)
137 { 270 {
138 *enc->cur++ = major | 26; 271 *enc->cur++ = major | LENGTH_EXT4;
139 *enc->cur++ = len >> 24; 272 *enc->cur++ = len >> 24;
140 *enc->cur++ = len >> 16; 273 *enc->cur++ = len >> 16;
141 *enc->cur++ = len >> 8; 274 *enc->cur++ = len >> 8;
142 *enc->cur++ = len; 275 *enc->cur++ = len;
143 } 276 }
144 else 277 else
145 { 278 {
146 *enc->cur++ = major | 27; 279 *enc->cur++ = major | LENGTH_EXT8;
147 *enc->cur++ = len >> 56; 280 *enc->cur++ = len >> 56;
148 *enc->cur++ = len >> 48; 281 *enc->cur++ = len >> 48;
149 *enc->cur++ = len >> 40; 282 *enc->cur++ = len >> 40;
150 *enc->cur++ = len >> 32; 283 *enc->cur++ = len >> 32;
151 *enc->cur++ = len >> 24; 284 *enc->cur++ = len >> 24;
153 *enc->cur++ = len >> 8; 286 *enc->cur++ = len >> 8;
154 *enc->cur++ = len; 287 *enc->cur++ = len;
155 } 288 }
156} 289}
157 290
158static void 291// encodes a perl value into a CBOR integer
292ecb_inline
293encode_int (enc_t *enc, SV *sv)
294{
295 if (SvIsUV (sv))
296 encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
297 else if (SvIVX (sv) >= 0)
298 encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
299 else
300 encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
301}
302
303ecb_inline void
304encode_tag (enc_t *enc, UV tag)
305{
306 encode_uint (enc, MAJOR_TAG, tag);
307}
308
309// exceptional (hopefully) slow path for byte strings that need to be utf8-encoded
310ecb_noinline static void
159encode_str (enc_t *enc, int utf8, char *str, STRLEN len) 311encode_str_utf8 (enc_t *enc, int utf8, char *str, STRLEN len)
160{ 312{
161 encode_uint (enc, utf8 ? 0x60 : 0x40, len); 313 STRLEN ulen = len;
314 U8 *p, *pend = (U8 *)str + len;
315
316 for (p = (U8 *)str; p < pend; ++p)
317 ulen += *p >> 7; // count set high bits
318
319 encode_uint (enc, MAJOR_TEXT, ulen);
320
321 need (enc, ulen);
322 for (p = (U8 *)str; p < pend; ++p)
323 if (*p < 0x80)
324 *enc->cur++ = *p;
325 else
326 {
327 *enc->cur++ = 0xc0 + (*p >> 6);
328 *enc->cur++ = 0x80 + (*p & 63);
329 }
330}
331
332ecb_inline void
333encode_str (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
334{
335 if (ecb_expect_false (upgrade_utf8))
336 if (!utf8)
337 {
338 encode_str_utf8 (enc, utf8, str, len);
339 return;
340 }
341
342 encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
162 need (enc, len); 343 need (enc, len);
163 memcpy (enc->cur, str, len); 344 memcpy (enc->cur, str, len);
164 enc->cur += len; 345 enc->cur += len;
165} 346}
166 347
348ecb_inline void
349encode_strref (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
350{
351 if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
352 {
353 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
354
355 if (SvOK (*svp))
356 {
357 // already registered, use stringref
358 encode_tag (enc, CBOR_TAG_STRINGREF);
359 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
360 return;
361 }
362 else if (len >= minimum_string_length (enc->stringref_idx))
363 {
364 // register only
365 sv_setuv (*svp, enc->stringref_idx);
366 ++enc->stringref_idx;
367 }
368 }
369
370 encode_str (enc, upgrade_utf8, utf8, str, len);
371}
372
373ecb_inline void
374encode_float16 (enc_t *enc, NV nv)
375{
376 need (enc, 1+2);
377
378 *enc->cur++ = MAJOR_MISC | MISC_FLOAT16;
379
380 uint16_t fp = ecb_float_to_binary16 (nv);
381
382 if (!ecb_big_endian ())
383 fp = ecb_bswap16 (fp);
384
385 memcpy (enc->cur, &fp, 2);
386 enc->cur += 2;
387}
388
389ecb_inline void
390encode_float32 (enc_t *enc, NV nv)
391{
392 need (enc, 1+4);
393
394 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
395
396 uint32_t fp = ecb_float_to_binary32 (nv);
397
398 if (!ecb_big_endian ())
399 fp = ecb_bswap32 (fp);
400
401 memcpy (enc->cur, &fp, 4);
402 enc->cur += 4;
403}
404
405ecb_inline void
406encode_float64 (enc_t *enc, NV nv)
407{
408 need (enc, 1+8);
409
410 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
411
412 uint64_t fp = ecb_double_to_binary64 (nv);
413
414 if (!ecb_big_endian ())
415 fp = ecb_bswap64 (fp);
416
417 memcpy (enc->cur, &fp, 8);
418 enc->cur += 8;
419}
420
421ecb_inline void
422encode_forced (enc_t *enc, UV type, SV *sv)
423{
424 switch (type)
425 {
426 case AS_CBOR:
427 {
428 STRLEN len;
429 char *str = SvPVbyte (sv, len);
430
431 need (enc, len);
432 memcpy (enc->cur, str, len);
433 enc->cur += len;
434 }
435 break;
436
437 case AS_BYTES:
438 {
439 STRLEN len;
440 char *str = SvPVbyte (sv, len);
441 encode_strref (enc, 0, 0, str, len);
442 }
443 break;
444
445 case AS_TEXT:
446 {
447 STRLEN len;
448 char *str = SvPVutf8 (sv, len);
449 encode_strref (enc, 1, 1, str, len);
450 }
451 break;
452
453 case AS_FLOAT16: encode_float16 (enc, SvNV (sv)); break;
454 case AS_FLOAT32: encode_float32 (enc, SvNV (sv)); break;
455 case AS_FLOAT64: encode_float64 (enc, SvNV (sv)); break;
456
457 default:
458 croak ("encountered malformed CBOR::XS::Tagged object");
459 }
460}
461
167static void encode_sv (enc_t *enc, SV *sv); 462static void encode_sv (enc_t *enc, SV *sv);
168 463
169static void 464static void
170encode_av (enc_t *enc, AV *av) 465encode_av (enc_t *enc, AV *av)
171{ 466{
174 if (enc->depth >= enc->cbor.max_depth) 469 if (enc->depth >= enc->cbor.max_depth)
175 croak (ERR_NESTING_EXCEEDED); 470 croak (ERR_NESTING_EXCEEDED);
176 471
177 ++enc->depth; 472 ++enc->depth;
178 473
179 encode_uint (enc, 0x80, len + 1); 474 encode_uint (enc, MAJOR_ARRAY, len + 1);
180 475
476 if (ecb_expect_false (SvMAGICAL (av)))
181 for (i = 0; i <= len; ++i) 477 for (i = 0; i <= len; ++i)
182 { 478 {
183 SV **svp = av_fetch (av, i, 0); 479 SV **svp = av_fetch (av, i, 0);
184 encode_sv (enc, svp ? *svp : &PL_sv_undef); 480 encode_sv (enc, svp ? *svp : &PL_sv_undef);
185 } 481 }
482 else
483 for (i = 0; i <= len; ++i)
484 {
485 SV *sv = AvARRAY (av)[i];
486 encode_sv (enc, sv ? sv : &PL_sv_undef);
487 }
186 488
187 --enc->depth; 489 --enc->depth;
188} 490}
189 491
190static void 492static void
198 ++enc->depth; 500 ++enc->depth;
199 501
200 int pairs = hv_iterinit (hv); 502 int pairs = hv_iterinit (hv);
201 int mg = SvMAGICAL (hv); 503 int mg = SvMAGICAL (hv);
202 504
203 if (mg) 505 if (ecb_expect_false (mg))
204 encode_ch (enc, 0xa0 | 31); 506 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
205 else 507 else
206 encode_uint (enc, 0xa0, pairs); 508 encode_uint (enc, MAJOR_MAP, pairs);
207 509
208 while ((he = hv_iternext (hv))) 510 while ((he = hv_iternext (hv)))
209 { 511 {
210 if (HeKLEN (he) == HEf_SVKEY) 512 if (HeKLEN (he) == HEf_SVKEY)
211 encode_sv (enc, HeSVKEY (he)); 513 encode_sv (enc, HeSVKEY (he));
212 else 514 else
213 encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); 515 encode_strref (enc, enc->cbor.flags & (F_TEXT_KEYS | F_TEXT_STRINGS), HeKUTF8 (he), HeKEY (he), HeKLEN (he));
214 516
215 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); 517 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
216 } 518 }
217 519
218 if (mg) 520 if (ecb_expect_false (mg))
219 encode_ch (enc, 0xe0 | 31); 521 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
220 522
221 --enc->depth; 523 --enc->depth;
222} 524}
223 525
224// encode objects, arrays and special \0=false and \1=true values. 526// encode objects, arrays and special \0=false and \1=true values.
225static void 527static void
226encode_rv (enc_t *enc, SV *sv) 528encode_rv (enc_t *enc, SV *sv)
227{ 529{
228 svtype svt;
229
230 SvGETMAGIC (sv); 530 SvGETMAGIC (sv);
531
231 svt = SvTYPE (sv); 532 svtype svt = SvTYPE (sv);
232 533
233 if (ecb_expect_false (SvOBJECT (sv))) 534 if (ecb_expect_false (SvOBJECT (sv)))
234 { 535 {
235 HV *boolean_stash = !CBOR_SLOW || cbor_boolean_stash 536 HV *boolean_stash = !CBOR_SLOW || types_boolean_stash
236 ? cbor_boolean_stash 537 ? types_boolean_stash
237 : gv_stashpv ("CBOR::XS::Boolean", 1); 538 : gv_stashpv ("Types::Serialiser::Boolean", 1);
539 HV *error_stash = !CBOR_SLOW || types_error_stash
540 ? types_error_stash
541 : gv_stashpv ("Types::Serialiser::Error", 1);
238 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 542 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
239 ? cbor_tagged_stash 543 ? cbor_tagged_stash
240 : gv_stashpv ("CBOR::XS::Tagged" , 1); 544 : gv_stashpv ("CBOR::XS::Tagged" , 1);
241 545
546 HV *stash = SvSTASH (sv);
547
242 if (SvSTASH (sv) == boolean_stash) 548 if (stash == boolean_stash)
243 encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20); 549 {
550 encode_ch (enc, SvIV (sv) ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE);
551 return;
552 }
553 else if (stash == error_stash)
554 {
555 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
556 return;
557 }
244 else if (SvSTASH (sv) == tagged_stash) 558 else if (stash == tagged_stash)
245 { 559 {
246 if (svt != SVt_PVAV) 560 if (svt != SVt_PVAV)
247 croak ("encountered CBOR::XS::Tagged object that isn't an array"); 561 croak ("encountered CBOR::XS::Tagged object that isn't an array");
248 562
563 switch (av_len ((AV *)sv))
564 {
565 case 2-1:
566 // actually a tagged value
249 encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); 567 encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
250 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); 568 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1));
569 break;
570
571 case 3-1:
572 // a forced type [value, type, undef]
573 encode_forced (enc, SvUV (*av_fetch ((AV *)sv, 1, 1)), *av_fetch ((AV *)sv, 0, 1));
574 break;
575
576 default:
577 croak ("encountered malformed CBOR::XS::Tagged object");
578 }
579
580 return;
581 }
582 }
583
584 if (ecb_expect_false (SvREFCNT (sv) > 1)
585 && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
586 {
587 if (ecb_expect_false (!enc->shareable))
588 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
589
590 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
591
592 if (SvOK (*svp))
593 {
594 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
595 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
596 return;
251 } 597 }
252 else 598 else
253 { 599 {
600 sv_setuv (*svp, enc->shareable_idx);
601 ++enc->shareable_idx;
602 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
603 }
604 }
605
606 if (ecb_expect_false (SvOBJECT (sv)))
607 {
608 HV *stash = SvSTASH (sv);
609 GV *method;
610
611 if (enc->cbor.flags & F_FORBID_OBJECTS)
612 croak ("encountered object '%s', but forbid_objects is enabled",
613 SvPV_nolen (sv_2mortal (newRV_inc (sv))));
614 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
615 {
616 dSP;
617
618 ENTER; SAVETMPS;
619 PUSHMARK (SP);
254 // we re-bless the reference to get overload and other niceties right 620 // we re-bless the reference to get overload and other niceties right
255 GV *to_cbor = gv_fetchmethod_autoload (SvSTASH (sv), "TO_CBOR", 0);
256
257 if (to_cbor)
258 {
259 dSP;
260
261 ENTER; SAVETMPS; PUSHMARK (SP);
262 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), SvSTASH (sv))); 621 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
263 622
264 // calling with G_SCALAR ensures that we always get a 1 return value
265 PUTBACK; 623 PUTBACK;
624 // G_SCALAR ensures that return value is 1
266 call_sv ((SV *)GvCV (to_cbor), G_SCALAR); 625 call_sv ((SV *)GvCV (method), G_SCALAR);
267 SPAGAIN; 626 SPAGAIN;
268 627
269 // catch this surprisingly common error 628 // catch this surprisingly common error
270 if (SvROK (TOPs) && SvRV (TOPs) == sv) 629 if (SvROK (TOPs) && SvRV (TOPs) == sv)
271 croak ("%s::TO_CBOR method returned same object as was passed instead of a new one", HvNAME (SvSTASH (sv))); 630 croak ("%s::TO_CBOR method returned same object as was passed instead of a new one", HvNAME (stash));
272 631
273 sv = POPs; 632 encode_sv (enc, POPs);
633
274 PUTBACK; 634 PUTBACK;
275 635
276 encode_sv (enc, sv);
277
278 FREETMPS; LEAVE; 636 FREETMPS; LEAVE;
637 }
638 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
639 {
640 dSP;
641
642 ENTER; SAVETMPS;
643 PUSHMARK (SP);
644 EXTEND (SP, 2);
645 // we re-bless the reference to get overload and other niceties right
646 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
647 PUSHs (sv_cbor);
648
649 PUTBACK;
650 int count = call_sv ((SV *)GvCV (method), G_ARRAY);
651 SPAGAIN;
652
653 // catch this surprisingly common error
654 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
655 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
656
657 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
658 encode_uint (enc, MAJOR_ARRAY, count + 1);
659 encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
660
661 {
662 int i;
663
664 for (i = 0; i < count; ++i)
665 encode_sv (enc, SP[i + 1 - count]);
666
667 SP -= count;
279 } 668 }
669
670 PUTBACK;
671
672 FREETMPS; LEAVE;
673 }
280 else 674 else
281 croak ("encountered object '%s', but no TO_CBOR method available on it", 675 croak ("encountered object '%s', but no TO_CBOR or FREEZE methods available on it",
282 SvPV_nolen (sv_2mortal (newRV_inc (sv)))); 676 SvPV_nolen (sv_2mortal (newRV_inc (sv))));
283 }
284 } 677 }
285 else if (svt == SVt_PVHV) 678 else if (svt == SVt_PVHV)
286 encode_hv (enc, (HV *)sv); 679 encode_hv (enc, (HV *)sv);
287 else if (svt == SVt_PVAV) 680 else if (svt == SVt_PVAV)
288 encode_av (enc, (AV *)sv); 681 encode_av (enc, (AV *)sv);
289 else if (svt < SVt_PVAV)
290 {
291 STRLEN len = 0;
292 char *pv = svt ? SvPV (sv, len) : 0;
293
294 if (len == 1 && *pv == '1')
295 encode_ch (enc, 0xe0 | 21);
296 else if (len == 1 && *pv == '0')
297 encode_ch (enc, 0xe0 | 20);
298 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
299 encode_ch (enc, 0xe0 | 23);
300 else
301 croak ("cannot encode reference to scalar '%s' unless the scalar is 0 or 1",
302 SvPV_nolen (sv_2mortal (newRV_inc (sv))));
303 }
304 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
305 encode_ch (enc, 0xe0 | 23);
306 else 682 else
307 croak ("encountered %s, but CBOR can only represent references to arrays or hashes", 683 {
308 SvPV_nolen (sv_2mortal (newRV_inc (sv)))); 684 encode_tag (enc, CBOR_TAG_INDIRECTION);
685 encode_sv (enc, sv);
686 }
309} 687}
310 688
311static void 689static void
312encode_nv (enc_t *enc, SV *sv) 690encode_nv (enc_t *enc, SV *sv)
313{ 691{
314 double nv = SvNVX (sv); 692 double nv = SvNVX (sv);
315 693
316 need (enc, 9); 694 need (enc, 9);
317 695
318 if (ecb_expect_false (nv == (U32)nv)) 696 if (ecb_expect_false (nv == (NV)(U32)nv))
319 encode_uint (enc, 0x00, (U32)nv); 697 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
320 //TODO: maybe I32? 698 //TODO: maybe I32?
321 else if (ecb_expect_false (nv == (float)nv)) 699 else if (ecb_expect_false (nv == (float)nv))
322 { 700 encode_float32 (enc, nv);
323 uint32_t fp = ecb_float_to_binary32 (nv);
324
325 *enc->cur++ = 0xe0 | 26;
326
327 if (!ecb_big_endian ())
328 fp = ecb_bswap32 (fp);
329
330 memcpy (enc->cur, &fp, 4);
331 enc->cur += 4;
332 }
333 else 701 else
334 { 702 encode_float64 (enc, nv);
335 uint64_t fp = ecb_double_to_binary64 (nv);
336
337 *enc->cur++ = 0xe0 | 27;
338
339 if (!ecb_big_endian ())
340 fp = ecb_bswap64 (fp);
341
342 memcpy (enc->cur, &fp, 8);
343 enc->cur += 8;
344 }
345} 703}
346 704
347static void 705static void
348encode_sv (enc_t *enc, SV *sv) 706encode_sv (enc_t *enc, SV *sv)
349{ 707{
351 709
352 if (SvPOKp (sv)) 710 if (SvPOKp (sv))
353 { 711 {
354 STRLEN len; 712 STRLEN len;
355 char *str = SvPV (sv, len); 713 char *str = SvPV (sv, len);
356 encode_str (enc, SvUTF8 (sv), str, len); 714 encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len);
357 } 715 }
358 else if (SvNOKp (sv)) 716 else if (SvNOKp (sv))
359 encode_nv (enc, sv); 717 encode_nv (enc, sv);
360 else if (SvIOKp (sv)) 718 else if (SvIOKp (sv))
361 { 719 encode_int (e,v sv);
362 if (SvIsUV (sv))
363 encode_uint (enc, 0x00, SvUVX (sv));
364 else if (SvIVX (sv) >= 0)
365 encode_uint (enc, 0x00, SvIVX (sv));
366 else
367 encode_uint (enc, 0x20, -(SvIVX (sv) + 1));
368 }
369 else if (SvROK (sv)) 720 else if (SvROK (sv))
370 encode_rv (enc, SvRV (sv)); 721 encode_rv (enc, SvRV (sv));
371 else if (!SvOK (sv)) 722 else if (!SvOK (sv))
372 encode_ch (enc, 0xe0 | 22); 723 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
373 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 724 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
374 encode_ch (enc, 0xe0 | 23); 725 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
375 else 726 else
376 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data", 727 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
377 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 728 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
378} 729}
379 730
380static SV * 731static SV *
381encode_cbor (SV *scalar, CBOR *cbor) 732encode_cbor (SV *scalar, CBOR *cbor)
382{ 733{
383 enc_t enc; 734 enc_t enc = { 0 };
384 735
385 enc.cbor = *cbor; 736 enc.cbor = *cbor;
386 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 737 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
387 enc.cur = SvPVX (enc.sv); 738 enc.cur = SvPVX (enc.sv);
388 enc.end = SvEND (enc.sv); 739 enc.end = SvEND (enc.sv);
389 enc.depth = 0;
390 740
391 SvPOK_only (enc.sv); 741 SvPOK_only (enc.sv);
742
743 if (cbor->flags & F_PACK_STRINGS)
744 {
745 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
746 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
747 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
748 }
749
392 encode_sv (&enc, scalar); 750 encode_sv (&enc, scalar);
393 751
394 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv)); 752 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv));
395 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings 753 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings
396 754
410 U8 *end; // end of input string 768 U8 *end; // end of input string
411 const char *err; // parse error, if != 0 769 const char *err; // parse error, if != 0
412 CBOR cbor; 770 CBOR cbor;
413 U32 depth; // recursion depth 771 U32 depth; // recursion depth
414 U32 maxdepth; // recursion depth limit 772 U32 maxdepth; // recursion depth limit
773 AV *shareable;
774 AV *stringref;
775 SV *decode_tagged;
776 SV *err_sv; // optional sv for error, needs to be freed
415} dec_t; 777} dec_t;
416 778
417#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE 779// set dec->err to ERRSV
780ecb_cold static void
781err_errsv (dec_t *dec)
782{
783 if (!dec->err)
784 {
785 dec->err_sv = newSVsv (ERRSV);
418 786
419#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data") 787 // chop off the trailing \n
788 SvCUR_set (dec->err_sv, SvCUR (dec->err_sv) - 1);
789 *SvEND (dec->err_sv) = 0;
420 790
791 dec->err = SvPVutf8_nolen (dec->err_sv);
792 }
793}
794
795// the following functions are used to reduce code size and help the compiler to optimise
796ecb_cold static void
797err_set (dec_t *dec, const char *reason)
798{
799 if (!dec->err)
800 dec->err = reason;
801}
802
803ecb_cold static void
804err_unexpected_end (dec_t *dec)
805{
806 err_set (dec, "unexpected end of CBOR data");
807}
808
809#define ERR_DO(do) SB do; goto fail; SE
810#define ERR(reason) ERR_DO (err_set (dec, reason))
811#define ERR_ERRSV ERR_DO (err_errsv (dec))
812
813#define WANT(len) if (ecb_expect_false ((uintptr_t)(dec->end - dec->cur) < (STRLEN)len)) ERR_DO (err_unexpected_end (dec))
814
421#define DEC_INC_DEPTH if (++dec->depth > dec->cbor.max_depth) ERR (ERR_NESTING_EXCEEDED) 815#define DEC_INC_DEPTH if (ecb_expect_false (++dec->depth > dec->cbor.max_depth)) ERR (ERR_NESTING_EXCEEDED)
422#define DEC_DEC_DEPTH --dec->depth 816#define DEC_DEC_DEPTH --dec->depth
423 817
424static UV 818static UV
425decode_uint (dec_t *dec) 819decode_uint (dec_t *dec)
426{ 820{
427 switch (*dec->cur & 31) 821 U8 m = *dec->cur & MINOR_MASK;
428 { 822 ++dec->cur;
429 case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
430 case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
431 case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
432 return *dec->cur++ & 31;
433 823
434 case 24: 824 if (ecb_expect_true (m < LENGTH_EXT1))
825 return m;
826 else if (ecb_expect_true (m == LENGTH_EXT1))
827 {
435 WANT (2); 828 WANT (1);
436 dec->cur += 2; 829 dec->cur += 1;
437 return dec->cur[-1]; 830 return dec->cur[-1];
438 831 }
439 case 25: 832 else if (ecb_expect_true (m == LENGTH_EXT2))
833 {
440 WANT (3); 834 WANT (2);
441 dec->cur += 3; 835 dec->cur += 2;
442 return (((UV)dec->cur[-2]) << 8) 836 return (((UV)dec->cur[-2]) << 8)
443 | ((UV)dec->cur[-1]); 837 | ((UV)dec->cur[-1]);
444 838 }
445 case 26: 839 else if (ecb_expect_true (m == LENGTH_EXT4))
840 {
446 WANT (5); 841 WANT (4);
447 dec->cur += 5; 842 dec->cur += 4;
448 return (((UV)dec->cur[-4]) << 24) 843 return (((UV)dec->cur[-4]) << 24)
449 | (((UV)dec->cur[-3]) << 16) 844 | (((UV)dec->cur[-3]) << 16)
450 | (((UV)dec->cur[-2]) << 8) 845 | (((UV)dec->cur[-2]) << 8)
451 | ((UV)dec->cur[-1]); 846 | ((UV)dec->cur[-1]);
452 847 }
453 case 27: 848 else if (ecb_expect_true (m == LENGTH_EXT8))
849 {
454 WANT (9); 850 WANT (8);
455 dec->cur += 9; 851 dec->cur += 8;
852
853 return
854#if UVSIZE < 8
855 0
856#else
456 return (((UV)dec->cur[-8]) << 56) 857 (((UV)dec->cur[-8]) << 56)
457 | (((UV)dec->cur[-7]) << 48) 858 | (((UV)dec->cur[-7]) << 48)
458 | (((UV)dec->cur[-6]) << 40) 859 | (((UV)dec->cur[-6]) << 40)
459 | (((UV)dec->cur[-5]) << 32) 860 | (((UV)dec->cur[-5]) << 32)
861#endif
460 | (((UV)dec->cur[-4]) << 24) 862 | (((UV)dec->cur[-4]) << 24)
461 | (((UV)dec->cur[-3]) << 16) 863 | (((UV)dec->cur[-3]) << 16)
462 | (((UV)dec->cur[-2]) << 8) 864 | (((UV)dec->cur[-2]) << 8)
463 | ((UV)dec->cur[-1]); 865 | ((UV)dec->cur[-1]);
464 866 }
465 default: 867 else
466 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 868 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
467 }
468 869
469fail: 870fail:
470 return 0; 871 return 0;
471} 872}
472 873
477{ 878{
478 AV *av = newAV (); 879 AV *av = newAV ();
479 880
480 DEC_INC_DEPTH; 881 DEC_INC_DEPTH;
481 882
482 if ((*dec->cur & 31) == 31) 883 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
483 { 884 {
484 ++dec->cur; 885 ++dec->cur;
485 886
486 for (;;) 887 for (;;)
487 { 888 {
488 WANT (1); 889 WANT (1);
489 890
490 if (*dec->cur == (0xe0 | 31)) 891 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
491 { 892 {
492 ++dec->cur; 893 ++dec->cur;
493 break; 894 break;
494 } 895 }
495 896
496 av_push (av, decode_sv (dec)); 897 av_push (av, decode_sv (dec));
497 } 898 }
498 } 899 }
499 else 900 else
500 { 901 {
501 int i, len = decode_uint (dec); 902 UV i, len = decode_uint (dec);
502 903
904 WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
503 av_fill (av, len - 1); 905 av_fill (av, len - 1);
504 906
505 for (i = 0; i < len; ++i) 907 for (i = 0; i < len; ++i)
506 AvARRAY (av)[i] = decode_sv (dec); 908 AvARRAY (av)[i] = decode_sv (dec);
507 } 909 }
508 910
509 DEC_DEC_DEPTH; 911 DEC_DEC_DEPTH;
510 return newRV_noinc ((SV *)av); 912 return newRV_noinc ((SV *)av);
511 913
512fail: 914fail:
513 SvREFCNT_dec (av); 915 SvREFCNT_dec_NN (av);
514 DEC_DEC_DEPTH; 916 DEC_DEC_DEPTH;
515 return &PL_sv_undef; 917 return &PL_sv_undef;
516} 918}
517 919
920static void
921decode_he (dec_t *dec, HV *hv)
922{
923 // for speed reasons, we specialcase single-string
924 // byte or utf-8 strings as keys, but only when !stringref
925
926 if (ecb_expect_true (!dec->stringref))
927 if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
928 {
929 STRLEN len = decode_uint (dec);
930 char *key = (char *)dec->cur;
931
932 WANT (len);
933 dec->cur += len;
934
935 hv_store (hv, key, len, decode_sv (dec), 0);
936
937 return;
938 }
939 else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
940 {
941 STRLEN len = decode_uint (dec);
942 char *key = (char *)dec->cur;
943
944 WANT (len);
945 dec->cur += len;
946
947 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
948 if (!is_utf8_string (key, len))
949 ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
950
951 hv_store (hv, key, -len, decode_sv (dec), 0);
952
953 return;
954 }
955
956 SV *k = decode_sv (dec);
957 SV *v = decode_sv (dec);
958
959 // we leak memory if uncaught exceptions are thrown by random magical
960 // methods, and this is hopefully the only place where it can happen,
961 // so if there is a chance of an exception, take the very slow path.
962 // since catching exceptions is "undocumented/internal/forbidden" by
963 // the new p5p powers, we need to call out to a perl function :/
964 if (ecb_expect_false (SvAMAGIC (k)))
965 {
966 dSP;
967
968 ENTER; SAVETMPS;
969 PUSHMARK (SP);
970 EXTEND (SP, 3);
971 PUSHs (sv_2mortal (newRV_inc ((SV *)hv)));
972 PUSHs (sv_2mortal (k));
973 PUSHs (sv_2mortal (v));
974
975 PUTBACK;
976 call_pv ("CBOR::XS::_hv_store", G_VOID | G_DISCARD | G_EVAL);
977 SPAGAIN;
978
979 FREETMPS; LEAVE;
980
981 if (SvTRUE (ERRSV))
982 ERR_ERRSV;
983
984 return;
985 }
986
987 hv_store_ent (hv, k, v, 0);
988 SvREFCNT_dec_NN (k);
989
990fail:
991 ;
992}
993
518static SV * 994static SV *
519decode_hv (dec_t *dec) 995decode_hv (dec_t *dec)
520{ 996{
521 HV *hv = newHV (); 997 HV *hv = newHV ();
522 998
523 DEC_INC_DEPTH; 999 DEC_INC_DEPTH;
524 1000
525 if ((*dec->cur & 31) == 31) 1001 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
526 { 1002 {
527 ++dec->cur; 1003 ++dec->cur;
528 1004
529 for (;;) 1005 for (;;)
530 { 1006 {
531 WANT (1); 1007 WANT (1);
532 1008
533 if (*dec->cur == (0xe0 | 31)) 1009 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
534 { 1010 {
535 ++dec->cur; 1011 ++dec->cur;
536 break; 1012 break;
537 } 1013 }
538 1014
539 SV *k = decode_sv (dec); 1015 decode_he (dec, hv);
540 SV *v = decode_sv (dec);
541
542 hv_store_ent (hv, k, v, 0);
543 } 1016 }
544 } 1017 }
545 else 1018 else
546 { 1019 {
547 int len = decode_uint (dec); 1020 UV pairs = decode_uint (dec);
548 1021
1022 WANT (pairs); // complexity check - need at least one byte per value, do not allow supersize hashes
1023
549 while (len--) 1024 while (pairs--)
550 { 1025 decode_he (dec, hv);
551 SV *k = decode_sv (dec);
552 SV *v = decode_sv (dec);
553
554 hv_store_ent (hv, k, v, 0);
555 }
556 } 1026 }
557 1027
558 DEC_DEC_DEPTH; 1028 DEC_DEC_DEPTH;
559 return newRV_noinc ((SV *)hv); 1029 return newRV_noinc ((SV *)hv);
560 1030
561fail: 1031fail:
562 SvREFCNT_dec (hv); 1032 SvREFCNT_dec_NN (hv);
563 DEC_DEC_DEPTH; 1033 DEC_DEC_DEPTH;
564 return &PL_sv_undef; 1034 return &PL_sv_undef;
565} 1035}
566 1036
567static SV * 1037static SV *
568decode_str (dec_t *dec, int utf8) 1038decode_str (dec_t *dec, int utf8)
569{ 1039{
570 SV *sv = 0; 1040 SV *sv = 0;
571 1041
572 if ((*dec->cur & 31) == 31) 1042 if (ecb_expect_false ((*dec->cur & MINOR_MASK) == MINOR_INDEF))
573 { 1043 {
1044 // indefinite length strings
574 ++dec->cur; 1045 ++dec->cur;
575 1046
1047 U8 major = *dec->cur & MAJOR_MISC;
1048
576 sv = newSVpvn ("", 0); 1049 sv = newSVpvn ("", 0);
577 1050
578 // not very fast, and certainly not robust against illegal input
579 for (;;) 1051 for (;;)
580 { 1052 {
581 WANT (1); 1053 WANT (1);
582 1054
583 if (*dec->cur == (0xe0 | 31)) 1055 if ((*dec->cur - major) > LENGTH_EXT8)
1056 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
584 { 1057 {
585 ++dec->cur; 1058 ++dec->cur;
586 break; 1059 break;
587 } 1060 }
1061 else
1062 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
588 1063
589 sv_catsv (sv, decode_sv (dec)); 1064 STRLEN len = decode_uint (dec);
1065
1066 WANT (len);
1067 sv_catpvn (sv, dec->cur, len);
1068 dec->cur += len;
590 } 1069 }
591 } 1070 }
592 else 1071 else
593 { 1072 {
594 STRLEN len = decode_uint (dec); 1073 STRLEN len = decode_uint (dec);
595 1074
596 WANT (len); 1075 WANT (len);
597 sv = newSVpvn (dec->cur, len); 1076 sv = newSVpvn (dec->cur, len);
598 dec->cur += len; 1077 dec->cur += len;
1078
1079 if (ecb_expect_false (dec->stringref)
1080 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
1081 av_push (dec->stringref, SvREFCNT_inc_NN (sv));
599 } 1082 }
600 1083
601 if (utf8) 1084 if (utf8)
1085 {
1086 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
1087 if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
1088 ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
1089
602 SvUTF8_on (sv); 1090 SvUTF8_on (sv);
1091 }
603 1092
604 return sv; 1093 return sv;
605 1094
606fail: 1095fail:
607 SvREFCNT_dec (sv); 1096 SvREFCNT_dec (sv);
609} 1098}
610 1099
611static SV * 1100static SV *
612decode_tagged (dec_t *dec) 1101decode_tagged (dec_t *dec)
613{ 1102{
1103 SV *sv = 0;
614 UV tag = decode_uint (dec); 1104 UV tag = decode_uint (dec);
1105
1106 WANT (1);
1107
1108 switch (tag)
1109 {
1110 case CBOR_TAG_MAGIC:
615 SV *sv = decode_sv (dec); 1111 sv = decode_sv (dec);
1112 break;
616 1113
617 if (tag == 55799) // 2.4.5 Self-Describe CBOR 1114 case CBOR_TAG_INDIRECTION:
1115 sv = newRV_noinc (decode_sv (dec));
1116 break;
1117
1118 case CBOR_TAG_STRINGREF_NAMESPACE:
1119 {
1120 // do not use SAVETMPS/FREETMPS, as these will
1121 // erase mortalised caches, e.g. "shareable"
1122 ENTER;
1123
1124 SAVESPTR (dec->stringref);
1125 dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());
1126
1127 sv = decode_sv (dec);
1128
1129 LEAVE;
1130 }
1131 break;
1132
1133 case CBOR_TAG_STRINGREF:
1134 {
1135 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
1136 ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
1137
1138 UV idx = decode_uint (dec);
1139
1140 if (!dec->stringref || idx >= (UV)(1 + AvFILLp (dec->stringref)))
1141 ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)");
1142
1143 sv = newSVsv (AvARRAY (dec->stringref)[idx]);
1144 }
1145 break;
1146
1147 case CBOR_TAG_VALUE_SHAREABLE:
1148 {
1149 if (ecb_expect_false (!dec->shareable))
1150 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
1151
1152 if (dec->cbor.flags & F_ALLOW_CYCLES)
1153 {
1154 sv = newSV (0);
1155 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
1156
1157 SV *osv = decode_sv (dec);
1158 sv_setsv (sv, osv);
1159 SvREFCNT_dec_NN (osv);
1160 }
1161 else
1162 {
1163 av_push (dec->shareable, &PL_sv_undef);
1164 int idx = AvFILLp (dec->shareable);
1165 sv = decode_sv (dec);
1166 av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
1167 }
1168 }
1169 break;
1170
1171 case CBOR_TAG_VALUE_SHAREDREF:
1172 {
1173 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
1174 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
1175
1176 UV idx = decode_uint (dec);
1177
1178 if (!dec->shareable || idx >= (UV)(1 + AvFILLp (dec->shareable)))
1179 ERR ("corrupted CBOR data (sharedref index out of bounds)");
1180
1181 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
1182
1183 if (sv == &PL_sv_undef)
1184 ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
1185 }
1186 break;
1187
1188 case CBOR_TAG_PERL_OBJECT:
1189 {
1190 if (dec->cbor.flags & F_FORBID_OBJECTS)
1191 goto filter;
1192
1193 sv = decode_sv (dec);
1194
1195 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
1196 ERR ("corrupted CBOR data (non-array perl object)");
1197
1198 AV *av = (AV *)SvRV (sv);
1199 int len = av_len (av) + 1;
1200 HV *stash = gv_stashsv (*av_fetch (av, 0, 1), 0);
1201
1202 if (!stash)
1203 ERR ("cannot decode perl-object (package does not exist)");
1204
1205 GV *method = gv_fetchmethod_autoload (stash, "THAW", 0);
1206
1207 if (!method)
1208 ERR ("cannot decode perl-object (package does not have a THAW method)");
1209
1210 dSP;
1211
1212 ENTER; SAVETMPS;
1213 PUSHMARK (SP);
1214 EXTEND (SP, len + 1);
1215 // we re-bless the reference to get overload and other niceties right
1216 PUSHs (*av_fetch (av, 0, 1));
1217 PUSHs (sv_cbor);
1218
1219 int i;
1220
1221 for (i = 1; i < len; ++i)
1222 PUSHs (*av_fetch (av, i, 1));
1223
1224 PUTBACK;
1225 call_sv ((SV *)GvCV (method), G_SCALAR | G_EVAL);
1226 SPAGAIN;
1227
1228 if (SvTRUE (ERRSV))
1229 {
1230 FREETMPS; LEAVE;
1231 ERR_ERRSV;
1232 }
1233
1234 SvREFCNT_dec_NN (sv);
1235 sv = SvREFCNT_inc (POPs);
1236
1237 PUTBACK;
1238
1239 FREETMPS; LEAVE;
1240 }
1241 break;
1242
1243 default:
1244 filter:
1245 {
1246 SV *tag_sv = newSVuv (tag);
1247
1248 sv = decode_sv (dec);
1249
1250 dSP;
1251 ENTER; SAVETMPS;
1252 PUSHMARK (SP);
1253 EXTEND (SP, 2);
1254 PUSHs (tag_sv);
1255 PUSHs (sv);
1256
1257 PUTBACK;
1258 int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL);
1259 SPAGAIN;
1260
1261 if (SvTRUE (ERRSV))
1262 {
1263 SvREFCNT_dec_NN (tag_sv);
1264 FREETMPS; LEAVE;
1265 ERR_ERRSV;
1266 }
1267
1268 if (count)
1269 {
1270 SvREFCNT_dec_NN (tag_sv);
1271 SvREFCNT_dec_NN (sv);
1272 sv = SvREFCNT_inc_NN (TOPs);
1273 SP -= count;
1274 }
1275 else
1276 {
1277 AV *av = newAV ();
1278 av_push (av, tag_sv);
1279 av_push (av, sv);
1280
1281 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
1282 ? cbor_tagged_stash
1283 : gv_stashpv ("CBOR::XS::Tagged" , 1);
1284 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
1285 }
1286
1287 PUTBACK;
1288
1289 FREETMPS; LEAVE;
1290 }
1291 break;
1292 }
1293
618 return sv; 1294 return sv;
619 1295
620 AV *av = newAV (); 1296fail:
621 av_push (av, newSVuv (tag)); 1297 SvREFCNT_dec (sv);
622 av_push (av, sv); 1298 return &PL_sv_undef;
623
624 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
625 ? cbor_tagged_stash
626 : gv_stashpv ("CBOR::XS::Tagged" , 1);
627
628 return sv_bless (newRV_noinc ((SV *)av), tagged_stash);
629} 1299}
630 1300
631static SV * 1301static SV *
632decode_sv (dec_t *dec) 1302decode_sv (dec_t *dec)
633{ 1303{
634 WANT (1); 1304 WANT (1);
635 1305
636 switch (*dec->cur >> 5) 1306 switch (*dec->cur >> MAJOR_SHIFT)
637 { 1307 {
638 case 0: // unsigned int 1308 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
639 return newSVuv (decode_uint (dec)); 1309 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
640 case 1: // negative int 1310 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
641 return newSViv (-1 - (IV)decode_uint (dec)); 1311 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
642 case 2: // octet string 1312 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
643 return decode_str (dec, 0); 1313 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
644 case 3: // utf-8 string 1314 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
645 return decode_str (dec, 1); 1315
646 case 4: // array 1316 case MAJOR_MISC >> MAJOR_SHIFT:
647 return decode_av (dec);
648 case 5: // map
649 return decode_hv (dec);
650 case 6: // tag
651 return decode_tagged (dec);
652 case 7: // misc
653 switch (*dec->cur++ & 31) 1317 switch (*dec->cur++ & MINOR_MASK)
654 { 1318 {
655 case 20: 1319 case SIMPLE_FALSE:
656#if CBOR_SLOW 1320#if CBOR_SLOW
657 cbor_false = get_bool ("CBOR::XS::false"); 1321 types_false = get_bool ("Types::Serialiser::false");
658#endif 1322#endif
659 return newSVsv (cbor_false); 1323 return newSVsv (types_false);
660 case 21: 1324 case SIMPLE_TRUE:
661#if CBOR_SLOW 1325#if CBOR_SLOW
662 cbor_true = get_bool ("CBOR::XS::true"); 1326 types_true = get_bool ("Types::Serialiser::true");
663#endif 1327#endif
664 return newSVsv (cbor_true); 1328 return newSVsv (types_true);
665 case 22: 1329 case SIMPLE_NULL:
666 return newSVsv (&PL_sv_undef); 1330 return newSVsv (&PL_sv_undef);
1331 case SIMPLE_UNDEF:
1332#if CBOR_SLOW
1333 types_error = get_bool ("Types::Serialiser::error");
1334#endif
1335 return newSVsv (types_error);
667 1336
668 case 25: 1337 case MISC_FLOAT16:
669 { 1338 {
670 WANT (2); 1339 WANT (2);
671 1340
672 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1341 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
673 dec->cur += 2; 1342 dec->cur += 2;
674 1343
675 return newSVnv (ecb_binary16_to_float (fp)); 1344 return newSVnv (ecb_binary16_to_float (fp));
676 } 1345 }
677 1346
678 case 26: 1347 case MISC_FLOAT32:
679 { 1348 {
680 uint32_t fp; 1349 uint32_t fp;
681 WANT (4); 1350 WANT (4);
682 memcpy (&fp, dec->cur, 4); 1351 memcpy (&fp, dec->cur, 4);
683 dec->cur += 4; 1352 dec->cur += 4;
686 fp = ecb_bswap32 (fp); 1355 fp = ecb_bswap32 (fp);
687 1356
688 return newSVnv (ecb_binary32_to_float (fp)); 1357 return newSVnv (ecb_binary32_to_float (fp));
689 } 1358 }
690 1359
691 case 27: 1360 case MISC_FLOAT64:
692 { 1361 {
693 uint64_t fp; 1362 uint64_t fp;
694 WANT (8); 1363 WANT (8);
695 memcpy (&fp, dec->cur, 8); 1364 memcpy (&fp, dec->cur, 8);
696 dec->cur += 8; 1365 dec->cur += 8;
699 fp = ecb_bswap64 (fp); 1368 fp = ecb_bswap64 (fp);
700 1369
701 return newSVnv (ecb_binary64_to_double (fp)); 1370 return newSVnv (ecb_binary64_to_double (fp));
702 } 1371 }
703 1372
704 // 0..19 unassigned 1373 // 0..19 unassigned simple
705 // 24 reserved + unassigned (reserved values are not encodable) 1374 // 24 reserved + unassigned simple (reserved values are not encodable)
1375 // 28-30 unassigned misc
1376 // 31 break code
706 default: 1377 default:
707 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1378 ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
708 } 1379 }
709 1380
710 break; 1381 break;
711 } 1382 }
712 1383
715} 1386}
716 1387
717static SV * 1388static SV *
718decode_cbor (SV *string, CBOR *cbor, char **offset_return) 1389decode_cbor (SV *string, CBOR *cbor, char **offset_return)
719{ 1390{
720 dec_t dec; 1391 dec_t dec = { 0 };
721 SV *sv; 1392 SV *sv;
1393 STRLEN len;
1394 char *data = SvPVbyte (string, len);
722 1395
723 /* work around bugs in 5.10 where manipulating magic values
724 * makes perl ignore the magic in subsequent accesses.
725 * also make a copy of non-PV values, to get them into a clean
726 * state (SvPV should do that, but it's buggy, see below).
727 */
728 /*SvGETMAGIC (string);*/
729 if (SvMAGICAL (string) || !SvPOK (string))
730 string = sv_2mortal (newSVsv (string));
731
732 SvUPGRADE (string, SVt_PV);
733
734 /* work around a bug in perl 5.10, which causes SvCUR to fail an
735 * assertion with -DDEBUGGING, although SvCUR is documented to
736 * return the xpv_cur field which certainly exists after upgrading.
737 * according to nicholas clark, calling SvPOK fixes this.
738 * But it doesn't fix it, so try another workaround, call SvPV_nolen
739 * and hope for the best.
740 * Damnit, SvPV_nolen still trips over yet another assertion. This
741 * assertion business is seriously broken, try yet another workaround
742 * for the broken -DDEBUGGING.
743 */
744 {
745#ifdef DEBUGGING
746 STRLEN offset = SvOK (string) ? sv_len (string) : 0;
747#else
748 STRLEN offset = SvCUR (string);
749#endif
750
751 if (offset > cbor->max_size && cbor->max_size) 1396 if (len > cbor->max_size && cbor->max_size)
752 croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu", 1397 croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
753 (unsigned long)SvCUR (string), (unsigned long)cbor->max_size); 1398 (unsigned long)len, (unsigned long)cbor->max_size);
754 }
755
756 sv_utf8_downgrade (string, 0);
757 1399
758 dec.cbor = *cbor; 1400 dec.cbor = *cbor;
759 dec.cur = (U8 *)SvPVX (string); 1401 dec.cur = (U8 *)data;
760 dec.end = (U8 *)SvEND (string); 1402 dec.end = (U8 *)data + len;
761 dec.err = 0;
762 dec.depth = 0;
763 1403
764 sv = decode_sv (&dec); 1404 sv = decode_sv (&dec);
765 1405
766 if (offset_return) 1406 if (offset_return)
767 *offset_return = dec.cur; 1407 *offset_return = dec.cur;
770 if (dec.cur != dec.end && !dec.err) 1410 if (dec.cur != dec.end && !dec.err)
771 dec.err = "garbage after CBOR object"; 1411 dec.err = "garbage after CBOR object";
772 1412
773 if (dec.err) 1413 if (dec.err)
774 { 1414 {
1415 if (dec.shareable)
1416 {
1417 // need to break cyclic links, which would all be in shareable
1418 int i;
1419 SV **svp;
1420
1421 for (i = av_len (dec.shareable) + 1; i--; )
1422 if ((svp = av_fetch (dec.shareable, i, 0)))
1423 sv_setsv (*svp, &PL_sv_undef);
1424 }
1425
775 SvREFCNT_dec (sv); 1426 SvREFCNT_dec_NN (sv);
1427
1428 if (dec.err_sv)
1429 sv_2mortal (dec.err_sv);
1430
776 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)SvPVX (string), (int)(uint8_t)*dec.cur); 1431 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur);
777 } 1432 }
778 1433
779 sv = sv_2mortal (sv); 1434 sv = sv_2mortal (sv);
780 1435
781 return sv; 1436 return sv;
782} 1437}
783 1438
1439/////////////////////////////////////////////////////////////////////////////
1440// incremental parser
1441
1442#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
1443
1444// returns 0 for notyet, 1 for success or error
1445static int
1446incr_parse (CBOR *self, SV *cborstr)
1447{
1448 STRLEN cur;
1449 SvPV (cborstr, cur);
1450
1451 while (ecb_expect_true (self->incr_need <= cur))
1452 {
1453 // table of integer count bytes
1454 static I8 incr_len[MINOR_MASK + 1] = {
1455 0, 0, 0, 0, 0, 0, 0, 0,
1456 0, 0, 0, 0, 0, 0, 0, 0,
1457 0, 0, 0, 0, 0, 0, 0, 0,
1458 1, 2, 4, 8,-1,-1,-1,-2
1459 };
1460
1461 const U8 *p = SvPVX (cborstr) + self->incr_pos;
1462 U8 m = *p & MINOR_MASK;
1463 IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1464 I8 ilen = incr_len[m];
1465
1466 self->incr_need = self->incr_pos + 1;
1467
1468 if (ecb_expect_false (ilen < 0))
1469 {
1470 if (m != MINOR_INDEF)
1471 return 1; // error
1472
1473 if (*p == (MAJOR_MISC | MINOR_INDEF))
1474 {
1475 if (count >= 0)
1476 return 1; // error
1477
1478 count = 1;
1479 }
1480 else
1481 {
1482 av_push (self->incr_count, newSViv (-1)); //TODO: nest
1483 count = -1;
1484 }
1485 }
1486 else
1487 {
1488 self->incr_need += ilen;
1489 if (ecb_expect_false (self->incr_need > cur))
1490 return 0;
1491
1492 int major = *p >> MAJOR_SHIFT;
1493
1494 switch (major)
1495 {
1496 case MAJOR_TAG >> MAJOR_SHIFT:
1497 ++count; // tags merely prefix another value
1498 break;
1499
1500 case MAJOR_BYTES >> MAJOR_SHIFT:
1501 case MAJOR_TEXT >> MAJOR_SHIFT:
1502 case MAJOR_ARRAY >> MAJOR_SHIFT:
1503 case MAJOR_MAP >> MAJOR_SHIFT:
1504 {
1505 UV len;
1506
1507 if (ecb_expect_false (ilen))
1508 {
1509 len = 0;
1510
1511 do {
1512 len = (len << 8) | *++p;
1513 } while (--ilen);
1514 }
1515 else
1516 len = m;
1517
1518 switch (major)
1519 {
1520 case MAJOR_BYTES >> MAJOR_SHIFT:
1521 case MAJOR_TEXT >> MAJOR_SHIFT:
1522 self->incr_need += len;
1523 if (ecb_expect_false (self->incr_need > cur))
1524 return 0;
1525
1526 break;
1527
1528 case MAJOR_MAP >> MAJOR_SHIFT:
1529 len <<= 1;
1530 case MAJOR_ARRAY >> MAJOR_SHIFT:
1531 if (len)
1532 {
1533 av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
1534 count = len + 1;
1535 }
1536 break;
1537 }
1538 }
1539 }
1540 }
1541
1542 self->incr_pos = self->incr_need;
1543
1544 if (count > 0)
1545 {
1546 while (!--count)
1547 {
1548 if (!AvFILLp (self->incr_count))
1549 return 1; // done
1550
1551 SvREFCNT_dec_NN (av_pop (self->incr_count));
1552 count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1553 }
1554
1555 SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
1556 }
1557 }
1558
1559 return 0;
1560}
1561
1562
784///////////////////////////////////////////////////////////////////////////// 1563/////////////////////////////////////////////////////////////////////////////
785// XS interface functions 1564// XS interface functions
786 1565
787MODULE = CBOR::XS PACKAGE = CBOR::XS 1566MODULE = CBOR::XS PACKAGE = CBOR::XS
788 1567
789BOOT: 1568BOOT:
790{ 1569{
791 cbor_stash = gv_stashpv ("CBOR::XS" , 1); 1570 cbor_stash = gv_stashpv ("CBOR::XS" , 1);
792 cbor_boolean_stash = gv_stashpv ("CBOR::XS::Boolean", 1);
793 cbor_tagged_stash = gv_stashpv ("CBOR::XS::Tagged" , 1); 1571 cbor_tagged_stash = gv_stashpv ("CBOR::XS::Tagged" , 1);
794 1572
795 cbor_true = get_bool ("CBOR::XS::true"); 1573 types_boolean_stash = gv_stashpv ("Types::Serialiser::Boolean", 1);
796 cbor_false = get_bool ("CBOR::XS::false"); 1574 types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1);
1575
1576 types_true = get_bool ("Types::Serialiser::true" );
1577 types_false = get_bool ("Types::Serialiser::false");
1578 types_error = get_bool ("Types::Serialiser::error");
1579
1580 default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1581
1582 sv_cbor = newSVpv ("CBOR", 0);
1583 SvREADONLY_on (sv_cbor);
1584
1585 assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1));
797} 1586}
798 1587
799PROTOTYPES: DISABLE 1588PROTOTYPES: DISABLE
800 1589
801void CLONE (...) 1590void CLONE (...)
802 CODE: 1591 CODE:
803 cbor_stash = 0; 1592 cbor_stash = 0;
804 cbor_boolean_stash = 0;
805 cbor_tagged_stash = 0; 1593 cbor_tagged_stash = 0;
1594 types_error_stash = 0;
1595 types_boolean_stash = 0;
806 1596
807void new (char *klass) 1597void new (char *klass)
808 PPCODE: 1598 PPCODE:
809{ 1599{
810 SV *pv = NEWSV (0, sizeof (CBOR)); 1600 SV *pv = NEWSV (0, sizeof (CBOR));
818 1608
819void shrink (CBOR *self, int enable = 1) 1609void shrink (CBOR *self, int enable = 1)
820 ALIAS: 1610 ALIAS:
821 shrink = F_SHRINK 1611 shrink = F_SHRINK
822 allow_unknown = F_ALLOW_UNKNOWN 1612 allow_unknown = F_ALLOW_UNKNOWN
1613 allow_sharing = F_ALLOW_SHARING
1614 allow_cycles = F_ALLOW_CYCLES
1615 forbid_objects = F_FORBID_OBJECTS
1616 pack_strings = F_PACK_STRINGS
1617 text_keys = F_TEXT_KEYS
1618 text_strings = F_TEXT_STRINGS
1619 validate_utf8 = F_VALIDATE_UTF8
823 PPCODE: 1620 PPCODE:
824{ 1621{
825 if (enable) 1622 if (enable)
826 self->flags |= ix; 1623 self->flags |= ix;
827 else 1624 else
832 1629
833void get_shrink (CBOR *self) 1630void get_shrink (CBOR *self)
834 ALIAS: 1631 ALIAS:
835 get_shrink = F_SHRINK 1632 get_shrink = F_SHRINK
836 get_allow_unknown = F_ALLOW_UNKNOWN 1633 get_allow_unknown = F_ALLOW_UNKNOWN
1634 get_allow_sharing = F_ALLOW_SHARING
1635 get_allow_cycles = F_ALLOW_CYCLES
1636 get_forbid_objects = F_FORBID_OBJECTS
1637 get_pack_strings = F_PACK_STRINGS
1638 get_text_keys = F_TEXT_KEYS
1639 get_text_strings = F_TEXT_STRINGS
1640 get_validate_utf8 = F_VALIDATE_UTF8
837 PPCODE: 1641 PPCODE:
838 XPUSHs (boolSV (self->flags & ix)); 1642 XPUSHs (boolSV (self->flags & ix));
839 1643
840void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1644void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
841 PPCODE: 1645 PPCODE:
857 CODE: 1661 CODE:
858 RETVAL = self->max_size; 1662 RETVAL = self->max_size;
859 OUTPUT: 1663 OUTPUT:
860 RETVAL 1664 RETVAL
861 1665
862#if 0 //TODO 1666void filter (CBOR *self, SV *filter = 0)
863
864void filter_cbor_object (CBOR *self, SV *cb = &PL_sv_undef)
865 PPCODE: 1667 PPCODE:
866{
867 SvREFCNT_dec (self->cb_object); 1668 SvREFCNT_dec (self->filter);
868 self->cb_object = SvOK (cb) ? newSVsv (cb) : 0; 1669 self->filter = filter ? newSVsv (filter) : filter;
869
870 XPUSHs (ST (0)); 1670 XPUSHs (ST (0));
871}
872 1671
873void filter_cbor_single_key_object (CBOR *self, SV *key, SV *cb = &PL_sv_undef) 1672SV *get_filter (CBOR *self)
874 PPCODE: 1673 CODE:
875{ 1674 RETVAL = self->filter ? self->filter : NEWSV (0, 0);
876 if (!self->cb_sk_object) 1675 OUTPUT:
877 self->cb_sk_object = newHV (); 1676 RETVAL
878
879 if (SvOK (cb))
880 hv_store_ent (self->cb_sk_object, key, newSVsv (cb), 0);
881 else
882 {
883 hv_delete_ent (self->cb_sk_object, key, G_DISCARD, 0);
884
885 if (!HvKEYS (self->cb_sk_object))
886 {
887 SvREFCNT_dec (self->cb_sk_object);
888 self->cb_sk_object = 0;
889 }
890 }
891
892 XPUSHs (ST (0));
893}
894
895#endif
896 1677
897void encode (CBOR *self, SV *scalar) 1678void encode (CBOR *self, SV *scalar)
898 PPCODE: 1679 PPCODE:
899 PUTBACK; scalar = encode_cbor (scalar, self); SPAGAIN; 1680 PUTBACK; scalar = encode_cbor (scalar, self); SPAGAIN;
900 XPUSHs (scalar); 1681 XPUSHs (scalar);
913 EXTEND (SP, 2); 1694 EXTEND (SP, 2);
914 PUSHs (sv); 1695 PUSHs (sv);
915 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1696 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
916} 1697}
917 1698
918#if 0 1699void incr_parse (CBOR *self, SV *cborstr)
1700 ALIAS:
1701 incr_parse_multiple = 1
1702 PPCODE:
1703{
1704 if (SvUTF8 (cborstr))
1705 sv_utf8_downgrade (cborstr, 0);
1706
1707 if (!self->incr_count)
1708 {
1709 self->incr_count = newAV ();
1710 self->incr_pos = 0;
1711 self->incr_need = 1;
1712
1713 av_push (self->incr_count, newSViv (1));
1714 }
1715
1716 do
1717 {
1718 if (!incr_parse (self, cborstr))
1719 {
1720 if (self->incr_need > self->max_size && self->max_size)
1721 croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
1722 (unsigned long)self->incr_need, (unsigned long)self->max_size);
1723
1724 break;
1725 }
1726
1727 SV *sv;
1728 char *offset;
1729
1730 PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
1731 XPUSHs (sv);
1732
1733 sv_chop (cborstr, offset);
1734
1735 av_clear (self->incr_count);
1736 av_push (self->incr_count, newSViv (1));
1737
1738 self->incr_pos = 0;
1739 self->incr_need = self->incr_pos + 1;
1740 }
1741 while (ix);
1742}
1743
1744void incr_reset (CBOR *self)
1745 CODE:
1746{
1747 SvREFCNT_dec (self->incr_count);
1748 self->incr_count = 0;
1749}
919 1750
920void DESTROY (CBOR *self) 1751void DESTROY (CBOR *self)
921 CODE: 1752 PPCODE:
922 SvREFCNT_dec (self->cb_sk_object); 1753 cbor_free (self);
923 SvREFCNT_dec (self->cb_object);
924
925#endif
926 1754
927PROTOTYPES: ENABLE 1755PROTOTYPES: ENABLE
928 1756
929void encode_cbor (SV *scalar) 1757void encode_cbor (SV *scalar)
1758 ALIAS:
1759 encode_cbor = 0
1760 encode_cbor_sharing = F_ALLOW_SHARING
930 PPCODE: 1761 PPCODE:
931{ 1762{
932 CBOR cbor; 1763 CBOR cbor;
933 cbor_init (&cbor); 1764 cbor_init (&cbor);
1765 cbor.flags |= ix;
934 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1766 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
935 XPUSHs (scalar); 1767 XPUSHs (scalar);
936} 1768}
937 1769
938void decode_cbor (SV *cborstr) 1770void decode_cbor (SV *cborstr)
942 cbor_init (&cbor); 1774 cbor_init (&cbor);
943 PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN; 1775 PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN;
944 XPUSHs (cborstr); 1776 XPUSHs (cborstr);
945} 1777}
946 1778
1779#ifdef __AFL_COMPILER
1780
1781void
1782afl_init ()
1783 CODE:
1784 __AFL_INIT ();
1785
1786int
1787afl_loop (unsigned int count = 10000)
1788 CODE:
1789 RETVAL = __AFL_LOOP (count);
1790 OUTPUT:
1791 RETVAL
1792
1793#endif
1794

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines