ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/CBOR-XS/XS.xs
(Generate patch)

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.18 by root, Wed Nov 20 01:09:46 2013 UTC vs.
Revision 1.69 by root, Mon Nov 30 20:38:25 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_SHARABLE = 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
63#define F_DEDUP_KEYS 0x00000010UL //TODO 123#define F_TEXT_KEYS 0x00000040UL
124#define F_TEXT_STRINGS 0x00000080UL
125#define F_VALIDATE_UTF8 0x00000100UL
64 126
65#define INIT_SIZE 32 // initial scalar size to be allocated 127#define INIT_SIZE 32 // initial scalar size to be allocated
66 128
67#define SB do { 129#define SB do {
68#define SE } while (0) 130#define SE } while (0)
80# define CBOR_SLOW 0 142# define CBOR_SLOW 0
81# define CBOR_STASH cbor_stash 143# define CBOR_STASH cbor_stash
82#endif 144#endif
83 145
84static 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::
85static SV *types_true, *types_false, *types_error, *sv_cbor; 147static SV *types_true, *types_false, *types_error, *sv_cbor, *default_filter;
86 148
87typedef struct { 149typedef struct {
88 U32 flags; 150 U32 flags;
89 U32 max_depth; 151 U32 max_depth;
90 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.
91} CBOR; 159} CBOR;
92 160
93ecb_inline void 161ecb_inline void
94cbor_init (CBOR *cbor) 162cbor_init (CBOR *cbor)
95{ 163{
96 Zero (cbor, 1, CBOR); 164 Zero (cbor, 1, CBOR);
97 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);
98} 173}
99 174
100///////////////////////////////////////////////////////////////////////////// 175/////////////////////////////////////////////////////////////////////////////
101// utility functions 176// utility functions
102 177
124 SvPV_renew (sv, SvCUR (sv) + 1); 199 SvPV_renew (sv, SvCUR (sv) + 1);
125#endif 200#endif
126 } 201 }
127} 202}
128 203
204// minimum length of a string to be registered for stringref
205ecb_inline int
206minimum_string_length (UV idx)
207{
208 return idx <= 23 ? 3
209 : idx <= 0xffU ? 4
210 : idx <= 0xffffU ? 5
211 : idx <= 0xffffffffU ? 7
212 : 11;
213}
214
129///////////////////////////////////////////////////////////////////////////// 215/////////////////////////////////////////////////////////////////////////////
130// encoder 216// encoder
131 217
132// structure used for encoding CBOR 218// structure used for encoding CBOR
133typedef struct 219typedef struct
135 char *cur; // SvPVX (sv) + current output position 221 char *cur; // SvPVX (sv) + current output position
136 char *end; // SvEND (sv) 222 char *end; // SvEND (sv)
137 SV *sv; // result scalar 223 SV *sv; // result scalar
138 CBOR cbor; 224 CBOR cbor;
139 U32 depth; // recursion level 225 U32 depth; // recursion level
140 HV *stringref; // string => index, or 0 226 HV *stringref[2]; // string => index, or 0 ([0] = bytes, [1] = utf-8)
227 UV stringref_idx;
141 HV *sharable; // ptr => index, or 0 228 HV *shareable; // ptr => index, or 0
142 HV *sharable_idx; 229 UV shareable_idx;
143} enc_t; 230} enc_t;
144 231
145ecb_inline void 232ecb_inline void
146need (enc_t *enc, STRLEN len) 233need (enc_t *enc, STRLEN len)
147{ 234{
148 if (ecb_expect_false (enc->cur + len >= enc->end)) 235 if (ecb_expect_false ((uintptr_t)(enc->end - enc->cur) < len))
149 { 236 {
150 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv); 237 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv);
151 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1); 238 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1);
152 enc->cur = SvPVX (enc->sv) + cur; 239 enc->cur = SvPVX (enc->sv) + cur;
153 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1; 240 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;
159{ 246{
160 need (enc, 1); 247 need (enc, 1);
161 *enc->cur++ = ch; 248 *enc->cur++ = ch;
162} 249}
163 250
251// used for tags, intregers, element counts and so on
164static void 252static void
165encode_uint (enc_t *enc, int major, UV len) 253encode_uint (enc_t *enc, int major, UV len)
166{ 254{
167 need (enc, 9); 255 need (enc, 9);
168 256
169 if (len < 24) 257 if (ecb_expect_true (len < LENGTH_EXT1))
170 *enc->cur++ = major | len; 258 *enc->cur++ = major | len;
171 else if (len <= 0xff) 259 else if (ecb_expect_true (len <= 0xffU))
172 { 260 {
173 *enc->cur++ = major | 24; 261 *enc->cur++ = major | LENGTH_EXT1;
174 *enc->cur++ = len; 262 *enc->cur++ = len;
175 } 263 }
176 else if (len <= 0xffff) 264 else if (len <= 0xffffU)
177 { 265 {
178 *enc->cur++ = major | 25; 266 *enc->cur++ = major | LENGTH_EXT2;
179 *enc->cur++ = len >> 8; 267 *enc->cur++ = len >> 8;
180 *enc->cur++ = len; 268 *enc->cur++ = len;
181 } 269 }
182 else if (len <= 0xffffffff) 270 else if (len <= 0xffffffffU)
183 { 271 {
184 *enc->cur++ = major | 26; 272 *enc->cur++ = major | LENGTH_EXT4;
185 *enc->cur++ = len >> 24; 273 *enc->cur++ = len >> 24;
186 *enc->cur++ = len >> 16; 274 *enc->cur++ = len >> 16;
187 *enc->cur++ = len >> 8; 275 *enc->cur++ = len >> 8;
188 *enc->cur++ = len; 276 *enc->cur++ = len;
189 } 277 }
190 else 278 else
191 { 279 {
192 *enc->cur++ = major | 27; 280 *enc->cur++ = major | LENGTH_EXT8;
193 *enc->cur++ = len >> 56; 281 *enc->cur++ = len >> 56;
194 *enc->cur++ = len >> 48; 282 *enc->cur++ = len >> 48;
195 *enc->cur++ = len >> 40; 283 *enc->cur++ = len >> 40;
196 *enc->cur++ = len >> 32; 284 *enc->cur++ = len >> 32;
197 *enc->cur++ = len >> 24; 285 *enc->cur++ = len >> 24;
199 *enc->cur++ = len >> 8; 287 *enc->cur++ = len >> 8;
200 *enc->cur++ = len; 288 *enc->cur++ = len;
201 } 289 }
202} 290}
203 291
204static void 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
304ecb_inline void
305encode_tag (enc_t *enc, UV tag)
306{
307 encode_uint (enc, MAJOR_TAG, tag);
308}
309
310// exceptional (hopefully) slow path for byte strings that need to be utf8-encoded
311ecb_noinline static void
205encode_str (enc_t *enc, int utf8, char *str, STRLEN len) 312encode_str_utf8 (enc_t *enc, int utf8, char *str, STRLEN len)
206{ 313{
207 encode_uint (enc, utf8 ? 0x60 : 0x40, len); 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);
208 need (enc, len); 344 need (enc, len);
209 memcpy (enc->cur, str, len); 345 memcpy (enc->cur, str, len);
210 enc->cur += len; 346 enc->cur += len;
211} 347}
212 348
213ecb_inline void 349ecb_inline void
214encode_tag (enc_t *enc, UV tag) 350encode_strref (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
215{ 351{
216 encode_uint (enc, 0xc0, tag); 352 if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
217} 353 {
354 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
218 355
219static int
220encode_sharable2 (enc_t *enc, SV *sv)
221{
222 if (!enc->sharable)
223 enc->sharable = (HV *)sv_2mortal ((SV *)newHV ());
224
225 SV **svp = hv_fetch (enc->sharable, &sv, sizeof (sv), 1);
226
227 if (SvOK (*svp)) 356 if (SvOK (*svp))
228 { 357 {
358 // already registered, use stringref
229 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF); 359 encode_tag (enc, CBOR_TAG_STRINGREF);
230 encode_uint (enc, 0x00, SvUV (*svp)); 360 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
231
232 return 1; 361 return;
233 } 362 }
234 else 363 else if (len >= minimum_string_length (enc->stringref_idx))
235 { 364 {
365 // register only
236 sv_setuv (*svp, enc->sharable_idx++); 366 sv_setuv (*svp, enc->stringref_idx);
237 encode_tag (enc, CBOR_TAG_VALUE_SHARABLE); 367 ++enc->stringref_idx;
238 368 }
239 return 0;
240 } 369 }
241}
242 370
371 encode_str (enc, upgrade_utf8, utf8, str, len);
372}
373
243ecb_inline int 374ecb_inline void
375encode_float16 (enc_t *enc, NV nv)
376{
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
386 memcpy (enc->cur, &fp, 2);
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_bool (enc_t *enc, int istrue)
424{
425 encode_ch (enc, istrue ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE);
426}
427
428ecb_inline void
244encode_sharable (enc_t *enc, SV *sv) 429encode_forced (enc_t *enc, UV type, SV *sv)
245{ 430{
246 if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING) 431 switch (type)
247 && ecb_expect_false (SvREFCNT (sv) > 1)) 432 {
248 return encode_sharable2 (enc, sv); 433 case AS_CBOR:
434 {
435 STRLEN len;
436 char *str = SvPVbyte (sv, len);
249 437
250 return 0; 438 need (enc, len);
439 memcpy (enc->cur, str, len);
440 enc->cur += len;
441 }
442 break;
443
444 case AS_BYTES:
445 {
446 STRLEN len;
447 char *str = SvPVbyte (sv, len);
448 encode_strref (enc, 0, 0, str, len);
449 }
450 break;
451
452 case AS_TEXT:
453 {
454 STRLEN len;
455 char *str = SvPVutf8 (sv, len);
456 encode_strref (enc, 1, 1, str, len);
457 }
458 break;
459
460 case AS_INT: encode_int (enc, sv); break;
461
462 case AS_FLOAT16: encode_float16 (enc, SvNV (sv)); break;
463 case AS_FLOAT32: encode_float32 (enc, SvNV (sv)); break;
464 case AS_FLOAT64: encode_float64 (enc, SvNV (sv)); break;
465
466 default:
467 croak ("encountered malformed CBOR::XS::Tagged object");
468 }
251} 469}
252 470
253static void encode_sv (enc_t *enc, SV *sv); 471static void encode_sv (enc_t *enc, SV *sv);
254 472
255static void 473static void
260 if (enc->depth >= enc->cbor.max_depth) 478 if (enc->depth >= enc->cbor.max_depth)
261 croak (ERR_NESTING_EXCEEDED); 479 croak (ERR_NESTING_EXCEEDED);
262 480
263 ++enc->depth; 481 ++enc->depth;
264 482
265 encode_uint (enc, 0x80, len + 1); 483 encode_uint (enc, MAJOR_ARRAY, len + 1);
266 484
485 if (ecb_expect_false (SvMAGICAL (av)))
267 for (i = 0; i <= len; ++i) 486 for (i = 0; i <= len; ++i)
268 { 487 {
269 SV **svp = av_fetch (av, i, 0); 488 SV **svp = av_fetch (av, i, 0);
270 encode_sv (enc, svp ? *svp : &PL_sv_undef); 489 encode_sv (enc, svp ? *svp : &PL_sv_undef);
271 } 490 }
491 else
492 for (i = 0; i <= len; ++i)
493 {
494 SV *sv = AvARRAY (av)[i];
495 encode_sv (enc, sv ? sv : &PL_sv_undef);
496 }
272 497
273 --enc->depth; 498 --enc->depth;
274} 499}
275 500
276static void 501static void
284 ++enc->depth; 509 ++enc->depth;
285 510
286 int pairs = hv_iterinit (hv); 511 int pairs = hv_iterinit (hv);
287 int mg = SvMAGICAL (hv); 512 int mg = SvMAGICAL (hv);
288 513
289 if (mg) 514 if (ecb_expect_false (mg))
290 encode_ch (enc, 0xa0 | 31); 515 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
291 else 516 else
292 encode_uint (enc, 0xa0, pairs); 517 encode_uint (enc, MAJOR_MAP, pairs);
293 518
294 while ((he = hv_iternext (hv))) 519 while ((he = hv_iternext (hv)))
295 { 520 {
296 if (HeKLEN (he) == HEf_SVKEY) 521 if (HeKLEN (he) == HEf_SVKEY)
297 encode_sv (enc, HeSVKEY (he)); 522 encode_sv (enc, HeSVKEY (he));
298 else 523 else
299 encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); 524 encode_strref (enc, enc->cbor.flags & (F_TEXT_KEYS | F_TEXT_STRINGS), HeKUTF8 (he), HeKEY (he), HeKLEN (he));
300 525
301 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); 526 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
302 } 527 }
303 528
304 if (mg) 529 if (ecb_expect_false (mg))
305 encode_ch (enc, 0xe0 | 31); 530 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
306 531
307 --enc->depth; 532 --enc->depth;
308} 533}
309 534
310// encode objects, arrays and special \0=false and \1=true values. 535// encode objects, arrays and special \0=false and \1=true values.
311static void 536static void
312encode_rv (enc_t *enc, SV *sv) 537encode_rv (enc_t *enc, SV *sv)
313{ 538{
314 svtype svt;
315
316 SvGETMAGIC (sv); 539 SvGETMAGIC (sv);
540
317 svt = SvTYPE (sv); 541 svtype svt = SvTYPE (sv);
318
319 if (encode_sharable (enc, sv))
320 return;
321 542
322 if (ecb_expect_false (SvOBJECT (sv))) 543 if (ecb_expect_false (SvOBJECT (sv)))
323 { 544 {
324 HV *boolean_stash = !CBOR_SLOW || types_boolean_stash 545 HV *boolean_stash = !CBOR_SLOW || types_boolean_stash
325 ? types_boolean_stash 546 ? types_boolean_stash
330 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 551 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
331 ? cbor_tagged_stash 552 ? cbor_tagged_stash
332 : gv_stashpv ("CBOR::XS::Tagged" , 1); 553 : gv_stashpv ("CBOR::XS::Tagged" , 1);
333 554
334 HV *stash = SvSTASH (sv); 555 HV *stash = SvSTASH (sv);
335 GV *method;
336 556
337 if (stash == boolean_stash) 557 if (stash == boolean_stash)
338 encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20); 558 {
559 encode_bool (enc, SvIV (sv));
560 return;
561 }
339 else if (stash == error_stash) 562 else if (stash == error_stash)
340 encode_ch (enc, 0xe0 | 23); 563 {
564 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
565 return;
566 }
341 else if (stash == tagged_stash) 567 else if (stash == tagged_stash)
342 { 568 {
343 if (svt != SVt_PVAV) 569 if (svt != SVt_PVAV)
344 croak ("encountered CBOR::XS::Tagged object that isn't an array"); 570 croak ("encountered CBOR::XS::Tagged object that isn't an array");
345 571
572 switch (av_len ((AV *)sv))
573 {
574 case 2-1:
575 // actually a tagged value
346 encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); 576 encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
347 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); 577 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1));
578 break;
579
580 case 3-1:
581 // a forced type [value, type, undef]
582 encode_forced (enc, SvUV (*av_fetch ((AV *)sv, 1, 1)), *av_fetch ((AV *)sv, 0, 1));
583 break;
584
585 default:
586 croak ("encountered malformed CBOR::XS::Tagged object");
587 }
588
589 return;
590 }
591 }
592
593 if (ecb_expect_false (SvREFCNT (sv) > 1)
594 && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
595 {
596 if (ecb_expect_false (!enc->shareable))
597 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
598
599 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
600
601 if (SvOK (*svp))
348 } 602 {
603 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
604 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
605 return;
606 }
607 else
608 {
609 sv_setuv (*svp, enc->shareable_idx);
610 ++enc->shareable_idx;
611 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
612 }
613 }
614
615 if (ecb_expect_false (SvOBJECT (sv)))
616 {
617 HV *stash = SvSTASH (sv);
618 GV *method;
619
620 if (enc->cbor.flags & F_FORBID_OBJECTS)
621 croak ("encountered object '%s', but forbid_objects is enabled",
622 SvPV_nolen (sv_2mortal (newRV_inc (sv))));
349 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) 623 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
350 { 624 {
351 dSP; 625 dSP;
352 626
353 ENTER; SAVETMPS; PUSHMARK (SP); 627 ENTER; SAVETMPS;
628 PUSHMARK (SP);
354 // we re-bless the reference to get overload and other niceties right 629 // we re-bless the reference to get overload and other niceties right
355 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 630 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
356 631
357 PUTBACK; 632 PUTBACK;
358 // G_SCALAR ensures that return value is 1 633 // G_SCALAR ensures that return value is 1
371 } 646 }
372 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0) 647 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
373 { 648 {
374 dSP; 649 dSP;
375 650
376 ENTER; SAVETMPS; PUSHMARK (SP); 651 ENTER; SAVETMPS;
652 PUSHMARK (SP);
377 EXTEND (SP, 2); 653 EXTEND (SP, 2);
378 // we re-bless the reference to get overload and other niceties right 654 // we re-bless the reference to get overload and other niceties right
379 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 655 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
380 PUSHs (sv_cbor); 656 PUSHs (sv_cbor);
381 657
386 // catch this surprisingly common error 662 // catch this surprisingly common error
387 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) 663 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
388 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash)); 664 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
389 665
390 encode_tag (enc, CBOR_TAG_PERL_OBJECT); 666 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
391 encode_uint (enc, 0x80, count + 1); 667 encode_uint (enc, MAJOR_ARRAY, count + 1);
392 encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); 668 encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
393 669
394 while (count) 670 {
671 int i;
672
673 for (i = 0; i < count; ++i)
395 encode_sv (enc, SP[1 - count--]); 674 encode_sv (enc, SP[i + 1 - count]);
675
676 SP -= count;
677 }
396 678
397 PUTBACK; 679 PUTBACK;
398 680
399 FREETMPS; LEAVE; 681 FREETMPS; LEAVE;
400 } 682 }
418{ 700{
419 double nv = SvNVX (sv); 701 double nv = SvNVX (sv);
420 702
421 need (enc, 9); 703 need (enc, 9);
422 704
423 if (ecb_expect_false (nv == (U32)nv)) 705 if (ecb_expect_false (nv == (NV)(U32)nv))
424 encode_uint (enc, 0x00, (U32)nv); 706 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
425 //TODO: maybe I32? 707 //TODO: maybe I32?
426 else if (ecb_expect_false (nv == (float)nv)) 708 else if (ecb_expect_false (nv == (float)nv))
427 { 709 encode_float32 (enc, nv);
428 uint32_t fp = ecb_float_to_binary32 (nv);
429
430 *enc->cur++ = 0xe0 | 26;
431
432 if (!ecb_big_endian ())
433 fp = ecb_bswap32 (fp);
434
435 memcpy (enc->cur, &fp, 4);
436 enc->cur += 4;
437 }
438 else 710 else
439 { 711 encode_float64 (enc, nv);
440 uint64_t fp = ecb_double_to_binary64 (nv);
441
442 *enc->cur++ = 0xe0 | 27;
443
444 if (!ecb_big_endian ())
445 fp = ecb_bswap64 (fp);
446
447 memcpy (enc->cur, &fp, 8);
448 enc->cur += 8;
449 }
450} 712}
451 713
452static void 714static void
453encode_sv (enc_t *enc, SV *sv) 715encode_sv (enc_t *enc, SV *sv)
454{ 716{
455 SvGETMAGIC (sv); 717 SvGETMAGIC (sv);
456 718
457 if (encode_sharable (enc, sv))
458 return;
459
460 if (SvPOKp (sv)) 719 if (SvPOKp (sv))
461 { 720 {
462 STRLEN len; 721 STRLEN len;
463 char *str = SvPV (sv, len); 722 char *str = SvPV (sv, len);
464 encode_str (enc, SvUTF8 (sv), str, len); 723 encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len);
465 } 724 }
466 else if (SvNOKp (sv)) 725 else if (SvNOKp (sv))
467 encode_nv (enc, sv); 726 encode_nv (enc, sv);
468 else if (SvIOKp (sv)) 727 else if (SvIOKp (sv))
469 { 728 encode_int (enc, sv);
470 if (SvIsUV (sv))
471 encode_uint (enc, 0x00, SvUVX (sv));
472 else if (SvIVX (sv) >= 0)
473 encode_uint (enc, 0x00, SvIVX (sv));
474 else
475 encode_uint (enc, 0x20, -(SvIVX (sv) + 1));
476 }
477 else if (SvROK (sv)) 729 else if (SvROK (sv))
478 encode_rv (enc, SvRV (sv)); 730 encode_rv (enc, SvRV (sv));
479 else if (!SvOK (sv)) 731 else if (!SvOK (sv))
480 encode_ch (enc, 0xe0 | 22); 732 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
481 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 733 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
482 encode_ch (enc, 0xe0 | 23); 734 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
483 else 735 else
484 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data", 736 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
485 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 737 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
486} 738}
487 739
488static SV * 740static SV *
489encode_cbor (SV *scalar, CBOR *cbor) 741encode_cbor (SV *scalar, CBOR *cbor)
490{ 742{
491 enc_t enc = { }; 743 enc_t enc = { 0 };
492 744
493 enc.cbor = *cbor; 745 enc.cbor = *cbor;
494 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 746 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
495 enc.cur = SvPVX (enc.sv); 747 enc.cur = SvPVX (enc.sv);
496 enc.end = SvEND (enc.sv); 748 enc.end = SvEND (enc.sv);
497 749
498 SvPOK_only (enc.sv); 750 SvPOK_only (enc.sv);
751
752 if (cbor->flags & F_PACK_STRINGS)
753 {
754 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
755 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
756 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
757 }
758
499 encode_sv (&enc, scalar); 759 encode_sv (&enc, scalar);
500 760
501 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv)); 761 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv));
502 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings 762 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings
503 763
517 U8 *end; // end of input string 777 U8 *end; // end of input string
518 const char *err; // parse error, if != 0 778 const char *err; // parse error, if != 0
519 CBOR cbor; 779 CBOR cbor;
520 U32 depth; // recursion depth 780 U32 depth; // recursion depth
521 U32 maxdepth; // recursion depth limit 781 U32 maxdepth; // recursion depth limit
522 AV *sharable; 782 AV *shareable;
783 AV *stringref;
784 SV *decode_tagged;
785 SV *err_sv; // optional sv for error, needs to be freed
523} dec_t; 786} dec_t;
524 787
525#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE 788// set dec->err to ERRSV
789ecb_cold static void
790err_errsv (dec_t *dec)
791{
792 if (!dec->err)
793 {
794 dec->err_sv = newSVsv (ERRSV);
526 795
527#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data") 796 // chop off the trailing \n
797 SvCUR_set (dec->err_sv, SvCUR (dec->err_sv) - 1);
798 *SvEND (dec->err_sv) = 0;
528 799
800 dec->err = SvPVutf8_nolen (dec->err_sv);
801 }
802}
803
804// the following functions are used to reduce code size and help the compiler to optimise
805ecb_cold static void
806err_set (dec_t *dec, const char *reason)
807{
808 if (!dec->err)
809 dec->err = reason;
810}
811
812ecb_cold static void
813err_unexpected_end (dec_t *dec)
814{
815 err_set (dec, "unexpected end of CBOR data");
816}
817
818#define ERR_DO(do) SB do; goto fail; SE
819#define ERR(reason) ERR_DO (err_set (dec, reason))
820#define ERR_ERRSV ERR_DO (err_errsv (dec))
821
822#define WANT(len) if (ecb_expect_false ((uintptr_t)(dec->end - dec->cur) < (STRLEN)len)) ERR_DO (err_unexpected_end (dec))
823
529#define DEC_INC_DEPTH if (++dec->depth > dec->cbor.max_depth) ERR (ERR_NESTING_EXCEEDED) 824#define DEC_INC_DEPTH if (ecb_expect_false (++dec->depth > dec->cbor.max_depth)) ERR (ERR_NESTING_EXCEEDED)
530#define DEC_DEC_DEPTH --dec->depth 825#define DEC_DEC_DEPTH --dec->depth
531 826
532static UV 827static UV
533decode_uint (dec_t *dec) 828decode_uint (dec_t *dec)
534{ 829{
535 switch (*dec->cur & 31) 830 U8 m = *dec->cur & MINOR_MASK;
536 { 831 ++dec->cur;
537 case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
538 case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
539 case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
540 return *dec->cur++ & 31;
541 832
542 case 24: 833 if (ecb_expect_true (m < LENGTH_EXT1))
834 return m;
835 else if (ecb_expect_true (m == LENGTH_EXT1))
836 {
543 WANT (2); 837 WANT (1);
544 dec->cur += 2; 838 dec->cur += 1;
545 return dec->cur[-1]; 839 return dec->cur[-1];
546 840 }
547 case 25: 841 else if (ecb_expect_true (m == LENGTH_EXT2))
842 {
548 WANT (3); 843 WANT (2);
549 dec->cur += 3; 844 dec->cur += 2;
550 return (((UV)dec->cur[-2]) << 8) 845 return (((UV)dec->cur[-2]) << 8)
551 | ((UV)dec->cur[-1]); 846 | ((UV)dec->cur[-1]);
552 847 }
553 case 26: 848 else if (ecb_expect_true (m == LENGTH_EXT4))
849 {
554 WANT (5); 850 WANT (4);
555 dec->cur += 5; 851 dec->cur += 4;
556 return (((UV)dec->cur[-4]) << 24) 852 return (((UV)dec->cur[-4]) << 24)
557 | (((UV)dec->cur[-3]) << 16) 853 | (((UV)dec->cur[-3]) << 16)
558 | (((UV)dec->cur[-2]) << 8) 854 | (((UV)dec->cur[-2]) << 8)
559 | ((UV)dec->cur[-1]); 855 | ((UV)dec->cur[-1]);
560 856 }
561 case 27: 857 else if (ecb_expect_true (m == LENGTH_EXT8))
858 {
562 WANT (9); 859 WANT (8);
563 dec->cur += 9; 860 dec->cur += 8;
861
862 return
863#if UVSIZE < 8
864 0
865#else
564 return (((UV)dec->cur[-8]) << 56) 866 (((UV)dec->cur[-8]) << 56)
565 | (((UV)dec->cur[-7]) << 48) 867 | (((UV)dec->cur[-7]) << 48)
566 | (((UV)dec->cur[-6]) << 40) 868 | (((UV)dec->cur[-6]) << 40)
567 | (((UV)dec->cur[-5]) << 32) 869 | (((UV)dec->cur[-5]) << 32)
870#endif
568 | (((UV)dec->cur[-4]) << 24) 871 | (((UV)dec->cur[-4]) << 24)
569 | (((UV)dec->cur[-3]) << 16) 872 | (((UV)dec->cur[-3]) << 16)
570 | (((UV)dec->cur[-2]) << 8) 873 | (((UV)dec->cur[-2]) << 8)
571 | ((UV)dec->cur[-1]); 874 | ((UV)dec->cur[-1]);
572 875 }
573 default: 876 else
574 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 877 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
575 }
576 878
577fail: 879fail:
578 return 0; 880 return 0;
579} 881}
580 882
585{ 887{
586 AV *av = newAV (); 888 AV *av = newAV ();
587 889
588 DEC_INC_DEPTH; 890 DEC_INC_DEPTH;
589 891
590 if ((*dec->cur & 31) == 31) 892 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
591 { 893 {
592 ++dec->cur; 894 ++dec->cur;
593 895
594 for (;;) 896 for (;;)
595 { 897 {
596 WANT (1); 898 WANT (1);
597 899
598 if (*dec->cur == (0xe0 | 31)) 900 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
599 { 901 {
600 ++dec->cur; 902 ++dec->cur;
601 break; 903 break;
602 } 904 }
603 905
604 av_push (av, decode_sv (dec)); 906 av_push (av, decode_sv (dec));
605 } 907 }
606 } 908 }
607 else 909 else
608 { 910 {
609 int i, len = decode_uint (dec); 911 UV i, len = decode_uint (dec);
610 912
913 WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
611 av_fill (av, len - 1); 914 av_fill (av, len - 1);
612 915
613 for (i = 0; i < len; ++i) 916 for (i = 0; i < len; ++i)
614 AvARRAY (av)[i] = decode_sv (dec); 917 AvARRAY (av)[i] = decode_sv (dec);
615 } 918 }
616 919
617 DEC_DEC_DEPTH; 920 DEC_DEC_DEPTH;
618 return newRV_noinc ((SV *)av); 921 return newRV_noinc ((SV *)av);
619 922
620fail: 923fail:
621 SvREFCNT_dec (av); 924 SvREFCNT_dec_NN (av);
622 DEC_DEC_DEPTH; 925 DEC_DEC_DEPTH;
623 return &PL_sv_undef; 926 return &PL_sv_undef;
624} 927}
625 928
626static void 929static void
627decode_he (dec_t *dec, HV *hv) 930decode_he (dec_t *dec, HV *hv)
628{ 931{
629 // for speed reasons, we specialcase single-string 932 // for speed reasons, we specialcase single-string
630 // byte or utf-8 strings as keys. 933 // byte or utf-8 strings as keys, but only when !stringref
631 934
632 if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27) 935 if (ecb_expect_true (!dec->stringref))
936 if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
633 { 937 {
634 I32 len = decode_uint (dec); 938 STRLEN len = decode_uint (dec);
635 char *key = (char *)dec->cur; 939 char *key = (char *)dec->cur;
636 940
941 WANT (len);
637 dec->cur += len; 942 dec->cur += len;
638 943
639 hv_store (hv, key, len, decode_sv (dec), 0); 944 hv_store (hv, key, len, decode_sv (dec), 0);
945
946 return;
640 } 947 }
641 else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27) 948 else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
642 { 949 {
643 I32 len = decode_uint (dec); 950 STRLEN len = decode_uint (dec);
644 char *key = (char *)dec->cur; 951 char *key = (char *)dec->cur;
645 952
953 WANT (len);
646 dec->cur += len; 954 dec->cur += len;
647 955
956 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
957 if (!is_utf8_string (key, len))
958 ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
959
648 hv_store (hv, key, -len, decode_sv (dec), 0); 960 hv_store (hv, key, -len, decode_sv (dec), 0);
961
962 return;
649 } 963 }
650 else 964
651 {
652 SV *k = decode_sv (dec); 965 SV *k = decode_sv (dec);
653 SV *v = decode_sv (dec); 966 SV *v = decode_sv (dec);
654 967
968 // we leak memory if uncaught exceptions are thrown by random magical
969 // methods, and this is hopefully the only place where it can happen,
970 // so if there is a chance of an exception, take the very slow path.
971 // since catching exceptions is "undocumented/internal/forbidden" by
972 // the new p5p powers, we need to call out to a perl function :/
973 if (ecb_expect_false (SvAMAGIC (k)))
974 {
975 dSP;
976
977 ENTER; SAVETMPS;
978 PUSHMARK (SP);
979 EXTEND (SP, 3);
980 PUSHs (sv_2mortal (newRV_inc ((SV *)hv)));
981 PUSHs (sv_2mortal (k));
982 PUSHs (sv_2mortal (v));
983
984 PUTBACK;
985 call_pv ("CBOR::XS::_hv_store", G_VOID | G_DISCARD | G_EVAL);
986 SPAGAIN;
987
988 FREETMPS; LEAVE;
989
990 if (SvTRUE (ERRSV))
991 ERR_ERRSV;
992
993 return;
994 }
995
655 hv_store_ent (hv, k, v, 0); 996 hv_store_ent (hv, k, v, 0);
656 SvREFCNT_dec (k); 997 SvREFCNT_dec_NN (k);
657 } 998
999fail:
1000 ;
658} 1001}
659 1002
660static SV * 1003static SV *
661decode_hv (dec_t *dec) 1004decode_hv (dec_t *dec)
662{ 1005{
663 HV *hv = newHV (); 1006 HV *hv = newHV ();
664 1007
665 DEC_INC_DEPTH; 1008 DEC_INC_DEPTH;
666 1009
667 if ((*dec->cur & 31) == 31) 1010 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
668 { 1011 {
669 ++dec->cur; 1012 ++dec->cur;
670 1013
671 for (;;) 1014 for (;;)
672 { 1015 {
673 WANT (1); 1016 WANT (1);
674 1017
675 if (*dec->cur == (0xe0 | 31)) 1018 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
676 { 1019 {
677 ++dec->cur; 1020 ++dec->cur;
678 break; 1021 break;
679 } 1022 }
680 1023
681 decode_he (dec, hv); 1024 decode_he (dec, hv);
682 } 1025 }
683 } 1026 }
684 else 1027 else
685 { 1028 {
686 int pairs = decode_uint (dec); 1029 UV pairs = decode_uint (dec);
1030
1031 WANT (pairs); // complexity check - need at least one byte per value, do not allow supersize hashes
687 1032
688 while (pairs--) 1033 while (pairs--)
689 decode_he (dec, hv); 1034 decode_he (dec, hv);
690 } 1035 }
691 1036
692 DEC_DEC_DEPTH; 1037 DEC_DEC_DEPTH;
693 return newRV_noinc ((SV *)hv); 1038 return newRV_noinc ((SV *)hv);
694 1039
695fail: 1040fail:
696 SvREFCNT_dec (hv); 1041 SvREFCNT_dec_NN (hv);
697 DEC_DEC_DEPTH; 1042 DEC_DEC_DEPTH;
698 return &PL_sv_undef; 1043 return &PL_sv_undef;
699} 1044}
700 1045
701static SV * 1046static SV *
702decode_str (dec_t *dec, int utf8) 1047decode_str (dec_t *dec, int utf8)
703{ 1048{
704 SV *sv = 0; 1049 SV *sv = 0;
705 1050
706 if ((*dec->cur & 31) == 31) 1051 if (ecb_expect_false ((*dec->cur & MINOR_MASK) == MINOR_INDEF))
707 { 1052 {
1053 // indefinite length strings
708 ++dec->cur; 1054 ++dec->cur;
709 1055
1056 U8 major = *dec->cur & MAJOR_MISC;
1057
710 sv = newSVpvn ("", 0); 1058 sv = newSVpvn ("", 0);
711 1059
712 // not very fast, and certainly not robust against illegal input
713 for (;;) 1060 for (;;)
714 { 1061 {
715 WANT (1); 1062 WANT (1);
716 1063
717 if (*dec->cur == (0xe0 | 31)) 1064 if ((*dec->cur - major) > LENGTH_EXT8)
1065 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
718 { 1066 {
719 ++dec->cur; 1067 ++dec->cur;
720 break; 1068 break;
721 } 1069 }
1070 else
1071 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
722 1072
723 sv_catsv (sv, decode_sv (dec)); 1073 STRLEN len = decode_uint (dec);
1074
1075 WANT (len);
1076 sv_catpvn (sv, dec->cur, len);
1077 dec->cur += len;
724 } 1078 }
725 } 1079 }
726 else 1080 else
727 { 1081 {
728 STRLEN len = decode_uint (dec); 1082 STRLEN len = decode_uint (dec);
729 1083
730 WANT (len); 1084 WANT (len);
731 sv = newSVpvn (dec->cur, len); 1085 sv = newSVpvn (dec->cur, len);
732 dec->cur += len; 1086 dec->cur += len;
1087
1088 if (ecb_expect_false (dec->stringref)
1089 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
1090 av_push (dec->stringref, SvREFCNT_inc_NN (sv));
733 } 1091 }
734 1092
735 if (utf8) 1093 if (utf8)
1094 {
1095 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
1096 if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
1097 ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
1098
736 SvUTF8_on (sv); 1099 SvUTF8_on (sv);
1100 }
737 1101
738 return sv; 1102 return sv;
739 1103
740fail: 1104fail:
741 SvREFCNT_dec (sv); 1105 SvREFCNT_dec (sv);
743} 1107}
744 1108
745static SV * 1109static SV *
746decode_tagged (dec_t *dec) 1110decode_tagged (dec_t *dec)
747{ 1111{
1112 SV *sv = 0;
748 UV tag = decode_uint (dec); 1113 UV tag = decode_uint (dec);
749 SV *sv = decode_sv (dec); 1114
1115 WANT (1);
750 1116
751 switch (tag) 1117 switch (tag)
752 { 1118 {
753 case CBOR_TAG_MAGIC: 1119 case CBOR_TAG_MAGIC:
754 return sv; 1120 sv = decode_sv (dec);
1121 break;
755 1122
756 case CBOR_TAG_INDIRECTION: 1123 case CBOR_TAG_INDIRECTION:
757 return newRV_noinc (sv); 1124 sv = newRV_noinc (decode_sv (dec));
1125 break;
758 1126
1127 case CBOR_TAG_STRINGREF_NAMESPACE:
1128 {
1129 // do not use SAVETMPS/FREETMPS, as these will
1130 // erase mortalised caches, e.g. "shareable"
1131 ENTER;
1132
1133 SAVESPTR (dec->stringref);
1134 dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());
1135
1136 sv = decode_sv (dec);
1137
1138 LEAVE;
1139 }
1140 break;
1141
1142 case CBOR_TAG_STRINGREF:
1143 {
1144 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
1145 ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
1146
1147 UV idx = decode_uint (dec);
1148
1149 if (!dec->stringref || idx >= (UV)(1 + AvFILLp (dec->stringref)))
1150 ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)");
1151
1152 sv = newSVsv (AvARRAY (dec->stringref)[idx]);
1153 }
1154 break;
1155
759 case CBOR_TAG_VALUE_SHARABLE: 1156 case CBOR_TAG_VALUE_SHAREABLE:
1157 {
760 if (ecb_expect_false (!dec->sharable)) 1158 if (ecb_expect_false (!dec->shareable))
761 dec->sharable = (AV *)sv_2mortal ((SV *)newAV ()); 1159 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
762 1160
1161 if (dec->cbor.flags & F_ALLOW_CYCLES)
1162 {
1163 sv = newSV (0);
763 av_push (dec->sharable, SvREFCNT_inc_NN (sv)); 1164 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
764 1165
765 return sv; 1166 SV *osv = decode_sv (dec);
1167 sv_setsv (sv, osv);
1168 SvREFCNT_dec_NN (osv);
1169 }
1170 else
1171 {
1172 av_push (dec->shareable, &PL_sv_undef);
1173 int idx = AvFILLp (dec->shareable);
1174 sv = decode_sv (dec);
1175 av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
1176 }
1177 }
1178 break;
766 1179
767 case CBOR_TAG_VALUE_SHAREDREF: 1180 case CBOR_TAG_VALUE_SHAREDREF:
768 { 1181 {
769 // TODO: should verify that the sv atcually was a CBOR unsigned integer 1182 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
770 UV idx = SvUV (sv); 1183 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
771 1184
1185 UV idx = decode_uint (dec);
1186
772 if (!dec->sharable || idx > AvFILLp (dec->sharable)) 1187 if (!dec->shareable || idx >= (UV)(1 + AvFILLp (dec->shareable)))
773 ERR ("corrupted CBOR data (sharedref index out of bounds)"); 1188 ERR ("corrupted CBOR data (sharedref index out of bounds)");
774 1189
775 SvREFCNT_dec (sv);
776
777 return SvREFCNT_inc_NN (AvARRAY (dec->sharable)[idx]); 1190 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
1191
1192 if (sv == &PL_sv_undef)
1193 ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
778 } 1194 }
1195 break;
779 1196
780 case CBOR_TAG_PERL_OBJECT: 1197 case CBOR_TAG_PERL_OBJECT:
781 { 1198 {
1199 if (dec->cbor.flags & F_FORBID_OBJECTS)
1200 goto filter;
1201
1202 sv = decode_sv (dec);
1203
782 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) 1204 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
783 ERR ("corrupted CBOR data (non-array perl object)"); 1205 ERR ("corrupted CBOR data (non-array perl object)");
784 1206
785 AV *av = (AV *)SvRV (sv); 1207 AV *av = (AV *)SvRV (sv);
786 int len = av_len (av) + 1; 1208 int len = av_len (av) + 1;
794 if (!method) 1216 if (!method)
795 ERR ("cannot decode perl-object (package does not have a THAW method)"); 1217 ERR ("cannot decode perl-object (package does not have a THAW method)");
796 1218
797 dSP; 1219 dSP;
798 1220
799 ENTER; SAVETMPS; PUSHMARK (SP); 1221 ENTER; SAVETMPS;
1222 PUSHMARK (SP);
800 EXTEND (SP, len + 1); 1223 EXTEND (SP, len + 1);
801 // we re-bless the reference to get overload and other niceties right 1224 // we re-bless the reference to get overload and other niceties right
802 PUSHs (*av_fetch (av, 0, 1)); 1225 PUSHs (*av_fetch (av, 0, 1));
803 PUSHs (sv_cbor); 1226 PUSHs (sv_cbor);
804 1227
812 SPAGAIN; 1235 SPAGAIN;
813 1236
814 if (SvTRUE (ERRSV)) 1237 if (SvTRUE (ERRSV))
815 { 1238 {
816 FREETMPS; LEAVE; 1239 FREETMPS; LEAVE;
817 ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV)))); 1240 ERR_ERRSV;
818 } 1241 }
819 1242
820 SvREFCNT_dec (sv); 1243 SvREFCNT_dec_NN (sv);
821 sv = SvREFCNT_inc (POPs); 1244 sv = SvREFCNT_inc (POPs);
822 1245
823 PUTBACK; 1246 PUTBACK;
824 1247
825 FREETMPS; LEAVE; 1248 FREETMPS; LEAVE;
826
827 return sv;
828 } 1249 }
1250 break;
829 1251
830 default: 1252 default:
1253 filter:
1254 {
1255 SV *tag_sv = newSVuv (tag);
1256
1257 sv = decode_sv (dec);
1258
1259 dSP;
1260 ENTER; SAVETMPS;
1261 PUSHMARK (SP);
1262 EXTEND (SP, 2);
1263 PUSHs (tag_sv);
1264 PUSHs (sv);
1265
1266 PUTBACK;
1267 int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL);
1268 SPAGAIN;
1269
1270 if (SvTRUE (ERRSV))
831 { 1271 {
1272 SvREFCNT_dec_NN (tag_sv);
1273 FREETMPS; LEAVE;
1274 ERR_ERRSV;
1275 }
1276
1277 if (count)
1278 {
1279 SvREFCNT_dec_NN (tag_sv);
1280 SvREFCNT_dec_NN (sv);
1281 sv = SvREFCNT_inc_NN (TOPs);
1282 SP -= count;
1283 }
1284 else
1285 {
832 AV *av = newAV (); 1286 AV *av = newAV ();
833 av_push (av, newSVuv (tag)); 1287 av_push (av, tag_sv);
834 av_push (av, sv); 1288 av_push (av, sv);
835 1289
836 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 1290 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
837 ? cbor_tagged_stash 1291 ? cbor_tagged_stash
838 : gv_stashpv ("CBOR::XS::Tagged" , 1); 1292 : gv_stashpv ("CBOR::XS::Tagged" , 1);
839
840 return sv_bless (newRV_noinc ((SV *)av), tagged_stash); 1293 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
841 } 1294 }
1295
1296 PUTBACK;
1297
1298 FREETMPS; LEAVE;
1299 }
1300 break;
842 } 1301 }
1302
1303 return sv;
843 1304
844fail: 1305fail:
845 SvREFCNT_dec (sv); 1306 SvREFCNT_dec (sv);
846 return &PL_sv_undef; 1307 return &PL_sv_undef;
847} 1308}
849static SV * 1310static SV *
850decode_sv (dec_t *dec) 1311decode_sv (dec_t *dec)
851{ 1312{
852 WANT (1); 1313 WANT (1);
853 1314
854 switch (*dec->cur >> 5) 1315 switch (*dec->cur >> MAJOR_SHIFT)
855 { 1316 {
856 case 0: // unsigned int 1317 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
857 return newSVuv (decode_uint (dec)); 1318 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
858 case 1: // negative int 1319 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
859 return newSViv (-1 - (IV)decode_uint (dec)); 1320 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
860 case 2: // octet string 1321 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
861 return decode_str (dec, 0); 1322 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
862 case 3: // utf-8 string 1323 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
863 return decode_str (dec, 1); 1324
864 case 4: // array 1325 case MAJOR_MISC >> MAJOR_SHIFT:
865 return decode_av (dec);
866 case 5: // map
867 return decode_hv (dec);
868 case 6: // tag
869 return decode_tagged (dec);
870 case 7: // misc
871 switch (*dec->cur++ & 31) 1326 switch (*dec->cur++ & MINOR_MASK)
872 { 1327 {
873 case 20: 1328 case SIMPLE_FALSE:
874#if CBOR_SLOW 1329#if CBOR_SLOW
875 types_false = get_bool ("Types::Serialiser::false"); 1330 types_false = get_bool ("Types::Serialiser::false");
876#endif 1331#endif
877 return newSVsv (types_false); 1332 return newSVsv (types_false);
878 case 21: 1333 case SIMPLE_TRUE:
879#if CBOR_SLOW 1334#if CBOR_SLOW
880 types_true = get_bool ("Types::Serialiser::true"); 1335 types_true = get_bool ("Types::Serialiser::true");
881#endif 1336#endif
882 return newSVsv (types_true); 1337 return newSVsv (types_true);
883 case 22: 1338 case SIMPLE_NULL:
884 return newSVsv (&PL_sv_undef); 1339 return newSVsv (&PL_sv_undef);
885 case 23: 1340 case SIMPLE_UNDEF:
886#if CBOR_SLOW 1341#if CBOR_SLOW
887 types_error = get_bool ("Types::Serialiser::error"); 1342 types_error = get_bool ("Types::Serialiser::error");
888#endif 1343#endif
889 return newSVsv (types_error); 1344 return newSVsv (types_error);
890 1345
891 case 25: 1346 case MISC_FLOAT16:
892 { 1347 {
893 WANT (2); 1348 WANT (2);
894 1349
895 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1350 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
896 dec->cur += 2; 1351 dec->cur += 2;
897 1352
898 return newSVnv (ecb_binary16_to_float (fp)); 1353 return newSVnv (ecb_binary16_to_float (fp));
899 } 1354 }
900 1355
901 case 26: 1356 case MISC_FLOAT32:
902 { 1357 {
903 uint32_t fp; 1358 uint32_t fp;
904 WANT (4); 1359 WANT (4);
905 memcpy (&fp, dec->cur, 4); 1360 memcpy (&fp, dec->cur, 4);
906 dec->cur += 4; 1361 dec->cur += 4;
909 fp = ecb_bswap32 (fp); 1364 fp = ecb_bswap32 (fp);
910 1365
911 return newSVnv (ecb_binary32_to_float (fp)); 1366 return newSVnv (ecb_binary32_to_float (fp));
912 } 1367 }
913 1368
914 case 27: 1369 case MISC_FLOAT64:
915 { 1370 {
916 uint64_t fp; 1371 uint64_t fp;
917 WANT (8); 1372 WANT (8);
918 memcpy (&fp, dec->cur, 8); 1373 memcpy (&fp, dec->cur, 8);
919 dec->cur += 8; 1374 dec->cur += 8;
922 fp = ecb_bswap64 (fp); 1377 fp = ecb_bswap64 (fp);
923 1378
924 return newSVnv (ecb_binary64_to_double (fp)); 1379 return newSVnv (ecb_binary64_to_double (fp));
925 } 1380 }
926 1381
927 // 0..19 unassigned 1382 // 0..19 unassigned simple
928 // 24 reserved + unassigned (reserved values are not encodable) 1383 // 24 reserved + unassigned simple (reserved values are not encodable)
1384 // 28-30 unassigned misc
1385 // 31 break code
929 default: 1386 default:
930 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1387 ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
931 } 1388 }
932 1389
933 break; 1390 break;
934 } 1391 }
935 1392
938} 1395}
939 1396
940static SV * 1397static SV *
941decode_cbor (SV *string, CBOR *cbor, char **offset_return) 1398decode_cbor (SV *string, CBOR *cbor, char **offset_return)
942{ 1399{
943 dec_t dec = { }; 1400 dec_t dec = { 0 };
944 SV *sv; 1401 SV *sv;
945 STRLEN len; 1402 STRLEN len;
946 char *data = SvPVbyte (string, len); 1403 char *data = SvPVbyte (string, len);
947 1404
948 if (len > cbor->max_size && cbor->max_size) 1405 if (len > cbor->max_size && cbor->max_size)
962 if (dec.cur != dec.end && !dec.err) 1419 if (dec.cur != dec.end && !dec.err)
963 dec.err = "garbage after CBOR object"; 1420 dec.err = "garbage after CBOR object";
964 1421
965 if (dec.err) 1422 if (dec.err)
966 { 1423 {
1424 if (dec.shareable)
1425 {
1426 // need to break cyclic links, which would all be in shareable
1427 int i;
1428 SV **svp;
1429
1430 for (i = av_len (dec.shareable) + 1; i--; )
1431 if ((svp = av_fetch (dec.shareable, i, 0)))
1432 sv_setsv (*svp, &PL_sv_undef);
1433 }
1434
967 SvREFCNT_dec (sv); 1435 SvREFCNT_dec_NN (sv);
1436
1437 if (dec.err_sv)
1438 sv_2mortal (dec.err_sv);
1439
968 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur); 1440 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur);
969 } 1441 }
970 1442
971 sv = sv_2mortal (sv); 1443 sv = sv_2mortal (sv);
972 1444
973 return sv; 1445 return sv;
974} 1446}
975 1447
1448/////////////////////////////////////////////////////////////////////////////
1449// incremental parser
1450
1451#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
1452
1453// returns 0 for notyet, 1 for success or error
1454static int
1455incr_parse (CBOR *self, SV *cborstr)
1456{
1457 STRLEN cur;
1458 SvPV (cborstr, cur);
1459
1460 while (ecb_expect_true (self->incr_need <= cur))
1461 {
1462 // table of integer count bytes
1463 static I8 incr_len[MINOR_MASK + 1] = {
1464 0, 0, 0, 0, 0, 0, 0, 0,
1465 0, 0, 0, 0, 0, 0, 0, 0,
1466 0, 0, 0, 0, 0, 0, 0, 0,
1467 1, 2, 4, 8,-1,-1,-1,-2
1468 };
1469
1470 const U8 *p = SvPVX (cborstr) + self->incr_pos;
1471 U8 m = *p & MINOR_MASK;
1472 IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1473 I8 ilen = incr_len[m];
1474
1475 self->incr_need = self->incr_pos + 1;
1476
1477 if (ecb_expect_false (ilen < 0))
1478 {
1479 if (m != MINOR_INDEF)
1480 return 1; // error
1481
1482 if (*p == (MAJOR_MISC | MINOR_INDEF))
1483 {
1484 if (count >= 0)
1485 return 1; // error
1486
1487 count = 1;
1488 }
1489 else
1490 {
1491 av_push (self->incr_count, newSViv (-1)); //TODO: nest
1492 count = -1;
1493 }
1494 }
1495 else
1496 {
1497 self->incr_need += ilen;
1498 if (ecb_expect_false (self->incr_need > cur))
1499 return 0;
1500
1501 int major = *p >> MAJOR_SHIFT;
1502
1503 switch (major)
1504 {
1505 case MAJOR_TAG >> MAJOR_SHIFT:
1506 ++count; // tags merely prefix another value
1507 break;
1508
1509 case MAJOR_BYTES >> MAJOR_SHIFT:
1510 case MAJOR_TEXT >> MAJOR_SHIFT:
1511 case MAJOR_ARRAY >> MAJOR_SHIFT:
1512 case MAJOR_MAP >> MAJOR_SHIFT:
1513 {
1514 UV len;
1515
1516 if (ecb_expect_false (ilen))
1517 {
1518 len = 0;
1519
1520 do {
1521 len = (len << 8) | *++p;
1522 } while (--ilen);
1523 }
1524 else
1525 len = m;
1526
1527 switch (major)
1528 {
1529 case MAJOR_BYTES >> MAJOR_SHIFT:
1530 case MAJOR_TEXT >> MAJOR_SHIFT:
1531 self->incr_need += len;
1532 if (ecb_expect_false (self->incr_need > cur))
1533 return 0;
1534
1535 break;
1536
1537 case MAJOR_MAP >> MAJOR_SHIFT:
1538 len <<= 1;
1539 case MAJOR_ARRAY >> MAJOR_SHIFT:
1540 if (len)
1541 {
1542 av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
1543 count = len + 1;
1544 }
1545 break;
1546 }
1547 }
1548 }
1549 }
1550
1551 self->incr_pos = self->incr_need;
1552
1553 if (count > 0)
1554 {
1555 while (!--count)
1556 {
1557 if (!AvFILLp (self->incr_count))
1558 return 1; // done
1559
1560 SvREFCNT_dec_NN (av_pop (self->incr_count));
1561 count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1562 }
1563
1564 SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
1565 }
1566 }
1567
1568 return 0;
1569}
1570
1571
976///////////////////////////////////////////////////////////////////////////// 1572/////////////////////////////////////////////////////////////////////////////
977// XS interface functions 1573// XS interface functions
978 1574
979MODULE = CBOR::XS PACKAGE = CBOR::XS 1575MODULE = CBOR::XS PACKAGE = CBOR::XS
980 1576
988 1584
989 types_true = get_bool ("Types::Serialiser::true" ); 1585 types_true = get_bool ("Types::Serialiser::true" );
990 types_false = get_bool ("Types::Serialiser::false"); 1586 types_false = get_bool ("Types::Serialiser::false");
991 types_error = get_bool ("Types::Serialiser::error"); 1587 types_error = get_bool ("Types::Serialiser::error");
992 1588
1589 default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1590
993 sv_cbor = newSVpv ("CBOR", 0); 1591 sv_cbor = newSVpv ("CBOR", 0);
994 SvREADONLY_on (sv_cbor); 1592 SvREADONLY_on (sv_cbor);
1593
1594 assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1));
995} 1595}
996 1596
997PROTOTYPES: DISABLE 1597PROTOTYPES: DISABLE
998 1598
999void CLONE (...) 1599void CLONE (...)
1018void shrink (CBOR *self, int enable = 1) 1618void shrink (CBOR *self, int enable = 1)
1019 ALIAS: 1619 ALIAS:
1020 shrink = F_SHRINK 1620 shrink = F_SHRINK
1021 allow_unknown = F_ALLOW_UNKNOWN 1621 allow_unknown = F_ALLOW_UNKNOWN
1022 allow_sharing = F_ALLOW_SHARING 1622 allow_sharing = F_ALLOW_SHARING
1023 dedup_keys = F_DEDUP_KEYS 1623 allow_cycles = F_ALLOW_CYCLES
1624 forbid_objects = F_FORBID_OBJECTS
1024 dedup_strings = F_DEDUP_STRINGS 1625 pack_strings = F_PACK_STRINGS
1626 text_keys = F_TEXT_KEYS
1627 text_strings = F_TEXT_STRINGS
1628 validate_utf8 = F_VALIDATE_UTF8
1025 PPCODE: 1629 PPCODE:
1026{ 1630{
1027 if (enable) 1631 if (enable)
1028 self->flags |= ix; 1632 self->flags |= ix;
1029 else 1633 else
1035void get_shrink (CBOR *self) 1639void get_shrink (CBOR *self)
1036 ALIAS: 1640 ALIAS:
1037 get_shrink = F_SHRINK 1641 get_shrink = F_SHRINK
1038 get_allow_unknown = F_ALLOW_UNKNOWN 1642 get_allow_unknown = F_ALLOW_UNKNOWN
1039 get_allow_sharing = F_ALLOW_SHARING 1643 get_allow_sharing = F_ALLOW_SHARING
1040 get_dedup_keys = F_DEDUP_KEYS 1644 get_allow_cycles = F_ALLOW_CYCLES
1645 get_forbid_objects = F_FORBID_OBJECTS
1041 get_dedup_strings = F_DEDUP_STRINGS 1646 get_pack_strings = F_PACK_STRINGS
1647 get_text_keys = F_TEXT_KEYS
1648 get_text_strings = F_TEXT_STRINGS
1649 get_validate_utf8 = F_VALIDATE_UTF8
1042 PPCODE: 1650 PPCODE:
1043 XPUSHs (boolSV (self->flags & ix)); 1651 XPUSHs (boolSV (self->flags & ix));
1044 1652
1045void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1653void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1046 PPCODE: 1654 PPCODE:
1059 XPUSHs (ST (0)); 1667 XPUSHs (ST (0));
1060 1668
1061int get_max_size (CBOR *self) 1669int get_max_size (CBOR *self)
1062 CODE: 1670 CODE:
1063 RETVAL = self->max_size; 1671 RETVAL = self->max_size;
1672 OUTPUT:
1673 RETVAL
1674
1675void filter (CBOR *self, SV *filter = 0)
1676 PPCODE:
1677 SvREFCNT_dec (self->filter);
1678 self->filter = filter ? newSVsv (filter) : filter;
1679 XPUSHs (ST (0));
1680
1681SV *get_filter (CBOR *self)
1682 CODE:
1683 RETVAL = self->filter ? self->filter : NEWSV (0, 0);
1064 OUTPUT: 1684 OUTPUT:
1065 RETVAL 1685 RETVAL
1066 1686
1067void encode (CBOR *self, SV *scalar) 1687void encode (CBOR *self, SV *scalar)
1068 PPCODE: 1688 PPCODE:
1083 EXTEND (SP, 2); 1703 EXTEND (SP, 2);
1084 PUSHs (sv); 1704 PUSHs (sv);
1085 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1705 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1086} 1706}
1087 1707
1708void incr_parse (CBOR *self, SV *cborstr)
1709 ALIAS:
1710 incr_parse_multiple = 1
1711 PPCODE:
1712{
1713 if (SvUTF8 (cborstr))
1714 sv_utf8_downgrade (cborstr, 0);
1715
1716 if (!self->incr_count)
1717 {
1718 self->incr_count = newAV ();
1719 self->incr_pos = 0;
1720 self->incr_need = 1;
1721
1722 av_push (self->incr_count, newSViv (1));
1723 }
1724
1725 do
1726 {
1727 if (!incr_parse (self, cborstr))
1728 {
1729 if (self->incr_need > self->max_size && self->max_size)
1730 croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
1731 (unsigned long)self->incr_need, (unsigned long)self->max_size);
1732
1733 break;
1734 }
1735
1736 SV *sv;
1737 char *offset;
1738
1739 PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
1740 XPUSHs (sv);
1741
1742 sv_chop (cborstr, offset);
1743
1744 av_clear (self->incr_count);
1745 av_push (self->incr_count, newSViv (1));
1746
1747 self->incr_pos = 0;
1748 self->incr_need = self->incr_pos + 1;
1749 }
1750 while (ix);
1751}
1752
1753void incr_reset (CBOR *self)
1754 CODE:
1755{
1756 SvREFCNT_dec (self->incr_count);
1757 self->incr_count = 0;
1758}
1759
1760void DESTROY (CBOR *self)
1761 PPCODE:
1762 cbor_free (self);
1763
1088PROTOTYPES: ENABLE 1764PROTOTYPES: ENABLE
1089 1765
1090void encode_cbor (SV *scalar) 1766void encode_cbor (SV *scalar)
1767 ALIAS:
1768 encode_cbor = 0
1769 encode_cbor_sharing = F_ALLOW_SHARING
1091 PPCODE: 1770 PPCODE:
1092{ 1771{
1093 CBOR cbor; 1772 CBOR cbor;
1094 cbor_init (&cbor); 1773 cbor_init (&cbor);
1774 cbor.flags |= ix;
1095 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1775 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1096 XPUSHs (scalar); 1776 XPUSHs (scalar);
1097} 1777}
1098 1778
1099void decode_cbor (SV *cborstr) 1779void decode_cbor (SV *cborstr)
1103 cbor_init (&cbor); 1783 cbor_init (&cbor);
1104 PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN; 1784 PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN;
1105 XPUSHs (cborstr); 1785 XPUSHs (cborstr);
1106} 1786}
1107 1787
1788#ifdef __AFL_COMPILER
1789
1790void
1791afl_init ()
1792 CODE:
1793 __AFL_INIT ();
1794
1795int
1796afl_loop (unsigned int count = 10000)
1797 CODE:
1798 RETVAL = __AFL_LOOP (count);
1799 OUTPUT:
1800 RETVAL
1801
1802#endif
1803

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines