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 1.19 by root, Wed Nov 20 02:03:09 2013 UTC vs.
Revision 1.67 by root, Sun Nov 29 21:35:06 2020 UTC

6#include <string.h> 6#include <string.h>
7#include <stdlib.h> 7#include <stdlib.h>
8#include <stdio.h> 8#include <stdio.h>
9#include <limits.h> 9#include <limits.h>
10#include <float.h> 10#include <float.h>
11#include <inttypes.h>
11 12
13#define ECB_NO_THREADS 1
12#include "ecb.h" 14#include "ecb.h"
13 15
14// compatibility with perl <5.18 16// compatibility with perl <5.18
15#ifndef HvNAMELEN_get 17#ifndef HvNAMELEN_get
16# define HvNAMELEN_get(hv) strlen (HvNAME (hv)) 18# define HvNAMELEN_get(hv) strlen (HvNAME (hv))
19# define HvNAMELEN(hv) HvNAMELEN_get (hv) 21# define HvNAMELEN(hv) HvNAMELEN_get (hv)
20#endif 22#endif
21#ifndef HvNAMEUTF8 23#ifndef HvNAMEUTF8
22# define HvNAMEUTF8(hv) 0 24# define HvNAMEUTF8(hv) 0
23#endif 25#endif
26#ifndef SvREFCNT_inc_NN
27# define SvREFCNT_inc_NN(sv) SvREFCNT_inc (sv)
28#endif
29#ifndef SvREFCNT_dec_NN
30# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
31#endif
32
33// known major and minor types
34enum cbor_type
35{
36 MAJOR_SHIFT = 5,
37 MINOR_MASK = 0x1f,
38
39 MAJOR_POS_INT = 0 << MAJOR_SHIFT,
40 MAJOR_NEG_INT = 1 << MAJOR_SHIFT,
41 MAJOR_BYTES = 2 << MAJOR_SHIFT,
42 MAJOR_TEXT = 3 << MAJOR_SHIFT,
43 MAJOR_ARRAY = 4 << MAJOR_SHIFT,
44 MAJOR_MAP = 5 << MAJOR_SHIFT,
45 MAJOR_TAG = 6 << MAJOR_SHIFT,
46 MAJOR_MISC = 7 << MAJOR_SHIFT,
47
48 // INT/STRING/ARRAY/MAP subtypes
49 LENGTH_EXT1 = 24,
50 LENGTH_EXT2 = 25,
51 LENGTH_EXT4 = 26,
52 LENGTH_EXT8 = 27,
53
54 // SIMPLE types (effectively MISC subtypes)
55 SIMPLE_FALSE = 20,
56 SIMPLE_TRUE = 21,
57 SIMPLE_NULL = 22,
58 SIMPLE_UNDEF = 23,
59
60 // MISC subtype (unused)
61 MISC_EXT1 = 24,
62 MISC_FLOAT16 = 25,
63 MISC_FLOAT32 = 26,
64 MISC_FLOAT64 = 27,
65
66 // BYTES/TEXT/ARRAY/MAP
67 MINOR_INDEF = 31,
68};
24 69
25// known tags 70// known tags
26enum cbor_tag 71enum cbor_tag
27{ 72{
28 // inofficial extensions (pending iana registration) 73 // extensions
74 CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
29 CBOR_TAG_PERL_OBJECT = 24, // http://cbor.schmorp.de/perl-object 75 CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
30 CBOR_TAG_GENERIC_OBJECT = 25, // http://cbor.schmorp.de/generic-object 76 CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
31 CBOR_TAG_VALUE_SHAREABLE = 26, // http://cbor.schmorp.de/value-sharing 77 CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
32 CBOR_TAG_VALUE_SHAREDREF = 27, // http://cbor.schmorp.de/value-sharing 78 CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
33 CBOR_TAG_STRINGREF_NAMESPACE = 65537, // http://cbor.schmorp.de/stringref 79 CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref
34 CBOR_TAG_STRINGREF = 28, // http://cbor.schmorp.de/stringref
35 CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection 80 CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
36 81
37 // rfc7049 82 // rfc7049
38 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8 83 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
39 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any 84 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
40 CBOR_TAG_POS_BIGNUM = 2, // byte string 85 CBOR_TAG_POS_BIGNUM = 2, // byte string
41 CBOR_TAG_NEG_BIGNUM = 3, // byte string 86 CBOR_TAG_NEG_BIGNUM = 3, // byte string
42 CBOR_TAG_DECIMAL = 4, // decimal fraction, array 87 CBOR_TAG_DECIMAL = 4, // decimal fraction, array
43 CBOR_TAG_BIGFLOAT = 5, // array 88 CBOR_TAG_BIGFLOAT = 5, // array
44 89
45 CBOR_TAG_CONV_B64U = 21, // base64url, any 90 CBOR_TAG_CONV_B64U = 21, // base64url, any
46 CBOR_TAG_CONV_B64 = 22, // base64, any 91 CBOR_TAG_CONV_B64 = 22, // base64, any
47 CBOR_TAG_CONV_HEX = 23, // base16, any 92 CBOR_TAG_CONV_HEX = 23, // base16, any
48 CBOR_TAG_CBOR = 24, // embedded cbor, byte string 93 CBOR_TAG_CBOR = 24, // embedded cbor, byte string
49 94
50 CBOR_TAG_URI = 32, // URI rfc3986, utf-8 95 CBOR_TAG_URI = 32, // URI rfc3986, utf-8
51 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8 96 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
52 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8 97 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
53 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8 98 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
54 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8 99 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
55 100
56 CBOR_TAG_MAGIC = 55799 // self-describe cbor 101 CBOR_TAG_MAGIC = 55799, // self-describe cbor
57}; 102};
58 103
104// known forced types, also hardcoded in CBOR.pm
105enum
106{
107 AS_CBOR = 0,
108 AS_BYTES = 1,
109 AS_TEXT = 2,
110 AS_INT = 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 *shareable; // ptr => index, or 0 228 HV *shareable; // ptr => index, or 0
142 UV shareable_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
350encode_strref (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
351{
352 if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
353 {
354 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
355
356 if (SvOK (*svp))
357 {
358 // already registered, use stringref
359 encode_tag (enc, CBOR_TAG_STRINGREF);
360 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
361 return;
362 }
363 else if (len >= minimum_string_length (enc->stringref_idx))
364 {
365 // register only
366 sv_setuv (*svp, enc->stringref_idx);
367 ++enc->stringref_idx;
368 }
369 }
370
371 encode_str (enc, upgrade_utf8, utf8, str, len);
372}
373
374ecb_inline void
214encode_tag (enc_t *enc, UV tag) 375encode_float16 (enc_t *enc, NV nv)
215{ 376{
216 encode_uint (enc, 0xc0, tag); 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_forced (enc_t *enc, UV type, SV *sv)
424{
425 switch (type)
426 {
427 case AS_CBOR:
428 {
429 STRLEN len;
430 char *str = SvPVbyte (sv, len);
431
432 need (enc, len);
433 memcpy (enc->cur, str, len);
434 enc->cur += len;
435 }
436 break;
437
438 case AS_BYTES:
439 {
440 STRLEN len;
441 char *str = SvPVbyte (sv, len);
442 encode_strref (enc, 0, 0, str, len);
443 }
444 break;
445
446 case AS_TEXT:
447 {
448 STRLEN len;
449 char *str = SvPVutf8 (sv, len);
450 encode_strref (enc, 1, 1, str, len);
451 }
452 break;
453
454 case AS_INT: encode_int (enc, sv); break;
455 case AS_FLOAT16: encode_float16 (enc, SvNV (sv)); break;
456 case AS_FLOAT32: encode_float32 (enc, SvNV (sv)); break;
457 case AS_FLOAT64: encode_float64 (enc, SvNV (sv)); break;
458
459 default:
460 croak ("encountered malformed CBOR::XS::Tagged object");
461 }
217} 462}
218 463
219static void encode_sv (enc_t *enc, SV *sv); 464static void encode_sv (enc_t *enc, SV *sv);
220 465
221static void 466static void
226 if (enc->depth >= enc->cbor.max_depth) 471 if (enc->depth >= enc->cbor.max_depth)
227 croak (ERR_NESTING_EXCEEDED); 472 croak (ERR_NESTING_EXCEEDED);
228 473
229 ++enc->depth; 474 ++enc->depth;
230 475
231 encode_uint (enc, 0x80, len + 1); 476 encode_uint (enc, MAJOR_ARRAY, len + 1);
232 477
478 if (ecb_expect_false (SvMAGICAL (av)))
233 for (i = 0; i <= len; ++i) 479 for (i = 0; i <= len; ++i)
234 { 480 {
235 SV **svp = av_fetch (av, i, 0); 481 SV **svp = av_fetch (av, i, 0);
236 encode_sv (enc, svp ? *svp : &PL_sv_undef); 482 encode_sv (enc, svp ? *svp : &PL_sv_undef);
237 } 483 }
484 else
485 for (i = 0; i <= len; ++i)
486 {
487 SV *sv = AvARRAY (av)[i];
488 encode_sv (enc, sv ? sv : &PL_sv_undef);
489 }
238 490
239 --enc->depth; 491 --enc->depth;
240} 492}
241 493
242static void 494static void
250 ++enc->depth; 502 ++enc->depth;
251 503
252 int pairs = hv_iterinit (hv); 504 int pairs = hv_iterinit (hv);
253 int mg = SvMAGICAL (hv); 505 int mg = SvMAGICAL (hv);
254 506
255 if (mg) 507 if (ecb_expect_false (mg))
256 encode_ch (enc, 0xa0 | 31); 508 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
257 else 509 else
258 encode_uint (enc, 0xa0, pairs); 510 encode_uint (enc, MAJOR_MAP, pairs);
259 511
260 while ((he = hv_iternext (hv))) 512 while ((he = hv_iternext (hv)))
261 { 513 {
262 if (HeKLEN (he) == HEf_SVKEY) 514 if (HeKLEN (he) == HEf_SVKEY)
263 encode_sv (enc, HeSVKEY (he)); 515 encode_sv (enc, HeSVKEY (he));
264 else 516 else
265 encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); 517 encode_strref (enc, enc->cbor.flags & (F_TEXT_KEYS | F_TEXT_STRINGS), HeKUTF8 (he), HeKEY (he), HeKLEN (he));
266 518
267 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); 519 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
268 } 520 }
269 521
270 if (mg) 522 if (ecb_expect_false (mg))
271 encode_ch (enc, 0xe0 | 31); 523 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
272 524
273 --enc->depth; 525 --enc->depth;
274} 526}
275 527
276// encode objects, arrays and special \0=false and \1=true values. 528// encode objects, arrays and special \0=false and \1=true values.
277static void 529static void
278encode_rv (enc_t *enc, SV *sv) 530encode_rv (enc_t *enc, SV *sv)
279{ 531{
280 SvGETMAGIC (sv); 532 SvGETMAGIC (sv);
281
282 if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)
283 && ecb_expect_false (SvREFCNT (sv) > 1))
284 {
285 if (!enc->shareable)
286 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
287
288 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
289
290 if (SvOK (*svp))
291 {
292 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
293 encode_uint (enc, 0x00, SvUV (*svp));
294 return;
295 }
296 else
297 {
298 sv_setuv (*svp, enc->shareable_idx);
299 ++enc->shareable_idx;
300 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
301 }
302 }
303 533
304 svtype svt = SvTYPE (sv); 534 svtype svt = SvTYPE (sv);
305 535
306 if (ecb_expect_false (SvOBJECT (sv))) 536 if (ecb_expect_false (SvOBJECT (sv)))
307 { 537 {
314 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 544 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
315 ? cbor_tagged_stash 545 ? cbor_tagged_stash
316 : gv_stashpv ("CBOR::XS::Tagged" , 1); 546 : gv_stashpv ("CBOR::XS::Tagged" , 1);
317 547
318 HV *stash = SvSTASH (sv); 548 HV *stash = SvSTASH (sv);
319 GV *method;
320 549
321 if (stash == boolean_stash) 550 if (stash == boolean_stash)
322 encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20); 551 {
552 encode_ch (enc, SvIV (sv) ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE);
553 return;
554 }
323 else if (stash == error_stash) 555 else if (stash == error_stash)
324 encode_ch (enc, 0xe0 | 23); 556 {
557 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
558 return;
559 }
325 else if (stash == tagged_stash) 560 else if (stash == tagged_stash)
326 { 561 {
327 if (svt != SVt_PVAV) 562 if (svt != SVt_PVAV)
328 croak ("encountered CBOR::XS::Tagged object that isn't an array"); 563 croak ("encountered CBOR::XS::Tagged object that isn't an array");
329 564
565 switch (av_len ((AV *)sv))
566 {
567 case 2-1:
568 // actually a tagged value
330 encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); 569 encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
331 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); 570 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1));
571 break;
572
573 case 3-1:
574 // a forced type [value, type, undef]
575 encode_forced (enc, SvUV (*av_fetch ((AV *)sv, 1, 1)), *av_fetch ((AV *)sv, 0, 1));
576 break;
577
578 default:
579 croak ("encountered malformed CBOR::XS::Tagged object");
580 }
581
582 return;
583 }
584 }
585
586 if (ecb_expect_false (SvREFCNT (sv) > 1)
587 && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
588 {
589 if (ecb_expect_false (!enc->shareable))
590 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
591
592 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
593
594 if (SvOK (*svp))
332 } 595 {
596 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
597 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
598 return;
599 }
600 else
601 {
602 sv_setuv (*svp, enc->shareable_idx);
603 ++enc->shareable_idx;
604 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
605 }
606 }
607
608 if (ecb_expect_false (SvOBJECT (sv)))
609 {
610 HV *stash = SvSTASH (sv);
611 GV *method;
612
613 if (enc->cbor.flags & F_FORBID_OBJECTS)
614 croak ("encountered object '%s', but forbid_objects is enabled",
615 SvPV_nolen (sv_2mortal (newRV_inc (sv))));
333 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) 616 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
334 { 617 {
335 dSP; 618 dSP;
336 619
337 ENTER; SAVETMPS; PUSHMARK (SP); 620 ENTER; SAVETMPS;
621 PUSHMARK (SP);
338 // we re-bless the reference to get overload and other niceties right 622 // we re-bless the reference to get overload and other niceties right
339 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 623 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
340 624
341 PUTBACK; 625 PUTBACK;
342 // G_SCALAR ensures that return value is 1 626 // G_SCALAR ensures that return value is 1
355 } 639 }
356 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0) 640 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
357 { 641 {
358 dSP; 642 dSP;
359 643
360 ENTER; SAVETMPS; PUSHMARK (SP); 644 ENTER; SAVETMPS;
645 PUSHMARK (SP);
361 EXTEND (SP, 2); 646 EXTEND (SP, 2);
362 // we re-bless the reference to get overload and other niceties right 647 // we re-bless the reference to get overload and other niceties right
363 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 648 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
364 PUSHs (sv_cbor); 649 PUSHs (sv_cbor);
365 650
370 // catch this surprisingly common error 655 // catch this surprisingly common error
371 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) 656 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
372 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash)); 657 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
373 658
374 encode_tag (enc, CBOR_TAG_PERL_OBJECT); 659 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
375 encode_uint (enc, 0x80, count + 1); 660 encode_uint (enc, MAJOR_ARRAY, count + 1);
376 encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); 661 encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
377 662
378 while (count) 663 {
664 int i;
665
666 for (i = 0; i < count; ++i)
379 encode_sv (enc, SP[1 - count--]); 667 encode_sv (enc, SP[i + 1 - count]);
668
669 SP -= count;
670 }
380 671
381 PUTBACK; 672 PUTBACK;
382 673
383 FREETMPS; LEAVE; 674 FREETMPS; LEAVE;
384 } 675 }
402{ 693{
403 double nv = SvNVX (sv); 694 double nv = SvNVX (sv);
404 695
405 need (enc, 9); 696 need (enc, 9);
406 697
407 if (ecb_expect_false (nv == (U32)nv)) 698 if (ecb_expect_false (nv == (NV)(U32)nv))
408 encode_uint (enc, 0x00, (U32)nv); 699 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
409 //TODO: maybe I32? 700 //TODO: maybe I32?
410 else if (ecb_expect_false (nv == (float)nv)) 701 else if (ecb_expect_false (nv == (float)nv))
411 { 702 encode_float32 (enc, nv);
412 uint32_t fp = ecb_float_to_binary32 (nv);
413
414 *enc->cur++ = 0xe0 | 26;
415
416 if (!ecb_big_endian ())
417 fp = ecb_bswap32 (fp);
418
419 memcpy (enc->cur, &fp, 4);
420 enc->cur += 4;
421 }
422 else 703 else
423 { 704 encode_float64 (enc, nv);
424 uint64_t fp = ecb_double_to_binary64 (nv);
425
426 *enc->cur++ = 0xe0 | 27;
427
428 if (!ecb_big_endian ())
429 fp = ecb_bswap64 (fp);
430
431 memcpy (enc->cur, &fp, 8);
432 enc->cur += 8;
433 }
434} 705}
435 706
436static void 707static void
437encode_sv (enc_t *enc, SV *sv) 708encode_sv (enc_t *enc, SV *sv)
438{ 709{
440 711
441 if (SvPOKp (sv)) 712 if (SvPOKp (sv))
442 { 713 {
443 STRLEN len; 714 STRLEN len;
444 char *str = SvPV (sv, len); 715 char *str = SvPV (sv, len);
445 encode_str (enc, SvUTF8 (sv), str, len); 716 encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len);
446 } 717 }
447 else if (SvNOKp (sv)) 718 else if (SvNOKp (sv))
448 encode_nv (enc, sv); 719 encode_nv (enc, sv);
449 else if (SvIOKp (sv)) 720 else if (SvIOKp (sv))
450 { 721 encode_int (enc, sv);
451 if (SvIsUV (sv))
452 encode_uint (enc, 0x00, SvUVX (sv));
453 else if (SvIVX (sv) >= 0)
454 encode_uint (enc, 0x00, SvIVX (sv));
455 else
456 encode_uint (enc, 0x20, -(SvIVX (sv) + 1));
457 }
458 else if (SvROK (sv)) 722 else if (SvROK (sv))
459 encode_rv (enc, SvRV (sv)); 723 encode_rv (enc, SvRV (sv));
460 else if (!SvOK (sv)) 724 else if (!SvOK (sv))
461 encode_ch (enc, 0xe0 | 22); 725 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
462 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 726 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
463 encode_ch (enc, 0xe0 | 23); 727 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
464 else 728 else
465 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data", 729 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
466 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 730 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
467} 731}
468 732
469static SV * 733static SV *
470encode_cbor (SV *scalar, CBOR *cbor) 734encode_cbor (SV *scalar, CBOR *cbor)
471{ 735{
472 enc_t enc = { }; 736 enc_t enc = { 0 };
473 737
474 enc.cbor = *cbor; 738 enc.cbor = *cbor;
475 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 739 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
476 enc.cur = SvPVX (enc.sv); 740 enc.cur = SvPVX (enc.sv);
477 enc.end = SvEND (enc.sv); 741 enc.end = SvEND (enc.sv);
478 742
479 SvPOK_only (enc.sv); 743 SvPOK_only (enc.sv);
744
745 if (cbor->flags & F_PACK_STRINGS)
746 {
747 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
748 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
749 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
750 }
751
480 encode_sv (&enc, scalar); 752 encode_sv (&enc, scalar);
481 753
482 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv)); 754 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv));
483 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings 755 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings
484 756
499 const char *err; // parse error, if != 0 771 const char *err; // parse error, if != 0
500 CBOR cbor; 772 CBOR cbor;
501 U32 depth; // recursion depth 773 U32 depth; // recursion depth
502 U32 maxdepth; // recursion depth limit 774 U32 maxdepth; // recursion depth limit
503 AV *shareable; 775 AV *shareable;
776 AV *stringref;
777 SV *decode_tagged;
778 SV *err_sv; // optional sv for error, needs to be freed
504} dec_t; 779} dec_t;
505 780
506#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE 781// set dec->err to ERRSV
782ecb_cold static void
783err_errsv (dec_t *dec)
784{
785 if (!dec->err)
786 {
787 dec->err_sv = newSVsv (ERRSV);
507 788
508#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data") 789 // chop off the trailing \n
790 SvCUR_set (dec->err_sv, SvCUR (dec->err_sv) - 1);
791 *SvEND (dec->err_sv) = 0;
509 792
793 dec->err = SvPVutf8_nolen (dec->err_sv);
794 }
795}
796
797// the following functions are used to reduce code size and help the compiler to optimise
798ecb_cold static void
799err_set (dec_t *dec, const char *reason)
800{
801 if (!dec->err)
802 dec->err = reason;
803}
804
805ecb_cold static void
806err_unexpected_end (dec_t *dec)
807{
808 err_set (dec, "unexpected end of CBOR data");
809}
810
811#define ERR_DO(do) SB do; goto fail; SE
812#define ERR(reason) ERR_DO (err_set (dec, reason))
813#define ERR_ERRSV ERR_DO (err_errsv (dec))
814
815#define WANT(len) if (ecb_expect_false ((uintptr_t)(dec->end - dec->cur) < (STRLEN)len)) ERR_DO (err_unexpected_end (dec))
816
510#define DEC_INC_DEPTH if (++dec->depth > dec->cbor.max_depth) ERR (ERR_NESTING_EXCEEDED) 817#define DEC_INC_DEPTH if (ecb_expect_false (++dec->depth > dec->cbor.max_depth)) ERR (ERR_NESTING_EXCEEDED)
511#define DEC_DEC_DEPTH --dec->depth 818#define DEC_DEC_DEPTH --dec->depth
512 819
513static UV 820static UV
514decode_uint (dec_t *dec) 821decode_uint (dec_t *dec)
515{ 822{
516 switch (*dec->cur & 31) 823 U8 m = *dec->cur & MINOR_MASK;
517 { 824 ++dec->cur;
518 case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
519 case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
520 case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
521 return *dec->cur++ & 31;
522 825
523 case 24: 826 if (ecb_expect_true (m < LENGTH_EXT1))
827 return m;
828 else if (ecb_expect_true (m == LENGTH_EXT1))
829 {
524 WANT (2); 830 WANT (1);
525 dec->cur += 2; 831 dec->cur += 1;
526 return dec->cur[-1]; 832 return dec->cur[-1];
527 833 }
528 case 25: 834 else if (ecb_expect_true (m == LENGTH_EXT2))
835 {
529 WANT (3); 836 WANT (2);
530 dec->cur += 3; 837 dec->cur += 2;
531 return (((UV)dec->cur[-2]) << 8) 838 return (((UV)dec->cur[-2]) << 8)
532 | ((UV)dec->cur[-1]); 839 | ((UV)dec->cur[-1]);
533 840 }
534 case 26: 841 else if (ecb_expect_true (m == LENGTH_EXT4))
842 {
535 WANT (5); 843 WANT (4);
536 dec->cur += 5; 844 dec->cur += 4;
537 return (((UV)dec->cur[-4]) << 24) 845 return (((UV)dec->cur[-4]) << 24)
538 | (((UV)dec->cur[-3]) << 16) 846 | (((UV)dec->cur[-3]) << 16)
539 | (((UV)dec->cur[-2]) << 8) 847 | (((UV)dec->cur[-2]) << 8)
540 | ((UV)dec->cur[-1]); 848 | ((UV)dec->cur[-1]);
541 849 }
542 case 27: 850 else if (ecb_expect_true (m == LENGTH_EXT8))
851 {
543 WANT (9); 852 WANT (8);
544 dec->cur += 9; 853 dec->cur += 8;
854
855 return
856#if UVSIZE < 8
857 0
858#else
545 return (((UV)dec->cur[-8]) << 56) 859 (((UV)dec->cur[-8]) << 56)
546 | (((UV)dec->cur[-7]) << 48) 860 | (((UV)dec->cur[-7]) << 48)
547 | (((UV)dec->cur[-6]) << 40) 861 | (((UV)dec->cur[-6]) << 40)
548 | (((UV)dec->cur[-5]) << 32) 862 | (((UV)dec->cur[-5]) << 32)
863#endif
549 | (((UV)dec->cur[-4]) << 24) 864 | (((UV)dec->cur[-4]) << 24)
550 | (((UV)dec->cur[-3]) << 16) 865 | (((UV)dec->cur[-3]) << 16)
551 | (((UV)dec->cur[-2]) << 8) 866 | (((UV)dec->cur[-2]) << 8)
552 | ((UV)dec->cur[-1]); 867 | ((UV)dec->cur[-1]);
553 868 }
554 default: 869 else
555 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 870 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
556 }
557 871
558fail: 872fail:
559 return 0; 873 return 0;
560} 874}
561 875
566{ 880{
567 AV *av = newAV (); 881 AV *av = newAV ();
568 882
569 DEC_INC_DEPTH; 883 DEC_INC_DEPTH;
570 884
571 if ((*dec->cur & 31) == 31) 885 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
572 { 886 {
573 ++dec->cur; 887 ++dec->cur;
574 888
575 for (;;) 889 for (;;)
576 { 890 {
577 WANT (1); 891 WANT (1);
578 892
579 if (*dec->cur == (0xe0 | 31)) 893 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
580 { 894 {
581 ++dec->cur; 895 ++dec->cur;
582 break; 896 break;
583 } 897 }
584 898
585 av_push (av, decode_sv (dec)); 899 av_push (av, decode_sv (dec));
586 } 900 }
587 } 901 }
588 else 902 else
589 { 903 {
590 int i, len = decode_uint (dec); 904 UV i, len = decode_uint (dec);
591 905
906 WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
592 av_fill (av, len - 1); 907 av_fill (av, len - 1);
593 908
594 for (i = 0; i < len; ++i) 909 for (i = 0; i < len; ++i)
595 AvARRAY (av)[i] = decode_sv (dec); 910 AvARRAY (av)[i] = decode_sv (dec);
596 } 911 }
597 912
598 DEC_DEC_DEPTH; 913 DEC_DEC_DEPTH;
599 return newRV_noinc ((SV *)av); 914 return newRV_noinc ((SV *)av);
600 915
601fail: 916fail:
602 SvREFCNT_dec (av); 917 SvREFCNT_dec_NN (av);
603 DEC_DEC_DEPTH; 918 DEC_DEC_DEPTH;
604 return &PL_sv_undef; 919 return &PL_sv_undef;
605} 920}
606 921
607static void 922static void
608decode_he (dec_t *dec, HV *hv) 923decode_he (dec_t *dec, HV *hv)
609{ 924{
610 // for speed reasons, we specialcase single-string 925 // for speed reasons, we specialcase single-string
611 // byte or utf-8 strings as keys. 926 // byte or utf-8 strings as keys, but only when !stringref
612 927
613 if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27) 928 if (ecb_expect_true (!dec->stringref))
929 if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
614 { 930 {
615 I32 len = decode_uint (dec); 931 STRLEN len = decode_uint (dec);
616 char *key = (char *)dec->cur; 932 char *key = (char *)dec->cur;
617 933
934 WANT (len);
618 dec->cur += len; 935 dec->cur += len;
619 936
620 hv_store (hv, key, len, decode_sv (dec), 0); 937 hv_store (hv, key, len, decode_sv (dec), 0);
938
939 return;
621 } 940 }
622 else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27) 941 else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
623 { 942 {
624 I32 len = decode_uint (dec); 943 STRLEN len = decode_uint (dec);
625 char *key = (char *)dec->cur; 944 char *key = (char *)dec->cur;
626 945
946 WANT (len);
627 dec->cur += len; 947 dec->cur += len;
628 948
949 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
950 if (!is_utf8_string (key, len))
951 ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
952
629 hv_store (hv, key, -len, decode_sv (dec), 0); 953 hv_store (hv, key, -len, decode_sv (dec), 0);
954
955 return;
630 } 956 }
631 else 957
632 {
633 SV *k = decode_sv (dec); 958 SV *k = decode_sv (dec);
634 SV *v = decode_sv (dec); 959 SV *v = decode_sv (dec);
635 960
961 // we leak memory if uncaught exceptions are thrown by random magical
962 // methods, and this is hopefully the only place where it can happen,
963 // so if there is a chance of an exception, take the very slow path.
964 // since catching exceptions is "undocumented/internal/forbidden" by
965 // the new p5p powers, we need to call out to a perl function :/
966 if (ecb_expect_false (SvAMAGIC (k)))
967 {
968 dSP;
969
970 ENTER; SAVETMPS;
971 PUSHMARK (SP);
972 EXTEND (SP, 3);
973 PUSHs (sv_2mortal (newRV_inc ((SV *)hv)));
974 PUSHs (sv_2mortal (k));
975 PUSHs (sv_2mortal (v));
976
977 PUTBACK;
978 call_pv ("CBOR::XS::_hv_store", G_VOID | G_DISCARD | G_EVAL);
979 SPAGAIN;
980
981 FREETMPS; LEAVE;
982
983 if (SvTRUE (ERRSV))
984 ERR_ERRSV;
985
986 return;
987 }
988
636 hv_store_ent (hv, k, v, 0); 989 hv_store_ent (hv, k, v, 0);
637 SvREFCNT_dec (k); 990 SvREFCNT_dec_NN (k);
638 } 991
992fail:
993 ;
639} 994}
640 995
641static SV * 996static SV *
642decode_hv (dec_t *dec) 997decode_hv (dec_t *dec)
643{ 998{
644 HV *hv = newHV (); 999 HV *hv = newHV ();
645 1000
646 DEC_INC_DEPTH; 1001 DEC_INC_DEPTH;
647 1002
648 if ((*dec->cur & 31) == 31) 1003 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
649 { 1004 {
650 ++dec->cur; 1005 ++dec->cur;
651 1006
652 for (;;) 1007 for (;;)
653 { 1008 {
654 WANT (1); 1009 WANT (1);
655 1010
656 if (*dec->cur == (0xe0 | 31)) 1011 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
657 { 1012 {
658 ++dec->cur; 1013 ++dec->cur;
659 break; 1014 break;
660 } 1015 }
661 1016
662 decode_he (dec, hv); 1017 decode_he (dec, hv);
663 } 1018 }
664 } 1019 }
665 else 1020 else
666 { 1021 {
667 int pairs = decode_uint (dec); 1022 UV pairs = decode_uint (dec);
1023
1024 WANT (pairs); // complexity check - need at least one byte per value, do not allow supersize hashes
668 1025
669 while (pairs--) 1026 while (pairs--)
670 decode_he (dec, hv); 1027 decode_he (dec, hv);
671 } 1028 }
672 1029
673 DEC_DEC_DEPTH; 1030 DEC_DEC_DEPTH;
674 return newRV_noinc ((SV *)hv); 1031 return newRV_noinc ((SV *)hv);
675 1032
676fail: 1033fail:
677 SvREFCNT_dec (hv); 1034 SvREFCNT_dec_NN (hv);
678 DEC_DEC_DEPTH; 1035 DEC_DEC_DEPTH;
679 return &PL_sv_undef; 1036 return &PL_sv_undef;
680} 1037}
681 1038
682static SV * 1039static SV *
683decode_str (dec_t *dec, int utf8) 1040decode_str (dec_t *dec, int utf8)
684{ 1041{
685 SV *sv = 0; 1042 SV *sv = 0;
686 1043
687 if ((*dec->cur & 31) == 31) 1044 if (ecb_expect_false ((*dec->cur & MINOR_MASK) == MINOR_INDEF))
688 { 1045 {
1046 // indefinite length strings
689 ++dec->cur; 1047 ++dec->cur;
690 1048
1049 U8 major = *dec->cur & MAJOR_MISC;
1050
691 sv = newSVpvn ("", 0); 1051 sv = newSVpvn ("", 0);
692 1052
693 // not very fast, and certainly not robust against illegal input
694 for (;;) 1053 for (;;)
695 { 1054 {
696 WANT (1); 1055 WANT (1);
697 1056
698 if (*dec->cur == (0xe0 | 31)) 1057 if ((*dec->cur - major) > LENGTH_EXT8)
1058 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
699 { 1059 {
700 ++dec->cur; 1060 ++dec->cur;
701 break; 1061 break;
702 } 1062 }
1063 else
1064 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
703 1065
704 sv_catsv (sv, decode_sv (dec)); 1066 STRLEN len = decode_uint (dec);
1067
1068 WANT (len);
1069 sv_catpvn (sv, dec->cur, len);
1070 dec->cur += len;
705 } 1071 }
706 } 1072 }
707 else 1073 else
708 { 1074 {
709 STRLEN len = decode_uint (dec); 1075 STRLEN len = decode_uint (dec);
710 1076
711 WANT (len); 1077 WANT (len);
712 sv = newSVpvn (dec->cur, len); 1078 sv = newSVpvn (dec->cur, len);
713 dec->cur += len; 1079 dec->cur += len;
1080
1081 if (ecb_expect_false (dec->stringref)
1082 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
1083 av_push (dec->stringref, SvREFCNT_inc_NN (sv));
714 } 1084 }
715 1085
716 if (utf8) 1086 if (utf8)
1087 {
1088 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
1089 if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
1090 ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
1091
717 SvUTF8_on (sv); 1092 SvUTF8_on (sv);
1093 }
718 1094
719 return sv; 1095 return sv;
720 1096
721fail: 1097fail:
722 SvREFCNT_dec (sv); 1098 SvREFCNT_dec (sv);
732 WANT (1); 1108 WANT (1);
733 1109
734 switch (tag) 1110 switch (tag)
735 { 1111 {
736 case CBOR_TAG_MAGIC: 1112 case CBOR_TAG_MAGIC:
737 return decode_sv (dec); 1113 sv = decode_sv (dec);
1114 break;
738 1115
739 case CBOR_TAG_INDIRECTION: 1116 case CBOR_TAG_INDIRECTION:
740 return newRV_noinc (decode_sv (dec)); 1117 sv = newRV_noinc (decode_sv (dec));
1118 break;
1119
1120 case CBOR_TAG_STRINGREF_NAMESPACE:
1121 {
1122 // do not use SAVETMPS/FREETMPS, as these will
1123 // erase mortalised caches, e.g. "shareable"
1124 ENTER;
1125
1126 SAVESPTR (dec->stringref);
1127 dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());
1128
1129 sv = decode_sv (dec);
1130
1131 LEAVE;
1132 }
1133 break;
1134
1135 case CBOR_TAG_STRINGREF:
1136 {
1137 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
1138 ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
1139
1140 UV idx = decode_uint (dec);
1141
1142 if (!dec->stringref || idx >= (UV)(1 + AvFILLp (dec->stringref)))
1143 ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)");
1144
1145 sv = newSVsv (AvARRAY (dec->stringref)[idx]);
1146 }
1147 break;
741 1148
742 case CBOR_TAG_VALUE_SHAREABLE: 1149 case CBOR_TAG_VALUE_SHAREABLE:
743 { 1150 {
744 if (ecb_expect_false (!dec->shareable)) 1151 if (ecb_expect_false (!dec->shareable))
745 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ()); 1152 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
746 1153
1154 if (dec->cbor.flags & F_ALLOW_CYCLES)
1155 {
747 sv = newSV (0); 1156 sv = newSV (0);
748 av_push (dec->shareable, SvREFCNT_inc_NN (sv)); 1157 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
749 1158
750 SV *osv = decode_sv (dec); 1159 SV *osv = decode_sv (dec);
751 sv_setsv (sv, osv); 1160 sv_setsv (sv, osv);
752 SvREFCNT_dec_NN (osv); 1161 SvREFCNT_dec_NN (osv);
753 } 1162 }
754 1163 else
755 return sv; 1164 {
1165 av_push (dec->shareable, &PL_sv_undef);
1166 int idx = AvFILLp (dec->shareable);
1167 sv = decode_sv (dec);
1168 av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
1169 }
1170 }
1171 break;
756 1172
757 case CBOR_TAG_VALUE_SHAREDREF: 1173 case CBOR_TAG_VALUE_SHAREDREF:
758 { 1174 {
759 if ((*dec->cur >> 5) != 0) 1175 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
760 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)"); 1176 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
761 1177
762 UV idx = decode_uint (dec); 1178 UV idx = decode_uint (dec);
763 1179
764 if (!dec->shareable || idx > AvFILLp (dec->shareable)) 1180 if (!dec->shareable || idx >= (UV)(1 + AvFILLp (dec->shareable)))
765 ERR ("corrupted CBOR data (sharedref index out of bounds)"); 1181 ERR ("corrupted CBOR data (sharedref index out of bounds)");
766 1182
767 return SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]); 1183 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
1184
1185 if (sv == &PL_sv_undef)
1186 ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
768 } 1187 }
1188 break;
769 1189
770 case CBOR_TAG_PERL_OBJECT: 1190 case CBOR_TAG_PERL_OBJECT:
771 { 1191 {
1192 if (dec->cbor.flags & F_FORBID_OBJECTS)
1193 goto filter;
1194
772 sv = decode_sv (dec); 1195 sv = decode_sv (dec);
773 1196
774 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) 1197 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
775 ERR ("corrupted CBOR data (non-array perl object)"); 1198 ERR ("corrupted CBOR data (non-array perl object)");
776 1199
786 if (!method) 1209 if (!method)
787 ERR ("cannot decode perl-object (package does not have a THAW method)"); 1210 ERR ("cannot decode perl-object (package does not have a THAW method)");
788 1211
789 dSP; 1212 dSP;
790 1213
791 ENTER; SAVETMPS; PUSHMARK (SP); 1214 ENTER; SAVETMPS;
1215 PUSHMARK (SP);
792 EXTEND (SP, len + 1); 1216 EXTEND (SP, len + 1);
793 // we re-bless the reference to get overload and other niceties right 1217 // we re-bless the reference to get overload and other niceties right
794 PUSHs (*av_fetch (av, 0, 1)); 1218 PUSHs (*av_fetch (av, 0, 1));
795 PUSHs (sv_cbor); 1219 PUSHs (sv_cbor);
796 1220
804 SPAGAIN; 1228 SPAGAIN;
805 1229
806 if (SvTRUE (ERRSV)) 1230 if (SvTRUE (ERRSV))
807 { 1231 {
808 FREETMPS; LEAVE; 1232 FREETMPS; LEAVE;
809 ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV)))); 1233 ERR_ERRSV;
810 } 1234 }
811 1235
812 SvREFCNT_dec (sv); 1236 SvREFCNT_dec_NN (sv);
813 sv = SvREFCNT_inc (POPs); 1237 sv = SvREFCNT_inc (POPs);
814 1238
815 PUTBACK; 1239 PUTBACK;
816 1240
817 FREETMPS; LEAVE; 1241 FREETMPS; LEAVE;
818
819 return sv;
820 } 1242 }
1243 break;
821 1244
822 default: 1245 default:
1246 filter:
823 { 1247 {
1248 SV *tag_sv = newSVuv (tag);
1249
824 sv = decode_sv (dec); 1250 sv = decode_sv (dec);
825 1251
1252 dSP;
1253 ENTER; SAVETMPS;
1254 PUSHMARK (SP);
1255 EXTEND (SP, 2);
1256 PUSHs (tag_sv);
1257 PUSHs (sv);
1258
1259 PUTBACK;
1260 int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL);
1261 SPAGAIN;
1262
1263 if (SvTRUE (ERRSV))
1264 {
1265 SvREFCNT_dec_NN (tag_sv);
1266 FREETMPS; LEAVE;
1267 ERR_ERRSV;
1268 }
1269
1270 if (count)
1271 {
1272 SvREFCNT_dec_NN (tag_sv);
1273 SvREFCNT_dec_NN (sv);
1274 sv = SvREFCNT_inc_NN (TOPs);
1275 SP -= count;
1276 }
1277 else
1278 {
826 AV *av = newAV (); 1279 AV *av = newAV ();
827 av_push (av, newSVuv (tag)); 1280 av_push (av, tag_sv);
828 av_push (av, sv); 1281 av_push (av, sv);
829 1282
830 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 1283 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
831 ? cbor_tagged_stash 1284 ? cbor_tagged_stash
832 : gv_stashpv ("CBOR::XS::Tagged" , 1); 1285 : gv_stashpv ("CBOR::XS::Tagged" , 1);
833
834 return sv_bless (newRV_noinc ((SV *)av), tagged_stash); 1286 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
835 } 1287 }
1288
1289 PUTBACK;
1290
1291 FREETMPS; LEAVE;
1292 }
1293 break;
836 } 1294 }
1295
1296 return sv;
837 1297
838fail: 1298fail:
839 SvREFCNT_dec (sv); 1299 SvREFCNT_dec (sv);
840 return &PL_sv_undef; 1300 return &PL_sv_undef;
841} 1301}
843static SV * 1303static SV *
844decode_sv (dec_t *dec) 1304decode_sv (dec_t *dec)
845{ 1305{
846 WANT (1); 1306 WANT (1);
847 1307
848 switch (*dec->cur >> 5) 1308 switch (*dec->cur >> MAJOR_SHIFT)
849 { 1309 {
850 case 0: // unsigned int 1310 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
851 return newSVuv (decode_uint (dec)); 1311 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
852 case 1: // negative int 1312 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
853 return newSViv (-1 - (IV)decode_uint (dec)); 1313 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
854 case 2: // octet string 1314 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
855 return decode_str (dec, 0); 1315 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
856 case 3: // utf-8 string 1316 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
857 return decode_str (dec, 1); 1317
858 case 4: // array 1318 case MAJOR_MISC >> MAJOR_SHIFT:
859 return decode_av (dec);
860 case 5: // map
861 return decode_hv (dec);
862 case 6: // tag
863 return decode_tagged (dec);
864 case 7: // misc
865 switch (*dec->cur++ & 31) 1319 switch (*dec->cur++ & MINOR_MASK)
866 { 1320 {
867 case 20: 1321 case SIMPLE_FALSE:
868#if CBOR_SLOW 1322#if CBOR_SLOW
869 types_false = get_bool ("Types::Serialiser::false"); 1323 types_false = get_bool ("Types::Serialiser::false");
870#endif 1324#endif
871 return newSVsv (types_false); 1325 return newSVsv (types_false);
872 case 21: 1326 case SIMPLE_TRUE:
873#if CBOR_SLOW 1327#if CBOR_SLOW
874 types_true = get_bool ("Types::Serialiser::true"); 1328 types_true = get_bool ("Types::Serialiser::true");
875#endif 1329#endif
876 return newSVsv (types_true); 1330 return newSVsv (types_true);
877 case 22: 1331 case SIMPLE_NULL:
878 return newSVsv (&PL_sv_undef); 1332 return newSVsv (&PL_sv_undef);
879 case 23: 1333 case SIMPLE_UNDEF:
880#if CBOR_SLOW 1334#if CBOR_SLOW
881 types_error = get_bool ("Types::Serialiser::error"); 1335 types_error = get_bool ("Types::Serialiser::error");
882#endif 1336#endif
883 return newSVsv (types_error); 1337 return newSVsv (types_error);
884 1338
885 case 25: 1339 case MISC_FLOAT16:
886 { 1340 {
887 WANT (2); 1341 WANT (2);
888 1342
889 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1343 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
890 dec->cur += 2; 1344 dec->cur += 2;
891 1345
892 return newSVnv (ecb_binary16_to_float (fp)); 1346 return newSVnv (ecb_binary16_to_float (fp));
893 } 1347 }
894 1348
895 case 26: 1349 case MISC_FLOAT32:
896 { 1350 {
897 uint32_t fp; 1351 uint32_t fp;
898 WANT (4); 1352 WANT (4);
899 memcpy (&fp, dec->cur, 4); 1353 memcpy (&fp, dec->cur, 4);
900 dec->cur += 4; 1354 dec->cur += 4;
903 fp = ecb_bswap32 (fp); 1357 fp = ecb_bswap32 (fp);
904 1358
905 return newSVnv (ecb_binary32_to_float (fp)); 1359 return newSVnv (ecb_binary32_to_float (fp));
906 } 1360 }
907 1361
908 case 27: 1362 case MISC_FLOAT64:
909 { 1363 {
910 uint64_t fp; 1364 uint64_t fp;
911 WANT (8); 1365 WANT (8);
912 memcpy (&fp, dec->cur, 8); 1366 memcpy (&fp, dec->cur, 8);
913 dec->cur += 8; 1367 dec->cur += 8;
916 fp = ecb_bswap64 (fp); 1370 fp = ecb_bswap64 (fp);
917 1371
918 return newSVnv (ecb_binary64_to_double (fp)); 1372 return newSVnv (ecb_binary64_to_double (fp));
919 } 1373 }
920 1374
921 // 0..19 unassigned 1375 // 0..19 unassigned simple
922 // 24 reserved + unassigned (reserved values are not encodable) 1376 // 24 reserved + unassigned simple (reserved values are not encodable)
1377 // 28-30 unassigned misc
1378 // 31 break code
923 default: 1379 default:
924 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1380 ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
925 } 1381 }
926 1382
927 break; 1383 break;
928 } 1384 }
929 1385
932} 1388}
933 1389
934static SV * 1390static SV *
935decode_cbor (SV *string, CBOR *cbor, char **offset_return) 1391decode_cbor (SV *string, CBOR *cbor, char **offset_return)
936{ 1392{
937 dec_t dec = { }; 1393 dec_t dec = { 0 };
938 SV *sv; 1394 SV *sv;
939 STRLEN len; 1395 STRLEN len;
940 char *data = SvPVbyte (string, len); 1396 char *data = SvPVbyte (string, len);
941 1397
942 if (len > cbor->max_size && cbor->max_size) 1398 if (len > cbor->max_size && cbor->max_size)
956 if (dec.cur != dec.end && !dec.err) 1412 if (dec.cur != dec.end && !dec.err)
957 dec.err = "garbage after CBOR object"; 1413 dec.err = "garbage after CBOR object";
958 1414
959 if (dec.err) 1415 if (dec.err)
960 { 1416 {
1417 if (dec.shareable)
1418 {
1419 // need to break cyclic links, which would all be in shareable
1420 int i;
1421 SV **svp;
1422
1423 for (i = av_len (dec.shareable) + 1; i--; )
1424 if ((svp = av_fetch (dec.shareable, i, 0)))
1425 sv_setsv (*svp, &PL_sv_undef);
1426 }
1427
961 SvREFCNT_dec (sv); 1428 SvREFCNT_dec_NN (sv);
1429
1430 if (dec.err_sv)
1431 sv_2mortal (dec.err_sv);
1432
962 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur); 1433 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur);
963 } 1434 }
964 1435
965 sv = sv_2mortal (sv); 1436 sv = sv_2mortal (sv);
966 1437
967 return sv; 1438 return sv;
968} 1439}
969 1440
1441/////////////////////////////////////////////////////////////////////////////
1442// incremental parser
1443
1444#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
1445
1446// returns 0 for notyet, 1 for success or error
1447static int
1448incr_parse (CBOR *self, SV *cborstr)
1449{
1450 STRLEN cur;
1451 SvPV (cborstr, cur);
1452
1453 while (ecb_expect_true (self->incr_need <= cur))
1454 {
1455 // table of integer count bytes
1456 static I8 incr_len[MINOR_MASK + 1] = {
1457 0, 0, 0, 0, 0, 0, 0, 0,
1458 0, 0, 0, 0, 0, 0, 0, 0,
1459 0, 0, 0, 0, 0, 0, 0, 0,
1460 1, 2, 4, 8,-1,-1,-1,-2
1461 };
1462
1463 const U8 *p = SvPVX (cborstr) + self->incr_pos;
1464 U8 m = *p & MINOR_MASK;
1465 IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1466 I8 ilen = incr_len[m];
1467
1468 self->incr_need = self->incr_pos + 1;
1469
1470 if (ecb_expect_false (ilen < 0))
1471 {
1472 if (m != MINOR_INDEF)
1473 return 1; // error
1474
1475 if (*p == (MAJOR_MISC | MINOR_INDEF))
1476 {
1477 if (count >= 0)
1478 return 1; // error
1479
1480 count = 1;
1481 }
1482 else
1483 {
1484 av_push (self->incr_count, newSViv (-1)); //TODO: nest
1485 count = -1;
1486 }
1487 }
1488 else
1489 {
1490 self->incr_need += ilen;
1491 if (ecb_expect_false (self->incr_need > cur))
1492 return 0;
1493
1494 int major = *p >> MAJOR_SHIFT;
1495
1496 switch (major)
1497 {
1498 case MAJOR_TAG >> MAJOR_SHIFT:
1499 ++count; // tags merely prefix another value
1500 break;
1501
1502 case MAJOR_BYTES >> MAJOR_SHIFT:
1503 case MAJOR_TEXT >> MAJOR_SHIFT:
1504 case MAJOR_ARRAY >> MAJOR_SHIFT:
1505 case MAJOR_MAP >> MAJOR_SHIFT:
1506 {
1507 UV len;
1508
1509 if (ecb_expect_false (ilen))
1510 {
1511 len = 0;
1512
1513 do {
1514 len = (len << 8) | *++p;
1515 } while (--ilen);
1516 }
1517 else
1518 len = m;
1519
1520 switch (major)
1521 {
1522 case MAJOR_BYTES >> MAJOR_SHIFT:
1523 case MAJOR_TEXT >> MAJOR_SHIFT:
1524 self->incr_need += len;
1525 if (ecb_expect_false (self->incr_need > cur))
1526 return 0;
1527
1528 break;
1529
1530 case MAJOR_MAP >> MAJOR_SHIFT:
1531 len <<= 1;
1532 case MAJOR_ARRAY >> MAJOR_SHIFT:
1533 if (len)
1534 {
1535 av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
1536 count = len + 1;
1537 }
1538 break;
1539 }
1540 }
1541 }
1542 }
1543
1544 self->incr_pos = self->incr_need;
1545
1546 if (count > 0)
1547 {
1548 while (!--count)
1549 {
1550 if (!AvFILLp (self->incr_count))
1551 return 1; // done
1552
1553 SvREFCNT_dec_NN (av_pop (self->incr_count));
1554 count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1555 }
1556
1557 SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
1558 }
1559 }
1560
1561 return 0;
1562}
1563
1564
970///////////////////////////////////////////////////////////////////////////// 1565/////////////////////////////////////////////////////////////////////////////
971// XS interface functions 1566// XS interface functions
972 1567
973MODULE = CBOR::XS PACKAGE = CBOR::XS 1568MODULE = CBOR::XS PACKAGE = CBOR::XS
974 1569
982 1577
983 types_true = get_bool ("Types::Serialiser::true" ); 1578 types_true = get_bool ("Types::Serialiser::true" );
984 types_false = get_bool ("Types::Serialiser::false"); 1579 types_false = get_bool ("Types::Serialiser::false");
985 types_error = get_bool ("Types::Serialiser::error"); 1580 types_error = get_bool ("Types::Serialiser::error");
986 1581
1582 default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1583
987 sv_cbor = newSVpv ("CBOR", 0); 1584 sv_cbor = newSVpv ("CBOR", 0);
988 SvREADONLY_on (sv_cbor); 1585 SvREADONLY_on (sv_cbor);
1586
1587 assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1));
989} 1588}
990 1589
991PROTOTYPES: DISABLE 1590PROTOTYPES: DISABLE
992 1591
993void CLONE (...) 1592void CLONE (...)
1012void shrink (CBOR *self, int enable = 1) 1611void shrink (CBOR *self, int enable = 1)
1013 ALIAS: 1612 ALIAS:
1014 shrink = F_SHRINK 1613 shrink = F_SHRINK
1015 allow_unknown = F_ALLOW_UNKNOWN 1614 allow_unknown = F_ALLOW_UNKNOWN
1016 allow_sharing = F_ALLOW_SHARING 1615 allow_sharing = F_ALLOW_SHARING
1017 dedup_keys = F_DEDUP_KEYS 1616 allow_cycles = F_ALLOW_CYCLES
1617 forbid_objects = F_FORBID_OBJECTS
1018 dedup_strings = F_DEDUP_STRINGS 1618 pack_strings = F_PACK_STRINGS
1619 text_keys = F_TEXT_KEYS
1620 text_strings = F_TEXT_STRINGS
1621 validate_utf8 = F_VALIDATE_UTF8
1019 PPCODE: 1622 PPCODE:
1020{ 1623{
1021 if (enable) 1624 if (enable)
1022 self->flags |= ix; 1625 self->flags |= ix;
1023 else 1626 else
1029void get_shrink (CBOR *self) 1632void get_shrink (CBOR *self)
1030 ALIAS: 1633 ALIAS:
1031 get_shrink = F_SHRINK 1634 get_shrink = F_SHRINK
1032 get_allow_unknown = F_ALLOW_UNKNOWN 1635 get_allow_unknown = F_ALLOW_UNKNOWN
1033 get_allow_sharing = F_ALLOW_SHARING 1636 get_allow_sharing = F_ALLOW_SHARING
1034 get_dedup_keys = F_DEDUP_KEYS 1637 get_allow_cycles = F_ALLOW_CYCLES
1638 get_forbid_objects = F_FORBID_OBJECTS
1035 get_dedup_strings = F_DEDUP_STRINGS 1639 get_pack_strings = F_PACK_STRINGS
1640 get_text_keys = F_TEXT_KEYS
1641 get_text_strings = F_TEXT_STRINGS
1642 get_validate_utf8 = F_VALIDATE_UTF8
1036 PPCODE: 1643 PPCODE:
1037 XPUSHs (boolSV (self->flags & ix)); 1644 XPUSHs (boolSV (self->flags & ix));
1038 1645
1039void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1646void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1040 PPCODE: 1647 PPCODE:
1053 XPUSHs (ST (0)); 1660 XPUSHs (ST (0));
1054 1661
1055int get_max_size (CBOR *self) 1662int get_max_size (CBOR *self)
1056 CODE: 1663 CODE:
1057 RETVAL = self->max_size; 1664 RETVAL = self->max_size;
1665 OUTPUT:
1666 RETVAL
1667
1668void filter (CBOR *self, SV *filter = 0)
1669 PPCODE:
1670 SvREFCNT_dec (self->filter);
1671 self->filter = filter ? newSVsv (filter) : filter;
1672 XPUSHs (ST (0));
1673
1674SV *get_filter (CBOR *self)
1675 CODE:
1676 RETVAL = self->filter ? self->filter : NEWSV (0, 0);
1058 OUTPUT: 1677 OUTPUT:
1059 RETVAL 1678 RETVAL
1060 1679
1061void encode (CBOR *self, SV *scalar) 1680void encode (CBOR *self, SV *scalar)
1062 PPCODE: 1681 PPCODE:
1077 EXTEND (SP, 2); 1696 EXTEND (SP, 2);
1078 PUSHs (sv); 1697 PUSHs (sv);
1079 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1698 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1080} 1699}
1081 1700
1701void incr_parse (CBOR *self, SV *cborstr)
1702 ALIAS:
1703 incr_parse_multiple = 1
1704 PPCODE:
1705{
1706 if (SvUTF8 (cborstr))
1707 sv_utf8_downgrade (cborstr, 0);
1708
1709 if (!self->incr_count)
1710 {
1711 self->incr_count = newAV ();
1712 self->incr_pos = 0;
1713 self->incr_need = 1;
1714
1715 av_push (self->incr_count, newSViv (1));
1716 }
1717
1718 do
1719 {
1720 if (!incr_parse (self, cborstr))
1721 {
1722 if (self->incr_need > self->max_size && self->max_size)
1723 croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
1724 (unsigned long)self->incr_need, (unsigned long)self->max_size);
1725
1726 break;
1727 }
1728
1729 SV *sv;
1730 char *offset;
1731
1732 PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
1733 XPUSHs (sv);
1734
1735 sv_chop (cborstr, offset);
1736
1737 av_clear (self->incr_count);
1738 av_push (self->incr_count, newSViv (1));
1739
1740 self->incr_pos = 0;
1741 self->incr_need = self->incr_pos + 1;
1742 }
1743 while (ix);
1744}
1745
1746void incr_reset (CBOR *self)
1747 CODE:
1748{
1749 SvREFCNT_dec (self->incr_count);
1750 self->incr_count = 0;
1751}
1752
1753void DESTROY (CBOR *self)
1754 PPCODE:
1755 cbor_free (self);
1756
1082PROTOTYPES: ENABLE 1757PROTOTYPES: ENABLE
1083 1758
1084void encode_cbor (SV *scalar) 1759void encode_cbor (SV *scalar)
1760 ALIAS:
1761 encode_cbor = 0
1762 encode_cbor_sharing = F_ALLOW_SHARING
1085 PPCODE: 1763 PPCODE:
1086{ 1764{
1087 CBOR cbor; 1765 CBOR cbor;
1088 cbor_init (&cbor); 1766 cbor_init (&cbor);
1767 cbor.flags |= ix;
1089 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1768 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1090 XPUSHs (scalar); 1769 XPUSHs (scalar);
1091} 1770}
1092 1771
1093void decode_cbor (SV *cborstr) 1772void decode_cbor (SV *cborstr)
1097 cbor_init (&cbor); 1776 cbor_init (&cbor);
1098 PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN; 1777 PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN;
1099 XPUSHs (cborstr); 1778 XPUSHs (cborstr);
1100} 1779}
1101 1780
1781#ifdef __AFL_COMPILER
1782
1783void
1784afl_init ()
1785 CODE:
1786 __AFL_INIT ();
1787
1788int
1789afl_loop (unsigned int count = 10000)
1790 CODE:
1791 RETVAL = __AFL_LOOP (count);
1792 OUTPUT:
1793 RETVAL
1794
1795#endif
1796

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines