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

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.22 by root, Wed Nov 20 15:05:46 2013 UTC vs.
Revision 1.66 by root, Sun Nov 29 21:32:01 2020 UTC

6#include <string.h> 6#include <string.h>
7#include <stdlib.h> 7#include <stdlib.h>
8#include <stdio.h> 8#include <stdio.h>
9#include <limits.h> 9#include <limits.h>
10#include <float.h> 10#include <float.h>
11#include <inttypes.h>
11 12
13#define ECB_NO_THREADS 1
12#include "ecb.h" 14#include "ecb.h"
13 15
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_FLOAT16 = 3,
111 AS_FLOAT32 = 4,
112 AS_FLOAT64 = 5,
113 // possibly future enhancements: float, integer
114};
115
59#define F_SHRINK 0x00000001UL 116#define F_SHRINK 0x00000001UL
60#define F_ALLOW_UNKNOWN 0x00000002UL 117#define F_ALLOW_UNKNOWN 0x00000002UL
61#define F_ALLOW_SHARING 0x00000004UL //TODO 118#define F_ALLOW_SHARING 0x00000004UL
119#define F_ALLOW_CYCLES 0x00000008UL
120#define F_FORBID_OBJECTS 0x00000010UL
62#define F_DEDUP_STRINGS 0x00000008UL //TODO 121#define F_PACK_STRINGS 0x00000020UL
122#define F_TEXT_KEYS 0x00000040UL
123#define F_TEXT_STRINGS 0x00000080UL
124#define F_VALIDATE_UTF8 0x00000100UL
63 125
64#define INIT_SIZE 32 // initial scalar size to be allocated 126#define INIT_SIZE 32 // initial scalar size to be allocated
65 127
66#define SB do { 128#define SB do {
67#define SE } while (0) 129#define SE } while (0)
79# define CBOR_SLOW 0 141# define CBOR_SLOW 0
80# define CBOR_STASH cbor_stash 142# define CBOR_STASH cbor_stash
81#endif 143#endif
82 144
83static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS:: 145static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS::
84static SV *types_true, *types_false, *types_error, *sv_cbor; 146static SV *types_true, *types_false, *types_error, *sv_cbor, *default_filter;
85 147
86typedef struct { 148typedef struct {
87 U32 flags; 149 U32 flags;
88 U32 max_depth; 150 U32 max_depth;
89 STRLEN max_size; 151 STRLEN max_size;
152 SV *filter;
153
154 // for the incremental parser
155 STRLEN incr_pos; // the current offset into the text
156 STRLEN incr_need; // minimum bytes needed to decode
157 AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef.
90} CBOR; 158} CBOR;
91 159
92ecb_inline void 160ecb_inline void
93cbor_init (CBOR *cbor) 161cbor_init (CBOR *cbor)
94{ 162{
95 Zero (cbor, 1, CBOR); 163 Zero (cbor, 1, CBOR);
96 cbor->max_depth = 512; 164 cbor->max_depth = 512;
165}
166
167ecb_inline void
168cbor_free (CBOR *cbor)
169{
170 SvREFCNT_dec (cbor->filter);
171 SvREFCNT_dec (cbor->incr_count);
97} 172}
98 173
99///////////////////////////////////////////////////////////////////////////// 174/////////////////////////////////////////////////////////////////////////////
100// utility functions 175// utility functions
101 176
127 202
128// minimum length of a string to be registered for stringref 203// minimum length of a string to be registered for stringref
129ecb_inline int 204ecb_inline int
130minimum_string_length (UV idx) 205minimum_string_length (UV idx)
131{ 206{
132 return idx > 23 207 return idx <= 23 ? 3
133 ? idx > 0xffU 208 : idx <= 0xffU ? 4
134 ? idx > 0xffffU 209 : idx <= 0xffffU ? 5
135 ? idx > 0xffffffffU 210 : idx <= 0xffffffffU ? 7
136 ? 7 211 : 11;
137 : 6
138 : 5
139 : 4
140 : 3;
141} 212}
142 213
143///////////////////////////////////////////////////////////////////////////// 214/////////////////////////////////////////////////////////////////////////////
144// encoder 215// encoder
145 216
158} enc_t; 229} enc_t;
159 230
160ecb_inline void 231ecb_inline void
161need (enc_t *enc, STRLEN len) 232need (enc_t *enc, STRLEN len)
162{ 233{
163 if (ecb_expect_false (enc->cur + len >= enc->end)) 234 if (ecb_expect_false ((uintptr_t)(enc->end - enc->cur) < len))
164 { 235 {
165 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv); 236 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv);
166 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1); 237 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1);
167 enc->cur = SvPVX (enc->sv) + cur; 238 enc->cur = SvPVX (enc->sv) + cur;
168 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1; 239 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;
174{ 245{
175 need (enc, 1); 246 need (enc, 1);
176 *enc->cur++ = ch; 247 *enc->cur++ = ch;
177} 248}
178 249
250// used for tags, intregers, element counts and so on
179static void 251static void
180encode_uint (enc_t *enc, int major, UV len) 252encode_uint (enc_t *enc, int major, UV len)
181{ 253{
182 need (enc, 9); 254 need (enc, 9);
183 255
184 if (len < 24) 256 if (ecb_expect_true (len < LENGTH_EXT1))
185 *enc->cur++ = major | len; 257 *enc->cur++ = major | len;
186 else if (len <= 0xff) 258 else if (ecb_expect_true (len <= 0xffU))
187 { 259 {
188 *enc->cur++ = major | 24; 260 *enc->cur++ = major | LENGTH_EXT1;
189 *enc->cur++ = len; 261 *enc->cur++ = len;
190 } 262 }
191 else if (len <= 0xffff) 263 else if (len <= 0xffffU)
192 { 264 {
193 *enc->cur++ = major | 25; 265 *enc->cur++ = major | LENGTH_EXT2;
194 *enc->cur++ = len >> 8; 266 *enc->cur++ = len >> 8;
195 *enc->cur++ = len; 267 *enc->cur++ = len;
196 } 268 }
197 else if (len <= 0xffffffff) 269 else if (len <= 0xffffffffU)
198 { 270 {
199 *enc->cur++ = major | 26; 271 *enc->cur++ = major | LENGTH_EXT4;
200 *enc->cur++ = len >> 24; 272 *enc->cur++ = len >> 24;
201 *enc->cur++ = len >> 16; 273 *enc->cur++ = len >> 16;
202 *enc->cur++ = len >> 8; 274 *enc->cur++ = len >> 8;
203 *enc->cur++ = len; 275 *enc->cur++ = len;
204 } 276 }
205 else 277 else
206 { 278 {
207 *enc->cur++ = major | 27; 279 *enc->cur++ = major | LENGTH_EXT8;
208 *enc->cur++ = len >> 56; 280 *enc->cur++ = len >> 56;
209 *enc->cur++ = len >> 48; 281 *enc->cur++ = len >> 48;
210 *enc->cur++ = len >> 40; 282 *enc->cur++ = len >> 40;
211 *enc->cur++ = len >> 32; 283 *enc->cur++ = len >> 32;
212 *enc->cur++ = len >> 24; 284 *enc->cur++ = len >> 24;
214 *enc->cur++ = len >> 8; 286 *enc->cur++ = len >> 8;
215 *enc->cur++ = len; 287 *enc->cur++ = len;
216 } 288 }
217} 289}
218 290
291// encodes a perl value into a CBOR integer
292ecb_inline
293encode_int (enc_t *enc, SV *sv)
294{
295 if (SvIsUV (sv))
296 encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
297 else if (SvIVX (sv) >= 0)
298 encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
299 else
300 encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
301}
302
219ecb_inline void 303ecb_inline void
220encode_tag (enc_t *enc, UV tag) 304encode_tag (enc_t *enc, UV tag)
221{ 305{
222 encode_uint (enc, 0xc0, tag); 306 encode_uint (enc, MAJOR_TAG, tag);
223} 307}
224 308
225static void 309// exceptional (hopefully) slow path for byte strings that need to be utf8-encoded
310ecb_noinline static void
226encode_str (enc_t *enc, int utf8, char *str, STRLEN len) 311encode_str_utf8 (enc_t *enc, int utf8, char *str, STRLEN len)
227{ 312{
313 STRLEN ulen = len;
314 U8 *p, *pend = (U8 *)str + len;
315
316 for (p = (U8 *)str; p < pend; ++p)
317 ulen += *p >> 7; // count set high bits
318
319 encode_uint (enc, MAJOR_TEXT, ulen);
320
321 need (enc, ulen);
322 for (p = (U8 *)str; p < pend; ++p)
323 if (*p < 0x80)
324 *enc->cur++ = *p;
325 else
326 {
327 *enc->cur++ = 0xc0 + (*p >> 6);
328 *enc->cur++ = 0x80 + (*p & 63);
329 }
330}
331
332ecb_inline void
333encode_str (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
334{
335 if (ecb_expect_false (upgrade_utf8))
336 if (!utf8)
337 {
338 encode_str_utf8 (enc, utf8, str, len);
339 return;
340 }
341
342 encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
343 need (enc, len);
344 memcpy (enc->cur, str, len);
345 enc->cur += len;
346}
347
348ecb_inline void
349encode_strref (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
350{
228 if (ecb_expect_false (enc->cbor.flags & F_DEDUP_STRINGS)) 351 if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
229 { 352 {
230 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1); 353 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
231 354
232 if (SvOK (*svp)) 355 if (SvOK (*svp))
233 { 356 {
234 // already registered, use stringref 357 // already registered, use stringref
235 encode_tag (enc, CBOR_TAG_STRINGREF); 358 encode_tag (enc, CBOR_TAG_STRINGREF);
236 encode_uint (enc, 0x00, SvUV (*svp)); 359 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
237 return; 360 return;
238 } 361 }
239 else if (len >= minimum_string_length (enc->stringref_idx)) 362 else if (len >= minimum_string_length (enc->stringref_idx))
240 { 363 {
241 // register only 364 // register only
242 sv_setuv (*svp, enc->stringref_idx); 365 sv_setuv (*svp, enc->stringref_idx);
243 ++enc->stringref_idx; 366 ++enc->stringref_idx;
244 } 367 }
245 } 368 }
246 369
247 encode_uint (enc, utf8 ? 0x60 : 0x40, len); 370 encode_str (enc, upgrade_utf8, utf8, str, len);
371}
372
373ecb_inline void
374encode_float16 (enc_t *enc, NV nv)
375{
248 need (enc, len); 376 need (enc, 1+2);
377
378 *enc->cur++ = MAJOR_MISC | MISC_FLOAT16;
379
380 uint16_t fp = ecb_float_to_binary16 (nv);
381
382 if (!ecb_big_endian ())
383 fp = ecb_bswap16 (fp);
384
249 memcpy (enc->cur, str, len); 385 memcpy (enc->cur, &fp, 2);
250 enc->cur += len; 386 enc->cur += 2;
387}
388
389ecb_inline void
390encode_float32 (enc_t *enc, NV nv)
391{
392 need (enc, 1+4);
393
394 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
395
396 uint32_t fp = ecb_float_to_binary32 (nv);
397
398 if (!ecb_big_endian ())
399 fp = ecb_bswap32 (fp);
400
401 memcpy (enc->cur, &fp, 4);
402 enc->cur += 4;
403}
404
405ecb_inline void
406encode_float64 (enc_t *enc, NV nv)
407{
408 need (enc, 1+8);
409
410 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
411
412 uint64_t fp = ecb_double_to_binary64 (nv);
413
414 if (!ecb_big_endian ())
415 fp = ecb_bswap64 (fp);
416
417 memcpy (enc->cur, &fp, 8);
418 enc->cur += 8;
419}
420
421ecb_inline void
422encode_forced (enc_t *enc, UV type, SV *sv)
423{
424 switch (type)
425 {
426 case AS_CBOR:
427 {
428 STRLEN len;
429 char *str = SvPVbyte (sv, len);
430
431 need (enc, len);
432 memcpy (enc->cur, str, len);
433 enc->cur += len;
434 }
435 break;
436
437 case AS_BYTES:
438 {
439 STRLEN len;
440 char *str = SvPVbyte (sv, len);
441 encode_strref (enc, 0, 0, str, len);
442 }
443 break;
444
445 case AS_TEXT:
446 {
447 STRLEN len;
448 char *str = SvPVutf8 (sv, len);
449 encode_strref (enc, 1, 1, str, len);
450 }
451 break;
452
453 case AS_FLOAT16: encode_float16 (enc, SvNV (sv)); break;
454 case AS_FLOAT32: encode_float32 (enc, SvNV (sv)); break;
455 case AS_FLOAT64: encode_float64 (enc, SvNV (sv)); break;
456
457 default:
458 croak ("encountered malformed CBOR::XS::Tagged object");
459 }
251} 460}
252 461
253static void encode_sv (enc_t *enc, SV *sv); 462static void encode_sv (enc_t *enc, SV *sv);
254 463
255static void 464static void
260 if (enc->depth >= enc->cbor.max_depth) 469 if (enc->depth >= enc->cbor.max_depth)
261 croak (ERR_NESTING_EXCEEDED); 470 croak (ERR_NESTING_EXCEEDED);
262 471
263 ++enc->depth; 472 ++enc->depth;
264 473
265 encode_uint (enc, 0x80, len + 1); 474 encode_uint (enc, MAJOR_ARRAY, len + 1);
266 475
476 if (ecb_expect_false (SvMAGICAL (av)))
267 for (i = 0; i <= len; ++i) 477 for (i = 0; i <= len; ++i)
268 { 478 {
269 SV **svp = av_fetch (av, i, 0); 479 SV **svp = av_fetch (av, i, 0);
270 encode_sv (enc, svp ? *svp : &PL_sv_undef); 480 encode_sv (enc, svp ? *svp : &PL_sv_undef);
271 } 481 }
482 else
483 for (i = 0; i <= len; ++i)
484 {
485 SV *sv = AvARRAY (av)[i];
486 encode_sv (enc, sv ? sv : &PL_sv_undef);
487 }
272 488
273 --enc->depth; 489 --enc->depth;
274}
275
276ecb_inline void
277encode_he (enc_t *enc, HE *he)
278{
279} 490}
280 491
281static void 492static void
282encode_hv (enc_t *enc, HV *hv) 493encode_hv (enc_t *enc, HV *hv)
283{ 494{
289 ++enc->depth; 500 ++enc->depth;
290 501
291 int pairs = hv_iterinit (hv); 502 int pairs = hv_iterinit (hv);
292 int mg = SvMAGICAL (hv); 503 int mg = SvMAGICAL (hv);
293 504
294 if (mg) 505 if (ecb_expect_false (mg))
295 encode_ch (enc, 0xa0 | 31); 506 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
296 else 507 else
297 encode_uint (enc, 0xa0, pairs); 508 encode_uint (enc, MAJOR_MAP, pairs);
298 509
299 while ((he = hv_iternext (hv))) 510 while ((he = hv_iternext (hv)))
300 { 511 {
301 if (HeKLEN (he) == HEf_SVKEY) 512 if (HeKLEN (he) == HEf_SVKEY)
302 encode_sv (enc, HeSVKEY (he)); 513 encode_sv (enc, HeSVKEY (he));
303 else 514 else
304 encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); 515 encode_strref (enc, enc->cbor.flags & (F_TEXT_KEYS | F_TEXT_STRINGS), HeKUTF8 (he), HeKEY (he), HeKLEN (he));
305 516
306 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); 517 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
307 } 518 }
308 519
309 if (mg) 520 if (ecb_expect_false (mg))
310 encode_ch (enc, 0xe0 | 31); 521 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
311 522
312 --enc->depth; 523 --enc->depth;
313} 524}
314 525
315// encode objects, arrays and special \0=false and \1=true values. 526// encode objects, arrays and special \0=false and \1=true values.
316static void 527static void
317encode_rv (enc_t *enc, SV *sv) 528encode_rv (enc_t *enc, SV *sv)
318{ 529{
319 SvGETMAGIC (sv); 530 SvGETMAGIC (sv);
320
321 if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)
322 && ecb_expect_false (SvREFCNT (sv) > 1))
323 {
324 if (!enc->shareable)
325 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
326
327 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
328
329 if (SvOK (*svp))
330 {
331 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
332 encode_uint (enc, 0x00, SvUV (*svp));
333 return;
334 }
335 else
336 {
337 sv_setuv (*svp, enc->shareable_idx);
338 ++enc->shareable_idx;
339 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
340 }
341 }
342 531
343 svtype svt = SvTYPE (sv); 532 svtype svt = SvTYPE (sv);
344 533
345 if (ecb_expect_false (SvOBJECT (sv))) 534 if (ecb_expect_false (SvOBJECT (sv)))
346 { 535 {
353 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 542 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
354 ? cbor_tagged_stash 543 ? cbor_tagged_stash
355 : gv_stashpv ("CBOR::XS::Tagged" , 1); 544 : gv_stashpv ("CBOR::XS::Tagged" , 1);
356 545
357 HV *stash = SvSTASH (sv); 546 HV *stash = SvSTASH (sv);
358 GV *method;
359 547
360 if (stash == boolean_stash) 548 if (stash == boolean_stash)
361 encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20); 549 {
550 encode_ch (enc, SvIV (sv) ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE);
551 return;
552 }
362 else if (stash == error_stash) 553 else if (stash == error_stash)
363 encode_ch (enc, 0xe0 | 23); 554 {
555 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
556 return;
557 }
364 else if (stash == tagged_stash) 558 else if (stash == tagged_stash)
365 { 559 {
366 if (svt != SVt_PVAV) 560 if (svt != SVt_PVAV)
367 croak ("encountered CBOR::XS::Tagged object that isn't an array"); 561 croak ("encountered CBOR::XS::Tagged object that isn't an array");
368 562
563 switch (av_len ((AV *)sv))
564 {
565 case 2-1:
566 // actually a tagged value
369 encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); 567 encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
370 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); 568 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1));
569 break;
570
571 case 3-1:
572 // a forced type [value, type, undef]
573 encode_forced (enc, SvUV (*av_fetch ((AV *)sv, 1, 1)), *av_fetch ((AV *)sv, 0, 1));
574 break;
575
576 default:
577 croak ("encountered malformed CBOR::XS::Tagged object");
578 }
579
580 return;
581 }
582 }
583
584 if (ecb_expect_false (SvREFCNT (sv) > 1)
585 && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
586 {
587 if (ecb_expect_false (!enc->shareable))
588 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
589
590 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
591
592 if (SvOK (*svp))
371 } 593 {
594 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
595 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
596 return;
597 }
598 else
599 {
600 sv_setuv (*svp, enc->shareable_idx);
601 ++enc->shareable_idx;
602 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
603 }
604 }
605
606 if (ecb_expect_false (SvOBJECT (sv)))
607 {
608 HV *stash = SvSTASH (sv);
609 GV *method;
610
611 if (enc->cbor.flags & F_FORBID_OBJECTS)
612 croak ("encountered object '%s', but forbid_objects is enabled",
613 SvPV_nolen (sv_2mortal (newRV_inc (sv))));
372 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) 614 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
373 { 615 {
374 dSP; 616 dSP;
375 617
376 ENTER; SAVETMPS; PUSHMARK (SP); 618 ENTER; SAVETMPS;
619 PUSHMARK (SP);
377 // we re-bless the reference to get overload and other niceties right 620 // we re-bless the reference to get overload and other niceties right
378 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 621 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
379 622
380 PUTBACK; 623 PUTBACK;
381 // G_SCALAR ensures that return value is 1 624 // G_SCALAR ensures that return value is 1
394 } 637 }
395 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0) 638 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
396 { 639 {
397 dSP; 640 dSP;
398 641
399 ENTER; SAVETMPS; PUSHMARK (SP); 642 ENTER; SAVETMPS;
643 PUSHMARK (SP);
400 EXTEND (SP, 2); 644 EXTEND (SP, 2);
401 // we re-bless the reference to get overload and other niceties right 645 // we re-bless the reference to get overload and other niceties right
402 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 646 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
403 PUSHs (sv_cbor); 647 PUSHs (sv_cbor);
404 648
409 // catch this surprisingly common error 653 // catch this surprisingly common error
410 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) 654 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
411 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash)); 655 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
412 656
413 encode_tag (enc, CBOR_TAG_PERL_OBJECT); 657 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
414 encode_uint (enc, 0x80, count + 1); 658 encode_uint (enc, MAJOR_ARRAY, count + 1);
415 encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); 659 encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
416 660
417 while (count) 661 {
662 int i;
663
664 for (i = 0; i < count; ++i)
418 encode_sv (enc, SP[1 - count--]); 665 encode_sv (enc, SP[i + 1 - count]);
666
667 SP -= count;
668 }
419 669
420 PUTBACK; 670 PUTBACK;
421 671
422 FREETMPS; LEAVE; 672 FREETMPS; LEAVE;
423 } 673 }
441{ 691{
442 double nv = SvNVX (sv); 692 double nv = SvNVX (sv);
443 693
444 need (enc, 9); 694 need (enc, 9);
445 695
446 if (ecb_expect_false (nv == (U32)nv)) 696 if (ecb_expect_false (nv == (NV)(U32)nv))
447 encode_uint (enc, 0x00, (U32)nv); 697 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
448 //TODO: maybe I32? 698 //TODO: maybe I32?
449 else if (ecb_expect_false (nv == (float)nv)) 699 else if (ecb_expect_false (nv == (float)nv))
450 { 700 encode_float32 (enc, nv);
451 uint32_t fp = ecb_float_to_binary32 (nv);
452
453 *enc->cur++ = 0xe0 | 26;
454
455 if (!ecb_big_endian ())
456 fp = ecb_bswap32 (fp);
457
458 memcpy (enc->cur, &fp, 4);
459 enc->cur += 4;
460 }
461 else 701 else
462 { 702 encode_float64 (enc, nv);
463 uint64_t fp = ecb_double_to_binary64 (nv);
464
465 *enc->cur++ = 0xe0 | 27;
466
467 if (!ecb_big_endian ())
468 fp = ecb_bswap64 (fp);
469
470 memcpy (enc->cur, &fp, 8);
471 enc->cur += 8;
472 }
473} 703}
474 704
475static void 705static void
476encode_sv (enc_t *enc, SV *sv) 706encode_sv (enc_t *enc, SV *sv)
477{ 707{
479 709
480 if (SvPOKp (sv)) 710 if (SvPOKp (sv))
481 { 711 {
482 STRLEN len; 712 STRLEN len;
483 char *str = SvPV (sv, len); 713 char *str = SvPV (sv, len);
484 encode_str (enc, SvUTF8 (sv), str, len); 714 encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len);
485 } 715 }
486 else if (SvNOKp (sv)) 716 else if (SvNOKp (sv))
487 encode_nv (enc, sv); 717 encode_nv (enc, sv);
488 else if (SvIOKp (sv)) 718 else if (SvIOKp (sv))
489 { 719 encode_int (e,v sv);
490 if (SvIsUV (sv))
491 encode_uint (enc, 0x00, SvUVX (sv));
492 else if (SvIVX (sv) >= 0)
493 encode_uint (enc, 0x00, SvIVX (sv));
494 else
495 encode_uint (enc, 0x20, -(SvIVX (sv) + 1));
496 }
497 else if (SvROK (sv)) 720 else if (SvROK (sv))
498 encode_rv (enc, SvRV (sv)); 721 encode_rv (enc, SvRV (sv));
499 else if (!SvOK (sv)) 722 else if (!SvOK (sv))
500 encode_ch (enc, 0xe0 | 22); 723 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
501 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 724 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
502 encode_ch (enc, 0xe0 | 23); 725 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
503 else 726 else
504 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data", 727 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
505 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 728 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
506} 729}
507 730
508static SV * 731static SV *
509encode_cbor (SV *scalar, CBOR *cbor) 732encode_cbor (SV *scalar, CBOR *cbor)
510{ 733{
511 enc_t enc = { }; 734 enc_t enc = { 0 };
512 735
513 enc.cbor = *cbor; 736 enc.cbor = *cbor;
514 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 737 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
515 enc.cur = SvPVX (enc.sv); 738 enc.cur = SvPVX (enc.sv);
516 enc.end = SvEND (enc.sv); 739 enc.end = SvEND (enc.sv);
517 740
518 SvPOK_only (enc.sv); 741 SvPOK_only (enc.sv);
519 742
520 if (cbor->flags & F_DEDUP_STRINGS) 743 if (cbor->flags & F_PACK_STRINGS)
521 { 744 {
522 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE); 745 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
523 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ()); 746 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
524 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ()); 747 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
525 } 748 }
547 CBOR cbor; 770 CBOR cbor;
548 U32 depth; // recursion depth 771 U32 depth; // recursion depth
549 U32 maxdepth; // recursion depth limit 772 U32 maxdepth; // recursion depth limit
550 AV *shareable; 773 AV *shareable;
551 AV *stringref; 774 AV *stringref;
775 SV *decode_tagged;
776 SV *err_sv; // optional sv for error, needs to be freed
552} dec_t; 777} dec_t;
553 778
554#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE 779// set dec->err to ERRSV
780ecb_cold static void
781err_errsv (dec_t *dec)
782{
783 if (!dec->err)
784 {
785 dec->err_sv = newSVsv (ERRSV);
555 786
556#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data") 787 // chop off the trailing \n
788 SvCUR_set (dec->err_sv, SvCUR (dec->err_sv) - 1);
789 *SvEND (dec->err_sv) = 0;
557 790
791 dec->err = SvPVutf8_nolen (dec->err_sv);
792 }
793}
794
795// the following functions are used to reduce code size and help the compiler to optimise
796ecb_cold static void
797err_set (dec_t *dec, const char *reason)
798{
799 if (!dec->err)
800 dec->err = reason;
801}
802
803ecb_cold static void
804err_unexpected_end (dec_t *dec)
805{
806 err_set (dec, "unexpected end of CBOR data");
807}
808
809#define ERR_DO(do) SB do; goto fail; SE
810#define ERR(reason) ERR_DO (err_set (dec, reason))
811#define ERR_ERRSV ERR_DO (err_errsv (dec))
812
813#define WANT(len) if (ecb_expect_false ((uintptr_t)(dec->end - dec->cur) < (STRLEN)len)) ERR_DO (err_unexpected_end (dec))
814
558#define DEC_INC_DEPTH if (++dec->depth > dec->cbor.max_depth) ERR (ERR_NESTING_EXCEEDED) 815#define DEC_INC_DEPTH if (ecb_expect_false (++dec->depth > dec->cbor.max_depth)) ERR (ERR_NESTING_EXCEEDED)
559#define DEC_DEC_DEPTH --dec->depth 816#define DEC_DEC_DEPTH --dec->depth
560 817
561static UV 818static UV
562decode_uint (dec_t *dec) 819decode_uint (dec_t *dec)
563{ 820{
564 switch (*dec->cur & 31) 821 U8 m = *dec->cur & MINOR_MASK;
565 { 822 ++dec->cur;
566 case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
567 case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
568 case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
569 return *dec->cur++ & 31;
570 823
571 case 24: 824 if (ecb_expect_true (m < LENGTH_EXT1))
825 return m;
826 else if (ecb_expect_true (m == LENGTH_EXT1))
827 {
572 WANT (2); 828 WANT (1);
573 dec->cur += 2; 829 dec->cur += 1;
574 return dec->cur[-1]; 830 return dec->cur[-1];
575 831 }
576 case 25: 832 else if (ecb_expect_true (m == LENGTH_EXT2))
833 {
577 WANT (3); 834 WANT (2);
578 dec->cur += 3; 835 dec->cur += 2;
579 return (((UV)dec->cur[-2]) << 8) 836 return (((UV)dec->cur[-2]) << 8)
580 | ((UV)dec->cur[-1]); 837 | ((UV)dec->cur[-1]);
581 838 }
582 case 26: 839 else if (ecb_expect_true (m == LENGTH_EXT4))
840 {
583 WANT (5); 841 WANT (4);
584 dec->cur += 5; 842 dec->cur += 4;
585 return (((UV)dec->cur[-4]) << 24) 843 return (((UV)dec->cur[-4]) << 24)
586 | (((UV)dec->cur[-3]) << 16) 844 | (((UV)dec->cur[-3]) << 16)
587 | (((UV)dec->cur[-2]) << 8) 845 | (((UV)dec->cur[-2]) << 8)
588 | ((UV)dec->cur[-1]); 846 | ((UV)dec->cur[-1]);
589 847 }
590 case 27: 848 else if (ecb_expect_true (m == LENGTH_EXT8))
849 {
591 WANT (9); 850 WANT (8);
592 dec->cur += 9; 851 dec->cur += 8;
852
853 return
854#if UVSIZE < 8
855 0
856#else
593 return (((UV)dec->cur[-8]) << 56) 857 (((UV)dec->cur[-8]) << 56)
594 | (((UV)dec->cur[-7]) << 48) 858 | (((UV)dec->cur[-7]) << 48)
595 | (((UV)dec->cur[-6]) << 40) 859 | (((UV)dec->cur[-6]) << 40)
596 | (((UV)dec->cur[-5]) << 32) 860 | (((UV)dec->cur[-5]) << 32)
861#endif
597 | (((UV)dec->cur[-4]) << 24) 862 | (((UV)dec->cur[-4]) << 24)
598 | (((UV)dec->cur[-3]) << 16) 863 | (((UV)dec->cur[-3]) << 16)
599 | (((UV)dec->cur[-2]) << 8) 864 | (((UV)dec->cur[-2]) << 8)
600 | ((UV)dec->cur[-1]); 865 | ((UV)dec->cur[-1]);
601 866 }
602 default: 867 else
603 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 868 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
604 }
605 869
606fail: 870fail:
607 return 0; 871 return 0;
608} 872}
609 873
614{ 878{
615 AV *av = newAV (); 879 AV *av = newAV ();
616 880
617 DEC_INC_DEPTH; 881 DEC_INC_DEPTH;
618 882
619 if ((*dec->cur & 31) == 31) 883 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
620 { 884 {
621 ++dec->cur; 885 ++dec->cur;
622 886
623 for (;;) 887 for (;;)
624 { 888 {
625 WANT (1); 889 WANT (1);
626 890
627 if (*dec->cur == (0xe0 | 31)) 891 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
628 { 892 {
629 ++dec->cur; 893 ++dec->cur;
630 break; 894 break;
631 } 895 }
632 896
633 av_push (av, decode_sv (dec)); 897 av_push (av, decode_sv (dec));
634 } 898 }
635 } 899 }
636 else 900 else
637 { 901 {
638 int i, len = decode_uint (dec); 902 UV i, len = decode_uint (dec);
639 903
904 WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
640 av_fill (av, len - 1); 905 av_fill (av, len - 1);
641 906
642 for (i = 0; i < len; ++i) 907 for (i = 0; i < len; ++i)
643 AvARRAY (av)[i] = decode_sv (dec); 908 AvARRAY (av)[i] = decode_sv (dec);
644 } 909 }
645 910
646 DEC_DEC_DEPTH; 911 DEC_DEC_DEPTH;
647 return newRV_noinc ((SV *)av); 912 return newRV_noinc ((SV *)av);
648 913
649fail: 914fail:
650 SvREFCNT_dec (av); 915 SvREFCNT_dec_NN (av);
651 DEC_DEC_DEPTH; 916 DEC_DEC_DEPTH;
652 return &PL_sv_undef; 917 return &PL_sv_undef;
653} 918}
654 919
655static void 920static void
656decode_he (dec_t *dec, HV *hv) 921decode_he (dec_t *dec, HV *hv)
657{ 922{
658 // for speed reasons, we specialcase single-string 923 // for speed reasons, we specialcase single-string
659 // byte or utf-8 strings as keys, but only when !stringref 924 // byte or utf-8 strings as keys, but only when !stringref
660 925
661 if (expect_true (!dec->stringref)) 926 if (ecb_expect_true (!dec->stringref))
662 if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27) 927 if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
663 { 928 {
664 I32 len = decode_uint (dec); 929 STRLEN len = decode_uint (dec);
665 char *key = (char *)dec->cur; 930 char *key = (char *)dec->cur;
666 931
932 WANT (len);
667 dec->cur += len; 933 dec->cur += len;
668 934
669 if (ecb_expect_false (dec->stringref))
670 av_push (dec->stringref, newSVpvn (key, len));
671
672 hv_store (hv, key, len, decode_sv (dec), 0); 935 hv_store (hv, key, len, decode_sv (dec), 0);
673 936
674 return; 937 return;
675 } 938 }
676 else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27) 939 else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
677 { 940 {
678 I32 len = decode_uint (dec); 941 STRLEN len = decode_uint (dec);
679 char *key = (char *)dec->cur; 942 char *key = (char *)dec->cur;
680 943
944 WANT (len);
681 dec->cur += len; 945 dec->cur += len;
682 946
683 if (ecb_expect_false (dec->stringref)) 947 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
684 av_push (dec->stringref, newSVpvn_utf8 (key, len, 1)); 948 if (!is_utf8_string (key, len))
949 ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
685 950
686 hv_store (hv, key, -len, decode_sv (dec), 0); 951 hv_store (hv, key, -len, decode_sv (dec), 0);
687 952
688 return; 953 return;
689 } 954 }
690 955
691 SV *k = decode_sv (dec); 956 SV *k = decode_sv (dec);
692 SV *v = decode_sv (dec); 957 SV *v = decode_sv (dec);
693 958
959 // we leak memory if uncaught exceptions are thrown by random magical
960 // methods, and this is hopefully the only place where it can happen,
961 // so if there is a chance of an exception, take the very slow path.
962 // since catching exceptions is "undocumented/internal/forbidden" by
963 // the new p5p powers, we need to call out to a perl function :/
964 if (ecb_expect_false (SvAMAGIC (k)))
965 {
966 dSP;
967
968 ENTER; SAVETMPS;
969 PUSHMARK (SP);
970 EXTEND (SP, 3);
971 PUSHs (sv_2mortal (newRV_inc ((SV *)hv)));
972 PUSHs (sv_2mortal (k));
973 PUSHs (sv_2mortal (v));
974
975 PUTBACK;
976 call_pv ("CBOR::XS::_hv_store", G_VOID | G_DISCARD | G_EVAL);
977 SPAGAIN;
978
979 FREETMPS; LEAVE;
980
981 if (SvTRUE (ERRSV))
982 ERR_ERRSV;
983
984 return;
985 }
986
694 hv_store_ent (hv, k, v, 0); 987 hv_store_ent (hv, k, v, 0);
695 SvREFCNT_dec (k); 988 SvREFCNT_dec_NN (k);
989
990fail:
991 ;
696} 992}
697 993
698static SV * 994static SV *
699decode_hv (dec_t *dec) 995decode_hv (dec_t *dec)
700{ 996{
701 HV *hv = newHV (); 997 HV *hv = newHV ();
702 998
703 DEC_INC_DEPTH; 999 DEC_INC_DEPTH;
704 1000
705 if ((*dec->cur & 31) == 31) 1001 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
706 { 1002 {
707 ++dec->cur; 1003 ++dec->cur;
708 1004
709 for (;;) 1005 for (;;)
710 { 1006 {
711 WANT (1); 1007 WANT (1);
712 1008
713 if (*dec->cur == (0xe0 | 31)) 1009 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
714 { 1010 {
715 ++dec->cur; 1011 ++dec->cur;
716 break; 1012 break;
717 } 1013 }
718 1014
719 decode_he (dec, hv); 1015 decode_he (dec, hv);
720 } 1016 }
721 } 1017 }
722 else 1018 else
723 { 1019 {
724 int pairs = decode_uint (dec); 1020 UV pairs = decode_uint (dec);
1021
1022 WANT (pairs); // complexity check - need at least one byte per value, do not allow supersize hashes
725 1023
726 while (pairs--) 1024 while (pairs--)
727 decode_he (dec, hv); 1025 decode_he (dec, hv);
728 } 1026 }
729 1027
730 DEC_DEC_DEPTH; 1028 DEC_DEC_DEPTH;
731 return newRV_noinc ((SV *)hv); 1029 return newRV_noinc ((SV *)hv);
732 1030
733fail: 1031fail:
734 SvREFCNT_dec (hv); 1032 SvREFCNT_dec_NN (hv);
735 DEC_DEC_DEPTH; 1033 DEC_DEC_DEPTH;
736 return &PL_sv_undef; 1034 return &PL_sv_undef;
737} 1035}
738 1036
739static SV * 1037static SV *
740decode_str (dec_t *dec, int utf8) 1038decode_str (dec_t *dec, int utf8)
741{ 1039{
742 SV *sv = 0; 1040 SV *sv = 0;
743 1041
744 if ((*dec->cur & 31) == 31) 1042 if (ecb_expect_false ((*dec->cur & MINOR_MASK) == MINOR_INDEF))
745 { 1043 {
1044 // indefinite length strings
746 ++dec->cur; 1045 ++dec->cur;
747 1046
1047 U8 major = *dec->cur & MAJOR_MISC;
1048
748 sv = newSVpvn ("", 0); 1049 sv = newSVpvn ("", 0);
749 1050
750 // not very fast, and certainly not robust against illegal input
751 for (;;) 1051 for (;;)
752 { 1052 {
753 WANT (1); 1053 WANT (1);
754 1054
755 if (*dec->cur == (0xe0 | 31)) 1055 if ((*dec->cur - major) > LENGTH_EXT8)
1056 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
756 { 1057 {
757 ++dec->cur; 1058 ++dec->cur;
758 break; 1059 break;
759 } 1060 }
1061 else
1062 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
760 1063
761 sv_catsv (sv, decode_sv (dec)); 1064 STRLEN len = decode_uint (dec);
1065
1066 WANT (len);
1067 sv_catpvn (sv, dec->cur, len);
1068 dec->cur += len;
762 } 1069 }
763 } 1070 }
764 else 1071 else
765 { 1072 {
766 STRLEN len = decode_uint (dec); 1073 STRLEN len = decode_uint (dec);
767 1074
768 WANT (len); 1075 WANT (len);
769 sv = newSVpvn (dec->cur, len); 1076 sv = newSVpvn (dec->cur, len);
770 dec->cur += len; 1077 dec->cur += len;
1078
1079 if (ecb_expect_false (dec->stringref)
1080 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
1081 av_push (dec->stringref, SvREFCNT_inc_NN (sv));
771 } 1082 }
772 1083
773 if (utf8) 1084 if (utf8)
1085 {
1086 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
1087 if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
1088 ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
1089
774 SvUTF8_on (sv); 1090 SvUTF8_on (sv);
775 1091 }
776 if (ecb_expect_false (dec->stringref)
777 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
778 av_push (dec->stringref, SvREFCNT_inc_NN (sv));
779 1092
780 return sv; 1093 return sv;
781 1094
782fail: 1095fail:
783 SvREFCNT_dec (sv); 1096 SvREFCNT_dec (sv);
802 sv = newRV_noinc (decode_sv (dec)); 1115 sv = newRV_noinc (decode_sv (dec));
803 break; 1116 break;
804 1117
805 case CBOR_TAG_STRINGREF_NAMESPACE: 1118 case CBOR_TAG_STRINGREF_NAMESPACE:
806 { 1119 {
1120 // do not use SAVETMPS/FREETMPS, as these will
1121 // erase mortalised caches, e.g. "shareable"
807 ENTER; SAVETMPS; 1122 ENTER;
808 1123
809 SAVESPTR (dec->stringref); 1124 SAVESPTR (dec->stringref);
810 dec->stringref = (AV *)sv_2mortal ((SV *)newAV ()); 1125 dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());
811 1126
812 sv = decode_sv (dec); 1127 sv = decode_sv (dec);
813 1128
814 FREETMPS; LEAVE; 1129 LEAVE;
815 } 1130 }
816 break; 1131 break;
817 1132
818 case CBOR_TAG_STRINGREF: 1133 case CBOR_TAG_STRINGREF:
819 { 1134 {
820 if ((*dec->cur >> 5) != 0) 1135 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
821 ERR ("corrupted CBOR data (stringref index not an unsigned integer)"); 1136 ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
822 1137
823 UV idx = decode_uint (dec); 1138 UV idx = decode_uint (dec);
824 1139
825 if (!dec->stringref || (int)idx > AvFILLp (dec->stringref)) 1140 if (!dec->stringref || idx >= (UV)(1 + AvFILLp (dec->stringref)))
826 ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)"); 1141 ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)");
827 1142
828 sv = newSVsv (AvARRAY (dec->stringref)[idx]); 1143 sv = newSVsv (AvARRAY (dec->stringref)[idx]);
829 } 1144 }
830 break; 1145 break;
832 case CBOR_TAG_VALUE_SHAREABLE: 1147 case CBOR_TAG_VALUE_SHAREABLE:
833 { 1148 {
834 if (ecb_expect_false (!dec->shareable)) 1149 if (ecb_expect_false (!dec->shareable))
835 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ()); 1150 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
836 1151
1152 if (dec->cbor.flags & F_ALLOW_CYCLES)
1153 {
837 sv = newSV (0); 1154 sv = newSV (0);
838 av_push (dec->shareable, SvREFCNT_inc_NN (sv)); 1155 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
839 1156
840 SV *osv = decode_sv (dec); 1157 SV *osv = decode_sv (dec);
841 sv_setsv (sv, osv); 1158 sv_setsv (sv, osv);
842 SvREFCNT_dec_NN (osv); 1159 SvREFCNT_dec_NN (osv);
1160 }
1161 else
1162 {
1163 av_push (dec->shareable, &PL_sv_undef);
1164 int idx = AvFILLp (dec->shareable);
1165 sv = decode_sv (dec);
1166 av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
1167 }
843 } 1168 }
844 break; 1169 break;
845 1170
846 case CBOR_TAG_VALUE_SHAREDREF: 1171 case CBOR_TAG_VALUE_SHAREDREF:
847 { 1172 {
848 if ((*dec->cur >> 5) != 0) 1173 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
849 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)"); 1174 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
850 1175
851 UV idx = decode_uint (dec); 1176 UV idx = decode_uint (dec);
852 1177
853 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable)) 1178 if (!dec->shareable || idx >= (UV)(1 + AvFILLp (dec->shareable)))
854 ERR ("corrupted CBOR data (sharedref index out of bounds)"); 1179 ERR ("corrupted CBOR data (sharedref index out of bounds)");
855 1180
856 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]); 1181 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
1182
1183 if (sv == &PL_sv_undef)
1184 ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
857 } 1185 }
858 break; 1186 break;
859 1187
860 case CBOR_TAG_PERL_OBJECT: 1188 case CBOR_TAG_PERL_OBJECT:
861 { 1189 {
1190 if (dec->cbor.flags & F_FORBID_OBJECTS)
1191 goto filter;
1192
862 sv = decode_sv (dec); 1193 sv = decode_sv (dec);
863 1194
864 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) 1195 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
865 ERR ("corrupted CBOR data (non-array perl object)"); 1196 ERR ("corrupted CBOR data (non-array perl object)");
866 1197
876 if (!method) 1207 if (!method)
877 ERR ("cannot decode perl-object (package does not have a THAW method)"); 1208 ERR ("cannot decode perl-object (package does not have a THAW method)");
878 1209
879 dSP; 1210 dSP;
880 1211
881 ENTER; SAVETMPS; PUSHMARK (SP); 1212 ENTER; SAVETMPS;
1213 PUSHMARK (SP);
882 EXTEND (SP, len + 1); 1214 EXTEND (SP, len + 1);
883 // we re-bless the reference to get overload and other niceties right 1215 // we re-bless the reference to get overload and other niceties right
884 PUSHs (*av_fetch (av, 0, 1)); 1216 PUSHs (*av_fetch (av, 0, 1));
885 PUSHs (sv_cbor); 1217 PUSHs (sv_cbor);
886 1218
894 SPAGAIN; 1226 SPAGAIN;
895 1227
896 if (SvTRUE (ERRSV)) 1228 if (SvTRUE (ERRSV))
897 { 1229 {
898 FREETMPS; LEAVE; 1230 FREETMPS; LEAVE;
899 ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV)))); 1231 ERR_ERRSV;
900 } 1232 }
901 1233
902 SvREFCNT_dec (sv); 1234 SvREFCNT_dec_NN (sv);
903 sv = SvREFCNT_inc (POPs); 1235 sv = SvREFCNT_inc (POPs);
904 1236
905 PUTBACK; 1237 PUTBACK;
906 1238
907 FREETMPS; LEAVE; 1239 FREETMPS; LEAVE;
908 } 1240 }
909 break; 1241 break;
910 1242
911 default: 1243 default:
1244 filter:
912 { 1245 {
1246 SV *tag_sv = newSVuv (tag);
1247
913 sv = decode_sv (dec); 1248 sv = decode_sv (dec);
914 1249
1250 dSP;
1251 ENTER; SAVETMPS;
1252 PUSHMARK (SP);
1253 EXTEND (SP, 2);
1254 PUSHs (tag_sv);
1255 PUSHs (sv);
1256
1257 PUTBACK;
1258 int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL);
1259 SPAGAIN;
1260
1261 if (SvTRUE (ERRSV))
1262 {
1263 SvREFCNT_dec_NN (tag_sv);
1264 FREETMPS; LEAVE;
1265 ERR_ERRSV;
1266 }
1267
1268 if (count)
1269 {
1270 SvREFCNT_dec_NN (tag_sv);
1271 SvREFCNT_dec_NN (sv);
1272 sv = SvREFCNT_inc_NN (TOPs);
1273 SP -= count;
1274 }
1275 else
1276 {
915 AV *av = newAV (); 1277 AV *av = newAV ();
916 av_push (av, newSVuv (tag)); 1278 av_push (av, tag_sv);
917 av_push (av, sv); 1279 av_push (av, sv);
918 1280
919 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 1281 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
920 ? cbor_tagged_stash 1282 ? cbor_tagged_stash
921 : gv_stashpv ("CBOR::XS::Tagged" , 1); 1283 : gv_stashpv ("CBOR::XS::Tagged" , 1);
922
923 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash); 1284 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
1285 }
1286
1287 PUTBACK;
1288
1289 FREETMPS; LEAVE;
924 } 1290 }
925 break; 1291 break;
926 } 1292 }
927 1293
928 return sv; 1294 return sv;
935static SV * 1301static SV *
936decode_sv (dec_t *dec) 1302decode_sv (dec_t *dec)
937{ 1303{
938 WANT (1); 1304 WANT (1);
939 1305
940 switch (*dec->cur >> 5) 1306 switch (*dec->cur >> MAJOR_SHIFT)
941 { 1307 {
942 case 0: // unsigned int 1308 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
943 return newSVuv (decode_uint (dec)); 1309 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
944 case 1: // negative int 1310 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
945 return newSViv (-1 - (IV)decode_uint (dec)); 1311 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
946 case 2: // octet string 1312 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
947 return decode_str (dec, 0); 1313 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
948 case 3: // utf-8 string 1314 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
949 return decode_str (dec, 1); 1315
950 case 4: // array 1316 case MAJOR_MISC >> MAJOR_SHIFT:
951 return decode_av (dec);
952 case 5: // map
953 return decode_hv (dec);
954 case 6: // tag
955 return decode_tagged (dec);
956 case 7: // misc
957 switch (*dec->cur++ & 31) 1317 switch (*dec->cur++ & MINOR_MASK)
958 { 1318 {
959 case 20: 1319 case SIMPLE_FALSE:
960#if CBOR_SLOW 1320#if CBOR_SLOW
961 types_false = get_bool ("Types::Serialiser::false"); 1321 types_false = get_bool ("Types::Serialiser::false");
962#endif 1322#endif
963 return newSVsv (types_false); 1323 return newSVsv (types_false);
964 case 21: 1324 case SIMPLE_TRUE:
965#if CBOR_SLOW 1325#if CBOR_SLOW
966 types_true = get_bool ("Types::Serialiser::true"); 1326 types_true = get_bool ("Types::Serialiser::true");
967#endif 1327#endif
968 return newSVsv (types_true); 1328 return newSVsv (types_true);
969 case 22: 1329 case SIMPLE_NULL:
970 return newSVsv (&PL_sv_undef); 1330 return newSVsv (&PL_sv_undef);
971 case 23: 1331 case SIMPLE_UNDEF:
972#if CBOR_SLOW 1332#if CBOR_SLOW
973 types_error = get_bool ("Types::Serialiser::error"); 1333 types_error = get_bool ("Types::Serialiser::error");
974#endif 1334#endif
975 return newSVsv (types_error); 1335 return newSVsv (types_error);
976 1336
977 case 25: 1337 case MISC_FLOAT16:
978 { 1338 {
979 WANT (2); 1339 WANT (2);
980 1340
981 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1341 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
982 dec->cur += 2; 1342 dec->cur += 2;
983 1343
984 return newSVnv (ecb_binary16_to_float (fp)); 1344 return newSVnv (ecb_binary16_to_float (fp));
985 } 1345 }
986 1346
987 case 26: 1347 case MISC_FLOAT32:
988 { 1348 {
989 uint32_t fp; 1349 uint32_t fp;
990 WANT (4); 1350 WANT (4);
991 memcpy (&fp, dec->cur, 4); 1351 memcpy (&fp, dec->cur, 4);
992 dec->cur += 4; 1352 dec->cur += 4;
995 fp = ecb_bswap32 (fp); 1355 fp = ecb_bswap32 (fp);
996 1356
997 return newSVnv (ecb_binary32_to_float (fp)); 1357 return newSVnv (ecb_binary32_to_float (fp));
998 } 1358 }
999 1359
1000 case 27: 1360 case MISC_FLOAT64:
1001 { 1361 {
1002 uint64_t fp; 1362 uint64_t fp;
1003 WANT (8); 1363 WANT (8);
1004 memcpy (&fp, dec->cur, 8); 1364 memcpy (&fp, dec->cur, 8);
1005 dec->cur += 8; 1365 dec->cur += 8;
1008 fp = ecb_bswap64 (fp); 1368 fp = ecb_bswap64 (fp);
1009 1369
1010 return newSVnv (ecb_binary64_to_double (fp)); 1370 return newSVnv (ecb_binary64_to_double (fp));
1011 } 1371 }
1012 1372
1013 // 0..19 unassigned 1373 // 0..19 unassigned simple
1014 // 24 reserved + unassigned (reserved values are not encodable) 1374 // 24 reserved + unassigned simple (reserved values are not encodable)
1375 // 28-30 unassigned misc
1376 // 31 break code
1015 default: 1377 default:
1016 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1378 ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1017 } 1379 }
1018 1380
1019 break; 1381 break;
1020 } 1382 }
1021 1383
1024} 1386}
1025 1387
1026static SV * 1388static SV *
1027decode_cbor (SV *string, CBOR *cbor, char **offset_return) 1389decode_cbor (SV *string, CBOR *cbor, char **offset_return)
1028{ 1390{
1029 dec_t dec = { }; 1391 dec_t dec = { 0 };
1030 SV *sv; 1392 SV *sv;
1031 STRLEN len; 1393 STRLEN len;
1032 char *data = SvPVbyte (string, len); 1394 char *data = SvPVbyte (string, len);
1033 1395
1034 if (len > cbor->max_size && cbor->max_size) 1396 if (len > cbor->max_size && cbor->max_size)
1048 if (dec.cur != dec.end && !dec.err) 1410 if (dec.cur != dec.end && !dec.err)
1049 dec.err = "garbage after CBOR object"; 1411 dec.err = "garbage after CBOR object";
1050 1412
1051 if (dec.err) 1413 if (dec.err)
1052 { 1414 {
1415 if (dec.shareable)
1416 {
1417 // need to break cyclic links, which would all be in shareable
1418 int i;
1419 SV **svp;
1420
1421 for (i = av_len (dec.shareable) + 1; i--; )
1422 if ((svp = av_fetch (dec.shareable, i, 0)))
1423 sv_setsv (*svp, &PL_sv_undef);
1424 }
1425
1053 SvREFCNT_dec (sv); 1426 SvREFCNT_dec_NN (sv);
1427
1428 if (dec.err_sv)
1429 sv_2mortal (dec.err_sv);
1430
1054 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur); 1431 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur);
1055 } 1432 }
1056 1433
1057 sv = sv_2mortal (sv); 1434 sv = sv_2mortal (sv);
1058 1435
1059 return sv; 1436 return sv;
1060} 1437}
1061 1438
1439/////////////////////////////////////////////////////////////////////////////
1440// incremental parser
1441
1442#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
1443
1444// returns 0 for notyet, 1 for success or error
1445static int
1446incr_parse (CBOR *self, SV *cborstr)
1447{
1448 STRLEN cur;
1449 SvPV (cborstr, cur);
1450
1451 while (ecb_expect_true (self->incr_need <= cur))
1452 {
1453 // table of integer count bytes
1454 static I8 incr_len[MINOR_MASK + 1] = {
1455 0, 0, 0, 0, 0, 0, 0, 0,
1456 0, 0, 0, 0, 0, 0, 0, 0,
1457 0, 0, 0, 0, 0, 0, 0, 0,
1458 1, 2, 4, 8,-1,-1,-1,-2
1459 };
1460
1461 const U8 *p = SvPVX (cborstr) + self->incr_pos;
1462 U8 m = *p & MINOR_MASK;
1463 IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1464 I8 ilen = incr_len[m];
1465
1466 self->incr_need = self->incr_pos + 1;
1467
1468 if (ecb_expect_false (ilen < 0))
1469 {
1470 if (m != MINOR_INDEF)
1471 return 1; // error
1472
1473 if (*p == (MAJOR_MISC | MINOR_INDEF))
1474 {
1475 if (count >= 0)
1476 return 1; // error
1477
1478 count = 1;
1479 }
1480 else
1481 {
1482 av_push (self->incr_count, newSViv (-1)); //TODO: nest
1483 count = -1;
1484 }
1485 }
1486 else
1487 {
1488 self->incr_need += ilen;
1489 if (ecb_expect_false (self->incr_need > cur))
1490 return 0;
1491
1492 int major = *p >> MAJOR_SHIFT;
1493
1494 switch (major)
1495 {
1496 case MAJOR_TAG >> MAJOR_SHIFT:
1497 ++count; // tags merely prefix another value
1498 break;
1499
1500 case MAJOR_BYTES >> MAJOR_SHIFT:
1501 case MAJOR_TEXT >> MAJOR_SHIFT:
1502 case MAJOR_ARRAY >> MAJOR_SHIFT:
1503 case MAJOR_MAP >> MAJOR_SHIFT:
1504 {
1505 UV len;
1506
1507 if (ecb_expect_false (ilen))
1508 {
1509 len = 0;
1510
1511 do {
1512 len = (len << 8) | *++p;
1513 } while (--ilen);
1514 }
1515 else
1516 len = m;
1517
1518 switch (major)
1519 {
1520 case MAJOR_BYTES >> MAJOR_SHIFT:
1521 case MAJOR_TEXT >> MAJOR_SHIFT:
1522 self->incr_need += len;
1523 if (ecb_expect_false (self->incr_need > cur))
1524 return 0;
1525
1526 break;
1527
1528 case MAJOR_MAP >> MAJOR_SHIFT:
1529 len <<= 1;
1530 case MAJOR_ARRAY >> MAJOR_SHIFT:
1531 if (len)
1532 {
1533 av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
1534 count = len + 1;
1535 }
1536 break;
1537 }
1538 }
1539 }
1540 }
1541
1542 self->incr_pos = self->incr_need;
1543
1544 if (count > 0)
1545 {
1546 while (!--count)
1547 {
1548 if (!AvFILLp (self->incr_count))
1549 return 1; // done
1550
1551 SvREFCNT_dec_NN (av_pop (self->incr_count));
1552 count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1553 }
1554
1555 SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
1556 }
1557 }
1558
1559 return 0;
1560}
1561
1562
1062///////////////////////////////////////////////////////////////////////////// 1563/////////////////////////////////////////////////////////////////////////////
1063// XS interface functions 1564// XS interface functions
1064 1565
1065MODULE = CBOR::XS PACKAGE = CBOR::XS 1566MODULE = CBOR::XS PACKAGE = CBOR::XS
1066 1567
1074 1575
1075 types_true = get_bool ("Types::Serialiser::true" ); 1576 types_true = get_bool ("Types::Serialiser::true" );
1076 types_false = get_bool ("Types::Serialiser::false"); 1577 types_false = get_bool ("Types::Serialiser::false");
1077 types_error = get_bool ("Types::Serialiser::error"); 1578 types_error = get_bool ("Types::Serialiser::error");
1078 1579
1580 default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1581
1079 sv_cbor = newSVpv ("CBOR", 0); 1582 sv_cbor = newSVpv ("CBOR", 0);
1080 SvREADONLY_on (sv_cbor); 1583 SvREADONLY_on (sv_cbor);
1584
1585 assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1));
1081} 1586}
1082 1587
1083PROTOTYPES: DISABLE 1588PROTOTYPES: DISABLE
1084 1589
1085void CLONE (...) 1590void CLONE (...)
1104void shrink (CBOR *self, int enable = 1) 1609void shrink (CBOR *self, int enable = 1)
1105 ALIAS: 1610 ALIAS:
1106 shrink = F_SHRINK 1611 shrink = F_SHRINK
1107 allow_unknown = F_ALLOW_UNKNOWN 1612 allow_unknown = F_ALLOW_UNKNOWN
1108 allow_sharing = F_ALLOW_SHARING 1613 allow_sharing = F_ALLOW_SHARING
1614 allow_cycles = F_ALLOW_CYCLES
1615 forbid_objects = F_FORBID_OBJECTS
1109 dedup_strings = F_DEDUP_STRINGS 1616 pack_strings = F_PACK_STRINGS
1617 text_keys = F_TEXT_KEYS
1618 text_strings = F_TEXT_STRINGS
1619 validate_utf8 = F_VALIDATE_UTF8
1110 PPCODE: 1620 PPCODE:
1111{ 1621{
1112 if (enable) 1622 if (enable)
1113 self->flags |= ix; 1623 self->flags |= ix;
1114 else 1624 else
1120void get_shrink (CBOR *self) 1630void get_shrink (CBOR *self)
1121 ALIAS: 1631 ALIAS:
1122 get_shrink = F_SHRINK 1632 get_shrink = F_SHRINK
1123 get_allow_unknown = F_ALLOW_UNKNOWN 1633 get_allow_unknown = F_ALLOW_UNKNOWN
1124 get_allow_sharing = F_ALLOW_SHARING 1634 get_allow_sharing = F_ALLOW_SHARING
1635 get_allow_cycles = F_ALLOW_CYCLES
1636 get_forbid_objects = F_FORBID_OBJECTS
1125 get_dedup_strings = F_DEDUP_STRINGS 1637 get_pack_strings = F_PACK_STRINGS
1638 get_text_keys = F_TEXT_KEYS
1639 get_text_strings = F_TEXT_STRINGS
1640 get_validate_utf8 = F_VALIDATE_UTF8
1126 PPCODE: 1641 PPCODE:
1127 XPUSHs (boolSV (self->flags & ix)); 1642 XPUSHs (boolSV (self->flags & ix));
1128 1643
1129void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1644void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1130 PPCODE: 1645 PPCODE:
1143 XPUSHs (ST (0)); 1658 XPUSHs (ST (0));
1144 1659
1145int get_max_size (CBOR *self) 1660int get_max_size (CBOR *self)
1146 CODE: 1661 CODE:
1147 RETVAL = self->max_size; 1662 RETVAL = self->max_size;
1663 OUTPUT:
1664 RETVAL
1665
1666void filter (CBOR *self, SV *filter = 0)
1667 PPCODE:
1668 SvREFCNT_dec (self->filter);
1669 self->filter = filter ? newSVsv (filter) : filter;
1670 XPUSHs (ST (0));
1671
1672SV *get_filter (CBOR *self)
1673 CODE:
1674 RETVAL = self->filter ? self->filter : NEWSV (0, 0);
1148 OUTPUT: 1675 OUTPUT:
1149 RETVAL 1676 RETVAL
1150 1677
1151void encode (CBOR *self, SV *scalar) 1678void encode (CBOR *self, SV *scalar)
1152 PPCODE: 1679 PPCODE:
1167 EXTEND (SP, 2); 1694 EXTEND (SP, 2);
1168 PUSHs (sv); 1695 PUSHs (sv);
1169 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1696 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1170} 1697}
1171 1698
1699void incr_parse (CBOR *self, SV *cborstr)
1700 ALIAS:
1701 incr_parse_multiple = 1
1702 PPCODE:
1703{
1704 if (SvUTF8 (cborstr))
1705 sv_utf8_downgrade (cborstr, 0);
1706
1707 if (!self->incr_count)
1708 {
1709 self->incr_count = newAV ();
1710 self->incr_pos = 0;
1711 self->incr_need = 1;
1712
1713 av_push (self->incr_count, newSViv (1));
1714 }
1715
1716 do
1717 {
1718 if (!incr_parse (self, cborstr))
1719 {
1720 if (self->incr_need > self->max_size && self->max_size)
1721 croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
1722 (unsigned long)self->incr_need, (unsigned long)self->max_size);
1723
1724 break;
1725 }
1726
1727 SV *sv;
1728 char *offset;
1729
1730 PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
1731 XPUSHs (sv);
1732
1733 sv_chop (cborstr, offset);
1734
1735 av_clear (self->incr_count);
1736 av_push (self->incr_count, newSViv (1));
1737
1738 self->incr_pos = 0;
1739 self->incr_need = self->incr_pos + 1;
1740 }
1741 while (ix);
1742}
1743
1744void incr_reset (CBOR *self)
1745 CODE:
1746{
1747 SvREFCNT_dec (self->incr_count);
1748 self->incr_count = 0;
1749}
1750
1751void DESTROY (CBOR *self)
1752 PPCODE:
1753 cbor_free (self);
1754
1172PROTOTYPES: ENABLE 1755PROTOTYPES: ENABLE
1173 1756
1174void encode_cbor (SV *scalar) 1757void encode_cbor (SV *scalar)
1758 ALIAS:
1759 encode_cbor = 0
1760 encode_cbor_sharing = F_ALLOW_SHARING
1175 PPCODE: 1761 PPCODE:
1176{ 1762{
1177 CBOR cbor; 1763 CBOR cbor;
1178 cbor_init (&cbor); 1764 cbor_init (&cbor);
1765 cbor.flags |= ix;
1179 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1766 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1180 XPUSHs (scalar); 1767 XPUSHs (scalar);
1181} 1768}
1182 1769
1183void decode_cbor (SV *cborstr) 1770void decode_cbor (SV *cborstr)
1187 cbor_init (&cbor); 1774 cbor_init (&cbor);
1188 PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN; 1775 PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN;
1189 XPUSHs (cborstr); 1776 XPUSHs (cborstr);
1190} 1777}
1191 1778
1779#ifdef __AFL_COMPILER
1780
1781void
1782afl_init ()
1783 CODE:
1784 __AFL_INIT ();
1785
1786int
1787afl_loop (unsigned int count = 10000)
1788 CODE:
1789 RETVAL = __AFL_LOOP (count);
1790 OUTPUT:
1791 RETVAL
1792
1793#endif
1794

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines