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 1.17 by root, Tue Oct 29 22:04:52 2013 UTC vs.
Revision 1.69 by root, Mon Nov 30 20:38:25 2020 UTC

6#include <string.h> 6#include <string.h>
7#include <stdlib.h> 7#include <stdlib.h>
8#include <stdio.h> 8#include <stdio.h>
9#include <limits.h> 9#include <limits.h>
10#include <float.h> 10#include <float.h>
11#include <inttypes.h>
11 12
13#define ECB_NO_THREADS 1
12#include "ecb.h" 14#include "ecb.h"
13 15
14// compatibility with perl <5.18 16// compatibility with perl <5.18
15#ifndef HvNAMELEN_get 17#ifndef HvNAMELEN_get
16# define HvNAMELEN_get(hv) strlen (HvNAME (hv)) 18# define HvNAMELEN_get(hv) strlen (HvNAME (hv))
19# define HvNAMELEN(hv) HvNAMELEN_get (hv) 21# define HvNAMELEN(hv) HvNAMELEN_get (hv)
20#endif 22#endif
21#ifndef HvNAMEUTF8 23#ifndef HvNAMEUTF8
22# define HvNAMEUTF8(hv) 0 24# define HvNAMEUTF8(hv) 0
23#endif 25#endif
26#ifndef SvREFCNT_inc_NN
27# define SvREFCNT_inc_NN(sv) SvREFCNT_inc (sv)
28#endif
29#ifndef SvREFCNT_dec_NN
30# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
31#endif
32
33// known major and minor types
34enum cbor_type
35{
36 MAJOR_SHIFT = 5,
37 MINOR_MASK = 0x1f,
38
39 MAJOR_POS_INT = 0 << MAJOR_SHIFT,
40 MAJOR_NEG_INT = 1 << MAJOR_SHIFT,
41 MAJOR_BYTES = 2 << MAJOR_SHIFT,
42 MAJOR_TEXT = 3 << MAJOR_SHIFT,
43 MAJOR_ARRAY = 4 << MAJOR_SHIFT,
44 MAJOR_MAP = 5 << MAJOR_SHIFT,
45 MAJOR_TAG = 6 << MAJOR_SHIFT,
46 MAJOR_MISC = 7 << MAJOR_SHIFT,
47
48 // INT/STRING/ARRAY/MAP subtypes
49 LENGTH_EXT1 = 24,
50 LENGTH_EXT2 = 25,
51 LENGTH_EXT4 = 26,
52 LENGTH_EXT8 = 27,
53
54 // SIMPLE types (effectively MISC subtypes)
55 SIMPLE_FALSE = 20,
56 SIMPLE_TRUE = 21,
57 SIMPLE_NULL = 22,
58 SIMPLE_UNDEF = 23,
59
60 // MISC subtype (unused)
61 MISC_EXT1 = 24,
62 MISC_FLOAT16 = 25,
63 MISC_FLOAT32 = 26,
64 MISC_FLOAT64 = 27,
65
66 // BYTES/TEXT/ARRAY/MAP
67 MINOR_INDEF = 31,
68};
24 69
25// known tags 70// known tags
26enum cbor_tag 71enum cbor_tag
27{ 72{
28 // inofficial extensions (pending iana registration) 73 // extensions
29 CBOR_TAG_PERL_OBJECT = 256, 74 CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
30 CBOR_TAG_GENERIC_OBJECT = 257, 75 CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
76 CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
77 CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
78 CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
79 CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref
80 CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
31 81
32 // rfc7049 82 // rfc7049
33 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8 83 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
34 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any 84 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
35 CBOR_TAG_POS_BIGNUM = 2, // byte string 85 CBOR_TAG_POS_BIGNUM = 2, // byte string
36 CBOR_TAG_NEG_BIGNUM = 3, // byte string 86 CBOR_TAG_NEG_BIGNUM = 3, // byte string
37 CBOR_TAG_DECIMAL = 4, // decimal fraction, array 87 CBOR_TAG_DECIMAL = 4, // decimal fraction, array
38 CBOR_TAG_BIGFLOAT = 5, // array 88 CBOR_TAG_BIGFLOAT = 5, // array
39 89
40 CBOR_TAG_CONV_B64U = 21, // base64url, any 90 CBOR_TAG_CONV_B64U = 21, // base64url, any
41 CBOR_TAG_CONV_B64 = 22, // base64, any 91 CBOR_TAG_CONV_B64 = 22, // base64, any
42 CBOR_TAG_CONV_HEX = 23, // base16, any 92 CBOR_TAG_CONV_HEX = 23, // base16, any
43 CBOR_TAG_CBOR = 24, // embedded cbor, byte string 93 CBOR_TAG_CBOR = 24, // embedded cbor, byte string
44 94
45 CBOR_TAG_URI = 32, // URI rfc3986, utf-8 95 CBOR_TAG_URI = 32, // URI rfc3986, utf-8
46 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8 96 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
47 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8 97 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
48 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8 98 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
49 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8 99 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
50 100
51 CBOR_TAG_MAGIC = 55799 // self-describe cbor 101 CBOR_TAG_MAGIC = 55799, // self-describe cbor
52}; 102};
53 103
104// known forced types, also hardcoded in CBOR.pm
105enum
106{
107 AS_CBOR = 0,
108 AS_INT = 1,
109 AS_BYTES = 2,
110 AS_TEXT = 3,
111 AS_FLOAT16 = 4,
112 AS_FLOAT32 = 5,
113 AS_FLOAT64 = 6,
114 // possibly future enhancements: (generic) float, (generic) string
115};
116
54#define F_SHRINK 0x00000200UL 117#define F_SHRINK 0x00000001UL
55#define F_ALLOW_UNKNOWN 0x00002000UL 118#define F_ALLOW_UNKNOWN 0x00000002UL
119#define F_ALLOW_SHARING 0x00000004UL
120#define F_ALLOW_CYCLES 0x00000008UL
121#define F_FORBID_OBJECTS 0x00000010UL
122#define F_PACK_STRINGS 0x00000020UL
123#define F_TEXT_KEYS 0x00000040UL
124#define F_TEXT_STRINGS 0x00000080UL
125#define F_VALIDATE_UTF8 0x00000100UL
56 126
57#define INIT_SIZE 32 // initial scalar size to be allocated 127#define INIT_SIZE 32 // initial scalar size to be allocated
58 128
59#define SB do { 129#define SB do {
60#define SE } while (0) 130#define SE } while (0)
72# define CBOR_SLOW 0 142# define CBOR_SLOW 0
73# define CBOR_STASH cbor_stash 143# define CBOR_STASH cbor_stash
74#endif 144#endif
75 145
76static 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::
77static SV *types_true, *types_false, *types_error, *sv_cbor; 147static SV *types_true, *types_false, *types_error, *sv_cbor, *default_filter;
78 148
79typedef struct { 149typedef struct {
80 U32 flags; 150 U32 flags;
81 U32 max_depth; 151 U32 max_depth;
82 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.
83} CBOR; 159} CBOR;
84 160
85ecb_inline void 161ecb_inline void
86cbor_init (CBOR *cbor) 162cbor_init (CBOR *cbor)
87{ 163{
88 Zero (cbor, 1, CBOR); 164 Zero (cbor, 1, CBOR);
89 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);
90} 173}
91 174
92///////////////////////////////////////////////////////////////////////////// 175/////////////////////////////////////////////////////////////////////////////
93// utility functions 176// utility functions
94 177
116 SvPV_renew (sv, SvCUR (sv) + 1); 199 SvPV_renew (sv, SvCUR (sv) + 1);
117#endif 200#endif
118 } 201 }
119} 202}
120 203
121///////////////////////////////////////////////////////////////////////////// 204// minimum length of a string to be registered for stringref
122// fp hell 205ecb_inline int
123 206minimum_string_length (UV idx)
124//TODO 207{
208 return idx <= 23 ? 3
209 : idx <= 0xffU ? 4
210 : idx <= 0xffffU ? 5
211 : idx <= 0xffffffffU ? 7
212 : 11;
213}
125 214
126///////////////////////////////////////////////////////////////////////////// 215/////////////////////////////////////////////////////////////////////////////
127// encoder 216// encoder
128 217
129// structure used for encoding CBOR 218// structure used for encoding CBOR
132 char *cur; // SvPVX (sv) + current output position 221 char *cur; // SvPVX (sv) + current output position
133 char *end; // SvEND (sv) 222 char *end; // SvEND (sv)
134 SV *sv; // result scalar 223 SV *sv; // result scalar
135 CBOR cbor; 224 CBOR cbor;
136 U32 depth; // recursion level 225 U32 depth; // recursion level
226 HV *stringref[2]; // string => index, or 0 ([0] = bytes, [1] = utf-8)
227 UV stringref_idx;
228 HV *shareable; // ptr => index, or 0
229 UV shareable_idx;
137} enc_t; 230} enc_t;
138 231
139ecb_inline void 232ecb_inline void
140need (enc_t *enc, STRLEN len) 233need (enc_t *enc, STRLEN len)
141{ 234{
142 if (ecb_expect_false (enc->cur + len >= enc->end)) 235 if (ecb_expect_false ((uintptr_t)(enc->end - enc->cur) < len))
143 { 236 {
144 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv); 237 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv);
145 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1); 238 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1);
146 enc->cur = SvPVX (enc->sv) + cur; 239 enc->cur = SvPVX (enc->sv) + cur;
147 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1; 240 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;
153{ 246{
154 need (enc, 1); 247 need (enc, 1);
155 *enc->cur++ = ch; 248 *enc->cur++ = ch;
156} 249}
157 250
251// used for tags, intregers, element counts and so on
158static void 252static void
159encode_uint (enc_t *enc, int major, UV len) 253encode_uint (enc_t *enc, int major, UV len)
160{ 254{
161 need (enc, 9); 255 need (enc, 9);
162 256
163 if (len < 24) 257 if (ecb_expect_true (len < LENGTH_EXT1))
164 *enc->cur++ = major | len; 258 *enc->cur++ = major | len;
165 else if (len <= 0xff) 259 else if (ecb_expect_true (len <= 0xffU))
166 { 260 {
167 *enc->cur++ = major | 24; 261 *enc->cur++ = major | LENGTH_EXT1;
168 *enc->cur++ = len; 262 *enc->cur++ = len;
169 } 263 }
170 else if (len <= 0xffff) 264 else if (len <= 0xffffU)
171 { 265 {
172 *enc->cur++ = major | 25; 266 *enc->cur++ = major | LENGTH_EXT2;
173 *enc->cur++ = len >> 8; 267 *enc->cur++ = len >> 8;
174 *enc->cur++ = len; 268 *enc->cur++ = len;
175 } 269 }
176 else if (len <= 0xffffffff) 270 else if (len <= 0xffffffffU)
177 { 271 {
178 *enc->cur++ = major | 26; 272 *enc->cur++ = major | LENGTH_EXT4;
179 *enc->cur++ = len >> 24; 273 *enc->cur++ = len >> 24;
180 *enc->cur++ = len >> 16; 274 *enc->cur++ = len >> 16;
181 *enc->cur++ = len >> 8; 275 *enc->cur++ = len >> 8;
182 *enc->cur++ = len; 276 *enc->cur++ = len;
183 } 277 }
184 else 278 else
185 { 279 {
186 *enc->cur++ = major | 27; 280 *enc->cur++ = major | LENGTH_EXT8;
187 *enc->cur++ = len >> 56; 281 *enc->cur++ = len >> 56;
188 *enc->cur++ = len >> 48; 282 *enc->cur++ = len >> 48;
189 *enc->cur++ = len >> 40; 283 *enc->cur++ = len >> 40;
190 *enc->cur++ = len >> 32; 284 *enc->cur++ = len >> 32;
191 *enc->cur++ = len >> 24; 285 *enc->cur++ = len >> 24;
193 *enc->cur++ = len >> 8; 287 *enc->cur++ = len >> 8;
194 *enc->cur++ = len; 288 *enc->cur++ = len;
195 } 289 }
196} 290}
197 291
198static 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
199encode_str (enc_t *enc, int utf8, char *str, STRLEN len) 312encode_str_utf8 (enc_t *enc, int utf8, char *str, STRLEN len)
200{ 313{
201 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);
202 need (enc, len); 344 need (enc, len);
203 memcpy (enc->cur, str, len); 345 memcpy (enc->cur, str, len);
204 enc->cur += len; 346 enc->cur += len;
205} 347}
206 348
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
375encode_float16 (enc_t *enc, NV nv)
376{
377 need (enc, 1+2);
378
379 *enc->cur++ = MAJOR_MISC | MISC_FLOAT16;
380
381 uint16_t fp = ecb_float_to_binary16 (nv);
382
383 if (!ecb_big_endian ())
384 fp = ecb_bswap16 (fp);
385
386 memcpy (enc->cur, &fp, 2);
387 enc->cur += 2;
388}
389
390ecb_inline void
391encode_float32 (enc_t *enc, NV nv)
392{
393 need (enc, 1+4);
394
395 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
396
397 uint32_t fp = ecb_float_to_binary32 (nv);
398
399 if (!ecb_big_endian ())
400 fp = ecb_bswap32 (fp);
401
402 memcpy (enc->cur, &fp, 4);
403 enc->cur += 4;
404}
405
406ecb_inline void
407encode_float64 (enc_t *enc, NV nv)
408{
409 need (enc, 1+8);
410
411 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
412
413 uint64_t fp = ecb_double_to_binary64 (nv);
414
415 if (!ecb_big_endian ())
416 fp = ecb_bswap64 (fp);
417
418 memcpy (enc->cur, &fp, 8);
419 enc->cur += 8;
420}
421
422ecb_inline void
423encode_bool (enc_t *enc, int istrue)
424{
425 encode_ch (enc, istrue ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE);
426}
427
428ecb_inline void
429encode_forced (enc_t *enc, UV type, SV *sv)
430{
431 switch (type)
432 {
433 case AS_CBOR:
434 {
435 STRLEN len;
436 char *str = SvPVbyte (sv, len);
437
438 need (enc, len);
439 memcpy (enc->cur, str, len);
440 enc->cur += len;
441 }
442 break;
443
444 case AS_BYTES:
445 {
446 STRLEN len;
447 char *str = SvPVbyte (sv, len);
448 encode_strref (enc, 0, 0, str, len);
449 }
450 break;
451
452 case AS_TEXT:
453 {
454 STRLEN len;
455 char *str = SvPVutf8 (sv, len);
456 encode_strref (enc, 1, 1, str, len);
457 }
458 break;
459
460 case AS_INT: encode_int (enc, sv); break;
461
462 case AS_FLOAT16: encode_float16 (enc, SvNV (sv)); break;
463 case AS_FLOAT32: encode_float32 (enc, SvNV (sv)); break;
464 case AS_FLOAT64: encode_float64 (enc, SvNV (sv)); break;
465
466 default:
467 croak ("encountered malformed CBOR::XS::Tagged object");
468 }
469}
470
207static void encode_sv (enc_t *enc, SV *sv); 471static void encode_sv (enc_t *enc, SV *sv);
208 472
209static void 473static void
210encode_av (enc_t *enc, AV *av) 474encode_av (enc_t *enc, AV *av)
211{ 475{
214 if (enc->depth >= enc->cbor.max_depth) 478 if (enc->depth >= enc->cbor.max_depth)
215 croak (ERR_NESTING_EXCEEDED); 479 croak (ERR_NESTING_EXCEEDED);
216 480
217 ++enc->depth; 481 ++enc->depth;
218 482
219 encode_uint (enc, 0x80, len + 1); 483 encode_uint (enc, MAJOR_ARRAY, len + 1);
220 484
485 if (ecb_expect_false (SvMAGICAL (av)))
221 for (i = 0; i <= len; ++i) 486 for (i = 0; i <= len; ++i)
222 { 487 {
223 SV **svp = av_fetch (av, i, 0); 488 SV **svp = av_fetch (av, i, 0);
224 encode_sv (enc, svp ? *svp : &PL_sv_undef); 489 encode_sv (enc, svp ? *svp : &PL_sv_undef);
225 } 490 }
491 else
492 for (i = 0; i <= len; ++i)
493 {
494 SV *sv = AvARRAY (av)[i];
495 encode_sv (enc, sv ? sv : &PL_sv_undef);
496 }
226 497
227 --enc->depth; 498 --enc->depth;
228} 499}
229 500
230static void 501static void
238 ++enc->depth; 509 ++enc->depth;
239 510
240 int pairs = hv_iterinit (hv); 511 int pairs = hv_iterinit (hv);
241 int mg = SvMAGICAL (hv); 512 int mg = SvMAGICAL (hv);
242 513
243 if (mg) 514 if (ecb_expect_false (mg))
244 encode_ch (enc, 0xa0 | 31); 515 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
245 else 516 else
246 encode_uint (enc, 0xa0, pairs); 517 encode_uint (enc, MAJOR_MAP, pairs);
247 518
248 while ((he = hv_iternext (hv))) 519 while ((he = hv_iternext (hv)))
249 { 520 {
250 if (HeKLEN (he) == HEf_SVKEY) 521 if (HeKLEN (he) == HEf_SVKEY)
251 encode_sv (enc, HeSVKEY (he)); 522 encode_sv (enc, HeSVKEY (he));
252 else 523 else
253 encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); 524 encode_strref (enc, enc->cbor.flags & (F_TEXT_KEYS | F_TEXT_STRINGS), HeKUTF8 (he), HeKEY (he), HeKLEN (he));
254 525
255 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); 526 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
256 } 527 }
257 528
258 if (mg) 529 if (ecb_expect_false (mg))
259 encode_ch (enc, 0xe0 | 31); 530 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
260 531
261 --enc->depth; 532 --enc->depth;
262} 533}
263 534
264// encode objects, arrays and special \0=false and \1=true values. 535// encode objects, arrays and special \0=false and \1=true values.
265static void 536static void
266encode_rv (enc_t *enc, SV *sv) 537encode_rv (enc_t *enc, SV *sv)
267{ 538{
268 svtype svt;
269
270 SvGETMAGIC (sv); 539 SvGETMAGIC (sv);
540
271 svt = SvTYPE (sv); 541 svtype svt = SvTYPE (sv);
272 542
273 if (ecb_expect_false (SvOBJECT (sv))) 543 if (ecb_expect_false (SvOBJECT (sv)))
274 { 544 {
275 HV *boolean_stash = !CBOR_SLOW || types_boolean_stash 545 HV *boolean_stash = !CBOR_SLOW || types_boolean_stash
276 ? types_boolean_stash 546 ? types_boolean_stash
281 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 551 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
282 ? cbor_tagged_stash 552 ? cbor_tagged_stash
283 : gv_stashpv ("CBOR::XS::Tagged" , 1); 553 : gv_stashpv ("CBOR::XS::Tagged" , 1);
284 554
285 HV *stash = SvSTASH (sv); 555 HV *stash = SvSTASH (sv);
286 GV *method;
287 556
288 if (stash == boolean_stash) 557 if (stash == boolean_stash)
289 encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20); 558 {
559 encode_bool (enc, SvIV (sv));
560 return;
561 }
290 else if (stash == error_stash) 562 else if (stash == error_stash)
291 encode_ch (enc, 0xe0 | 23); 563 {
564 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
565 return;
566 }
292 else if (stash == tagged_stash) 567 else if (stash == tagged_stash)
293 { 568 {
294 if (svt != SVt_PVAV) 569 if (svt != SVt_PVAV)
295 croak ("encountered CBOR::XS::Tagged object that isn't an array"); 570 croak ("encountered CBOR::XS::Tagged object that isn't an array");
296 571
572 switch (av_len ((AV *)sv))
573 {
574 case 2-1:
575 // actually a tagged value
297 encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); 576 encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
298 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); 577 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1));
578 break;
579
580 case 3-1:
581 // a forced type [value, type, undef]
582 encode_forced (enc, SvUV (*av_fetch ((AV *)sv, 1, 1)), *av_fetch ((AV *)sv, 0, 1));
583 break;
584
585 default:
586 croak ("encountered malformed CBOR::XS::Tagged object");
587 }
588
589 return;
590 }
591 }
592
593 if (ecb_expect_false (SvREFCNT (sv) > 1)
594 && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
595 {
596 if (ecb_expect_false (!enc->shareable))
597 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
598
599 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
600
601 if (SvOK (*svp))
299 } 602 {
603 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
604 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
605 return;
606 }
607 else
608 {
609 sv_setuv (*svp, enc->shareable_idx);
610 ++enc->shareable_idx;
611 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
612 }
613 }
614
615 if (ecb_expect_false (SvOBJECT (sv)))
616 {
617 HV *stash = SvSTASH (sv);
618 GV *method;
619
620 if (enc->cbor.flags & F_FORBID_OBJECTS)
621 croak ("encountered object '%s', but forbid_objects is enabled",
622 SvPV_nolen (sv_2mortal (newRV_inc (sv))));
300 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) 623 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
301 { 624 {
302 dSP; 625 dSP;
303 626
304 ENTER; SAVETMPS; PUSHMARK (SP); 627 ENTER; SAVETMPS;
628 PUSHMARK (SP);
305 // we re-bless the reference to get overload and other niceties right 629 // we re-bless the reference to get overload and other niceties right
306 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 630 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
307 631
308 PUTBACK; 632 PUTBACK;
309 // G_SCALAR ensures that return value is 1 633 // G_SCALAR ensures that return value is 1
322 } 646 }
323 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0) 647 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
324 { 648 {
325 dSP; 649 dSP;
326 650
327 ENTER; SAVETMPS; PUSHMARK (SP); 651 ENTER; SAVETMPS;
652 PUSHMARK (SP);
328 EXTEND (SP, 2); 653 EXTEND (SP, 2);
329 // we re-bless the reference to get overload and other niceties right 654 // we re-bless the reference to get overload and other niceties right
330 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 655 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
331 PUSHs (sv_cbor); 656 PUSHs (sv_cbor);
332 657
336 661
337 // catch this surprisingly common error 662 // catch this surprisingly common error
338 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) 663 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
339 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash)); 664 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
340 665
341 encode_uint (enc, 0xc0, CBOR_TAG_PERL_OBJECT); 666 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
342 encode_uint (enc, 0x80, count + 1); 667 encode_uint (enc, MAJOR_ARRAY, count + 1);
343 encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); 668 encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
344 669
345 while (count) 670 {
671 int i;
672
673 for (i = 0; i < count; ++i)
346 encode_sv (enc, SP[1 - count--]); 674 encode_sv (enc, SP[i + 1 - count]);
675
676 SP -= count;
677 }
347 678
348 PUTBACK; 679 PUTBACK;
349 680
350 FREETMPS; LEAVE; 681 FREETMPS; LEAVE;
351 } 682 }
355 } 686 }
356 else if (svt == SVt_PVHV) 687 else if (svt == SVt_PVHV)
357 encode_hv (enc, (HV *)sv); 688 encode_hv (enc, (HV *)sv);
358 else if (svt == SVt_PVAV) 689 else if (svt == SVt_PVAV)
359 encode_av (enc, (AV *)sv); 690 encode_av (enc, (AV *)sv);
360 else if (svt < SVt_PVAV)
361 {
362 STRLEN len = 0;
363 char *pv = svt ? SvPV (sv, len) : 0;
364
365 if (len == 1 && *pv == '1')
366 encode_ch (enc, 0xe0 | 21);
367 else if (len == 1 && *pv == '0')
368 encode_ch (enc, 0xe0 | 20);
369 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
370 encode_ch (enc, 0xe0 | 23);
371 else
372 croak ("cannot encode reference to scalar '%s' unless the scalar is 0 or 1",
373 SvPV_nolen (sv_2mortal (newRV_inc (sv))));
374 }
375 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
376 encode_ch (enc, 0xe0 | 23);
377 else 691 else
378 croak ("encountered %s, but CBOR can only represent references to arrays or hashes", 692 {
379 SvPV_nolen (sv_2mortal (newRV_inc (sv)))); 693 encode_tag (enc, CBOR_TAG_INDIRECTION);
694 encode_sv (enc, sv);
695 }
380} 696}
381 697
382static void 698static void
383encode_nv (enc_t *enc, SV *sv) 699encode_nv (enc_t *enc, SV *sv)
384{ 700{
385 double nv = SvNVX (sv); 701 double nv = SvNVX (sv);
386 702
387 need (enc, 9); 703 need (enc, 9);
388 704
389 if (ecb_expect_false (nv == (U32)nv)) 705 if (ecb_expect_false (nv == (NV)(U32)nv))
390 encode_uint (enc, 0x00, (U32)nv); 706 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
391 //TODO: maybe I32? 707 //TODO: maybe I32?
392 else if (ecb_expect_false (nv == (float)nv)) 708 else if (ecb_expect_false (nv == (float)nv))
393 { 709 encode_float32 (enc, nv);
394 uint32_t fp = ecb_float_to_binary32 (nv);
395
396 *enc->cur++ = 0xe0 | 26;
397
398 if (!ecb_big_endian ())
399 fp = ecb_bswap32 (fp);
400
401 memcpy (enc->cur, &fp, 4);
402 enc->cur += 4;
403 }
404 else 710 else
405 { 711 encode_float64 (enc, nv);
406 uint64_t fp = ecb_double_to_binary64 (nv);
407
408 *enc->cur++ = 0xe0 | 27;
409
410 if (!ecb_big_endian ())
411 fp = ecb_bswap64 (fp);
412
413 memcpy (enc->cur, &fp, 8);
414 enc->cur += 8;
415 }
416} 712}
417 713
418static void 714static void
419encode_sv (enc_t *enc, SV *sv) 715encode_sv (enc_t *enc, SV *sv)
420{ 716{
422 718
423 if (SvPOKp (sv)) 719 if (SvPOKp (sv))
424 { 720 {
425 STRLEN len; 721 STRLEN len;
426 char *str = SvPV (sv, len); 722 char *str = SvPV (sv, len);
427 encode_str (enc, SvUTF8 (sv), str, len); 723 encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len);
428 } 724 }
429 else if (SvNOKp (sv)) 725 else if (SvNOKp (sv))
430 encode_nv (enc, sv); 726 encode_nv (enc, sv);
431 else if (SvIOKp (sv)) 727 else if (SvIOKp (sv))
432 { 728 encode_int (enc, sv);
433 if (SvIsUV (sv))
434 encode_uint (enc, 0x00, SvUVX (sv));
435 else if (SvIVX (sv) >= 0)
436 encode_uint (enc, 0x00, SvIVX (sv));
437 else
438 encode_uint (enc, 0x20, -(SvIVX (sv) + 1));
439 }
440 else if (SvROK (sv)) 729 else if (SvROK (sv))
441 encode_rv (enc, SvRV (sv)); 730 encode_rv (enc, SvRV (sv));
442 else if (!SvOK (sv)) 731 else if (!SvOK (sv))
443 encode_ch (enc, 0xe0 | 22); 732 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
444 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 733 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
445 encode_ch (enc, 0xe0 | 23); 734 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
446 else 735 else
447 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data", 736 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
448 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 737 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
449} 738}
450 739
451static SV * 740static SV *
452encode_cbor (SV *scalar, CBOR *cbor) 741encode_cbor (SV *scalar, CBOR *cbor)
453{ 742{
454 enc_t enc; 743 enc_t enc = { 0 };
455 744
456 enc.cbor = *cbor; 745 enc.cbor = *cbor;
457 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 746 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
458 enc.cur = SvPVX (enc.sv); 747 enc.cur = SvPVX (enc.sv);
459 enc.end = SvEND (enc.sv); 748 enc.end = SvEND (enc.sv);
460 enc.depth = 0;
461 749
462 SvPOK_only (enc.sv); 750 SvPOK_only (enc.sv);
751
752 if (cbor->flags & F_PACK_STRINGS)
753 {
754 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
755 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
756 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
757 }
758
463 encode_sv (&enc, scalar); 759 encode_sv (&enc, scalar);
464 760
465 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv)); 761 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv));
466 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings 762 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings
467 763
481 U8 *end; // end of input string 777 U8 *end; // end of input string
482 const char *err; // parse error, if != 0 778 const char *err; // parse error, if != 0
483 CBOR cbor; 779 CBOR cbor;
484 U32 depth; // recursion depth 780 U32 depth; // recursion depth
485 U32 maxdepth; // recursion depth limit 781 U32 maxdepth; // recursion depth limit
782 AV *shareable;
783 AV *stringref;
784 SV *decode_tagged;
785 SV *err_sv; // optional sv for error, needs to be freed
486} dec_t; 786} dec_t;
487 787
488#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE 788// set dec->err to ERRSV
789ecb_cold static void
790err_errsv (dec_t *dec)
791{
792 if (!dec->err)
793 {
794 dec->err_sv = newSVsv (ERRSV);
489 795
490#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data") 796 // chop off the trailing \n
797 SvCUR_set (dec->err_sv, SvCUR (dec->err_sv) - 1);
798 *SvEND (dec->err_sv) = 0;
491 799
800 dec->err = SvPVutf8_nolen (dec->err_sv);
801 }
802}
803
804// the following functions are used to reduce code size and help the compiler to optimise
805ecb_cold static void
806err_set (dec_t *dec, const char *reason)
807{
808 if (!dec->err)
809 dec->err = reason;
810}
811
812ecb_cold static void
813err_unexpected_end (dec_t *dec)
814{
815 err_set (dec, "unexpected end of CBOR data");
816}
817
818#define ERR_DO(do) SB do; goto fail; SE
819#define ERR(reason) ERR_DO (err_set (dec, reason))
820#define ERR_ERRSV ERR_DO (err_errsv (dec))
821
822#define WANT(len) if (ecb_expect_false ((uintptr_t)(dec->end - dec->cur) < (STRLEN)len)) ERR_DO (err_unexpected_end (dec))
823
492#define DEC_INC_DEPTH if (++dec->depth > dec->cbor.max_depth) ERR (ERR_NESTING_EXCEEDED) 824#define DEC_INC_DEPTH if (ecb_expect_false (++dec->depth > dec->cbor.max_depth)) ERR (ERR_NESTING_EXCEEDED)
493#define DEC_DEC_DEPTH --dec->depth 825#define DEC_DEC_DEPTH --dec->depth
494 826
495static UV 827static UV
496decode_uint (dec_t *dec) 828decode_uint (dec_t *dec)
497{ 829{
498 switch (*dec->cur & 31) 830 U8 m = *dec->cur & MINOR_MASK;
499 { 831 ++dec->cur;
500 case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
501 case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
502 case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
503 return *dec->cur++ & 31;
504 832
505 case 24: 833 if (ecb_expect_true (m < LENGTH_EXT1))
834 return m;
835 else if (ecb_expect_true (m == LENGTH_EXT1))
836 {
506 WANT (2); 837 WANT (1);
507 dec->cur += 2; 838 dec->cur += 1;
508 return dec->cur[-1]; 839 return dec->cur[-1];
509 840 }
510 case 25: 841 else if (ecb_expect_true (m == LENGTH_EXT2))
842 {
511 WANT (3); 843 WANT (2);
512 dec->cur += 3; 844 dec->cur += 2;
513 return (((UV)dec->cur[-2]) << 8) 845 return (((UV)dec->cur[-2]) << 8)
514 | ((UV)dec->cur[-1]); 846 | ((UV)dec->cur[-1]);
515 847 }
516 case 26: 848 else if (ecb_expect_true (m == LENGTH_EXT4))
849 {
517 WANT (5); 850 WANT (4);
518 dec->cur += 5; 851 dec->cur += 4;
519 return (((UV)dec->cur[-4]) << 24) 852 return (((UV)dec->cur[-4]) << 24)
520 | (((UV)dec->cur[-3]) << 16) 853 | (((UV)dec->cur[-3]) << 16)
521 | (((UV)dec->cur[-2]) << 8) 854 | (((UV)dec->cur[-2]) << 8)
522 | ((UV)dec->cur[-1]); 855 | ((UV)dec->cur[-1]);
523 856 }
524 case 27: 857 else if (ecb_expect_true (m == LENGTH_EXT8))
858 {
525 WANT (9); 859 WANT (8);
526 dec->cur += 9; 860 dec->cur += 8;
861
862 return
863#if UVSIZE < 8
864 0
865#else
527 return (((UV)dec->cur[-8]) << 56) 866 (((UV)dec->cur[-8]) << 56)
528 | (((UV)dec->cur[-7]) << 48) 867 | (((UV)dec->cur[-7]) << 48)
529 | (((UV)dec->cur[-6]) << 40) 868 | (((UV)dec->cur[-6]) << 40)
530 | (((UV)dec->cur[-5]) << 32) 869 | (((UV)dec->cur[-5]) << 32)
870#endif
531 | (((UV)dec->cur[-4]) << 24) 871 | (((UV)dec->cur[-4]) << 24)
532 | (((UV)dec->cur[-3]) << 16) 872 | (((UV)dec->cur[-3]) << 16)
533 | (((UV)dec->cur[-2]) << 8) 873 | (((UV)dec->cur[-2]) << 8)
534 | ((UV)dec->cur[-1]); 874 | ((UV)dec->cur[-1]);
535 875 }
536 default: 876 else
537 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 877 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
538 }
539 878
540fail: 879fail:
541 return 0; 880 return 0;
542} 881}
543 882
548{ 887{
549 AV *av = newAV (); 888 AV *av = newAV ();
550 889
551 DEC_INC_DEPTH; 890 DEC_INC_DEPTH;
552 891
553 if ((*dec->cur & 31) == 31) 892 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
554 { 893 {
555 ++dec->cur; 894 ++dec->cur;
556 895
557 for (;;) 896 for (;;)
558 { 897 {
559 WANT (1); 898 WANT (1);
560 899
561 if (*dec->cur == (0xe0 | 31)) 900 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
562 { 901 {
563 ++dec->cur; 902 ++dec->cur;
564 break; 903 break;
565 } 904 }
566 905
567 av_push (av, decode_sv (dec)); 906 av_push (av, decode_sv (dec));
568 } 907 }
569 } 908 }
570 else 909 else
571 { 910 {
572 int i, len = decode_uint (dec); 911 UV i, len = decode_uint (dec);
573 912
913 WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
574 av_fill (av, len - 1); 914 av_fill (av, len - 1);
575 915
576 for (i = 0; i < len; ++i) 916 for (i = 0; i < len; ++i)
577 AvARRAY (av)[i] = decode_sv (dec); 917 AvARRAY (av)[i] = decode_sv (dec);
578 } 918 }
579 919
580 DEC_DEC_DEPTH; 920 DEC_DEC_DEPTH;
581 return newRV_noinc ((SV *)av); 921 return newRV_noinc ((SV *)av);
582 922
583fail: 923fail:
584 SvREFCNT_dec (av); 924 SvREFCNT_dec_NN (av);
585 DEC_DEC_DEPTH; 925 DEC_DEC_DEPTH;
586 return &PL_sv_undef; 926 return &PL_sv_undef;
587} 927}
588 928
589static void 929static void
590decode_he (dec_t *dec, HV *hv) 930decode_he (dec_t *dec, HV *hv)
591{ 931{
592 // for speed reasons, we specialcase single-string 932 // for speed reasons, we specialcase single-string
593 // byte or utf-8 strings as keys. 933 // byte or utf-8 strings as keys, but only when !stringref
594 934
595 if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27) 935 if (ecb_expect_true (!dec->stringref))
936 if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
596 { 937 {
597 I32 len = decode_uint (dec); 938 STRLEN len = decode_uint (dec);
598 char *key = (char *)dec->cur; 939 char *key = (char *)dec->cur;
599 940
941 WANT (len);
600 dec->cur += len; 942 dec->cur += len;
601 943
602 hv_store (hv, key, len, decode_sv (dec), 0); 944 hv_store (hv, key, len, decode_sv (dec), 0);
945
946 return;
603 } 947 }
604 else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27) 948 else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
605 { 949 {
606 I32 len = decode_uint (dec); 950 STRLEN len = decode_uint (dec);
607 char *key = (char *)dec->cur; 951 char *key = (char *)dec->cur;
608 952
953 WANT (len);
609 dec->cur += len; 954 dec->cur += len;
610 955
956 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
957 if (!is_utf8_string (key, len))
958 ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
959
611 hv_store (hv, key, -len, decode_sv (dec), 0); 960 hv_store (hv, key, -len, decode_sv (dec), 0);
961
962 return;
612 } 963 }
613 else 964
614 {
615 SV *k = decode_sv (dec); 965 SV *k = decode_sv (dec);
616 SV *v = decode_sv (dec); 966 SV *v = decode_sv (dec);
617 967
968 // we leak memory if uncaught exceptions are thrown by random magical
969 // methods, and this is hopefully the only place where it can happen,
970 // so if there is a chance of an exception, take the very slow path.
971 // since catching exceptions is "undocumented/internal/forbidden" by
972 // the new p5p powers, we need to call out to a perl function :/
973 if (ecb_expect_false (SvAMAGIC (k)))
974 {
975 dSP;
976
977 ENTER; SAVETMPS;
978 PUSHMARK (SP);
979 EXTEND (SP, 3);
980 PUSHs (sv_2mortal (newRV_inc ((SV *)hv)));
981 PUSHs (sv_2mortal (k));
982 PUSHs (sv_2mortal (v));
983
984 PUTBACK;
985 call_pv ("CBOR::XS::_hv_store", G_VOID | G_DISCARD | G_EVAL);
986 SPAGAIN;
987
988 FREETMPS; LEAVE;
989
990 if (SvTRUE (ERRSV))
991 ERR_ERRSV;
992
993 return;
994 }
995
618 hv_store_ent (hv, k, v, 0); 996 hv_store_ent (hv, k, v, 0);
619 SvREFCNT_dec (k); 997 SvREFCNT_dec_NN (k);
620 } 998
999fail:
1000 ;
621} 1001}
622 1002
623static SV * 1003static SV *
624decode_hv (dec_t *dec) 1004decode_hv (dec_t *dec)
625{ 1005{
626 HV *hv = newHV (); 1006 HV *hv = newHV ();
627 1007
628 DEC_INC_DEPTH; 1008 DEC_INC_DEPTH;
629 1009
630 if ((*dec->cur & 31) == 31) 1010 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
631 { 1011 {
632 ++dec->cur; 1012 ++dec->cur;
633 1013
634 for (;;) 1014 for (;;)
635 { 1015 {
636 WANT (1); 1016 WANT (1);
637 1017
638 if (*dec->cur == (0xe0 | 31)) 1018 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
639 { 1019 {
640 ++dec->cur; 1020 ++dec->cur;
641 break; 1021 break;
642 } 1022 }
643 1023
644 decode_he (dec, hv); 1024 decode_he (dec, hv);
645 } 1025 }
646 } 1026 }
647 else 1027 else
648 { 1028 {
649 int pairs = decode_uint (dec); 1029 UV pairs = decode_uint (dec);
1030
1031 WANT (pairs); // complexity check - need at least one byte per value, do not allow supersize hashes
650 1032
651 while (pairs--) 1033 while (pairs--)
652 decode_he (dec, hv); 1034 decode_he (dec, hv);
653 } 1035 }
654 1036
655 DEC_DEC_DEPTH; 1037 DEC_DEC_DEPTH;
656 return newRV_noinc ((SV *)hv); 1038 return newRV_noinc ((SV *)hv);
657 1039
658fail: 1040fail:
659 SvREFCNT_dec (hv); 1041 SvREFCNT_dec_NN (hv);
660 DEC_DEC_DEPTH; 1042 DEC_DEC_DEPTH;
661 return &PL_sv_undef; 1043 return &PL_sv_undef;
662} 1044}
663 1045
664static SV * 1046static SV *
665decode_str (dec_t *dec, int utf8) 1047decode_str (dec_t *dec, int utf8)
666{ 1048{
667 SV *sv = 0; 1049 SV *sv = 0;
668 1050
669 if ((*dec->cur & 31) == 31) 1051 if (ecb_expect_false ((*dec->cur & MINOR_MASK) == MINOR_INDEF))
670 { 1052 {
1053 // indefinite length strings
671 ++dec->cur; 1054 ++dec->cur;
672 1055
1056 U8 major = *dec->cur & MAJOR_MISC;
1057
673 sv = newSVpvn ("", 0); 1058 sv = newSVpvn ("", 0);
674 1059
675 // not very fast, and certainly not robust against illegal input
676 for (;;) 1060 for (;;)
677 { 1061 {
678 WANT (1); 1062 WANT (1);
679 1063
680 if (*dec->cur == (0xe0 | 31)) 1064 if ((*dec->cur - major) > LENGTH_EXT8)
1065 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
681 { 1066 {
682 ++dec->cur; 1067 ++dec->cur;
683 break; 1068 break;
684 } 1069 }
1070 else
1071 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
685 1072
686 sv_catsv (sv, decode_sv (dec)); 1073 STRLEN len = decode_uint (dec);
1074
1075 WANT (len);
1076 sv_catpvn (sv, dec->cur, len);
1077 dec->cur += len;
687 } 1078 }
688 } 1079 }
689 else 1080 else
690 { 1081 {
691 STRLEN len = decode_uint (dec); 1082 STRLEN len = decode_uint (dec);
692 1083
693 WANT (len); 1084 WANT (len);
694 sv = newSVpvn (dec->cur, len); 1085 sv = newSVpvn (dec->cur, len);
695 dec->cur += len; 1086 dec->cur += len;
1087
1088 if (ecb_expect_false (dec->stringref)
1089 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
1090 av_push (dec->stringref, SvREFCNT_inc_NN (sv));
696 } 1091 }
697 1092
698 if (utf8) 1093 if (utf8)
1094 {
1095 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
1096 if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
1097 ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
1098
699 SvUTF8_on (sv); 1099 SvUTF8_on (sv);
1100 }
700 1101
701 return sv; 1102 return sv;
702 1103
703fail: 1104fail:
704 SvREFCNT_dec (sv); 1105 SvREFCNT_dec (sv);
706} 1107}
707 1108
708static SV * 1109static SV *
709decode_tagged (dec_t *dec) 1110decode_tagged (dec_t *dec)
710{ 1111{
1112 SV *sv = 0;
711 UV tag = decode_uint (dec); 1113 UV tag = decode_uint (dec);
1114
1115 WANT (1);
1116
1117 switch (tag)
1118 {
1119 case CBOR_TAG_MAGIC:
712 SV *sv = decode_sv (dec); 1120 sv = decode_sv (dec);
1121 break;
713 1122
714 if (tag == CBOR_TAG_MAGIC) 1123 case CBOR_TAG_INDIRECTION:
715 return sv; 1124 sv = newRV_noinc (decode_sv (dec));
716 else if (tag == CBOR_TAG_PERL_OBJECT) 1125 break;
717 { 1126
1127 case CBOR_TAG_STRINGREF_NAMESPACE:
1128 {
1129 // do not use SAVETMPS/FREETMPS, as these will
1130 // erase mortalised caches, e.g. "shareable"
1131 ENTER;
1132
1133 SAVESPTR (dec->stringref);
1134 dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());
1135
1136 sv = decode_sv (dec);
1137
1138 LEAVE;
1139 }
1140 break;
1141
1142 case CBOR_TAG_STRINGREF:
1143 {
1144 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
1145 ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
1146
1147 UV idx = decode_uint (dec);
1148
1149 if (!dec->stringref || idx >= (UV)(1 + AvFILLp (dec->stringref)))
1150 ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)");
1151
1152 sv = newSVsv (AvARRAY (dec->stringref)[idx]);
1153 }
1154 break;
1155
1156 case CBOR_TAG_VALUE_SHAREABLE:
1157 {
1158 if (ecb_expect_false (!dec->shareable))
1159 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
1160
1161 if (dec->cbor.flags & F_ALLOW_CYCLES)
1162 {
1163 sv = newSV (0);
1164 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
1165
1166 SV *osv = decode_sv (dec);
1167 sv_setsv (sv, osv);
1168 SvREFCNT_dec_NN (osv);
1169 }
1170 else
1171 {
1172 av_push (dec->shareable, &PL_sv_undef);
1173 int idx = AvFILLp (dec->shareable);
1174 sv = decode_sv (dec);
1175 av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
1176 }
1177 }
1178 break;
1179
1180 case CBOR_TAG_VALUE_SHAREDREF:
1181 {
1182 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
1183 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
1184
1185 UV idx = decode_uint (dec);
1186
1187 if (!dec->shareable || idx >= (UV)(1 + AvFILLp (dec->shareable)))
1188 ERR ("corrupted CBOR data (sharedref index out of bounds)");
1189
1190 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
1191
1192 if (sv == &PL_sv_undef)
1193 ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
1194 }
1195 break;
1196
1197 case CBOR_TAG_PERL_OBJECT:
1198 {
1199 if (dec->cbor.flags & F_FORBID_OBJECTS)
1200 goto filter;
1201
1202 sv = decode_sv (dec);
1203
718 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) 1204 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
719 ERR ("corrupted CBOR data (non-array perl object)"); 1205 ERR ("corrupted CBOR data (non-array perl object)");
720 1206
721 AV *av = (AV *)SvRV (sv); 1207 AV *av = (AV *)SvRV (sv);
722 int len = av_len (av) + 1; 1208 int len = av_len (av) + 1;
723 HV *stash = gv_stashsv (*av_fetch (av, 0, 1), 0); 1209 HV *stash = gv_stashsv (*av_fetch (av, 0, 1), 0);
724 1210
725 if (!stash) 1211 if (!stash)
726 ERR ("cannot decode perl-object (package does not exist)"); 1212 ERR ("cannot decode perl-object (package does not exist)");
727 1213
728 GV *method = gv_fetchmethod_autoload (stash, "THAW", 0); 1214 GV *method = gv_fetchmethod_autoload (stash, "THAW", 0);
729 1215
730 if (!method) 1216 if (!method)
731 ERR ("cannot decode perl-object (package does not have a THAW method)"); 1217 ERR ("cannot decode perl-object (package does not have a THAW method)");
732 1218
733 dSP; 1219 dSP;
734 1220
735 ENTER; SAVETMPS; PUSHMARK (SP); 1221 ENTER; SAVETMPS;
1222 PUSHMARK (SP);
736 EXTEND (SP, len + 1); 1223 EXTEND (SP, len + 1);
737 // we re-bless the reference to get overload and other niceties right 1224 // we re-bless the reference to get overload and other niceties right
738 PUSHs (*av_fetch (av, 0, 1)); 1225 PUSHs (*av_fetch (av, 0, 1));
739 PUSHs (sv_cbor); 1226 PUSHs (sv_cbor);
740 1227
741 int i; 1228 int i;
742 1229
743 for (i = 1; i < len; ++i) 1230 for (i = 1; i < len; ++i)
744 PUSHs (*av_fetch (av, i, 1)); 1231 PUSHs (*av_fetch (av, i, 1));
745 1232
746 PUTBACK; 1233 PUTBACK;
747 call_sv ((SV *)GvCV (method), G_SCALAR | G_EVAL); 1234 call_sv ((SV *)GvCV (method), G_SCALAR | G_EVAL);
748 SPAGAIN; 1235 SPAGAIN;
749 1236
750 if (SvTRUE (ERRSV)) 1237 if (SvTRUE (ERRSV))
751 { 1238 {
1239 FREETMPS; LEAVE;
1240 ERR_ERRSV;
1241 }
1242
1243 SvREFCNT_dec_NN (sv);
1244 sv = SvREFCNT_inc (POPs);
1245
1246 PUTBACK;
1247
752 FREETMPS; LEAVE; 1248 FREETMPS; LEAVE;
753 ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV)))); 1249 }
1250 break;
1251
1252 default:
1253 filter:
754 } 1254 {
1255 SV *tag_sv = newSVuv (tag);
755 1256
756 SvREFCNT_dec (sv); 1257 sv = decode_sv (dec);
757 sv = SvREFCNT_inc (POPs);
758 1258
1259 dSP;
1260 ENTER; SAVETMPS;
1261 PUSHMARK (SP);
1262 EXTEND (SP, 2);
1263 PUSHs (tag_sv);
1264 PUSHs (sv);
1265
759 PUTBACK; 1266 PUTBACK;
1267 int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL);
1268 SPAGAIN;
760 1269
1270 if (SvTRUE (ERRSV))
1271 {
1272 SvREFCNT_dec_NN (tag_sv);
761 FREETMPS; LEAVE; 1273 FREETMPS; LEAVE;
1274 ERR_ERRSV;
1275 }
762 1276
763 return sv; 1277 if (count)
764 } 1278 {
765 else 1279 SvREFCNT_dec_NN (tag_sv);
766 { 1280 SvREFCNT_dec_NN (sv);
1281 sv = SvREFCNT_inc_NN (TOPs);
1282 SP -= count;
1283 }
1284 else
1285 {
767 AV *av = newAV (); 1286 AV *av = newAV ();
768 av_push (av, newSVuv (tag)); 1287 av_push (av, tag_sv);
769 av_push (av, sv); 1288 av_push (av, sv);
770 1289
771 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 1290 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
772 ? cbor_tagged_stash 1291 ? cbor_tagged_stash
773 : gv_stashpv ("CBOR::XS::Tagged" , 1); 1292 : gv_stashpv ("CBOR::XS::Tagged" , 1);
774
775 return sv_bless (newRV_noinc ((SV *)av), tagged_stash); 1293 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
1294 }
1295
1296 PUTBACK;
1297
1298 FREETMPS; LEAVE;
1299 }
1300 break;
776 } 1301 }
1302
1303 return sv;
777 1304
778fail: 1305fail:
779 SvREFCNT_dec (sv); 1306 SvREFCNT_dec (sv);
780 return &PL_sv_undef; 1307 return &PL_sv_undef;
781} 1308}
783static SV * 1310static SV *
784decode_sv (dec_t *dec) 1311decode_sv (dec_t *dec)
785{ 1312{
786 WANT (1); 1313 WANT (1);
787 1314
788 switch (*dec->cur >> 5) 1315 switch (*dec->cur >> MAJOR_SHIFT)
789 { 1316 {
790 case 0: // unsigned int 1317 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
791 return newSVuv (decode_uint (dec)); 1318 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
792 case 1: // negative int 1319 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
793 return newSViv (-1 - (IV)decode_uint (dec)); 1320 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
794 case 2: // octet string 1321 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
795 return decode_str (dec, 0); 1322 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
796 case 3: // utf-8 string 1323 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
797 return decode_str (dec, 1); 1324
798 case 4: // array 1325 case MAJOR_MISC >> MAJOR_SHIFT:
799 return decode_av (dec);
800 case 5: // map
801 return decode_hv (dec);
802 case 6: // tag
803 return decode_tagged (dec);
804 case 7: // misc
805 switch (*dec->cur++ & 31) 1326 switch (*dec->cur++ & MINOR_MASK)
806 { 1327 {
807 case 20: 1328 case SIMPLE_FALSE:
808#if CBOR_SLOW 1329#if CBOR_SLOW
809 types_false = get_bool ("Types::Serialiser::false"); 1330 types_false = get_bool ("Types::Serialiser::false");
810#endif 1331#endif
811 return newSVsv (types_false); 1332 return newSVsv (types_false);
812 case 21: 1333 case SIMPLE_TRUE:
813#if CBOR_SLOW 1334#if CBOR_SLOW
814 types_true = get_bool ("Types::Serialiser::true"); 1335 types_true = get_bool ("Types::Serialiser::true");
815#endif 1336#endif
816 return newSVsv (types_true); 1337 return newSVsv (types_true);
817 case 22: 1338 case SIMPLE_NULL:
818 return newSVsv (&PL_sv_undef); 1339 return newSVsv (&PL_sv_undef);
819 case 23: 1340 case SIMPLE_UNDEF:
820#if CBOR_SLOW 1341#if CBOR_SLOW
821 types_error = get_bool ("Types::Serialiser::error"); 1342 types_error = get_bool ("Types::Serialiser::error");
822#endif 1343#endif
823 return newSVsv (types_error); 1344 return newSVsv (types_error);
824 1345
825 case 25: 1346 case MISC_FLOAT16:
826 { 1347 {
827 WANT (2); 1348 WANT (2);
828 1349
829 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1350 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
830 dec->cur += 2; 1351 dec->cur += 2;
831 1352
832 return newSVnv (ecb_binary16_to_float (fp)); 1353 return newSVnv (ecb_binary16_to_float (fp));
833 } 1354 }
834 1355
835 case 26: 1356 case MISC_FLOAT32:
836 { 1357 {
837 uint32_t fp; 1358 uint32_t fp;
838 WANT (4); 1359 WANT (4);
839 memcpy (&fp, dec->cur, 4); 1360 memcpy (&fp, dec->cur, 4);
840 dec->cur += 4; 1361 dec->cur += 4;
843 fp = ecb_bswap32 (fp); 1364 fp = ecb_bswap32 (fp);
844 1365
845 return newSVnv (ecb_binary32_to_float (fp)); 1366 return newSVnv (ecb_binary32_to_float (fp));
846 } 1367 }
847 1368
848 case 27: 1369 case MISC_FLOAT64:
849 { 1370 {
850 uint64_t fp; 1371 uint64_t fp;
851 WANT (8); 1372 WANT (8);
852 memcpy (&fp, dec->cur, 8); 1373 memcpy (&fp, dec->cur, 8);
853 dec->cur += 8; 1374 dec->cur += 8;
856 fp = ecb_bswap64 (fp); 1377 fp = ecb_bswap64 (fp);
857 1378
858 return newSVnv (ecb_binary64_to_double (fp)); 1379 return newSVnv (ecb_binary64_to_double (fp));
859 } 1380 }
860 1381
861 // 0..19 unassigned 1382 // 0..19 unassigned simple
862 // 24 reserved + unassigned (reserved values are not encodable) 1383 // 24 reserved + unassigned simple (reserved values are not encodable)
1384 // 28-30 unassigned misc
1385 // 31 break code
863 default: 1386 default:
864 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1387 ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
865 } 1388 }
866 1389
867 break; 1390 break;
868 } 1391 }
869 1392
872} 1395}
873 1396
874static SV * 1397static SV *
875decode_cbor (SV *string, CBOR *cbor, char **offset_return) 1398decode_cbor (SV *string, CBOR *cbor, char **offset_return)
876{ 1399{
877 dec_t dec; 1400 dec_t dec = { 0 };
878 SV *sv; 1401 SV *sv;
879 STRLEN len; 1402 STRLEN len;
880 char *data = SvPVbyte (string, len); 1403 char *data = SvPVbyte (string, len);
881 1404
882 if (len > cbor->max_size && cbor->max_size) 1405 if (len > cbor->max_size && cbor->max_size)
884 (unsigned long)len, (unsigned long)cbor->max_size); 1407 (unsigned long)len, (unsigned long)cbor->max_size);
885 1408
886 dec.cbor = *cbor; 1409 dec.cbor = *cbor;
887 dec.cur = (U8 *)data; 1410 dec.cur = (U8 *)data;
888 dec.end = (U8 *)data + len; 1411 dec.end = (U8 *)data + len;
889 dec.err = 0;
890 dec.depth = 0;
891 1412
892 sv = decode_sv (&dec); 1413 sv = decode_sv (&dec);
893 1414
894 if (offset_return) 1415 if (offset_return)
895 *offset_return = dec.cur; 1416 *offset_return = dec.cur;
898 if (dec.cur != dec.end && !dec.err) 1419 if (dec.cur != dec.end && !dec.err)
899 dec.err = "garbage after CBOR object"; 1420 dec.err = "garbage after CBOR object";
900 1421
901 if (dec.err) 1422 if (dec.err)
902 { 1423 {
1424 if (dec.shareable)
1425 {
1426 // need to break cyclic links, which would all be in shareable
1427 int i;
1428 SV **svp;
1429
1430 for (i = av_len (dec.shareable) + 1; i--; )
1431 if ((svp = av_fetch (dec.shareable, i, 0)))
1432 sv_setsv (*svp, &PL_sv_undef);
1433 }
1434
903 SvREFCNT_dec (sv); 1435 SvREFCNT_dec_NN (sv);
1436
1437 if (dec.err_sv)
1438 sv_2mortal (dec.err_sv);
1439
904 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur); 1440 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur);
905 } 1441 }
906 1442
907 sv = sv_2mortal (sv); 1443 sv = sv_2mortal (sv);
908 1444
909 return sv; 1445 return sv;
910} 1446}
911 1447
1448/////////////////////////////////////////////////////////////////////////////
1449// incremental parser
1450
1451#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
1452
1453// returns 0 for notyet, 1 for success or error
1454static int
1455incr_parse (CBOR *self, SV *cborstr)
1456{
1457 STRLEN cur;
1458 SvPV (cborstr, cur);
1459
1460 while (ecb_expect_true (self->incr_need <= cur))
1461 {
1462 // table of integer count bytes
1463 static I8 incr_len[MINOR_MASK + 1] = {
1464 0, 0, 0, 0, 0, 0, 0, 0,
1465 0, 0, 0, 0, 0, 0, 0, 0,
1466 0, 0, 0, 0, 0, 0, 0, 0,
1467 1, 2, 4, 8,-1,-1,-1,-2
1468 };
1469
1470 const U8 *p = SvPVX (cborstr) + self->incr_pos;
1471 U8 m = *p & MINOR_MASK;
1472 IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1473 I8 ilen = incr_len[m];
1474
1475 self->incr_need = self->incr_pos + 1;
1476
1477 if (ecb_expect_false (ilen < 0))
1478 {
1479 if (m != MINOR_INDEF)
1480 return 1; // error
1481
1482 if (*p == (MAJOR_MISC | MINOR_INDEF))
1483 {
1484 if (count >= 0)
1485 return 1; // error
1486
1487 count = 1;
1488 }
1489 else
1490 {
1491 av_push (self->incr_count, newSViv (-1)); //TODO: nest
1492 count = -1;
1493 }
1494 }
1495 else
1496 {
1497 self->incr_need += ilen;
1498 if (ecb_expect_false (self->incr_need > cur))
1499 return 0;
1500
1501 int major = *p >> MAJOR_SHIFT;
1502
1503 switch (major)
1504 {
1505 case MAJOR_TAG >> MAJOR_SHIFT:
1506 ++count; // tags merely prefix another value
1507 break;
1508
1509 case MAJOR_BYTES >> MAJOR_SHIFT:
1510 case MAJOR_TEXT >> MAJOR_SHIFT:
1511 case MAJOR_ARRAY >> MAJOR_SHIFT:
1512 case MAJOR_MAP >> MAJOR_SHIFT:
1513 {
1514 UV len;
1515
1516 if (ecb_expect_false (ilen))
1517 {
1518 len = 0;
1519
1520 do {
1521 len = (len << 8) | *++p;
1522 } while (--ilen);
1523 }
1524 else
1525 len = m;
1526
1527 switch (major)
1528 {
1529 case MAJOR_BYTES >> MAJOR_SHIFT:
1530 case MAJOR_TEXT >> MAJOR_SHIFT:
1531 self->incr_need += len;
1532 if (ecb_expect_false (self->incr_need > cur))
1533 return 0;
1534
1535 break;
1536
1537 case MAJOR_MAP >> MAJOR_SHIFT:
1538 len <<= 1;
1539 case MAJOR_ARRAY >> MAJOR_SHIFT:
1540 if (len)
1541 {
1542 av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
1543 count = len + 1;
1544 }
1545 break;
1546 }
1547 }
1548 }
1549 }
1550
1551 self->incr_pos = self->incr_need;
1552
1553 if (count > 0)
1554 {
1555 while (!--count)
1556 {
1557 if (!AvFILLp (self->incr_count))
1558 return 1; // done
1559
1560 SvREFCNT_dec_NN (av_pop (self->incr_count));
1561 count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1562 }
1563
1564 SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
1565 }
1566 }
1567
1568 return 0;
1569}
1570
1571
912///////////////////////////////////////////////////////////////////////////// 1572/////////////////////////////////////////////////////////////////////////////
913// XS interface functions 1573// XS interface functions
914 1574
915MODULE = CBOR::XS PACKAGE = CBOR::XS 1575MODULE = CBOR::XS PACKAGE = CBOR::XS
916 1576
924 1584
925 types_true = get_bool ("Types::Serialiser::true" ); 1585 types_true = get_bool ("Types::Serialiser::true" );
926 types_false = get_bool ("Types::Serialiser::false"); 1586 types_false = get_bool ("Types::Serialiser::false");
927 types_error = get_bool ("Types::Serialiser::error"); 1587 types_error = get_bool ("Types::Serialiser::error");
928 1588
1589 default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1590
929 sv_cbor = newSVpv ("CBOR", 0); 1591 sv_cbor = newSVpv ("CBOR", 0);
930 SvREADONLY_on (sv_cbor); 1592 SvREADONLY_on (sv_cbor);
1593
1594 assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1));
931} 1595}
932 1596
933PROTOTYPES: DISABLE 1597PROTOTYPES: DISABLE
934 1598
935void CLONE (...) 1599void CLONE (...)
953 1617
954void shrink (CBOR *self, int enable = 1) 1618void shrink (CBOR *self, int enable = 1)
955 ALIAS: 1619 ALIAS:
956 shrink = F_SHRINK 1620 shrink = F_SHRINK
957 allow_unknown = F_ALLOW_UNKNOWN 1621 allow_unknown = F_ALLOW_UNKNOWN
1622 allow_sharing = F_ALLOW_SHARING
1623 allow_cycles = F_ALLOW_CYCLES
1624 forbid_objects = F_FORBID_OBJECTS
1625 pack_strings = F_PACK_STRINGS
1626 text_keys = F_TEXT_KEYS
1627 text_strings = F_TEXT_STRINGS
1628 validate_utf8 = F_VALIDATE_UTF8
958 PPCODE: 1629 PPCODE:
959{ 1630{
960 if (enable) 1631 if (enable)
961 self->flags |= ix; 1632 self->flags |= ix;
962 else 1633 else
967 1638
968void get_shrink (CBOR *self) 1639void get_shrink (CBOR *self)
969 ALIAS: 1640 ALIAS:
970 get_shrink = F_SHRINK 1641 get_shrink = F_SHRINK
971 get_allow_unknown = F_ALLOW_UNKNOWN 1642 get_allow_unknown = F_ALLOW_UNKNOWN
1643 get_allow_sharing = F_ALLOW_SHARING
1644 get_allow_cycles = F_ALLOW_CYCLES
1645 get_forbid_objects = F_FORBID_OBJECTS
1646 get_pack_strings = F_PACK_STRINGS
1647 get_text_keys = F_TEXT_KEYS
1648 get_text_strings = F_TEXT_STRINGS
1649 get_validate_utf8 = F_VALIDATE_UTF8
972 PPCODE: 1650 PPCODE:
973 XPUSHs (boolSV (self->flags & ix)); 1651 XPUSHs (boolSV (self->flags & ix));
974 1652
975void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1653void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
976 PPCODE: 1654 PPCODE:
989 XPUSHs (ST (0)); 1667 XPUSHs (ST (0));
990 1668
991int get_max_size (CBOR *self) 1669int get_max_size (CBOR *self)
992 CODE: 1670 CODE:
993 RETVAL = self->max_size; 1671 RETVAL = self->max_size;
1672 OUTPUT:
1673 RETVAL
1674
1675void filter (CBOR *self, SV *filter = 0)
1676 PPCODE:
1677 SvREFCNT_dec (self->filter);
1678 self->filter = filter ? newSVsv (filter) : filter;
1679 XPUSHs (ST (0));
1680
1681SV *get_filter (CBOR *self)
1682 CODE:
1683 RETVAL = self->filter ? self->filter : NEWSV (0, 0);
994 OUTPUT: 1684 OUTPUT:
995 RETVAL 1685 RETVAL
996 1686
997void encode (CBOR *self, SV *scalar) 1687void encode (CBOR *self, SV *scalar)
998 PPCODE: 1688 PPCODE:
1013 EXTEND (SP, 2); 1703 EXTEND (SP, 2);
1014 PUSHs (sv); 1704 PUSHs (sv);
1015 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1705 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1016} 1706}
1017 1707
1708void incr_parse (CBOR *self, SV *cborstr)
1709 ALIAS:
1710 incr_parse_multiple = 1
1711 PPCODE:
1712{
1713 if (SvUTF8 (cborstr))
1714 sv_utf8_downgrade (cborstr, 0);
1715
1716 if (!self->incr_count)
1717 {
1718 self->incr_count = newAV ();
1719 self->incr_pos = 0;
1720 self->incr_need = 1;
1721
1722 av_push (self->incr_count, newSViv (1));
1723 }
1724
1725 do
1726 {
1727 if (!incr_parse (self, cborstr))
1728 {
1729 if (self->incr_need > self->max_size && self->max_size)
1730 croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
1731 (unsigned long)self->incr_need, (unsigned long)self->max_size);
1732
1733 break;
1734 }
1735
1736 SV *sv;
1737 char *offset;
1738
1739 PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
1740 XPUSHs (sv);
1741
1742 sv_chop (cborstr, offset);
1743
1744 av_clear (self->incr_count);
1745 av_push (self->incr_count, newSViv (1));
1746
1747 self->incr_pos = 0;
1748 self->incr_need = self->incr_pos + 1;
1749 }
1750 while (ix);
1751}
1752
1753void incr_reset (CBOR *self)
1754 CODE:
1755{
1756 SvREFCNT_dec (self->incr_count);
1757 self->incr_count = 0;
1758}
1759
1760void DESTROY (CBOR *self)
1761 PPCODE:
1762 cbor_free (self);
1763
1018PROTOTYPES: ENABLE 1764PROTOTYPES: ENABLE
1019 1765
1020void encode_cbor (SV *scalar) 1766void encode_cbor (SV *scalar)
1767 ALIAS:
1768 encode_cbor = 0
1769 encode_cbor_sharing = F_ALLOW_SHARING
1021 PPCODE: 1770 PPCODE:
1022{ 1771{
1023 CBOR cbor; 1772 CBOR cbor;
1024 cbor_init (&cbor); 1773 cbor_init (&cbor);
1774 cbor.flags |= ix;
1025 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1775 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1026 XPUSHs (scalar); 1776 XPUSHs (scalar);
1027} 1777}
1028 1778
1029void decode_cbor (SV *cborstr) 1779void decode_cbor (SV *cborstr)
1033 cbor_init (&cbor); 1783 cbor_init (&cbor);
1034 PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN; 1784 PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN;
1035 XPUSHs (cborstr); 1785 XPUSHs (cborstr);
1036} 1786}
1037 1787
1788#ifdef __AFL_COMPILER
1789
1790void
1791afl_init ()
1792 CODE:
1793 __AFL_INIT ();
1794
1795int
1796afl_loop (unsigned int count = 10000)
1797 CODE:
1798 RETVAL = __AFL_LOOP (count);
1799 OUTPUT:
1800 RETVAL
1801
1802#endif
1803

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines