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

Comparing cvsroot/CBOR-XS/XS.xs (file contents):
Revision 1.20 by root, Wed Nov 20 11:06:42 2013 UTC vs.
Revision 1.69 by root, Mon Nov 30 20:38:25 2020 UTC

6#include <string.h> 6#include <string.h>
7#include <stdlib.h> 7#include <stdlib.h>
8#include <stdio.h> 8#include <stdio.h>
9#include <limits.h> 9#include <limits.h>
10#include <float.h> 10#include <float.h>
11#include <inttypes.h>
11 12
13#define ECB_NO_THREADS 1
12#include "ecb.h" 14#include "ecb.h"
13 15
14// compatibility with perl <5.18 16// compatibility with perl <5.18
15#ifndef HvNAMELEN_get 17#ifndef HvNAMELEN_get
16# define HvNAMELEN_get(hv) strlen (HvNAME (hv)) 18# define HvNAMELEN_get(hv) strlen (HvNAME (hv))
19# define HvNAMELEN(hv) HvNAMELEN_get (hv) 21# define HvNAMELEN(hv) HvNAMELEN_get (hv)
20#endif 22#endif
21#ifndef HvNAMEUTF8 23#ifndef HvNAMEUTF8
22# define HvNAMEUTF8(hv) 0 24# define HvNAMEUTF8(hv) 0
23#endif 25#endif
26#ifndef SvREFCNT_inc_NN
27# define SvREFCNT_inc_NN(sv) SvREFCNT_inc (sv)
28#endif
29#ifndef SvREFCNT_dec_NN
30# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
31#endif
32
33// known major and minor types
34enum cbor_type
35{
36 MAJOR_SHIFT = 5,
37 MINOR_MASK = 0x1f,
38
39 MAJOR_POS_INT = 0 << MAJOR_SHIFT,
40 MAJOR_NEG_INT = 1 << MAJOR_SHIFT,
41 MAJOR_BYTES = 2 << MAJOR_SHIFT,
42 MAJOR_TEXT = 3 << MAJOR_SHIFT,
43 MAJOR_ARRAY = 4 << MAJOR_SHIFT,
44 MAJOR_MAP = 5 << MAJOR_SHIFT,
45 MAJOR_TAG = 6 << MAJOR_SHIFT,
46 MAJOR_MISC = 7 << MAJOR_SHIFT,
47
48 // INT/STRING/ARRAY/MAP subtypes
49 LENGTH_EXT1 = 24,
50 LENGTH_EXT2 = 25,
51 LENGTH_EXT4 = 26,
52 LENGTH_EXT8 = 27,
53
54 // SIMPLE types (effectively MISC subtypes)
55 SIMPLE_FALSE = 20,
56 SIMPLE_TRUE = 21,
57 SIMPLE_NULL = 22,
58 SIMPLE_UNDEF = 23,
59
60 // MISC subtype (unused)
61 MISC_EXT1 = 24,
62 MISC_FLOAT16 = 25,
63 MISC_FLOAT32 = 26,
64 MISC_FLOAT64 = 27,
65
66 // BYTES/TEXT/ARRAY/MAP
67 MINOR_INDEF = 31,
68};
24 69
25// known tags 70// known tags
26enum cbor_tag 71enum cbor_tag
27{ 72{
28 // inofficial extensions (pending iana registration) 73 // extensions
74 CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
29 CBOR_TAG_PERL_OBJECT = 24, // http://cbor.schmorp.de/perl-object 75 CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
30 CBOR_TAG_GENERIC_OBJECT = 25, // http://cbor.schmorp.de/generic-object 76 CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
31 CBOR_TAG_VALUE_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_INT = 1,
109 AS_BYTES = 2,
110 AS_TEXT = 3,
111 AS_FLOAT16 = 4,
112 AS_FLOAT32 = 5,
113 AS_FLOAT64 = 6,
114 // possibly future enhancements: (generic) float, (generic) string
115};
116
59#define F_SHRINK 0x00000001UL 117#define F_SHRINK 0x00000001UL
60#define F_ALLOW_UNKNOWN 0x00000002UL 118#define F_ALLOW_UNKNOWN 0x00000002UL
61#define F_ALLOW_SHARING 0x00000004UL //TODO 119#define F_ALLOW_SHARING 0x00000004UL
120#define F_ALLOW_CYCLES 0x00000008UL
121#define F_FORBID_OBJECTS 0x00000010UL
62#define F_DEDUP_STRINGS 0x00000008UL //TODO 122#define F_PACK_STRINGS 0x00000020UL
63#define F_DEDUP_KEYS 0x00000010UL //TODO 123#define F_TEXT_KEYS 0x00000040UL
124#define F_TEXT_STRINGS 0x00000080UL
125#define F_VALIDATE_UTF8 0x00000100UL
64 126
65#define INIT_SIZE 32 // initial scalar size to be allocated 127#define INIT_SIZE 32 // initial scalar size to be allocated
66 128
67#define SB do { 129#define SB do {
68#define SE } while (0) 130#define SE } while (0)
80# define CBOR_SLOW 0 142# define CBOR_SLOW 0
81# define CBOR_STASH cbor_stash 143# define CBOR_STASH cbor_stash
82#endif 144#endif
83 145
84static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS:: 146static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS::
85static SV *types_true, *types_false, *types_error, *sv_cbor; 147static SV *types_true, *types_false, *types_error, *sv_cbor, *default_filter;
86 148
87typedef struct { 149typedef struct {
88 U32 flags; 150 U32 flags;
89 U32 max_depth; 151 U32 max_depth;
90 STRLEN max_size; 152 STRLEN max_size;
153 SV *filter;
154
155 // for the incremental parser
156 STRLEN incr_pos; // the current offset into the text
157 STRLEN incr_need; // minimum bytes needed to decode
158 AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef.
91} CBOR; 159} CBOR;
92 160
93ecb_inline void 161ecb_inline void
94cbor_init (CBOR *cbor) 162cbor_init (CBOR *cbor)
95{ 163{
96 Zero (cbor, 1, CBOR); 164 Zero (cbor, 1, CBOR);
97 cbor->max_depth = 512; 165 cbor->max_depth = 512;
166}
167
168ecb_inline void
169cbor_free (CBOR *cbor)
170{
171 SvREFCNT_dec (cbor->filter);
172 SvREFCNT_dec (cbor->incr_count);
98} 173}
99 174
100///////////////////////////////////////////////////////////////////////////// 175/////////////////////////////////////////////////////////////////////////////
101// utility functions 176// utility functions
102 177
122 SvPV_shrink_to_cur (sv); 197 SvPV_shrink_to_cur (sv);
123#elif defined (SvPV_renew) 198#elif defined (SvPV_renew)
124 SvPV_renew (sv, SvCUR (sv) + 1); 199 SvPV_renew (sv, SvCUR (sv) + 1);
125#endif 200#endif
126 } 201 }
202}
203
204// minimum length of a string to be registered for stringref
205ecb_inline int
206minimum_string_length (UV idx)
207{
208 return idx <= 23 ? 3
209 : idx <= 0xffU ? 4
210 : idx <= 0xffffU ? 5
211 : idx <= 0xffffffffU ? 7
212 : 11;
127} 213}
128 214
129///////////////////////////////////////////////////////////////////////////// 215/////////////////////////////////////////////////////////////////////////////
130// encoder 216// encoder
131 217
144} enc_t; 230} enc_t;
145 231
146ecb_inline void 232ecb_inline void
147need (enc_t *enc, STRLEN len) 233need (enc_t *enc, STRLEN len)
148{ 234{
149 if (ecb_expect_false (enc->cur + len >= enc->end)) 235 if (ecb_expect_false ((uintptr_t)(enc->end - enc->cur) < len))
150 { 236 {
151 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv); 237 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv);
152 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1); 238 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1);
153 enc->cur = SvPVX (enc->sv) + cur; 239 enc->cur = SvPVX (enc->sv) + cur;
154 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1; 240 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;
160{ 246{
161 need (enc, 1); 247 need (enc, 1);
162 *enc->cur++ = ch; 248 *enc->cur++ = ch;
163} 249}
164 250
251// used for tags, intregers, element counts and so on
165static void 252static void
166encode_uint (enc_t *enc, int major, UV len) 253encode_uint (enc_t *enc, int major, UV len)
167{ 254{
168 need (enc, 9); 255 need (enc, 9);
169 256
170 if (len < 24) 257 if (ecb_expect_true (len < LENGTH_EXT1))
171 *enc->cur++ = major | len; 258 *enc->cur++ = major | len;
172 else if (len <= 0xff) 259 else if (ecb_expect_true (len <= 0xffU))
173 { 260 {
174 *enc->cur++ = major | 24; 261 *enc->cur++ = major | LENGTH_EXT1;
175 *enc->cur++ = len; 262 *enc->cur++ = len;
176 } 263 }
177 else if (len <= 0xffff) 264 else if (len <= 0xffffU)
178 { 265 {
179 *enc->cur++ = major | 25; 266 *enc->cur++ = major | LENGTH_EXT2;
180 *enc->cur++ = len >> 8; 267 *enc->cur++ = len >> 8;
181 *enc->cur++ = len; 268 *enc->cur++ = len;
182 } 269 }
183 else if (len <= 0xffffffff) 270 else if (len <= 0xffffffffU)
184 { 271 {
185 *enc->cur++ = major | 26; 272 *enc->cur++ = major | LENGTH_EXT4;
186 *enc->cur++ = len >> 24; 273 *enc->cur++ = len >> 24;
187 *enc->cur++ = len >> 16; 274 *enc->cur++ = len >> 16;
188 *enc->cur++ = len >> 8; 275 *enc->cur++ = len >> 8;
189 *enc->cur++ = len; 276 *enc->cur++ = len;
190 } 277 }
191 else 278 else
192 { 279 {
193 *enc->cur++ = major | 27; 280 *enc->cur++ = major | LENGTH_EXT8;
194 *enc->cur++ = len >> 56; 281 *enc->cur++ = len >> 56;
195 *enc->cur++ = len >> 48; 282 *enc->cur++ = len >> 48;
196 *enc->cur++ = len >> 40; 283 *enc->cur++ = len >> 40;
197 *enc->cur++ = len >> 32; 284 *enc->cur++ = len >> 32;
198 *enc->cur++ = len >> 24; 285 *enc->cur++ = len >> 24;
200 *enc->cur++ = len >> 8; 287 *enc->cur++ = len >> 8;
201 *enc->cur++ = len; 288 *enc->cur++ = len;
202 } 289 }
203} 290}
204 291
205static 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
206encode_str (enc_t *enc, int utf8, char *str, STRLEN len) 312encode_str_utf8 (enc_t *enc, int utf8, char *str, STRLEN len)
207{ 313{
208 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);
209 need (enc, len); 344 need (enc, len);
210 memcpy (enc->cur, str, len); 345 memcpy (enc->cur, str, len);
211 enc->cur += len; 346 enc->cur += len;
212} 347}
213 348
214ecb_inline void 349ecb_inline void
350encode_strref (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
351{
352 if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
353 {
354 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
355
356 if (SvOK (*svp))
357 {
358 // already registered, use stringref
359 encode_tag (enc, CBOR_TAG_STRINGREF);
360 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
361 return;
362 }
363 else if (len >= minimum_string_length (enc->stringref_idx))
364 {
365 // register only
366 sv_setuv (*svp, enc->stringref_idx);
367 ++enc->stringref_idx;
368 }
369 }
370
371 encode_str (enc, upgrade_utf8, utf8, str, len);
372}
373
374ecb_inline void
215encode_tag (enc_t *enc, UV tag) 375encode_float16 (enc_t *enc, NV nv)
216{ 376{
217 encode_uint (enc, 0xc0, tag); 377 need (enc, 1+2);
378
379 *enc->cur++ = MAJOR_MISC | MISC_FLOAT16;
380
381 uint16_t fp = ecb_float_to_binary16 (nv);
382
383 if (!ecb_big_endian ())
384 fp = ecb_bswap16 (fp);
385
386 memcpy (enc->cur, &fp, 2);
387 enc->cur += 2;
388}
389
390ecb_inline void
391encode_float32 (enc_t *enc, NV nv)
392{
393 need (enc, 1+4);
394
395 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
396
397 uint32_t fp = ecb_float_to_binary32 (nv);
398
399 if (!ecb_big_endian ())
400 fp = ecb_bswap32 (fp);
401
402 memcpy (enc->cur, &fp, 4);
403 enc->cur += 4;
404}
405
406ecb_inline void
407encode_float64 (enc_t *enc, NV nv)
408{
409 need (enc, 1+8);
410
411 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
412
413 uint64_t fp = ecb_double_to_binary64 (nv);
414
415 if (!ecb_big_endian ())
416 fp = ecb_bswap64 (fp);
417
418 memcpy (enc->cur, &fp, 8);
419 enc->cur += 8;
420}
421
422ecb_inline void
423encode_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 }
218} 469}
219 470
220static void encode_sv (enc_t *enc, SV *sv); 471static void encode_sv (enc_t *enc, SV *sv);
221 472
222static void 473static void
227 if (enc->depth >= enc->cbor.max_depth) 478 if (enc->depth >= enc->cbor.max_depth)
228 croak (ERR_NESTING_EXCEEDED); 479 croak (ERR_NESTING_EXCEEDED);
229 480
230 ++enc->depth; 481 ++enc->depth;
231 482
232 encode_uint (enc, 0x80, len + 1); 483 encode_uint (enc, MAJOR_ARRAY, len + 1);
233 484
485 if (ecb_expect_false (SvMAGICAL (av)))
234 for (i = 0; i <= len; ++i) 486 for (i = 0; i <= len; ++i)
235 { 487 {
236 SV **svp = av_fetch (av, i, 0); 488 SV **svp = av_fetch (av, i, 0);
237 encode_sv (enc, svp ? *svp : &PL_sv_undef); 489 encode_sv (enc, svp ? *svp : &PL_sv_undef);
238 } 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 }
239 497
240 --enc->depth; 498 --enc->depth;
241}
242
243ecb_inline void
244encode_he (enc_t *enc, HE *he)
245{
246 if (HeKLEN (he) == HEf_SVKEY)
247 encode_sv (enc, HeSVKEY (he));
248 else
249 encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
250} 499}
251 500
252static void 501static void
253encode_hv (enc_t *enc, HV *hv) 502encode_hv (enc_t *enc, HV *hv)
254{ 503{
260 ++enc->depth; 509 ++enc->depth;
261 510
262 int pairs = hv_iterinit (hv); 511 int pairs = hv_iterinit (hv);
263 int mg = SvMAGICAL (hv); 512 int mg = SvMAGICAL (hv);
264 513
265 if (mg) 514 if (ecb_expect_false (mg))
266 encode_ch (enc, 0xa0 | 31); 515 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
267 else 516 else
268 encode_uint (enc, 0xa0, pairs); 517 encode_uint (enc, MAJOR_MAP, pairs);
269 518
270 while ((he = hv_iternext (hv))) 519 while ((he = hv_iternext (hv)))
271 { 520 {
272 if (ecb_expect_false (enc->cbor.flags & (F_DEDUP_STRINGS | F_DEDUP_KEYS)))
273 {
274 SV **svp;
275
276 if (HeKLEN (he) == HEf_SVKEY) 521 if (HeKLEN (he) == HEf_SVKEY)
277 svp = hv_fetch_ent (enc->stringref[!! SvUTF8 (HeSVKEY (he))], HeSVKEY (he) , 1, 0);//TODO return HE :/ 522 encode_sv (enc, HeSVKEY (he));
278 else
279 svp = hv_fetch (enc->stringref[!! HeKUTF8 (he) ], HeKEY (he), HeKLEN (he), 1);
280
281 if (SvOK (*svp))
282 {
283 encode_tag (enc, CBOR_TAG_STRINGREF);
284 encode_uint (enc, 0x00, SvUV (*svp));
285 }
286 else
287 {
288 sv_setuv (*svp, enc->stringref_idx);
289 ++enc->stringref_idx;
290 encode_he (enc, he);
291 }
292 }
293 else 523 else
294 encode_he (enc, he); 524 encode_strref (enc, enc->cbor.flags & (F_TEXT_KEYS | F_TEXT_STRINGS), HeKUTF8 (he), HeKEY (he), HeKLEN (he));
295 525
296 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));
297 } 527 }
298 528
299 if (mg) 529 if (ecb_expect_false (mg))
300 encode_ch (enc, 0xe0 | 31); 530 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
301 531
302 --enc->depth; 532 --enc->depth;
303} 533}
304 534
305// encode objects, arrays and special \0=false and \1=true values. 535// encode objects, arrays and special \0=false and \1=true values.
306static void 536static void
307encode_rv (enc_t *enc, SV *sv) 537encode_rv (enc_t *enc, SV *sv)
308{ 538{
309 SvGETMAGIC (sv); 539 SvGETMAGIC (sv);
310
311 if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)
312 && ecb_expect_false (SvREFCNT (sv) > 1))
313 {
314 if (!enc->shareable)
315 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
316
317 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
318
319 if (SvOK (*svp))
320 {
321 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
322 encode_uint (enc, 0x00, SvUV (*svp));
323 return;
324 }
325 else
326 {
327 sv_setuv (*svp, enc->shareable_idx);
328 ++enc->shareable_idx;
329 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
330 }
331 }
332 540
333 svtype svt = SvTYPE (sv); 541 svtype svt = SvTYPE (sv);
334 542
335 if (ecb_expect_false (SvOBJECT (sv))) 543 if (ecb_expect_false (SvOBJECT (sv)))
336 { 544 {
343 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 551 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
344 ? cbor_tagged_stash 552 ? cbor_tagged_stash
345 : gv_stashpv ("CBOR::XS::Tagged" , 1); 553 : gv_stashpv ("CBOR::XS::Tagged" , 1);
346 554
347 HV *stash = SvSTASH (sv); 555 HV *stash = SvSTASH (sv);
348 GV *method;
349 556
350 if (stash == boolean_stash) 557 if (stash == boolean_stash)
351 encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20); 558 {
559 encode_bool (enc, SvIV (sv));
560 return;
561 }
352 else if (stash == error_stash) 562 else if (stash == error_stash)
353 encode_ch (enc, 0xe0 | 23); 563 {
564 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
565 return;
566 }
354 else if (stash == tagged_stash) 567 else if (stash == tagged_stash)
355 { 568 {
356 if (svt != SVt_PVAV) 569 if (svt != SVt_PVAV)
357 croak ("encountered CBOR::XS::Tagged object that isn't an array"); 570 croak ("encountered CBOR::XS::Tagged object that isn't an array");
358 571
572 switch (av_len ((AV *)sv))
573 {
574 case 2-1:
575 // actually a tagged value
359 encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); 576 encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
360 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))
361 } 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))));
362 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) 623 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
363 { 624 {
364 dSP; 625 dSP;
365 626
366 ENTER; SAVETMPS; PUSHMARK (SP); 627 ENTER; SAVETMPS;
628 PUSHMARK (SP);
367 // 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
368 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 630 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
369 631
370 PUTBACK; 632 PUTBACK;
371 // G_SCALAR ensures that return value is 1 633 // G_SCALAR ensures that return value is 1
384 } 646 }
385 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0) 647 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
386 { 648 {
387 dSP; 649 dSP;
388 650
389 ENTER; SAVETMPS; PUSHMARK (SP); 651 ENTER; SAVETMPS;
652 PUSHMARK (SP);
390 EXTEND (SP, 2); 653 EXTEND (SP, 2);
391 // 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
392 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 655 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
393 PUSHs (sv_cbor); 656 PUSHs (sv_cbor);
394 657
399 // catch this surprisingly common error 662 // catch this surprisingly common error
400 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) 663 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
401 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));
402 665
403 encode_tag (enc, CBOR_TAG_PERL_OBJECT); 666 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
404 encode_uint (enc, 0x80, count + 1); 667 encode_uint (enc, MAJOR_ARRAY, count + 1);
405 encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); 668 encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
406 669
407 while (count) 670 {
671 int i;
672
673 for (i = 0; i < count; ++i)
408 encode_sv (enc, SP[1 - count--]); 674 encode_sv (enc, SP[i + 1 - count]);
675
676 SP -= count;
677 }
409 678
410 PUTBACK; 679 PUTBACK;
411 680
412 FREETMPS; LEAVE; 681 FREETMPS; LEAVE;
413 } 682 }
431{ 700{
432 double nv = SvNVX (sv); 701 double nv = SvNVX (sv);
433 702
434 need (enc, 9); 703 need (enc, 9);
435 704
436 if (ecb_expect_false (nv == (U32)nv)) 705 if (ecb_expect_false (nv == (NV)(U32)nv))
437 encode_uint (enc, 0x00, (U32)nv); 706 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
438 //TODO: maybe I32? 707 //TODO: maybe I32?
439 else if (ecb_expect_false (nv == (float)nv)) 708 else if (ecb_expect_false (nv == (float)nv))
440 { 709 encode_float32 (enc, nv);
441 uint32_t fp = ecb_float_to_binary32 (nv);
442
443 *enc->cur++ = 0xe0 | 26;
444
445 if (!ecb_big_endian ())
446 fp = ecb_bswap32 (fp);
447
448 memcpy (enc->cur, &fp, 4);
449 enc->cur += 4;
450 }
451 else 710 else
452 { 711 encode_float64 (enc, nv);
453 uint64_t fp = ecb_double_to_binary64 (nv);
454
455 *enc->cur++ = 0xe0 | 27;
456
457 if (!ecb_big_endian ())
458 fp = ecb_bswap64 (fp);
459
460 memcpy (enc->cur, &fp, 8);
461 enc->cur += 8;
462 }
463} 712}
464 713
465static void 714static void
466encode_sv (enc_t *enc, SV *sv) 715encode_sv (enc_t *enc, SV *sv)
467{ 716{
469 718
470 if (SvPOKp (sv)) 719 if (SvPOKp (sv))
471 { 720 {
472 STRLEN len; 721 STRLEN len;
473 char *str = SvPV (sv, len); 722 char *str = SvPV (sv, len);
474 encode_str (enc, SvUTF8 (sv), str, len); 723 encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len);
475 } 724 }
476 else if (SvNOKp (sv)) 725 else if (SvNOKp (sv))
477 encode_nv (enc, sv); 726 encode_nv (enc, sv);
478 else if (SvIOKp (sv)) 727 else if (SvIOKp (sv))
479 { 728 encode_int (enc, sv);
480 if (SvIsUV (sv))
481 encode_uint (enc, 0x00, SvUVX (sv));
482 else if (SvIVX (sv) >= 0)
483 encode_uint (enc, 0x00, SvIVX (sv));
484 else
485 encode_uint (enc, 0x20, -(SvIVX (sv) + 1));
486 }
487 else if (SvROK (sv)) 729 else if (SvROK (sv))
488 encode_rv (enc, SvRV (sv)); 730 encode_rv (enc, SvRV (sv));
489 else if (!SvOK (sv)) 731 else if (!SvOK (sv))
490 encode_ch (enc, 0xe0 | 22); 732 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
491 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 733 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
492 encode_ch (enc, 0xe0 | 23); 734 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
493 else 735 else
494 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",
495 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 737 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
496} 738}
497 739
498static SV * 740static SV *
499encode_cbor (SV *scalar, CBOR *cbor) 741encode_cbor (SV *scalar, CBOR *cbor)
500{ 742{
501 enc_t enc = { }; 743 enc_t enc = { 0 };
502 744
503 enc.cbor = *cbor; 745 enc.cbor = *cbor;
504 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 746 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
505 enc.cur = SvPVX (enc.sv); 747 enc.cur = SvPVX (enc.sv);
506 enc.end = SvEND (enc.sv); 748 enc.end = SvEND (enc.sv);
507 749
508 SvPOK_only (enc.sv); 750 SvPOK_only (enc.sv);
509 751
510 if (cbor->flags & (F_DEDUP_STRINGS | F_DEDUP_KEYS)) 752 if (cbor->flags & F_PACK_STRINGS)
511 { 753 {
512 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE); 754 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
513 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ()); 755 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
514 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ()); 756 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
515 } 757 }
537 CBOR cbor; 779 CBOR cbor;
538 U32 depth; // recursion depth 780 U32 depth; // recursion depth
539 U32 maxdepth; // recursion depth limit 781 U32 maxdepth; // recursion depth limit
540 AV *shareable; 782 AV *shareable;
541 AV *stringref; 783 AV *stringref;
784 SV *decode_tagged;
785 SV *err_sv; // optional sv for error, needs to be freed
542} dec_t; 786} dec_t;
543 787
544#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);
545 795
546#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;
547 799
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
548#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)
549#define DEC_DEC_DEPTH --dec->depth 825#define DEC_DEC_DEPTH --dec->depth
550 826
551static UV 827static UV
552decode_uint (dec_t *dec) 828decode_uint (dec_t *dec)
553{ 829{
554 switch (*dec->cur & 31) 830 U8 m = *dec->cur & MINOR_MASK;
555 { 831 ++dec->cur;
556 case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
557 case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
558 case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
559 return *dec->cur++ & 31;
560 832
561 case 24: 833 if (ecb_expect_true (m < LENGTH_EXT1))
834 return m;
835 else if (ecb_expect_true (m == LENGTH_EXT1))
836 {
562 WANT (2); 837 WANT (1);
563 dec->cur += 2; 838 dec->cur += 1;
564 return dec->cur[-1]; 839 return dec->cur[-1];
565 840 }
566 case 25: 841 else if (ecb_expect_true (m == LENGTH_EXT2))
842 {
567 WANT (3); 843 WANT (2);
568 dec->cur += 3; 844 dec->cur += 2;
569 return (((UV)dec->cur[-2]) << 8) 845 return (((UV)dec->cur[-2]) << 8)
570 | ((UV)dec->cur[-1]); 846 | ((UV)dec->cur[-1]);
571 847 }
572 case 26: 848 else if (ecb_expect_true (m == LENGTH_EXT4))
849 {
573 WANT (5); 850 WANT (4);
574 dec->cur += 5; 851 dec->cur += 4;
575 return (((UV)dec->cur[-4]) << 24) 852 return (((UV)dec->cur[-4]) << 24)
576 | (((UV)dec->cur[-3]) << 16) 853 | (((UV)dec->cur[-3]) << 16)
577 | (((UV)dec->cur[-2]) << 8) 854 | (((UV)dec->cur[-2]) << 8)
578 | ((UV)dec->cur[-1]); 855 | ((UV)dec->cur[-1]);
579 856 }
580 case 27: 857 else if (ecb_expect_true (m == LENGTH_EXT8))
858 {
581 WANT (9); 859 WANT (8);
582 dec->cur += 9; 860 dec->cur += 8;
861
862 return
863#if UVSIZE < 8
864 0
865#else
583 return (((UV)dec->cur[-8]) << 56) 866 (((UV)dec->cur[-8]) << 56)
584 | (((UV)dec->cur[-7]) << 48) 867 | (((UV)dec->cur[-7]) << 48)
585 | (((UV)dec->cur[-6]) << 40) 868 | (((UV)dec->cur[-6]) << 40)
586 | (((UV)dec->cur[-5]) << 32) 869 | (((UV)dec->cur[-5]) << 32)
870#endif
587 | (((UV)dec->cur[-4]) << 24) 871 | (((UV)dec->cur[-4]) << 24)
588 | (((UV)dec->cur[-3]) << 16) 872 | (((UV)dec->cur[-3]) << 16)
589 | (((UV)dec->cur[-2]) << 8) 873 | (((UV)dec->cur[-2]) << 8)
590 | ((UV)dec->cur[-1]); 874 | ((UV)dec->cur[-1]);
591 875 }
592 default: 876 else
593 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 877 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
594 }
595 878
596fail: 879fail:
597 return 0; 880 return 0;
598} 881}
599 882
604{ 887{
605 AV *av = newAV (); 888 AV *av = newAV ();
606 889
607 DEC_INC_DEPTH; 890 DEC_INC_DEPTH;
608 891
609 if ((*dec->cur & 31) == 31) 892 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
610 { 893 {
611 ++dec->cur; 894 ++dec->cur;
612 895
613 for (;;) 896 for (;;)
614 { 897 {
615 WANT (1); 898 WANT (1);
616 899
617 if (*dec->cur == (0xe0 | 31)) 900 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
618 { 901 {
619 ++dec->cur; 902 ++dec->cur;
620 break; 903 break;
621 } 904 }
622 905
623 av_push (av, decode_sv (dec)); 906 av_push (av, decode_sv (dec));
624 } 907 }
625 } 908 }
626 else 909 else
627 { 910 {
628 int i, len = decode_uint (dec); 911 UV i, len = decode_uint (dec);
629 912
913 WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
630 av_fill (av, len - 1); 914 av_fill (av, len - 1);
631 915
632 for (i = 0; i < len; ++i) 916 for (i = 0; i < len; ++i)
633 AvARRAY (av)[i] = decode_sv (dec); 917 AvARRAY (av)[i] = decode_sv (dec);
634 } 918 }
635 919
636 DEC_DEC_DEPTH; 920 DEC_DEC_DEPTH;
637 return newRV_noinc ((SV *)av); 921 return newRV_noinc ((SV *)av);
638 922
639fail: 923fail:
640 SvREFCNT_dec (av); 924 SvREFCNT_dec_NN (av);
641 DEC_DEC_DEPTH; 925 DEC_DEC_DEPTH;
642 return &PL_sv_undef; 926 return &PL_sv_undef;
643} 927}
644 928
645static void 929static void
646decode_he (dec_t *dec, HV *hv) 930decode_he (dec_t *dec, HV *hv)
647{ 931{
648 // for speed reasons, we specialcase single-string 932 // for speed reasons, we specialcase single-string
649 // byte or utf-8 strings as keys. 933 // byte or utf-8 strings as keys, but only when !stringref
650 934
651 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))
652 { 937 {
653 I32 len = decode_uint (dec); 938 STRLEN len = decode_uint (dec);
654 char *key = (char *)dec->cur; 939 char *key = (char *)dec->cur;
655 940
941 WANT (len);
656 dec->cur += len; 942 dec->cur += len;
657 943
658 if (ecb_expect_false (dec->stringref))
659 av_push (dec->stringref, newSVpvn (key, len));
660
661 hv_store (hv, key, len, decode_sv (dec), 0); 944 hv_store (hv, key, len, decode_sv (dec), 0);
945
946 return;
662 } 947 }
663 else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27) 948 else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
664 { 949 {
665 I32 len = decode_uint (dec); 950 STRLEN len = decode_uint (dec);
666 char *key = (char *)dec->cur; 951 char *key = (char *)dec->cur;
667 952
953 WANT (len);
668 dec->cur += len; 954 dec->cur += len;
669 955
670 if (ecb_expect_false (dec->stringref)) 956 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
671 av_push (dec->stringref, newSVpvn_utf8 (key, len, 1)); 957 if (!is_utf8_string (key, len))
958 ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
672 959
673 hv_store (hv, key, -len, decode_sv (dec), 0); 960 hv_store (hv, key, -len, decode_sv (dec), 0);
961
962 return;
674 } 963 }
675 else 964
676 {
677 SV *k = decode_sv (dec); 965 SV *k = decode_sv (dec);
678 SV *v = decode_sv (dec); 966 SV *v = decode_sv (dec);
679 967
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
680 hv_store_ent (hv, k, v, 0); 996 hv_store_ent (hv, k, v, 0);
681 SvREFCNT_dec (k); 997 SvREFCNT_dec_NN (k);
682 } 998
999fail:
1000 ;
683} 1001}
684 1002
685static SV * 1003static SV *
686decode_hv (dec_t *dec) 1004decode_hv (dec_t *dec)
687{ 1005{
688 HV *hv = newHV (); 1006 HV *hv = newHV ();
689 1007
690 DEC_INC_DEPTH; 1008 DEC_INC_DEPTH;
691 1009
692 if ((*dec->cur & 31) == 31) 1010 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
693 { 1011 {
694 ++dec->cur; 1012 ++dec->cur;
695 1013
696 for (;;) 1014 for (;;)
697 { 1015 {
698 WANT (1); 1016 WANT (1);
699 1017
700 if (*dec->cur == (0xe0 | 31)) 1018 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
701 { 1019 {
702 ++dec->cur; 1020 ++dec->cur;
703 break; 1021 break;
704 } 1022 }
705 1023
706 decode_he (dec, hv); 1024 decode_he (dec, hv);
707 } 1025 }
708 } 1026 }
709 else 1027 else
710 { 1028 {
711 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
712 1032
713 while (pairs--) 1033 while (pairs--)
714 decode_he (dec, hv); 1034 decode_he (dec, hv);
715 } 1035 }
716 1036
717 DEC_DEC_DEPTH; 1037 DEC_DEC_DEPTH;
718 return newRV_noinc ((SV *)hv); 1038 return newRV_noinc ((SV *)hv);
719 1039
720fail: 1040fail:
721 SvREFCNT_dec (hv); 1041 SvREFCNT_dec_NN (hv);
722 DEC_DEC_DEPTH; 1042 DEC_DEC_DEPTH;
723 return &PL_sv_undef; 1043 return &PL_sv_undef;
724} 1044}
725 1045
726static SV * 1046static SV *
727decode_str (dec_t *dec, int utf8) 1047decode_str (dec_t *dec, int utf8)
728{ 1048{
729 SV *sv = 0; 1049 SV *sv = 0;
730 1050
731 if ((*dec->cur & 31) == 31) 1051 if (ecb_expect_false ((*dec->cur & MINOR_MASK) == MINOR_INDEF))
732 { 1052 {
1053 // indefinite length strings
733 ++dec->cur; 1054 ++dec->cur;
734 1055
1056 U8 major = *dec->cur & MAJOR_MISC;
1057
735 sv = newSVpvn ("", 0); 1058 sv = newSVpvn ("", 0);
736 1059
737 // not very fast, and certainly not robust against illegal input
738 for (;;) 1060 for (;;)
739 { 1061 {
740 WANT (1); 1062 WANT (1);
741 1063
742 if (*dec->cur == (0xe0 | 31)) 1064 if ((*dec->cur - major) > LENGTH_EXT8)
1065 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
743 { 1066 {
744 ++dec->cur; 1067 ++dec->cur;
745 break; 1068 break;
746 } 1069 }
1070 else
1071 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
747 1072
748 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;
749 } 1078 }
750 } 1079 }
751 else 1080 else
752 { 1081 {
753 STRLEN len = decode_uint (dec); 1082 STRLEN len = decode_uint (dec);
754 1083
755 WANT (len); 1084 WANT (len);
756 sv = newSVpvn (dec->cur, len); 1085 sv = newSVpvn (dec->cur, len);
757 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));
758 } 1091 }
759 1092
760 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
761 SvUTF8_on (sv); 1099 SvUTF8_on (sv);
762 1100 }
763 if (ecb_expect_false (dec->stringref))
764 av_push (dec->stringref, SvREFCNT_inc_NN (sv));
765 1101
766 return sv; 1102 return sv;
767 1103
768fail: 1104fail:
769 SvREFCNT_dec (sv); 1105 SvREFCNT_dec (sv);
788 sv = newRV_noinc (decode_sv (dec)); 1124 sv = newRV_noinc (decode_sv (dec));
789 break; 1125 break;
790 1126
791 case CBOR_TAG_STRINGREF_NAMESPACE: 1127 case CBOR_TAG_STRINGREF_NAMESPACE:
792 { 1128 {
1129 // do not use SAVETMPS/FREETMPS, as these will
1130 // erase mortalised caches, e.g. "shareable"
793 ENTER; SAVETMPS; 1131 ENTER;
794 1132
795 SAVESPTR (dec->stringref); 1133 SAVESPTR (dec->stringref);
796 dec->stringref = (AV *)sv_2mortal ((SV *)newAV ()); 1134 dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());
797 1135
798 sv = decode_sv (dec); 1136 sv = decode_sv (dec);
799 1137
800 FREETMPS; LEAVE; 1138 LEAVE;
801 } 1139 }
802 break; 1140 break;
803 1141
804 case CBOR_TAG_STRINGREF: 1142 case CBOR_TAG_STRINGREF:
805 { 1143 {
806 if ((*dec->cur >> 5) != 0) 1144 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
807 ERR ("corrupted CBOR data (stringref index not an unsigned integer)"); 1145 ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
808 1146
809 UV idx = decode_uint (dec); 1147 UV idx = decode_uint (dec);
810 1148
811 if (!dec->stringref || (int)idx > AvFILLp (dec->stringref)) 1149 if (!dec->stringref || idx >= (UV)(1 + AvFILLp (dec->stringref)))
812 ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)"); 1150 ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)");
813 1151
814 sv = newSVsv (AvARRAY (dec->stringref)[idx]); 1152 sv = newSVsv (AvARRAY (dec->stringref)[idx]);
815 } 1153 }
816 break; 1154 break;
818 case CBOR_TAG_VALUE_SHAREABLE: 1156 case CBOR_TAG_VALUE_SHAREABLE:
819 { 1157 {
820 if (ecb_expect_false (!dec->shareable)) 1158 if (ecb_expect_false (!dec->shareable))
821 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ()); 1159 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
822 1160
1161 if (dec->cbor.flags & F_ALLOW_CYCLES)
1162 {
823 sv = newSV (0); 1163 sv = newSV (0);
824 av_push (dec->shareable, SvREFCNT_inc_NN (sv)); 1164 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
825 1165
826 SV *osv = decode_sv (dec); 1166 SV *osv = decode_sv (dec);
827 sv_setsv (sv, osv); 1167 sv_setsv (sv, osv);
828 SvREFCNT_dec_NN (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 }
829 } 1177 }
830 break; 1178 break;
831 1179
832 case CBOR_TAG_VALUE_SHAREDREF: 1180 case CBOR_TAG_VALUE_SHAREDREF:
833 { 1181 {
834 if ((*dec->cur >> 5) != 0) 1182 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
835 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)"); 1183 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
836 1184
837 UV idx = decode_uint (dec); 1185 UV idx = decode_uint (dec);
838 1186
839 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable)) 1187 if (!dec->shareable || idx >= (UV)(1 + AvFILLp (dec->shareable)))
840 ERR ("corrupted CBOR data (sharedref index out of bounds)"); 1188 ERR ("corrupted CBOR data (sharedref index out of bounds)");
841 1189
842 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]); 1190 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
1191
1192 if (sv == &PL_sv_undef)
1193 ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
843 } 1194 }
844 break; 1195 break;
845 1196
846 case CBOR_TAG_PERL_OBJECT: 1197 case CBOR_TAG_PERL_OBJECT:
847 { 1198 {
1199 if (dec->cbor.flags & F_FORBID_OBJECTS)
1200 goto filter;
1201
848 sv = decode_sv (dec); 1202 sv = decode_sv (dec);
849 1203
850 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) 1204 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
851 ERR ("corrupted CBOR data (non-array perl object)"); 1205 ERR ("corrupted CBOR data (non-array perl object)");
852 1206
862 if (!method) 1216 if (!method)
863 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)");
864 1218
865 dSP; 1219 dSP;
866 1220
867 ENTER; SAVETMPS; PUSHMARK (SP); 1221 ENTER; SAVETMPS;
1222 PUSHMARK (SP);
868 EXTEND (SP, len + 1); 1223 EXTEND (SP, len + 1);
869 // 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
870 PUSHs (*av_fetch (av, 0, 1)); 1225 PUSHs (*av_fetch (av, 0, 1));
871 PUSHs (sv_cbor); 1226 PUSHs (sv_cbor);
872 1227
880 SPAGAIN; 1235 SPAGAIN;
881 1236
882 if (SvTRUE (ERRSV)) 1237 if (SvTRUE (ERRSV))
883 { 1238 {
884 FREETMPS; LEAVE; 1239 FREETMPS; LEAVE;
885 ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV)))); 1240 ERR_ERRSV;
886 } 1241 }
887 1242
888 SvREFCNT_dec (sv); 1243 SvREFCNT_dec_NN (sv);
889 sv = SvREFCNT_inc (POPs); 1244 sv = SvREFCNT_inc (POPs);
890 1245
891 PUTBACK; 1246 PUTBACK;
892 1247
893 FREETMPS; LEAVE; 1248 FREETMPS; LEAVE;
894 } 1249 }
895 break; 1250 break;
896 1251
897 default: 1252 default:
1253 filter:
898 { 1254 {
1255 SV *tag_sv = newSVuv (tag);
1256
899 sv = decode_sv (dec); 1257 sv = decode_sv (dec);
900 1258
1259 dSP;
1260 ENTER; SAVETMPS;
1261 PUSHMARK (SP);
1262 EXTEND (SP, 2);
1263 PUSHs (tag_sv);
1264 PUSHs (sv);
1265
1266 PUTBACK;
1267 int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL);
1268 SPAGAIN;
1269
1270 if (SvTRUE (ERRSV))
1271 {
1272 SvREFCNT_dec_NN (tag_sv);
1273 FREETMPS; LEAVE;
1274 ERR_ERRSV;
1275 }
1276
1277 if (count)
1278 {
1279 SvREFCNT_dec_NN (tag_sv);
1280 SvREFCNT_dec_NN (sv);
1281 sv = SvREFCNT_inc_NN (TOPs);
1282 SP -= count;
1283 }
1284 else
1285 {
901 AV *av = newAV (); 1286 AV *av = newAV ();
902 av_push (av, newSVuv (tag)); 1287 av_push (av, tag_sv);
903 av_push (av, sv); 1288 av_push (av, sv);
904 1289
905 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 1290 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
906 ? cbor_tagged_stash 1291 ? cbor_tagged_stash
907 : gv_stashpv ("CBOR::XS::Tagged" , 1); 1292 : gv_stashpv ("CBOR::XS::Tagged" , 1);
908
909 sv = 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;
910 } 1299 }
911 break; 1300 break;
912 } 1301 }
913 1302
914 return sv; 1303 return sv;
921static SV * 1310static SV *
922decode_sv (dec_t *dec) 1311decode_sv (dec_t *dec)
923{ 1312{
924 WANT (1); 1313 WANT (1);
925 1314
926 switch (*dec->cur >> 5) 1315 switch (*dec->cur >> MAJOR_SHIFT)
927 { 1316 {
928 case 0: // unsigned int 1317 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
929 return newSVuv (decode_uint (dec)); 1318 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
930 case 1: // negative int 1319 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
931 return newSViv (-1 - (IV)decode_uint (dec)); 1320 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
932 case 2: // octet string 1321 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
933 return decode_str (dec, 0); 1322 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
934 case 3: // utf-8 string 1323 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
935 return decode_str (dec, 1); 1324
936 case 4: // array 1325 case MAJOR_MISC >> MAJOR_SHIFT:
937 return decode_av (dec);
938 case 5: // map
939 return decode_hv (dec);
940 case 6: // tag
941 return decode_tagged (dec);
942 case 7: // misc
943 switch (*dec->cur++ & 31) 1326 switch (*dec->cur++ & MINOR_MASK)
944 { 1327 {
945 case 20: 1328 case SIMPLE_FALSE:
946#if CBOR_SLOW 1329#if CBOR_SLOW
947 types_false = get_bool ("Types::Serialiser::false"); 1330 types_false = get_bool ("Types::Serialiser::false");
948#endif 1331#endif
949 return newSVsv (types_false); 1332 return newSVsv (types_false);
950 case 21: 1333 case SIMPLE_TRUE:
951#if CBOR_SLOW 1334#if CBOR_SLOW
952 types_true = get_bool ("Types::Serialiser::true"); 1335 types_true = get_bool ("Types::Serialiser::true");
953#endif 1336#endif
954 return newSVsv (types_true); 1337 return newSVsv (types_true);
955 case 22: 1338 case SIMPLE_NULL:
956 return newSVsv (&PL_sv_undef); 1339 return newSVsv (&PL_sv_undef);
957 case 23: 1340 case SIMPLE_UNDEF:
958#if CBOR_SLOW 1341#if CBOR_SLOW
959 types_error = get_bool ("Types::Serialiser::error"); 1342 types_error = get_bool ("Types::Serialiser::error");
960#endif 1343#endif
961 return newSVsv (types_error); 1344 return newSVsv (types_error);
962 1345
963 case 25: 1346 case MISC_FLOAT16:
964 { 1347 {
965 WANT (2); 1348 WANT (2);
966 1349
967 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1350 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
968 dec->cur += 2; 1351 dec->cur += 2;
969 1352
970 return newSVnv (ecb_binary16_to_float (fp)); 1353 return newSVnv (ecb_binary16_to_float (fp));
971 } 1354 }
972 1355
973 case 26: 1356 case MISC_FLOAT32:
974 { 1357 {
975 uint32_t fp; 1358 uint32_t fp;
976 WANT (4); 1359 WANT (4);
977 memcpy (&fp, dec->cur, 4); 1360 memcpy (&fp, dec->cur, 4);
978 dec->cur += 4; 1361 dec->cur += 4;
981 fp = ecb_bswap32 (fp); 1364 fp = ecb_bswap32 (fp);
982 1365
983 return newSVnv (ecb_binary32_to_float (fp)); 1366 return newSVnv (ecb_binary32_to_float (fp));
984 } 1367 }
985 1368
986 case 27: 1369 case MISC_FLOAT64:
987 { 1370 {
988 uint64_t fp; 1371 uint64_t fp;
989 WANT (8); 1372 WANT (8);
990 memcpy (&fp, dec->cur, 8); 1373 memcpy (&fp, dec->cur, 8);
991 dec->cur += 8; 1374 dec->cur += 8;
994 fp = ecb_bswap64 (fp); 1377 fp = ecb_bswap64 (fp);
995 1378
996 return newSVnv (ecb_binary64_to_double (fp)); 1379 return newSVnv (ecb_binary64_to_double (fp));
997 } 1380 }
998 1381
999 // 0..19 unassigned 1382 // 0..19 unassigned simple
1000 // 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
1001 default: 1386 default:
1002 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1387 ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1003 } 1388 }
1004 1389
1005 break; 1390 break;
1006 } 1391 }
1007 1392
1010} 1395}
1011 1396
1012static SV * 1397static SV *
1013decode_cbor (SV *string, CBOR *cbor, char **offset_return) 1398decode_cbor (SV *string, CBOR *cbor, char **offset_return)
1014{ 1399{
1015 dec_t dec = { }; 1400 dec_t dec = { 0 };
1016 SV *sv; 1401 SV *sv;
1017 STRLEN len; 1402 STRLEN len;
1018 char *data = SvPVbyte (string, len); 1403 char *data = SvPVbyte (string, len);
1019 1404
1020 if (len > cbor->max_size && cbor->max_size) 1405 if (len > cbor->max_size && cbor->max_size)
1034 if (dec.cur != dec.end && !dec.err) 1419 if (dec.cur != dec.end && !dec.err)
1035 dec.err = "garbage after CBOR object"; 1420 dec.err = "garbage after CBOR object";
1036 1421
1037 if (dec.err) 1422 if (dec.err)
1038 { 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
1039 SvREFCNT_dec (sv); 1435 SvREFCNT_dec_NN (sv);
1436
1437 if (dec.err_sv)
1438 sv_2mortal (dec.err_sv);
1439
1040 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);
1041 } 1441 }
1042 1442
1043 sv = sv_2mortal (sv); 1443 sv = sv_2mortal (sv);
1044 1444
1045 return sv; 1445 return sv;
1046} 1446}
1047 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
1048///////////////////////////////////////////////////////////////////////////// 1572/////////////////////////////////////////////////////////////////////////////
1049// XS interface functions 1573// XS interface functions
1050 1574
1051MODULE = CBOR::XS PACKAGE = CBOR::XS 1575MODULE = CBOR::XS PACKAGE = CBOR::XS
1052 1576
1060 1584
1061 types_true = get_bool ("Types::Serialiser::true" ); 1585 types_true = get_bool ("Types::Serialiser::true" );
1062 types_false = get_bool ("Types::Serialiser::false"); 1586 types_false = get_bool ("Types::Serialiser::false");
1063 types_error = get_bool ("Types::Serialiser::error"); 1587 types_error = get_bool ("Types::Serialiser::error");
1064 1588
1589 default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1590
1065 sv_cbor = newSVpv ("CBOR", 0); 1591 sv_cbor = newSVpv ("CBOR", 0);
1066 SvREADONLY_on (sv_cbor); 1592 SvREADONLY_on (sv_cbor);
1593
1594 assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1));
1067} 1595}
1068 1596
1069PROTOTYPES: DISABLE 1597PROTOTYPES: DISABLE
1070 1598
1071void CLONE (...) 1599void CLONE (...)
1090void shrink (CBOR *self, int enable = 1) 1618void shrink (CBOR *self, int enable = 1)
1091 ALIAS: 1619 ALIAS:
1092 shrink = F_SHRINK 1620 shrink = F_SHRINK
1093 allow_unknown = F_ALLOW_UNKNOWN 1621 allow_unknown = F_ALLOW_UNKNOWN
1094 allow_sharing = F_ALLOW_SHARING 1622 allow_sharing = F_ALLOW_SHARING
1095 dedup_keys = F_DEDUP_KEYS 1623 allow_cycles = F_ALLOW_CYCLES
1624 forbid_objects = F_FORBID_OBJECTS
1096 dedup_strings = F_DEDUP_STRINGS 1625 pack_strings = F_PACK_STRINGS
1626 text_keys = F_TEXT_KEYS
1627 text_strings = F_TEXT_STRINGS
1628 validate_utf8 = F_VALIDATE_UTF8
1097 PPCODE: 1629 PPCODE:
1098{ 1630{
1099 if (enable) 1631 if (enable)
1100 self->flags |= ix; 1632 self->flags |= ix;
1101 else 1633 else
1107void get_shrink (CBOR *self) 1639void get_shrink (CBOR *self)
1108 ALIAS: 1640 ALIAS:
1109 get_shrink = F_SHRINK 1641 get_shrink = F_SHRINK
1110 get_allow_unknown = F_ALLOW_UNKNOWN 1642 get_allow_unknown = F_ALLOW_UNKNOWN
1111 get_allow_sharing = F_ALLOW_SHARING 1643 get_allow_sharing = F_ALLOW_SHARING
1112 get_dedup_keys = F_DEDUP_KEYS 1644 get_allow_cycles = F_ALLOW_CYCLES
1645 get_forbid_objects = F_FORBID_OBJECTS
1113 get_dedup_strings = F_DEDUP_STRINGS 1646 get_pack_strings = F_PACK_STRINGS
1647 get_text_keys = F_TEXT_KEYS
1648 get_text_strings = F_TEXT_STRINGS
1649 get_validate_utf8 = F_VALIDATE_UTF8
1114 PPCODE: 1650 PPCODE:
1115 XPUSHs (boolSV (self->flags & ix)); 1651 XPUSHs (boolSV (self->flags & ix));
1116 1652
1117void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1653void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1118 PPCODE: 1654 PPCODE:
1131 XPUSHs (ST (0)); 1667 XPUSHs (ST (0));
1132 1668
1133int get_max_size (CBOR *self) 1669int get_max_size (CBOR *self)
1134 CODE: 1670 CODE:
1135 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);
1136 OUTPUT: 1684 OUTPUT:
1137 RETVAL 1685 RETVAL
1138 1686
1139void encode (CBOR *self, SV *scalar) 1687void encode (CBOR *self, SV *scalar)
1140 PPCODE: 1688 PPCODE:
1155 EXTEND (SP, 2); 1703 EXTEND (SP, 2);
1156 PUSHs (sv); 1704 PUSHs (sv);
1157 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1705 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1158} 1706}
1159 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
1160PROTOTYPES: ENABLE 1764PROTOTYPES: ENABLE
1161 1765
1162void encode_cbor (SV *scalar) 1766void encode_cbor (SV *scalar)
1767 ALIAS:
1768 encode_cbor = 0
1769 encode_cbor_sharing = F_ALLOW_SHARING
1163 PPCODE: 1770 PPCODE:
1164{ 1771{
1165 CBOR cbor; 1772 CBOR cbor;
1166 cbor_init (&cbor); 1773 cbor_init (&cbor);
1774 cbor.flags |= ix;
1167 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1775 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1168 XPUSHs (scalar); 1776 XPUSHs (scalar);
1169} 1777}
1170 1778
1171void decode_cbor (SV *cborstr) 1779void decode_cbor (SV *cborstr)
1175 cbor_init (&cbor); 1783 cbor_init (&cbor);
1176 PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN; 1784 PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN;
1177 XPUSHs (cborstr); 1785 XPUSHs (cborstr);
1178} 1786}
1179 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