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

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.18 by root, Wed Nov 20 01:09:46 2013 UTC vs.
Revision 1.61 by root, Sat Nov 26 02:10:19 2016 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_dec_NN
27# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
28#endif
29
30// known major and minor types
31enum cbor_type
32{
33 MAJOR_SHIFT = 5,
34 MINOR_MASK = 0x1f,
35
36 MAJOR_POS_INT = 0 << MAJOR_SHIFT,
37 MAJOR_NEG_INT = 1 << MAJOR_SHIFT,
38 MAJOR_BYTES = 2 << MAJOR_SHIFT,
39 MAJOR_TEXT = 3 << MAJOR_SHIFT,
40 MAJOR_ARRAY = 4 << MAJOR_SHIFT,
41 MAJOR_MAP = 5 << MAJOR_SHIFT,
42 MAJOR_TAG = 6 << MAJOR_SHIFT,
43 MAJOR_MISC = 7 << MAJOR_SHIFT,
44
45 // INT/STRING/ARRAY/MAP subtypes
46 LENGTH_EXT1 = 24,
47 LENGTH_EXT2 = 25,
48 LENGTH_EXT4 = 26,
49 LENGTH_EXT8 = 27,
50
51 // SIMPLE types (effectively MISC subtypes)
52 SIMPLE_FALSE = 20,
53 SIMPLE_TRUE = 21,
54 SIMPLE_NULL = 22,
55 SIMPLE_UNDEF = 23,
56
57 // MISC subtype (unused)
58 MISC_EXT1 = 24,
59 MISC_FLOAT16 = 25,
60 MISC_FLOAT32 = 26,
61 MISC_FLOAT64 = 27,
62
63 // BYTES/TEXT/ARRAY/MAP
64 MINOR_INDEF = 31,
65};
24 66
25// known tags 67// known tags
26enum cbor_tag 68enum cbor_tag
27{ 69{
28 // inofficial extensions (pending iana registration) 70 // extensions
71 CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
29 CBOR_TAG_PERL_OBJECT = 24, // http://cbor.schmorp.de/perl-object 72 CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
30 CBOR_TAG_GENERIC_OBJECT = 25, // http://cbor.schmorp.de/generic-object 73 CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
31 CBOR_TAG_VALUE_SHARABLE = 26, // http://cbor.schmorp.de/value-sharing 74 CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
32 CBOR_TAG_VALUE_SHAREDREF = 27, // http://cbor.schmorp.de/value-sharing 75 CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
33 CBOR_TAG_STRINGREF_NAMESPACE = 65537, // http://cbor.schmorp.de/stringref 76 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 77 CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
36 78
37 // rfc7049 79 // rfc7049
38 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8 80 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
39 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any 81 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
40 CBOR_TAG_POS_BIGNUM = 2, // byte string 82 CBOR_TAG_POS_BIGNUM = 2, // byte string
41 CBOR_TAG_NEG_BIGNUM = 3, // byte string 83 CBOR_TAG_NEG_BIGNUM = 3, // byte string
42 CBOR_TAG_DECIMAL = 4, // decimal fraction, array 84 CBOR_TAG_DECIMAL = 4, // decimal fraction, array
43 CBOR_TAG_BIGFLOAT = 5, // array 85 CBOR_TAG_BIGFLOAT = 5, // array
44 86
45 CBOR_TAG_CONV_B64U = 21, // base64url, any 87 CBOR_TAG_CONV_B64U = 21, // base64url, any
46 CBOR_TAG_CONV_B64 = 22, // base64, any 88 CBOR_TAG_CONV_B64 = 22, // base64, any
47 CBOR_TAG_CONV_HEX = 23, // base16, any 89 CBOR_TAG_CONV_HEX = 23, // base16, any
48 CBOR_TAG_CBOR = 24, // embedded cbor, byte string 90 CBOR_TAG_CBOR = 24, // embedded cbor, byte string
49 91
50 CBOR_TAG_URI = 32, // URI rfc3986, utf-8 92 CBOR_TAG_URI = 32, // URI rfc3986, utf-8
51 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8 93 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
52 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8 94 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
53 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8 95 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
54 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8 96 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
55 97
56 CBOR_TAG_MAGIC = 55799 // self-describe cbor 98 CBOR_TAG_MAGIC = 55799, // self-describe cbor
57}; 99};
58 100
59#define F_SHRINK 0x00000001UL 101#define F_SHRINK 0x00000001UL
60#define F_ALLOW_UNKNOWN 0x00000002UL 102#define F_ALLOW_UNKNOWN 0x00000002UL
61#define F_ALLOW_SHARING 0x00000004UL //TODO 103#define F_ALLOW_SHARING 0x00000004UL
104#define F_ALLOW_CYCLES 0x00000008UL
105#define F_FORBID_OBJECTS 0x00000010UL
62#define F_DEDUP_STRINGS 0x00000008UL //TODO 106#define F_PACK_STRINGS 0x00000020UL
63#define F_DEDUP_KEYS 0x00000010UL //TODO 107#define F_TEXT_KEYS 0x00000040UL
108#define F_TEXT_STRINGS 0x00000080UL
109#define F_VALIDATE_UTF8 0x00000100UL
64 110
65#define INIT_SIZE 32 // initial scalar size to be allocated 111#define INIT_SIZE 32 // initial scalar size to be allocated
66 112
67#define SB do { 113#define SB do {
68#define SE } while (0) 114#define SE } while (0)
80# define CBOR_SLOW 0 126# define CBOR_SLOW 0
81# define CBOR_STASH cbor_stash 127# define CBOR_STASH cbor_stash
82#endif 128#endif
83 129
84static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS:: 130static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS::
85static SV *types_true, *types_false, *types_error, *sv_cbor; 131static SV *types_true, *types_false, *types_error, *sv_cbor, *default_filter;
86 132
87typedef struct { 133typedef struct {
88 U32 flags; 134 U32 flags;
89 U32 max_depth; 135 U32 max_depth;
90 STRLEN max_size; 136 STRLEN max_size;
137 SV *filter;
138
139 // for the incremental parser
140 STRLEN incr_pos; // the current offset into the text
141 STRLEN incr_need; // minimum bytes needed to decode
142 AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef.
91} CBOR; 143} CBOR;
92 144
93ecb_inline void 145ecb_inline void
94cbor_init (CBOR *cbor) 146cbor_init (CBOR *cbor)
95{ 147{
96 Zero (cbor, 1, CBOR); 148 Zero (cbor, 1, CBOR);
97 cbor->max_depth = 512; 149 cbor->max_depth = 512;
150}
151
152ecb_inline void
153cbor_free (CBOR *cbor)
154{
155 SvREFCNT_dec (cbor->filter);
156 SvREFCNT_dec (cbor->incr_count);
98} 157}
99 158
100///////////////////////////////////////////////////////////////////////////// 159/////////////////////////////////////////////////////////////////////////////
101// utility functions 160// utility functions
102 161
124 SvPV_renew (sv, SvCUR (sv) + 1); 183 SvPV_renew (sv, SvCUR (sv) + 1);
125#endif 184#endif
126 } 185 }
127} 186}
128 187
188// minimum length of a string to be registered for stringref
189ecb_inline int
190minimum_string_length (UV idx)
191{
192
193 return idx <= 23 ? 3
194 : idx <= 0xffU ? 4
195 : idx <= 0xffffU ? 5
196 : idx <= 0xffffffffU ? 7
197 : 11;
198}
199
129///////////////////////////////////////////////////////////////////////////// 200/////////////////////////////////////////////////////////////////////////////
130// encoder 201// encoder
131 202
132// structure used for encoding CBOR 203// structure used for encoding CBOR
133typedef struct 204typedef struct
135 char *cur; // SvPVX (sv) + current output position 206 char *cur; // SvPVX (sv) + current output position
136 char *end; // SvEND (sv) 207 char *end; // SvEND (sv)
137 SV *sv; // result scalar 208 SV *sv; // result scalar
138 CBOR cbor; 209 CBOR cbor;
139 U32 depth; // recursion level 210 U32 depth; // recursion level
140 HV *stringref; // string => index, or 0 211 HV *stringref[2]; // string => index, or 0 ([0] = bytes, [1] = utf-8)
212 UV stringref_idx;
141 HV *sharable; // ptr => index, or 0 213 HV *shareable; // ptr => index, or 0
142 HV *sharable_idx; 214 UV shareable_idx;
143} enc_t; 215} enc_t;
144 216
145ecb_inline void 217ecb_inline void
146need (enc_t *enc, STRLEN len) 218need (enc_t *enc, STRLEN len)
147{ 219{
148 if (ecb_expect_false (enc->cur + len >= enc->end)) 220 if (ecb_expect_false ((uintptr_t)(enc->end - enc->cur) < len))
149 { 221 {
150 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv); 222 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv);
151 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1); 223 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1);
152 enc->cur = SvPVX (enc->sv) + cur; 224 enc->cur = SvPVX (enc->sv) + cur;
153 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1; 225 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;
164static void 236static void
165encode_uint (enc_t *enc, int major, UV len) 237encode_uint (enc_t *enc, int major, UV len)
166{ 238{
167 need (enc, 9); 239 need (enc, 9);
168 240
169 if (len < 24) 241 if (ecb_expect_true (len < LENGTH_EXT1))
170 *enc->cur++ = major | len; 242 *enc->cur++ = major | len;
171 else if (len <= 0xff) 243 else if (ecb_expect_true (len <= 0xffU))
172 { 244 {
173 *enc->cur++ = major | 24; 245 *enc->cur++ = major | LENGTH_EXT1;
174 *enc->cur++ = len; 246 *enc->cur++ = len;
175 } 247 }
176 else if (len <= 0xffff) 248 else if (len <= 0xffffU)
177 { 249 {
178 *enc->cur++ = major | 25; 250 *enc->cur++ = major | LENGTH_EXT2;
179 *enc->cur++ = len >> 8; 251 *enc->cur++ = len >> 8;
180 *enc->cur++ = len; 252 *enc->cur++ = len;
181 } 253 }
182 else if (len <= 0xffffffff) 254 else if (len <= 0xffffffffU)
183 { 255 {
184 *enc->cur++ = major | 26; 256 *enc->cur++ = major | LENGTH_EXT4;
185 *enc->cur++ = len >> 24; 257 *enc->cur++ = len >> 24;
186 *enc->cur++ = len >> 16; 258 *enc->cur++ = len >> 16;
187 *enc->cur++ = len >> 8; 259 *enc->cur++ = len >> 8;
188 *enc->cur++ = len; 260 *enc->cur++ = len;
189 } 261 }
190 else 262 else
191 { 263 {
192 *enc->cur++ = major | 27; 264 *enc->cur++ = major | LENGTH_EXT8;
193 *enc->cur++ = len >> 56; 265 *enc->cur++ = len >> 56;
194 *enc->cur++ = len >> 48; 266 *enc->cur++ = len >> 48;
195 *enc->cur++ = len >> 40; 267 *enc->cur++ = len >> 40;
196 *enc->cur++ = len >> 32; 268 *enc->cur++ = len >> 32;
197 *enc->cur++ = len >> 24; 269 *enc->cur++ = len >> 24;
199 *enc->cur++ = len >> 8; 271 *enc->cur++ = len >> 8;
200 *enc->cur++ = len; 272 *enc->cur++ = len;
201 } 273 }
202} 274}
203 275
204static void 276ecb_inline void
277encode_tag (enc_t *enc, UV tag)
278{
279 encode_uint (enc, MAJOR_TAG, tag);
280}
281
282// exceptional (hopefully) slow path for byte strings that need to be utf8-encoded
283ecb_noinline static void
205encode_str (enc_t *enc, int utf8, char *str, STRLEN len) 284encode_str_utf8 (enc_t *enc, int utf8, char *str, STRLEN len)
206{ 285{
207 encode_uint (enc, utf8 ? 0x60 : 0x40, len); 286 STRLEN ulen = len;
287 U8 *p, *pend = (U8 *)str + len;
288
289 for (p = (U8 *)str; p < pend; ++p)
290 ulen += *p >> 7; // count set high bits
291
292 encode_uint (enc, MAJOR_TEXT, ulen);
293
294 need (enc, ulen);
295 for (p = (U8 *)str; p < pend; ++p)
296 if (*p < 0x80)
297 *enc->cur++ = *p;
298 else
299 {
300 *enc->cur++ = 0xc0 + (*p >> 6);
301 *enc->cur++ = 0x80 + (*p & 63);
302 }
303}
304
305ecb_inline void
306encode_str (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
307{
308 if (ecb_expect_false (upgrade_utf8))
309 if (!utf8)
310 {
311 encode_str_utf8 (enc, utf8, str, len);
312 return;
313 }
314
315 encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
208 need (enc, len); 316 need (enc, len);
209 memcpy (enc->cur, str, len); 317 memcpy (enc->cur, str, len);
210 enc->cur += len; 318 enc->cur += len;
211} 319}
212 320
213ecb_inline void 321ecb_inline void
214encode_tag (enc_t *enc, UV tag) 322encode_strref (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
215{ 323{
216 encode_uint (enc, 0xc0, tag); 324 if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
217} 325 {
326 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
218 327
219static int
220encode_sharable2 (enc_t *enc, SV *sv)
221{
222 if (!enc->sharable)
223 enc->sharable = (HV *)sv_2mortal ((SV *)newHV ());
224
225 SV **svp = hv_fetch (enc->sharable, &sv, sizeof (sv), 1);
226
227 if (SvOK (*svp)) 328 if (SvOK (*svp))
228 { 329 {
330 // already registered, use stringref
229 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF); 331 encode_tag (enc, CBOR_TAG_STRINGREF);
230 encode_uint (enc, 0x00, SvUV (*svp)); 332 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
231
232 return 1; 333 return;
233 } 334 }
234 else 335 else if (len >= minimum_string_length (enc->stringref_idx))
235 { 336 {
337 // register only
236 sv_setuv (*svp, enc->sharable_idx++); 338 sv_setuv (*svp, enc->stringref_idx);
237 encode_tag (enc, CBOR_TAG_VALUE_SHARABLE); 339 ++enc->stringref_idx;
238 340 }
239 return 0;
240 } 341 }
241}
242 342
243ecb_inline int 343 encode_str (enc, upgrade_utf8, utf8, str, len);
244encode_sharable (enc_t *enc, SV *sv)
245{
246 if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)
247 && ecb_expect_false (SvREFCNT (sv) > 1))
248 return encode_sharable2 (enc, sv);
249
250 return 0;
251} 344}
252 345
253static void encode_sv (enc_t *enc, SV *sv); 346static void encode_sv (enc_t *enc, SV *sv);
254 347
255static void 348static void
260 if (enc->depth >= enc->cbor.max_depth) 353 if (enc->depth >= enc->cbor.max_depth)
261 croak (ERR_NESTING_EXCEEDED); 354 croak (ERR_NESTING_EXCEEDED);
262 355
263 ++enc->depth; 356 ++enc->depth;
264 357
265 encode_uint (enc, 0x80, len + 1); 358 encode_uint (enc, MAJOR_ARRAY, len + 1);
266 359
360 if (SvMAGICAL (av))
267 for (i = 0; i <= len; ++i) 361 for (i = 0; i <= len; ++i)
268 { 362 {
269 SV **svp = av_fetch (av, i, 0); 363 SV **svp = av_fetch (av, i, 0);
270 encode_sv (enc, svp ? *svp : &PL_sv_undef); 364 encode_sv (enc, svp ? *svp : &PL_sv_undef);
271 } 365 }
366 else
367 for (i = 0; i <= len; ++i)
368 {
369 SV *sv = AvARRAY (av)[i];
370 encode_sv (enc, sv ? sv : &PL_sv_undef);
371 }
272 372
273 --enc->depth; 373 --enc->depth;
274} 374}
275 375
276static void 376static void
285 385
286 int pairs = hv_iterinit (hv); 386 int pairs = hv_iterinit (hv);
287 int mg = SvMAGICAL (hv); 387 int mg = SvMAGICAL (hv);
288 388
289 if (mg) 389 if (mg)
290 encode_ch (enc, 0xa0 | 31); 390 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
291 else 391 else
292 encode_uint (enc, 0xa0, pairs); 392 encode_uint (enc, MAJOR_MAP, pairs);
293 393
294 while ((he = hv_iternext (hv))) 394 while ((he = hv_iternext (hv)))
295 { 395 {
296 if (HeKLEN (he) == HEf_SVKEY) 396 if (HeKLEN (he) == HEf_SVKEY)
297 encode_sv (enc, HeSVKEY (he)); 397 encode_sv (enc, HeSVKEY (he));
298 else 398 else
299 encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); 399 encode_strref (enc, enc->cbor.flags & (F_TEXT_KEYS | F_TEXT_STRINGS), HeKUTF8 (he), HeKEY (he), HeKLEN (he));
300 400
301 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); 401 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
302 } 402 }
303 403
304 if (mg) 404 if (mg)
305 encode_ch (enc, 0xe0 | 31); 405 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
306 406
307 --enc->depth; 407 --enc->depth;
308} 408}
309 409
310// encode objects, arrays and special \0=false and \1=true values. 410// encode objects, arrays and special \0=false and \1=true values.
311static void 411static void
312encode_rv (enc_t *enc, SV *sv) 412encode_rv (enc_t *enc, SV *sv)
313{ 413{
314 svtype svt;
315
316 SvGETMAGIC (sv); 414 SvGETMAGIC (sv);
415
317 svt = SvTYPE (sv); 416 svtype svt = SvTYPE (sv);
318
319 if (encode_sharable (enc, sv))
320 return;
321 417
322 if (ecb_expect_false (SvOBJECT (sv))) 418 if (ecb_expect_false (SvOBJECT (sv)))
323 { 419 {
324 HV *boolean_stash = !CBOR_SLOW || types_boolean_stash 420 HV *boolean_stash = !CBOR_SLOW || types_boolean_stash
325 ? types_boolean_stash 421 ? types_boolean_stash
330 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 426 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
331 ? cbor_tagged_stash 427 ? cbor_tagged_stash
332 : gv_stashpv ("CBOR::XS::Tagged" , 1); 428 : gv_stashpv ("CBOR::XS::Tagged" , 1);
333 429
334 HV *stash = SvSTASH (sv); 430 HV *stash = SvSTASH (sv);
335 GV *method;
336 431
337 if (stash == boolean_stash) 432 if (stash == boolean_stash)
338 encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20); 433 {
434 encode_ch (enc, SvIV (sv) ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE);
435 return;
436 }
339 else if (stash == error_stash) 437 else if (stash == error_stash)
340 encode_ch (enc, 0xe0 | 23); 438 {
439 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
440 return;
441 }
341 else if (stash == tagged_stash) 442 else if (stash == tagged_stash)
342 { 443 {
343 if (svt != SVt_PVAV) 444 if (svt != SVt_PVAV)
344 croak ("encountered CBOR::XS::Tagged object that isn't an array"); 445 croak ("encountered CBOR::XS::Tagged object that isn't an array");
345 446
346 encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); 447 encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
347 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); 448 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1));
449
450 return;
451 }
452 }
453
454 if (ecb_expect_false (SvREFCNT (sv) > 1)
455 && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
456 {
457 if (!enc->shareable)
458 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
459
460 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
461
462 if (SvOK (*svp))
348 } 463 {
464 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
465 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
466 return;
467 }
468 else
469 {
470 sv_setuv (*svp, enc->shareable_idx);
471 ++enc->shareable_idx;
472 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
473 }
474 }
475
476 if (ecb_expect_false (SvOBJECT (sv)))
477 {
478 HV *stash = SvSTASH (sv);
479 GV *method;
480
481 if (enc->cbor.flags & F_FORBID_OBJECTS)
482 croak ("encountered object '%s', but forbid_objects is enabled",
483 SvPV_nolen (sv_2mortal (newRV_inc (sv))));
349 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) 484 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
350 { 485 {
351 dSP; 486 dSP;
352 487
353 ENTER; SAVETMPS; PUSHMARK (SP); 488 ENTER; SAVETMPS;
489 PUSHMARK (SP);
354 // we re-bless the reference to get overload and other niceties right 490 // we re-bless the reference to get overload and other niceties right
355 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 491 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
356 492
357 PUTBACK; 493 PUTBACK;
358 // G_SCALAR ensures that return value is 1 494 // G_SCALAR ensures that return value is 1
371 } 507 }
372 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0) 508 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
373 { 509 {
374 dSP; 510 dSP;
375 511
376 ENTER; SAVETMPS; PUSHMARK (SP); 512 ENTER; SAVETMPS;
513 SAVESTACK_POS ();
514 PUSHMARK (SP);
377 EXTEND (SP, 2); 515 EXTEND (SP, 2);
378 // we re-bless the reference to get overload and other niceties right 516 // we re-bless the reference to get overload and other niceties right
379 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 517 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
380 PUSHs (sv_cbor); 518 PUSHs (sv_cbor);
381 519
386 // catch this surprisingly common error 524 // catch this surprisingly common error
387 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) 525 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
388 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash)); 526 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
389 527
390 encode_tag (enc, CBOR_TAG_PERL_OBJECT); 528 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
391 encode_uint (enc, 0x80, count + 1); 529 encode_uint (enc, MAJOR_ARRAY, count + 1);
392 encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); 530 encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
393 531
394 while (count) 532 while (count)
395 encode_sv (enc, SP[1 - count--]); 533 encode_sv (enc, SP[1 - count--]);
396 534
397 PUTBACK; 535 PUTBACK;
418{ 556{
419 double nv = SvNVX (sv); 557 double nv = SvNVX (sv);
420 558
421 need (enc, 9); 559 need (enc, 9);
422 560
423 if (ecb_expect_false (nv == (U32)nv)) 561 if (ecb_expect_false (nv == (NV)(U32)nv))
424 encode_uint (enc, 0x00, (U32)nv); 562 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
425 //TODO: maybe I32? 563 //TODO: maybe I32?
426 else if (ecb_expect_false (nv == (float)nv)) 564 else if (ecb_expect_false (nv == (float)nv))
427 { 565 {
428 uint32_t fp = ecb_float_to_binary32 (nv); 566 uint32_t fp = ecb_float_to_binary32 (nv);
429 567
430 *enc->cur++ = 0xe0 | 26; 568 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
431 569
432 if (!ecb_big_endian ()) 570 if (!ecb_big_endian ())
433 fp = ecb_bswap32 (fp); 571 fp = ecb_bswap32 (fp);
434 572
435 memcpy (enc->cur, &fp, 4); 573 memcpy (enc->cur, &fp, 4);
437 } 575 }
438 else 576 else
439 { 577 {
440 uint64_t fp = ecb_double_to_binary64 (nv); 578 uint64_t fp = ecb_double_to_binary64 (nv);
441 579
442 *enc->cur++ = 0xe0 | 27; 580 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
443 581
444 if (!ecb_big_endian ()) 582 if (!ecb_big_endian ())
445 fp = ecb_bswap64 (fp); 583 fp = ecb_bswap64 (fp);
446 584
447 memcpy (enc->cur, &fp, 8); 585 memcpy (enc->cur, &fp, 8);
452static void 590static void
453encode_sv (enc_t *enc, SV *sv) 591encode_sv (enc_t *enc, SV *sv)
454{ 592{
455 SvGETMAGIC (sv); 593 SvGETMAGIC (sv);
456 594
457 if (encode_sharable (enc, sv))
458 return;
459
460 if (SvPOKp (sv)) 595 if (SvPOKp (sv))
461 { 596 {
462 STRLEN len; 597 STRLEN len;
463 char *str = SvPV (sv, len); 598 char *str = SvPV (sv, len);
464 encode_str (enc, SvUTF8 (sv), str, len); 599 encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len);
465 } 600 }
466 else if (SvNOKp (sv)) 601 else if (SvNOKp (sv))
467 encode_nv (enc, sv); 602 encode_nv (enc, sv);
468 else if (SvIOKp (sv)) 603 else if (SvIOKp (sv))
469 { 604 {
470 if (SvIsUV (sv)) 605 if (SvIsUV (sv))
471 encode_uint (enc, 0x00, SvUVX (sv)); 606 encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
472 else if (SvIVX (sv) >= 0) 607 else if (SvIVX (sv) >= 0)
473 encode_uint (enc, 0x00, SvIVX (sv)); 608 encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
474 else 609 else
475 encode_uint (enc, 0x20, -(SvIVX (sv) + 1)); 610 encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
476 } 611 }
477 else if (SvROK (sv)) 612 else if (SvROK (sv))
478 encode_rv (enc, SvRV (sv)); 613 encode_rv (enc, SvRV (sv));
479 else if (!SvOK (sv)) 614 else if (!SvOK (sv))
480 encode_ch (enc, 0xe0 | 22); 615 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
481 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 616 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
482 encode_ch (enc, 0xe0 | 23); 617 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
483 else 618 else
484 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data", 619 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
485 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 620 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
486} 621}
487 622
488static SV * 623static SV *
489encode_cbor (SV *scalar, CBOR *cbor) 624encode_cbor (SV *scalar, CBOR *cbor)
490{ 625{
491 enc_t enc = { }; 626 enc_t enc = { 0 };
492 627
493 enc.cbor = *cbor; 628 enc.cbor = *cbor;
494 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 629 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
495 enc.cur = SvPVX (enc.sv); 630 enc.cur = SvPVX (enc.sv);
496 enc.end = SvEND (enc.sv); 631 enc.end = SvEND (enc.sv);
497 632
498 SvPOK_only (enc.sv); 633 SvPOK_only (enc.sv);
634
635 if (cbor->flags & F_PACK_STRINGS)
636 {
637 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
638 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
639 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
640 }
641
499 encode_sv (&enc, scalar); 642 encode_sv (&enc, scalar);
500 643
501 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv)); 644 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv));
502 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings 645 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings
503 646
517 U8 *end; // end of input string 660 U8 *end; // end of input string
518 const char *err; // parse error, if != 0 661 const char *err; // parse error, if != 0
519 CBOR cbor; 662 CBOR cbor;
520 U32 depth; // recursion depth 663 U32 depth; // recursion depth
521 U32 maxdepth; // recursion depth limit 664 U32 maxdepth; // recursion depth limit
522 AV *sharable; 665 AV *shareable;
666 AV *stringref;
667 SV *decode_tagged;
668 SV *err_sv; // optional sv for error, needs to be freed
523} dec_t; 669} dec_t;
524 670
525#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE 671// set dec->err to ERRSV
672ecb_cold static void
673err_errsv (dec_t *dec)
674{
675 if (!dec->err)
676 {
677 dec->err_sv = newSVsv (ERRSV);
526 678
527#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data") 679 // chop off the trailing \n
680 SvCUR_set (dec->err_sv, SvCUR (dec->err_sv) - 1);
681 *SvEND (dec->err_sv) = 0;
528 682
683 dec->err = SvPVutf8_nolen (dec->err_sv);
684 }
685}
686
687// the following functions are used to reduce code size and help the compiler to optimise
688ecb_cold static void
689err_set (dec_t *dec, const char *reason)
690{
691 if (!dec->err)
692 dec->err = reason;
693}
694
695ecb_cold static void
696err_unexpected_end (dec_t *dec)
697{
698 err_set (dec, "unexpected end of CBOR data");
699}
700
701ecb_cold static void
702err_nesting_exceeded (dec_t *dec)
703{
704 err_set (dec, ERR_NESTING_EXCEEDED);
705}
706
707#define ERR_DO(do) SB do; goto fail; SE
708#define ERR(reason) ERR_DO (err_set (dec, reason))
709#define ERR_ERRSV ERR_DO (err_errsv (dec))
710
711#define WANT(len) if (ecb_expect_false ((uintptr_t)(dec->end - dec->cur) < (STRLEN)len)) ERR_DO (err_unexpected_end (dec))
712
529#define DEC_INC_DEPTH if (++dec->depth > dec->cbor.max_depth) ERR (ERR_NESTING_EXCEEDED) 713#define DEC_INC_DEPTH if (ecb_expect_false (++dec->depth > dec->cbor.max_depth)) ERR (ERR_NESTING_EXCEEDED)
530#define DEC_DEC_DEPTH --dec->depth 714#define DEC_DEC_DEPTH --dec->depth
531 715
532static UV 716static UV
533decode_uint (dec_t *dec) 717decode_uint (dec_t *dec)
534{ 718{
535 switch (*dec->cur & 31) 719 U8 m = *dec->cur & MINOR_MASK;
536 { 720 ++dec->cur;
537 case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
538 case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
539 case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
540 return *dec->cur++ & 31;
541 721
542 case 24: 722 if (ecb_expect_true (m < LENGTH_EXT1))
723 return m;
724 else if (ecb_expect_true (m == LENGTH_EXT1))
725 {
543 WANT (2); 726 WANT (1);
544 dec->cur += 2; 727 dec->cur += 1;
545 return dec->cur[-1]; 728 return dec->cur[-1];
546 729 }
547 case 25: 730 else if (ecb_expect_true (m == LENGTH_EXT2))
731 {
548 WANT (3); 732 WANT (2);
549 dec->cur += 3; 733 dec->cur += 2;
550 return (((UV)dec->cur[-2]) << 8) 734 return (((UV)dec->cur[-2]) << 8)
551 | ((UV)dec->cur[-1]); 735 | ((UV)dec->cur[-1]);
552 736 }
553 case 26: 737 else if (ecb_expect_true (m == LENGTH_EXT4))
738 {
554 WANT (5); 739 WANT (4);
555 dec->cur += 5; 740 dec->cur += 4;
556 return (((UV)dec->cur[-4]) << 24) 741 return (((UV)dec->cur[-4]) << 24)
557 | (((UV)dec->cur[-3]) << 16) 742 | (((UV)dec->cur[-3]) << 16)
558 | (((UV)dec->cur[-2]) << 8) 743 | (((UV)dec->cur[-2]) << 8)
559 | ((UV)dec->cur[-1]); 744 | ((UV)dec->cur[-1]);
560 745 }
561 case 27: 746 else if (ecb_expect_true (m == LENGTH_EXT8))
747 {
562 WANT (9); 748 WANT (8);
563 dec->cur += 9; 749 dec->cur += 8;
750
751 return
752#if UVSIZE < 8
753 0
754#else
564 return (((UV)dec->cur[-8]) << 56) 755 (((UV)dec->cur[-8]) << 56)
565 | (((UV)dec->cur[-7]) << 48) 756 | (((UV)dec->cur[-7]) << 48)
566 | (((UV)dec->cur[-6]) << 40) 757 | (((UV)dec->cur[-6]) << 40)
567 | (((UV)dec->cur[-5]) << 32) 758 | (((UV)dec->cur[-5]) << 32)
759#endif
568 | (((UV)dec->cur[-4]) << 24) 760 | (((UV)dec->cur[-4]) << 24)
569 | (((UV)dec->cur[-3]) << 16) 761 | (((UV)dec->cur[-3]) << 16)
570 | (((UV)dec->cur[-2]) << 8) 762 | (((UV)dec->cur[-2]) << 8)
571 | ((UV)dec->cur[-1]); 763 | ((UV)dec->cur[-1]);
572 764 }
573 default: 765 else
574 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 766 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
575 }
576 767
577fail: 768fail:
578 return 0; 769 return 0;
579} 770}
580 771
585{ 776{
586 AV *av = newAV (); 777 AV *av = newAV ();
587 778
588 DEC_INC_DEPTH; 779 DEC_INC_DEPTH;
589 780
590 if ((*dec->cur & 31) == 31) 781 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
591 { 782 {
592 ++dec->cur; 783 ++dec->cur;
593 784
594 for (;;) 785 for (;;)
595 { 786 {
596 WANT (1); 787 WANT (1);
597 788
598 if (*dec->cur == (0xe0 | 31)) 789 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
599 { 790 {
600 ++dec->cur; 791 ++dec->cur;
601 break; 792 break;
602 } 793 }
603 794
604 av_push (av, decode_sv (dec)); 795 av_push (av, decode_sv (dec));
605 } 796 }
606 } 797 }
607 else 798 else
608 { 799 {
609 int i, len = decode_uint (dec); 800 UV i, len = decode_uint (dec);
610 801
802 WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
611 av_fill (av, len - 1); 803 av_fill (av, len - 1);
612 804
613 for (i = 0; i < len; ++i) 805 for (i = 0; i < len; ++i)
614 AvARRAY (av)[i] = decode_sv (dec); 806 AvARRAY (av)[i] = decode_sv (dec);
615 } 807 }
625 817
626static void 818static void
627decode_he (dec_t *dec, HV *hv) 819decode_he (dec_t *dec, HV *hv)
628{ 820{
629 // for speed reasons, we specialcase single-string 821 // for speed reasons, we specialcase single-string
630 // byte or utf-8 strings as keys. 822 // byte or utf-8 strings as keys, but only when !stringref
631 823
632 if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27) 824 if (ecb_expect_true (!dec->stringref))
825 if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
633 { 826 {
634 I32 len = decode_uint (dec); 827 STRLEN len = decode_uint (dec);
635 char *key = (char *)dec->cur; 828 char *key = (char *)dec->cur;
636 829
830 WANT (len);
637 dec->cur += len; 831 dec->cur += len;
638 832
639 hv_store (hv, key, len, decode_sv (dec), 0); 833 hv_store (hv, key, len, decode_sv (dec), 0);
834
835 return;
640 } 836 }
641 else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27) 837 else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
642 { 838 {
643 I32 len = decode_uint (dec); 839 STRLEN len = decode_uint (dec);
644 char *key = (char *)dec->cur; 840 char *key = (char *)dec->cur;
645 841
842 WANT (len);
646 dec->cur += len; 843 dec->cur += len;
647 844
845 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
846 if (!is_utf8_string (key, len))
847 ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
848
648 hv_store (hv, key, -len, decode_sv (dec), 0); 849 hv_store (hv, key, -len, decode_sv (dec), 0);
850
851 return;
649 } 852 }
650 else 853
651 {
652 SV *k = decode_sv (dec); 854 SV *k = decode_sv (dec);
653 SV *v = decode_sv (dec); 855 SV *v = decode_sv (dec);
654 856
857 // we leak memory if uncaught exceptions are thrown by random magical
858 // methods, and this is hopefully the only place where it can happen,
859 // so if there is a chance of an exception, take the very slow path.
860 // since catching exceptions is "undocumented/internal/forbidden" by
861 // the new p5p powers, we need to call out to a perl function :/
862 if (ecb_expect_false (SvAMAGIC (k)))
863 {
864 dSP;
865
866 ENTER; SAVETMPS;
867 PUSHMARK (SP);
868 EXTEND (SP, 3);
869 PUSHs (sv_2mortal (newRV_inc ((SV *)hv)));
870 PUSHs (sv_2mortal (k));
871 PUSHs (sv_2mortal (v));
872
873 PUTBACK;
874 call_pv ("CBOR::XS::_hv_store", G_VOID | G_DISCARD | G_EVAL);
875 SPAGAIN;
876
877 FREETMPS; LEAVE;
878
879 if (SvTRUE (ERRSV))
880 ERR_ERRSV;
881
882 return;
883 }
884
655 hv_store_ent (hv, k, v, 0); 885 hv_store_ent (hv, k, v, 0);
656 SvREFCNT_dec (k); 886 SvREFCNT_dec (k);
657 } 887
888fail:
889 ;
658} 890}
659 891
660static SV * 892static SV *
661decode_hv (dec_t *dec) 893decode_hv (dec_t *dec)
662{ 894{
663 HV *hv = newHV (); 895 HV *hv = newHV ();
664 896
665 DEC_INC_DEPTH; 897 DEC_INC_DEPTH;
666 898
667 if ((*dec->cur & 31) == 31) 899 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
668 { 900 {
669 ++dec->cur; 901 ++dec->cur;
670 902
671 for (;;) 903 for (;;)
672 { 904 {
673 WANT (1); 905 WANT (1);
674 906
675 if (*dec->cur == (0xe0 | 31)) 907 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
676 { 908 {
677 ++dec->cur; 909 ++dec->cur;
678 break; 910 break;
679 } 911 }
680 912
681 decode_he (dec, hv); 913 decode_he (dec, hv);
682 } 914 }
683 } 915 }
684 else 916 else
685 { 917 {
686 int pairs = decode_uint (dec); 918 UV pairs = decode_uint (dec);
919
920 WANT (pairs); // complexity check - need at least one byte per value, do not allow supersize hashes
687 921
688 while (pairs--) 922 while (pairs--)
689 decode_he (dec, hv); 923 decode_he (dec, hv);
690 } 924 }
691 925
701static SV * 935static SV *
702decode_str (dec_t *dec, int utf8) 936decode_str (dec_t *dec, int utf8)
703{ 937{
704 SV *sv = 0; 938 SV *sv = 0;
705 939
706 if ((*dec->cur & 31) == 31) 940 if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
707 { 941 {
942 // indefinite length strings
708 ++dec->cur; 943 ++dec->cur;
709 944
945 U8 major = *dec->cur & MAJOR_MISC;
946
710 sv = newSVpvn ("", 0); 947 sv = newSVpvn ("", 0);
711 948
712 // not very fast, and certainly not robust against illegal input
713 for (;;) 949 for (;;)
714 { 950 {
715 WANT (1); 951 WANT (1);
716 952
717 if (*dec->cur == (0xe0 | 31)) 953 if ((*dec->cur - major) > LENGTH_EXT8)
954 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
718 { 955 {
719 ++dec->cur; 956 ++dec->cur;
720 break; 957 break;
721 } 958 }
959 else
960 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
722 961
723 sv_catsv (sv, decode_sv (dec)); 962 STRLEN len = decode_uint (dec);
963
964 WANT (len);
965 sv_catpvn (sv, dec->cur, len);
966 dec->cur += len;
724 } 967 }
725 } 968 }
726 else 969 else
727 { 970 {
728 STRLEN len = decode_uint (dec); 971 STRLEN len = decode_uint (dec);
729 972
730 WANT (len); 973 WANT (len);
731 sv = newSVpvn (dec->cur, len); 974 sv = newSVpvn (dec->cur, len);
732 dec->cur += len; 975 dec->cur += len;
976
977 if (ecb_expect_false (dec->stringref)
978 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
979 av_push (dec->stringref, SvREFCNT_inc_NN (sv));
733 } 980 }
734 981
735 if (utf8) 982 if (utf8)
983 {
984 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
985 if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
986 ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
987
736 SvUTF8_on (sv); 988 SvUTF8_on (sv);
989 }
737 990
738 return sv; 991 return sv;
739 992
740fail: 993fail:
741 SvREFCNT_dec (sv); 994 SvREFCNT_dec (sv);
743} 996}
744 997
745static SV * 998static SV *
746decode_tagged (dec_t *dec) 999decode_tagged (dec_t *dec)
747{ 1000{
1001 SV *sv = 0;
748 UV tag = decode_uint (dec); 1002 UV tag = decode_uint (dec);
749 SV *sv = decode_sv (dec); 1003
1004 WANT (1);
750 1005
751 switch (tag) 1006 switch (tag)
752 { 1007 {
753 case CBOR_TAG_MAGIC: 1008 case CBOR_TAG_MAGIC:
754 return sv; 1009 sv = decode_sv (dec);
1010 break;
755 1011
756 case CBOR_TAG_INDIRECTION: 1012 case CBOR_TAG_INDIRECTION:
757 return newRV_noinc (sv); 1013 sv = newRV_noinc (decode_sv (dec));
1014 break;
758 1015
1016 case CBOR_TAG_STRINGREF_NAMESPACE:
1017 {
1018 // do nmot use SAVETMPS/FREETMPS, as these will
1019 // erase mortalised caches, e.g. "shareable"
1020 ENTER;
1021
1022 SAVESPTR (dec->stringref);
1023 dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());
1024
1025 sv = decode_sv (dec);
1026
1027 LEAVE;
1028 }
1029 break;
1030
1031 case CBOR_TAG_STRINGREF:
1032 {
1033 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
1034 ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
1035
1036 UV idx = decode_uint (dec);
1037
1038 if (!dec->stringref || (int)idx > AvFILLp (dec->stringref))
1039 ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)");
1040
1041 sv = newSVsv (AvARRAY (dec->stringref)[idx]);
1042 }
1043 break;
1044
759 case CBOR_TAG_VALUE_SHARABLE: 1045 case CBOR_TAG_VALUE_SHAREABLE:
1046 {
760 if (ecb_expect_false (!dec->sharable)) 1047 if (ecb_expect_false (!dec->shareable))
761 dec->sharable = (AV *)sv_2mortal ((SV *)newAV ()); 1048 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
762 1049
1050 if (dec->cbor.flags & F_ALLOW_CYCLES)
1051 {
1052 sv = newSV (0);
763 av_push (dec->sharable, SvREFCNT_inc_NN (sv)); 1053 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
764 1054
765 return sv; 1055 SV *osv = decode_sv (dec);
1056 sv_setsv (sv, osv);
1057 SvREFCNT_dec_NN (osv);
1058 }
1059 else
1060 {
1061 av_push (dec->shareable, &PL_sv_undef);
1062 int idx = AvFILLp (dec->shareable);
1063 sv = decode_sv (dec);
1064 av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
1065 }
1066 }
1067 break;
766 1068
767 case CBOR_TAG_VALUE_SHAREDREF: 1069 case CBOR_TAG_VALUE_SHAREDREF:
768 { 1070 {
769 // TODO: should verify that the sv atcually was a CBOR unsigned integer 1071 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
770 UV idx = SvUV (sv); 1072 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
771 1073
1074 UV idx = decode_uint (dec);
1075
772 if (!dec->sharable || idx > AvFILLp (dec->sharable)) 1076 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable))
773 ERR ("corrupted CBOR data (sharedref index out of bounds)"); 1077 ERR ("corrupted CBOR data (sharedref index out of bounds)");
774 1078
775 SvREFCNT_dec (sv);
776
777 return SvREFCNT_inc_NN (AvARRAY (dec->sharable)[idx]); 1079 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
1080
1081 if (sv == &PL_sv_undef)
1082 ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
778 } 1083 }
1084 break;
779 1085
780 case CBOR_TAG_PERL_OBJECT: 1086 case CBOR_TAG_PERL_OBJECT:
781 { 1087 {
1088 if (dec->cbor.flags & F_FORBID_OBJECTS)
1089 goto filter;
1090
1091 sv = decode_sv (dec);
1092
782 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) 1093 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
783 ERR ("corrupted CBOR data (non-array perl object)"); 1094 ERR ("corrupted CBOR data (non-array perl object)");
784 1095
785 AV *av = (AV *)SvRV (sv); 1096 AV *av = (AV *)SvRV (sv);
786 int len = av_len (av) + 1; 1097 int len = av_len (av) + 1;
794 if (!method) 1105 if (!method)
795 ERR ("cannot decode perl-object (package does not have a THAW method)"); 1106 ERR ("cannot decode perl-object (package does not have a THAW method)");
796 1107
797 dSP; 1108 dSP;
798 1109
799 ENTER; SAVETMPS; PUSHMARK (SP); 1110 ENTER; SAVETMPS;
1111 PUSHMARK (SP);
800 EXTEND (SP, len + 1); 1112 EXTEND (SP, len + 1);
801 // we re-bless the reference to get overload and other niceties right 1113 // we re-bless the reference to get overload and other niceties right
802 PUSHs (*av_fetch (av, 0, 1)); 1114 PUSHs (*av_fetch (av, 0, 1));
803 PUSHs (sv_cbor); 1115 PUSHs (sv_cbor);
804 1116
812 SPAGAIN; 1124 SPAGAIN;
813 1125
814 if (SvTRUE (ERRSV)) 1126 if (SvTRUE (ERRSV))
815 { 1127 {
816 FREETMPS; LEAVE; 1128 FREETMPS; LEAVE;
817 ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV)))); 1129 ERR_ERRSV;
818 } 1130 }
819 1131
820 SvREFCNT_dec (sv); 1132 SvREFCNT_dec (sv);
821 sv = SvREFCNT_inc (POPs); 1133 sv = SvREFCNT_inc (POPs);
822 1134
823 PUTBACK; 1135 PUTBACK;
824 1136
825 FREETMPS; LEAVE; 1137 FREETMPS; LEAVE;
826
827 return sv;
828 } 1138 }
1139 break;
829 1140
830 default: 1141 default:
1142 filter:
1143 {
1144 SV *tag_sv = newSVuv (tag);
1145
1146 sv = decode_sv (dec);
1147
1148 dSP;
1149 ENTER; SAVETMPS;
1150 SAVESTACK_POS ();
1151 PUSHMARK (SP);
1152 EXTEND (SP, 2);
1153 PUSHs (tag_sv);
1154 PUSHs (sv);
1155
1156 PUTBACK;
1157 int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL);
1158 SPAGAIN;
1159
1160 if (SvTRUE (ERRSV))
831 { 1161 {
1162 SvREFCNT_dec (tag_sv);
1163 FREETMPS; LEAVE;
1164 ERR_ERRSV;
1165 }
1166
1167 if (count)
1168 {
1169 SvREFCNT_dec (tag_sv);
1170 SvREFCNT_dec (sv);
1171 sv = SvREFCNT_inc (POPs);
1172 }
1173 else
1174 {
832 AV *av = newAV (); 1175 AV *av = newAV ();
833 av_push (av, newSVuv (tag)); 1176 av_push (av, tag_sv);
834 av_push (av, sv); 1177 av_push (av, sv);
835 1178
836 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 1179 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
837 ? cbor_tagged_stash 1180 ? cbor_tagged_stash
838 : gv_stashpv ("CBOR::XS::Tagged" , 1); 1181 : gv_stashpv ("CBOR::XS::Tagged" , 1);
839
840 return sv_bless (newRV_noinc ((SV *)av), tagged_stash); 1182 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
841 } 1183 }
1184
1185 PUTBACK;
1186
1187 FREETMPS; LEAVE;
1188 }
1189 break;
842 } 1190 }
1191
1192 return sv;
843 1193
844fail: 1194fail:
845 SvREFCNT_dec (sv); 1195 SvREFCNT_dec (sv);
846 return &PL_sv_undef; 1196 return &PL_sv_undef;
847} 1197}
849static SV * 1199static SV *
850decode_sv (dec_t *dec) 1200decode_sv (dec_t *dec)
851{ 1201{
852 WANT (1); 1202 WANT (1);
853 1203
854 switch (*dec->cur >> 5) 1204 switch (*dec->cur >> MAJOR_SHIFT)
855 { 1205 {
856 case 0: // unsigned int 1206 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
857 return newSVuv (decode_uint (dec)); 1207 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
858 case 1: // negative int 1208 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
859 return newSViv (-1 - (IV)decode_uint (dec)); 1209 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
860 case 2: // octet string 1210 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
861 return decode_str (dec, 0); 1211 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
862 case 3: // utf-8 string 1212 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
863 return decode_str (dec, 1); 1213
864 case 4: // array 1214 case MAJOR_MISC >> MAJOR_SHIFT:
865 return decode_av (dec);
866 case 5: // map
867 return decode_hv (dec);
868 case 6: // tag
869 return decode_tagged (dec);
870 case 7: // misc
871 switch (*dec->cur++ & 31) 1215 switch (*dec->cur++ & MINOR_MASK)
872 { 1216 {
873 case 20: 1217 case SIMPLE_FALSE:
874#if CBOR_SLOW 1218#if CBOR_SLOW
875 types_false = get_bool ("Types::Serialiser::false"); 1219 types_false = get_bool ("Types::Serialiser::false");
876#endif 1220#endif
877 return newSVsv (types_false); 1221 return newSVsv (types_false);
878 case 21: 1222 case SIMPLE_TRUE:
879#if CBOR_SLOW 1223#if CBOR_SLOW
880 types_true = get_bool ("Types::Serialiser::true"); 1224 types_true = get_bool ("Types::Serialiser::true");
881#endif 1225#endif
882 return newSVsv (types_true); 1226 return newSVsv (types_true);
883 case 22: 1227 case SIMPLE_NULL:
884 return newSVsv (&PL_sv_undef); 1228 return newSVsv (&PL_sv_undef);
885 case 23: 1229 case SIMPLE_UNDEF:
886#if CBOR_SLOW 1230#if CBOR_SLOW
887 types_error = get_bool ("Types::Serialiser::error"); 1231 types_error = get_bool ("Types::Serialiser::error");
888#endif 1232#endif
889 return newSVsv (types_error); 1233 return newSVsv (types_error);
890 1234
891 case 25: 1235 case MISC_FLOAT16:
892 { 1236 {
893 WANT (2); 1237 WANT (2);
894 1238
895 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1239 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
896 dec->cur += 2; 1240 dec->cur += 2;
897 1241
898 return newSVnv (ecb_binary16_to_float (fp)); 1242 return newSVnv (ecb_binary16_to_float (fp));
899 } 1243 }
900 1244
901 case 26: 1245 case MISC_FLOAT32:
902 { 1246 {
903 uint32_t fp; 1247 uint32_t fp;
904 WANT (4); 1248 WANT (4);
905 memcpy (&fp, dec->cur, 4); 1249 memcpy (&fp, dec->cur, 4);
906 dec->cur += 4; 1250 dec->cur += 4;
909 fp = ecb_bswap32 (fp); 1253 fp = ecb_bswap32 (fp);
910 1254
911 return newSVnv (ecb_binary32_to_float (fp)); 1255 return newSVnv (ecb_binary32_to_float (fp));
912 } 1256 }
913 1257
914 case 27: 1258 case MISC_FLOAT64:
915 { 1259 {
916 uint64_t fp; 1260 uint64_t fp;
917 WANT (8); 1261 WANT (8);
918 memcpy (&fp, dec->cur, 8); 1262 memcpy (&fp, dec->cur, 8);
919 dec->cur += 8; 1263 dec->cur += 8;
922 fp = ecb_bswap64 (fp); 1266 fp = ecb_bswap64 (fp);
923 1267
924 return newSVnv (ecb_binary64_to_double (fp)); 1268 return newSVnv (ecb_binary64_to_double (fp));
925 } 1269 }
926 1270
927 // 0..19 unassigned 1271 // 0..19 unassigned simple
928 // 24 reserved + unassigned (reserved values are not encodable) 1272 // 24 reserved + unassigned simple (reserved values are not encodable)
1273 // 28-30 unassigned misc
1274 // 31 break code
929 default: 1275 default:
930 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1276 ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
931 } 1277 }
932 1278
933 break; 1279 break;
934 } 1280 }
935 1281
938} 1284}
939 1285
940static SV * 1286static SV *
941decode_cbor (SV *string, CBOR *cbor, char **offset_return) 1287decode_cbor (SV *string, CBOR *cbor, char **offset_return)
942{ 1288{
943 dec_t dec = { }; 1289 dec_t dec = { 0 };
944 SV *sv; 1290 SV *sv;
945 STRLEN len; 1291 STRLEN len;
946 char *data = SvPVbyte (string, len); 1292 char *data = SvPVbyte (string, len);
947 1293
948 if (len > cbor->max_size && cbor->max_size) 1294 if (len > cbor->max_size && cbor->max_size)
962 if (dec.cur != dec.end && !dec.err) 1308 if (dec.cur != dec.end && !dec.err)
963 dec.err = "garbage after CBOR object"; 1309 dec.err = "garbage after CBOR object";
964 1310
965 if (dec.err) 1311 if (dec.err)
966 { 1312 {
1313 if (dec.shareable)
1314 {
1315 // need to break cyclic links, which would all be in shareable
1316 int i;
1317 SV **svp;
1318
1319 for (i = av_len (dec.shareable) + 1; i--; )
1320 if ((svp = av_fetch (dec.shareable, i, 0)))
1321 sv_setsv (*svp, &PL_sv_undef);
1322 }
1323
967 SvREFCNT_dec (sv); 1324 SvREFCNT_dec (sv);
1325
1326 if (dec.err_sv)
1327 sv_2mortal (dec.err_sv);
1328
968 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur); 1329 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur);
969 } 1330 }
970 1331
971 sv = sv_2mortal (sv); 1332 sv = sv_2mortal (sv);
972 1333
973 return sv; 1334 return sv;
974} 1335}
975 1336
1337/////////////////////////////////////////////////////////////////////////////
1338// incremental parser
1339
1340#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
1341
1342// returns 0 for notyet, 1 for success or error
1343static int
1344incr_parse (CBOR *self, SV *cborstr)
1345{
1346 STRLEN cur;
1347 SvPV (cborstr, cur);
1348
1349 while (ecb_expect_true (self->incr_need <= cur))
1350 {
1351 // table of integer count bytes
1352 static I8 incr_len[MINOR_MASK + 1] = {
1353 0, 0, 0, 0, 0, 0, 0, 0,
1354 0, 0, 0, 0, 0, 0, 0, 0,
1355 0, 0, 0, 0, 0, 0, 0, 0,
1356 1, 2, 4, 8,-1,-1,-1,-2
1357 };
1358
1359 const U8 *p = SvPVX (cborstr) + self->incr_pos;
1360 U8 m = *p & MINOR_MASK;
1361 IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1362 I8 ilen = incr_len[m];
1363
1364 self->incr_need = self->incr_pos + 1;
1365
1366 if (ecb_expect_false (ilen < 0))
1367 {
1368 if (m != MINOR_INDEF)
1369 return 1; // error
1370
1371 if (*p == (MAJOR_MISC | MINOR_INDEF))
1372 {
1373 if (count >= 0)
1374 return 1; // error
1375
1376 count = 1;
1377 }
1378 else
1379 {
1380 av_push (self->incr_count, newSViv (-1)); //TODO: nest
1381 count = -1;
1382 }
1383 }
1384 else
1385 {
1386 self->incr_need += ilen;
1387 if (ecb_expect_false (self->incr_need > cur))
1388 return 0;
1389
1390 int major = *p >> MAJOR_SHIFT;
1391
1392 switch (major)
1393 {
1394 case MAJOR_TAG >> MAJOR_SHIFT:
1395 ++count; // tags merely prefix another value
1396 break;
1397
1398 case MAJOR_BYTES >> MAJOR_SHIFT:
1399 case MAJOR_TEXT >> MAJOR_SHIFT:
1400 case MAJOR_ARRAY >> MAJOR_SHIFT:
1401 case MAJOR_MAP >> MAJOR_SHIFT:
1402 {
1403 UV len;
1404
1405 if (ecb_expect_false (ilen))
1406 {
1407 len = 0;
1408
1409 do {
1410 len = (len << 8) | *++p;
1411 } while (--ilen);
1412 }
1413 else
1414 len = m;
1415
1416 switch (major)
1417 {
1418 case MAJOR_BYTES >> MAJOR_SHIFT:
1419 case MAJOR_TEXT >> MAJOR_SHIFT:
1420 self->incr_need += len;
1421 if (ecb_expect_false (self->incr_need > cur))
1422 return 0;
1423
1424 break;
1425
1426 case MAJOR_MAP >> MAJOR_SHIFT:
1427 len <<= 1;
1428 case MAJOR_ARRAY >> MAJOR_SHIFT:
1429 if (len)
1430 {
1431 av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
1432 count = len + 1;
1433 }
1434 break;
1435 }
1436 }
1437 }
1438 }
1439
1440 self->incr_pos = self->incr_need;
1441
1442 if (count > 0)
1443 {
1444 while (!--count)
1445 {
1446 if (!AvFILLp (self->incr_count))
1447 return 1; // done
1448
1449 SvREFCNT_dec_NN (av_pop (self->incr_count));
1450 count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1451 }
1452
1453 SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
1454 }
1455 }
1456
1457 return 0;
1458}
1459
1460
976///////////////////////////////////////////////////////////////////////////// 1461/////////////////////////////////////////////////////////////////////////////
977// XS interface functions 1462// XS interface functions
978 1463
979MODULE = CBOR::XS PACKAGE = CBOR::XS 1464MODULE = CBOR::XS PACKAGE = CBOR::XS
980 1465
988 1473
989 types_true = get_bool ("Types::Serialiser::true" ); 1474 types_true = get_bool ("Types::Serialiser::true" );
990 types_false = get_bool ("Types::Serialiser::false"); 1475 types_false = get_bool ("Types::Serialiser::false");
991 types_error = get_bool ("Types::Serialiser::error"); 1476 types_error = get_bool ("Types::Serialiser::error");
992 1477
1478 default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1479
993 sv_cbor = newSVpv ("CBOR", 0); 1480 sv_cbor = newSVpv ("CBOR", 0);
994 SvREADONLY_on (sv_cbor); 1481 SvREADONLY_on (sv_cbor);
1482
1483 assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1));
995} 1484}
996 1485
997PROTOTYPES: DISABLE 1486PROTOTYPES: DISABLE
998 1487
999void CLONE (...) 1488void CLONE (...)
1018void shrink (CBOR *self, int enable = 1) 1507void shrink (CBOR *self, int enable = 1)
1019 ALIAS: 1508 ALIAS:
1020 shrink = F_SHRINK 1509 shrink = F_SHRINK
1021 allow_unknown = F_ALLOW_UNKNOWN 1510 allow_unknown = F_ALLOW_UNKNOWN
1022 allow_sharing = F_ALLOW_SHARING 1511 allow_sharing = F_ALLOW_SHARING
1023 dedup_keys = F_DEDUP_KEYS 1512 allow_cycles = F_ALLOW_CYCLES
1513 forbid_objects = F_FORBID_OBJECTS
1024 dedup_strings = F_DEDUP_STRINGS 1514 pack_strings = F_PACK_STRINGS
1515 text_keys = F_TEXT_KEYS
1516 text_strings = F_TEXT_STRINGS
1517 validate_utf8 = F_VALIDATE_UTF8
1025 PPCODE: 1518 PPCODE:
1026{ 1519{
1027 if (enable) 1520 if (enable)
1028 self->flags |= ix; 1521 self->flags |= ix;
1029 else 1522 else
1035void get_shrink (CBOR *self) 1528void get_shrink (CBOR *self)
1036 ALIAS: 1529 ALIAS:
1037 get_shrink = F_SHRINK 1530 get_shrink = F_SHRINK
1038 get_allow_unknown = F_ALLOW_UNKNOWN 1531 get_allow_unknown = F_ALLOW_UNKNOWN
1039 get_allow_sharing = F_ALLOW_SHARING 1532 get_allow_sharing = F_ALLOW_SHARING
1040 get_dedup_keys = F_DEDUP_KEYS 1533 get_allow_cycles = F_ALLOW_CYCLES
1534 get_forbid_objects = F_FORBID_OBJECTS
1041 get_dedup_strings = F_DEDUP_STRINGS 1535 get_pack_strings = F_PACK_STRINGS
1536 get_text_keys = F_TEXT_KEYS
1537 get_text_strings = F_TEXT_STRINGS
1538 get_validate_utf8 = F_VALIDATE_UTF8
1042 PPCODE: 1539 PPCODE:
1043 XPUSHs (boolSV (self->flags & ix)); 1540 XPUSHs (boolSV (self->flags & ix));
1044 1541
1045void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1542void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1046 PPCODE: 1543 PPCODE:
1059 XPUSHs (ST (0)); 1556 XPUSHs (ST (0));
1060 1557
1061int get_max_size (CBOR *self) 1558int get_max_size (CBOR *self)
1062 CODE: 1559 CODE:
1063 RETVAL = self->max_size; 1560 RETVAL = self->max_size;
1561 OUTPUT:
1562 RETVAL
1563
1564void filter (CBOR *self, SV *filter = 0)
1565 PPCODE:
1566 SvREFCNT_dec (self->filter);
1567 self->filter = filter ? newSVsv (filter) : filter;
1568 XPUSHs (ST (0));
1569
1570SV *get_filter (CBOR *self)
1571 CODE:
1572 RETVAL = self->filter ? self->filter : NEWSV (0, 0);
1064 OUTPUT: 1573 OUTPUT:
1065 RETVAL 1574 RETVAL
1066 1575
1067void encode (CBOR *self, SV *scalar) 1576void encode (CBOR *self, SV *scalar)
1068 PPCODE: 1577 PPCODE:
1083 EXTEND (SP, 2); 1592 EXTEND (SP, 2);
1084 PUSHs (sv); 1593 PUSHs (sv);
1085 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1594 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1086} 1595}
1087 1596
1597void incr_parse (CBOR *self, SV *cborstr)
1598 ALIAS:
1599 incr_parse_multiple = 1
1600 PPCODE:
1601{
1602 if (SvUTF8 (cborstr))
1603 sv_utf8_downgrade (cborstr, 0);
1604
1605 if (!self->incr_count)
1606 {
1607 self->incr_count = newAV ();
1608 self->incr_pos = 0;
1609 self->incr_need = 1;
1610
1611 av_push (self->incr_count, newSViv (1));
1612 }
1613
1614 do
1615 {
1616 if (!incr_parse (self, cborstr))
1617 {
1618 if (self->incr_need > self->max_size && self->max_size)
1619 croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
1620 (unsigned long)self->incr_need, (unsigned long)self->max_size);
1621
1622 break;
1623 }
1624
1625 SV *sv;
1626 char *offset;
1627
1628 PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
1629 XPUSHs (sv);
1630
1631 sv_chop (cborstr, offset);
1632
1633 av_clear (self->incr_count);
1634 av_push (self->incr_count, newSViv (1));
1635
1636 self->incr_pos = 0;
1637 self->incr_need = self->incr_pos + 1;
1638 }
1639 while (ix);
1640}
1641
1642void incr_reset (CBOR *self)
1643 CODE:
1644{
1645 SvREFCNT_dec (self->incr_count);
1646 self->incr_count = 0;
1647}
1648
1649void DESTROY (CBOR *self)
1650 PPCODE:
1651 cbor_free (self);
1652
1088PROTOTYPES: ENABLE 1653PROTOTYPES: ENABLE
1089 1654
1090void encode_cbor (SV *scalar) 1655void encode_cbor (SV *scalar)
1656 ALIAS:
1657 encode_cbor = 0
1658 encode_cbor_sharing = F_ALLOW_SHARING
1091 PPCODE: 1659 PPCODE:
1092{ 1660{
1093 CBOR cbor; 1661 CBOR cbor;
1094 cbor_init (&cbor); 1662 cbor_init (&cbor);
1663 cbor.flags |= ix;
1095 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1664 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1096 XPUSHs (scalar); 1665 XPUSHs (scalar);
1097} 1666}
1098 1667
1099void decode_cbor (SV *cborstr) 1668void decode_cbor (SV *cborstr)

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines