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 1.64 by root, Tue Jun 27 02:03:24 2017 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_SHARABLE = 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
59#define F_SHRINK 0x00000001UL 104#define F_SHRINK 0x00000001UL
60#define F_ALLOW_UNKNOWN 0x00000002UL 105#define F_ALLOW_UNKNOWN 0x00000002UL
61#define F_ALLOW_SHARING 0x00000004UL //TODO 106#define F_ALLOW_SHARING 0x00000004UL
107#define F_ALLOW_CYCLES 0x00000008UL
108#define F_FORBID_OBJECTS 0x00000010UL
62#define F_DEDUP_STRINGS 0x00000008UL //TODO 109#define F_PACK_STRINGS 0x00000020UL
63#define F_DEDUP_KEYS 0x00000010UL //TODO 110#define F_TEXT_KEYS 0x00000040UL
111#define F_TEXT_STRINGS 0x00000080UL
112#define F_VALIDATE_UTF8 0x00000100UL
64 113
65#define INIT_SIZE 32 // initial scalar size to be allocated 114#define INIT_SIZE 32 // initial scalar size to be allocated
66 115
67#define SB do { 116#define SB do {
68#define SE } while (0) 117#define SE } while (0)
80# define CBOR_SLOW 0 129# define CBOR_SLOW 0
81# define CBOR_STASH cbor_stash 130# define CBOR_STASH cbor_stash
82#endif 131#endif
83 132
84static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS:: 133static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS::
85static SV *types_true, *types_false, *types_error, *sv_cbor; 134static SV *types_true, *types_false, *types_error, *sv_cbor, *default_filter;
86 135
87typedef struct { 136typedef struct {
88 U32 flags; 137 U32 flags;
89 U32 max_depth; 138 U32 max_depth;
90 STRLEN max_size; 139 STRLEN max_size;
140 SV *filter;
141
142 // for the incremental parser
143 STRLEN incr_pos; // the current offset into the text
144 STRLEN incr_need; // minimum bytes needed to decode
145 AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef.
91} CBOR; 146} CBOR;
92 147
93ecb_inline void 148ecb_inline void
94cbor_init (CBOR *cbor) 149cbor_init (CBOR *cbor)
95{ 150{
96 Zero (cbor, 1, CBOR); 151 Zero (cbor, 1, CBOR);
97 cbor->max_depth = 512; 152 cbor->max_depth = 512;
153}
154
155ecb_inline void
156cbor_free (CBOR *cbor)
157{
158 SvREFCNT_dec (cbor->filter);
159 SvREFCNT_dec (cbor->incr_count);
98} 160}
99 161
100///////////////////////////////////////////////////////////////////////////// 162/////////////////////////////////////////////////////////////////////////////
101// utility functions 163// utility functions
102 164
124 SvPV_renew (sv, SvCUR (sv) + 1); 186 SvPV_renew (sv, SvCUR (sv) + 1);
125#endif 187#endif
126 } 188 }
127} 189}
128 190
191// minimum length of a string to be registered for stringref
192ecb_inline int
193minimum_string_length (UV idx)
194{
195 return idx <= 23 ? 3
196 : idx <= 0xffU ? 4
197 : idx <= 0xffffU ? 5
198 : idx <= 0xffffffffU ? 7
199 : 11;
200}
201
129///////////////////////////////////////////////////////////////////////////// 202/////////////////////////////////////////////////////////////////////////////
130// encoder 203// encoder
131 204
132// structure used for encoding CBOR 205// structure used for encoding CBOR
133typedef struct 206typedef struct
135 char *cur; // SvPVX (sv) + current output position 208 char *cur; // SvPVX (sv) + current output position
136 char *end; // SvEND (sv) 209 char *end; // SvEND (sv)
137 SV *sv; // result scalar 210 SV *sv; // result scalar
138 CBOR cbor; 211 CBOR cbor;
139 U32 depth; // recursion level 212 U32 depth; // recursion level
140 HV *stringref; // string => index, or 0 213 HV *stringref[2]; // string => index, or 0 ([0] = bytes, [1] = utf-8)
214 UV stringref_idx;
141 HV *sharable; // ptr => index, or 0 215 HV *shareable; // ptr => index, or 0
142 HV *sharable_idx; 216 UV shareable_idx;
143} enc_t; 217} enc_t;
144 218
145ecb_inline void 219ecb_inline void
146need (enc_t *enc, STRLEN len) 220need (enc_t *enc, STRLEN len)
147{ 221{
148 if (ecb_expect_false (enc->cur + len >= enc->end)) 222 if (ecb_expect_false ((uintptr_t)(enc->end - enc->cur) < len))
149 { 223 {
150 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv); 224 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv);
151 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1); 225 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1);
152 enc->cur = SvPVX (enc->sv) + cur; 226 enc->cur = SvPVX (enc->sv) + cur;
153 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1; 227 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;
164static void 238static void
165encode_uint (enc_t *enc, int major, UV len) 239encode_uint (enc_t *enc, int major, UV len)
166{ 240{
167 need (enc, 9); 241 need (enc, 9);
168 242
169 if (len < 24) 243 if (ecb_expect_true (len < LENGTH_EXT1))
170 *enc->cur++ = major | len; 244 *enc->cur++ = major | len;
171 else if (len <= 0xff) 245 else if (ecb_expect_true (len <= 0xffU))
172 { 246 {
173 *enc->cur++ = major | 24; 247 *enc->cur++ = major | LENGTH_EXT1;
174 *enc->cur++ = len; 248 *enc->cur++ = len;
175 } 249 }
176 else if (len <= 0xffff) 250 else if (len <= 0xffffU)
177 { 251 {
178 *enc->cur++ = major | 25; 252 *enc->cur++ = major | LENGTH_EXT2;
179 *enc->cur++ = len >> 8; 253 *enc->cur++ = len >> 8;
180 *enc->cur++ = len; 254 *enc->cur++ = len;
181 } 255 }
182 else if (len <= 0xffffffff) 256 else if (len <= 0xffffffffU)
183 { 257 {
184 *enc->cur++ = major | 26; 258 *enc->cur++ = major | LENGTH_EXT4;
185 *enc->cur++ = len >> 24; 259 *enc->cur++ = len >> 24;
186 *enc->cur++ = len >> 16; 260 *enc->cur++ = len >> 16;
187 *enc->cur++ = len >> 8; 261 *enc->cur++ = len >> 8;
188 *enc->cur++ = len; 262 *enc->cur++ = len;
189 } 263 }
190 else 264 else
191 { 265 {
192 *enc->cur++ = major | 27; 266 *enc->cur++ = major | LENGTH_EXT8;
193 *enc->cur++ = len >> 56; 267 *enc->cur++ = len >> 56;
194 *enc->cur++ = len >> 48; 268 *enc->cur++ = len >> 48;
195 *enc->cur++ = len >> 40; 269 *enc->cur++ = len >> 40;
196 *enc->cur++ = len >> 32; 270 *enc->cur++ = len >> 32;
197 *enc->cur++ = len >> 24; 271 *enc->cur++ = len >> 24;
199 *enc->cur++ = len >> 8; 273 *enc->cur++ = len >> 8;
200 *enc->cur++ = len; 274 *enc->cur++ = len;
201 } 275 }
202} 276}
203 277
204static void 278ecb_inline void
279encode_tag (enc_t *enc, UV tag)
280{
281 encode_uint (enc, MAJOR_TAG, tag);
282}
283
284// exceptional (hopefully) slow path for byte strings that need to be utf8-encoded
285ecb_noinline static void
205encode_str (enc_t *enc, int utf8, char *str, STRLEN len) 286encode_str_utf8 (enc_t *enc, int utf8, char *str, STRLEN len)
206{ 287{
207 encode_uint (enc, utf8 ? 0x60 : 0x40, len); 288 STRLEN ulen = len;
289 U8 *p, *pend = (U8 *)str + len;
290
291 for (p = (U8 *)str; p < pend; ++p)
292 ulen += *p >> 7; // count set high bits
293
294 encode_uint (enc, MAJOR_TEXT, ulen);
295
296 need (enc, ulen);
297 for (p = (U8 *)str; p < pend; ++p)
298 if (*p < 0x80)
299 *enc->cur++ = *p;
300 else
301 {
302 *enc->cur++ = 0xc0 + (*p >> 6);
303 *enc->cur++ = 0x80 + (*p & 63);
304 }
305}
306
307ecb_inline void
308encode_str (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
309{
310 if (ecb_expect_false (upgrade_utf8))
311 if (!utf8)
312 {
313 encode_str_utf8 (enc, utf8, str, len);
314 return;
315 }
316
317 encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
208 need (enc, len); 318 need (enc, len);
209 memcpy (enc->cur, str, len); 319 memcpy (enc->cur, str, len);
210 enc->cur += len; 320 enc->cur += len;
211} 321}
212 322
213ecb_inline void 323ecb_inline void
214encode_tag (enc_t *enc, UV tag) 324encode_strref (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
215{ 325{
216 encode_uint (enc, 0xc0, tag); 326 if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
217} 327 {
328 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
218 329
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)) 330 if (SvOK (*svp))
228 { 331 {
332 // already registered, use stringref
229 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF); 333 encode_tag (enc, CBOR_TAG_STRINGREF);
230 encode_uint (enc, 0x00, SvUV (*svp)); 334 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
231
232 return 1; 335 return;
233 } 336 }
234 else 337 else if (len >= minimum_string_length (enc->stringref_idx))
235 { 338 {
339 // register only
236 sv_setuv (*svp, enc->sharable_idx++); 340 sv_setuv (*svp, enc->stringref_idx);
237 encode_tag (enc, CBOR_TAG_VALUE_SHARABLE); 341 ++enc->stringref_idx;
238 342 }
239 return 0;
240 } 343 }
241}
242 344
243ecb_inline int 345 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} 346}
252 347
253static void encode_sv (enc_t *enc, SV *sv); 348static void encode_sv (enc_t *enc, SV *sv);
254 349
255static void 350static void
260 if (enc->depth >= enc->cbor.max_depth) 355 if (enc->depth >= enc->cbor.max_depth)
261 croak (ERR_NESTING_EXCEEDED); 356 croak (ERR_NESTING_EXCEEDED);
262 357
263 ++enc->depth; 358 ++enc->depth;
264 359
265 encode_uint (enc, 0x80, len + 1); 360 encode_uint (enc, MAJOR_ARRAY, len + 1);
266 361
362 if (ecb_expect_false (SvMAGICAL (av)))
267 for (i = 0; i <= len; ++i) 363 for (i = 0; i <= len; ++i)
268 { 364 {
269 SV **svp = av_fetch (av, i, 0); 365 SV **svp = av_fetch (av, i, 0);
270 encode_sv (enc, svp ? *svp : &PL_sv_undef); 366 encode_sv (enc, svp ? *svp : &PL_sv_undef);
271 } 367 }
368 else
369 for (i = 0; i <= len; ++i)
370 {
371 SV *sv = AvARRAY (av)[i];
372 encode_sv (enc, sv ? sv : &PL_sv_undef);
373 }
272 374
273 --enc->depth; 375 --enc->depth;
274} 376}
275 377
276static void 378static void
284 ++enc->depth; 386 ++enc->depth;
285 387
286 int pairs = hv_iterinit (hv); 388 int pairs = hv_iterinit (hv);
287 int mg = SvMAGICAL (hv); 389 int mg = SvMAGICAL (hv);
288 390
289 if (mg) 391 if (ecb_expect_false (mg))
290 encode_ch (enc, 0xa0 | 31); 392 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
291 else 393 else
292 encode_uint (enc, 0xa0, pairs); 394 encode_uint (enc, MAJOR_MAP, pairs);
293 395
294 while ((he = hv_iternext (hv))) 396 while ((he = hv_iternext (hv)))
295 { 397 {
296 if (HeKLEN (he) == HEf_SVKEY) 398 if (HeKLEN (he) == HEf_SVKEY)
297 encode_sv (enc, HeSVKEY (he)); 399 encode_sv (enc, HeSVKEY (he));
298 else 400 else
299 encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); 401 encode_strref (enc, enc->cbor.flags & (F_TEXT_KEYS | F_TEXT_STRINGS), HeKUTF8 (he), HeKEY (he), HeKLEN (he));
300 402
301 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); 403 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
302 } 404 }
303 405
304 if (mg) 406 if (ecb_expect_false (mg))
305 encode_ch (enc, 0xe0 | 31); 407 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
306 408
307 --enc->depth; 409 --enc->depth;
308} 410}
309 411
310// encode objects, arrays and special \0=false and \1=true values. 412// encode objects, arrays and special \0=false and \1=true values.
311static void 413static void
312encode_rv (enc_t *enc, SV *sv) 414encode_rv (enc_t *enc, SV *sv)
313{ 415{
314 svtype svt;
315
316 SvGETMAGIC (sv); 416 SvGETMAGIC (sv);
417
317 svt = SvTYPE (sv); 418 svtype svt = SvTYPE (sv);
318
319 if (encode_sharable (enc, sv))
320 return;
321 419
322 if (ecb_expect_false (SvOBJECT (sv))) 420 if (ecb_expect_false (SvOBJECT (sv)))
323 { 421 {
324 HV *boolean_stash = !CBOR_SLOW || types_boolean_stash 422 HV *boolean_stash = !CBOR_SLOW || types_boolean_stash
325 ? types_boolean_stash 423 ? types_boolean_stash
330 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 428 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
331 ? cbor_tagged_stash 429 ? cbor_tagged_stash
332 : gv_stashpv ("CBOR::XS::Tagged" , 1); 430 : gv_stashpv ("CBOR::XS::Tagged" , 1);
333 431
334 HV *stash = SvSTASH (sv); 432 HV *stash = SvSTASH (sv);
335 GV *method;
336 433
337 if (stash == boolean_stash) 434 if (stash == boolean_stash)
338 encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20); 435 {
436 encode_ch (enc, SvIV (sv) ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE);
437 return;
438 }
339 else if (stash == error_stash) 439 else if (stash == error_stash)
340 encode_ch (enc, 0xe0 | 23); 440 {
441 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
442 return;
443 }
341 else if (stash == tagged_stash) 444 else if (stash == tagged_stash)
342 { 445 {
343 if (svt != SVt_PVAV) 446 if (svt != SVt_PVAV)
344 croak ("encountered CBOR::XS::Tagged object that isn't an array"); 447 croak ("encountered CBOR::XS::Tagged object that isn't an array");
345 448
346 encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); 449 encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
347 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); 450 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1));
451
452 return;
453 }
454 }
455
456 if (ecb_expect_false (SvREFCNT (sv) > 1)
457 && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
458 {
459 if (ecb_expect_false (!enc->shareable))
460 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
461
462 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
463
464 if (SvOK (*svp))
348 } 465 {
466 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
467 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
468 return;
469 }
470 else
471 {
472 sv_setuv (*svp, enc->shareable_idx);
473 ++enc->shareable_idx;
474 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
475 }
476 }
477
478 if (ecb_expect_false (SvOBJECT (sv)))
479 {
480 HV *stash = SvSTASH (sv);
481 GV *method;
482
483 if (enc->cbor.flags & F_FORBID_OBJECTS)
484 croak ("encountered object '%s', but forbid_objects is enabled",
485 SvPV_nolen (sv_2mortal (newRV_inc (sv))));
349 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) 486 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
350 { 487 {
351 dSP; 488 dSP;
352 489
353 ENTER; SAVETMPS; PUSHMARK (SP); 490 ENTER; SAVETMPS;
491 PUSHMARK (SP);
354 // we re-bless the reference to get overload and other niceties right 492 // we re-bless the reference to get overload and other niceties right
355 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 493 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
356 494
357 PUTBACK; 495 PUTBACK;
358 // G_SCALAR ensures that return value is 1 496 // G_SCALAR ensures that return value is 1
371 } 509 }
372 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0) 510 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
373 { 511 {
374 dSP; 512 dSP;
375 513
376 ENTER; SAVETMPS; PUSHMARK (SP); 514 ENTER; SAVETMPS;
515 SAVESTACK_POS ();
516 PUSHMARK (SP);
377 EXTEND (SP, 2); 517 EXTEND (SP, 2);
378 // we re-bless the reference to get overload and other niceties right 518 // we re-bless the reference to get overload and other niceties right
379 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 519 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
380 PUSHs (sv_cbor); 520 PUSHs (sv_cbor);
381 521
386 // catch this surprisingly common error 526 // catch this surprisingly common error
387 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) 527 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)); 528 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
389 529
390 encode_tag (enc, CBOR_TAG_PERL_OBJECT); 530 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
391 encode_uint (enc, 0x80, count + 1); 531 encode_uint (enc, MAJOR_ARRAY, count + 1);
392 encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); 532 encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
393 533
394 while (count) 534 while (count)
395 encode_sv (enc, SP[1 - count--]); 535 encode_sv (enc, SP[1 - count--]);
396 536
397 PUTBACK; 537 PUTBACK;
418{ 558{
419 double nv = SvNVX (sv); 559 double nv = SvNVX (sv);
420 560
421 need (enc, 9); 561 need (enc, 9);
422 562
423 if (ecb_expect_false (nv == (U32)nv)) 563 if (ecb_expect_false (nv == (NV)(U32)nv))
424 encode_uint (enc, 0x00, (U32)nv); 564 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
425 //TODO: maybe I32? 565 //TODO: maybe I32?
426 else if (ecb_expect_false (nv == (float)nv)) 566 else if (ecb_expect_false (nv == (float)nv))
427 { 567 {
568 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
569
428 uint32_t fp = ecb_float_to_binary32 (nv); 570 uint32_t fp = ecb_float_to_binary32 (nv);
429
430 *enc->cur++ = 0xe0 | 26;
431 571
432 if (!ecb_big_endian ()) 572 if (!ecb_big_endian ())
433 fp = ecb_bswap32 (fp); 573 fp = ecb_bswap32 (fp);
434 574
435 memcpy (enc->cur, &fp, 4); 575 memcpy (enc->cur, &fp, 4);
436 enc->cur += 4; 576 enc->cur += 4;
437 } 577 }
438 else 578 else
439 { 579 {
580 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
581
440 uint64_t fp = ecb_double_to_binary64 (nv); 582 uint64_t fp = ecb_double_to_binary64 (nv);
441
442 *enc->cur++ = 0xe0 | 27;
443 583
444 if (!ecb_big_endian ()) 584 if (!ecb_big_endian ())
445 fp = ecb_bswap64 (fp); 585 fp = ecb_bswap64 (fp);
446 586
447 memcpy (enc->cur, &fp, 8); 587 memcpy (enc->cur, &fp, 8);
452static void 592static void
453encode_sv (enc_t *enc, SV *sv) 593encode_sv (enc_t *enc, SV *sv)
454{ 594{
455 SvGETMAGIC (sv); 595 SvGETMAGIC (sv);
456 596
457 if (encode_sharable (enc, sv))
458 return;
459
460 if (SvPOKp (sv)) 597 if (SvPOKp (sv))
461 { 598 {
462 STRLEN len; 599 STRLEN len;
463 char *str = SvPV (sv, len); 600 char *str = SvPV (sv, len);
464 encode_str (enc, SvUTF8 (sv), str, len); 601 encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len);
465 } 602 }
466 else if (SvNOKp (sv)) 603 else if (SvNOKp (sv))
467 encode_nv (enc, sv); 604 encode_nv (enc, sv);
468 else if (SvIOKp (sv)) 605 else if (SvIOKp (sv))
469 { 606 {
470 if (SvIsUV (sv)) 607 if (SvIsUV (sv))
471 encode_uint (enc, 0x00, SvUVX (sv)); 608 encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
472 else if (SvIVX (sv) >= 0) 609 else if (SvIVX (sv) >= 0)
473 encode_uint (enc, 0x00, SvIVX (sv)); 610 encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
474 else 611 else
475 encode_uint (enc, 0x20, -(SvIVX (sv) + 1)); 612 encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
476 } 613 }
477 else if (SvROK (sv)) 614 else if (SvROK (sv))
478 encode_rv (enc, SvRV (sv)); 615 encode_rv (enc, SvRV (sv));
479 else if (!SvOK (sv)) 616 else if (!SvOK (sv))
480 encode_ch (enc, 0xe0 | 22); 617 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
481 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 618 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
482 encode_ch (enc, 0xe0 | 23); 619 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
483 else 620 else
484 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data", 621 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
485 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 622 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
486} 623}
487 624
488static SV * 625static SV *
489encode_cbor (SV *scalar, CBOR *cbor) 626encode_cbor (SV *scalar, CBOR *cbor)
490{ 627{
491 enc_t enc = { }; 628 enc_t enc = { 0 };
492 629
493 enc.cbor = *cbor; 630 enc.cbor = *cbor;
494 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 631 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
495 enc.cur = SvPVX (enc.sv); 632 enc.cur = SvPVX (enc.sv);
496 enc.end = SvEND (enc.sv); 633 enc.end = SvEND (enc.sv);
497 634
498 SvPOK_only (enc.sv); 635 SvPOK_only (enc.sv);
636
637 if (cbor->flags & F_PACK_STRINGS)
638 {
639 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
640 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
641 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
642 }
643
499 encode_sv (&enc, scalar); 644 encode_sv (&enc, scalar);
500 645
501 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv)); 646 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv));
502 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings 647 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings
503 648
517 U8 *end; // end of input string 662 U8 *end; // end of input string
518 const char *err; // parse error, if != 0 663 const char *err; // parse error, if != 0
519 CBOR cbor; 664 CBOR cbor;
520 U32 depth; // recursion depth 665 U32 depth; // recursion depth
521 U32 maxdepth; // recursion depth limit 666 U32 maxdepth; // recursion depth limit
522 AV *sharable; 667 AV *shareable;
668 AV *stringref;
669 SV *decode_tagged;
670 SV *err_sv; // optional sv for error, needs to be freed
523} dec_t; 671} dec_t;
524 672
525#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE 673// set dec->err to ERRSV
674ecb_cold static void
675err_errsv (dec_t *dec)
676{
677 if (!dec->err)
678 {
679 dec->err_sv = newSVsv (ERRSV);
526 680
527#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data") 681 // chop off the trailing \n
682 SvCUR_set (dec->err_sv, SvCUR (dec->err_sv) - 1);
683 *SvEND (dec->err_sv) = 0;
528 684
685 dec->err = SvPVutf8_nolen (dec->err_sv);
686 }
687}
688
689// the following functions are used to reduce code size and help the compiler to optimise
690ecb_cold static void
691err_set (dec_t *dec, const char *reason)
692{
693 if (!dec->err)
694 dec->err = reason;
695}
696
697ecb_cold static void
698err_unexpected_end (dec_t *dec)
699{
700 err_set (dec, "unexpected end of CBOR data");
701}
702
703#define ERR_DO(do) SB do; goto fail; SE
704#define ERR(reason) ERR_DO (err_set (dec, reason))
705#define ERR_ERRSV ERR_DO (err_errsv (dec))
706
707#define WANT(len) if (ecb_expect_false ((uintptr_t)(dec->end - dec->cur) < (STRLEN)len)) ERR_DO (err_unexpected_end (dec))
708
529#define DEC_INC_DEPTH if (++dec->depth > dec->cbor.max_depth) ERR (ERR_NESTING_EXCEEDED) 709#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 710#define DEC_DEC_DEPTH --dec->depth
531 711
532static UV 712static UV
533decode_uint (dec_t *dec) 713decode_uint (dec_t *dec)
534{ 714{
535 switch (*dec->cur & 31) 715 U8 m = *dec->cur & MINOR_MASK;
536 { 716 ++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;
541 717
542 case 24: 718 if (ecb_expect_true (m < LENGTH_EXT1))
719 return m;
720 else if (ecb_expect_true (m == LENGTH_EXT1))
721 {
543 WANT (2); 722 WANT (1);
544 dec->cur += 2; 723 dec->cur += 1;
545 return dec->cur[-1]; 724 return dec->cur[-1];
546 725 }
547 case 25: 726 else if (ecb_expect_true (m == LENGTH_EXT2))
727 {
548 WANT (3); 728 WANT (2);
549 dec->cur += 3; 729 dec->cur += 2;
550 return (((UV)dec->cur[-2]) << 8) 730 return (((UV)dec->cur[-2]) << 8)
551 | ((UV)dec->cur[-1]); 731 | ((UV)dec->cur[-1]);
552 732 }
553 case 26: 733 else if (ecb_expect_true (m == LENGTH_EXT4))
734 {
554 WANT (5); 735 WANT (4);
555 dec->cur += 5; 736 dec->cur += 4;
556 return (((UV)dec->cur[-4]) << 24) 737 return (((UV)dec->cur[-4]) << 24)
557 | (((UV)dec->cur[-3]) << 16) 738 | (((UV)dec->cur[-3]) << 16)
558 | (((UV)dec->cur[-2]) << 8) 739 | (((UV)dec->cur[-2]) << 8)
559 | ((UV)dec->cur[-1]); 740 | ((UV)dec->cur[-1]);
560 741 }
561 case 27: 742 else if (ecb_expect_true (m == LENGTH_EXT8))
743 {
562 WANT (9); 744 WANT (8);
563 dec->cur += 9; 745 dec->cur += 8;
746
747 return
748#if UVSIZE < 8
749 0
750#else
564 return (((UV)dec->cur[-8]) << 56) 751 (((UV)dec->cur[-8]) << 56)
565 | (((UV)dec->cur[-7]) << 48) 752 | (((UV)dec->cur[-7]) << 48)
566 | (((UV)dec->cur[-6]) << 40) 753 | (((UV)dec->cur[-6]) << 40)
567 | (((UV)dec->cur[-5]) << 32) 754 | (((UV)dec->cur[-5]) << 32)
755#endif
568 | (((UV)dec->cur[-4]) << 24) 756 | (((UV)dec->cur[-4]) << 24)
569 | (((UV)dec->cur[-3]) << 16) 757 | (((UV)dec->cur[-3]) << 16)
570 | (((UV)dec->cur[-2]) << 8) 758 | (((UV)dec->cur[-2]) << 8)
571 | ((UV)dec->cur[-1]); 759 | ((UV)dec->cur[-1]);
572 760 }
573 default: 761 else
574 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 762 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
575 }
576 763
577fail: 764fail:
578 return 0; 765 return 0;
579} 766}
580 767
585{ 772{
586 AV *av = newAV (); 773 AV *av = newAV ();
587 774
588 DEC_INC_DEPTH; 775 DEC_INC_DEPTH;
589 776
590 if ((*dec->cur & 31) == 31) 777 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
591 { 778 {
592 ++dec->cur; 779 ++dec->cur;
593 780
594 for (;;) 781 for (;;)
595 { 782 {
596 WANT (1); 783 WANT (1);
597 784
598 if (*dec->cur == (0xe0 | 31)) 785 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
599 { 786 {
600 ++dec->cur; 787 ++dec->cur;
601 break; 788 break;
602 } 789 }
603 790
604 av_push (av, decode_sv (dec)); 791 av_push (av, decode_sv (dec));
605 } 792 }
606 } 793 }
607 else 794 else
608 { 795 {
609 int i, len = decode_uint (dec); 796 UV i, len = decode_uint (dec);
610 797
798 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); 799 av_fill (av, len - 1);
612 800
613 for (i = 0; i < len; ++i) 801 for (i = 0; i < len; ++i)
614 AvARRAY (av)[i] = decode_sv (dec); 802 AvARRAY (av)[i] = decode_sv (dec);
615 } 803 }
616 804
617 DEC_DEC_DEPTH; 805 DEC_DEC_DEPTH;
618 return newRV_noinc ((SV *)av); 806 return newRV_noinc ((SV *)av);
619 807
620fail: 808fail:
621 SvREFCNT_dec (av); 809 SvREFCNT_dec_NN (av);
622 DEC_DEC_DEPTH; 810 DEC_DEC_DEPTH;
623 return &PL_sv_undef; 811 return &PL_sv_undef;
624} 812}
625 813
626static void 814static void
627decode_he (dec_t *dec, HV *hv) 815decode_he (dec_t *dec, HV *hv)
628{ 816{
629 // for speed reasons, we specialcase single-string 817 // for speed reasons, we specialcase single-string
630 // byte or utf-8 strings as keys. 818 // byte or utf-8 strings as keys, but only when !stringref
631 819
632 if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27) 820 if (ecb_expect_true (!dec->stringref))
821 if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
633 { 822 {
634 I32 len = decode_uint (dec); 823 STRLEN len = decode_uint (dec);
635 char *key = (char *)dec->cur; 824 char *key = (char *)dec->cur;
636 825
826 WANT (len);
637 dec->cur += len; 827 dec->cur += len;
638 828
639 hv_store (hv, key, len, decode_sv (dec), 0); 829 hv_store (hv, key, len, decode_sv (dec), 0);
830
831 return;
640 } 832 }
641 else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27) 833 else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
642 { 834 {
643 I32 len = decode_uint (dec); 835 STRLEN len = decode_uint (dec);
644 char *key = (char *)dec->cur; 836 char *key = (char *)dec->cur;
645 837
838 WANT (len);
646 dec->cur += len; 839 dec->cur += len;
647 840
841 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
842 if (!is_utf8_string (key, len))
843 ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
844
648 hv_store (hv, key, -len, decode_sv (dec), 0); 845 hv_store (hv, key, -len, decode_sv (dec), 0);
846
847 return;
649 } 848 }
650 else 849
651 {
652 SV *k = decode_sv (dec); 850 SV *k = decode_sv (dec);
653 SV *v = decode_sv (dec); 851 SV *v = decode_sv (dec);
654 852
853 // we leak memory if uncaught exceptions are thrown by random magical
854 // methods, and this is hopefully the only place where it can happen,
855 // so if there is a chance of an exception, take the very slow path.
856 // since catching exceptions is "undocumented/internal/forbidden" by
857 // the new p5p powers, we need to call out to a perl function :/
858 if (ecb_expect_false (SvAMAGIC (k)))
859 {
860 dSP;
861
862 ENTER; SAVETMPS;
863 PUSHMARK (SP);
864 EXTEND (SP, 3);
865 PUSHs (sv_2mortal (newRV_inc ((SV *)hv)));
866 PUSHs (sv_2mortal (k));
867 PUSHs (sv_2mortal (v));
868
869 PUTBACK;
870 call_pv ("CBOR::XS::_hv_store", G_VOID | G_DISCARD | G_EVAL);
871 SPAGAIN;
872
873 FREETMPS; LEAVE;
874
875 if (SvTRUE (ERRSV))
876 ERR_ERRSV;
877
878 return;
879 }
880
655 hv_store_ent (hv, k, v, 0); 881 hv_store_ent (hv, k, v, 0);
656 SvREFCNT_dec (k); 882 SvREFCNT_dec_NN (k);
657 } 883
884fail:
885 ;
658} 886}
659 887
660static SV * 888static SV *
661decode_hv (dec_t *dec) 889decode_hv (dec_t *dec)
662{ 890{
663 HV *hv = newHV (); 891 HV *hv = newHV ();
664 892
665 DEC_INC_DEPTH; 893 DEC_INC_DEPTH;
666 894
667 if ((*dec->cur & 31) == 31) 895 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
668 { 896 {
669 ++dec->cur; 897 ++dec->cur;
670 898
671 for (;;) 899 for (;;)
672 { 900 {
673 WANT (1); 901 WANT (1);
674 902
675 if (*dec->cur == (0xe0 | 31)) 903 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
676 { 904 {
677 ++dec->cur; 905 ++dec->cur;
678 break; 906 break;
679 } 907 }
680 908
681 decode_he (dec, hv); 909 decode_he (dec, hv);
682 } 910 }
683 } 911 }
684 else 912 else
685 { 913 {
686 int pairs = decode_uint (dec); 914 UV pairs = decode_uint (dec);
915
916 WANT (pairs); // complexity check - need at least one byte per value, do not allow supersize hashes
687 917
688 while (pairs--) 918 while (pairs--)
689 decode_he (dec, hv); 919 decode_he (dec, hv);
690 } 920 }
691 921
692 DEC_DEC_DEPTH; 922 DEC_DEC_DEPTH;
693 return newRV_noinc ((SV *)hv); 923 return newRV_noinc ((SV *)hv);
694 924
695fail: 925fail:
696 SvREFCNT_dec (hv); 926 SvREFCNT_dec_NN (hv);
697 DEC_DEC_DEPTH; 927 DEC_DEC_DEPTH;
698 return &PL_sv_undef; 928 return &PL_sv_undef;
699} 929}
700 930
701static SV * 931static SV *
702decode_str (dec_t *dec, int utf8) 932decode_str (dec_t *dec, int utf8)
703{ 933{
704 SV *sv = 0; 934 SV *sv = 0;
705 935
706 if ((*dec->cur & 31) == 31) 936 if (ecb_expect_false ((*dec->cur & MINOR_MASK) == MINOR_INDEF))
707 { 937 {
938 // indefinite length strings
708 ++dec->cur; 939 ++dec->cur;
709 940
941 U8 major = *dec->cur & MAJOR_MISC;
942
710 sv = newSVpvn ("", 0); 943 sv = newSVpvn ("", 0);
711 944
712 // not very fast, and certainly not robust against illegal input
713 for (;;) 945 for (;;)
714 { 946 {
715 WANT (1); 947 WANT (1);
716 948
717 if (*dec->cur == (0xe0 | 31)) 949 if ((*dec->cur - major) > LENGTH_EXT8)
950 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
718 { 951 {
719 ++dec->cur; 952 ++dec->cur;
720 break; 953 break;
721 } 954 }
955 else
956 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
722 957
723 sv_catsv (sv, decode_sv (dec)); 958 STRLEN len = decode_uint (dec);
959
960 WANT (len);
961 sv_catpvn (sv, dec->cur, len);
962 dec->cur += len;
724 } 963 }
725 } 964 }
726 else 965 else
727 { 966 {
728 STRLEN len = decode_uint (dec); 967 STRLEN len = decode_uint (dec);
729 968
730 WANT (len); 969 WANT (len);
731 sv = newSVpvn (dec->cur, len); 970 sv = newSVpvn (dec->cur, len);
732 dec->cur += len; 971 dec->cur += len;
972
973 if (ecb_expect_false (dec->stringref)
974 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
975 av_push (dec->stringref, SvREFCNT_inc_NN (sv));
733 } 976 }
734 977
735 if (utf8) 978 if (utf8)
979 {
980 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
981 if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
982 ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
983
736 SvUTF8_on (sv); 984 SvUTF8_on (sv);
985 }
737 986
738 return sv; 987 return sv;
739 988
740fail: 989fail:
741 SvREFCNT_dec (sv); 990 SvREFCNT_dec (sv);
743} 992}
744 993
745static SV * 994static SV *
746decode_tagged (dec_t *dec) 995decode_tagged (dec_t *dec)
747{ 996{
997 SV *sv = 0;
748 UV tag = decode_uint (dec); 998 UV tag = decode_uint (dec);
749 SV *sv = decode_sv (dec); 999
1000 WANT (1);
750 1001
751 switch (tag) 1002 switch (tag)
752 { 1003 {
753 case CBOR_TAG_MAGIC: 1004 case CBOR_TAG_MAGIC:
754 return sv; 1005 sv = decode_sv (dec);
1006 break;
755 1007
756 case CBOR_TAG_INDIRECTION: 1008 case CBOR_TAG_INDIRECTION:
757 return newRV_noinc (sv); 1009 sv = newRV_noinc (decode_sv (dec));
1010 break;
758 1011
1012 case CBOR_TAG_STRINGREF_NAMESPACE:
1013 {
1014 // do not use SAVETMPS/FREETMPS, as these will
1015 // erase mortalised caches, e.g. "shareable"
1016 ENTER;
1017
1018 SAVESPTR (dec->stringref);
1019 dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());
1020
1021 sv = decode_sv (dec);
1022
1023 LEAVE;
1024 }
1025 break;
1026
1027 case CBOR_TAG_STRINGREF:
1028 {
1029 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
1030 ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
1031
1032 UV idx = decode_uint (dec);
1033
1034 if (!dec->stringref || idx >= (UV)(1 + AvFILLp (dec->stringref)))
1035 ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)");
1036
1037 sv = newSVsv (AvARRAY (dec->stringref)[idx]);
1038 }
1039 break;
1040
759 case CBOR_TAG_VALUE_SHARABLE: 1041 case CBOR_TAG_VALUE_SHAREABLE:
1042 {
760 if (ecb_expect_false (!dec->sharable)) 1043 if (ecb_expect_false (!dec->shareable))
761 dec->sharable = (AV *)sv_2mortal ((SV *)newAV ()); 1044 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
762 1045
1046 if (dec->cbor.flags & F_ALLOW_CYCLES)
1047 {
1048 sv = newSV (0);
763 av_push (dec->sharable, SvREFCNT_inc_NN (sv)); 1049 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
764 1050
765 return sv; 1051 SV *osv = decode_sv (dec);
1052 sv_setsv (sv, osv);
1053 SvREFCNT_dec_NN (osv);
1054 }
1055 else
1056 {
1057 av_push (dec->shareable, &PL_sv_undef);
1058 int idx = AvFILLp (dec->shareable);
1059 sv = decode_sv (dec);
1060 av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
1061 }
1062 }
1063 break;
766 1064
767 case CBOR_TAG_VALUE_SHAREDREF: 1065 case CBOR_TAG_VALUE_SHAREDREF:
768 { 1066 {
769 // TODO: should verify that the sv atcually was a CBOR unsigned integer 1067 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
770 UV idx = SvUV (sv); 1068 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
771 1069
1070 UV idx = decode_uint (dec);
1071
772 if (!dec->sharable || idx > AvFILLp (dec->sharable)) 1072 if (!dec->shareable || idx >= (UV)(1 + AvFILLp (dec->shareable)))
773 ERR ("corrupted CBOR data (sharedref index out of bounds)"); 1073 ERR ("corrupted CBOR data (sharedref index out of bounds)");
774 1074
775 SvREFCNT_dec (sv);
776
777 return SvREFCNT_inc_NN (AvARRAY (dec->sharable)[idx]); 1075 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
1076
1077 if (sv == &PL_sv_undef)
1078 ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
778 } 1079 }
1080 break;
779 1081
780 case CBOR_TAG_PERL_OBJECT: 1082 case CBOR_TAG_PERL_OBJECT:
781 { 1083 {
1084 if (dec->cbor.flags & F_FORBID_OBJECTS)
1085 goto filter;
1086
1087 sv = decode_sv (dec);
1088
782 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) 1089 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
783 ERR ("corrupted CBOR data (non-array perl object)"); 1090 ERR ("corrupted CBOR data (non-array perl object)");
784 1091
785 AV *av = (AV *)SvRV (sv); 1092 AV *av = (AV *)SvRV (sv);
786 int len = av_len (av) + 1; 1093 int len = av_len (av) + 1;
794 if (!method) 1101 if (!method)
795 ERR ("cannot decode perl-object (package does not have a THAW method)"); 1102 ERR ("cannot decode perl-object (package does not have a THAW method)");
796 1103
797 dSP; 1104 dSP;
798 1105
799 ENTER; SAVETMPS; PUSHMARK (SP); 1106 ENTER; SAVETMPS;
1107 PUSHMARK (SP);
800 EXTEND (SP, len + 1); 1108 EXTEND (SP, len + 1);
801 // we re-bless the reference to get overload and other niceties right 1109 // we re-bless the reference to get overload and other niceties right
802 PUSHs (*av_fetch (av, 0, 1)); 1110 PUSHs (*av_fetch (av, 0, 1));
803 PUSHs (sv_cbor); 1111 PUSHs (sv_cbor);
804 1112
812 SPAGAIN; 1120 SPAGAIN;
813 1121
814 if (SvTRUE (ERRSV)) 1122 if (SvTRUE (ERRSV))
815 { 1123 {
816 FREETMPS; LEAVE; 1124 FREETMPS; LEAVE;
817 ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV)))); 1125 ERR_ERRSV;
818 } 1126 }
819 1127
820 SvREFCNT_dec (sv); 1128 SvREFCNT_dec_NN (sv);
821 sv = SvREFCNT_inc (POPs); 1129 sv = SvREFCNT_inc (POPs);
822 1130
823 PUTBACK; 1131 PUTBACK;
824 1132
825 FREETMPS; LEAVE; 1133 FREETMPS; LEAVE;
826
827 return sv;
828 } 1134 }
1135 break;
829 1136
830 default: 1137 default:
1138 filter:
1139 {
1140 SV *tag_sv = newSVuv (tag);
1141
1142 sv = decode_sv (dec);
1143
1144 dSP;
1145 ENTER; SAVETMPS;
1146 SAVESTACK_POS ();
1147 PUSHMARK (SP);
1148 EXTEND (SP, 2);
1149 PUSHs (tag_sv);
1150 PUSHs (sv);
1151
1152 PUTBACK;
1153 int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL);
1154 SPAGAIN;
1155
1156 if (SvTRUE (ERRSV))
831 { 1157 {
1158 SvREFCNT_dec_NN (tag_sv);
1159 FREETMPS; LEAVE;
1160 ERR_ERRSV;
1161 }
1162
1163 if (count)
1164 {
1165 SvREFCNT_dec_NN (tag_sv);
1166 SvREFCNT_dec_NN (sv);
1167 sv = SvREFCNT_inc_NN (POPs);
1168 }
1169 else
1170 {
832 AV *av = newAV (); 1171 AV *av = newAV ();
833 av_push (av, newSVuv (tag)); 1172 av_push (av, tag_sv);
834 av_push (av, sv); 1173 av_push (av, sv);
835 1174
836 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 1175 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
837 ? cbor_tagged_stash 1176 ? cbor_tagged_stash
838 : gv_stashpv ("CBOR::XS::Tagged" , 1); 1177 : gv_stashpv ("CBOR::XS::Tagged" , 1);
839
840 return sv_bless (newRV_noinc ((SV *)av), tagged_stash); 1178 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
841 } 1179 }
1180
1181 PUTBACK;
1182
1183 FREETMPS; LEAVE;
1184 }
1185 break;
842 } 1186 }
1187
1188 return sv;
843 1189
844fail: 1190fail:
845 SvREFCNT_dec (sv); 1191 SvREFCNT_dec (sv);
846 return &PL_sv_undef; 1192 return &PL_sv_undef;
847} 1193}
849static SV * 1195static SV *
850decode_sv (dec_t *dec) 1196decode_sv (dec_t *dec)
851{ 1197{
852 WANT (1); 1198 WANT (1);
853 1199
854 switch (*dec->cur >> 5) 1200 switch (*dec->cur >> MAJOR_SHIFT)
855 { 1201 {
856 case 0: // unsigned int 1202 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
857 return newSVuv (decode_uint (dec)); 1203 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
858 case 1: // negative int 1204 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
859 return newSViv (-1 - (IV)decode_uint (dec)); 1205 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
860 case 2: // octet string 1206 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
861 return decode_str (dec, 0); 1207 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
862 case 3: // utf-8 string 1208 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
863 return decode_str (dec, 1); 1209
864 case 4: // array 1210 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) 1211 switch (*dec->cur++ & MINOR_MASK)
872 { 1212 {
873 case 20: 1213 case SIMPLE_FALSE:
874#if CBOR_SLOW 1214#if CBOR_SLOW
875 types_false = get_bool ("Types::Serialiser::false"); 1215 types_false = get_bool ("Types::Serialiser::false");
876#endif 1216#endif
877 return newSVsv (types_false); 1217 return newSVsv (types_false);
878 case 21: 1218 case SIMPLE_TRUE:
879#if CBOR_SLOW 1219#if CBOR_SLOW
880 types_true = get_bool ("Types::Serialiser::true"); 1220 types_true = get_bool ("Types::Serialiser::true");
881#endif 1221#endif
882 return newSVsv (types_true); 1222 return newSVsv (types_true);
883 case 22: 1223 case SIMPLE_NULL:
884 return newSVsv (&PL_sv_undef); 1224 return newSVsv (&PL_sv_undef);
885 case 23: 1225 case SIMPLE_UNDEF:
886#if CBOR_SLOW 1226#if CBOR_SLOW
887 types_error = get_bool ("Types::Serialiser::error"); 1227 types_error = get_bool ("Types::Serialiser::error");
888#endif 1228#endif
889 return newSVsv (types_error); 1229 return newSVsv (types_error);
890 1230
891 case 25: 1231 case MISC_FLOAT16:
892 { 1232 {
893 WANT (2); 1233 WANT (2);
894 1234
895 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1235 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
896 dec->cur += 2; 1236 dec->cur += 2;
897 1237
898 return newSVnv (ecb_binary16_to_float (fp)); 1238 return newSVnv (ecb_binary16_to_float (fp));
899 } 1239 }
900 1240
901 case 26: 1241 case MISC_FLOAT32:
902 { 1242 {
903 uint32_t fp; 1243 uint32_t fp;
904 WANT (4); 1244 WANT (4);
905 memcpy (&fp, dec->cur, 4); 1245 memcpy (&fp, dec->cur, 4);
906 dec->cur += 4; 1246 dec->cur += 4;
909 fp = ecb_bswap32 (fp); 1249 fp = ecb_bswap32 (fp);
910 1250
911 return newSVnv (ecb_binary32_to_float (fp)); 1251 return newSVnv (ecb_binary32_to_float (fp));
912 } 1252 }
913 1253
914 case 27: 1254 case MISC_FLOAT64:
915 { 1255 {
916 uint64_t fp; 1256 uint64_t fp;
917 WANT (8); 1257 WANT (8);
918 memcpy (&fp, dec->cur, 8); 1258 memcpy (&fp, dec->cur, 8);
919 dec->cur += 8; 1259 dec->cur += 8;
922 fp = ecb_bswap64 (fp); 1262 fp = ecb_bswap64 (fp);
923 1263
924 return newSVnv (ecb_binary64_to_double (fp)); 1264 return newSVnv (ecb_binary64_to_double (fp));
925 } 1265 }
926 1266
927 // 0..19 unassigned 1267 // 0..19 unassigned simple
928 // 24 reserved + unassigned (reserved values are not encodable) 1268 // 24 reserved + unassigned simple (reserved values are not encodable)
1269 // 28-30 unassigned misc
1270 // 31 break code
929 default: 1271 default:
930 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1272 ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
931 } 1273 }
932 1274
933 break; 1275 break;
934 } 1276 }
935 1277
938} 1280}
939 1281
940static SV * 1282static SV *
941decode_cbor (SV *string, CBOR *cbor, char **offset_return) 1283decode_cbor (SV *string, CBOR *cbor, char **offset_return)
942{ 1284{
943 dec_t dec = { }; 1285 dec_t dec = { 0 };
944 SV *sv; 1286 SV *sv;
945 STRLEN len; 1287 STRLEN len;
946 char *data = SvPVbyte (string, len); 1288 char *data = SvPVbyte (string, len);
947 1289
948 if (len > cbor->max_size && cbor->max_size) 1290 if (len > cbor->max_size && cbor->max_size)
962 if (dec.cur != dec.end && !dec.err) 1304 if (dec.cur != dec.end && !dec.err)
963 dec.err = "garbage after CBOR object"; 1305 dec.err = "garbage after CBOR object";
964 1306
965 if (dec.err) 1307 if (dec.err)
966 { 1308 {
1309 if (dec.shareable)
1310 {
1311 // need to break cyclic links, which would all be in shareable
1312 int i;
1313 SV **svp;
1314
1315 for (i = av_len (dec.shareable) + 1; i--; )
1316 if ((svp = av_fetch (dec.shareable, i, 0)))
1317 sv_setsv (*svp, &PL_sv_undef);
1318 }
1319
967 SvREFCNT_dec (sv); 1320 SvREFCNT_dec_NN (sv);
1321
1322 if (dec.err_sv)
1323 sv_2mortal (dec.err_sv);
1324
968 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur); 1325 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur);
969 } 1326 }
970 1327
971 sv = sv_2mortal (sv); 1328 sv = sv_2mortal (sv);
972 1329
973 return sv; 1330 return sv;
974} 1331}
975 1332
1333/////////////////////////////////////////////////////////////////////////////
1334// incremental parser
1335
1336#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
1337
1338// returns 0 for notyet, 1 for success or error
1339static int
1340incr_parse (CBOR *self, SV *cborstr)
1341{
1342 STRLEN cur;
1343 SvPV (cborstr, cur);
1344
1345 while (ecb_expect_true (self->incr_need <= cur))
1346 {
1347 // table of integer count bytes
1348 static I8 incr_len[MINOR_MASK + 1] = {
1349 0, 0, 0, 0, 0, 0, 0, 0,
1350 0, 0, 0, 0, 0, 0, 0, 0,
1351 0, 0, 0, 0, 0, 0, 0, 0,
1352 1, 2, 4, 8,-1,-1,-1,-2
1353 };
1354
1355 const U8 *p = SvPVX (cborstr) + self->incr_pos;
1356 U8 m = *p & MINOR_MASK;
1357 IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1358 I8 ilen = incr_len[m];
1359
1360 self->incr_need = self->incr_pos + 1;
1361
1362 if (ecb_expect_false (ilen < 0))
1363 {
1364 if (m != MINOR_INDEF)
1365 return 1; // error
1366
1367 if (*p == (MAJOR_MISC | MINOR_INDEF))
1368 {
1369 if (count >= 0)
1370 return 1; // error
1371
1372 count = 1;
1373 }
1374 else
1375 {
1376 av_push (self->incr_count, newSViv (-1)); //TODO: nest
1377 count = -1;
1378 }
1379 }
1380 else
1381 {
1382 self->incr_need += ilen;
1383 if (ecb_expect_false (self->incr_need > cur))
1384 return 0;
1385
1386 int major = *p >> MAJOR_SHIFT;
1387
1388 switch (major)
1389 {
1390 case MAJOR_TAG >> MAJOR_SHIFT:
1391 ++count; // tags merely prefix another value
1392 break;
1393
1394 case MAJOR_BYTES >> MAJOR_SHIFT:
1395 case MAJOR_TEXT >> MAJOR_SHIFT:
1396 case MAJOR_ARRAY >> MAJOR_SHIFT:
1397 case MAJOR_MAP >> MAJOR_SHIFT:
1398 {
1399 UV len;
1400
1401 if (ecb_expect_false (ilen))
1402 {
1403 len = 0;
1404
1405 do {
1406 len = (len << 8) | *++p;
1407 } while (--ilen);
1408 }
1409 else
1410 len = m;
1411
1412 switch (major)
1413 {
1414 case MAJOR_BYTES >> MAJOR_SHIFT:
1415 case MAJOR_TEXT >> MAJOR_SHIFT:
1416 self->incr_need += len;
1417 if (ecb_expect_false (self->incr_need > cur))
1418 return 0;
1419
1420 break;
1421
1422 case MAJOR_MAP >> MAJOR_SHIFT:
1423 len <<= 1;
1424 case MAJOR_ARRAY >> MAJOR_SHIFT:
1425 if (len)
1426 {
1427 av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
1428 count = len + 1;
1429 }
1430 break;
1431 }
1432 }
1433 }
1434 }
1435
1436 self->incr_pos = self->incr_need;
1437
1438 if (count > 0)
1439 {
1440 while (!--count)
1441 {
1442 if (!AvFILLp (self->incr_count))
1443 return 1; // done
1444
1445 SvREFCNT_dec_NN (av_pop (self->incr_count));
1446 count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1447 }
1448
1449 SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
1450 }
1451 }
1452
1453 return 0;
1454}
1455
1456
976///////////////////////////////////////////////////////////////////////////// 1457/////////////////////////////////////////////////////////////////////////////
977// XS interface functions 1458// XS interface functions
978 1459
979MODULE = CBOR::XS PACKAGE = CBOR::XS 1460MODULE = CBOR::XS PACKAGE = CBOR::XS
980 1461
988 1469
989 types_true = get_bool ("Types::Serialiser::true" ); 1470 types_true = get_bool ("Types::Serialiser::true" );
990 types_false = get_bool ("Types::Serialiser::false"); 1471 types_false = get_bool ("Types::Serialiser::false");
991 types_error = get_bool ("Types::Serialiser::error"); 1472 types_error = get_bool ("Types::Serialiser::error");
992 1473
1474 default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1475
993 sv_cbor = newSVpv ("CBOR", 0); 1476 sv_cbor = newSVpv ("CBOR", 0);
994 SvREADONLY_on (sv_cbor); 1477 SvREADONLY_on (sv_cbor);
1478
1479 assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1));
995} 1480}
996 1481
997PROTOTYPES: DISABLE 1482PROTOTYPES: DISABLE
998 1483
999void CLONE (...) 1484void CLONE (...)
1018void shrink (CBOR *self, int enable = 1) 1503void shrink (CBOR *self, int enable = 1)
1019 ALIAS: 1504 ALIAS:
1020 shrink = F_SHRINK 1505 shrink = F_SHRINK
1021 allow_unknown = F_ALLOW_UNKNOWN 1506 allow_unknown = F_ALLOW_UNKNOWN
1022 allow_sharing = F_ALLOW_SHARING 1507 allow_sharing = F_ALLOW_SHARING
1023 dedup_keys = F_DEDUP_KEYS 1508 allow_cycles = F_ALLOW_CYCLES
1509 forbid_objects = F_FORBID_OBJECTS
1024 dedup_strings = F_DEDUP_STRINGS 1510 pack_strings = F_PACK_STRINGS
1511 text_keys = F_TEXT_KEYS
1512 text_strings = F_TEXT_STRINGS
1513 validate_utf8 = F_VALIDATE_UTF8
1025 PPCODE: 1514 PPCODE:
1026{ 1515{
1027 if (enable) 1516 if (enable)
1028 self->flags |= ix; 1517 self->flags |= ix;
1029 else 1518 else
1035void get_shrink (CBOR *self) 1524void get_shrink (CBOR *self)
1036 ALIAS: 1525 ALIAS:
1037 get_shrink = F_SHRINK 1526 get_shrink = F_SHRINK
1038 get_allow_unknown = F_ALLOW_UNKNOWN 1527 get_allow_unknown = F_ALLOW_UNKNOWN
1039 get_allow_sharing = F_ALLOW_SHARING 1528 get_allow_sharing = F_ALLOW_SHARING
1040 get_dedup_keys = F_DEDUP_KEYS 1529 get_allow_cycles = F_ALLOW_CYCLES
1530 get_forbid_objects = F_FORBID_OBJECTS
1041 get_dedup_strings = F_DEDUP_STRINGS 1531 get_pack_strings = F_PACK_STRINGS
1532 get_text_keys = F_TEXT_KEYS
1533 get_text_strings = F_TEXT_STRINGS
1534 get_validate_utf8 = F_VALIDATE_UTF8
1042 PPCODE: 1535 PPCODE:
1043 XPUSHs (boolSV (self->flags & ix)); 1536 XPUSHs (boolSV (self->flags & ix));
1044 1537
1045void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1538void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1046 PPCODE: 1539 PPCODE:
1059 XPUSHs (ST (0)); 1552 XPUSHs (ST (0));
1060 1553
1061int get_max_size (CBOR *self) 1554int get_max_size (CBOR *self)
1062 CODE: 1555 CODE:
1063 RETVAL = self->max_size; 1556 RETVAL = self->max_size;
1557 OUTPUT:
1558 RETVAL
1559
1560void filter (CBOR *self, SV *filter = 0)
1561 PPCODE:
1562 SvREFCNT_dec (self->filter);
1563 self->filter = filter ? newSVsv (filter) : filter;
1564 XPUSHs (ST (0));
1565
1566SV *get_filter (CBOR *self)
1567 CODE:
1568 RETVAL = self->filter ? self->filter : NEWSV (0, 0);
1064 OUTPUT: 1569 OUTPUT:
1065 RETVAL 1570 RETVAL
1066 1571
1067void encode (CBOR *self, SV *scalar) 1572void encode (CBOR *self, SV *scalar)
1068 PPCODE: 1573 PPCODE:
1083 EXTEND (SP, 2); 1588 EXTEND (SP, 2);
1084 PUSHs (sv); 1589 PUSHs (sv);
1085 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1590 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1086} 1591}
1087 1592
1593void incr_parse (CBOR *self, SV *cborstr)
1594 ALIAS:
1595 incr_parse_multiple = 1
1596 PPCODE:
1597{
1598 if (SvUTF8 (cborstr))
1599 sv_utf8_downgrade (cborstr, 0);
1600
1601 if (!self->incr_count)
1602 {
1603 self->incr_count = newAV ();
1604 self->incr_pos = 0;
1605 self->incr_need = 1;
1606
1607 av_push (self->incr_count, newSViv (1));
1608 }
1609
1610 do
1611 {
1612 if (!incr_parse (self, cborstr))
1613 {
1614 if (self->incr_need > self->max_size && self->max_size)
1615 croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
1616 (unsigned long)self->incr_need, (unsigned long)self->max_size);
1617
1618 break;
1619 }
1620
1621 SV *sv;
1622 char *offset;
1623
1624 PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
1625 XPUSHs (sv);
1626
1627 sv_chop (cborstr, offset);
1628
1629 av_clear (self->incr_count);
1630 av_push (self->incr_count, newSViv (1));
1631
1632 self->incr_pos = 0;
1633 self->incr_need = self->incr_pos + 1;
1634 }
1635 while (ix);
1636}
1637
1638void incr_reset (CBOR *self)
1639 CODE:
1640{
1641 SvREFCNT_dec (self->incr_count);
1642 self->incr_count = 0;
1643}
1644
1645void DESTROY (CBOR *self)
1646 PPCODE:
1647 cbor_free (self);
1648
1088PROTOTYPES: ENABLE 1649PROTOTYPES: ENABLE
1089 1650
1090void encode_cbor (SV *scalar) 1651void encode_cbor (SV *scalar)
1652 ALIAS:
1653 encode_cbor = 0
1654 encode_cbor_sharing = F_ALLOW_SHARING
1091 PPCODE: 1655 PPCODE:
1092{ 1656{
1093 CBOR cbor; 1657 CBOR cbor;
1094 cbor_init (&cbor); 1658 cbor_init (&cbor);
1659 cbor.flags |= ix;
1095 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1660 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1096 XPUSHs (scalar); 1661 XPUSHs (scalar);
1097} 1662}
1098 1663
1099void decode_cbor (SV *cborstr) 1664void decode_cbor (SV *cborstr)
1103 cbor_init (&cbor); 1668 cbor_init (&cbor);
1104 PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN; 1669 PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN;
1105 XPUSHs (cborstr); 1670 XPUSHs (cborstr);
1106} 1671}
1107 1672
1673#ifdef __AFL_COMPILER
1674
1675void
1676afl_init ()
1677 CODE:
1678 __AFL_INIT ();
1679
1680int
1681afl_loop (unsigned int count = 10000)
1682 CODE:
1683 RETVAL = __AFL_LOOP (count);
1684 OUTPUT:
1685 RETVAL
1686
1687#endif
1688

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines