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

Comparing cvsroot/CBOR-XS/XS.xs (file contents):
Revision 1.18 by root, Wed Nov 20 01:09:46 2013 UTC vs.
Revision 1.65 by root, Thu Nov 15 19:52:41 2018 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 PUSHMARK (SP);
377 EXTEND (SP, 2); 516 EXTEND (SP, 2);
378 // we re-bless the reference to get overload and other niceties right 517 // we re-bless the reference to get overload and other niceties right
379 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 518 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
380 PUSHs (sv_cbor); 519 PUSHs (sv_cbor);
381 520
386 // catch this surprisingly common error 525 // catch this surprisingly common error
387 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) 526 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)); 527 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
389 528
390 encode_tag (enc, CBOR_TAG_PERL_OBJECT); 529 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
391 encode_uint (enc, 0x80, count + 1); 530 encode_uint (enc, MAJOR_ARRAY, count + 1);
392 encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); 531 encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
393 532
394 while (count) 533 {
534 int i;
535
536 for (i = 0; i < count; ++i)
395 encode_sv (enc, SP[1 - count--]); 537 encode_sv (enc, SP[i + 1 - count]);
538
539 SP -= count;
540 }
396 541
397 PUTBACK; 542 PUTBACK;
398 543
399 FREETMPS; LEAVE; 544 FREETMPS; LEAVE;
400 } 545 }
418{ 563{
419 double nv = SvNVX (sv); 564 double nv = SvNVX (sv);
420 565
421 need (enc, 9); 566 need (enc, 9);
422 567
423 if (ecb_expect_false (nv == (U32)nv)) 568 if (ecb_expect_false (nv == (NV)(U32)nv))
424 encode_uint (enc, 0x00, (U32)nv); 569 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
425 //TODO: maybe I32? 570 //TODO: maybe I32?
426 else if (ecb_expect_false (nv == (float)nv)) 571 else if (ecb_expect_false (nv == (float)nv))
427 { 572 {
573 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
574
428 uint32_t fp = ecb_float_to_binary32 (nv); 575 uint32_t fp = ecb_float_to_binary32 (nv);
429
430 *enc->cur++ = 0xe0 | 26;
431 576
432 if (!ecb_big_endian ()) 577 if (!ecb_big_endian ())
433 fp = ecb_bswap32 (fp); 578 fp = ecb_bswap32 (fp);
434 579
435 memcpy (enc->cur, &fp, 4); 580 memcpy (enc->cur, &fp, 4);
436 enc->cur += 4; 581 enc->cur += 4;
437 } 582 }
438 else 583 else
439 { 584 {
585 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
586
440 uint64_t fp = ecb_double_to_binary64 (nv); 587 uint64_t fp = ecb_double_to_binary64 (nv);
441
442 *enc->cur++ = 0xe0 | 27;
443 588
444 if (!ecb_big_endian ()) 589 if (!ecb_big_endian ())
445 fp = ecb_bswap64 (fp); 590 fp = ecb_bswap64 (fp);
446 591
447 memcpy (enc->cur, &fp, 8); 592 memcpy (enc->cur, &fp, 8);
452static void 597static void
453encode_sv (enc_t *enc, SV *sv) 598encode_sv (enc_t *enc, SV *sv)
454{ 599{
455 SvGETMAGIC (sv); 600 SvGETMAGIC (sv);
456 601
457 if (encode_sharable (enc, sv))
458 return;
459
460 if (SvPOKp (sv)) 602 if (SvPOKp (sv))
461 { 603 {
462 STRLEN len; 604 STRLEN len;
463 char *str = SvPV (sv, len); 605 char *str = SvPV (sv, len);
464 encode_str (enc, SvUTF8 (sv), str, len); 606 encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len);
465 } 607 }
466 else if (SvNOKp (sv)) 608 else if (SvNOKp (sv))
467 encode_nv (enc, sv); 609 encode_nv (enc, sv);
468 else if (SvIOKp (sv)) 610 else if (SvIOKp (sv))
469 { 611 {
470 if (SvIsUV (sv)) 612 if (SvIsUV (sv))
471 encode_uint (enc, 0x00, SvUVX (sv)); 613 encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
472 else if (SvIVX (sv) >= 0) 614 else if (SvIVX (sv) >= 0)
473 encode_uint (enc, 0x00, SvIVX (sv)); 615 encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
474 else 616 else
475 encode_uint (enc, 0x20, -(SvIVX (sv) + 1)); 617 encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
476 } 618 }
477 else if (SvROK (sv)) 619 else if (SvROK (sv))
478 encode_rv (enc, SvRV (sv)); 620 encode_rv (enc, SvRV (sv));
479 else if (!SvOK (sv)) 621 else if (!SvOK (sv))
480 encode_ch (enc, 0xe0 | 22); 622 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
481 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 623 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
482 encode_ch (enc, 0xe0 | 23); 624 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
483 else 625 else
484 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data", 626 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
485 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 627 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
486} 628}
487 629
488static SV * 630static SV *
489encode_cbor (SV *scalar, CBOR *cbor) 631encode_cbor (SV *scalar, CBOR *cbor)
490{ 632{
491 enc_t enc = { }; 633 enc_t enc = { 0 };
492 634
493 enc.cbor = *cbor; 635 enc.cbor = *cbor;
494 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 636 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
495 enc.cur = SvPVX (enc.sv); 637 enc.cur = SvPVX (enc.sv);
496 enc.end = SvEND (enc.sv); 638 enc.end = SvEND (enc.sv);
497 639
498 SvPOK_only (enc.sv); 640 SvPOK_only (enc.sv);
641
642 if (cbor->flags & F_PACK_STRINGS)
643 {
644 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
645 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
646 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
647 }
648
499 encode_sv (&enc, scalar); 649 encode_sv (&enc, scalar);
500 650
501 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv)); 651 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv));
502 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings 652 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings
503 653
517 U8 *end; // end of input string 667 U8 *end; // end of input string
518 const char *err; // parse error, if != 0 668 const char *err; // parse error, if != 0
519 CBOR cbor; 669 CBOR cbor;
520 U32 depth; // recursion depth 670 U32 depth; // recursion depth
521 U32 maxdepth; // recursion depth limit 671 U32 maxdepth; // recursion depth limit
522 AV *sharable; 672 AV *shareable;
673 AV *stringref;
674 SV *decode_tagged;
675 SV *err_sv; // optional sv for error, needs to be freed
523} dec_t; 676} dec_t;
524 677
525#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE 678// set dec->err to ERRSV
679ecb_cold static void
680err_errsv (dec_t *dec)
681{
682 if (!dec->err)
683 {
684 dec->err_sv = newSVsv (ERRSV);
526 685
527#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data") 686 // chop off the trailing \n
687 SvCUR_set (dec->err_sv, SvCUR (dec->err_sv) - 1);
688 *SvEND (dec->err_sv) = 0;
528 689
690 dec->err = SvPVutf8_nolen (dec->err_sv);
691 }
692}
693
694// the following functions are used to reduce code size and help the compiler to optimise
695ecb_cold static void
696err_set (dec_t *dec, const char *reason)
697{
698 if (!dec->err)
699 dec->err = reason;
700}
701
702ecb_cold static void
703err_unexpected_end (dec_t *dec)
704{
705 err_set (dec, "unexpected end of CBOR data");
706}
707
708#define ERR_DO(do) SB do; goto fail; SE
709#define ERR(reason) ERR_DO (err_set (dec, reason))
710#define ERR_ERRSV ERR_DO (err_errsv (dec))
711
712#define WANT(len) if (ecb_expect_false ((uintptr_t)(dec->end - dec->cur) < (STRLEN)len)) ERR_DO (err_unexpected_end (dec))
713
529#define DEC_INC_DEPTH if (++dec->depth > dec->cbor.max_depth) ERR (ERR_NESTING_EXCEEDED) 714#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 715#define DEC_DEC_DEPTH --dec->depth
531 716
532static UV 717static UV
533decode_uint (dec_t *dec) 718decode_uint (dec_t *dec)
534{ 719{
535 switch (*dec->cur & 31) 720 U8 m = *dec->cur & MINOR_MASK;
536 { 721 ++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;
541 722
542 case 24: 723 if (ecb_expect_true (m < LENGTH_EXT1))
724 return m;
725 else if (ecb_expect_true (m == LENGTH_EXT1))
726 {
543 WANT (2); 727 WANT (1);
544 dec->cur += 2; 728 dec->cur += 1;
545 return dec->cur[-1]; 729 return dec->cur[-1];
546 730 }
547 case 25: 731 else if (ecb_expect_true (m == LENGTH_EXT2))
732 {
548 WANT (3); 733 WANT (2);
549 dec->cur += 3; 734 dec->cur += 2;
550 return (((UV)dec->cur[-2]) << 8) 735 return (((UV)dec->cur[-2]) << 8)
551 | ((UV)dec->cur[-1]); 736 | ((UV)dec->cur[-1]);
552 737 }
553 case 26: 738 else if (ecb_expect_true (m == LENGTH_EXT4))
739 {
554 WANT (5); 740 WANT (4);
555 dec->cur += 5; 741 dec->cur += 4;
556 return (((UV)dec->cur[-4]) << 24) 742 return (((UV)dec->cur[-4]) << 24)
557 | (((UV)dec->cur[-3]) << 16) 743 | (((UV)dec->cur[-3]) << 16)
558 | (((UV)dec->cur[-2]) << 8) 744 | (((UV)dec->cur[-2]) << 8)
559 | ((UV)dec->cur[-1]); 745 | ((UV)dec->cur[-1]);
560 746 }
561 case 27: 747 else if (ecb_expect_true (m == LENGTH_EXT8))
748 {
562 WANT (9); 749 WANT (8);
563 dec->cur += 9; 750 dec->cur += 8;
751
752 return
753#if UVSIZE < 8
754 0
755#else
564 return (((UV)dec->cur[-8]) << 56) 756 (((UV)dec->cur[-8]) << 56)
565 | (((UV)dec->cur[-7]) << 48) 757 | (((UV)dec->cur[-7]) << 48)
566 | (((UV)dec->cur[-6]) << 40) 758 | (((UV)dec->cur[-6]) << 40)
567 | (((UV)dec->cur[-5]) << 32) 759 | (((UV)dec->cur[-5]) << 32)
760#endif
568 | (((UV)dec->cur[-4]) << 24) 761 | (((UV)dec->cur[-4]) << 24)
569 | (((UV)dec->cur[-3]) << 16) 762 | (((UV)dec->cur[-3]) << 16)
570 | (((UV)dec->cur[-2]) << 8) 763 | (((UV)dec->cur[-2]) << 8)
571 | ((UV)dec->cur[-1]); 764 | ((UV)dec->cur[-1]);
572 765 }
573 default: 766 else
574 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 767 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
575 }
576 768
577fail: 769fail:
578 return 0; 770 return 0;
579} 771}
580 772
585{ 777{
586 AV *av = newAV (); 778 AV *av = newAV ();
587 779
588 DEC_INC_DEPTH; 780 DEC_INC_DEPTH;
589 781
590 if ((*dec->cur & 31) == 31) 782 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
591 { 783 {
592 ++dec->cur; 784 ++dec->cur;
593 785
594 for (;;) 786 for (;;)
595 { 787 {
596 WANT (1); 788 WANT (1);
597 789
598 if (*dec->cur == (0xe0 | 31)) 790 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
599 { 791 {
600 ++dec->cur; 792 ++dec->cur;
601 break; 793 break;
602 } 794 }
603 795
604 av_push (av, decode_sv (dec)); 796 av_push (av, decode_sv (dec));
605 } 797 }
606 } 798 }
607 else 799 else
608 { 800 {
609 int i, len = decode_uint (dec); 801 UV i, len = decode_uint (dec);
610 802
803 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); 804 av_fill (av, len - 1);
612 805
613 for (i = 0; i < len; ++i) 806 for (i = 0; i < len; ++i)
614 AvARRAY (av)[i] = decode_sv (dec); 807 AvARRAY (av)[i] = decode_sv (dec);
615 } 808 }
616 809
617 DEC_DEC_DEPTH; 810 DEC_DEC_DEPTH;
618 return newRV_noinc ((SV *)av); 811 return newRV_noinc ((SV *)av);
619 812
620fail: 813fail:
621 SvREFCNT_dec (av); 814 SvREFCNT_dec_NN (av);
622 DEC_DEC_DEPTH; 815 DEC_DEC_DEPTH;
623 return &PL_sv_undef; 816 return &PL_sv_undef;
624} 817}
625 818
626static void 819static void
627decode_he (dec_t *dec, HV *hv) 820decode_he (dec_t *dec, HV *hv)
628{ 821{
629 // for speed reasons, we specialcase single-string 822 // for speed reasons, we specialcase single-string
630 // byte or utf-8 strings as keys. 823 // byte or utf-8 strings as keys, but only when !stringref
631 824
632 if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27) 825 if (ecb_expect_true (!dec->stringref))
826 if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
633 { 827 {
634 I32 len = decode_uint (dec); 828 STRLEN len = decode_uint (dec);
635 char *key = (char *)dec->cur; 829 char *key = (char *)dec->cur;
636 830
831 WANT (len);
637 dec->cur += len; 832 dec->cur += len;
638 833
639 hv_store (hv, key, len, decode_sv (dec), 0); 834 hv_store (hv, key, len, decode_sv (dec), 0);
835
836 return;
640 } 837 }
641 else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27) 838 else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
642 { 839 {
643 I32 len = decode_uint (dec); 840 STRLEN len = decode_uint (dec);
644 char *key = (char *)dec->cur; 841 char *key = (char *)dec->cur;
645 842
843 WANT (len);
646 dec->cur += len; 844 dec->cur += len;
647 845
846 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
847 if (!is_utf8_string (key, len))
848 ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
849
648 hv_store (hv, key, -len, decode_sv (dec), 0); 850 hv_store (hv, key, -len, decode_sv (dec), 0);
851
852 return;
649 } 853 }
650 else 854
651 {
652 SV *k = decode_sv (dec); 855 SV *k = decode_sv (dec);
653 SV *v = decode_sv (dec); 856 SV *v = decode_sv (dec);
654 857
858 // we leak memory if uncaught exceptions are thrown by random magical
859 // methods, and this is hopefully the only place where it can happen,
860 // so if there is a chance of an exception, take the very slow path.
861 // since catching exceptions is "undocumented/internal/forbidden" by
862 // the new p5p powers, we need to call out to a perl function :/
863 if (ecb_expect_false (SvAMAGIC (k)))
864 {
865 dSP;
866
867 ENTER; SAVETMPS;
868 PUSHMARK (SP);
869 EXTEND (SP, 3);
870 PUSHs (sv_2mortal (newRV_inc ((SV *)hv)));
871 PUSHs (sv_2mortal (k));
872 PUSHs (sv_2mortal (v));
873
874 PUTBACK;
875 call_pv ("CBOR::XS::_hv_store", G_VOID | G_DISCARD | G_EVAL);
876 SPAGAIN;
877
878 FREETMPS; LEAVE;
879
880 if (SvTRUE (ERRSV))
881 ERR_ERRSV;
882
883 return;
884 }
885
655 hv_store_ent (hv, k, v, 0); 886 hv_store_ent (hv, k, v, 0);
656 SvREFCNT_dec (k); 887 SvREFCNT_dec_NN (k);
657 } 888
889fail:
890 ;
658} 891}
659 892
660static SV * 893static SV *
661decode_hv (dec_t *dec) 894decode_hv (dec_t *dec)
662{ 895{
663 HV *hv = newHV (); 896 HV *hv = newHV ();
664 897
665 DEC_INC_DEPTH; 898 DEC_INC_DEPTH;
666 899
667 if ((*dec->cur & 31) == 31) 900 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
668 { 901 {
669 ++dec->cur; 902 ++dec->cur;
670 903
671 for (;;) 904 for (;;)
672 { 905 {
673 WANT (1); 906 WANT (1);
674 907
675 if (*dec->cur == (0xe0 | 31)) 908 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
676 { 909 {
677 ++dec->cur; 910 ++dec->cur;
678 break; 911 break;
679 } 912 }
680 913
681 decode_he (dec, hv); 914 decode_he (dec, hv);
682 } 915 }
683 } 916 }
684 else 917 else
685 { 918 {
686 int pairs = decode_uint (dec); 919 UV pairs = decode_uint (dec);
920
921 WANT (pairs); // complexity check - need at least one byte per value, do not allow supersize hashes
687 922
688 while (pairs--) 923 while (pairs--)
689 decode_he (dec, hv); 924 decode_he (dec, hv);
690 } 925 }
691 926
692 DEC_DEC_DEPTH; 927 DEC_DEC_DEPTH;
693 return newRV_noinc ((SV *)hv); 928 return newRV_noinc ((SV *)hv);
694 929
695fail: 930fail:
696 SvREFCNT_dec (hv); 931 SvREFCNT_dec_NN (hv);
697 DEC_DEC_DEPTH; 932 DEC_DEC_DEPTH;
698 return &PL_sv_undef; 933 return &PL_sv_undef;
699} 934}
700 935
701static SV * 936static SV *
702decode_str (dec_t *dec, int utf8) 937decode_str (dec_t *dec, int utf8)
703{ 938{
704 SV *sv = 0; 939 SV *sv = 0;
705 940
706 if ((*dec->cur & 31) == 31) 941 if (ecb_expect_false ((*dec->cur & MINOR_MASK) == MINOR_INDEF))
707 { 942 {
943 // indefinite length strings
708 ++dec->cur; 944 ++dec->cur;
709 945
946 U8 major = *dec->cur & MAJOR_MISC;
947
710 sv = newSVpvn ("", 0); 948 sv = newSVpvn ("", 0);
711 949
712 // not very fast, and certainly not robust against illegal input
713 for (;;) 950 for (;;)
714 { 951 {
715 WANT (1); 952 WANT (1);
716 953
717 if (*dec->cur == (0xe0 | 31)) 954 if ((*dec->cur - major) > LENGTH_EXT8)
955 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
718 { 956 {
719 ++dec->cur; 957 ++dec->cur;
720 break; 958 break;
721 } 959 }
960 else
961 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
722 962
723 sv_catsv (sv, decode_sv (dec)); 963 STRLEN len = decode_uint (dec);
964
965 WANT (len);
966 sv_catpvn (sv, dec->cur, len);
967 dec->cur += len;
724 } 968 }
725 } 969 }
726 else 970 else
727 { 971 {
728 STRLEN len = decode_uint (dec); 972 STRLEN len = decode_uint (dec);
729 973
730 WANT (len); 974 WANT (len);
731 sv = newSVpvn (dec->cur, len); 975 sv = newSVpvn (dec->cur, len);
732 dec->cur += len; 976 dec->cur += len;
977
978 if (ecb_expect_false (dec->stringref)
979 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
980 av_push (dec->stringref, SvREFCNT_inc_NN (sv));
733 } 981 }
734 982
735 if (utf8) 983 if (utf8)
984 {
985 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
986 if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
987 ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
988
736 SvUTF8_on (sv); 989 SvUTF8_on (sv);
990 }
737 991
738 return sv; 992 return sv;
739 993
740fail: 994fail:
741 SvREFCNT_dec (sv); 995 SvREFCNT_dec (sv);
743} 997}
744 998
745static SV * 999static SV *
746decode_tagged (dec_t *dec) 1000decode_tagged (dec_t *dec)
747{ 1001{
1002 SV *sv = 0;
748 UV tag = decode_uint (dec); 1003 UV tag = decode_uint (dec);
749 SV *sv = decode_sv (dec); 1004
1005 WANT (1);
750 1006
751 switch (tag) 1007 switch (tag)
752 { 1008 {
753 case CBOR_TAG_MAGIC: 1009 case CBOR_TAG_MAGIC:
754 return sv; 1010 sv = decode_sv (dec);
1011 break;
755 1012
756 case CBOR_TAG_INDIRECTION: 1013 case CBOR_TAG_INDIRECTION:
757 return newRV_noinc (sv); 1014 sv = newRV_noinc (decode_sv (dec));
1015 break;
758 1016
1017 case CBOR_TAG_STRINGREF_NAMESPACE:
1018 {
1019 // do not use SAVETMPS/FREETMPS, as these will
1020 // erase mortalised caches, e.g. "shareable"
1021 ENTER;
1022
1023 SAVESPTR (dec->stringref);
1024 dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());
1025
1026 sv = decode_sv (dec);
1027
1028 LEAVE;
1029 }
1030 break;
1031
1032 case CBOR_TAG_STRINGREF:
1033 {
1034 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
1035 ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
1036
1037 UV idx = decode_uint (dec);
1038
1039 if (!dec->stringref || idx >= (UV)(1 + AvFILLp (dec->stringref)))
1040 ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)");
1041
1042 sv = newSVsv (AvARRAY (dec->stringref)[idx]);
1043 }
1044 break;
1045
759 case CBOR_TAG_VALUE_SHARABLE: 1046 case CBOR_TAG_VALUE_SHAREABLE:
1047 {
760 if (ecb_expect_false (!dec->sharable)) 1048 if (ecb_expect_false (!dec->shareable))
761 dec->sharable = (AV *)sv_2mortal ((SV *)newAV ()); 1049 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
762 1050
1051 if (dec->cbor.flags & F_ALLOW_CYCLES)
1052 {
1053 sv = newSV (0);
763 av_push (dec->sharable, SvREFCNT_inc_NN (sv)); 1054 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
764 1055
765 return sv; 1056 SV *osv = decode_sv (dec);
1057 sv_setsv (sv, osv);
1058 SvREFCNT_dec_NN (osv);
1059 }
1060 else
1061 {
1062 av_push (dec->shareable, &PL_sv_undef);
1063 int idx = AvFILLp (dec->shareable);
1064 sv = decode_sv (dec);
1065 av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
1066 }
1067 }
1068 break;
766 1069
767 case CBOR_TAG_VALUE_SHAREDREF: 1070 case CBOR_TAG_VALUE_SHAREDREF:
768 { 1071 {
769 // TODO: should verify that the sv atcually was a CBOR unsigned integer 1072 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
770 UV idx = SvUV (sv); 1073 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
771 1074
1075 UV idx = decode_uint (dec);
1076
772 if (!dec->sharable || idx > AvFILLp (dec->sharable)) 1077 if (!dec->shareable || idx >= (UV)(1 + AvFILLp (dec->shareable)))
773 ERR ("corrupted CBOR data (sharedref index out of bounds)"); 1078 ERR ("corrupted CBOR data (sharedref index out of bounds)");
774 1079
775 SvREFCNT_dec (sv);
776
777 return SvREFCNT_inc_NN (AvARRAY (dec->sharable)[idx]); 1080 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
1081
1082 if (sv == &PL_sv_undef)
1083 ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
778 } 1084 }
1085 break;
779 1086
780 case CBOR_TAG_PERL_OBJECT: 1087 case CBOR_TAG_PERL_OBJECT:
781 { 1088 {
1089 if (dec->cbor.flags & F_FORBID_OBJECTS)
1090 goto filter;
1091
1092 sv = decode_sv (dec);
1093
782 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) 1094 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
783 ERR ("corrupted CBOR data (non-array perl object)"); 1095 ERR ("corrupted CBOR data (non-array perl object)");
784 1096
785 AV *av = (AV *)SvRV (sv); 1097 AV *av = (AV *)SvRV (sv);
786 int len = av_len (av) + 1; 1098 int len = av_len (av) + 1;
794 if (!method) 1106 if (!method)
795 ERR ("cannot decode perl-object (package does not have a THAW method)"); 1107 ERR ("cannot decode perl-object (package does not have a THAW method)");
796 1108
797 dSP; 1109 dSP;
798 1110
799 ENTER; SAVETMPS; PUSHMARK (SP); 1111 ENTER; SAVETMPS;
1112 PUSHMARK (SP);
800 EXTEND (SP, len + 1); 1113 EXTEND (SP, len + 1);
801 // we re-bless the reference to get overload and other niceties right 1114 // we re-bless the reference to get overload and other niceties right
802 PUSHs (*av_fetch (av, 0, 1)); 1115 PUSHs (*av_fetch (av, 0, 1));
803 PUSHs (sv_cbor); 1116 PUSHs (sv_cbor);
804 1117
812 SPAGAIN; 1125 SPAGAIN;
813 1126
814 if (SvTRUE (ERRSV)) 1127 if (SvTRUE (ERRSV))
815 { 1128 {
816 FREETMPS; LEAVE; 1129 FREETMPS; LEAVE;
817 ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV)))); 1130 ERR_ERRSV;
818 } 1131 }
819 1132
820 SvREFCNT_dec (sv); 1133 SvREFCNT_dec_NN (sv);
821 sv = SvREFCNT_inc (POPs); 1134 sv = SvREFCNT_inc (POPs);
822 1135
823 PUTBACK; 1136 PUTBACK;
824 1137
825 FREETMPS; LEAVE; 1138 FREETMPS; LEAVE;
826
827 return sv;
828 } 1139 }
1140 break;
829 1141
830 default: 1142 default:
1143 filter:
1144 {
1145 SV *tag_sv = newSVuv (tag);
1146
1147 sv = decode_sv (dec);
1148
1149 dSP;
1150 ENTER; SAVETMPS;
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_NN (tag_sv);
1163 FREETMPS; LEAVE;
1164 ERR_ERRSV;
1165 }
1166
1167 if (count)
1168 {
1169 SvREFCNT_dec_NN (tag_sv);
1170 SvREFCNT_dec_NN (sv);
1171 sv = SvREFCNT_inc_NN (TOPs);
1172 SP -= count;
1173 }
1174 else
1175 {
832 AV *av = newAV (); 1176 AV *av = newAV ();
833 av_push (av, newSVuv (tag)); 1177 av_push (av, tag_sv);
834 av_push (av, sv); 1178 av_push (av, sv);
835 1179
836 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 1180 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
837 ? cbor_tagged_stash 1181 ? cbor_tagged_stash
838 : gv_stashpv ("CBOR::XS::Tagged" , 1); 1182 : gv_stashpv ("CBOR::XS::Tagged" , 1);
839
840 return sv_bless (newRV_noinc ((SV *)av), tagged_stash); 1183 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
841 } 1184 }
1185
1186 PUTBACK;
1187
1188 FREETMPS; LEAVE;
1189 }
1190 break;
842 } 1191 }
1192
1193 return sv;
843 1194
844fail: 1195fail:
845 SvREFCNT_dec (sv); 1196 SvREFCNT_dec (sv);
846 return &PL_sv_undef; 1197 return &PL_sv_undef;
847} 1198}
849static SV * 1200static SV *
850decode_sv (dec_t *dec) 1201decode_sv (dec_t *dec)
851{ 1202{
852 WANT (1); 1203 WANT (1);
853 1204
854 switch (*dec->cur >> 5) 1205 switch (*dec->cur >> MAJOR_SHIFT)
855 { 1206 {
856 case 0: // unsigned int 1207 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
857 return newSVuv (decode_uint (dec)); 1208 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
858 case 1: // negative int 1209 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
859 return newSViv (-1 - (IV)decode_uint (dec)); 1210 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
860 case 2: // octet string 1211 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
861 return decode_str (dec, 0); 1212 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
862 case 3: // utf-8 string 1213 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
863 return decode_str (dec, 1); 1214
864 case 4: // array 1215 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) 1216 switch (*dec->cur++ & MINOR_MASK)
872 { 1217 {
873 case 20: 1218 case SIMPLE_FALSE:
874#if CBOR_SLOW 1219#if CBOR_SLOW
875 types_false = get_bool ("Types::Serialiser::false"); 1220 types_false = get_bool ("Types::Serialiser::false");
876#endif 1221#endif
877 return newSVsv (types_false); 1222 return newSVsv (types_false);
878 case 21: 1223 case SIMPLE_TRUE:
879#if CBOR_SLOW 1224#if CBOR_SLOW
880 types_true = get_bool ("Types::Serialiser::true"); 1225 types_true = get_bool ("Types::Serialiser::true");
881#endif 1226#endif
882 return newSVsv (types_true); 1227 return newSVsv (types_true);
883 case 22: 1228 case SIMPLE_NULL:
884 return newSVsv (&PL_sv_undef); 1229 return newSVsv (&PL_sv_undef);
885 case 23: 1230 case SIMPLE_UNDEF:
886#if CBOR_SLOW 1231#if CBOR_SLOW
887 types_error = get_bool ("Types::Serialiser::error"); 1232 types_error = get_bool ("Types::Serialiser::error");
888#endif 1233#endif
889 return newSVsv (types_error); 1234 return newSVsv (types_error);
890 1235
891 case 25: 1236 case MISC_FLOAT16:
892 { 1237 {
893 WANT (2); 1238 WANT (2);
894 1239
895 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1240 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
896 dec->cur += 2; 1241 dec->cur += 2;
897 1242
898 return newSVnv (ecb_binary16_to_float (fp)); 1243 return newSVnv (ecb_binary16_to_float (fp));
899 } 1244 }
900 1245
901 case 26: 1246 case MISC_FLOAT32:
902 { 1247 {
903 uint32_t fp; 1248 uint32_t fp;
904 WANT (4); 1249 WANT (4);
905 memcpy (&fp, dec->cur, 4); 1250 memcpy (&fp, dec->cur, 4);
906 dec->cur += 4; 1251 dec->cur += 4;
909 fp = ecb_bswap32 (fp); 1254 fp = ecb_bswap32 (fp);
910 1255
911 return newSVnv (ecb_binary32_to_float (fp)); 1256 return newSVnv (ecb_binary32_to_float (fp));
912 } 1257 }
913 1258
914 case 27: 1259 case MISC_FLOAT64:
915 { 1260 {
916 uint64_t fp; 1261 uint64_t fp;
917 WANT (8); 1262 WANT (8);
918 memcpy (&fp, dec->cur, 8); 1263 memcpy (&fp, dec->cur, 8);
919 dec->cur += 8; 1264 dec->cur += 8;
922 fp = ecb_bswap64 (fp); 1267 fp = ecb_bswap64 (fp);
923 1268
924 return newSVnv (ecb_binary64_to_double (fp)); 1269 return newSVnv (ecb_binary64_to_double (fp));
925 } 1270 }
926 1271
927 // 0..19 unassigned 1272 // 0..19 unassigned simple
928 // 24 reserved + unassigned (reserved values are not encodable) 1273 // 24 reserved + unassigned simple (reserved values are not encodable)
1274 // 28-30 unassigned misc
1275 // 31 break code
929 default: 1276 default:
930 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1277 ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
931 } 1278 }
932 1279
933 break; 1280 break;
934 } 1281 }
935 1282
938} 1285}
939 1286
940static SV * 1287static SV *
941decode_cbor (SV *string, CBOR *cbor, char **offset_return) 1288decode_cbor (SV *string, CBOR *cbor, char **offset_return)
942{ 1289{
943 dec_t dec = { }; 1290 dec_t dec = { 0 };
944 SV *sv; 1291 SV *sv;
945 STRLEN len; 1292 STRLEN len;
946 char *data = SvPVbyte (string, len); 1293 char *data = SvPVbyte (string, len);
947 1294
948 if (len > cbor->max_size && cbor->max_size) 1295 if (len > cbor->max_size && cbor->max_size)
962 if (dec.cur != dec.end && !dec.err) 1309 if (dec.cur != dec.end && !dec.err)
963 dec.err = "garbage after CBOR object"; 1310 dec.err = "garbage after CBOR object";
964 1311
965 if (dec.err) 1312 if (dec.err)
966 { 1313 {
1314 if (dec.shareable)
1315 {
1316 // need to break cyclic links, which would all be in shareable
1317 int i;
1318 SV **svp;
1319
1320 for (i = av_len (dec.shareable) + 1; i--; )
1321 if ((svp = av_fetch (dec.shareable, i, 0)))
1322 sv_setsv (*svp, &PL_sv_undef);
1323 }
1324
967 SvREFCNT_dec (sv); 1325 SvREFCNT_dec_NN (sv);
1326
1327 if (dec.err_sv)
1328 sv_2mortal (dec.err_sv);
1329
968 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur); 1330 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur);
969 } 1331 }
970 1332
971 sv = sv_2mortal (sv); 1333 sv = sv_2mortal (sv);
972 1334
973 return sv; 1335 return sv;
974} 1336}
975 1337
1338/////////////////////////////////////////////////////////////////////////////
1339// incremental parser
1340
1341#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
1342
1343// returns 0 for notyet, 1 for success or error
1344static int
1345incr_parse (CBOR *self, SV *cborstr)
1346{
1347 STRLEN cur;
1348 SvPV (cborstr, cur);
1349
1350 while (ecb_expect_true (self->incr_need <= cur))
1351 {
1352 // table of integer count bytes
1353 static I8 incr_len[MINOR_MASK + 1] = {
1354 0, 0, 0, 0, 0, 0, 0, 0,
1355 0, 0, 0, 0, 0, 0, 0, 0,
1356 0, 0, 0, 0, 0, 0, 0, 0,
1357 1, 2, 4, 8,-1,-1,-1,-2
1358 };
1359
1360 const U8 *p = SvPVX (cborstr) + self->incr_pos;
1361 U8 m = *p & MINOR_MASK;
1362 IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1363 I8 ilen = incr_len[m];
1364
1365 self->incr_need = self->incr_pos + 1;
1366
1367 if (ecb_expect_false (ilen < 0))
1368 {
1369 if (m != MINOR_INDEF)
1370 return 1; // error
1371
1372 if (*p == (MAJOR_MISC | MINOR_INDEF))
1373 {
1374 if (count >= 0)
1375 return 1; // error
1376
1377 count = 1;
1378 }
1379 else
1380 {
1381 av_push (self->incr_count, newSViv (-1)); //TODO: nest
1382 count = -1;
1383 }
1384 }
1385 else
1386 {
1387 self->incr_need += ilen;
1388 if (ecb_expect_false (self->incr_need > cur))
1389 return 0;
1390
1391 int major = *p >> MAJOR_SHIFT;
1392
1393 switch (major)
1394 {
1395 case MAJOR_TAG >> MAJOR_SHIFT:
1396 ++count; // tags merely prefix another value
1397 break;
1398
1399 case MAJOR_BYTES >> MAJOR_SHIFT:
1400 case MAJOR_TEXT >> MAJOR_SHIFT:
1401 case MAJOR_ARRAY >> MAJOR_SHIFT:
1402 case MAJOR_MAP >> MAJOR_SHIFT:
1403 {
1404 UV len;
1405
1406 if (ecb_expect_false (ilen))
1407 {
1408 len = 0;
1409
1410 do {
1411 len = (len << 8) | *++p;
1412 } while (--ilen);
1413 }
1414 else
1415 len = m;
1416
1417 switch (major)
1418 {
1419 case MAJOR_BYTES >> MAJOR_SHIFT:
1420 case MAJOR_TEXT >> MAJOR_SHIFT:
1421 self->incr_need += len;
1422 if (ecb_expect_false (self->incr_need > cur))
1423 return 0;
1424
1425 break;
1426
1427 case MAJOR_MAP >> MAJOR_SHIFT:
1428 len <<= 1;
1429 case MAJOR_ARRAY >> MAJOR_SHIFT:
1430 if (len)
1431 {
1432 av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
1433 count = len + 1;
1434 }
1435 break;
1436 }
1437 }
1438 }
1439 }
1440
1441 self->incr_pos = self->incr_need;
1442
1443 if (count > 0)
1444 {
1445 while (!--count)
1446 {
1447 if (!AvFILLp (self->incr_count))
1448 return 1; // done
1449
1450 SvREFCNT_dec_NN (av_pop (self->incr_count));
1451 count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1452 }
1453
1454 SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
1455 }
1456 }
1457
1458 return 0;
1459}
1460
1461
976///////////////////////////////////////////////////////////////////////////// 1462/////////////////////////////////////////////////////////////////////////////
977// XS interface functions 1463// XS interface functions
978 1464
979MODULE = CBOR::XS PACKAGE = CBOR::XS 1465MODULE = CBOR::XS PACKAGE = CBOR::XS
980 1466
988 1474
989 types_true = get_bool ("Types::Serialiser::true" ); 1475 types_true = get_bool ("Types::Serialiser::true" );
990 types_false = get_bool ("Types::Serialiser::false"); 1476 types_false = get_bool ("Types::Serialiser::false");
991 types_error = get_bool ("Types::Serialiser::error"); 1477 types_error = get_bool ("Types::Serialiser::error");
992 1478
1479 default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1480
993 sv_cbor = newSVpv ("CBOR", 0); 1481 sv_cbor = newSVpv ("CBOR", 0);
994 SvREADONLY_on (sv_cbor); 1482 SvREADONLY_on (sv_cbor);
1483
1484 assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1));
995} 1485}
996 1486
997PROTOTYPES: DISABLE 1487PROTOTYPES: DISABLE
998 1488
999void CLONE (...) 1489void CLONE (...)
1018void shrink (CBOR *self, int enable = 1) 1508void shrink (CBOR *self, int enable = 1)
1019 ALIAS: 1509 ALIAS:
1020 shrink = F_SHRINK 1510 shrink = F_SHRINK
1021 allow_unknown = F_ALLOW_UNKNOWN 1511 allow_unknown = F_ALLOW_UNKNOWN
1022 allow_sharing = F_ALLOW_SHARING 1512 allow_sharing = F_ALLOW_SHARING
1023 dedup_keys = F_DEDUP_KEYS 1513 allow_cycles = F_ALLOW_CYCLES
1514 forbid_objects = F_FORBID_OBJECTS
1024 dedup_strings = F_DEDUP_STRINGS 1515 pack_strings = F_PACK_STRINGS
1516 text_keys = F_TEXT_KEYS
1517 text_strings = F_TEXT_STRINGS
1518 validate_utf8 = F_VALIDATE_UTF8
1025 PPCODE: 1519 PPCODE:
1026{ 1520{
1027 if (enable) 1521 if (enable)
1028 self->flags |= ix; 1522 self->flags |= ix;
1029 else 1523 else
1035void get_shrink (CBOR *self) 1529void get_shrink (CBOR *self)
1036 ALIAS: 1530 ALIAS:
1037 get_shrink = F_SHRINK 1531 get_shrink = F_SHRINK
1038 get_allow_unknown = F_ALLOW_UNKNOWN 1532 get_allow_unknown = F_ALLOW_UNKNOWN
1039 get_allow_sharing = F_ALLOW_SHARING 1533 get_allow_sharing = F_ALLOW_SHARING
1040 get_dedup_keys = F_DEDUP_KEYS 1534 get_allow_cycles = F_ALLOW_CYCLES
1535 get_forbid_objects = F_FORBID_OBJECTS
1041 get_dedup_strings = F_DEDUP_STRINGS 1536 get_pack_strings = F_PACK_STRINGS
1537 get_text_keys = F_TEXT_KEYS
1538 get_text_strings = F_TEXT_STRINGS
1539 get_validate_utf8 = F_VALIDATE_UTF8
1042 PPCODE: 1540 PPCODE:
1043 XPUSHs (boolSV (self->flags & ix)); 1541 XPUSHs (boolSV (self->flags & ix));
1044 1542
1045void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1543void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1046 PPCODE: 1544 PPCODE:
1059 XPUSHs (ST (0)); 1557 XPUSHs (ST (0));
1060 1558
1061int get_max_size (CBOR *self) 1559int get_max_size (CBOR *self)
1062 CODE: 1560 CODE:
1063 RETVAL = self->max_size; 1561 RETVAL = self->max_size;
1562 OUTPUT:
1563 RETVAL
1564
1565void filter (CBOR *self, SV *filter = 0)
1566 PPCODE:
1567 SvREFCNT_dec (self->filter);
1568 self->filter = filter ? newSVsv (filter) : filter;
1569 XPUSHs (ST (0));
1570
1571SV *get_filter (CBOR *self)
1572 CODE:
1573 RETVAL = self->filter ? self->filter : NEWSV (0, 0);
1064 OUTPUT: 1574 OUTPUT:
1065 RETVAL 1575 RETVAL
1066 1576
1067void encode (CBOR *self, SV *scalar) 1577void encode (CBOR *self, SV *scalar)
1068 PPCODE: 1578 PPCODE:
1083 EXTEND (SP, 2); 1593 EXTEND (SP, 2);
1084 PUSHs (sv); 1594 PUSHs (sv);
1085 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1595 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1086} 1596}
1087 1597
1598void incr_parse (CBOR *self, SV *cborstr)
1599 ALIAS:
1600 incr_parse_multiple = 1
1601 PPCODE:
1602{
1603 if (SvUTF8 (cborstr))
1604 sv_utf8_downgrade (cborstr, 0);
1605
1606 if (!self->incr_count)
1607 {
1608 self->incr_count = newAV ();
1609 self->incr_pos = 0;
1610 self->incr_need = 1;
1611
1612 av_push (self->incr_count, newSViv (1));
1613 }
1614
1615 do
1616 {
1617 if (!incr_parse (self, cborstr))
1618 {
1619 if (self->incr_need > self->max_size && self->max_size)
1620 croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
1621 (unsigned long)self->incr_need, (unsigned long)self->max_size);
1622
1623 break;
1624 }
1625
1626 SV *sv;
1627 char *offset;
1628
1629 PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
1630 XPUSHs (sv);
1631
1632 sv_chop (cborstr, offset);
1633
1634 av_clear (self->incr_count);
1635 av_push (self->incr_count, newSViv (1));
1636
1637 self->incr_pos = 0;
1638 self->incr_need = self->incr_pos + 1;
1639 }
1640 while (ix);
1641}
1642
1643void incr_reset (CBOR *self)
1644 CODE:
1645{
1646 SvREFCNT_dec (self->incr_count);
1647 self->incr_count = 0;
1648}
1649
1650void DESTROY (CBOR *self)
1651 PPCODE:
1652 cbor_free (self);
1653
1088PROTOTYPES: ENABLE 1654PROTOTYPES: ENABLE
1089 1655
1090void encode_cbor (SV *scalar) 1656void encode_cbor (SV *scalar)
1657 ALIAS:
1658 encode_cbor = 0
1659 encode_cbor_sharing = F_ALLOW_SHARING
1091 PPCODE: 1660 PPCODE:
1092{ 1661{
1093 CBOR cbor; 1662 CBOR cbor;
1094 cbor_init (&cbor); 1663 cbor_init (&cbor);
1664 cbor.flags |= ix;
1095 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1665 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1096 XPUSHs (scalar); 1666 XPUSHs (scalar);
1097} 1667}
1098 1668
1099void decode_cbor (SV *cborstr) 1669void decode_cbor (SV *cborstr)
1103 cbor_init (&cbor); 1673 cbor_init (&cbor);
1104 PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN; 1674 PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN;
1105 XPUSHs (cborstr); 1675 XPUSHs (cborstr);
1106} 1676}
1107 1677
1678#ifdef __AFL_COMPILER
1679
1680void
1681afl_init ()
1682 CODE:
1683 __AFL_INIT ();
1684
1685int
1686afl_loop (unsigned int count = 10000)
1687 CODE:
1688 RETVAL = __AFL_LOOP (count);
1689 OUTPUT:
1690 RETVAL
1691
1692#endif
1693

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines