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 1.19 by root, Wed Nov 20 02:03:09 2013 UTC vs.
Revision 1.63 by root, Sat Nov 26 04:50:58 2016 UTC

6#include <string.h> 6#include <string.h>
7#include <stdlib.h> 7#include <stdlib.h>
8#include <stdio.h> 8#include <stdio.h>
9#include <limits.h> 9#include <limits.h>
10#include <float.h> 10#include <float.h>
11#include <inttypes.h>
11 12
13#define ECB_NO_THREADS 1
12#include "ecb.h" 14#include "ecb.h"
13 15
14// compatibility with perl <5.18 16// compatibility with perl <5.18
15#ifndef HvNAMELEN_get 17#ifndef HvNAMELEN_get
16# define HvNAMELEN_get(hv) strlen (HvNAME (hv)) 18# define HvNAMELEN_get(hv) strlen (HvNAME (hv))
19# define HvNAMELEN(hv) HvNAMELEN_get (hv) 21# define HvNAMELEN(hv) HvNAMELEN_get (hv)
20#endif 22#endif
21#ifndef HvNAMEUTF8 23#ifndef HvNAMEUTF8
22# define HvNAMEUTF8(hv) 0 24# define HvNAMEUTF8(hv) 0
23#endif 25#endif
26#ifndef SvREFCNT_inc_NN
27# define SvREFCNT_inc_NN(sv) SvREFCNT_inc (sv)
28#endif
29#ifndef SvREFCNT_dec_NN
30# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
31#endif
32
33// known major and minor types
34enum cbor_type
35{
36 MAJOR_SHIFT = 5,
37 MINOR_MASK = 0x1f,
38
39 MAJOR_POS_INT = 0 << MAJOR_SHIFT,
40 MAJOR_NEG_INT = 1 << MAJOR_SHIFT,
41 MAJOR_BYTES = 2 << MAJOR_SHIFT,
42 MAJOR_TEXT = 3 << MAJOR_SHIFT,
43 MAJOR_ARRAY = 4 << MAJOR_SHIFT,
44 MAJOR_MAP = 5 << MAJOR_SHIFT,
45 MAJOR_TAG = 6 << MAJOR_SHIFT,
46 MAJOR_MISC = 7 << MAJOR_SHIFT,
47
48 // INT/STRING/ARRAY/MAP subtypes
49 LENGTH_EXT1 = 24,
50 LENGTH_EXT2 = 25,
51 LENGTH_EXT4 = 26,
52 LENGTH_EXT8 = 27,
53
54 // SIMPLE types (effectively MISC subtypes)
55 SIMPLE_FALSE = 20,
56 SIMPLE_TRUE = 21,
57 SIMPLE_NULL = 22,
58 SIMPLE_UNDEF = 23,
59
60 // MISC subtype (unused)
61 MISC_EXT1 = 24,
62 MISC_FLOAT16 = 25,
63 MISC_FLOAT32 = 26,
64 MISC_FLOAT64 = 27,
65
66 // BYTES/TEXT/ARRAY/MAP
67 MINOR_INDEF = 31,
68};
24 69
25// known tags 70// known tags
26enum cbor_tag 71enum cbor_tag
27{ 72{
28 // inofficial extensions (pending iana registration) 73 // extensions
74 CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
29 CBOR_TAG_PERL_OBJECT = 24, // http://cbor.schmorp.de/perl-object 75 CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
30 CBOR_TAG_GENERIC_OBJECT = 25, // http://cbor.schmorp.de/generic-object 76 CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
31 CBOR_TAG_VALUE_SHAREABLE = 26, // http://cbor.schmorp.de/value-sharing 77 CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
32 CBOR_TAG_VALUE_SHAREDREF = 27, // http://cbor.schmorp.de/value-sharing 78 CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
33 CBOR_TAG_STRINGREF_NAMESPACE = 65537, // http://cbor.schmorp.de/stringref 79 CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref
34 CBOR_TAG_STRINGREF = 28, // http://cbor.schmorp.de/stringref
35 CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection 80 CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
36 81
37 // rfc7049 82 // rfc7049
38 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8 83 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
39 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any 84 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
40 CBOR_TAG_POS_BIGNUM = 2, // byte string 85 CBOR_TAG_POS_BIGNUM = 2, // byte string
41 CBOR_TAG_NEG_BIGNUM = 3, // byte string 86 CBOR_TAG_NEG_BIGNUM = 3, // byte string
42 CBOR_TAG_DECIMAL = 4, // decimal fraction, array 87 CBOR_TAG_DECIMAL = 4, // decimal fraction, array
43 CBOR_TAG_BIGFLOAT = 5, // array 88 CBOR_TAG_BIGFLOAT = 5, // array
44 89
45 CBOR_TAG_CONV_B64U = 21, // base64url, any 90 CBOR_TAG_CONV_B64U = 21, // base64url, any
46 CBOR_TAG_CONV_B64 = 22, // base64, any 91 CBOR_TAG_CONV_B64 = 22, // base64, any
47 CBOR_TAG_CONV_HEX = 23, // base16, any 92 CBOR_TAG_CONV_HEX = 23, // base16, any
48 CBOR_TAG_CBOR = 24, // embedded cbor, byte string 93 CBOR_TAG_CBOR = 24, // embedded cbor, byte string
49 94
50 CBOR_TAG_URI = 32, // URI rfc3986, utf-8 95 CBOR_TAG_URI = 32, // URI rfc3986, utf-8
51 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8 96 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
52 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8 97 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
53 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8 98 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
54 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8 99 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
55 100
56 CBOR_TAG_MAGIC = 55799 // self-describe cbor 101 CBOR_TAG_MAGIC = 55799, // self-describe cbor
57}; 102};
58 103
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 *shareable; // ptr => index, or 0 215 HV *shareable; // ptr => index, or 0
142 UV shareable_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))
327 {
328 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
329
330 if (SvOK (*svp))
331 {
332 // already registered, use stringref
333 encode_tag (enc, CBOR_TAG_STRINGREF);
334 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
335 return;
336 }
337 else if (len >= minimum_string_length (enc->stringref_idx))
338 {
339 // register only
340 sv_setuv (*svp, enc->stringref_idx);
341 ++enc->stringref_idx;
342 }
343 }
344
345 encode_str (enc, upgrade_utf8, utf8, str, len);
217} 346}
218 347
219static void encode_sv (enc_t *enc, SV *sv); 348static void encode_sv (enc_t *enc, SV *sv);
220 349
221static void 350static void
226 if (enc->depth >= enc->cbor.max_depth) 355 if (enc->depth >= enc->cbor.max_depth)
227 croak (ERR_NESTING_EXCEEDED); 356 croak (ERR_NESTING_EXCEEDED);
228 357
229 ++enc->depth; 358 ++enc->depth;
230 359
231 encode_uint (enc, 0x80, len + 1); 360 encode_uint (enc, MAJOR_ARRAY, len + 1);
232 361
362 if (ecb_expect_false (SvMAGICAL (av)))
233 for (i = 0; i <= len; ++i) 363 for (i = 0; i <= len; ++i)
234 { 364 {
235 SV **svp = av_fetch (av, i, 0); 365 SV **svp = av_fetch (av, i, 0);
236 encode_sv (enc, svp ? *svp : &PL_sv_undef); 366 encode_sv (enc, svp ? *svp : &PL_sv_undef);
237 } 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 }
238 374
239 --enc->depth; 375 --enc->depth;
240} 376}
241 377
242static void 378static void
250 ++enc->depth; 386 ++enc->depth;
251 387
252 int pairs = hv_iterinit (hv); 388 int pairs = hv_iterinit (hv);
253 int mg = SvMAGICAL (hv); 389 int mg = SvMAGICAL (hv);
254 390
255 if (mg) 391 if (ecb_expect_false (mg))
256 encode_ch (enc, 0xa0 | 31); 392 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
257 else 393 else
258 encode_uint (enc, 0xa0, pairs); 394 encode_uint (enc, MAJOR_MAP, pairs);
259 395
260 while ((he = hv_iternext (hv))) 396 while ((he = hv_iternext (hv)))
261 { 397 {
262 if (HeKLEN (he) == HEf_SVKEY) 398 if (HeKLEN (he) == HEf_SVKEY)
263 encode_sv (enc, HeSVKEY (he)); 399 encode_sv (enc, HeSVKEY (he));
264 else 400 else
265 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));
266 402
267 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));
268 } 404 }
269 405
270 if (mg) 406 if (ecb_expect_false (mg))
271 encode_ch (enc, 0xe0 | 31); 407 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
272 408
273 --enc->depth; 409 --enc->depth;
274} 410}
275 411
276// encode objects, arrays and special \0=false and \1=true values. 412// encode objects, arrays and special \0=false and \1=true values.
277static void 413static void
278encode_rv (enc_t *enc, SV *sv) 414encode_rv (enc_t *enc, SV *sv)
279{ 415{
280 SvGETMAGIC (sv); 416 SvGETMAGIC (sv);
281
282 if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)
283 && ecb_expect_false (SvREFCNT (sv) > 1))
284 {
285 if (!enc->shareable)
286 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
287
288 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
289
290 if (SvOK (*svp))
291 {
292 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
293 encode_uint (enc, 0x00, SvUV (*svp));
294 return;
295 }
296 else
297 {
298 sv_setuv (*svp, enc->shareable_idx);
299 ++enc->shareable_idx;
300 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
301 }
302 }
303 417
304 svtype svt = SvTYPE (sv); 418 svtype svt = SvTYPE (sv);
305 419
306 if (ecb_expect_false (SvOBJECT (sv))) 420 if (ecb_expect_false (SvOBJECT (sv)))
307 { 421 {
314 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 428 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
315 ? cbor_tagged_stash 429 ? cbor_tagged_stash
316 : gv_stashpv ("CBOR::XS::Tagged" , 1); 430 : gv_stashpv ("CBOR::XS::Tagged" , 1);
317 431
318 HV *stash = SvSTASH (sv); 432 HV *stash = SvSTASH (sv);
319 GV *method;
320 433
321 if (stash == boolean_stash) 434 if (stash == boolean_stash)
322 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 }
323 else if (stash == error_stash) 439 else if (stash == error_stash)
324 encode_ch (enc, 0xe0 | 23); 440 {
441 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
442 return;
443 }
325 else if (stash == tagged_stash) 444 else if (stash == tagged_stash)
326 { 445 {
327 if (svt != SVt_PVAV) 446 if (svt != SVt_PVAV)
328 croak ("encountered CBOR::XS::Tagged object that isn't an array"); 447 croak ("encountered CBOR::XS::Tagged object that isn't an array");
329 448
330 encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); 449 encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
331 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))
332 } 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))));
333 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) 486 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
334 { 487 {
335 dSP; 488 dSP;
336 489
337 ENTER; SAVETMPS; PUSHMARK (SP); 490 ENTER; SAVETMPS;
491 PUSHMARK (SP);
338 // 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
339 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 493 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
340 494
341 PUTBACK; 495 PUTBACK;
342 // G_SCALAR ensures that return value is 1 496 // G_SCALAR ensures that return value is 1
355 } 509 }
356 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0) 510 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
357 { 511 {
358 dSP; 512 dSP;
359 513
360 ENTER; SAVETMPS; PUSHMARK (SP); 514 ENTER; SAVETMPS;
515 SAVESTACK_POS ();
516 PUSHMARK (SP);
361 EXTEND (SP, 2); 517 EXTEND (SP, 2);
362 // 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
363 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 519 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
364 PUSHs (sv_cbor); 520 PUSHs (sv_cbor);
365 521
370 // catch this surprisingly common error 526 // catch this surprisingly common error
371 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) 527 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
372 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));
373 529
374 encode_tag (enc, CBOR_TAG_PERL_OBJECT); 530 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
375 encode_uint (enc, 0x80, count + 1); 531 encode_uint (enc, MAJOR_ARRAY, count + 1);
376 encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); 532 encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
377 533
378 while (count) 534 while (count)
379 encode_sv (enc, SP[1 - count--]); 535 encode_sv (enc, SP[1 - count--]);
380 536
381 PUTBACK; 537 PUTBACK;
402{ 558{
403 double nv = SvNVX (sv); 559 double nv = SvNVX (sv);
404 560
405 need (enc, 9); 561 need (enc, 9);
406 562
407 if (ecb_expect_false (nv == (U32)nv)) 563 if (ecb_expect_false (nv == (NV)(U32)nv))
408 encode_uint (enc, 0x00, (U32)nv); 564 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
409 //TODO: maybe I32? 565 //TODO: maybe I32?
410 else if (ecb_expect_false (nv == (float)nv)) 566 else if (ecb_expect_false (nv == (float)nv))
411 { 567 {
568 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
569
412 uint32_t fp = ecb_float_to_binary32 (nv); 570 uint32_t fp = ecb_float_to_binary32 (nv);
413
414 *enc->cur++ = 0xe0 | 26;
415 571
416 if (!ecb_big_endian ()) 572 if (!ecb_big_endian ())
417 fp = ecb_bswap32 (fp); 573 fp = ecb_bswap32 (fp);
418 574
419 memcpy (enc->cur, &fp, 4); 575 memcpy (enc->cur, &fp, 4);
420 enc->cur += 4; 576 enc->cur += 4;
421 } 577 }
422 else 578 else
423 { 579 {
580 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
581
424 uint64_t fp = ecb_double_to_binary64 (nv); 582 uint64_t fp = ecb_double_to_binary64 (nv);
425
426 *enc->cur++ = 0xe0 | 27;
427 583
428 if (!ecb_big_endian ()) 584 if (!ecb_big_endian ())
429 fp = ecb_bswap64 (fp); 585 fp = ecb_bswap64 (fp);
430 586
431 memcpy (enc->cur, &fp, 8); 587 memcpy (enc->cur, &fp, 8);
440 596
441 if (SvPOKp (sv)) 597 if (SvPOKp (sv))
442 { 598 {
443 STRLEN len; 599 STRLEN len;
444 char *str = SvPV (sv, len); 600 char *str = SvPV (sv, len);
445 encode_str (enc, SvUTF8 (sv), str, len); 601 encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len);
446 } 602 }
447 else if (SvNOKp (sv)) 603 else if (SvNOKp (sv))
448 encode_nv (enc, sv); 604 encode_nv (enc, sv);
449 else if (SvIOKp (sv)) 605 else if (SvIOKp (sv))
450 { 606 {
451 if (SvIsUV (sv)) 607 if (SvIsUV (sv))
452 encode_uint (enc, 0x00, SvUVX (sv)); 608 encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
453 else if (SvIVX (sv) >= 0) 609 else if (SvIVX (sv) >= 0)
454 encode_uint (enc, 0x00, SvIVX (sv)); 610 encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
455 else 611 else
456 encode_uint (enc, 0x20, -(SvIVX (sv) + 1)); 612 encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
457 } 613 }
458 else if (SvROK (sv)) 614 else if (SvROK (sv))
459 encode_rv (enc, SvRV (sv)); 615 encode_rv (enc, SvRV (sv));
460 else if (!SvOK (sv)) 616 else if (!SvOK (sv))
461 encode_ch (enc, 0xe0 | 22); 617 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
462 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 618 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
463 encode_ch (enc, 0xe0 | 23); 619 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
464 else 620 else
465 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",
466 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 622 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
467} 623}
468 624
469static SV * 625static SV *
470encode_cbor (SV *scalar, CBOR *cbor) 626encode_cbor (SV *scalar, CBOR *cbor)
471{ 627{
472 enc_t enc = { }; 628 enc_t enc = { 0 };
473 629
474 enc.cbor = *cbor; 630 enc.cbor = *cbor;
475 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 631 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
476 enc.cur = SvPVX (enc.sv); 632 enc.cur = SvPVX (enc.sv);
477 enc.end = SvEND (enc.sv); 633 enc.end = SvEND (enc.sv);
478 634
479 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
480 encode_sv (&enc, scalar); 644 encode_sv (&enc, scalar);
481 645
482 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv)); 646 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv));
483 *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
484 648
499 const char *err; // parse error, if != 0 663 const char *err; // parse error, if != 0
500 CBOR cbor; 664 CBOR cbor;
501 U32 depth; // recursion depth 665 U32 depth; // recursion depth
502 U32 maxdepth; // recursion depth limit 666 U32 maxdepth; // recursion depth limit
503 AV *shareable; 667 AV *shareable;
668 AV *stringref;
669 SV *decode_tagged;
670 SV *err_sv; // optional sv for error, needs to be freed
504} dec_t; 671} dec_t;
505 672
506#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);
507 680
508#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;
509 684
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
510#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)
511#define DEC_DEC_DEPTH --dec->depth 710#define DEC_DEC_DEPTH --dec->depth
512 711
513static UV 712static UV
514decode_uint (dec_t *dec) 713decode_uint (dec_t *dec)
515{ 714{
516 switch (*dec->cur & 31) 715 U8 m = *dec->cur & MINOR_MASK;
517 { 716 ++dec->cur;
518 case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
519 case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
520 case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
521 return *dec->cur++ & 31;
522 717
523 case 24: 718 if (ecb_expect_true (m < LENGTH_EXT1))
719 return m;
720 else if (ecb_expect_true (m == LENGTH_EXT1))
721 {
524 WANT (2); 722 WANT (1);
525 dec->cur += 2; 723 dec->cur += 1;
526 return dec->cur[-1]; 724 return dec->cur[-1];
527 725 }
528 case 25: 726 else if (ecb_expect_true (m == LENGTH_EXT2))
727 {
529 WANT (3); 728 WANT (2);
530 dec->cur += 3; 729 dec->cur += 2;
531 return (((UV)dec->cur[-2]) << 8) 730 return (((UV)dec->cur[-2]) << 8)
532 | ((UV)dec->cur[-1]); 731 | ((UV)dec->cur[-1]);
533 732 }
534 case 26: 733 else if (ecb_expect_true (m == LENGTH_EXT4))
734 {
535 WANT (5); 735 WANT (4);
536 dec->cur += 5; 736 dec->cur += 4;
537 return (((UV)dec->cur[-4]) << 24) 737 return (((UV)dec->cur[-4]) << 24)
538 | (((UV)dec->cur[-3]) << 16) 738 | (((UV)dec->cur[-3]) << 16)
539 | (((UV)dec->cur[-2]) << 8) 739 | (((UV)dec->cur[-2]) << 8)
540 | ((UV)dec->cur[-1]); 740 | ((UV)dec->cur[-1]);
541 741 }
542 case 27: 742 else if (ecb_expect_true (m == LENGTH_EXT8))
743 {
543 WANT (9); 744 WANT (8);
544 dec->cur += 9; 745 dec->cur += 8;
746
747 return
748#if UVSIZE < 8
749 0
750#else
545 return (((UV)dec->cur[-8]) << 56) 751 (((UV)dec->cur[-8]) << 56)
546 | (((UV)dec->cur[-7]) << 48) 752 | (((UV)dec->cur[-7]) << 48)
547 | (((UV)dec->cur[-6]) << 40) 753 | (((UV)dec->cur[-6]) << 40)
548 | (((UV)dec->cur[-5]) << 32) 754 | (((UV)dec->cur[-5]) << 32)
755#endif
549 | (((UV)dec->cur[-4]) << 24) 756 | (((UV)dec->cur[-4]) << 24)
550 | (((UV)dec->cur[-3]) << 16) 757 | (((UV)dec->cur[-3]) << 16)
551 | (((UV)dec->cur[-2]) << 8) 758 | (((UV)dec->cur[-2]) << 8)
552 | ((UV)dec->cur[-1]); 759 | ((UV)dec->cur[-1]);
553 760 }
554 default: 761 else
555 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 762 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
556 }
557 763
558fail: 764fail:
559 return 0; 765 return 0;
560} 766}
561 767
566{ 772{
567 AV *av = newAV (); 773 AV *av = newAV ();
568 774
569 DEC_INC_DEPTH; 775 DEC_INC_DEPTH;
570 776
571 if ((*dec->cur & 31) == 31) 777 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
572 { 778 {
573 ++dec->cur; 779 ++dec->cur;
574 780
575 for (;;) 781 for (;;)
576 { 782 {
577 WANT (1); 783 WANT (1);
578 784
579 if (*dec->cur == (0xe0 | 31)) 785 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
580 { 786 {
581 ++dec->cur; 787 ++dec->cur;
582 break; 788 break;
583 } 789 }
584 790
585 av_push (av, decode_sv (dec)); 791 av_push (av, decode_sv (dec));
586 } 792 }
587 } 793 }
588 else 794 else
589 { 795 {
590 int i, len = decode_uint (dec); 796 UV i, len = decode_uint (dec);
591 797
798 WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
592 av_fill (av, len - 1); 799 av_fill (av, len - 1);
593 800
594 for (i = 0; i < len; ++i) 801 for (i = 0; i < len; ++i)
595 AvARRAY (av)[i] = decode_sv (dec); 802 AvARRAY (av)[i] = decode_sv (dec);
596 } 803 }
597 804
598 DEC_DEC_DEPTH; 805 DEC_DEC_DEPTH;
599 return newRV_noinc ((SV *)av); 806 return newRV_noinc ((SV *)av);
600 807
601fail: 808fail:
602 SvREFCNT_dec (av); 809 SvREFCNT_dec_NN (av);
603 DEC_DEC_DEPTH; 810 DEC_DEC_DEPTH;
604 return &PL_sv_undef; 811 return &PL_sv_undef;
605} 812}
606 813
607static void 814static void
608decode_he (dec_t *dec, HV *hv) 815decode_he (dec_t *dec, HV *hv)
609{ 816{
610 // for speed reasons, we specialcase single-string 817 // for speed reasons, we specialcase single-string
611 // byte or utf-8 strings as keys. 818 // byte or utf-8 strings as keys, but only when !stringref
612 819
613 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))
614 { 822 {
615 I32 len = decode_uint (dec); 823 STRLEN len = decode_uint (dec);
616 char *key = (char *)dec->cur; 824 char *key = (char *)dec->cur;
617 825
826 WANT (len);
618 dec->cur += len; 827 dec->cur += len;
619 828
620 hv_store (hv, key, len, decode_sv (dec), 0); 829 hv_store (hv, key, len, decode_sv (dec), 0);
830
831 return;
621 } 832 }
622 else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27) 833 else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
623 { 834 {
624 I32 len = decode_uint (dec); 835 STRLEN len = decode_uint (dec);
625 char *key = (char *)dec->cur; 836 char *key = (char *)dec->cur;
626 837
838 WANT (len);
627 dec->cur += len; 839 dec->cur += len;
628 840
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
629 hv_store (hv, key, -len, decode_sv (dec), 0); 845 hv_store (hv, key, -len, decode_sv (dec), 0);
846
847 return;
630 } 848 }
631 else 849
632 {
633 SV *k = decode_sv (dec); 850 SV *k = decode_sv (dec);
634 SV *v = decode_sv (dec); 851 SV *v = decode_sv (dec);
635 852
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
636 hv_store_ent (hv, k, v, 0); 881 hv_store_ent (hv, k, v, 0);
637 SvREFCNT_dec (k); 882 SvREFCNT_dec_NN (k);
638 } 883
884fail:
885 ;
639} 886}
640 887
641static SV * 888static SV *
642decode_hv (dec_t *dec) 889decode_hv (dec_t *dec)
643{ 890{
644 HV *hv = newHV (); 891 HV *hv = newHV ();
645 892
646 DEC_INC_DEPTH; 893 DEC_INC_DEPTH;
647 894
648 if ((*dec->cur & 31) == 31) 895 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
649 { 896 {
650 ++dec->cur; 897 ++dec->cur;
651 898
652 for (;;) 899 for (;;)
653 { 900 {
654 WANT (1); 901 WANT (1);
655 902
656 if (*dec->cur == (0xe0 | 31)) 903 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
657 { 904 {
658 ++dec->cur; 905 ++dec->cur;
659 break; 906 break;
660 } 907 }
661 908
662 decode_he (dec, hv); 909 decode_he (dec, hv);
663 } 910 }
664 } 911 }
665 else 912 else
666 { 913 {
667 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
668 917
669 while (pairs--) 918 while (pairs--)
670 decode_he (dec, hv); 919 decode_he (dec, hv);
671 } 920 }
672 921
673 DEC_DEC_DEPTH; 922 DEC_DEC_DEPTH;
674 return newRV_noinc ((SV *)hv); 923 return newRV_noinc ((SV *)hv);
675 924
676fail: 925fail:
677 SvREFCNT_dec (hv); 926 SvREFCNT_dec_NN (hv);
678 DEC_DEC_DEPTH; 927 DEC_DEC_DEPTH;
679 return &PL_sv_undef; 928 return &PL_sv_undef;
680} 929}
681 930
682static SV * 931static SV *
683decode_str (dec_t *dec, int utf8) 932decode_str (dec_t *dec, int utf8)
684{ 933{
685 SV *sv = 0; 934 SV *sv = 0;
686 935
687 if ((*dec->cur & 31) == 31) 936 if (ecb_expect_false ((*dec->cur & MINOR_MASK) == MINOR_INDEF))
688 { 937 {
938 // indefinite length strings
689 ++dec->cur; 939 ++dec->cur;
690 940
941 U8 major = *dec->cur & MAJOR_MISC;
942
691 sv = newSVpvn ("", 0); 943 sv = newSVpvn ("", 0);
692 944
693 // not very fast, and certainly not robust against illegal input
694 for (;;) 945 for (;;)
695 { 946 {
696 WANT (1); 947 WANT (1);
697 948
698 if (*dec->cur == (0xe0 | 31)) 949 if ((*dec->cur - major) > LENGTH_EXT8)
950 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
699 { 951 {
700 ++dec->cur; 952 ++dec->cur;
701 break; 953 break;
702 } 954 }
955 else
956 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
703 957
704 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;
705 } 963 }
706 } 964 }
707 else 965 else
708 { 966 {
709 STRLEN len = decode_uint (dec); 967 STRLEN len = decode_uint (dec);
710 968
711 WANT (len); 969 WANT (len);
712 sv = newSVpvn (dec->cur, len); 970 sv = newSVpvn (dec->cur, len);
713 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));
714 } 976 }
715 977
716 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
717 SvUTF8_on (sv); 984 SvUTF8_on (sv);
985 }
718 986
719 return sv; 987 return sv;
720 988
721fail: 989fail:
722 SvREFCNT_dec (sv); 990 SvREFCNT_dec (sv);
732 WANT (1); 1000 WANT (1);
733 1001
734 switch (tag) 1002 switch (tag)
735 { 1003 {
736 case CBOR_TAG_MAGIC: 1004 case CBOR_TAG_MAGIC:
737 return decode_sv (dec); 1005 sv = decode_sv (dec);
1006 break;
738 1007
739 case CBOR_TAG_INDIRECTION: 1008 case CBOR_TAG_INDIRECTION:
740 return newRV_noinc (decode_sv (dec)); 1009 sv = newRV_noinc (decode_sv (dec));
1010 break;
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 || (int)idx > 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;
741 1040
742 case CBOR_TAG_VALUE_SHAREABLE: 1041 case CBOR_TAG_VALUE_SHAREABLE:
743 { 1042 {
744 if (ecb_expect_false (!dec->shareable)) 1043 if (ecb_expect_false (!dec->shareable))
745 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ()); 1044 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
746 1045
1046 if (dec->cbor.flags & F_ALLOW_CYCLES)
1047 {
747 sv = newSV (0); 1048 sv = newSV (0);
748 av_push (dec->shareable, SvREFCNT_inc_NN (sv)); 1049 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
749 1050
750 SV *osv = decode_sv (dec); 1051 SV *osv = decode_sv (dec);
751 sv_setsv (sv, osv); 1052 sv_setsv (sv, osv);
752 SvREFCNT_dec_NN (osv); 1053 SvREFCNT_dec_NN (osv);
753 } 1054 }
754 1055 else
755 return sv; 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;
756 1064
757 case CBOR_TAG_VALUE_SHAREDREF: 1065 case CBOR_TAG_VALUE_SHAREDREF:
758 { 1066 {
759 if ((*dec->cur >> 5) != 0) 1067 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
760 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)"); 1068 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
761 1069
762 UV idx = decode_uint (dec); 1070 UV idx = decode_uint (dec);
763 1071
764 if (!dec->shareable || idx > AvFILLp (dec->shareable)) 1072 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable))
765 ERR ("corrupted CBOR data (sharedref index out of bounds)"); 1073 ERR ("corrupted CBOR data (sharedref index out of bounds)");
766 1074
767 return SvREFCNT_inc_NN (AvARRAY (dec->shareable)[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");
768 } 1079 }
1080 break;
769 1081
770 case CBOR_TAG_PERL_OBJECT: 1082 case CBOR_TAG_PERL_OBJECT:
771 { 1083 {
1084 if (dec->cbor.flags & F_FORBID_OBJECTS)
1085 goto filter;
1086
772 sv = decode_sv (dec); 1087 sv = decode_sv (dec);
773 1088
774 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) 1089 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
775 ERR ("corrupted CBOR data (non-array perl object)"); 1090 ERR ("corrupted CBOR data (non-array perl object)");
776 1091
786 if (!method) 1101 if (!method)
787 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)");
788 1103
789 dSP; 1104 dSP;
790 1105
791 ENTER; SAVETMPS; PUSHMARK (SP); 1106 ENTER; SAVETMPS;
1107 PUSHMARK (SP);
792 EXTEND (SP, len + 1); 1108 EXTEND (SP, len + 1);
793 // 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
794 PUSHs (*av_fetch (av, 0, 1)); 1110 PUSHs (*av_fetch (av, 0, 1));
795 PUSHs (sv_cbor); 1111 PUSHs (sv_cbor);
796 1112
804 SPAGAIN; 1120 SPAGAIN;
805 1121
806 if (SvTRUE (ERRSV)) 1122 if (SvTRUE (ERRSV))
807 { 1123 {
808 FREETMPS; LEAVE; 1124 FREETMPS; LEAVE;
809 ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV)))); 1125 ERR_ERRSV;
810 } 1126 }
811 1127
812 SvREFCNT_dec (sv); 1128 SvREFCNT_dec_NN (sv);
813 sv = SvREFCNT_inc (POPs); 1129 sv = SvREFCNT_inc (POPs);
814 1130
815 PUTBACK; 1131 PUTBACK;
816 1132
817 FREETMPS; LEAVE; 1133 FREETMPS; LEAVE;
818
819 return sv;
820 } 1134 }
1135 break;
821 1136
822 default: 1137 default:
1138 filter:
823 { 1139 {
1140 SV *tag_sv = newSVuv (tag);
1141
824 sv = decode_sv (dec); 1142 sv = decode_sv (dec);
825 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))
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 {
826 AV *av = newAV (); 1171 AV *av = newAV ();
827 av_push (av, newSVuv (tag)); 1172 av_push (av, tag_sv);
828 av_push (av, sv); 1173 av_push (av, sv);
829 1174
830 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 1175 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
831 ? cbor_tagged_stash 1176 ? cbor_tagged_stash
832 : gv_stashpv ("CBOR::XS::Tagged" , 1); 1177 : gv_stashpv ("CBOR::XS::Tagged" , 1);
833
834 return sv_bless (newRV_noinc ((SV *)av), tagged_stash); 1178 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
835 } 1179 }
1180
1181 PUTBACK;
1182
1183 FREETMPS; LEAVE;
1184 }
1185 break;
836 } 1186 }
1187
1188 return sv;
837 1189
838fail: 1190fail:
839 SvREFCNT_dec (sv); 1191 SvREFCNT_dec (sv);
840 return &PL_sv_undef; 1192 return &PL_sv_undef;
841} 1193}
843static SV * 1195static SV *
844decode_sv (dec_t *dec) 1196decode_sv (dec_t *dec)
845{ 1197{
846 WANT (1); 1198 WANT (1);
847 1199
848 switch (*dec->cur >> 5) 1200 switch (*dec->cur >> MAJOR_SHIFT)
849 { 1201 {
850 case 0: // unsigned int 1202 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
851 return newSVuv (decode_uint (dec)); 1203 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
852 case 1: // negative int 1204 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
853 return newSViv (-1 - (IV)decode_uint (dec)); 1205 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
854 case 2: // octet string 1206 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
855 return decode_str (dec, 0); 1207 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
856 case 3: // utf-8 string 1208 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
857 return decode_str (dec, 1); 1209
858 case 4: // array 1210 case MAJOR_MISC >> MAJOR_SHIFT:
859 return decode_av (dec);
860 case 5: // map
861 return decode_hv (dec);
862 case 6: // tag
863 return decode_tagged (dec);
864 case 7: // misc
865 switch (*dec->cur++ & 31) 1211 switch (*dec->cur++ & MINOR_MASK)
866 { 1212 {
867 case 20: 1213 case SIMPLE_FALSE:
868#if CBOR_SLOW 1214#if CBOR_SLOW
869 types_false = get_bool ("Types::Serialiser::false"); 1215 types_false = get_bool ("Types::Serialiser::false");
870#endif 1216#endif
871 return newSVsv (types_false); 1217 return newSVsv (types_false);
872 case 21: 1218 case SIMPLE_TRUE:
873#if CBOR_SLOW 1219#if CBOR_SLOW
874 types_true = get_bool ("Types::Serialiser::true"); 1220 types_true = get_bool ("Types::Serialiser::true");
875#endif 1221#endif
876 return newSVsv (types_true); 1222 return newSVsv (types_true);
877 case 22: 1223 case SIMPLE_NULL:
878 return newSVsv (&PL_sv_undef); 1224 return newSVsv (&PL_sv_undef);
879 case 23: 1225 case SIMPLE_UNDEF:
880#if CBOR_SLOW 1226#if CBOR_SLOW
881 types_error = get_bool ("Types::Serialiser::error"); 1227 types_error = get_bool ("Types::Serialiser::error");
882#endif 1228#endif
883 return newSVsv (types_error); 1229 return newSVsv (types_error);
884 1230
885 case 25: 1231 case MISC_FLOAT16:
886 { 1232 {
887 WANT (2); 1233 WANT (2);
888 1234
889 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1235 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
890 dec->cur += 2; 1236 dec->cur += 2;
891 1237
892 return newSVnv (ecb_binary16_to_float (fp)); 1238 return newSVnv (ecb_binary16_to_float (fp));
893 } 1239 }
894 1240
895 case 26: 1241 case MISC_FLOAT32:
896 { 1242 {
897 uint32_t fp; 1243 uint32_t fp;
898 WANT (4); 1244 WANT (4);
899 memcpy (&fp, dec->cur, 4); 1245 memcpy (&fp, dec->cur, 4);
900 dec->cur += 4; 1246 dec->cur += 4;
903 fp = ecb_bswap32 (fp); 1249 fp = ecb_bswap32 (fp);
904 1250
905 return newSVnv (ecb_binary32_to_float (fp)); 1251 return newSVnv (ecb_binary32_to_float (fp));
906 } 1252 }
907 1253
908 case 27: 1254 case MISC_FLOAT64:
909 { 1255 {
910 uint64_t fp; 1256 uint64_t fp;
911 WANT (8); 1257 WANT (8);
912 memcpy (&fp, dec->cur, 8); 1258 memcpy (&fp, dec->cur, 8);
913 dec->cur += 8; 1259 dec->cur += 8;
916 fp = ecb_bswap64 (fp); 1262 fp = ecb_bswap64 (fp);
917 1263
918 return newSVnv (ecb_binary64_to_double (fp)); 1264 return newSVnv (ecb_binary64_to_double (fp));
919 } 1265 }
920 1266
921 // 0..19 unassigned 1267 // 0..19 unassigned simple
922 // 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
923 default: 1271 default:
924 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1272 ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
925 } 1273 }
926 1274
927 break; 1275 break;
928 } 1276 }
929 1277
932} 1280}
933 1281
934static SV * 1282static SV *
935decode_cbor (SV *string, CBOR *cbor, char **offset_return) 1283decode_cbor (SV *string, CBOR *cbor, char **offset_return)
936{ 1284{
937 dec_t dec = { }; 1285 dec_t dec = { 0 };
938 SV *sv; 1286 SV *sv;
939 STRLEN len; 1287 STRLEN len;
940 char *data = SvPVbyte (string, len); 1288 char *data = SvPVbyte (string, len);
941 1289
942 if (len > cbor->max_size && cbor->max_size) 1290 if (len > cbor->max_size && cbor->max_size)
956 if (dec.cur != dec.end && !dec.err) 1304 if (dec.cur != dec.end && !dec.err)
957 dec.err = "garbage after CBOR object"; 1305 dec.err = "garbage after CBOR object";
958 1306
959 if (dec.err) 1307 if (dec.err)
960 { 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
961 SvREFCNT_dec (sv); 1320 SvREFCNT_dec_NN (sv);
1321
1322 if (dec.err_sv)
1323 sv_2mortal (dec.err_sv);
1324
962 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);
963 } 1326 }
964 1327
965 sv = sv_2mortal (sv); 1328 sv = sv_2mortal (sv);
966 1329
967 return sv; 1330 return sv;
968} 1331}
969 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
970///////////////////////////////////////////////////////////////////////////// 1457/////////////////////////////////////////////////////////////////////////////
971// XS interface functions 1458// XS interface functions
972 1459
973MODULE = CBOR::XS PACKAGE = CBOR::XS 1460MODULE = CBOR::XS PACKAGE = CBOR::XS
974 1461
982 1469
983 types_true = get_bool ("Types::Serialiser::true" ); 1470 types_true = get_bool ("Types::Serialiser::true" );
984 types_false = get_bool ("Types::Serialiser::false"); 1471 types_false = get_bool ("Types::Serialiser::false");
985 types_error = get_bool ("Types::Serialiser::error"); 1472 types_error = get_bool ("Types::Serialiser::error");
986 1473
1474 default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1475
987 sv_cbor = newSVpv ("CBOR", 0); 1476 sv_cbor = newSVpv ("CBOR", 0);
988 SvREADONLY_on (sv_cbor); 1477 SvREADONLY_on (sv_cbor);
1478
1479 assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1));
989} 1480}
990 1481
991PROTOTYPES: DISABLE 1482PROTOTYPES: DISABLE
992 1483
993void CLONE (...) 1484void CLONE (...)
1012void shrink (CBOR *self, int enable = 1) 1503void shrink (CBOR *self, int enable = 1)
1013 ALIAS: 1504 ALIAS:
1014 shrink = F_SHRINK 1505 shrink = F_SHRINK
1015 allow_unknown = F_ALLOW_UNKNOWN 1506 allow_unknown = F_ALLOW_UNKNOWN
1016 allow_sharing = F_ALLOW_SHARING 1507 allow_sharing = F_ALLOW_SHARING
1017 dedup_keys = F_DEDUP_KEYS 1508 allow_cycles = F_ALLOW_CYCLES
1509 forbid_objects = F_FORBID_OBJECTS
1018 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
1019 PPCODE: 1514 PPCODE:
1020{ 1515{
1021 if (enable) 1516 if (enable)
1022 self->flags |= ix; 1517 self->flags |= ix;
1023 else 1518 else
1029void get_shrink (CBOR *self) 1524void get_shrink (CBOR *self)
1030 ALIAS: 1525 ALIAS:
1031 get_shrink = F_SHRINK 1526 get_shrink = F_SHRINK
1032 get_allow_unknown = F_ALLOW_UNKNOWN 1527 get_allow_unknown = F_ALLOW_UNKNOWN
1033 get_allow_sharing = F_ALLOW_SHARING 1528 get_allow_sharing = F_ALLOW_SHARING
1034 get_dedup_keys = F_DEDUP_KEYS 1529 get_allow_cycles = F_ALLOW_CYCLES
1530 get_forbid_objects = F_FORBID_OBJECTS
1035 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
1036 PPCODE: 1535 PPCODE:
1037 XPUSHs (boolSV (self->flags & ix)); 1536 XPUSHs (boolSV (self->flags & ix));
1038 1537
1039void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1538void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1040 PPCODE: 1539 PPCODE:
1053 XPUSHs (ST (0)); 1552 XPUSHs (ST (0));
1054 1553
1055int get_max_size (CBOR *self) 1554int get_max_size (CBOR *self)
1056 CODE: 1555 CODE:
1057 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);
1058 OUTPUT: 1569 OUTPUT:
1059 RETVAL 1570 RETVAL
1060 1571
1061void encode (CBOR *self, SV *scalar) 1572void encode (CBOR *self, SV *scalar)
1062 PPCODE: 1573 PPCODE:
1077 EXTEND (SP, 2); 1588 EXTEND (SP, 2);
1078 PUSHs (sv); 1589 PUSHs (sv);
1079 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1590 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1080} 1591}
1081 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
1082PROTOTYPES: ENABLE 1649PROTOTYPES: ENABLE
1083 1650
1084void encode_cbor (SV *scalar) 1651void encode_cbor (SV *scalar)
1652 ALIAS:
1653 encode_cbor = 0
1654 encode_cbor_sharing = F_ALLOW_SHARING
1085 PPCODE: 1655 PPCODE:
1086{ 1656{
1087 CBOR cbor; 1657 CBOR cbor;
1088 cbor_init (&cbor); 1658 cbor_init (&cbor);
1659 cbor.flags |= ix;
1089 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1660 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1090 XPUSHs (scalar); 1661 XPUSHs (scalar);
1091} 1662}
1092 1663
1093void decode_cbor (SV *cborstr) 1664void decode_cbor (SV *cborstr)

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines