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 1.16 by root, Tue Oct 29 20:59:16 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
29 CBOR_TAG_PERL_OBJECT = 256, 74 CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
30 CBOR_TAG_GENERIC_OBJECT = 257, 75 CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
76 CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
77 CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
78 CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
79 CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref
80 CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
31 81
32 // rfc7049 82 // rfc7049
33 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8 83 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
34 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any 84 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
35 CBOR_TAG_POS_BIGNUM = 2, // byte string 85 CBOR_TAG_POS_BIGNUM = 2, // byte string
36 CBOR_TAG_NEG_BIGNUM = 3, // byte string 86 CBOR_TAG_NEG_BIGNUM = 3, // byte string
37 CBOR_TAG_DECIMAL = 4, // decimal fraction, array 87 CBOR_TAG_DECIMAL = 4, // decimal fraction, array
38 CBOR_TAG_BIGFLOAT = 5, // array 88 CBOR_TAG_BIGFLOAT = 5, // array
39 89
40 CBOR_TAG_CONV_B64U = 21, // base64url, any 90 CBOR_TAG_CONV_B64U = 21, // base64url, any
41 CBOR_TAG_CONV_B64 = 22, // base64, any 91 CBOR_TAG_CONV_B64 = 22, // base64, any
42 CBOR_TAG_CONV_HEX = 23, // base16, any 92 CBOR_TAG_CONV_HEX = 23, // base16, any
43 CBOR_TAG_CBOR = 24, // embedded cbor, byte string 93 CBOR_TAG_CBOR = 24, // embedded cbor, byte string
44 94
45 CBOR_TAG_URI = 32, // URI rfc3986, utf-8 95 CBOR_TAG_URI = 32, // URI rfc3986, utf-8
46 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8 96 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
47 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8 97 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
48 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8 98 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
49 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8 99 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
50 100
51 CBOR_TAG_MAGIC = 55799 // self-describe cbor 101 CBOR_TAG_MAGIC = 55799, // self-describe cbor
52}; 102};
53 103
54#define F_SHRINK 0x00000200UL 104#define F_SHRINK 0x00000001UL
55#define F_ALLOW_UNKNOWN 0x00002000UL 105#define F_ALLOW_UNKNOWN 0x00000002UL
106#define F_ALLOW_SHARING 0x00000004UL
107#define F_ALLOW_CYCLES 0x00000008UL
108#define F_FORBID_OBJECTS 0x00000010UL
109#define F_PACK_STRINGS 0x00000020UL
110#define F_TEXT_KEYS 0x00000040UL
111#define F_TEXT_STRINGS 0x00000080UL
112#define F_VALIDATE_UTF8 0x00000100UL
56 113
57#define INIT_SIZE 32 // initial scalar size to be allocated 114#define INIT_SIZE 32 // initial scalar size to be allocated
58 115
59#define SB do { 116#define SB do {
60#define SE } while (0) 117#define SE } while (0)
72# define CBOR_SLOW 0 129# define CBOR_SLOW 0
73# define CBOR_STASH cbor_stash 130# define CBOR_STASH cbor_stash
74#endif 131#endif
75 132
76static 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::
77static SV *types_true, *types_false, *types_error, *sv_cbor; 134static SV *types_true, *types_false, *types_error, *sv_cbor, *default_filter;
78 135
79typedef struct { 136typedef struct {
80 U32 flags; 137 U32 flags;
81 U32 max_depth; 138 U32 max_depth;
82 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.
83} CBOR; 146} CBOR;
84 147
85ecb_inline void 148ecb_inline void
86cbor_init (CBOR *cbor) 149cbor_init (CBOR *cbor)
87{ 150{
88 Zero (cbor, 1, CBOR); 151 Zero (cbor, 1, CBOR);
89 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);
90} 160}
91 161
92///////////////////////////////////////////////////////////////////////////// 162/////////////////////////////////////////////////////////////////////////////
93// utility functions 163// utility functions
94 164
116 SvPV_renew (sv, SvCUR (sv) + 1); 186 SvPV_renew (sv, SvCUR (sv) + 1);
117#endif 187#endif
118 } 188 }
119} 189}
120 190
121///////////////////////////////////////////////////////////////////////////// 191// minimum length of a string to be registered for stringref
122// fp hell 192ecb_inline int
123 193minimum_string_length (UV idx)
124//TODO 194{
195 return idx <= 23 ? 3
196 : idx <= 0xffU ? 4
197 : idx <= 0xffffU ? 5
198 : idx <= 0xffffffffU ? 7
199 : 11;
200}
125 201
126///////////////////////////////////////////////////////////////////////////// 202/////////////////////////////////////////////////////////////////////////////
127// encoder 203// encoder
128 204
129// structure used for encoding CBOR 205// structure used for encoding CBOR
132 char *cur; // SvPVX (sv) + current output position 208 char *cur; // SvPVX (sv) + current output position
133 char *end; // SvEND (sv) 209 char *end; // SvEND (sv)
134 SV *sv; // result scalar 210 SV *sv; // result scalar
135 CBOR cbor; 211 CBOR cbor;
136 U32 depth; // recursion level 212 U32 depth; // recursion level
213 HV *stringref[2]; // string => index, or 0 ([0] = bytes, [1] = utf-8)
214 UV stringref_idx;
215 HV *shareable; // ptr => index, or 0
216 UV shareable_idx;
137} enc_t; 217} enc_t;
138 218
139ecb_inline void 219ecb_inline void
140need (enc_t *enc, STRLEN len) 220need (enc_t *enc, STRLEN len)
141{ 221{
142 if (ecb_expect_false (enc->cur + len >= enc->end)) 222 if (ecb_expect_false ((uintptr_t)(enc->end - enc->cur) < len))
143 { 223 {
144 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv); 224 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv);
145 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1); 225 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1);
146 enc->cur = SvPVX (enc->sv) + cur; 226 enc->cur = SvPVX (enc->sv) + cur;
147 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1; 227 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;
158static void 238static void
159encode_uint (enc_t *enc, int major, UV len) 239encode_uint (enc_t *enc, int major, UV len)
160{ 240{
161 need (enc, 9); 241 need (enc, 9);
162 242
163 if (len < 24) 243 if (ecb_expect_true (len < LENGTH_EXT1))
164 *enc->cur++ = major | len; 244 *enc->cur++ = major | len;
165 else if (len <= 0xff) 245 else if (ecb_expect_true (len <= 0xffU))
166 { 246 {
167 *enc->cur++ = major | 24; 247 *enc->cur++ = major | LENGTH_EXT1;
168 *enc->cur++ = len; 248 *enc->cur++ = len;
169 } 249 }
170 else if (len <= 0xffff) 250 else if (len <= 0xffffU)
171 { 251 {
172 *enc->cur++ = major | 25; 252 *enc->cur++ = major | LENGTH_EXT2;
173 *enc->cur++ = len >> 8; 253 *enc->cur++ = len >> 8;
174 *enc->cur++ = len; 254 *enc->cur++ = len;
175 } 255 }
176 else if (len <= 0xffffffff) 256 else if (len <= 0xffffffffU)
177 { 257 {
178 *enc->cur++ = major | 26; 258 *enc->cur++ = major | LENGTH_EXT4;
179 *enc->cur++ = len >> 24; 259 *enc->cur++ = len >> 24;
180 *enc->cur++ = len >> 16; 260 *enc->cur++ = len >> 16;
181 *enc->cur++ = len >> 8; 261 *enc->cur++ = len >> 8;
182 *enc->cur++ = len; 262 *enc->cur++ = len;
183 } 263 }
184 else 264 else
185 { 265 {
186 *enc->cur++ = major | 27; 266 *enc->cur++ = major | LENGTH_EXT8;
187 *enc->cur++ = len >> 56; 267 *enc->cur++ = len >> 56;
188 *enc->cur++ = len >> 48; 268 *enc->cur++ = len >> 48;
189 *enc->cur++ = len >> 40; 269 *enc->cur++ = len >> 40;
190 *enc->cur++ = len >> 32; 270 *enc->cur++ = len >> 32;
191 *enc->cur++ = len >> 24; 271 *enc->cur++ = len >> 24;
193 *enc->cur++ = len >> 8; 273 *enc->cur++ = len >> 8;
194 *enc->cur++ = len; 274 *enc->cur++ = len;
195 } 275 }
196} 276}
197 277
198static 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
199encode_str (enc_t *enc, int utf8, char *str, STRLEN len) 286encode_str_utf8 (enc_t *enc, int utf8, char *str, STRLEN len)
200{ 287{
201 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);
202 need (enc, len); 318 need (enc, len);
203 memcpy (enc->cur, str, len); 319 memcpy (enc->cur, str, len);
204 enc->cur += len; 320 enc->cur += len;
205} 321}
206 322
323ecb_inline void
324encode_strref (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
325{
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);
346}
347
207static void encode_sv (enc_t *enc, SV *sv); 348static void encode_sv (enc_t *enc, SV *sv);
208 349
209static void 350static void
210encode_av (enc_t *enc, AV *av) 351encode_av (enc_t *enc, AV *av)
211{ 352{
214 if (enc->depth >= enc->cbor.max_depth) 355 if (enc->depth >= enc->cbor.max_depth)
215 croak (ERR_NESTING_EXCEEDED); 356 croak (ERR_NESTING_EXCEEDED);
216 357
217 ++enc->depth; 358 ++enc->depth;
218 359
219 encode_uint (enc, 0x80, len + 1); 360 encode_uint (enc, MAJOR_ARRAY, len + 1);
220 361
362 if (ecb_expect_false (SvMAGICAL (av)))
221 for (i = 0; i <= len; ++i) 363 for (i = 0; i <= len; ++i)
222 { 364 {
223 SV **svp = av_fetch (av, i, 0); 365 SV **svp = av_fetch (av, i, 0);
224 encode_sv (enc, svp ? *svp : &PL_sv_undef); 366 encode_sv (enc, svp ? *svp : &PL_sv_undef);
225 } 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 }
226 374
227 --enc->depth; 375 --enc->depth;
228} 376}
229 377
230static void 378static void
238 ++enc->depth; 386 ++enc->depth;
239 387
240 int pairs = hv_iterinit (hv); 388 int pairs = hv_iterinit (hv);
241 int mg = SvMAGICAL (hv); 389 int mg = SvMAGICAL (hv);
242 390
243 if (mg) 391 if (ecb_expect_false (mg))
244 encode_ch (enc, 0xa0 | 31); 392 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
245 else 393 else
246 encode_uint (enc, 0xa0, pairs); 394 encode_uint (enc, MAJOR_MAP, pairs);
247 395
248 while ((he = hv_iternext (hv))) 396 while ((he = hv_iternext (hv)))
249 { 397 {
250 if (HeKLEN (he) == HEf_SVKEY) 398 if (HeKLEN (he) == HEf_SVKEY)
251 encode_sv (enc, HeSVKEY (he)); 399 encode_sv (enc, HeSVKEY (he));
252 else 400 else
253 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));
254 402
255 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));
256 } 404 }
257 405
258 if (mg) 406 if (ecb_expect_false (mg))
259 encode_ch (enc, 0xe0 | 31); 407 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
260 408
261 --enc->depth; 409 --enc->depth;
262} 410}
263 411
264// encode objects, arrays and special \0=false and \1=true values. 412// encode objects, arrays and special \0=false and \1=true values.
265static void 413static void
266encode_rv (enc_t *enc, SV *sv) 414encode_rv (enc_t *enc, SV *sv)
267{ 415{
268 svtype svt;
269
270 SvGETMAGIC (sv); 416 SvGETMAGIC (sv);
417
271 svt = SvTYPE (sv); 418 svtype svt = SvTYPE (sv);
272 419
273 if (ecb_expect_false (SvOBJECT (sv))) 420 if (ecb_expect_false (SvOBJECT (sv)))
274 { 421 {
275 HV *boolean_stash = !CBOR_SLOW || types_boolean_stash 422 HV *boolean_stash = !CBOR_SLOW || types_boolean_stash
276 ? types_boolean_stash 423 ? types_boolean_stash
281 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 428 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
282 ? cbor_tagged_stash 429 ? cbor_tagged_stash
283 : gv_stashpv ("CBOR::XS::Tagged" , 1); 430 : gv_stashpv ("CBOR::XS::Tagged" , 1);
284 431
285 HV *stash = SvSTASH (sv); 432 HV *stash = SvSTASH (sv);
286 GV *method;
287 433
288 if (stash == boolean_stash) 434 if (stash == boolean_stash)
289 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 }
290 else if (stash == error_stash) 439 else if (stash == error_stash)
291 encode_ch (enc, 0xe0 | 23); 440 {
441 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
442 return;
443 }
292 else if (stash == tagged_stash) 444 else if (stash == tagged_stash)
293 { 445 {
294 if (svt != SVt_PVAV) 446 if (svt != SVt_PVAV)
295 croak ("encountered CBOR::XS::Tagged object that isn't an array"); 447 croak ("encountered CBOR::XS::Tagged object that isn't an array");
296 448
297 encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); 449 encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
298 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))
299 } 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))));
300 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) 486 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
301 { 487 {
302 dSP; 488 dSP;
303 489
304 ENTER; SAVETMPS; PUSHMARK (SP); 490 ENTER; SAVETMPS;
491 PUSHMARK (SP);
305 // 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
306 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 493 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
307 494
308 PUTBACK; 495 PUTBACK;
309 // G_SCALAR ensures that return value is 1 496 // G_SCALAR ensures that return value is 1
322 } 509 }
323 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0) 510 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
324 { 511 {
325 dSP; 512 dSP;
326 513
327 ENTER; SAVETMPS; PUSHMARK (SP); 514 ENTER; SAVETMPS;
515 PUSHMARK (SP);
328 EXTEND (SP, 2); 516 EXTEND (SP, 2);
329 // 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
330 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 518 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
331 PUSHs (sv_cbor); 519 PUSHs (sv_cbor);
332 520
336 524
337 // catch this surprisingly common error 525 // catch this surprisingly common error
338 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) 526 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
339 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));
340 528
341 encode_uint (enc, 0xc0, CBOR_TAG_PERL_OBJECT); 529 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
342 encode_uint (enc, 0x80, count + 1); 530 encode_uint (enc, MAJOR_ARRAY, count + 1);
343 encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); 531 encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
344 532
345 while (count) 533 {
534 int i;
535
536 for (i = 0; i < count; ++i)
346 encode_sv (enc, SP[1 - count--]); 537 encode_sv (enc, SP[i + 1 - count]);
538
539 SP -= count;
540 }
347 541
348 PUTBACK; 542 PUTBACK;
349 543
350 FREETMPS; LEAVE; 544 FREETMPS; LEAVE;
351 } 545 }
355 } 549 }
356 else if (svt == SVt_PVHV) 550 else if (svt == SVt_PVHV)
357 encode_hv (enc, (HV *)sv); 551 encode_hv (enc, (HV *)sv);
358 else if (svt == SVt_PVAV) 552 else if (svt == SVt_PVAV)
359 encode_av (enc, (AV *)sv); 553 encode_av (enc, (AV *)sv);
360 else if (svt < SVt_PVAV)
361 {
362 STRLEN len = 0;
363 char *pv = svt ? SvPV (sv, len) : 0;
364
365 if (len == 1 && *pv == '1')
366 encode_ch (enc, 0xe0 | 21);
367 else if (len == 1 && *pv == '0')
368 encode_ch (enc, 0xe0 | 20);
369 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
370 encode_ch (enc, 0xe0 | 23);
371 else
372 croak ("cannot encode reference to scalar '%s' unless the scalar is 0 or 1",
373 SvPV_nolen (sv_2mortal (newRV_inc (sv))));
374 }
375 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
376 encode_ch (enc, 0xe0 | 23);
377 else 554 else
378 croak ("encountered %s, but CBOR can only represent references to arrays or hashes", 555 {
379 SvPV_nolen (sv_2mortal (newRV_inc (sv)))); 556 encode_tag (enc, CBOR_TAG_INDIRECTION);
557 encode_sv (enc, sv);
558 }
380} 559}
381 560
382static void 561static void
383encode_nv (enc_t *enc, SV *sv) 562encode_nv (enc_t *enc, SV *sv)
384{ 563{
385 double nv = SvNVX (sv); 564 double nv = SvNVX (sv);
386 565
387 need (enc, 9); 566 need (enc, 9);
388 567
389 if (ecb_expect_false (nv == (U32)nv)) 568 if (ecb_expect_false (nv == (NV)(U32)nv))
390 encode_uint (enc, 0x00, (U32)nv); 569 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
391 //TODO: maybe I32? 570 //TODO: maybe I32?
392 else if (ecb_expect_false (nv == (float)nv)) 571 else if (ecb_expect_false (nv == (float)nv))
393 { 572 {
573 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
574
394 uint32_t fp = ecb_float_to_binary32 (nv); 575 uint32_t fp = ecb_float_to_binary32 (nv);
395
396 *enc->cur++ = 0xe0 | 26;
397 576
398 if (!ecb_big_endian ()) 577 if (!ecb_big_endian ())
399 fp = ecb_bswap32 (fp); 578 fp = ecb_bswap32 (fp);
400 579
401 memcpy (enc->cur, &fp, 4); 580 memcpy (enc->cur, &fp, 4);
402 enc->cur += 4; 581 enc->cur += 4;
403 } 582 }
404 else 583 else
405 { 584 {
585 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
586
406 uint64_t fp = ecb_double_to_binary64 (nv); 587 uint64_t fp = ecb_double_to_binary64 (nv);
407
408 *enc->cur++ = 0xe0 | 27;
409 588
410 if (!ecb_big_endian ()) 589 if (!ecb_big_endian ())
411 fp = ecb_bswap64 (fp); 590 fp = ecb_bswap64 (fp);
412 591
413 memcpy (enc->cur, &fp, 8); 592 memcpy (enc->cur, &fp, 8);
422 601
423 if (SvPOKp (sv)) 602 if (SvPOKp (sv))
424 { 603 {
425 STRLEN len; 604 STRLEN len;
426 char *str = SvPV (sv, len); 605 char *str = SvPV (sv, len);
427 encode_str (enc, SvUTF8 (sv), str, len); 606 encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len);
428 } 607 }
429 else if (SvNOKp (sv)) 608 else if (SvNOKp (sv))
430 encode_nv (enc, sv); 609 encode_nv (enc, sv);
431 else if (SvIOKp (sv)) 610 else if (SvIOKp (sv))
432 { 611 {
433 if (SvIsUV (sv)) 612 if (SvIsUV (sv))
434 encode_uint (enc, 0x00, SvUVX (sv)); 613 encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
435 else if (SvIVX (sv) >= 0) 614 else if (SvIVX (sv) >= 0)
436 encode_uint (enc, 0x00, SvIVX (sv)); 615 encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
437 else 616 else
438 encode_uint (enc, 0x20, -(SvIVX (sv) + 1)); 617 encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
439 } 618 }
440 else if (SvROK (sv)) 619 else if (SvROK (sv))
441 encode_rv (enc, SvRV (sv)); 620 encode_rv (enc, SvRV (sv));
442 else if (!SvOK (sv)) 621 else if (!SvOK (sv))
443 encode_ch (enc, 0xe0 | 22); 622 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
444 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 623 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
445 encode_ch (enc, 0xe0 | 23); 624 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
446 else 625 else
447 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",
448 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 627 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
449} 628}
450 629
451static SV * 630static SV *
452encode_cbor (SV *scalar, CBOR *cbor) 631encode_cbor (SV *scalar, CBOR *cbor)
453{ 632{
454 enc_t enc; 633 enc_t enc = { 0 };
455 634
456 enc.cbor = *cbor; 635 enc.cbor = *cbor;
457 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 636 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
458 enc.cur = SvPVX (enc.sv); 637 enc.cur = SvPVX (enc.sv);
459 enc.end = SvEND (enc.sv); 638 enc.end = SvEND (enc.sv);
460 enc.depth = 0;
461 639
462 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
463 encode_sv (&enc, scalar); 649 encode_sv (&enc, scalar);
464 650
465 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv)); 651 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv));
466 *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
467 653
481 U8 *end; // end of input string 667 U8 *end; // end of input string
482 const char *err; // parse error, if != 0 668 const char *err; // parse error, if != 0
483 CBOR cbor; 669 CBOR cbor;
484 U32 depth; // recursion depth 670 U32 depth; // recursion depth
485 U32 maxdepth; // recursion depth limit 671 U32 maxdepth; // recursion depth limit
672 AV *shareable;
673 AV *stringref;
674 SV *decode_tagged;
675 SV *err_sv; // optional sv for error, needs to be freed
486} dec_t; 676} dec_t;
487 677
488#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);
489 685
490#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;
491 689
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
492#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)
493#define DEC_DEC_DEPTH --dec->depth 715#define DEC_DEC_DEPTH --dec->depth
494 716
495static UV 717static UV
496decode_uint (dec_t *dec) 718decode_uint (dec_t *dec)
497{ 719{
498 switch (*dec->cur & 31) 720 U8 m = *dec->cur & MINOR_MASK;
499 { 721 ++dec->cur;
500 case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
501 case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
502 case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
503 return *dec->cur++ & 31;
504 722
505 case 24: 723 if (ecb_expect_true (m < LENGTH_EXT1))
724 return m;
725 else if (ecb_expect_true (m == LENGTH_EXT1))
726 {
506 WANT (2); 727 WANT (1);
507 dec->cur += 2; 728 dec->cur += 1;
508 return dec->cur[-1]; 729 return dec->cur[-1];
509 730 }
510 case 25: 731 else if (ecb_expect_true (m == LENGTH_EXT2))
732 {
511 WANT (3); 733 WANT (2);
512 dec->cur += 3; 734 dec->cur += 2;
513 return (((UV)dec->cur[-2]) << 8) 735 return (((UV)dec->cur[-2]) << 8)
514 | ((UV)dec->cur[-1]); 736 | ((UV)dec->cur[-1]);
515 737 }
516 case 26: 738 else if (ecb_expect_true (m == LENGTH_EXT4))
739 {
517 WANT (5); 740 WANT (4);
518 dec->cur += 5; 741 dec->cur += 4;
519 return (((UV)dec->cur[-4]) << 24) 742 return (((UV)dec->cur[-4]) << 24)
520 | (((UV)dec->cur[-3]) << 16) 743 | (((UV)dec->cur[-3]) << 16)
521 | (((UV)dec->cur[-2]) << 8) 744 | (((UV)dec->cur[-2]) << 8)
522 | ((UV)dec->cur[-1]); 745 | ((UV)dec->cur[-1]);
523 746 }
524 case 27: 747 else if (ecb_expect_true (m == LENGTH_EXT8))
748 {
525 WANT (9); 749 WANT (8);
526 dec->cur += 9; 750 dec->cur += 8;
751
752 return
753#if UVSIZE < 8
754 0
755#else
527 return (((UV)dec->cur[-8]) << 56) 756 (((UV)dec->cur[-8]) << 56)
528 | (((UV)dec->cur[-7]) << 48) 757 | (((UV)dec->cur[-7]) << 48)
529 | (((UV)dec->cur[-6]) << 40) 758 | (((UV)dec->cur[-6]) << 40)
530 | (((UV)dec->cur[-5]) << 32) 759 | (((UV)dec->cur[-5]) << 32)
760#endif
531 | (((UV)dec->cur[-4]) << 24) 761 | (((UV)dec->cur[-4]) << 24)
532 | (((UV)dec->cur[-3]) << 16) 762 | (((UV)dec->cur[-3]) << 16)
533 | (((UV)dec->cur[-2]) << 8) 763 | (((UV)dec->cur[-2]) << 8)
534 | ((UV)dec->cur[-1]); 764 | ((UV)dec->cur[-1]);
535 765 }
536 default: 766 else
537 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 767 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
538 }
539 768
540fail: 769fail:
541 return 0; 770 return 0;
542} 771}
543 772
548{ 777{
549 AV *av = newAV (); 778 AV *av = newAV ();
550 779
551 DEC_INC_DEPTH; 780 DEC_INC_DEPTH;
552 781
553 if ((*dec->cur & 31) == 31) 782 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
554 { 783 {
555 ++dec->cur; 784 ++dec->cur;
556 785
557 for (;;) 786 for (;;)
558 { 787 {
559 WANT (1); 788 WANT (1);
560 789
561 if (*dec->cur == (0xe0 | 31)) 790 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
562 { 791 {
563 ++dec->cur; 792 ++dec->cur;
564 break; 793 break;
565 } 794 }
566 795
567 av_push (av, decode_sv (dec)); 796 av_push (av, decode_sv (dec));
568 } 797 }
569 } 798 }
570 else 799 else
571 { 800 {
572 int i, len = decode_uint (dec); 801 UV i, len = decode_uint (dec);
573 802
803 WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
574 av_fill (av, len - 1); 804 av_fill (av, len - 1);
575 805
576 for (i = 0; i < len; ++i) 806 for (i = 0; i < len; ++i)
577 AvARRAY (av)[i] = decode_sv (dec); 807 AvARRAY (av)[i] = decode_sv (dec);
578 } 808 }
579 809
580 DEC_DEC_DEPTH; 810 DEC_DEC_DEPTH;
581 return newRV_noinc ((SV *)av); 811 return newRV_noinc ((SV *)av);
582 812
583fail: 813fail:
584 SvREFCNT_dec (av); 814 SvREFCNT_dec_NN (av);
585 DEC_DEC_DEPTH; 815 DEC_DEC_DEPTH;
586 return &PL_sv_undef; 816 return &PL_sv_undef;
587} 817}
588 818
589static void 819static void
590decode_he (dec_t *dec, HV *hv) 820decode_he (dec_t *dec, HV *hv)
591{ 821{
592 // for speed reasons, we specialcase single-string 822 // for speed reasons, we specialcase single-string
593 // byte or utf-8 strings as keys. 823 // byte or utf-8 strings as keys, but only when !stringref
594 824
595 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))
596 { 827 {
597 I32 len = decode_uint (dec); 828 STRLEN len = decode_uint (dec);
598 char *key = (char *)dec->cur; 829 char *key = (char *)dec->cur;
599 830
831 WANT (len);
600 dec->cur += len; 832 dec->cur += len;
601 833
602 hv_store (hv, key, len, decode_sv (dec), 0); 834 hv_store (hv, key, len, decode_sv (dec), 0);
835
836 return;
603 } 837 }
604 else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27) 838 else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
605 { 839 {
606 I32 len = decode_uint (dec); 840 STRLEN len = decode_uint (dec);
607 char *key = (char *)dec->cur; 841 char *key = (char *)dec->cur;
608 842
843 WANT (len);
609 dec->cur += len; 844 dec->cur += len;
610 845
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
611 hv_store (hv, key, -len, decode_sv (dec), 0); 850 hv_store (hv, key, -len, decode_sv (dec), 0);
851
852 return;
612 } 853 }
613 else 854
614 {
615 SV *k = decode_sv (dec); 855 SV *k = decode_sv (dec);
616 SV *v = decode_sv (dec); 856 SV *v = decode_sv (dec);
617 857
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
618 hv_store_ent (hv, k, v, 0); 886 hv_store_ent (hv, k, v, 0);
619 SvREFCNT_dec (k); 887 SvREFCNT_dec_NN (k);
620 } 888
889fail:
890 ;
621} 891}
622 892
623static SV * 893static SV *
624decode_hv (dec_t *dec) 894decode_hv (dec_t *dec)
625{ 895{
626 HV *hv = newHV (); 896 HV *hv = newHV ();
627 897
628 DEC_INC_DEPTH; 898 DEC_INC_DEPTH;
629 899
630 if ((*dec->cur & 31) == 31) 900 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
631 { 901 {
632 ++dec->cur; 902 ++dec->cur;
633 903
634 for (;;) 904 for (;;)
635 { 905 {
636 WANT (1); 906 WANT (1);
637 907
638 if (*dec->cur == (0xe0 | 31)) 908 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
639 { 909 {
640 ++dec->cur; 910 ++dec->cur;
641 break; 911 break;
642 } 912 }
643 913
644 decode_he (dec, hv); 914 decode_he (dec, hv);
645 } 915 }
646 } 916 }
647 else 917 else
648 { 918 {
649 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
650 922
651 while (pairs--) 923 while (pairs--)
652 decode_he (dec, hv); 924 decode_he (dec, hv);
653 } 925 }
654 926
655 DEC_DEC_DEPTH; 927 DEC_DEC_DEPTH;
656 return newRV_noinc ((SV *)hv); 928 return newRV_noinc ((SV *)hv);
657 929
658fail: 930fail:
659 SvREFCNT_dec (hv); 931 SvREFCNT_dec_NN (hv);
660 DEC_DEC_DEPTH; 932 DEC_DEC_DEPTH;
661 return &PL_sv_undef; 933 return &PL_sv_undef;
662} 934}
663 935
664static SV * 936static SV *
665decode_str (dec_t *dec, int utf8) 937decode_str (dec_t *dec, int utf8)
666{ 938{
667 SV *sv = 0; 939 SV *sv = 0;
668 940
669 if ((*dec->cur & 31) == 31) 941 if (ecb_expect_false ((*dec->cur & MINOR_MASK) == MINOR_INDEF))
670 { 942 {
943 // indefinite length strings
671 ++dec->cur; 944 ++dec->cur;
672 945
946 U8 major = *dec->cur & MAJOR_MISC;
947
673 sv = newSVpvn ("", 0); 948 sv = newSVpvn ("", 0);
674 949
675 // not very fast, and certainly not robust against illegal input
676 for (;;) 950 for (;;)
677 { 951 {
678 WANT (1); 952 WANT (1);
679 953
680 if (*dec->cur == (0xe0 | 31)) 954 if ((*dec->cur - major) > LENGTH_EXT8)
955 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
681 { 956 {
682 ++dec->cur; 957 ++dec->cur;
683 break; 958 break;
684 } 959 }
960 else
961 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
685 962
686 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;
687 } 968 }
688 } 969 }
689 else 970 else
690 { 971 {
691 STRLEN len = decode_uint (dec); 972 STRLEN len = decode_uint (dec);
692 973
693 WANT (len); 974 WANT (len);
694 sv = newSVpvn (dec->cur, len); 975 sv = newSVpvn (dec->cur, len);
695 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));
696 } 981 }
697 982
698 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
699 SvUTF8_on (sv); 989 SvUTF8_on (sv);
990 }
700 991
701 return sv; 992 return sv;
702 993
703fail: 994fail:
704 SvREFCNT_dec (sv); 995 SvREFCNT_dec (sv);
706} 997}
707 998
708static SV * 999static SV *
709decode_tagged (dec_t *dec) 1000decode_tagged (dec_t *dec)
710{ 1001{
1002 SV *sv = 0;
711 UV tag = decode_uint (dec); 1003 UV tag = decode_uint (dec);
1004
1005 WANT (1);
1006
1007 switch (tag)
1008 {
1009 case CBOR_TAG_MAGIC:
712 SV *sv = decode_sv (dec); 1010 sv = decode_sv (dec);
1011 break;
713 1012
714 if (tag == CBOR_TAG_MAGIC) 1013 case CBOR_TAG_INDIRECTION:
715 return sv; 1014 sv = newRV_noinc (decode_sv (dec));
716 else if (tag == CBOR_TAG_PERL_OBJECT) 1015 break;
717 { 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
1046 case CBOR_TAG_VALUE_SHAREABLE:
1047 {
1048 if (ecb_expect_false (!dec->shareable))
1049 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
1050
1051 if (dec->cbor.flags & F_ALLOW_CYCLES)
1052 {
1053 sv = newSV (0);
1054 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
1055
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;
1069
1070 case CBOR_TAG_VALUE_SHAREDREF:
1071 {
1072 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
1073 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
1074
1075 UV idx = decode_uint (dec);
1076
1077 if (!dec->shareable || idx >= (UV)(1 + AvFILLp (dec->shareable)))
1078 ERR ("corrupted CBOR data (sharedref index out of bounds)");
1079
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");
1084 }
1085 break;
1086
1087 case CBOR_TAG_PERL_OBJECT:
1088 {
1089 if (dec->cbor.flags & F_FORBID_OBJECTS)
1090 goto filter;
1091
1092 sv = decode_sv (dec);
1093
718 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) 1094 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
719 ERR ("corrupted CBOR data (non-array perl object)"); 1095 ERR ("corrupted CBOR data (non-array perl object)");
720 1096
721 AV *av = (AV *)SvRV (sv); 1097 AV *av = (AV *)SvRV (sv);
722 int len = av_len (av) + 1; 1098 int len = av_len (av) + 1;
723 HV *stash = gv_stashsv (*av_fetch (av, 0, 1), 0); 1099 HV *stash = gv_stashsv (*av_fetch (av, 0, 1), 0);
724 1100
725 if (!stash) 1101 if (!stash)
726 ERR ("cannot decode perl-object (package does not exist)"); 1102 ERR ("cannot decode perl-object (package does not exist)");
727 1103
728 GV *method = gv_fetchmethod_autoload (stash, "THAW", 0); 1104 GV *method = gv_fetchmethod_autoload (stash, "THAW", 0);
729 1105
730 if (!method) 1106 if (!method)
731 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)");
732 1108
733 dSP; 1109 dSP;
734 1110
735 ENTER; SAVETMPS; PUSHMARK (SP); 1111 ENTER; SAVETMPS;
1112 PUSHMARK (SP);
736 EXTEND (SP, len + 1); 1113 EXTEND (SP, len + 1);
737 // 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
738 PUSHs (*av_fetch (av, 0, 1)); 1115 PUSHs (*av_fetch (av, 0, 1));
739 PUSHs (sv_cbor); 1116 PUSHs (sv_cbor);
740 1117
741 int i; 1118 int i;
742 1119
743 for (i = 1; i < len; ++i) 1120 for (i = 1; i < len; ++i)
744 PUSHs (*av_fetch (av, i, 1)); 1121 PUSHs (*av_fetch (av, i, 1));
745 1122
746 PUTBACK; 1123 PUTBACK;
747 call_sv ((SV *)GvCV (method), G_SCALAR); 1124 call_sv ((SV *)GvCV (method), G_SCALAR | G_EVAL);
748 SPAGAIN; 1125 SPAGAIN;
749 1126
1127 if (SvTRUE (ERRSV))
1128 {
1129 FREETMPS; LEAVE;
1130 ERR_ERRSV;
1131 }
1132
750 SvREFCNT_dec (sv); 1133 SvREFCNT_dec_NN (sv);
751 sv = SvREFCNT_inc (POPs); 1134 sv = SvREFCNT_inc (POPs);
752 1135
753 PUTBACK; 1136 PUTBACK;
754 1137
755 FREETMPS; LEAVE; 1138 FREETMPS; LEAVE;
1139 }
1140 break;
756 1141
757 return sv; 1142 default:
758 } 1143 filter:
759 else 1144 {
760 { 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))
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 {
761 AV *av = newAV (); 1176 AV *av = newAV ();
762 av_push (av, newSVuv (tag)); 1177 av_push (av, tag_sv);
763 av_push (av, sv); 1178 av_push (av, sv);
764 1179
765 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 1180 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
766 ? cbor_tagged_stash 1181 ? cbor_tagged_stash
767 : gv_stashpv ("CBOR::XS::Tagged" , 1); 1182 : gv_stashpv ("CBOR::XS::Tagged" , 1);
768
769 return sv_bless (newRV_noinc ((SV *)av), tagged_stash); 1183 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
1184 }
1185
1186 PUTBACK;
1187
1188 FREETMPS; LEAVE;
1189 }
1190 break;
770 } 1191 }
1192
1193 return sv;
771 1194
772fail: 1195fail:
773 SvREFCNT_dec (sv); 1196 SvREFCNT_dec (sv);
774 return &PL_sv_undef; 1197 return &PL_sv_undef;
775} 1198}
777static SV * 1200static SV *
778decode_sv (dec_t *dec) 1201decode_sv (dec_t *dec)
779{ 1202{
780 WANT (1); 1203 WANT (1);
781 1204
782 switch (*dec->cur >> 5) 1205 switch (*dec->cur >> MAJOR_SHIFT)
783 { 1206 {
784 case 0: // unsigned int 1207 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
785 return newSVuv (decode_uint (dec)); 1208 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
786 case 1: // negative int 1209 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
787 return newSViv (-1 - (IV)decode_uint (dec)); 1210 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
788 case 2: // octet string 1211 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
789 return decode_str (dec, 0); 1212 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
790 case 3: // utf-8 string 1213 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
791 return decode_str (dec, 1); 1214
792 case 4: // array 1215 case MAJOR_MISC >> MAJOR_SHIFT:
793 return decode_av (dec);
794 case 5: // map
795 return decode_hv (dec);
796 case 6: // tag
797 return decode_tagged (dec);
798 case 7: // misc
799 switch (*dec->cur++ & 31) 1216 switch (*dec->cur++ & MINOR_MASK)
800 { 1217 {
801 case 20: 1218 case SIMPLE_FALSE:
802#if CBOR_SLOW 1219#if CBOR_SLOW
803 types_false = get_bool ("Types::Serialiser::false"); 1220 types_false = get_bool ("Types::Serialiser::false");
804#endif 1221#endif
805 return newSVsv (types_false); 1222 return newSVsv (types_false);
806 case 21: 1223 case SIMPLE_TRUE:
807#if CBOR_SLOW 1224#if CBOR_SLOW
808 types_true = get_bool ("Types::Serialiser::true"); 1225 types_true = get_bool ("Types::Serialiser::true");
809#endif 1226#endif
810 return newSVsv (types_true); 1227 return newSVsv (types_true);
811 case 22: 1228 case SIMPLE_NULL:
812 return newSVsv (&PL_sv_undef); 1229 return newSVsv (&PL_sv_undef);
813 case 23: 1230 case SIMPLE_UNDEF:
814#if CBOR_SLOW 1231#if CBOR_SLOW
815 types_error = get_bool ("Types::Serialiser::error"); 1232 types_error = get_bool ("Types::Serialiser::error");
816#endif 1233#endif
817 return newSVsv (types_error); 1234 return newSVsv (types_error);
818 1235
819 case 25: 1236 case MISC_FLOAT16:
820 { 1237 {
821 WANT (2); 1238 WANT (2);
822 1239
823 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1240 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
824 dec->cur += 2; 1241 dec->cur += 2;
825 1242
826 return newSVnv (ecb_binary16_to_float (fp)); 1243 return newSVnv (ecb_binary16_to_float (fp));
827 } 1244 }
828 1245
829 case 26: 1246 case MISC_FLOAT32:
830 { 1247 {
831 uint32_t fp; 1248 uint32_t fp;
832 WANT (4); 1249 WANT (4);
833 memcpy (&fp, dec->cur, 4); 1250 memcpy (&fp, dec->cur, 4);
834 dec->cur += 4; 1251 dec->cur += 4;
837 fp = ecb_bswap32 (fp); 1254 fp = ecb_bswap32 (fp);
838 1255
839 return newSVnv (ecb_binary32_to_float (fp)); 1256 return newSVnv (ecb_binary32_to_float (fp));
840 } 1257 }
841 1258
842 case 27: 1259 case MISC_FLOAT64:
843 { 1260 {
844 uint64_t fp; 1261 uint64_t fp;
845 WANT (8); 1262 WANT (8);
846 memcpy (&fp, dec->cur, 8); 1263 memcpy (&fp, dec->cur, 8);
847 dec->cur += 8; 1264 dec->cur += 8;
850 fp = ecb_bswap64 (fp); 1267 fp = ecb_bswap64 (fp);
851 1268
852 return newSVnv (ecb_binary64_to_double (fp)); 1269 return newSVnv (ecb_binary64_to_double (fp));
853 } 1270 }
854 1271
855 // 0..19 unassigned 1272 // 0..19 unassigned simple
856 // 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
857 default: 1276 default:
858 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1277 ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
859 } 1278 }
860 1279
861 break; 1280 break;
862 } 1281 }
863 1282
866} 1285}
867 1286
868static SV * 1287static SV *
869decode_cbor (SV *string, CBOR *cbor, char **offset_return) 1288decode_cbor (SV *string, CBOR *cbor, char **offset_return)
870{ 1289{
871 dec_t dec; 1290 dec_t dec = { 0 };
872 SV *sv; 1291 SV *sv;
873 STRLEN len; 1292 STRLEN len;
874 char *data = SvPVbyte (string, len); 1293 char *data = SvPVbyte (string, len);
875 1294
876 if (len > cbor->max_size && cbor->max_size) 1295 if (len > cbor->max_size && cbor->max_size)
878 (unsigned long)len, (unsigned long)cbor->max_size); 1297 (unsigned long)len, (unsigned long)cbor->max_size);
879 1298
880 dec.cbor = *cbor; 1299 dec.cbor = *cbor;
881 dec.cur = (U8 *)data; 1300 dec.cur = (U8 *)data;
882 dec.end = (U8 *)data + len; 1301 dec.end = (U8 *)data + len;
883 dec.err = 0;
884 dec.depth = 0;
885 1302
886 sv = decode_sv (&dec); 1303 sv = decode_sv (&dec);
887 1304
888 if (offset_return) 1305 if (offset_return)
889 *offset_return = dec.cur; 1306 *offset_return = dec.cur;
892 if (dec.cur != dec.end && !dec.err) 1309 if (dec.cur != dec.end && !dec.err)
893 dec.err = "garbage after CBOR object"; 1310 dec.err = "garbage after CBOR object";
894 1311
895 if (dec.err) 1312 if (dec.err)
896 { 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
897 SvREFCNT_dec (sv); 1325 SvREFCNT_dec_NN (sv);
1326
1327 if (dec.err_sv)
1328 sv_2mortal (dec.err_sv);
1329
898 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);
899 } 1331 }
900 1332
901 sv = sv_2mortal (sv); 1333 sv = sv_2mortal (sv);
902 1334
903 return sv; 1335 return sv;
904} 1336}
905 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
906///////////////////////////////////////////////////////////////////////////// 1462/////////////////////////////////////////////////////////////////////////////
907// XS interface functions 1463// XS interface functions
908 1464
909MODULE = CBOR::XS PACKAGE = CBOR::XS 1465MODULE = CBOR::XS PACKAGE = CBOR::XS
910 1466
918 1474
919 types_true = get_bool ("Types::Serialiser::true" ); 1475 types_true = get_bool ("Types::Serialiser::true" );
920 types_false = get_bool ("Types::Serialiser::false"); 1476 types_false = get_bool ("Types::Serialiser::false");
921 types_error = get_bool ("Types::Serialiser::error"); 1477 types_error = get_bool ("Types::Serialiser::error");
922 1478
1479 default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1480
923 sv_cbor = newSVpv ("CBOR", 0); 1481 sv_cbor = newSVpv ("CBOR", 0);
924 SvREADONLY_on (sv_cbor); 1482 SvREADONLY_on (sv_cbor);
1483
1484 assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1));
925} 1485}
926 1486
927PROTOTYPES: DISABLE 1487PROTOTYPES: DISABLE
928 1488
929void CLONE (...) 1489void CLONE (...)
947 1507
948void shrink (CBOR *self, int enable = 1) 1508void shrink (CBOR *self, int enable = 1)
949 ALIAS: 1509 ALIAS:
950 shrink = F_SHRINK 1510 shrink = F_SHRINK
951 allow_unknown = F_ALLOW_UNKNOWN 1511 allow_unknown = F_ALLOW_UNKNOWN
1512 allow_sharing = F_ALLOW_SHARING
1513 allow_cycles = F_ALLOW_CYCLES
1514 forbid_objects = F_FORBID_OBJECTS
1515 pack_strings = F_PACK_STRINGS
1516 text_keys = F_TEXT_KEYS
1517 text_strings = F_TEXT_STRINGS
1518 validate_utf8 = F_VALIDATE_UTF8
952 PPCODE: 1519 PPCODE:
953{ 1520{
954 if (enable) 1521 if (enable)
955 self->flags |= ix; 1522 self->flags |= ix;
956 else 1523 else
961 1528
962void get_shrink (CBOR *self) 1529void get_shrink (CBOR *self)
963 ALIAS: 1530 ALIAS:
964 get_shrink = F_SHRINK 1531 get_shrink = F_SHRINK
965 get_allow_unknown = F_ALLOW_UNKNOWN 1532 get_allow_unknown = F_ALLOW_UNKNOWN
1533 get_allow_sharing = F_ALLOW_SHARING
1534 get_allow_cycles = F_ALLOW_CYCLES
1535 get_forbid_objects = F_FORBID_OBJECTS
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
966 PPCODE: 1540 PPCODE:
967 XPUSHs (boolSV (self->flags & ix)); 1541 XPUSHs (boolSV (self->flags & ix));
968 1542
969void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1543void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
970 PPCODE: 1544 PPCODE:
983 XPUSHs (ST (0)); 1557 XPUSHs (ST (0));
984 1558
985int get_max_size (CBOR *self) 1559int get_max_size (CBOR *self)
986 CODE: 1560 CODE:
987 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);
988 OUTPUT: 1574 OUTPUT:
989 RETVAL 1575 RETVAL
990 1576
991void encode (CBOR *self, SV *scalar) 1577void encode (CBOR *self, SV *scalar)
992 PPCODE: 1578 PPCODE:
1007 EXTEND (SP, 2); 1593 EXTEND (SP, 2);
1008 PUSHs (sv); 1594 PUSHs (sv);
1009 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1595 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1010} 1596}
1011 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
1012PROTOTYPES: ENABLE 1654PROTOTYPES: ENABLE
1013 1655
1014void encode_cbor (SV *scalar) 1656void encode_cbor (SV *scalar)
1657 ALIAS:
1658 encode_cbor = 0
1659 encode_cbor_sharing = F_ALLOW_SHARING
1015 PPCODE: 1660 PPCODE:
1016{ 1661{
1017 CBOR cbor; 1662 CBOR cbor;
1018 cbor_init (&cbor); 1663 cbor_init (&cbor);
1664 cbor.flags |= ix;
1019 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1665 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1020 XPUSHs (scalar); 1666 XPUSHs (scalar);
1021} 1667}
1022 1668
1023void decode_cbor (SV *cborstr) 1669void decode_cbor (SV *cborstr)
1027 cbor_init (&cbor); 1673 cbor_init (&cbor);
1028 PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN; 1674 PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN;
1029 XPUSHs (cborstr); 1675 XPUSHs (cborstr);
1030} 1676}
1031 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