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 1.15 by root, Tue Oct 29 18:37:31 2013 UTC vs.
Revision 1.64 by root, Tue Jun 27 02:03:24 2017 UTC

6#include <string.h> 6#include <string.h>
7#include <stdlib.h> 7#include <stdlib.h>
8#include <stdio.h> 8#include <stdio.h>
9#include <limits.h> 9#include <limits.h>
10#include <float.h> 10#include <float.h>
11#include <inttypes.h>
11 12
13#define ECB_NO_THREADS 1
12#include "ecb.h" 14#include "ecb.h"
13 15
14// compatibility with perl <5.18 16// compatibility with perl <5.18
15#ifndef HvNAMELEN_get 17#ifndef HvNAMELEN_get
16# define HvNAMELEN_get(hv) strlen (HvNAME (hv)) 18# define HvNAMELEN_get(hv) strlen (HvNAME (hv))
19# define HvNAMELEN(hv) HvNAMELEN_get (hv) 21# define HvNAMELEN(hv) HvNAMELEN_get (hv)
20#endif 22#endif
21#ifndef HvNAMEUTF8 23#ifndef HvNAMEUTF8
22# define HvNAMEUTF8(hv) 0 24# define HvNAMEUTF8(hv) 0
23#endif 25#endif
26#ifndef SvREFCNT_inc_NN
27# define SvREFCNT_inc_NN(sv) SvREFCNT_inc (sv)
28#endif
29#ifndef SvREFCNT_dec_NN
30# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
31#endif
32
33// known major and minor types
34enum cbor_type
35{
36 MAJOR_SHIFT = 5,
37 MINOR_MASK = 0x1f,
38
39 MAJOR_POS_INT = 0 << MAJOR_SHIFT,
40 MAJOR_NEG_INT = 1 << MAJOR_SHIFT,
41 MAJOR_BYTES = 2 << MAJOR_SHIFT,
42 MAJOR_TEXT = 3 << MAJOR_SHIFT,
43 MAJOR_ARRAY = 4 << MAJOR_SHIFT,
44 MAJOR_MAP = 5 << MAJOR_SHIFT,
45 MAJOR_TAG = 6 << MAJOR_SHIFT,
46 MAJOR_MISC = 7 << MAJOR_SHIFT,
47
48 // INT/STRING/ARRAY/MAP subtypes
49 LENGTH_EXT1 = 24,
50 LENGTH_EXT2 = 25,
51 LENGTH_EXT4 = 26,
52 LENGTH_EXT8 = 27,
53
54 // SIMPLE types (effectively MISC subtypes)
55 SIMPLE_FALSE = 20,
56 SIMPLE_TRUE = 21,
57 SIMPLE_NULL = 22,
58 SIMPLE_UNDEF = 23,
59
60 // MISC subtype (unused)
61 MISC_EXT1 = 24,
62 MISC_FLOAT16 = 25,
63 MISC_FLOAT32 = 26,
64 MISC_FLOAT64 = 27,
65
66 // BYTES/TEXT/ARRAY/MAP
67 MINOR_INDEF = 31,
68};
24 69
25// known tags 70// known tags
26enum cbor_tag 71enum cbor_tag
27{ 72{
28 // inofficial extensions (pending iana registration) 73 // extensions
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 SAVESTACK_POS ();
516 PUSHMARK (SP);
328 EXTEND (SP, 2); 517 EXTEND (SP, 2);
329 // 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
330 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 519 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
331 PUSHs (sv_cbor); 520 PUSHs (sv_cbor);
332 521
336 525
337 // catch this surprisingly common error 526 // catch this surprisingly common error
338 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) 527 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)); 528 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
340 529
341 encode_uint (enc, 0xc0, CBOR_TAG_PERL_OBJECT); 530 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
342 encode_uint (enc, 0x80, count + 1); 531 encode_uint (enc, MAJOR_ARRAY, count + 1);
343 encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); 532 encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
344 533
345 while (count) 534 while (count)
346 encode_sv (enc, SP[1 - count--]); 535 encode_sv (enc, SP[1 - count--]);
347 536
348 PUTBACK; 537 PUTBACK;
355 } 544 }
356 else if (svt == SVt_PVHV) 545 else if (svt == SVt_PVHV)
357 encode_hv (enc, (HV *)sv); 546 encode_hv (enc, (HV *)sv);
358 else if (svt == SVt_PVAV) 547 else if (svt == SVt_PVAV)
359 encode_av (enc, (AV *)sv); 548 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 549 else
378 croak ("encountered %s, but CBOR can only represent references to arrays or hashes", 550 {
379 SvPV_nolen (sv_2mortal (newRV_inc (sv)))); 551 encode_tag (enc, CBOR_TAG_INDIRECTION);
552 encode_sv (enc, sv);
553 }
380} 554}
381 555
382static void 556static void
383encode_nv (enc_t *enc, SV *sv) 557encode_nv (enc_t *enc, SV *sv)
384{ 558{
385 double nv = SvNVX (sv); 559 double nv = SvNVX (sv);
386 560
387 need (enc, 9); 561 need (enc, 9);
388 562
389 if (ecb_expect_false (nv == (U32)nv)) 563 if (ecb_expect_false (nv == (NV)(U32)nv))
390 encode_uint (enc, 0x00, (U32)nv); 564 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
391 //TODO: maybe I32? 565 //TODO: maybe I32?
392 else if (ecb_expect_false (nv == (float)nv)) 566 else if (ecb_expect_false (nv == (float)nv))
393 { 567 {
568 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
569
394 uint32_t fp = ecb_float_to_binary32 (nv); 570 uint32_t fp = ecb_float_to_binary32 (nv);
395
396 *enc->cur++ = 0xe0 | 26;
397 571
398 if (!ecb_big_endian ()) 572 if (!ecb_big_endian ())
399 fp = ecb_bswap32 (fp); 573 fp = ecb_bswap32 (fp);
400 574
401 memcpy (enc->cur, &fp, 4); 575 memcpy (enc->cur, &fp, 4);
402 enc->cur += 4; 576 enc->cur += 4;
403 } 577 }
404 else 578 else
405 { 579 {
580 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
581
406 uint64_t fp = ecb_double_to_binary64 (nv); 582 uint64_t fp = ecb_double_to_binary64 (nv);
407
408 *enc->cur++ = 0xe0 | 27;
409 583
410 if (!ecb_big_endian ()) 584 if (!ecb_big_endian ())
411 fp = ecb_bswap64 (fp); 585 fp = ecb_bswap64 (fp);
412 586
413 memcpy (enc->cur, &fp, 8); 587 memcpy (enc->cur, &fp, 8);
422 596
423 if (SvPOKp (sv)) 597 if (SvPOKp (sv))
424 { 598 {
425 STRLEN len; 599 STRLEN len;
426 char *str = SvPV (sv, len); 600 char *str = SvPV (sv, len);
427 encode_str (enc, SvUTF8 (sv), str, len); 601 encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len);
428 } 602 }
429 else if (SvNOKp (sv)) 603 else if (SvNOKp (sv))
430 encode_nv (enc, sv); 604 encode_nv (enc, sv);
431 else if (SvIOKp (sv)) 605 else if (SvIOKp (sv))
432 { 606 {
433 if (SvIsUV (sv)) 607 if (SvIsUV (sv))
434 encode_uint (enc, 0x00, SvUVX (sv)); 608 encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
435 else if (SvIVX (sv) >= 0) 609 else if (SvIVX (sv) >= 0)
436 encode_uint (enc, 0x00, SvIVX (sv)); 610 encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
437 else 611 else
438 encode_uint (enc, 0x20, -(SvIVX (sv) + 1)); 612 encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
439 } 613 }
440 else if (SvROK (sv)) 614 else if (SvROK (sv))
441 encode_rv (enc, SvRV (sv)); 615 encode_rv (enc, SvRV (sv));
442 else if (!SvOK (sv)) 616 else if (!SvOK (sv))
443 encode_ch (enc, 0xe0 | 22); 617 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
444 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 618 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
445 encode_ch (enc, 0xe0 | 23); 619 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
446 else 620 else
447 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",
448 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 622 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
449} 623}
450 624
451static SV * 625static SV *
452encode_cbor (SV *scalar, CBOR *cbor) 626encode_cbor (SV *scalar, CBOR *cbor)
453{ 627{
454 enc_t enc; 628 enc_t enc = { 0 };
455 629
456 enc.cbor = *cbor; 630 enc.cbor = *cbor;
457 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 631 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
458 enc.cur = SvPVX (enc.sv); 632 enc.cur = SvPVX (enc.sv);
459 enc.end = SvEND (enc.sv); 633 enc.end = SvEND (enc.sv);
460 enc.depth = 0;
461 634
462 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
463 encode_sv (&enc, scalar); 644 encode_sv (&enc, scalar);
464 645
465 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv)); 646 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv));
466 *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
467 648
481 U8 *end; // end of input string 662 U8 *end; // end of input string
482 const char *err; // parse error, if != 0 663 const char *err; // parse error, if != 0
483 CBOR cbor; 664 CBOR cbor;
484 U32 depth; // recursion depth 665 U32 depth; // recursion depth
485 U32 maxdepth; // recursion depth limit 666 U32 maxdepth; // recursion depth limit
667 AV *shareable;
668 AV *stringref;
669 SV *decode_tagged;
670 SV *err_sv; // optional sv for error, needs to be freed
486} dec_t; 671} dec_t;
487 672
488#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);
489 680
490#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;
491 684
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
492#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)
493#define DEC_DEC_DEPTH --dec->depth 710#define DEC_DEC_DEPTH --dec->depth
494 711
495static UV 712static UV
496decode_uint (dec_t *dec) 713decode_uint (dec_t *dec)
497{ 714{
498 switch (*dec->cur & 31) 715 U8 m = *dec->cur & MINOR_MASK;
499 { 716 ++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;
504 717
505 case 24: 718 if (ecb_expect_true (m < LENGTH_EXT1))
719 return m;
720 else if (ecb_expect_true (m == LENGTH_EXT1))
721 {
506 WANT (2); 722 WANT (1);
507 dec->cur += 2; 723 dec->cur += 1;
508 return dec->cur[-1]; 724 return dec->cur[-1];
509 725 }
510 case 25: 726 else if (ecb_expect_true (m == LENGTH_EXT2))
727 {
511 WANT (3); 728 WANT (2);
512 dec->cur += 3; 729 dec->cur += 2;
513 return (((UV)dec->cur[-2]) << 8) 730 return (((UV)dec->cur[-2]) << 8)
514 | ((UV)dec->cur[-1]); 731 | ((UV)dec->cur[-1]);
515 732 }
516 case 26: 733 else if (ecb_expect_true (m == LENGTH_EXT4))
734 {
517 WANT (5); 735 WANT (4);
518 dec->cur += 5; 736 dec->cur += 4;
519 return (((UV)dec->cur[-4]) << 24) 737 return (((UV)dec->cur[-4]) << 24)
520 | (((UV)dec->cur[-3]) << 16) 738 | (((UV)dec->cur[-3]) << 16)
521 | (((UV)dec->cur[-2]) << 8) 739 | (((UV)dec->cur[-2]) << 8)
522 | ((UV)dec->cur[-1]); 740 | ((UV)dec->cur[-1]);
523 741 }
524 case 27: 742 else if (ecb_expect_true (m == LENGTH_EXT8))
743 {
525 WANT (9); 744 WANT (8);
526 dec->cur += 9; 745 dec->cur += 8;
746
747 return
748#if UVSIZE < 8
749 0
750#else
527 return (((UV)dec->cur[-8]) << 56) 751 (((UV)dec->cur[-8]) << 56)
528 | (((UV)dec->cur[-7]) << 48) 752 | (((UV)dec->cur[-7]) << 48)
529 | (((UV)dec->cur[-6]) << 40) 753 | (((UV)dec->cur[-6]) << 40)
530 | (((UV)dec->cur[-5]) << 32) 754 | (((UV)dec->cur[-5]) << 32)
755#endif
531 | (((UV)dec->cur[-4]) << 24) 756 | (((UV)dec->cur[-4]) << 24)
532 | (((UV)dec->cur[-3]) << 16) 757 | (((UV)dec->cur[-3]) << 16)
533 | (((UV)dec->cur[-2]) << 8) 758 | (((UV)dec->cur[-2]) << 8)
534 | ((UV)dec->cur[-1]); 759 | ((UV)dec->cur[-1]);
535 760 }
536 default: 761 else
537 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 762 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
538 }
539 763
540fail: 764fail:
541 return 0; 765 return 0;
542} 766}
543 767
548{ 772{
549 AV *av = newAV (); 773 AV *av = newAV ();
550 774
551 DEC_INC_DEPTH; 775 DEC_INC_DEPTH;
552 776
553 if ((*dec->cur & 31) == 31) 777 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
554 { 778 {
555 ++dec->cur; 779 ++dec->cur;
556 780
557 for (;;) 781 for (;;)
558 { 782 {
559 WANT (1); 783 WANT (1);
560 784
561 if (*dec->cur == (0xe0 | 31)) 785 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
562 { 786 {
563 ++dec->cur; 787 ++dec->cur;
564 break; 788 break;
565 } 789 }
566 790
567 av_push (av, decode_sv (dec)); 791 av_push (av, decode_sv (dec));
568 } 792 }
569 } 793 }
570 else 794 else
571 { 795 {
572 int i, len = decode_uint (dec); 796 UV i, len = decode_uint (dec);
573 797
798 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); 799 av_fill (av, len - 1);
575 800
576 for (i = 0; i < len; ++i) 801 for (i = 0; i < len; ++i)
577 AvARRAY (av)[i] = decode_sv (dec); 802 AvARRAY (av)[i] = decode_sv (dec);
578 } 803 }
579 804
580 DEC_DEC_DEPTH; 805 DEC_DEC_DEPTH;
581 return newRV_noinc ((SV *)av); 806 return newRV_noinc ((SV *)av);
582 807
583fail: 808fail:
584 SvREFCNT_dec (av); 809 SvREFCNT_dec_NN (av);
585 DEC_DEC_DEPTH; 810 DEC_DEC_DEPTH;
586 return &PL_sv_undef; 811 return &PL_sv_undef;
587} 812}
588 813
814static void
815decode_he (dec_t *dec, HV *hv)
816{
817 // for speed reasons, we specialcase single-string
818 // byte or utf-8 strings as keys, but only when !stringref
819
820 if (ecb_expect_true (!dec->stringref))
821 if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
822 {
823 STRLEN len = decode_uint (dec);
824 char *key = (char *)dec->cur;
825
826 WANT (len);
827 dec->cur += len;
828
829 hv_store (hv, key, len, decode_sv (dec), 0);
830
831 return;
832 }
833 else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
834 {
835 STRLEN len = decode_uint (dec);
836 char *key = (char *)dec->cur;
837
838 WANT (len);
839 dec->cur += len;
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
845 hv_store (hv, key, -len, decode_sv (dec), 0);
846
847 return;
848 }
849
850 SV *k = decode_sv (dec);
851 SV *v = decode_sv (dec);
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
881 hv_store_ent (hv, k, v, 0);
882 SvREFCNT_dec_NN (k);
883
884fail:
885 ;
886}
887
589static SV * 888static SV *
590decode_hv (dec_t *dec) 889decode_hv (dec_t *dec)
591{ 890{
592 HV *hv = newHV (); 891 HV *hv = newHV ();
593 892
594 DEC_INC_DEPTH; 893 DEC_INC_DEPTH;
595 894
596 if ((*dec->cur & 31) == 31) 895 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
597 { 896 {
598 ++dec->cur; 897 ++dec->cur;
599 898
600 for (;;) 899 for (;;)
601 { 900 {
602 WANT (1); 901 WANT (1);
603 902
604 if (*dec->cur == (0xe0 | 31)) 903 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
605 { 904 {
606 ++dec->cur; 905 ++dec->cur;
607 break; 906 break;
608 } 907 }
609 908
610 SV *k = decode_sv (dec); 909 decode_he (dec, hv);
611 SV *v = decode_sv (dec);
612
613 hv_store_ent (hv, k, v, 0);
614 SvREFCNT_dec (k);
615 } 910 }
616 } 911 }
617 else 912 else
618 { 913 {
619 int len = decode_uint (dec); 914 UV pairs = decode_uint (dec);
620 915
916 WANT (pairs); // complexity check - need at least one byte per value, do not allow supersize hashes
917
621 while (len--) 918 while (pairs--)
622 { 919 decode_he (dec, hv);
623 SV *k = decode_sv (dec);
624 SV *v = decode_sv (dec);
625
626 hv_store_ent (hv, k, v, 0);
627 SvREFCNT_dec (k);
628 }
629 } 920 }
630 921
631 DEC_DEC_DEPTH; 922 DEC_DEC_DEPTH;
632 return newRV_noinc ((SV *)hv); 923 return newRV_noinc ((SV *)hv);
633 924
634fail: 925fail:
635 SvREFCNT_dec (hv); 926 SvREFCNT_dec_NN (hv);
636 DEC_DEC_DEPTH; 927 DEC_DEC_DEPTH;
637 return &PL_sv_undef; 928 return &PL_sv_undef;
638} 929}
639 930
640static SV * 931static SV *
641decode_str (dec_t *dec, int utf8) 932decode_str (dec_t *dec, int utf8)
642{ 933{
643 SV *sv = 0; 934 SV *sv = 0;
644 935
645 if ((*dec->cur & 31) == 31) 936 if (ecb_expect_false ((*dec->cur & MINOR_MASK) == MINOR_INDEF))
646 { 937 {
938 // indefinite length strings
647 ++dec->cur; 939 ++dec->cur;
648 940
941 U8 major = *dec->cur & MAJOR_MISC;
942
649 sv = newSVpvn ("", 0); 943 sv = newSVpvn ("", 0);
650 944
651 // not very fast, and certainly not robust against illegal input
652 for (;;) 945 for (;;)
653 { 946 {
654 WANT (1); 947 WANT (1);
655 948
656 if (*dec->cur == (0xe0 | 31)) 949 if ((*dec->cur - major) > LENGTH_EXT8)
950 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
657 { 951 {
658 ++dec->cur; 952 ++dec->cur;
659 break; 953 break;
660 } 954 }
955 else
956 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
661 957
662 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;
663 } 963 }
664 } 964 }
665 else 965 else
666 { 966 {
667 STRLEN len = decode_uint (dec); 967 STRLEN len = decode_uint (dec);
668 968
669 WANT (len); 969 WANT (len);
670 sv = newSVpvn (dec->cur, len); 970 sv = newSVpvn (dec->cur, len);
671 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));
672 } 976 }
673 977
674 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
675 SvUTF8_on (sv); 984 SvUTF8_on (sv);
985 }
676 986
677 return sv; 987 return sv;
678 988
679fail: 989fail:
680 SvREFCNT_dec (sv); 990 SvREFCNT_dec (sv);
682} 992}
683 993
684static SV * 994static SV *
685decode_tagged (dec_t *dec) 995decode_tagged (dec_t *dec)
686{ 996{
997 SV *sv = 0;
687 UV tag = decode_uint (dec); 998 UV tag = decode_uint (dec);
999
1000 WANT (1);
1001
1002 switch (tag)
1003 {
1004 case CBOR_TAG_MAGIC:
688 SV *sv = decode_sv (dec); 1005 sv = decode_sv (dec);
1006 break;
689 1007
690 if (tag == CBOR_TAG_MAGIC) 1008 case CBOR_TAG_INDIRECTION:
691 return sv; 1009 sv = newRV_noinc (decode_sv (dec));
692 else if (tag == CBOR_TAG_PERL_OBJECT) 1010 break;
693 { 1011
1012 case CBOR_TAG_STRINGREF_NAMESPACE:
1013 {
1014 // do not use SAVETMPS/FREETMPS, as these will
1015 // erase mortalised caches, e.g. "shareable"
1016 ENTER;
1017
1018 SAVESPTR (dec->stringref);
1019 dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());
1020
1021 sv = decode_sv (dec);
1022
1023 LEAVE;
1024 }
1025 break;
1026
1027 case CBOR_TAG_STRINGREF:
1028 {
1029 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
1030 ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
1031
1032 UV idx = decode_uint (dec);
1033
1034 if (!dec->stringref || idx >= (UV)(1 + AvFILLp (dec->stringref)))
1035 ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)");
1036
1037 sv = newSVsv (AvARRAY (dec->stringref)[idx]);
1038 }
1039 break;
1040
1041 case CBOR_TAG_VALUE_SHAREABLE:
1042 {
1043 if (ecb_expect_false (!dec->shareable))
1044 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
1045
1046 if (dec->cbor.flags & F_ALLOW_CYCLES)
1047 {
1048 sv = newSV (0);
1049 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
1050
1051 SV *osv = decode_sv (dec);
1052 sv_setsv (sv, osv);
1053 SvREFCNT_dec_NN (osv);
1054 }
1055 else
1056 {
1057 av_push (dec->shareable, &PL_sv_undef);
1058 int idx = AvFILLp (dec->shareable);
1059 sv = decode_sv (dec);
1060 av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
1061 }
1062 }
1063 break;
1064
1065 case CBOR_TAG_VALUE_SHAREDREF:
1066 {
1067 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
1068 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
1069
1070 UV idx = decode_uint (dec);
1071
1072 if (!dec->shareable || idx >= (UV)(1 + AvFILLp (dec->shareable)))
1073 ERR ("corrupted CBOR data (sharedref index out of bounds)");
1074
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");
1079 }
1080 break;
1081
1082 case CBOR_TAG_PERL_OBJECT:
1083 {
1084 if (dec->cbor.flags & F_FORBID_OBJECTS)
1085 goto filter;
1086
1087 sv = decode_sv (dec);
1088
694 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) 1089 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
695 ERR ("corrupted CBOR data (non-array perl object)"); 1090 ERR ("corrupted CBOR data (non-array perl object)");
696 1091
697 AV *av = (AV *)SvRV (sv); 1092 AV *av = (AV *)SvRV (sv);
698 int len = av_len (av) + 1; 1093 int len = av_len (av) + 1;
699 HV *stash = gv_stashsv (*av_fetch (av, 0, 1), 0); 1094 HV *stash = gv_stashsv (*av_fetch (av, 0, 1), 0);
700 1095
701 if (!stash) 1096 if (!stash)
702 ERR ("cannot decode perl-object (package does not exist)"); 1097 ERR ("cannot decode perl-object (package does not exist)");
703 1098
704 GV *method = gv_fetchmethod_autoload (stash, "THAW", 0); 1099 GV *method = gv_fetchmethod_autoload (stash, "THAW", 0);
705 1100
706 if (!method) 1101 if (!method)
707 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)");
708 1103
709 dSP; 1104 dSP;
710 1105
711 ENTER; SAVETMPS; PUSHMARK (SP); 1106 ENTER; SAVETMPS;
1107 PUSHMARK (SP);
712 EXTEND (SP, len + 1); 1108 EXTEND (SP, len + 1);
713 // 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
714 PUSHs (*av_fetch (av, 0, 1)); 1110 PUSHs (*av_fetch (av, 0, 1));
715 PUSHs (sv_cbor); 1111 PUSHs (sv_cbor);
716 1112
717 int i; 1113 int i;
718 1114
719 for (i = 1; i < len; ++i) 1115 for (i = 1; i < len; ++i)
720 PUSHs (*av_fetch (av, i, 1)); 1116 PUSHs (*av_fetch (av, i, 1));
721 1117
722 PUTBACK; 1118 PUTBACK;
723 call_sv ((SV *)GvCV (method), G_SCALAR); 1119 call_sv ((SV *)GvCV (method), G_SCALAR | G_EVAL);
724 SPAGAIN; 1120 SPAGAIN;
725 1121
1122 if (SvTRUE (ERRSV))
1123 {
1124 FREETMPS; LEAVE;
1125 ERR_ERRSV;
1126 }
1127
726 SvREFCNT_dec (sv); 1128 SvREFCNT_dec_NN (sv);
727 sv = SvREFCNT_inc (POPs); 1129 sv = SvREFCNT_inc (POPs);
728 1130
729 PUTBACK; 1131 PUTBACK;
730 1132
731 FREETMPS; LEAVE; 1133 FREETMPS; LEAVE;
1134 }
1135 break;
732 1136
733 return sv; 1137 default:
734 } 1138 filter:
735 else 1139 {
736 { 1140 SV *tag_sv = newSVuv (tag);
1141
1142 sv = decode_sv (dec);
1143
1144 dSP;
1145 ENTER; SAVETMPS;
1146 SAVESTACK_POS ();
1147 PUSHMARK (SP);
1148 EXTEND (SP, 2);
1149 PUSHs (tag_sv);
1150 PUSHs (sv);
1151
1152 PUTBACK;
1153 int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL);
1154 SPAGAIN;
1155
1156 if (SvTRUE (ERRSV))
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 {
737 AV *av = newAV (); 1171 AV *av = newAV ();
738 av_push (av, newSVuv (tag)); 1172 av_push (av, tag_sv);
739 av_push (av, sv); 1173 av_push (av, sv);
740 1174
741 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 1175 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
742 ? cbor_tagged_stash 1176 ? cbor_tagged_stash
743 : gv_stashpv ("CBOR::XS::Tagged" , 1); 1177 : gv_stashpv ("CBOR::XS::Tagged" , 1);
744
745 return sv_bless (newRV_noinc ((SV *)av), tagged_stash); 1178 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
1179 }
1180
1181 PUTBACK;
1182
1183 FREETMPS; LEAVE;
1184 }
1185 break;
746 } 1186 }
1187
1188 return sv;
747 1189
748fail: 1190fail:
749 SvREFCNT_dec (sv); 1191 SvREFCNT_dec (sv);
750 return &PL_sv_undef; 1192 return &PL_sv_undef;
751} 1193}
753static SV * 1195static SV *
754decode_sv (dec_t *dec) 1196decode_sv (dec_t *dec)
755{ 1197{
756 WANT (1); 1198 WANT (1);
757 1199
758 switch (*dec->cur >> 5) 1200 switch (*dec->cur >> MAJOR_SHIFT)
759 { 1201 {
760 case 0: // unsigned int 1202 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
761 return newSVuv (decode_uint (dec)); 1203 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
762 case 1: // negative int 1204 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
763 return newSViv (-1 - (IV)decode_uint (dec)); 1205 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
764 case 2: // octet string 1206 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
765 return decode_str (dec, 0); 1207 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
766 case 3: // utf-8 string 1208 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
767 return decode_str (dec, 1); 1209
768 case 4: // array 1210 case MAJOR_MISC >> MAJOR_SHIFT:
769 return decode_av (dec);
770 case 5: // map
771 return decode_hv (dec);
772 case 6: // tag
773 return decode_tagged (dec);
774 case 7: // misc
775 switch (*dec->cur++ & 31) 1211 switch (*dec->cur++ & MINOR_MASK)
776 { 1212 {
777 case 20: 1213 case SIMPLE_FALSE:
778#if CBOR_SLOW 1214#if CBOR_SLOW
779 types_false = get_bool ("Types::Serialiser::false"); 1215 types_false = get_bool ("Types::Serialiser::false");
780#endif 1216#endif
781 return newSVsv (types_false); 1217 return newSVsv (types_false);
782 case 21: 1218 case SIMPLE_TRUE:
783#if CBOR_SLOW 1219#if CBOR_SLOW
784 types_true = get_bool ("Types::Serialiser::true"); 1220 types_true = get_bool ("Types::Serialiser::true");
785#endif 1221#endif
786 return newSVsv (types_true); 1222 return newSVsv (types_true);
787 case 22: 1223 case SIMPLE_NULL:
788 return newSVsv (&PL_sv_undef); 1224 return newSVsv (&PL_sv_undef);
789 case 23: 1225 case SIMPLE_UNDEF:
790#if CBOR_SLOW 1226#if CBOR_SLOW
791 types_error = get_bool ("Types::Serialiser::error"); 1227 types_error = get_bool ("Types::Serialiser::error");
792#endif 1228#endif
793 return newSVsv (types_error); 1229 return newSVsv (types_error);
794 1230
795 case 25: 1231 case MISC_FLOAT16:
796 { 1232 {
797 WANT (2); 1233 WANT (2);
798 1234
799 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1235 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
800 dec->cur += 2; 1236 dec->cur += 2;
801 1237
802 return newSVnv (ecb_binary16_to_float (fp)); 1238 return newSVnv (ecb_binary16_to_float (fp));
803 } 1239 }
804 1240
805 case 26: 1241 case MISC_FLOAT32:
806 { 1242 {
807 uint32_t fp; 1243 uint32_t fp;
808 WANT (4); 1244 WANT (4);
809 memcpy (&fp, dec->cur, 4); 1245 memcpy (&fp, dec->cur, 4);
810 dec->cur += 4; 1246 dec->cur += 4;
813 fp = ecb_bswap32 (fp); 1249 fp = ecb_bswap32 (fp);
814 1250
815 return newSVnv (ecb_binary32_to_float (fp)); 1251 return newSVnv (ecb_binary32_to_float (fp));
816 } 1252 }
817 1253
818 case 27: 1254 case MISC_FLOAT64:
819 { 1255 {
820 uint64_t fp; 1256 uint64_t fp;
821 WANT (8); 1257 WANT (8);
822 memcpy (&fp, dec->cur, 8); 1258 memcpy (&fp, dec->cur, 8);
823 dec->cur += 8; 1259 dec->cur += 8;
826 fp = ecb_bswap64 (fp); 1262 fp = ecb_bswap64 (fp);
827 1263
828 return newSVnv (ecb_binary64_to_double (fp)); 1264 return newSVnv (ecb_binary64_to_double (fp));
829 } 1265 }
830 1266
831 // 0..19 unassigned 1267 // 0..19 unassigned simple
832 // 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
833 default: 1271 default:
834 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1272 ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
835 } 1273 }
836 1274
837 break; 1275 break;
838 } 1276 }
839 1277
842} 1280}
843 1281
844static SV * 1282static SV *
845decode_cbor (SV *string, CBOR *cbor, char **offset_return) 1283decode_cbor (SV *string, CBOR *cbor, char **offset_return)
846{ 1284{
847 dec_t dec; 1285 dec_t dec = { 0 };
848 SV *sv; 1286 SV *sv;
1287 STRLEN len;
1288 char *data = SvPVbyte (string, len);
849 1289
850 /* work around bugs in 5.10 where manipulating magic values
851 * makes perl ignore the magic in subsequent accesses.
852 * also make a copy of non-PV values, to get them into a clean
853 * state (SvPV should do that, but it's buggy, see below).
854 */
855 /*SvGETMAGIC (string);*/
856 if (SvMAGICAL (string) || !SvPOK (string))
857 string = sv_2mortal (newSVsv (string));
858
859 SvUPGRADE (string, SVt_PV);
860
861 /* work around a bug in perl 5.10, which causes SvCUR to fail an
862 * assertion with -DDEBUGGING, although SvCUR is documented to
863 * return the xpv_cur field which certainly exists after upgrading.
864 * according to nicholas clark, calling SvPOK fixes this.
865 * But it doesn't fix it, so try another workaround, call SvPV_nolen
866 * and hope for the best.
867 * Damnit, SvPV_nolen still trips over yet another assertion. This
868 * assertion business is seriously broken, try yet another workaround
869 * for the broken -DDEBUGGING.
870 */
871 {
872#ifdef DEBUGGING
873 STRLEN offset = SvOK (string) ? sv_len (string) : 0;
874#else
875 STRLEN offset = SvCUR (string);
876#endif
877
878 if (offset > cbor->max_size && cbor->max_size) 1290 if (len > cbor->max_size && cbor->max_size)
879 croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu", 1291 croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
880 (unsigned long)SvCUR (string), (unsigned long)cbor->max_size); 1292 (unsigned long)len, (unsigned long)cbor->max_size);
881 }
882
883 sv_utf8_downgrade (string, 0);
884 1293
885 dec.cbor = *cbor; 1294 dec.cbor = *cbor;
886 dec.cur = (U8 *)SvPVX (string); 1295 dec.cur = (U8 *)data;
887 dec.end = (U8 *)SvEND (string); 1296 dec.end = (U8 *)data + len;
888 dec.err = 0;
889 dec.depth = 0;
890 1297
891 sv = decode_sv (&dec); 1298 sv = decode_sv (&dec);
892 1299
893 if (offset_return) 1300 if (offset_return)
894 *offset_return = dec.cur; 1301 *offset_return = dec.cur;
897 if (dec.cur != dec.end && !dec.err) 1304 if (dec.cur != dec.end && !dec.err)
898 dec.err = "garbage after CBOR object"; 1305 dec.err = "garbage after CBOR object";
899 1306
900 if (dec.err) 1307 if (dec.err)
901 { 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
902 SvREFCNT_dec (sv); 1320 SvREFCNT_dec_NN (sv);
1321
1322 if (dec.err_sv)
1323 sv_2mortal (dec.err_sv);
1324
903 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)SvPVX (string), (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);
904 } 1326 }
905 1327
906 sv = sv_2mortal (sv); 1328 sv = sv_2mortal (sv);
907 1329
908 return sv; 1330 return sv;
909} 1331}
910 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
911///////////////////////////////////////////////////////////////////////////// 1457/////////////////////////////////////////////////////////////////////////////
912// XS interface functions 1458// XS interface functions
913 1459
914MODULE = CBOR::XS PACKAGE = CBOR::XS 1460MODULE = CBOR::XS PACKAGE = CBOR::XS
915 1461
923 1469
924 types_true = get_bool ("Types::Serialiser::true" ); 1470 types_true = get_bool ("Types::Serialiser::true" );
925 types_false = get_bool ("Types::Serialiser::false"); 1471 types_false = get_bool ("Types::Serialiser::false");
926 types_error = get_bool ("Types::Serialiser::error"); 1472 types_error = get_bool ("Types::Serialiser::error");
927 1473
1474 default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1475
928 sv_cbor = newSVpv ("CBOR", 0); 1476 sv_cbor = newSVpv ("CBOR", 0);
929 SvREADONLY_on (sv_cbor); 1477 SvREADONLY_on (sv_cbor);
1478
1479 assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1));
930} 1480}
931 1481
932PROTOTYPES: DISABLE 1482PROTOTYPES: DISABLE
933 1483
934void CLONE (...) 1484void CLONE (...)
952 1502
953void shrink (CBOR *self, int enable = 1) 1503void shrink (CBOR *self, int enable = 1)
954 ALIAS: 1504 ALIAS:
955 shrink = F_SHRINK 1505 shrink = F_SHRINK
956 allow_unknown = F_ALLOW_UNKNOWN 1506 allow_unknown = F_ALLOW_UNKNOWN
1507 allow_sharing = F_ALLOW_SHARING
1508 allow_cycles = F_ALLOW_CYCLES
1509 forbid_objects = F_FORBID_OBJECTS
1510 pack_strings = F_PACK_STRINGS
1511 text_keys = F_TEXT_KEYS
1512 text_strings = F_TEXT_STRINGS
1513 validate_utf8 = F_VALIDATE_UTF8
957 PPCODE: 1514 PPCODE:
958{ 1515{
959 if (enable) 1516 if (enable)
960 self->flags |= ix; 1517 self->flags |= ix;
961 else 1518 else
966 1523
967void get_shrink (CBOR *self) 1524void get_shrink (CBOR *self)
968 ALIAS: 1525 ALIAS:
969 get_shrink = F_SHRINK 1526 get_shrink = F_SHRINK
970 get_allow_unknown = F_ALLOW_UNKNOWN 1527 get_allow_unknown = F_ALLOW_UNKNOWN
1528 get_allow_sharing = F_ALLOW_SHARING
1529 get_allow_cycles = F_ALLOW_CYCLES
1530 get_forbid_objects = F_FORBID_OBJECTS
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
971 PPCODE: 1535 PPCODE:
972 XPUSHs (boolSV (self->flags & ix)); 1536 XPUSHs (boolSV (self->flags & ix));
973 1537
974void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1538void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
975 PPCODE: 1539 PPCODE:
988 XPUSHs (ST (0)); 1552 XPUSHs (ST (0));
989 1553
990int get_max_size (CBOR *self) 1554int get_max_size (CBOR *self)
991 CODE: 1555 CODE:
992 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);
993 OUTPUT: 1569 OUTPUT:
994 RETVAL 1570 RETVAL
995 1571
996void encode (CBOR *self, SV *scalar) 1572void encode (CBOR *self, SV *scalar)
997 PPCODE: 1573 PPCODE:
1012 EXTEND (SP, 2); 1588 EXTEND (SP, 2);
1013 PUSHs (sv); 1589 PUSHs (sv);
1014 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1590 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1015} 1591}
1016 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
1017PROTOTYPES: ENABLE 1649PROTOTYPES: ENABLE
1018 1650
1019void encode_cbor (SV *scalar) 1651void encode_cbor (SV *scalar)
1652 ALIAS:
1653 encode_cbor = 0
1654 encode_cbor_sharing = F_ALLOW_SHARING
1020 PPCODE: 1655 PPCODE:
1021{ 1656{
1022 CBOR cbor; 1657 CBOR cbor;
1023 cbor_init (&cbor); 1658 cbor_init (&cbor);
1659 cbor.flags |= ix;
1024 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1660 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1025 XPUSHs (scalar); 1661 XPUSHs (scalar);
1026} 1662}
1027 1663
1028void decode_cbor (SV *cborstr) 1664void decode_cbor (SV *cborstr)
1032 cbor_init (&cbor); 1668 cbor_init (&cbor);
1033 PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN; 1669 PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN;
1034 XPUSHs (cborstr); 1670 XPUSHs (cborstr);
1035} 1671}
1036 1672
1673#ifdef __AFL_COMPILER
1674
1675void
1676afl_init ()
1677 CODE:
1678 __AFL_INIT ();
1679
1680int
1681afl_loop (unsigned int count = 10000)
1682 CODE:
1683 RETVAL = __AFL_LOOP (count);
1684 OUTPUT:
1685 RETVAL
1686
1687#endif
1688

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines