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 1.13 by root, Mon Oct 28 22:50:50 2013 UTC vs.
Revision 1.63 by root, Sat Nov 26 04:50:58 2016 UTC

6#include <string.h> 6#include <string.h>
7#include <stdlib.h> 7#include <stdlib.h>
8#include <stdio.h> 8#include <stdio.h>
9#include <limits.h> 9#include <limits.h>
10#include <float.h> 10#include <float.h>
11#include <inttypes.h>
11 12
13#define ECB_NO_THREADS 1
12#include "ecb.h" 14#include "ecb.h"
15
16// compatibility with perl <5.18
17#ifndef HvNAMELEN_get
18# define HvNAMELEN_get(hv) strlen (HvNAME (hv))
19#endif
20#ifndef HvNAMELEN
21# define HvNAMELEN(hv) HvNAMELEN_get (hv)
22#endif
23#ifndef HvNAMEUTF8
24# define HvNAMEUTF8(hv) 0
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};
13 69
14// known tags 70// known tags
15enum cbor_tag 71enum cbor_tag
16{ 72{
17 // inofficial extensions (pending iana registration) 73 // extensions
18 CBOR_TAG_PERL_OBJECT = 256, 74 CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
19 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
20 81
21 // rfc7049 82 // rfc7049
22 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8 83 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
23 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any 84 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
24 CBOR_TAG_POS_BIGNUM = 2, // byte string 85 CBOR_TAG_POS_BIGNUM = 2, // byte string
25 CBOR_TAG_NEG_BIGNUM = 3, // byte string 86 CBOR_TAG_NEG_BIGNUM = 3, // byte string
26 CBOR_TAG_DECIMAL = 4, // decimal fraction, array 87 CBOR_TAG_DECIMAL = 4, // decimal fraction, array
27 CBOR_TAG_BIGFLOAT = 5, // array 88 CBOR_TAG_BIGFLOAT = 5, // array
28 89
29 CBOR_TAG_CONV_B64U = 21, // base64url, any 90 CBOR_TAG_CONV_B64U = 21, // base64url, any
30 CBOR_TAG_CONV_B64 = 22, // base64, any 91 CBOR_TAG_CONV_B64 = 22, // base64, any
31 CBOR_TAG_CONV_HEX = 23, // base16, any 92 CBOR_TAG_CONV_HEX = 23, // base16, any
32 CBOR_TAG_CBOR = 24, // embedded cbor, byte string 93 CBOR_TAG_CBOR = 24, // embedded cbor, byte string
33 94
34 CBOR_TAG_URI = 32, // URI rfc3986, utf-8 95 CBOR_TAG_URI = 32, // URI rfc3986, utf-8
35 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8 96 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
36 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8 97 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
37 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8 98 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
38 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8 99 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
39 100
40 CBOR_TAG_MAGIC = 55799 // self-describe cbor 101 CBOR_TAG_MAGIC = 55799, // self-describe cbor
41}; 102};
42 103
43#define F_SHRINK 0x00000200UL 104#define F_SHRINK 0x00000001UL
44#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
45 113
46#define INIT_SIZE 32 // initial scalar size to be allocated 114#define INIT_SIZE 32 // initial scalar size to be allocated
47 115
48#define SB do { 116#define SB do {
49#define SE } while (0) 117#define SE } while (0)
61# define CBOR_SLOW 0 129# define CBOR_SLOW 0
62# define CBOR_STASH cbor_stash 130# define CBOR_STASH cbor_stash
63#endif 131#endif
64 132
65static 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::
66static SV *types_true, *types_false, *types_error, *sv_cbor; 134static SV *types_true, *types_false, *types_error, *sv_cbor, *default_filter;
67 135
68typedef struct { 136typedef struct {
69 U32 flags; 137 U32 flags;
70 U32 max_depth; 138 U32 max_depth;
71 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.
72} CBOR; 146} CBOR;
73 147
74ecb_inline void 148ecb_inline void
75cbor_init (CBOR *cbor) 149cbor_init (CBOR *cbor)
76{ 150{
77 Zero (cbor, 1, CBOR); 151 Zero (cbor, 1, CBOR);
78 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);
79} 160}
80 161
81///////////////////////////////////////////////////////////////////////////// 162/////////////////////////////////////////////////////////////////////////////
82// utility functions 163// utility functions
83 164
105 SvPV_renew (sv, SvCUR (sv) + 1); 186 SvPV_renew (sv, SvCUR (sv) + 1);
106#endif 187#endif
107 } 188 }
108} 189}
109 190
110///////////////////////////////////////////////////////////////////////////// 191// minimum length of a string to be registered for stringref
111// fp hell 192ecb_inline int
112 193minimum_string_length (UV idx)
113//TODO 194{
195 return idx <= 23 ? 3
196 : idx <= 0xffU ? 4
197 : idx <= 0xffffU ? 5
198 : idx <= 0xffffffffU ? 7
199 : 11;
200}
114 201
115///////////////////////////////////////////////////////////////////////////// 202/////////////////////////////////////////////////////////////////////////////
116// encoder 203// encoder
117 204
118// structure used for encoding CBOR 205// structure used for encoding CBOR
121 char *cur; // SvPVX (sv) + current output position 208 char *cur; // SvPVX (sv) + current output position
122 char *end; // SvEND (sv) 209 char *end; // SvEND (sv)
123 SV *sv; // result scalar 210 SV *sv; // result scalar
124 CBOR cbor; 211 CBOR cbor;
125 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;
126} enc_t; 217} enc_t;
127 218
128ecb_inline void 219ecb_inline void
129need (enc_t *enc, STRLEN len) 220need (enc_t *enc, STRLEN len)
130{ 221{
131 if (ecb_expect_false (enc->cur + len >= enc->end)) 222 if (ecb_expect_false ((uintptr_t)(enc->end - enc->cur) < len))
132 { 223 {
133 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv); 224 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv);
134 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1); 225 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1);
135 enc->cur = SvPVX (enc->sv) + cur; 226 enc->cur = SvPVX (enc->sv) + cur;
136 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1; 227 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;
147static void 238static void
148encode_uint (enc_t *enc, int major, UV len) 239encode_uint (enc_t *enc, int major, UV len)
149{ 240{
150 need (enc, 9); 241 need (enc, 9);
151 242
152 if (len < 24) 243 if (ecb_expect_true (len < LENGTH_EXT1))
153 *enc->cur++ = major | len; 244 *enc->cur++ = major | len;
154 else if (len <= 0xff) 245 else if (ecb_expect_true (len <= 0xffU))
155 { 246 {
156 *enc->cur++ = major | 24; 247 *enc->cur++ = major | LENGTH_EXT1;
157 *enc->cur++ = len; 248 *enc->cur++ = len;
158 } 249 }
159 else if (len <= 0xffff) 250 else if (len <= 0xffffU)
160 { 251 {
161 *enc->cur++ = major | 25; 252 *enc->cur++ = major | LENGTH_EXT2;
162 *enc->cur++ = len >> 8; 253 *enc->cur++ = len >> 8;
163 *enc->cur++ = len; 254 *enc->cur++ = len;
164 } 255 }
165 else if (len <= 0xffffffff) 256 else if (len <= 0xffffffffU)
166 { 257 {
167 *enc->cur++ = major | 26; 258 *enc->cur++ = major | LENGTH_EXT4;
168 *enc->cur++ = len >> 24; 259 *enc->cur++ = len >> 24;
169 *enc->cur++ = len >> 16; 260 *enc->cur++ = len >> 16;
170 *enc->cur++ = len >> 8; 261 *enc->cur++ = len >> 8;
171 *enc->cur++ = len; 262 *enc->cur++ = len;
172 } 263 }
173 else 264 else
174 { 265 {
175 *enc->cur++ = major | 27; 266 *enc->cur++ = major | LENGTH_EXT8;
176 *enc->cur++ = len >> 56; 267 *enc->cur++ = len >> 56;
177 *enc->cur++ = len >> 48; 268 *enc->cur++ = len >> 48;
178 *enc->cur++ = len >> 40; 269 *enc->cur++ = len >> 40;
179 *enc->cur++ = len >> 32; 270 *enc->cur++ = len >> 32;
180 *enc->cur++ = len >> 24; 271 *enc->cur++ = len >> 24;
182 *enc->cur++ = len >> 8; 273 *enc->cur++ = len >> 8;
183 *enc->cur++ = len; 274 *enc->cur++ = len;
184 } 275 }
185} 276}
186 277
187static 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
188encode_str (enc_t *enc, int utf8, char *str, STRLEN len) 286encode_str_utf8 (enc_t *enc, int utf8, char *str, STRLEN len)
189{ 287{
190 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);
191 need (enc, len); 318 need (enc, len);
192 memcpy (enc->cur, str, len); 319 memcpy (enc->cur, str, len);
193 enc->cur += len; 320 enc->cur += len;
194} 321}
195 322
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
196static void encode_sv (enc_t *enc, SV *sv); 348static void encode_sv (enc_t *enc, SV *sv);
197 349
198static void 350static void
199encode_av (enc_t *enc, AV *av) 351encode_av (enc_t *enc, AV *av)
200{ 352{
203 if (enc->depth >= enc->cbor.max_depth) 355 if (enc->depth >= enc->cbor.max_depth)
204 croak (ERR_NESTING_EXCEEDED); 356 croak (ERR_NESTING_EXCEEDED);
205 357
206 ++enc->depth; 358 ++enc->depth;
207 359
208 encode_uint (enc, 0x80, len + 1); 360 encode_uint (enc, MAJOR_ARRAY, len + 1);
209 361
362 if (ecb_expect_false (SvMAGICAL (av)))
210 for (i = 0; i <= len; ++i) 363 for (i = 0; i <= len; ++i)
211 { 364 {
212 SV **svp = av_fetch (av, i, 0); 365 SV **svp = av_fetch (av, i, 0);
213 encode_sv (enc, svp ? *svp : &PL_sv_undef); 366 encode_sv (enc, svp ? *svp : &PL_sv_undef);
214 } 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 }
215 374
216 --enc->depth; 375 --enc->depth;
217} 376}
218 377
219static void 378static void
227 ++enc->depth; 386 ++enc->depth;
228 387
229 int pairs = hv_iterinit (hv); 388 int pairs = hv_iterinit (hv);
230 int mg = SvMAGICAL (hv); 389 int mg = SvMAGICAL (hv);
231 390
232 if (mg) 391 if (ecb_expect_false (mg))
233 encode_ch (enc, 0xa0 | 31); 392 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
234 else 393 else
235 encode_uint (enc, 0xa0, pairs); 394 encode_uint (enc, MAJOR_MAP, pairs);
236 395
237 while ((he = hv_iternext (hv))) 396 while ((he = hv_iternext (hv)))
238 { 397 {
239 if (HeKLEN (he) == HEf_SVKEY) 398 if (HeKLEN (he) == HEf_SVKEY)
240 encode_sv (enc, HeSVKEY (he)); 399 encode_sv (enc, HeSVKEY (he));
241 else 400 else
242 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));
243 402
244 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));
245 } 404 }
246 405
247 if (mg) 406 if (ecb_expect_false (mg))
248 encode_ch (enc, 0xe0 | 31); 407 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
249 408
250 --enc->depth; 409 --enc->depth;
251} 410}
252 411
253// encode objects, arrays and special \0=false and \1=true values. 412// encode objects, arrays and special \0=false and \1=true values.
254static void 413static void
255encode_rv (enc_t *enc, SV *sv) 414encode_rv (enc_t *enc, SV *sv)
256{ 415{
257 svtype svt;
258
259 SvGETMAGIC (sv); 416 SvGETMAGIC (sv);
417
260 svt = SvTYPE (sv); 418 svtype svt = SvTYPE (sv);
261 419
262 if (ecb_expect_false (SvOBJECT (sv))) 420 if (ecb_expect_false (SvOBJECT (sv)))
263 { 421 {
264 HV *boolean_stash = !CBOR_SLOW || types_boolean_stash 422 HV *boolean_stash = !CBOR_SLOW || types_boolean_stash
265 ? types_boolean_stash 423 ? types_boolean_stash
270 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 428 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
271 ? cbor_tagged_stash 429 ? cbor_tagged_stash
272 : gv_stashpv ("CBOR::XS::Tagged" , 1); 430 : gv_stashpv ("CBOR::XS::Tagged" , 1);
273 431
274 HV *stash = SvSTASH (sv); 432 HV *stash = SvSTASH (sv);
275 GV *method;
276 433
277 if (stash == boolean_stash) 434 if (stash == boolean_stash)
278 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 }
279 else if (stash == error_stash) 439 else if (stash == error_stash)
280 encode_ch (enc, 0xe0 | 23); 440 {
441 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
442 return;
443 }
281 else if (stash == tagged_stash) 444 else if (stash == tagged_stash)
282 { 445 {
283 if (svt != SVt_PVAV) 446 if (svt != SVt_PVAV)
284 croak ("encountered CBOR::XS::Tagged object that isn't an array"); 447 croak ("encountered CBOR::XS::Tagged object that isn't an array");
285 448
286 encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); 449 encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
287 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))
288 } 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))));
289 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) 486 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
290 { 487 {
291 dSP; 488 dSP;
292 489
293 ENTER; SAVETMPS; PUSHMARK (SP); 490 ENTER; SAVETMPS;
491 PUSHMARK (SP);
294 // 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
295 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 493 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
296 494
297 PUTBACK; 495 PUTBACK;
298 // G_SCALAR ensures that return value is 1 496 // G_SCALAR ensures that return value is 1
311 } 509 }
312 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0) 510 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
313 { 511 {
314 dSP; 512 dSP;
315 513
316 ENTER; SAVETMPS; PUSHMARK (SP); 514 ENTER; SAVETMPS;
515 SAVESTACK_POS ();
516 PUSHMARK (SP);
317 EXTEND (SP, 2); 517 EXTEND (SP, 2);
318 // 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
319 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 519 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
320 PUSHs (sv_cbor); 520 PUSHs (sv_cbor);
321 521
325 525
326 // catch this surprisingly common error 526 // catch this surprisingly common error
327 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) 527 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
328 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));
329 529
330 encode_uint (enc, 0xc0, CBOR_TAG_PERL_OBJECT); 530 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
331 encode_uint (enc, 0x80, count + 1); 531 encode_uint (enc, MAJOR_ARRAY, count + 1);
332 encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); 532 encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
333 533
334 while (count) 534 while (count)
335 encode_sv (enc, SP[1 - count--]); 535 encode_sv (enc, SP[1 - count--]);
336 536
337 PUTBACK; 537 PUTBACK;
344 } 544 }
345 else if (svt == SVt_PVHV) 545 else if (svt == SVt_PVHV)
346 encode_hv (enc, (HV *)sv); 546 encode_hv (enc, (HV *)sv);
347 else if (svt == SVt_PVAV) 547 else if (svt == SVt_PVAV)
348 encode_av (enc, (AV *)sv); 548 encode_av (enc, (AV *)sv);
349 else if (svt < SVt_PVAV)
350 {
351 STRLEN len = 0;
352 char *pv = svt ? SvPV (sv, len) : 0;
353
354 if (len == 1 && *pv == '1')
355 encode_ch (enc, 0xe0 | 21);
356 else if (len == 1 && *pv == '0')
357 encode_ch (enc, 0xe0 | 20);
358 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
359 encode_ch (enc, 0xe0 | 23);
360 else
361 croak ("cannot encode reference to scalar '%s' unless the scalar is 0 or 1",
362 SvPV_nolen (sv_2mortal (newRV_inc (sv))));
363 }
364 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
365 encode_ch (enc, 0xe0 | 23);
366 else 549 else
367 croak ("encountered %s, but CBOR can only represent references to arrays or hashes", 550 {
368 SvPV_nolen (sv_2mortal (newRV_inc (sv)))); 551 encode_tag (enc, CBOR_TAG_INDIRECTION);
552 encode_sv (enc, sv);
553 }
369} 554}
370 555
371static void 556static void
372encode_nv (enc_t *enc, SV *sv) 557encode_nv (enc_t *enc, SV *sv)
373{ 558{
374 double nv = SvNVX (sv); 559 double nv = SvNVX (sv);
375 560
376 need (enc, 9); 561 need (enc, 9);
377 562
378 if (ecb_expect_false (nv == (U32)nv)) 563 if (ecb_expect_false (nv == (NV)(U32)nv))
379 encode_uint (enc, 0x00, (U32)nv); 564 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
380 //TODO: maybe I32? 565 //TODO: maybe I32?
381 else if (ecb_expect_false (nv == (float)nv)) 566 else if (ecb_expect_false (nv == (float)nv))
382 { 567 {
568 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
569
383 uint32_t fp = ecb_float_to_binary32 (nv); 570 uint32_t fp = ecb_float_to_binary32 (nv);
384
385 *enc->cur++ = 0xe0 | 26;
386 571
387 if (!ecb_big_endian ()) 572 if (!ecb_big_endian ())
388 fp = ecb_bswap32 (fp); 573 fp = ecb_bswap32 (fp);
389 574
390 memcpy (enc->cur, &fp, 4); 575 memcpy (enc->cur, &fp, 4);
391 enc->cur += 4; 576 enc->cur += 4;
392 } 577 }
393 else 578 else
394 { 579 {
580 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
581
395 uint64_t fp = ecb_double_to_binary64 (nv); 582 uint64_t fp = ecb_double_to_binary64 (nv);
396
397 *enc->cur++ = 0xe0 | 27;
398 583
399 if (!ecb_big_endian ()) 584 if (!ecb_big_endian ())
400 fp = ecb_bswap64 (fp); 585 fp = ecb_bswap64 (fp);
401 586
402 memcpy (enc->cur, &fp, 8); 587 memcpy (enc->cur, &fp, 8);
411 596
412 if (SvPOKp (sv)) 597 if (SvPOKp (sv))
413 { 598 {
414 STRLEN len; 599 STRLEN len;
415 char *str = SvPV (sv, len); 600 char *str = SvPV (sv, len);
416 encode_str (enc, SvUTF8 (sv), str, len); 601 encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len);
417 } 602 }
418 else if (SvNOKp (sv)) 603 else if (SvNOKp (sv))
419 encode_nv (enc, sv); 604 encode_nv (enc, sv);
420 else if (SvIOKp (sv)) 605 else if (SvIOKp (sv))
421 { 606 {
422 if (SvIsUV (sv)) 607 if (SvIsUV (sv))
423 encode_uint (enc, 0x00, SvUVX (sv)); 608 encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
424 else if (SvIVX (sv) >= 0) 609 else if (SvIVX (sv) >= 0)
425 encode_uint (enc, 0x00, SvIVX (sv)); 610 encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
426 else 611 else
427 encode_uint (enc, 0x20, -(SvIVX (sv) + 1)); 612 encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
428 } 613 }
429 else if (SvROK (sv)) 614 else if (SvROK (sv))
430 encode_rv (enc, SvRV (sv)); 615 encode_rv (enc, SvRV (sv));
431 else if (!SvOK (sv)) 616 else if (!SvOK (sv))
432 encode_ch (enc, 0xe0 | 22); 617 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
433 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 618 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
434 encode_ch (enc, 0xe0 | 23); 619 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
435 else 620 else
436 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",
437 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 622 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
438} 623}
439 624
440static SV * 625static SV *
441encode_cbor (SV *scalar, CBOR *cbor) 626encode_cbor (SV *scalar, CBOR *cbor)
442{ 627{
443 enc_t enc; 628 enc_t enc = { 0 };
444 629
445 enc.cbor = *cbor; 630 enc.cbor = *cbor;
446 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 631 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
447 enc.cur = SvPVX (enc.sv); 632 enc.cur = SvPVX (enc.sv);
448 enc.end = SvEND (enc.sv); 633 enc.end = SvEND (enc.sv);
449 enc.depth = 0;
450 634
451 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
452 encode_sv (&enc, scalar); 644 encode_sv (&enc, scalar);
453 645
454 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv)); 646 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv));
455 *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
456 648
470 U8 *end; // end of input string 662 U8 *end; // end of input string
471 const char *err; // parse error, if != 0 663 const char *err; // parse error, if != 0
472 CBOR cbor; 664 CBOR cbor;
473 U32 depth; // recursion depth 665 U32 depth; // recursion depth
474 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
475} dec_t; 671} dec_t;
476 672
477#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);
478 680
479#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;
480 684
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
481#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)
482#define DEC_DEC_DEPTH --dec->depth 710#define DEC_DEC_DEPTH --dec->depth
483 711
484static UV 712static UV
485decode_uint (dec_t *dec) 713decode_uint (dec_t *dec)
486{ 714{
487 switch (*dec->cur & 31) 715 U8 m = *dec->cur & MINOR_MASK;
488 { 716 ++dec->cur;
489 case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
490 case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
491 case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
492 return *dec->cur++ & 31;
493 717
494 case 24: 718 if (ecb_expect_true (m < LENGTH_EXT1))
719 return m;
720 else if (ecb_expect_true (m == LENGTH_EXT1))
721 {
495 WANT (2); 722 WANT (1);
496 dec->cur += 2; 723 dec->cur += 1;
497 return dec->cur[-1]; 724 return dec->cur[-1];
498 725 }
499 case 25: 726 else if (ecb_expect_true (m == LENGTH_EXT2))
727 {
500 WANT (3); 728 WANT (2);
501 dec->cur += 3; 729 dec->cur += 2;
502 return (((UV)dec->cur[-2]) << 8) 730 return (((UV)dec->cur[-2]) << 8)
503 | ((UV)dec->cur[-1]); 731 | ((UV)dec->cur[-1]);
504 732 }
505 case 26: 733 else if (ecb_expect_true (m == LENGTH_EXT4))
734 {
506 WANT (5); 735 WANT (4);
507 dec->cur += 5; 736 dec->cur += 4;
508 return (((UV)dec->cur[-4]) << 24) 737 return (((UV)dec->cur[-4]) << 24)
509 | (((UV)dec->cur[-3]) << 16) 738 | (((UV)dec->cur[-3]) << 16)
510 | (((UV)dec->cur[-2]) << 8) 739 | (((UV)dec->cur[-2]) << 8)
511 | ((UV)dec->cur[-1]); 740 | ((UV)dec->cur[-1]);
512 741 }
513 case 27: 742 else if (ecb_expect_true (m == LENGTH_EXT8))
743 {
514 WANT (9); 744 WANT (8);
515 dec->cur += 9; 745 dec->cur += 8;
746
747 return
748#if UVSIZE < 8
749 0
750#else
516 return (((UV)dec->cur[-8]) << 56) 751 (((UV)dec->cur[-8]) << 56)
517 | (((UV)dec->cur[-7]) << 48) 752 | (((UV)dec->cur[-7]) << 48)
518 | (((UV)dec->cur[-6]) << 40) 753 | (((UV)dec->cur[-6]) << 40)
519 | (((UV)dec->cur[-5]) << 32) 754 | (((UV)dec->cur[-5]) << 32)
755#endif
520 | (((UV)dec->cur[-4]) << 24) 756 | (((UV)dec->cur[-4]) << 24)
521 | (((UV)dec->cur[-3]) << 16) 757 | (((UV)dec->cur[-3]) << 16)
522 | (((UV)dec->cur[-2]) << 8) 758 | (((UV)dec->cur[-2]) << 8)
523 | ((UV)dec->cur[-1]); 759 | ((UV)dec->cur[-1]);
524 760 }
525 default: 761 else
526 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 762 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
527 }
528 763
529fail: 764fail:
530 return 0; 765 return 0;
531} 766}
532 767
537{ 772{
538 AV *av = newAV (); 773 AV *av = newAV ();
539 774
540 DEC_INC_DEPTH; 775 DEC_INC_DEPTH;
541 776
542 if ((*dec->cur & 31) == 31) 777 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
543 { 778 {
544 ++dec->cur; 779 ++dec->cur;
545 780
546 for (;;) 781 for (;;)
547 { 782 {
548 WANT (1); 783 WANT (1);
549 784
550 if (*dec->cur == (0xe0 | 31)) 785 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
551 { 786 {
552 ++dec->cur; 787 ++dec->cur;
553 break; 788 break;
554 } 789 }
555 790
556 av_push (av, decode_sv (dec)); 791 av_push (av, decode_sv (dec));
557 } 792 }
558 } 793 }
559 else 794 else
560 { 795 {
561 int i, len = decode_uint (dec); 796 UV i, len = decode_uint (dec);
562 797
798 WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
563 av_fill (av, len - 1); 799 av_fill (av, len - 1);
564 800
565 for (i = 0; i < len; ++i) 801 for (i = 0; i < len; ++i)
566 AvARRAY (av)[i] = decode_sv (dec); 802 AvARRAY (av)[i] = decode_sv (dec);
567 } 803 }
568 804
569 DEC_DEC_DEPTH; 805 DEC_DEC_DEPTH;
570 return newRV_noinc ((SV *)av); 806 return newRV_noinc ((SV *)av);
571 807
572fail: 808fail:
573 SvREFCNT_dec (av); 809 SvREFCNT_dec_NN (av);
574 DEC_DEC_DEPTH; 810 DEC_DEC_DEPTH;
575 return &PL_sv_undef; 811 return &PL_sv_undef;
576} 812}
577 813
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
578static SV * 888static SV *
579decode_hv (dec_t *dec) 889decode_hv (dec_t *dec)
580{ 890{
581 HV *hv = newHV (); 891 HV *hv = newHV ();
582 892
583 DEC_INC_DEPTH; 893 DEC_INC_DEPTH;
584 894
585 if ((*dec->cur & 31) == 31) 895 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
586 { 896 {
587 ++dec->cur; 897 ++dec->cur;
588 898
589 for (;;) 899 for (;;)
590 { 900 {
591 WANT (1); 901 WANT (1);
592 902
593 if (*dec->cur == (0xe0 | 31)) 903 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
594 { 904 {
595 ++dec->cur; 905 ++dec->cur;
596 break; 906 break;
597 } 907 }
598 908
599 SV *k = decode_sv (dec); 909 decode_he (dec, hv);
600 SV *v = decode_sv (dec);
601
602 hv_store_ent (hv, k, v, 0);
603 SvREFCNT_dec (k);
604 } 910 }
605 } 911 }
606 else 912 else
607 { 913 {
608 int len = decode_uint (dec); 914 UV pairs = decode_uint (dec);
609 915
916 WANT (pairs); // complexity check - need at least one byte per value, do not allow supersize hashes
917
610 while (len--) 918 while (pairs--)
611 { 919 decode_he (dec, hv);
612 SV *k = decode_sv (dec);
613 SV *v = decode_sv (dec);
614
615 hv_store_ent (hv, k, v, 0);
616 SvREFCNT_dec (k);
617 }
618 } 920 }
619 921
620 DEC_DEC_DEPTH; 922 DEC_DEC_DEPTH;
621 return newRV_noinc ((SV *)hv); 923 return newRV_noinc ((SV *)hv);
622 924
623fail: 925fail:
624 SvREFCNT_dec (hv); 926 SvREFCNT_dec_NN (hv);
625 DEC_DEC_DEPTH; 927 DEC_DEC_DEPTH;
626 return &PL_sv_undef; 928 return &PL_sv_undef;
627} 929}
628 930
629static SV * 931static SV *
630decode_str (dec_t *dec, int utf8) 932decode_str (dec_t *dec, int utf8)
631{ 933{
632 SV *sv = 0; 934 SV *sv = 0;
633 935
634 if ((*dec->cur & 31) == 31) 936 if (ecb_expect_false ((*dec->cur & MINOR_MASK) == MINOR_INDEF))
635 { 937 {
938 // indefinite length strings
636 ++dec->cur; 939 ++dec->cur;
637 940
941 U8 major = *dec->cur & MAJOR_MISC;
942
638 sv = newSVpvn ("", 0); 943 sv = newSVpvn ("", 0);
639 944
640 // not very fast, and certainly not robust against illegal input
641 for (;;) 945 for (;;)
642 { 946 {
643 WANT (1); 947 WANT (1);
644 948
645 if (*dec->cur == (0xe0 | 31)) 949 if ((*dec->cur - major) > LENGTH_EXT8)
950 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
646 { 951 {
647 ++dec->cur; 952 ++dec->cur;
648 break; 953 break;
649 } 954 }
955 else
956 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
650 957
651 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;
652 } 963 }
653 } 964 }
654 else 965 else
655 { 966 {
656 STRLEN len = decode_uint (dec); 967 STRLEN len = decode_uint (dec);
657 968
658 WANT (len); 969 WANT (len);
659 sv = newSVpvn (dec->cur, len); 970 sv = newSVpvn (dec->cur, len);
660 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));
661 } 976 }
662 977
663 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
664 SvUTF8_on (sv); 984 SvUTF8_on (sv);
985 }
665 986
666 return sv; 987 return sv;
667 988
668fail: 989fail:
669 SvREFCNT_dec (sv); 990 SvREFCNT_dec (sv);
671} 992}
672 993
673static SV * 994static SV *
674decode_tagged (dec_t *dec) 995decode_tagged (dec_t *dec)
675{ 996{
997 SV *sv = 0;
676 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:
677 SV *sv = decode_sv (dec); 1005 sv = decode_sv (dec);
1006 break;
678 1007
679 if (tag == CBOR_TAG_MAGIC) 1008 case CBOR_TAG_INDIRECTION:
680 return sv; 1009 sv = newRV_noinc (decode_sv (dec));
681 else if (tag == CBOR_TAG_PERL_OBJECT) 1010 break;
682 { 1011
1012 case CBOR_TAG_STRINGREF_NAMESPACE:
1013 {
1014 // do not use SAVETMPS/FREETMPS, as these will
1015 // erase mortalised caches, e.g. "shareable"
1016 ENTER;
1017
1018 SAVESPTR (dec->stringref);
1019 dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());
1020
1021 sv = decode_sv (dec);
1022
1023 LEAVE;
1024 }
1025 break;
1026
1027 case CBOR_TAG_STRINGREF:
1028 {
1029 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
1030 ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
1031
1032 UV idx = decode_uint (dec);
1033
1034 if (!dec->stringref || (int)idx > AvFILLp (dec->stringref))
1035 ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)");
1036
1037 sv = newSVsv (AvARRAY (dec->stringref)[idx]);
1038 }
1039 break;
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 || (int)idx > 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
683 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) 1089 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
684 ERR ("corrupted CBOR data (non-array perl object)"); 1090 ERR ("corrupted CBOR data (non-array perl object)");
685 1091
686 AV *av = (AV *)SvRV (sv); 1092 AV *av = (AV *)SvRV (sv);
687 int len = av_len (av) + 1; 1093 int len = av_len (av) + 1;
688 HV *stash = gv_stashsv (*av_fetch (av, 0, 1), 0); 1094 HV *stash = gv_stashsv (*av_fetch (av, 0, 1), 0);
689 1095
690 if (!stash) 1096 if (!stash)
691 ERR ("cannot decode perl-object (package does not exist)"); 1097 ERR ("cannot decode perl-object (package does not exist)");
692 1098
693 GV *method = gv_fetchmethod_autoload (stash, "THAW", 0); 1099 GV *method = gv_fetchmethod_autoload (stash, "THAW", 0);
694 1100
695 if (!method) 1101 if (!method)
696 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)");
697 1103
698 dSP; 1104 dSP;
699 1105
700 ENTER; SAVETMPS; PUSHMARK (SP); 1106 ENTER; SAVETMPS;
1107 PUSHMARK (SP);
701 EXTEND (SP, len + 1); 1108 EXTEND (SP, len + 1);
702 // 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
703 PUSHs (*av_fetch (av, 0, 1)); 1110 PUSHs (*av_fetch (av, 0, 1));
704 PUSHs (sv_cbor); 1111 PUSHs (sv_cbor);
705 1112
706 int i; 1113 int i;
707 1114
708 for (i = 1; i < len; ++i) 1115 for (i = 1; i < len; ++i)
709 PUSHs (*av_fetch (av, i, 1)); 1116 PUSHs (*av_fetch (av, i, 1));
710 1117
711 PUTBACK; 1118 PUTBACK;
712 call_sv ((SV *)GvCV (method), G_SCALAR); 1119 call_sv ((SV *)GvCV (method), G_SCALAR | G_EVAL);
713 SPAGAIN; 1120 SPAGAIN;
714 1121
1122 if (SvTRUE (ERRSV))
1123 {
1124 FREETMPS; LEAVE;
1125 ERR_ERRSV;
1126 }
1127
715 SvREFCNT_dec (sv); 1128 SvREFCNT_dec_NN (sv);
716 sv = SvREFCNT_inc (POPs); 1129 sv = SvREFCNT_inc (POPs);
717 1130
718 PUTBACK; 1131 PUTBACK;
719 1132
720 FREETMPS; LEAVE; 1133 FREETMPS; LEAVE;
1134 }
1135 break;
721 1136
722 return sv; 1137 default:
723 } 1138 filter:
724 else 1139 {
725 { 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 {
726 AV *av = newAV (); 1171 AV *av = newAV ();
727 av_push (av, newSVuv (tag)); 1172 av_push (av, tag_sv);
728 av_push (av, sv); 1173 av_push (av, sv);
729 1174
730 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 1175 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
731 ? cbor_tagged_stash 1176 ? cbor_tagged_stash
732 : gv_stashpv ("CBOR::XS::Tagged" , 1); 1177 : gv_stashpv ("CBOR::XS::Tagged" , 1);
733
734 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;
735 } 1186 }
1187
1188 return sv;
736 1189
737fail: 1190fail:
738 SvREFCNT_dec (sv); 1191 SvREFCNT_dec (sv);
739 return &PL_sv_undef; 1192 return &PL_sv_undef;
740} 1193}
742static SV * 1195static SV *
743decode_sv (dec_t *dec) 1196decode_sv (dec_t *dec)
744{ 1197{
745 WANT (1); 1198 WANT (1);
746 1199
747 switch (*dec->cur >> 5) 1200 switch (*dec->cur >> MAJOR_SHIFT)
748 { 1201 {
749 case 0: // unsigned int 1202 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
750 return newSVuv (decode_uint (dec)); 1203 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
751 case 1: // negative int 1204 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
752 return newSViv (-1 - (IV)decode_uint (dec)); 1205 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
753 case 2: // octet string 1206 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
754 return decode_str (dec, 0); 1207 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
755 case 3: // utf-8 string 1208 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
756 return decode_str (dec, 1); 1209
757 case 4: // array 1210 case MAJOR_MISC >> MAJOR_SHIFT:
758 return decode_av (dec);
759 case 5: // map
760 return decode_hv (dec);
761 case 6: // tag
762 return decode_tagged (dec);
763 case 7: // misc
764 switch (*dec->cur++ & 31) 1211 switch (*dec->cur++ & MINOR_MASK)
765 { 1212 {
766 case 20: 1213 case SIMPLE_FALSE:
767#if CBOR_SLOW 1214#if CBOR_SLOW
768 types_false = get_bool ("Types::Serialiser::false"); 1215 types_false = get_bool ("Types::Serialiser::false");
769#endif 1216#endif
770 return newSVsv (types_false); 1217 return newSVsv (types_false);
771 case 21: 1218 case SIMPLE_TRUE:
772#if CBOR_SLOW 1219#if CBOR_SLOW
773 types_true = get_bool ("Types::Serialiser::true"); 1220 types_true = get_bool ("Types::Serialiser::true");
774#endif 1221#endif
775 return newSVsv (types_true); 1222 return newSVsv (types_true);
776 case 22: 1223 case SIMPLE_NULL:
777 return newSVsv (&PL_sv_undef); 1224 return newSVsv (&PL_sv_undef);
778 case 23: 1225 case SIMPLE_UNDEF:
779#if CBOR_SLOW 1226#if CBOR_SLOW
780 types_error = get_bool ("Types::Serialiser::error"); 1227 types_error = get_bool ("Types::Serialiser::error");
781#endif 1228#endif
782 return newSVsv (types_error); 1229 return newSVsv (types_error);
783 1230
784 case 25: 1231 case MISC_FLOAT16:
785 { 1232 {
786 WANT (2); 1233 WANT (2);
787 1234
788 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1235 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
789 dec->cur += 2; 1236 dec->cur += 2;
790 1237
791 return newSVnv (ecb_binary16_to_float (fp)); 1238 return newSVnv (ecb_binary16_to_float (fp));
792 } 1239 }
793 1240
794 case 26: 1241 case MISC_FLOAT32:
795 { 1242 {
796 uint32_t fp; 1243 uint32_t fp;
797 WANT (4); 1244 WANT (4);
798 memcpy (&fp, dec->cur, 4); 1245 memcpy (&fp, dec->cur, 4);
799 dec->cur += 4; 1246 dec->cur += 4;
802 fp = ecb_bswap32 (fp); 1249 fp = ecb_bswap32 (fp);
803 1250
804 return newSVnv (ecb_binary32_to_float (fp)); 1251 return newSVnv (ecb_binary32_to_float (fp));
805 } 1252 }
806 1253
807 case 27: 1254 case MISC_FLOAT64:
808 { 1255 {
809 uint64_t fp; 1256 uint64_t fp;
810 WANT (8); 1257 WANT (8);
811 memcpy (&fp, dec->cur, 8); 1258 memcpy (&fp, dec->cur, 8);
812 dec->cur += 8; 1259 dec->cur += 8;
815 fp = ecb_bswap64 (fp); 1262 fp = ecb_bswap64 (fp);
816 1263
817 return newSVnv (ecb_binary64_to_double (fp)); 1264 return newSVnv (ecb_binary64_to_double (fp));
818 } 1265 }
819 1266
820 // 0..19 unassigned 1267 // 0..19 unassigned simple
821 // 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
822 default: 1271 default:
823 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1272 ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
824 } 1273 }
825 1274
826 break; 1275 break;
827 } 1276 }
828 1277
831} 1280}
832 1281
833static SV * 1282static SV *
834decode_cbor (SV *string, CBOR *cbor, char **offset_return) 1283decode_cbor (SV *string, CBOR *cbor, char **offset_return)
835{ 1284{
836 dec_t dec; 1285 dec_t dec = { 0 };
837 SV *sv; 1286 SV *sv;
1287 STRLEN len;
1288 char *data = SvPVbyte (string, len);
838 1289
839 /* work around bugs in 5.10 where manipulating magic values
840 * makes perl ignore the magic in subsequent accesses.
841 * also make a copy of non-PV values, to get them into a clean
842 * state (SvPV should do that, but it's buggy, see below).
843 */
844 /*SvGETMAGIC (string);*/
845 if (SvMAGICAL (string) || !SvPOK (string))
846 string = sv_2mortal (newSVsv (string));
847
848 SvUPGRADE (string, SVt_PV);
849
850 /* work around a bug in perl 5.10, which causes SvCUR to fail an
851 * assertion with -DDEBUGGING, although SvCUR is documented to
852 * return the xpv_cur field which certainly exists after upgrading.
853 * according to nicholas clark, calling SvPOK fixes this.
854 * But it doesn't fix it, so try another workaround, call SvPV_nolen
855 * and hope for the best.
856 * Damnit, SvPV_nolen still trips over yet another assertion. This
857 * assertion business is seriously broken, try yet another workaround
858 * for the broken -DDEBUGGING.
859 */
860 {
861#ifdef DEBUGGING
862 STRLEN offset = SvOK (string) ? sv_len (string) : 0;
863#else
864 STRLEN offset = SvCUR (string);
865#endif
866
867 if (offset > cbor->max_size && cbor->max_size) 1290 if (len > cbor->max_size && cbor->max_size)
868 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",
869 (unsigned long)SvCUR (string), (unsigned long)cbor->max_size); 1292 (unsigned long)len, (unsigned long)cbor->max_size);
870 }
871
872 sv_utf8_downgrade (string, 0);
873 1293
874 dec.cbor = *cbor; 1294 dec.cbor = *cbor;
875 dec.cur = (U8 *)SvPVX (string); 1295 dec.cur = (U8 *)data;
876 dec.end = (U8 *)SvEND (string); 1296 dec.end = (U8 *)data + len;
877 dec.err = 0;
878 dec.depth = 0;
879 1297
880 sv = decode_sv (&dec); 1298 sv = decode_sv (&dec);
881 1299
882 if (offset_return) 1300 if (offset_return)
883 *offset_return = dec.cur; 1301 *offset_return = dec.cur;
886 if (dec.cur != dec.end && !dec.err) 1304 if (dec.cur != dec.end && !dec.err)
887 dec.err = "garbage after CBOR object"; 1305 dec.err = "garbage after CBOR object";
888 1306
889 if (dec.err) 1307 if (dec.err)
890 { 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
891 SvREFCNT_dec (sv); 1320 SvREFCNT_dec_NN (sv);
1321
1322 if (dec.err_sv)
1323 sv_2mortal (dec.err_sv);
1324
892 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);
893 } 1326 }
894 1327
895 sv = sv_2mortal (sv); 1328 sv = sv_2mortal (sv);
896 1329
897 return sv; 1330 return sv;
898} 1331}
899 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
900///////////////////////////////////////////////////////////////////////////// 1457/////////////////////////////////////////////////////////////////////////////
901// XS interface functions 1458// XS interface functions
902 1459
903MODULE = CBOR::XS PACKAGE = CBOR::XS 1460MODULE = CBOR::XS PACKAGE = CBOR::XS
904 1461
912 1469
913 types_true = get_bool ("Types::Serialiser::true" ); 1470 types_true = get_bool ("Types::Serialiser::true" );
914 types_false = get_bool ("Types::Serialiser::false"); 1471 types_false = get_bool ("Types::Serialiser::false");
915 types_error = get_bool ("Types::Serialiser::error"); 1472 types_error = get_bool ("Types::Serialiser::error");
916 1473
1474 default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1475
917 sv_cbor = newSVpv ("CBOR", 0); 1476 sv_cbor = newSVpv ("CBOR", 0);
918 SvREADONLY_on (sv_cbor); 1477 SvREADONLY_on (sv_cbor);
1478
1479 assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1));
919} 1480}
920 1481
921PROTOTYPES: DISABLE 1482PROTOTYPES: DISABLE
922 1483
923void CLONE (...) 1484void CLONE (...)
941 1502
942void shrink (CBOR *self, int enable = 1) 1503void shrink (CBOR *self, int enable = 1)
943 ALIAS: 1504 ALIAS:
944 shrink = F_SHRINK 1505 shrink = F_SHRINK
945 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
946 PPCODE: 1514 PPCODE:
947{ 1515{
948 if (enable) 1516 if (enable)
949 self->flags |= ix; 1517 self->flags |= ix;
950 else 1518 else
955 1523
956void get_shrink (CBOR *self) 1524void get_shrink (CBOR *self)
957 ALIAS: 1525 ALIAS:
958 get_shrink = F_SHRINK 1526 get_shrink = F_SHRINK
959 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
960 PPCODE: 1535 PPCODE:
961 XPUSHs (boolSV (self->flags & ix)); 1536 XPUSHs (boolSV (self->flags & ix));
962 1537
963void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1538void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
964 PPCODE: 1539 PPCODE:
980 CODE: 1555 CODE:
981 RETVAL = self->max_size; 1556 RETVAL = self->max_size;
982 OUTPUT: 1557 OUTPUT:
983 RETVAL 1558 RETVAL
984 1559
985#if 0 //TODO 1560void filter (CBOR *self, SV *filter = 0)
986
987void filter_cbor_object (CBOR *self, SV *cb = &PL_sv_undef)
988 PPCODE: 1561 PPCODE:
989{
990 SvREFCNT_dec (self->cb_object); 1562 SvREFCNT_dec (self->filter);
991 self->cb_object = SvOK (cb) ? newSVsv (cb) : 0; 1563 self->filter = filter ? newSVsv (filter) : filter;
992
993 XPUSHs (ST (0)); 1564 XPUSHs (ST (0));
994}
995 1565
996void filter_cbor_single_key_object (CBOR *self, SV *key, SV *cb = &PL_sv_undef) 1566SV *get_filter (CBOR *self)
997 PPCODE: 1567 CODE:
998{ 1568 RETVAL = self->filter ? self->filter : NEWSV (0, 0);
999 if (!self->cb_sk_object) 1569 OUTPUT:
1000 self->cb_sk_object = newHV (); 1570 RETVAL
1001
1002 if (SvOK (cb))
1003 hv_store_ent (self->cb_sk_object, key, newSVsv (cb), 0);
1004 else
1005 {
1006 hv_delete_ent (self->cb_sk_object, key, G_DISCARD, 0);
1007
1008 if (!HvKEYS (self->cb_sk_object))
1009 {
1010 SvREFCNT_dec (self->cb_sk_object);
1011 self->cb_sk_object = 0;
1012 }
1013 }
1014
1015 XPUSHs (ST (0));
1016}
1017
1018#endif
1019 1571
1020void encode (CBOR *self, SV *scalar) 1572void encode (CBOR *self, SV *scalar)
1021 PPCODE: 1573 PPCODE:
1022 PUTBACK; scalar = encode_cbor (scalar, self); SPAGAIN; 1574 PUTBACK; scalar = encode_cbor (scalar, self); SPAGAIN;
1023 XPUSHs (scalar); 1575 XPUSHs (scalar);
1036 EXTEND (SP, 2); 1588 EXTEND (SP, 2);
1037 PUSHs (sv); 1589 PUSHs (sv);
1038 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1590 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1039} 1591}
1040 1592
1041#if 0 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}
1042 1644
1043void DESTROY (CBOR *self) 1645void DESTROY (CBOR *self)
1044 CODE: 1646 PPCODE:
1045 SvREFCNT_dec (self->cb_sk_object); 1647 cbor_free (self);
1046 SvREFCNT_dec (self->cb_object);
1047
1048#endif
1049 1648
1050PROTOTYPES: ENABLE 1649PROTOTYPES: ENABLE
1051 1650
1052void encode_cbor (SV *scalar) 1651void encode_cbor (SV *scalar)
1652 ALIAS:
1653 encode_cbor = 0
1654 encode_cbor_sharing = F_ALLOW_SHARING
1053 PPCODE: 1655 PPCODE:
1054{ 1656{
1055 CBOR cbor; 1657 CBOR cbor;
1056 cbor_init (&cbor); 1658 cbor_init (&cbor);
1659 cbor.flags |= ix;
1057 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1660 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1058 XPUSHs (scalar); 1661 XPUSHs (scalar);
1059} 1662}
1060 1663
1061void decode_cbor (SV *cborstr) 1664void decode_cbor (SV *cborstr)

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines