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 1.3 by root, Sat Oct 26 11:08:34 2013 UTC vs.
Revision 1.52 by root, Sun Apr 24 19:31:55 2016 UTC

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 11
12#define ECB_NO_THREADS 1
12#include "ecb.h" 13#include "ecb.h"
13 14
15// compatibility with perl <5.18
16#ifndef HvNAMELEN_get
17# define HvNAMELEN_get(hv) strlen (HvNAME (hv))
18#endif
19#ifndef HvNAMELEN
20# define HvNAMELEN(hv) HvNAMELEN_get (hv)
21#endif
22#ifndef HvNAMEUTF8
23# define HvNAMEUTF8(hv) 0
24#endif
25#ifndef SvREFCNT_dec_NN
26# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
27#endif
28
29// known major and minor types
30enum cbor_type
31{
32 MAJOR_SHIFT = 5,
33 MINOR_MASK = 0x1f,
34
35 MAJOR_POS_INT = 0 << MAJOR_SHIFT,
36 MAJOR_NEG_INT = 1 << MAJOR_SHIFT,
37 MAJOR_BYTES = 2 << MAJOR_SHIFT,
38 MAJOR_TEXT = 3 << MAJOR_SHIFT,
39 MAJOR_ARRAY = 4 << MAJOR_SHIFT,
40 MAJOR_MAP = 5 << MAJOR_SHIFT,
41 MAJOR_TAG = 6 << MAJOR_SHIFT,
42 MAJOR_MISC = 7 << MAJOR_SHIFT,
43
44 // INT/STRING/ARRAY/MAP subtypes
45 LENGTH_EXT1 = 24,
46 LENGTH_EXT2 = 25,
47 LENGTH_EXT4 = 26,
48 LENGTH_EXT8 = 27,
49
50 // SIMPLE types (effectively MISC subtypes)
51 SIMPLE_FALSE = 20,
52 SIMPLE_TRUE = 21,
53 SIMPLE_NULL = 22,
54 SIMPLE_UNDEF = 23,
55
56 // MISC subtype (unused)
57 MISC_EXT1 = 24,
58 MISC_FLOAT16 = 25,
59 MISC_FLOAT32 = 26,
60 MISC_FLOAT64 = 27,
61
62 // BYTES/TEXT/ARRAY/MAP
63 MINOR_INDEF = 31,
64};
65
66// known tags
67enum cbor_tag
68{
69 // extensions
70 CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
71 CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
72 CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
73 CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
74 CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
75 CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref
76 CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
77
78 // rfc7049
79 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
80 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
81 CBOR_TAG_POS_BIGNUM = 2, // byte string
82 CBOR_TAG_NEG_BIGNUM = 3, // byte string
83 CBOR_TAG_DECIMAL = 4, // decimal fraction, array
84 CBOR_TAG_BIGFLOAT = 5, // array
85
86 CBOR_TAG_CONV_B64U = 21, // base64url, any
87 CBOR_TAG_CONV_B64 = 22, // base64, any
88 CBOR_TAG_CONV_HEX = 23, // base16, any
89 CBOR_TAG_CBOR = 24, // embedded cbor, byte string
90
91 CBOR_TAG_URI = 32, // URI rfc3986, utf-8
92 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
93 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
94 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
95 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
96
97 CBOR_TAG_MAGIC = 55799, // self-describe cbor
98};
99
14#define F_SHRINK 0x00000200UL 100#define F_SHRINK 0x00000001UL
15#define F_ALLOW_UNKNOWN 0x00002000UL 101#define F_ALLOW_UNKNOWN 0x00000002UL
102#define F_ALLOW_SHARING 0x00000004UL
103#define F_ALLOW_CYCLES 0x00000008UL
104#define F_PACK_STRINGS 0x00000010UL
105#define F_UTF8_STRINGS 0x00000020UL
106#define F_VALIDATE_UTF8 0x00000040UL
16 107
17#define INIT_SIZE 32 // initial scalar size to be allocated 108#define INIT_SIZE 32 // initial scalar size to be allocated
18 109
19#define SB do { 110#define SB do {
20#define SE } while (0) 111#define SE } while (0)
21
22#if __GNUC__ >= 3
23# define expect(expr,value) __builtin_expect ((expr), (value))
24# define INLINE static inline
25#else
26# define expect(expr,value) (expr)
27# define INLINE static
28#endif
29
30#define expect_false(expr) expect ((expr) != 0, 0)
31#define expect_true(expr) expect ((expr) != 0, 1)
32 112
33#define IN_RANGE_INC(type,val,beg,end) \ 113#define IN_RANGE_INC(type,val,beg,end) \
34 ((unsigned type)((unsigned type)(val) - (unsigned type)(beg)) \ 114 ((unsigned type)((unsigned type)(val) - (unsigned type)(beg)) \
35 <= (unsigned type)((unsigned type)(end) - (unsigned type)(beg))) 115 <= (unsigned type)((unsigned type)(end) - (unsigned type)(beg)))
36 116
42#else 122#else
43# define CBOR_SLOW 0 123# define CBOR_SLOW 0
44# define CBOR_STASH cbor_stash 124# define CBOR_STASH cbor_stash
45#endif 125#endif
46 126
47static HV *cbor_stash, *cbor_boolean_stash; // CBOR::XS:: 127static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS::
48static SV *cbor_true, *cbor_false; 128static SV *types_true, *types_false, *types_error, *sv_cbor, *default_filter;
49 129
50typedef struct { 130typedef struct {
51 U32 flags; 131 U32 flags;
52 U32 max_depth; 132 U32 max_depth;
53 STRLEN max_size; 133 STRLEN max_size;
134 SV *filter;
54 135
55 SV *cb_object; 136 // for the incremental parser
56 HV *cb_sk_object; 137 STRLEN incr_pos; // the current offset into the text
138 STRLEN incr_need; // minimum bytes needed to decode
139 AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef.
57} CBOR; 140} CBOR;
58 141
59INLINE void 142ecb_inline void
60cbor_init (CBOR *cbor) 143cbor_init (CBOR *cbor)
61{ 144{
62 Zero (cbor, 1, CBOR); 145 Zero (cbor, 1, CBOR);
63 cbor->max_depth = 512; 146 cbor->max_depth = 512;
64} 147}
65 148
149ecb_inline void
150cbor_free (CBOR *cbor)
151{
152 SvREFCNT_dec (cbor->filter);
153 SvREFCNT_dec (cbor->incr_count);
154}
155
66///////////////////////////////////////////////////////////////////////////// 156/////////////////////////////////////////////////////////////////////////////
67// utility functions 157// utility functions
68 158
69INLINE SV * 159ecb_inline SV *
70get_bool (const char *name) 160get_bool (const char *name)
71{ 161{
72 SV *sv = get_sv (name, 1); 162 SV *sv = get_sv (name, 1);
73 163
74 SvREADONLY_on (sv); 164 SvREADONLY_on (sv);
75 SvREADONLY_on (SvRV (sv)); 165 SvREADONLY_on (SvRV (sv));
76 166
77 return sv; 167 return sv;
78} 168}
79 169
80INLINE void 170ecb_inline void
81shrink (SV *sv) 171shrink (SV *sv)
82{ 172{
83 sv_utf8_downgrade (sv, 1); 173 sv_utf8_downgrade (sv, 1);
84 174
85 if (SvLEN (sv) > SvCUR (sv) + 1) 175 if (SvLEN (sv) > SvCUR (sv) + 1)
90 SvPV_renew (sv, SvCUR (sv) + 1); 180 SvPV_renew (sv, SvCUR (sv) + 1);
91#endif 181#endif
92 } 182 }
93} 183}
94 184
95///////////////////////////////////////////////////////////////////////////// 185// minimum length of a string to be registered for stringref
96// fp hell 186ecb_inline int
97 187minimum_string_length (UV idx)
98//TODO 188{
189 return idx > 23
190 ? idx > 0xffU
191 ? idx > 0xffffU
192 ? idx > 0xffffffffU
193 ? 11
194 : 7
195 : 5
196 : 4
197 : 3;
198}
99 199
100///////////////////////////////////////////////////////////////////////////// 200/////////////////////////////////////////////////////////////////////////////
101// encoder 201// encoder
102 202
103// structure used for encoding CBOR 203// structure used for encoding CBOR
106 char *cur; // SvPVX (sv) + current output position 206 char *cur; // SvPVX (sv) + current output position
107 char *end; // SvEND (sv) 207 char *end; // SvEND (sv)
108 SV *sv; // result scalar 208 SV *sv; // result scalar
109 CBOR cbor; 209 CBOR cbor;
110 U32 depth; // recursion level 210 U32 depth; // recursion level
211 HV *stringref[2]; // string => index, or 0 ([0] = bytes, [1] = utf-8)
212 UV stringref_idx;
213 HV *shareable; // ptr => index, or 0
214 UV shareable_idx;
111} enc_t; 215} enc_t;
112 216
113INLINE void 217ecb_inline void
114need (enc_t *enc, STRLEN len) 218need (enc_t *enc, STRLEN len)
115{ 219{
116 if (expect_false (enc->cur + len >= enc->end)) 220 if (ecb_expect_false (enc->cur + len >= enc->end))
117 { 221 {
118 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv); 222 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv);
119 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1); 223 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1);
120 enc->cur = SvPVX (enc->sv) + cur; 224 enc->cur = SvPVX (enc->sv) + cur;
121 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1; 225 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;
122 } 226 }
123} 227}
124 228
125INLINE void 229ecb_inline void
126encode_ch (enc_t *enc, char ch) 230encode_ch (enc_t *enc, char ch)
127{ 231{
128 need (enc, 1); 232 need (enc, 1);
129 *enc->cur++ = ch; 233 *enc->cur++ = ch;
130} 234}
132static void 236static void
133encode_uint (enc_t *enc, int major, UV len) 237encode_uint (enc_t *enc, int major, UV len)
134{ 238{
135 need (enc, 9); 239 need (enc, 9);
136 240
137 if (len < 24) 241 if (ecb_expect_true (len < LENGTH_EXT1))
138 *enc->cur++ = major | len; 242 *enc->cur++ = major | len;
139 else if (len < 0x100) 243 else if (ecb_expect_true (len <= 0xffU))
140 { 244 {
141 *enc->cur++ = major | 24; 245 *enc->cur++ = major | LENGTH_EXT1;
142 *enc->cur++ = len; 246 *enc->cur++ = len;
143 } 247 }
144 else if (len < 0x10000) 248 else if (len <= 0xffffU)
145 { 249 {
146 *enc->cur++ = major | 25; 250 *enc->cur++ = major | LENGTH_EXT2;
147 *enc->cur++ = len >> 8; 251 *enc->cur++ = len >> 8;
148 *enc->cur++ = len; 252 *enc->cur++ = len;
149 } 253 }
150 else if (len < 0x100000000) 254 else if (len <= 0xffffffffU)
151 { 255 {
152 *enc->cur++ = major | 26; 256 *enc->cur++ = major | LENGTH_EXT4;
153 *enc->cur++ = len >> 24; 257 *enc->cur++ = len >> 24;
154 *enc->cur++ = len >> 16; 258 *enc->cur++ = len >> 16;
155 *enc->cur++ = len >> 8; 259 *enc->cur++ = len >> 8;
156 *enc->cur++ = len; 260 *enc->cur++ = len;
157 } 261 }
158 else if (len) 262 else
159 { 263 {
160 *enc->cur++ = major | 27; 264 *enc->cur++ = major | LENGTH_EXT8;
161 *enc->cur++ = len >> 56; 265 *enc->cur++ = len >> 56;
162 *enc->cur++ = len >> 48; 266 *enc->cur++ = len >> 48;
163 *enc->cur++ = len >> 40; 267 *enc->cur++ = len >> 40;
164 *enc->cur++ = len >> 32; 268 *enc->cur++ = len >> 32;
165 *enc->cur++ = len >> 24; 269 *enc->cur++ = len >> 24;
167 *enc->cur++ = len >> 8; 271 *enc->cur++ = len >> 8;
168 *enc->cur++ = len; 272 *enc->cur++ = len;
169 } 273 }
170} 274}
171 275
172static void 276ecb_inline void
277encode_tag (enc_t *enc, UV tag)
278{
279 encode_uint (enc, MAJOR_TAG, tag);
280}
281
282ecb_inline void
173encode_str (enc_t *enc, int utf8, char *str, STRLEN len) 283encode_str (enc_t *enc, int utf8, char *str, STRLEN len)
174{ 284{
175 encode_uint (enc, utf8 ? 0x60 : 0x40, len); 285 if (ecb_expect_false (enc->cbor.flags & F_UTF8_STRINGS))
286 if (!utf8)
287 {
288 // exceptional path for bytze strings that need to be utf8-encoded
289 STRLEN ulen = len;
290 U8 *p, *pend = (U8 *)str + len;
291
292 for (p = (U8 *)str; p < pend; ++p)
293 ulen += *p >> 7; // count set high bits
294
295 encode_uint (enc, MAJOR_TEXT, ulen);
296
297 need (enc, ulen);
298 for (p = (U8 *)str; p < pend; ++p)
299 if (*p < 0x80)
300 *enc->cur++ = *p;
301 else
302 {
303 *enc->cur++ = 0xc0 + (*p >> 6);
304 *enc->cur++ = 0x80 + (*p & 63);
305 }
306
307 return;
308 }
309
310 encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
176 need (enc, len); 311 need (enc, len);
177 memcpy (enc->cur, str, len); 312 memcpy (enc->cur, str, len);
178 enc->cur += len; 313 enc->cur += len;
179} 314}
180 315
316static void
317encode_strref (enc_t *enc, int utf8, char *str, STRLEN len)
318{
319 if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
320 {
321 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
322
323 if (SvOK (*svp))
324 {
325 // already registered, use stringref
326 encode_tag (enc, CBOR_TAG_STRINGREF);
327 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
328 return;
329 }
330 else if (len >= minimum_string_length (enc->stringref_idx))
331 {
332 // register only
333 sv_setuv (*svp, enc->stringref_idx);
334 ++enc->stringref_idx;
335 }
336 }
337
338 encode_str (enc, utf8, str, len);
339}
340
181static void encode_sv (enc_t *enc, SV *sv); 341static void encode_sv (enc_t *enc, SV *sv);
182 342
183static void 343static void
184encode_av (enc_t *enc, AV *av) 344encode_av (enc_t *enc, AV *av)
185{ 345{
188 if (enc->depth >= enc->cbor.max_depth) 348 if (enc->depth >= enc->cbor.max_depth)
189 croak (ERR_NESTING_EXCEEDED); 349 croak (ERR_NESTING_EXCEEDED);
190 350
191 ++enc->depth; 351 ++enc->depth;
192 352
193 encode_uint (enc, 0x80, len + 1); 353 encode_uint (enc, MAJOR_ARRAY, len + 1);
194 354
355 if (SvMAGICAL (av))
195 for (i = 0; i <= len; ++i) 356 for (i = 0; i <= len; ++i)
196 { 357 {
197 SV **svp = av_fetch (av, i, 0); 358 SV **svp = av_fetch (av, i, 0);
198 encode_sv (enc, svp ? *svp : &PL_sv_undef); 359 encode_sv (enc, svp ? *svp : &PL_sv_undef);
199 } 360 }
361 else
362 for (i = 0; i <= len; ++i)
363 {
364 SV *sv = AvARRAY (av)[i];
365 encode_sv (enc, sv ? sv : &PL_sv_undef);
366 }
200 367
201 --enc->depth; 368 --enc->depth;
202} 369}
203 370
204static void 371static void
213 380
214 int pairs = hv_iterinit (hv); 381 int pairs = hv_iterinit (hv);
215 int mg = SvMAGICAL (hv); 382 int mg = SvMAGICAL (hv);
216 383
217 if (mg) 384 if (mg)
218 encode_ch (enc, 0xa0 | 31); 385 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
219 else 386 else
220 encode_uint (enc, 0xa0, pairs); 387 encode_uint (enc, MAJOR_MAP, pairs);
221 388
222 while ((he = hv_iternext (hv))) 389 while ((he = hv_iternext (hv)))
223 { 390 {
224 if (HeKLEN (he) == HEf_SVKEY) 391 if (HeKLEN (he) == HEf_SVKEY)
225 encode_sv (enc, HeSVKEY (he)); 392 encode_sv (enc, HeSVKEY (he));
226 else 393 else
227 encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); 394 encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
228 395
229 encode_sv (enc, expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); 396 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
230 } 397 }
231 398
232 if (mg) 399 if (mg)
233 encode_ch (enc, 0xe0 | 31); 400 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
234 401
235 --enc->depth; 402 --enc->depth;
236} 403}
237 404
238// encode objects, arrays and special \0=false and \1=true values. 405// encode objects, arrays and special \0=false and \1=true values.
239static void 406static void
240encode_rv (enc_t *enc, SV *sv) 407encode_rv (enc_t *enc, SV *sv)
241{ 408{
242 svtype svt;
243
244 SvGETMAGIC (sv); 409 SvGETMAGIC (sv);
410
245 svt = SvTYPE (sv); 411 svtype svt = SvTYPE (sv);
246 412
247 if (expect_false (SvOBJECT (sv))) 413 if (ecb_expect_false (SvOBJECT (sv)))
248 { 414 {
249 HV *stash = !CBOR_SLOW || cbor_boolean_stash 415 HV *boolean_stash = !CBOR_SLOW || types_boolean_stash
250 ? cbor_boolean_stash 416 ? types_boolean_stash
417 : gv_stashpv ("Types::Serialiser::Boolean", 1);
418 HV *error_stash = !CBOR_SLOW || types_error_stash
419 ? types_error_stash
420 : gv_stashpv ("Types::Serialiser::Error", 1);
421 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
422 ? cbor_tagged_stash
251 : gv_stashpv ("CBOR::XS::Boolean", 1); 423 : gv_stashpv ("CBOR::XS::Tagged" , 1);
252 424
253 if (SvSTASH (sv) == stash) 425 HV *stash = SvSTASH (sv);
254 encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20); 426
427 if (stash == boolean_stash)
428 {
429 encode_ch (enc, SvIV (sv) ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE);
430 return;
431 }
432 else if (stash == error_stash)
433 {
434 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
435 return;
436 }
437 else if (stash == tagged_stash)
438 {
439 if (svt != SVt_PVAV)
440 croak ("encountered CBOR::XS::Tagged object that isn't an array");
441
442 encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
443 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1));
444
445 return;
446 }
447 }
448
449 if (ecb_expect_false (SvREFCNT (sv) > 1)
450 && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
451 {
452 if (!enc->shareable)
453 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
454
455 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
456
457 if (SvOK (*svp))
458 {
459 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
460 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
461 return;
462 }
255 else 463 else
256 { 464 {
257#if 0 //TODO 465 sv_setuv (*svp, enc->shareable_idx);
258 if (enc->cbor.flags & F_CONV_BLESSED) 466 ++enc->shareable_idx;
467 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
468 }
469 }
470
471 if (ecb_expect_false (SvOBJECT (sv)))
472 {
473 HV *stash = SvSTASH (sv);
474 GV *method;
475
476 if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
259 { 477 {
478 dSP;
479
480 ENTER; SAVETMPS;
481 PUSHMARK (SP);
260 // we re-bless the reference to get overload and other niceties right 482 // we re-bless the reference to get overload and other niceties right
261 GV *to_cbor = gv_fetchmethod_autoload (SvSTASH (sv), "TO_CBOR", 0);
262
263 if (to_cbor)
264 {
265 dSP;
266
267 ENTER; SAVETMPS; PUSHMARK (SP);
268 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), SvSTASH (sv))); 483 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
269 484
270 // calling with G_SCALAR ensures that we always get a 1 return value
271 PUTBACK; 485 PUTBACK;
486 // G_SCALAR ensures that return value is 1
272 call_sv ((SV *)GvCV (to_cbor), G_SCALAR); 487 call_sv ((SV *)GvCV (method), G_SCALAR);
273 SPAGAIN; 488 SPAGAIN;
274 489
275 // catch this surprisingly common error 490 // catch this surprisingly common error
276 if (SvROK (TOPs) && SvRV (TOPs) == sv) 491 if (SvROK (TOPs) && SvRV (TOPs) == sv)
277 croak ("%s::TO_CBOR method returned same object as was passed instead of a new one", HvNAME (SvSTASH (sv))); 492 croak ("%s::TO_CBOR method returned same object as was passed instead of a new one", HvNAME (stash));
278 493
279 sv = POPs;
280 PUTBACK;
281
282 encode_sv (enc, sv); 494 encode_sv (enc, POPs);
283 495
496 PUTBACK;
497
284 FREETMPS; LEAVE; 498 FREETMPS; LEAVE;
285 }
286 else if (enc->cbor.flags & F_ALLOW_BLESSED)
287 encode_str (enc, "null", 4, 0);
288 else
289 croak ("encountered object '%s', but neither allow_blessed enabled nor TO_CBOR method available on it",
290 SvPV_nolen (sv_2mortal (newRV_inc (sv))));
291 } 499 }
292 else if (enc->cbor.flags & F_ALLOW_BLESSED) 500 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
293 encode_str (enc, "null", 4, 0); 501 {
502 dSP;
503
504 ENTER; SAVETMPS;
505 SAVESTACK_POS ();
506 PUSHMARK (SP);
507 EXTEND (SP, 2);
508 // we re-bless the reference to get overload and other niceties right
509 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
510 PUSHs (sv_cbor);
511
512 PUTBACK;
513 int count = call_sv ((SV *)GvCV (method), G_ARRAY);
514 SPAGAIN;
515
516 // catch this surprisingly common error
517 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
518 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
519
520 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
521 encode_uint (enc, MAJOR_ARRAY, count + 1);
522 encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
523
524 while (count)
525 encode_sv (enc, SP[1 - count--]);
526
527 PUTBACK;
528
529 FREETMPS; LEAVE;
530 }
294 else 531 else
295 croak ("encountered object '%s', but neither allow_blessed nor convert_blessed settings are enabled", 532 croak ("encountered object '%s', but no TO_CBOR or FREEZE methods available on it",
296 SvPV_nolen (sv_2mortal (newRV_inc (sv)))); 533 SvPV_nolen (sv_2mortal (newRV_inc (sv))));
297#endif
298 }
299 } 534 }
300 else if (svt == SVt_PVHV) 535 else if (svt == SVt_PVHV)
301 encode_hv (enc, (HV *)sv); 536 encode_hv (enc, (HV *)sv);
302 else if (svt == SVt_PVAV) 537 else if (svt == SVt_PVAV)
303 encode_av (enc, (AV *)sv); 538 encode_av (enc, (AV *)sv);
304 else if (svt < SVt_PVAV)
305 {
306 STRLEN len = 0;
307 char *pv = svt ? SvPV (sv, len) : 0;
308
309 if (len == 1 && *pv == '1')
310 encode_ch (enc, 0xe0 | 21);
311 else if (len == 1 && *pv == '0')
312 encode_ch (enc, 0xe0 | 20);
313 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
314 encode_ch (enc, 0xe0 | 23);
315 else
316 croak ("cannot encode reference to scalar '%s' unless the scalar is 0 or 1",
317 SvPV_nolen (sv_2mortal (newRV_inc (sv))));
318 }
319 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
320 encode_ch (enc, 0xe0 | 23);
321 else 539 else
322 croak ("encountered %s, but CBOR can only represent references to arrays or hashes", 540 {
323 SvPV_nolen (sv_2mortal (newRV_inc (sv)))); 541 encode_tag (enc, CBOR_TAG_INDIRECTION);
542 encode_sv (enc, sv);
543 }
324} 544}
325 545
326static void 546static void
327encode_nv (enc_t *enc, SV *sv) 547encode_nv (enc_t *enc, SV *sv)
328{ 548{
329 double nv = SvNVX (sv); 549 double nv = SvNVX (sv);
330 550
331 need (enc, 9); 551 need (enc, 9);
332 552
333 if (expect_false (nv == (U32)nv)) 553 if (ecb_expect_false (nv == (NV)(U32)nv))
334 encode_uint (enc, 0x00, (U32)nv); 554 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
335 //TODO: maybe I32? 555 //TODO: maybe I32?
336 else if (expect_false (nv == (float)nv)) 556 else if (ecb_expect_false (nv == (float)nv))
337 { 557 {
338 uint32_t fp = ecb_float_to_binary32 (nv); 558 uint32_t fp = ecb_float_to_binary32 (nv);
339 559
340 *enc->cur++ = 0xe0 | 26; 560 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
341 561
342 if (!ecb_big_endian ()) 562 if (!ecb_big_endian ())
343 fp = ecb_bswap32 (fp); 563 fp = ecb_bswap32 (fp);
344 564
345 memcpy (enc->cur, &fp, 4); 565 memcpy (enc->cur, &fp, 4);
347 } 567 }
348 else 568 else
349 { 569 {
350 uint64_t fp = ecb_double_to_binary64 (nv); 570 uint64_t fp = ecb_double_to_binary64 (nv);
351 571
352 *enc->cur++ = 0xe0 | 27; 572 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
353 573
354 if (!ecb_big_endian ()) 574 if (!ecb_big_endian ())
355 fp = ecb_bswap64 (fp); 575 fp = ecb_bswap64 (fp);
356 576
357 memcpy (enc->cur, &fp, 8); 577 memcpy (enc->cur, &fp, 8);
366 586
367 if (SvPOKp (sv)) 587 if (SvPOKp (sv))
368 { 588 {
369 STRLEN len; 589 STRLEN len;
370 char *str = SvPV (sv, len); 590 char *str = SvPV (sv, len);
371 encode_str (enc, SvUTF8 (sv), str, len); 591 encode_strref (enc, SvUTF8 (sv), str, len);
372 } 592 }
373 else if (SvNOKp (sv)) 593 else if (SvNOKp (sv))
374 encode_nv (enc, sv); 594 encode_nv (enc, sv);
375 else if (SvIOKp (sv)) 595 else if (SvIOKp (sv))
376 { 596 {
377 if (SvIsUV (sv)) 597 if (SvIsUV (sv))
378 encode_uint (enc, 0x00, SvUVX (sv)); 598 encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
379 else if (SvIVX (sv) >= 0) 599 else if (SvIVX (sv) >= 0)
380 encode_uint (enc, 0x00, SvIVX (sv)); 600 encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
381 else 601 else
382 encode_uint (enc, 0x20, -(SvIVX (sv) + 1)); 602 encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
383 } 603 }
384 else if (SvROK (sv)) 604 else if (SvROK (sv))
385 encode_rv (enc, SvRV (sv)); 605 encode_rv (enc, SvRV (sv));
386 else if (!SvOK (sv)) 606 else if (!SvOK (sv))
387 encode_ch (enc, 0xe0 | 22); 607 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
388 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 608 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
389 encode_ch (enc, 0xe0 | 23); 609 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
390 else 610 else
391 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data", 611 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
392 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 612 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
393} 613}
394 614
395static SV * 615static SV *
396encode_cbor (SV *scalar, CBOR *cbor) 616encode_cbor (SV *scalar, CBOR *cbor)
397{ 617{
398 enc_t enc; 618 enc_t enc = { 0 };
399 619
400 enc.cbor = *cbor; 620 enc.cbor = *cbor;
401 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 621 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
402 enc.cur = SvPVX (enc.sv); 622 enc.cur = SvPVX (enc.sv);
403 enc.end = SvEND (enc.sv); 623 enc.end = SvEND (enc.sv);
404 enc.depth = 0;
405 624
406 SvPOK_only (enc.sv); 625 SvPOK_only (enc.sv);
626
627 if (cbor->flags & F_PACK_STRINGS)
628 {
629 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
630 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
631 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
632 }
633
407 encode_sv (&enc, scalar); 634 encode_sv (&enc, scalar);
408 635
409 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv)); 636 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv));
410 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings 637 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings
411 638
425 U8 *end; // end of input string 652 U8 *end; // end of input string
426 const char *err; // parse error, if != 0 653 const char *err; // parse error, if != 0
427 CBOR cbor; 654 CBOR cbor;
428 U32 depth; // recursion depth 655 U32 depth; // recursion depth
429 U32 maxdepth; // recursion depth limit 656 U32 maxdepth; // recursion depth limit
657 AV *shareable;
658 AV *stringref;
659 SV *decode_tagged;
430} dec_t; 660} dec_t;
431 661
432#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE 662#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE
433 663
434#define WANT(len) if (expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data"); 664#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data")
435 665
436#define DEC_INC_DEPTH if (++dec->depth > dec->cbor.max_depth) ERR (ERR_NESTING_EXCEEDED) 666#define DEC_INC_DEPTH if (++dec->depth > dec->cbor.max_depth) ERR (ERR_NESTING_EXCEEDED)
437#define DEC_DEC_DEPTH --dec->depth 667#define DEC_DEC_DEPTH --dec->depth
438 668
439static UV 669static UV
440decode_uint (dec_t *dec) 670decode_uint (dec_t *dec)
441{ 671{
442 switch (*dec->cur & 31) 672 U8 m = *dec->cur & MINOR_MASK;
443 { 673 ++dec->cur;
444 case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
445 case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
446 case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
447 return *dec->cur++ & 31;
448 674
449 case 24: 675 if (ecb_expect_true (m < LENGTH_EXT1))
676 return m;
677 else if (ecb_expect_true (m == LENGTH_EXT1))
678 {
450 WANT (2); 679 WANT (1);
451 dec->cur += 2; 680 dec->cur += 1;
452 return dec->cur[-1]; 681 return dec->cur[-1];
453 682 }
454 case 25: 683 else if (ecb_expect_true (m == LENGTH_EXT2))
684 {
455 WANT (3); 685 WANT (2);
456 dec->cur += 3; 686 dec->cur += 2;
457 return (((UV)dec->cur[-2]) << 8) 687 return (((UV)dec->cur[-2]) << 8)
458 | ((UV)dec->cur[-1]); 688 | ((UV)dec->cur[-1]);
459 689 }
460 case 26: 690 else if (ecb_expect_true (m == LENGTH_EXT4))
691 {
461 WANT (5); 692 WANT (4);
462 dec->cur += 5; 693 dec->cur += 4;
463 return (((UV)dec->cur[-4]) << 24) 694 return (((UV)dec->cur[-4]) << 24)
464 | (((UV)dec->cur[-3]) << 16) 695 | (((UV)dec->cur[-3]) << 16)
465 | (((UV)dec->cur[-2]) << 8) 696 | (((UV)dec->cur[-2]) << 8)
466 | ((UV)dec->cur[-1]); 697 | ((UV)dec->cur[-1]);
467 698 }
468 case 27: 699 else if (ecb_expect_true (m == LENGTH_EXT8))
700 {
469 WANT (9); 701 WANT (8);
470 dec->cur += 9; 702 dec->cur += 8;
703
704 return
705#if UVSIZE < 8
706 0
707#else
471 return (((UV)dec->cur[-8]) << 56) 708 (((UV)dec->cur[-8]) << 56)
472 | (((UV)dec->cur[-7]) << 48) 709 | (((UV)dec->cur[-7]) << 48)
473 | (((UV)dec->cur[-6]) << 40) 710 | (((UV)dec->cur[-6]) << 40)
474 | (((UV)dec->cur[-5]) << 32) 711 | (((UV)dec->cur[-5]) << 32)
712#endif
475 | (((UV)dec->cur[-4]) << 24) 713 | (((UV)dec->cur[-4]) << 24)
476 | (((UV)dec->cur[-3]) << 16) 714 | (((UV)dec->cur[-3]) << 16)
477 | (((UV)dec->cur[-2]) << 8) 715 | (((UV)dec->cur[-2]) << 8)
478 | ((UV)dec->cur[-1]); 716 | ((UV)dec->cur[-1]);
479 717 }
480 default: 718 else
481 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 719 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
482 }
483 720
484fail: 721fail:
485 return 0; 722 return 0;
486} 723}
487 724
492{ 729{
493 AV *av = newAV (); 730 AV *av = newAV ();
494 731
495 DEC_INC_DEPTH; 732 DEC_INC_DEPTH;
496 733
497 if ((*dec->cur & 31) == 31) 734 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
498 { 735 {
499 ++dec->cur; 736 ++dec->cur;
500 737
501 for (;;) 738 for (;;)
502 { 739 {
503 WANT (1); 740 WANT (1);
504 741
505 if (*dec->cur == (0xe0 | 31)) 742 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
506 { 743 {
507 ++dec->cur; 744 ++dec->cur;
508 break; 745 break;
509 } 746 }
510 747
513 } 750 }
514 else 751 else
515 { 752 {
516 int i, len = decode_uint (dec); 753 int i, len = decode_uint (dec);
517 754
755 WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
518 av_fill (av, len - 1); 756 av_fill (av, len - 1);
519 757
520 for (i = 0; i < len; ++i) 758 for (i = 0; i < len; ++i)
521 AvARRAY (av)[i] = decode_sv (dec); 759 AvARRAY (av)[i] = decode_sv (dec);
522 } 760 }
528 SvREFCNT_dec (av); 766 SvREFCNT_dec (av);
529 DEC_DEC_DEPTH; 767 DEC_DEC_DEPTH;
530 return &PL_sv_undef; 768 return &PL_sv_undef;
531} 769}
532 770
771static void
772decode_he (dec_t *dec, HV *hv)
773{
774 // for speed reasons, we specialcase single-string
775 // byte or utf-8 strings as keys, but only when !stringref
776
777 if (ecb_expect_true (!dec->stringref))
778 if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
779 {
780 I32 len = decode_uint (dec);
781 char *key = (char *)dec->cur;
782
783 WANT (len);
784 dec->cur += len;
785
786 hv_store (hv, key, len, decode_sv (dec), 0);
787
788 return;
789 }
790 else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
791 {
792 I32 len = decode_uint (dec);
793 char *key = (char *)dec->cur;
794
795 WANT (len);
796 dec->cur += len;
797
798 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
799 if (!is_utf8_string (key, len))
800 ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
801
802 hv_store (hv, key, -len, decode_sv (dec), 0);
803
804 return;
805 }
806
807 SV *k = decode_sv (dec);
808 SV *v = decode_sv (dec);
809
810 hv_store_ent (hv, k, v, 0);
811 SvREFCNT_dec (k);
812
813fail:
814 ;
815}
816
533static SV * 817static SV *
534decode_hv (dec_t *dec) 818decode_hv (dec_t *dec)
535{ 819{
536 HV *hv = newHV (); 820 HV *hv = newHV ();
537 821
538 DEC_INC_DEPTH; 822 DEC_INC_DEPTH;
539 823
540 if ((*dec->cur & 31) == 31) 824 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
541 { 825 {
542 ++dec->cur; 826 ++dec->cur;
543 827
544 for (;;) 828 for (;;)
545 { 829 {
546 WANT (1); 830 WANT (1);
547 831
548 if (*dec->cur == (0xe0 | 31)) 832 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
549 { 833 {
550 ++dec->cur; 834 ++dec->cur;
551 break; 835 break;
552 } 836 }
553 837
554 SV *k = decode_sv (dec); 838 decode_he (dec, hv);
555 SV *v = decode_sv (dec);
556
557 hv_store_ent (hv, k, v, 0);
558 } 839 }
559 } 840 }
560 else 841 else
561 { 842 {
562 int len = decode_uint (dec); 843 int pairs = decode_uint (dec);
563 844
564 while (len--) 845 while (pairs--)
565 { 846 decode_he (dec, hv);
566 SV *k = decode_sv (dec);
567 SV *v = decode_sv (dec);
568
569 hv_store_ent (hv, k, v, 0);
570 }
571 } 847 }
572 848
573 DEC_DEC_DEPTH; 849 DEC_DEC_DEPTH;
574 return newRV_noinc ((SV *)hv); 850 return newRV_noinc ((SV *)hv);
575
576#if 0
577 SV *sv;
578 HV *hv = newHV ();
579
580 DEC_INC_DEPTH;
581 decode_ws (dec);
582
583 for (;;)
584 {
585 // heuristic: assume that
586 // a) decode_str + hv_store_ent are abysmally slow.
587 // b) most hash keys are short, simple ascii text.
588 // => try to "fast-match" such strings to avoid
589 // the overhead of decode_str + hv_store_ent.
590 {
591 SV *value;
592 char *p = dec->cur;
593 char *e = p + 24; // only try up to 24 bytes
594
595 for (;;)
596 {
597 // the >= 0x80 is false on most architectures
598 if (p == e || *p < 0x20 || *p >= 0x80 || *p == '\\')
599 {
600 // slow path, back up and use decode_str
601 SV *key = decode_str (dec);
602 if (!key)
603 goto fail;
604
605 decode_ws (dec); EXPECT_CH (':');
606
607 decode_ws (dec);
608 value = decode_sv (dec);
609 if (!value)
610 {
611 SvREFCNT_dec (key);
612 goto fail;
613 }
614
615 hv_store_ent (hv, key, value, 0);
616 SvREFCNT_dec (key);
617
618 break;
619 }
620 else if (*p == '"')
621 {
622 // fast path, got a simple key
623 char *key = dec->cur;
624 int len = p - key;
625 dec->cur = p + 1;
626
627 decode_ws (dec); EXPECT_CH (':');
628
629 decode_ws (dec);
630 value = decode_sv (dec);
631 if (!value)
632 goto fail;
633
634 hv_store (hv, key, len, value, 0);
635
636 break;
637 }
638
639 ++p;
640 }
641 }
642
643 decode_ws (dec);
644
645 if (*dec->cur == '}')
646 {
647 ++dec->cur;
648 break;
649 }
650
651 if (*dec->cur != ',')
652 ERR (", or } expected while parsing object/hash");
653
654 ++dec->cur;
655
656 decode_ws (dec);
657
658 if (*dec->cur == '}' && dec->cbor.flags & F_RELAXED)
659 {
660 ++dec->cur;
661 break;
662 }
663 }
664
665 DEC_DEC_DEPTH;
666 sv = newRV_noinc ((SV *)hv);
667
668 // check filter callbacks
669 if (dec->cbor.flags & F_HOOK)
670 {
671 if (dec->cbor.cb_sk_object && HvKEYS (hv) == 1)
672 {
673 HE *cb, *he;
674
675 hv_iterinit (hv);
676 he = hv_iternext (hv);
677 hv_iterinit (hv);
678
679 // the next line creates a mortal sv each time its called.
680 // might want to optimise this for common cases.
681 cb = hv_fetch_ent (dec->cbor.cb_sk_object, hv_iterkeysv (he), 0, 0);
682
683 if (cb)
684 {
685 dSP;
686 int count;
687
688 ENTER; SAVETMPS; PUSHMARK (SP);
689 XPUSHs (HeVAL (he));
690 sv_2mortal (sv);
691
692 PUTBACK; count = call_sv (HeVAL (cb), G_ARRAY); SPAGAIN;
693
694 if (count == 1)
695 {
696 sv = newSVsv (POPs);
697 FREETMPS; LEAVE;
698 return sv;
699 }
700
701 SvREFCNT_inc (sv);
702 FREETMPS; LEAVE;
703 }
704 }
705
706 if (dec->cbor.cb_object)
707 {
708 dSP;
709 int count;
710
711 ENTER; SAVETMPS; PUSHMARK (SP);
712 XPUSHs (sv_2mortal (sv));
713
714 PUTBACK; count = call_sv (dec->cbor.cb_object, G_ARRAY); SPAGAIN;
715
716 if (count == 1)
717 {
718 sv = newSVsv (POPs);
719 FREETMPS; LEAVE;
720 return sv;
721 }
722
723 SvREFCNT_inc (sv);
724 FREETMPS; LEAVE;
725 }
726 }
727
728 return sv;
729#endif
730 851
731fail: 852fail:
732 SvREFCNT_dec (hv); 853 SvREFCNT_dec (hv);
733 DEC_DEC_DEPTH; 854 DEC_DEC_DEPTH;
734 return &PL_sv_undef; 855 return &PL_sv_undef;
735} 856}
736 857
737static SV * 858static SV *
738decode_str (dec_t *dec, int utf8) 859decode_str (dec_t *dec, int utf8)
739{ 860{
740 SV *sv; 861 SV *sv = 0;
741 862
742 if ((*dec->cur & 31) == 31) 863 if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
743 { 864 {
865 // indefinite length strings
744 ++dec->cur; 866 ++dec->cur;
745 867
868 U8 major = *dec->cur & MAJOR_MISC;
869
746 sv = newSVpvn ("", 0); 870 sv = newSVpvn ("", 0);
747 871
748 // not very fast, and certainly not robust against illegal input
749 for (;;) 872 for (;;)
750 { 873 {
751 WANT (1); 874 WANT (1);
752 875
753 if (*dec->cur == (0xe0 | 31)) 876 if ((*dec->cur - major) > LENGTH_EXT8)
877 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
754 { 878 {
755 ++dec->cur; 879 ++dec->cur;
756 break; 880 break;
757 } 881 }
882 else
883 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
758 884
759 SV *sv2 = decode_sv (dec); 885 STRLEN len = decode_uint (dec);
760 sv_catsv (sv, sv2); 886
887 WANT (len);
888 sv_catpvn (sv, dec->cur, len);
889 dec->cur += len;
761 } 890 }
762 } 891 }
763 else 892 else
764 { 893 {
765 STRLEN len = decode_uint (dec); 894 STRLEN len = decode_uint (dec);
766 895
767 WANT (len); 896 WANT (len);
768 sv = newSVpvn (dec->cur, len); 897 sv = newSVpvn (dec->cur, len);
769 dec->cur += len; 898 dec->cur += len;
899
900 if (ecb_expect_false (dec->stringref)
901 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
902 av_push (dec->stringref, SvREFCNT_inc_NN (sv));
770 } 903 }
771 904
772 if (utf8) 905 if (utf8)
906 {
907 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
908 if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
909 ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
910
773 SvUTF8_on (sv); 911 SvUTF8_on (sv);
912 }
774 913
775 return sv; 914 return sv;
776 915
777fail: 916fail:
917 SvREFCNT_dec (sv);
778 return &PL_sv_undef; 918 return &PL_sv_undef;
779} 919}
780 920
781static SV * 921static SV *
782decode_tagged (dec_t *dec) 922decode_tagged (dec_t *dec)
783{ 923{
924 SV *sv = 0;
784 UV tag = decode_uint (dec); 925 UV tag = decode_uint (dec);
926
927 WANT (1);
928
929 switch (tag)
930 {
931 case CBOR_TAG_MAGIC:
785 SV *sv = decode_sv (dec); 932 sv = decode_sv (dec);
933 break;
786 934
787 if (tag == 55799) // 2.4.5 Self-Describe CBOR 935 case CBOR_TAG_INDIRECTION:
936 sv = newRV_noinc (decode_sv (dec));
937 break;
938
939 case CBOR_TAG_STRINGREF_NAMESPACE:
940 {
941 ENTER; SAVETMPS;
942
943 SAVESPTR (dec->stringref);
944 dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());
945
946 sv = decode_sv (dec);
947
948 FREETMPS; LEAVE;
949 }
950 break;
951
952 case CBOR_TAG_STRINGREF:
953 {
954 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
955 ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
956
957 UV idx = decode_uint (dec);
958
959 if (!dec->stringref || (int)idx > AvFILLp (dec->stringref))
960 ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)");
961
962 sv = newSVsv (AvARRAY (dec->stringref)[idx]);
963 }
964 break;
965
966 case CBOR_TAG_VALUE_SHAREABLE:
967 {
968 if (ecb_expect_false (!dec->shareable))
969 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
970
971 if (dec->cbor.flags & F_ALLOW_CYCLES)
972 {
973 sv = newSV (0);
974 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
975
976 SV *osv = decode_sv (dec);
977 sv_setsv (sv, osv);
978 SvREFCNT_dec_NN (osv);
979 }
980 else
981 {
982 av_push (dec->shareable, &PL_sv_undef);
983 int idx = AvFILLp (dec->shareable);
984 sv = decode_sv (dec);
985 av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
986 }
987 }
988 break;
989
990 case CBOR_TAG_VALUE_SHAREDREF:
991 {
992 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
993 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
994
995 UV idx = decode_uint (dec);
996
997 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable))
998 ERR ("corrupted CBOR data (sharedref index out of bounds)");
999
1000 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
1001
1002 if (sv == &PL_sv_undef)
1003 ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
1004 }
1005 break;
1006
1007 case CBOR_TAG_PERL_OBJECT:
1008 {
1009 sv = decode_sv (dec);
1010
1011 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
1012 ERR ("corrupted CBOR data (non-array perl object)");
1013
1014 AV *av = (AV *)SvRV (sv);
1015 int len = av_len (av) + 1;
1016 HV *stash = gv_stashsv (*av_fetch (av, 0, 1), 0);
1017
1018 if (!stash)
1019 ERR ("cannot decode perl-object (package does not exist)");
1020
1021 GV *method = gv_fetchmethod_autoload (stash, "THAW", 0);
1022
1023 if (!method)
1024 ERR ("cannot decode perl-object (package does not have a THAW method)");
1025
1026 dSP;
1027
1028 ENTER; SAVETMPS;
1029 PUSHMARK (SP);
1030 EXTEND (SP, len + 1);
1031 // we re-bless the reference to get overload and other niceties right
1032 PUSHs (*av_fetch (av, 0, 1));
1033 PUSHs (sv_cbor);
1034
1035 int i;
1036
1037 for (i = 1; i < len; ++i)
1038 PUSHs (*av_fetch (av, i, 1));
1039
1040 PUTBACK;
1041 call_sv ((SV *)GvCV (method), G_SCALAR | G_EVAL);
1042 SPAGAIN;
1043
1044 if (SvTRUE (ERRSV))
1045 {
1046 FREETMPS; LEAVE;
1047 ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));
1048 }
1049
1050 SvREFCNT_dec (sv);
1051 sv = SvREFCNT_inc (POPs);
1052
1053 PUTBACK;
1054
1055 FREETMPS; LEAVE;
1056 }
1057 break;
1058
1059 default:
1060 {
1061 sv = decode_sv (dec);
1062
1063 dSP;
1064 ENTER; SAVETMPS;
1065 SAVESTACK_POS ();
1066 PUSHMARK (SP);
1067 EXTEND (SP, 2);
1068 PUSHs (newSVuv (tag));
1069 PUSHs (sv);
1070
1071 PUTBACK;
1072 int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL);
1073 SPAGAIN;
1074
1075 if (SvTRUE (ERRSV))
1076 {
1077 FREETMPS; LEAVE;
1078 ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));
1079 }
1080
1081 if (count)
1082 {
1083 SvREFCNT_dec (sv);
1084 sv = SvREFCNT_inc (POPs);
1085 }
1086 else
1087 {
1088 AV *av = newAV ();
1089 av_push (av, newSVuv (tag));
1090 av_push (av, sv);
1091
1092 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
1093 ? cbor_tagged_stash
1094 : gv_stashpv ("CBOR::XS::Tagged" , 1);
1095 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
1096 }
1097
1098 PUTBACK;
1099
1100 FREETMPS; LEAVE;
1101 }
1102 break;
1103 }
1104
788 return sv; 1105 return sv;
789 1106
790 AV *av = newAV (); 1107fail:
791 av_push (av, newSVuv (tag)); 1108 SvREFCNT_dec (sv);
792 av_push (av, sv); 1109 return &PL_sv_undef;
793 return newRV_noinc ((SV *)av);
794} 1110}
795 1111
796static SV * 1112static SV *
797decode_sv (dec_t *dec) 1113decode_sv (dec_t *dec)
798{ 1114{
799 WANT (1); 1115 WANT (1);
800 1116
801 switch (*dec->cur >> 5) 1117 switch (*dec->cur >> MAJOR_SHIFT)
802 { 1118 {
803 case 0: // unsigned int 1119 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
804 //TODO: 64 bit values on 3 2bit perls 1120 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
805 return newSVuv (decode_uint (dec)); 1121 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
806 case 1: // negative int 1122 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
807 return newSViv (-1 - (IV)decode_uint (dec)); 1123 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
808 case 2: // octet string 1124 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
809 return decode_str (dec, 0); 1125 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
810 case 3: // utf-8 string 1126
811 return decode_str (dec, 1); 1127 case MAJOR_MISC >> MAJOR_SHIFT:
812 case 4: // array
813 return decode_av (dec);
814 case 5: // map
815 return decode_hv (dec);
816 case 6: // tag
817 return decode_tagged (dec);
818 case 7: // misc
819 switch (*dec->cur++ & 31) 1128 switch (*dec->cur++ & MINOR_MASK)
820 { 1129 {
821 case 20: 1130 case SIMPLE_FALSE:
822#if CBOR_SLOW 1131#if CBOR_SLOW
823 cbor_false = get_bool ("CBOR::XS::false"); 1132 types_false = get_bool ("Types::Serialiser::false");
824#endif 1133#endif
825 return newSVsv (cbor_false); 1134 return newSVsv (types_false);
826 case 21: 1135 case SIMPLE_TRUE:
827#if CBOR_SLOW 1136#if CBOR_SLOW
828 cbor_true = get_bool ("CBOR::XS::true"); 1137 types_true = get_bool ("Types::Serialiser::true");
829#endif 1138#endif
830 return newSVsv (cbor_true); 1139 return newSVsv (types_true);
831 case 22: 1140 case SIMPLE_NULL:
832 return newSVsv (&PL_sv_undef); 1141 return newSVsv (&PL_sv_undef);
1142 case SIMPLE_UNDEF:
1143#if CBOR_SLOW
1144 types_error = get_bool ("Types::Serialiser::error");
1145#endif
1146 return newSVsv (types_error);
833 1147
834 case 25: 1148 case MISC_FLOAT16:
835 { 1149 {
836 WANT (2); 1150 WANT (2);
837 1151
838 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1152 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
839 dec->cur += 2; 1153 dec->cur += 2;
840 1154
841 return newSVnv (ecb_binary16_to_float (fp)); 1155 return newSVnv (ecb_binary16_to_float (fp));
842 } 1156 }
843 1157
844 case 26: 1158 case MISC_FLOAT32:
845 { 1159 {
846 uint32_t fp; 1160 uint32_t fp;
847 WANT (4); 1161 WANT (4);
848 memcpy (&fp, dec->cur, 4); 1162 memcpy (&fp, dec->cur, 4);
849 dec->cur += 4; 1163 dec->cur += 4;
852 fp = ecb_bswap32 (fp); 1166 fp = ecb_bswap32 (fp);
853 1167
854 return newSVnv (ecb_binary32_to_float (fp)); 1168 return newSVnv (ecb_binary32_to_float (fp));
855 } 1169 }
856 1170
857 case 27: 1171 case MISC_FLOAT64:
858 { 1172 {
859 uint64_t fp; 1173 uint64_t fp;
860 WANT (8); 1174 WANT (8);
861 memcpy (&fp, dec->cur, 8); 1175 memcpy (&fp, dec->cur, 8);
862 dec->cur += 8; 1176 dec->cur += 8;
865 fp = ecb_bswap64 (fp); 1179 fp = ecb_bswap64 (fp);
866 1180
867 return newSVnv (ecb_binary64_to_double (fp)); 1181 return newSVnv (ecb_binary64_to_double (fp));
868 } 1182 }
869 1183
870 // 0..19 unassigned 1184 // 0..19 unassigned simple
871 // 24 reserved + unassigned (reserved values are not encodable) 1185 // 24 reserved + unassigned simple (reserved values are not encodable)
1186 // 28-30 unassigned misc
1187 // 31 break code
872 default: 1188 default:
873 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1189 ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
874 } 1190 }
875 1191
876 break; 1192 break;
877 } 1193 }
878#if 0
879 switch (*dec->cur)
880 {
881 //case '"': ++dec->cur; return decode_str (dec);
882 case '[': ++dec->cur; return decode_av (dec);
883 case '{': ++dec->cur; return decode_hv (dec);
884
885 case '-':
886 case '0': case '1': case '2': case '3': case '4':
887 case '5': case '6': case '7': case '8': case '9':
888 //TODO return decode_num (dec);
889
890 case 't':
891 if (dec->end - dec->cur >= 4 && !memcmp (dec->cur, "true", 4))
892 {
893 dec->cur += 4;
894#if CBOR_SLOW
895 cbor_true = get_bool ("CBOR::XS::true");
896#endif
897 return newSVsv (cbor_true);
898 }
899 else
900 ERR ("'true' expected");
901
902 break;
903
904 case 'f':
905 if (dec->end - dec->cur >= 5 && !memcmp (dec->cur, "false", 5))
906 {
907 dec->cur += 5;
908#if CBOR_SLOW
909 cbor_false = get_bool ("CBOR::XS::false");
910#endif
911 return newSVsv (cbor_false);
912 }
913 else
914 ERR ("'false' expected");
915
916 break;
917
918 case 'n':
919 if (dec->end - dec->cur >= 4 && !memcmp (dec->cur, "null", 4))
920 {
921 dec->cur += 4;
922 return newSVsv (&PL_sv_undef);
923 }
924 else
925 ERR ("'null' expected");
926
927 break;
928
929 default:
930 ERR ("malformed CBOR string, neither array, object, number, string or atom");
931 break;
932 }
933#endif
934 1194
935fail: 1195fail:
936 return &PL_sv_undef; 1196 return &PL_sv_undef;
937} 1197}
938 1198
939static SV * 1199static SV *
940decode_cbor (SV *string, CBOR *cbor, char **offset_return) 1200decode_cbor (SV *string, CBOR *cbor, char **offset_return)
941{ 1201{
942 dec_t dec; 1202 dec_t dec = { 0 };
943 SV *sv; 1203 SV *sv;
1204 STRLEN len;
1205 char *data = SvPVbyte (string, len);
944 1206
945 /* work around bugs in 5.10 where manipulating magic values
946 * makes perl ignore the magic in subsequent accesses.
947 * also make a copy of non-PV values, to get them into a clean
948 * state (SvPV should do that, but it's buggy, see below).
949 */
950 /*SvGETMAGIC (string);*/
951 if (SvMAGICAL (string) || !SvPOK (string))
952 string = sv_2mortal (newSVsv (string));
953
954 SvUPGRADE (string, SVt_PV);
955
956 /* work around a bug in perl 5.10, which causes SvCUR to fail an
957 * assertion with -DDEBUGGING, although SvCUR is documented to
958 * return the xpv_cur field which certainly exists after upgrading.
959 * according to nicholas clark, calling SvPOK fixes this.
960 * But it doesn't fix it, so try another workaround, call SvPV_nolen
961 * and hope for the best.
962 * Damnit, SvPV_nolen still trips over yet another assertion. This
963 * assertion business is seriously broken, try yet another workaround
964 * for the broken -DDEBUGGING.
965 */
966 {
967#ifdef DEBUGGING
968 STRLEN offset = SvOK (string) ? sv_len (string) : 0;
969#else
970 STRLEN offset = SvCUR (string);
971#endif
972
973 if (offset > cbor->max_size && cbor->max_size) 1207 if (len > cbor->max_size && cbor->max_size)
974 croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu", 1208 croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
975 (unsigned long)SvCUR (string), (unsigned long)cbor->max_size); 1209 (unsigned long)len, (unsigned long)cbor->max_size);
976 }
977
978 sv_utf8_downgrade (string, 0);
979 1210
980 dec.cbor = *cbor; 1211 dec.cbor = *cbor;
981 dec.cur = (U8 *)SvPVX (string); 1212 dec.cur = (U8 *)data;
982 dec.end = (U8 *)SvEND (string); 1213 dec.end = (U8 *)data + len;
983 dec.err = 0;
984 dec.depth = 0;
985
986 if (dec.cbor.cb_object || dec.cbor.cb_sk_object)
987 ;//TODO dec.cbor.flags |= F_HOOK;
988 1214
989 sv = decode_sv (&dec); 1215 sv = decode_sv (&dec);
990 1216
991 if (offset_return) 1217 if (offset_return)
992 *offset_return = dec.cur; 1218 *offset_return = dec.cur;
995 if (dec.cur != dec.end && !dec.err) 1221 if (dec.cur != dec.end && !dec.err)
996 dec.err = "garbage after CBOR object"; 1222 dec.err = "garbage after CBOR object";
997 1223
998 if (dec.err) 1224 if (dec.err)
999 { 1225 {
1226 if (dec.shareable)
1227 {
1228 // need to break cyclic links, which whould all be in shareable
1229 int i;
1230 SV **svp;
1231
1232 for (i = av_len (dec.shareable) + 1; i--; )
1233 if ((svp = av_fetch (dec.shareable, i, 0)))
1234 sv_setsv (*svp, &PL_sv_undef);
1235 }
1236
1000 SvREFCNT_dec (sv); 1237 SvREFCNT_dec (sv);
1001 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)SvPVX (string), (int)(uint8_t)*dec.cur); 1238 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur);
1002 } 1239 }
1003 1240
1004 sv = sv_2mortal (sv); 1241 sv = sv_2mortal (sv);
1005 1242
1006 return sv; 1243 return sv;
1007} 1244}
1008 1245
1246/////////////////////////////////////////////////////////////////////////////
1247// incremental parser
1248
1249#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
1250
1251// returns 0 for notyet, 1 for success or error
1252static int
1253incr_parse (CBOR *self, SV *cborstr)
1254{
1255 STRLEN cur;
1256 SvPV (cborstr, cur);
1257
1258 while (ecb_expect_true (self->incr_need <= cur))
1259 {
1260 // table of integer count bytes
1261 static I8 incr_len[MINOR_MASK + 1] = {
1262 0, 0, 0, 0, 0, 0, 0, 0,
1263 0, 0, 0, 0, 0, 0, 0, 0,
1264 0, 0, 0, 0, 0, 0, 0, 0,
1265 1, 2, 4, 8,-1,-1,-1,-2
1266 };
1267
1268 const U8 *p = SvPVX (cborstr) + self->incr_pos;
1269 U8 m = *p & MINOR_MASK;
1270 IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1271 I8 ilen = incr_len[m];
1272
1273 self->incr_need = self->incr_pos + 1;
1274
1275 if (ecb_expect_false (ilen < 0))
1276 {
1277 if (m != MINOR_INDEF)
1278 return 1; // error
1279
1280 if (*p == (MAJOR_MISC | MINOR_INDEF))
1281 {
1282 if (count >= 0)
1283 return 1; // error
1284
1285 count = 1;
1286 }
1287 else
1288 {
1289 av_push (self->incr_count, newSViv (-1)); //TODO: nest
1290 count = -1;
1291 }
1292 }
1293 else
1294 {
1295 self->incr_need += ilen;
1296 if (ecb_expect_false (self->incr_need > cur))
1297 return 0;
1298
1299 int major = *p >> MAJOR_SHIFT;
1300
1301 switch (major)
1302 {
1303 case MAJOR_TAG >> MAJOR_SHIFT:
1304 ++count; // tags merely prefix another value
1305 break;
1306
1307 case MAJOR_BYTES >> MAJOR_SHIFT:
1308 case MAJOR_TEXT >> MAJOR_SHIFT:
1309 case MAJOR_ARRAY >> MAJOR_SHIFT:
1310 case MAJOR_MAP >> MAJOR_SHIFT:
1311 {
1312 UV len;
1313
1314 if (ecb_expect_false (ilen))
1315 {
1316 len = 0;
1317
1318 do {
1319 len = (len << 8) | *++p;
1320 } while (--ilen);
1321 }
1322 else
1323 len = m;
1324
1325 switch (major)
1326 {
1327 case MAJOR_BYTES >> MAJOR_SHIFT:
1328 case MAJOR_TEXT >> MAJOR_SHIFT:
1329 self->incr_need += len;
1330 if (ecb_expect_false (self->incr_need > cur))
1331 return 0;
1332
1333 break;
1334
1335 case MAJOR_MAP >> MAJOR_SHIFT:
1336 len <<= 1;
1337 case MAJOR_ARRAY >> MAJOR_SHIFT:
1338 if (len)
1339 {
1340 av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
1341 count = len + 1;
1342 }
1343 break;
1344 }
1345 }
1346 }
1347 }
1348
1349 self->incr_pos = self->incr_need;
1350
1351 if (count > 0)
1352 {
1353 while (!--count)
1354 {
1355 if (!AvFILLp (self->incr_count))
1356 return 1; // done
1357
1358 SvREFCNT_dec_NN (av_pop (self->incr_count));
1359 count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1360 }
1361
1362 SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
1363 }
1364 }
1365
1366 return 0;
1367}
1368
1369
1009///////////////////////////////////////////////////////////////////////////// 1370/////////////////////////////////////////////////////////////////////////////
1010// XS interface functions 1371// XS interface functions
1011 1372
1012MODULE = CBOR::XS PACKAGE = CBOR::XS 1373MODULE = CBOR::XS PACKAGE = CBOR::XS
1013 1374
1014BOOT: 1375BOOT:
1015{ 1376{
1016 cbor_stash = gv_stashpv ("CBOR::XS" , 1); 1377 cbor_stash = gv_stashpv ("CBOR::XS" , 1);
1017 cbor_boolean_stash = gv_stashpv ("CBOR::XS::Boolean", 1); 1378 cbor_tagged_stash = gv_stashpv ("CBOR::XS::Tagged" , 1);
1018 1379
1019 cbor_true = get_bool ("CBOR::XS::true"); 1380 types_boolean_stash = gv_stashpv ("Types::Serialiser::Boolean", 1);
1020 cbor_false = get_bool ("CBOR::XS::false"); 1381 types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1);
1382
1383 types_true = get_bool ("Types::Serialiser::true" );
1384 types_false = get_bool ("Types::Serialiser::false");
1385 types_error = get_bool ("Types::Serialiser::error");
1386
1387 default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1388
1389 sv_cbor = newSVpv ("CBOR", 0);
1390 SvREADONLY_on (sv_cbor);
1021} 1391}
1022 1392
1023PROTOTYPES: DISABLE 1393PROTOTYPES: DISABLE
1024 1394
1025void CLONE (...) 1395void CLONE (...)
1026 CODE: 1396 CODE:
1027 cbor_stash = 0; 1397 cbor_stash = 0;
1398 cbor_tagged_stash = 0;
1399 types_error_stash = 0;
1028 cbor_boolean_stash = 0; 1400 types_boolean_stash = 0;
1029 1401
1030void new (char *klass) 1402void new (char *klass)
1031 PPCODE: 1403 PPCODE:
1032{ 1404{
1033 SV *pv = NEWSV (0, sizeof (CBOR)); 1405 SV *pv = NEWSV (0, sizeof (CBOR));
1041 1413
1042void shrink (CBOR *self, int enable = 1) 1414void shrink (CBOR *self, int enable = 1)
1043 ALIAS: 1415 ALIAS:
1044 shrink = F_SHRINK 1416 shrink = F_SHRINK
1045 allow_unknown = F_ALLOW_UNKNOWN 1417 allow_unknown = F_ALLOW_UNKNOWN
1418 allow_sharing = F_ALLOW_SHARING
1419 allow_cycles = F_ALLOW_CYCLES
1420 pack_strings = F_PACK_STRINGS
1421 utf8_strings = F_UTF8_STRINGS
1422 validate_utf8 = F_VALIDATE_UTF8
1046 PPCODE: 1423 PPCODE:
1047{ 1424{
1048 if (enable) 1425 if (enable)
1049 self->flags |= ix; 1426 self->flags |= ix;
1050 else 1427 else
1055 1432
1056void get_shrink (CBOR *self) 1433void get_shrink (CBOR *self)
1057 ALIAS: 1434 ALIAS:
1058 get_shrink = F_SHRINK 1435 get_shrink = F_SHRINK
1059 get_allow_unknown = F_ALLOW_UNKNOWN 1436 get_allow_unknown = F_ALLOW_UNKNOWN
1437 get_allow_sharing = F_ALLOW_SHARING
1438 get_allow_cycles = F_ALLOW_CYCLES
1439 get_pack_strings = F_PACK_STRINGS
1440 get_validate_utf8 = F_VALIDATE_UTF8
1060 PPCODE: 1441 PPCODE:
1061 XPUSHs (boolSV (self->flags & ix)); 1442 XPUSHs (boolSV (self->flags & ix));
1062 1443
1063void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1444void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1064 PPCODE: 1445 PPCODE:
1080 CODE: 1461 CODE:
1081 RETVAL = self->max_size; 1462 RETVAL = self->max_size;
1082 OUTPUT: 1463 OUTPUT:
1083 RETVAL 1464 RETVAL
1084 1465
1085#if 0 //TODO 1466void filter (CBOR *self, SV *filter = 0)
1086
1087void filter_cbor_object (CBOR *self, SV *cb = &PL_sv_undef)
1088 PPCODE: 1467 PPCODE:
1089{
1090 SvREFCNT_dec (self->cb_object); 1468 SvREFCNT_dec (self->filter);
1091 self->cb_object = SvOK (cb) ? newSVsv (cb) : 0; 1469 self->filter = filter ? newSVsv (filter) : filter;
1092
1093 XPUSHs (ST (0)); 1470 XPUSHs (ST (0));
1094}
1095 1471
1096void filter_cbor_single_key_object (CBOR *self, SV *key, SV *cb = &PL_sv_undef) 1472SV *get_filter (CBOR *self)
1097 PPCODE: 1473 CODE:
1098{ 1474 RETVAL = self->filter ? self->filter : NEWSV (0, 0);
1099 if (!self->cb_sk_object) 1475 OUTPUT:
1100 self->cb_sk_object = newHV (); 1476 RETVAL
1101
1102 if (SvOK (cb))
1103 hv_store_ent (self->cb_sk_object, key, newSVsv (cb), 0);
1104 else
1105 {
1106 hv_delete_ent (self->cb_sk_object, key, G_DISCARD, 0);
1107
1108 if (!HvKEYS (self->cb_sk_object))
1109 {
1110 SvREFCNT_dec (self->cb_sk_object);
1111 self->cb_sk_object = 0;
1112 }
1113 }
1114
1115 XPUSHs (ST (0));
1116}
1117
1118#endif
1119 1477
1120void encode (CBOR *self, SV *scalar) 1478void encode (CBOR *self, SV *scalar)
1121 PPCODE: 1479 PPCODE:
1122 PUTBACK; scalar = encode_cbor (scalar, self); SPAGAIN; 1480 PUTBACK; scalar = encode_cbor (scalar, self); SPAGAIN;
1123 XPUSHs (scalar); 1481 XPUSHs (scalar);
1136 EXTEND (SP, 2); 1494 EXTEND (SP, 2);
1137 PUSHs (sv); 1495 PUSHs (sv);
1138 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1496 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1139} 1497}
1140 1498
1499void incr_parse (CBOR *self, SV *cborstr)
1500 ALIAS:
1501 incr_parse_multiple = 1
1502 PPCODE:
1503{
1504 if (SvUTF8 (cborstr))
1505 sv_utf8_downgrade (cborstr, 0);
1506
1507 if (!self->incr_count)
1508 {
1509 self->incr_count = newAV ();
1510 self->incr_pos = 0;
1511 self->incr_need = 1;
1512
1513 av_push (self->incr_count, newSViv (1));
1514 }
1515
1516 do
1517 {
1518 if (!incr_parse (self, cborstr))
1519 {
1520 if (self->incr_need > self->max_size && self->max_size)
1521 croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
1522 (unsigned long)self->incr_need, (unsigned long)self->max_size);
1523
1524 break;
1525 }
1526
1527 SV *sv;
1528 char *offset;
1529
1530 PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
1531 XPUSHs (sv);
1532
1533 sv_chop (cborstr, offset);
1534
1535 av_clear (self->incr_count);
1536 av_push (self->incr_count, newSViv (1));
1537
1538 self->incr_pos = 0;
1539 self->incr_need = self->incr_pos + 1;
1540 }
1541 while (ix);
1542}
1543
1544void incr_reset (CBOR *self)
1545 CODE:
1546{
1547 SvREFCNT_dec (self->incr_count);
1548 self->incr_count = 0;
1549}
1550
1141void DESTROY (CBOR *self) 1551void DESTROY (CBOR *self)
1142 CODE: 1552 PPCODE:
1143 SvREFCNT_dec (self->cb_sk_object); 1553 cbor_free (self);
1144 SvREFCNT_dec (self->cb_object);
1145 1554
1146PROTOTYPES: ENABLE 1555PROTOTYPES: ENABLE
1147 1556
1148void encode_cbor (SV *scalar) 1557void encode_cbor (SV *scalar)
1558 ALIAS:
1559 encode_cbor = 0
1560 encode_cbor_sharing = F_ALLOW_SHARING
1149 PPCODE: 1561 PPCODE:
1150{ 1562{
1151 CBOR cbor; 1563 CBOR cbor;
1152 cbor_init (&cbor); 1564 cbor_init (&cbor);
1565 cbor.flags |= ix;
1153 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1566 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1154 XPUSHs (scalar); 1567 XPUSHs (scalar);
1155} 1568}
1156 1569
1157void decode_cbor (SV *cborstr) 1570void decode_cbor (SV *cborstr)

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines