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 1.1 by root, Fri Oct 25 23:09:45 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
14#if defined(__BORLANDC__) || defined(_MSC_VER) 15// compatibility with perl <5.18
15# define snprintf _snprintf // C compilers have this in stdio.h 16#ifndef HvNAMELEN_get
17# define HvNAMELEN_get(hv) strlen (HvNAME (hv))
16#endif 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
17 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
18#define F_SHRINK 0x00000200UL 100#define F_SHRINK 0x00000001UL
19#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
20 107
21#define INIT_SIZE 32 // initial scalar size to be allocated 108#define INIT_SIZE 32 // initial scalar size to be allocated
22 109
23#define SB do { 110#define SB do {
24#define SE } while (0) 111#define SE } while (0)
25
26#if __GNUC__ >= 3
27# define expect(expr,value) __builtin_expect ((expr), (value))
28# define INLINE static inline
29#else
30# define expect(expr,value) (expr)
31# define INLINE static
32#endif
33
34#define expect_false(expr) expect ((expr) != 0, 0)
35#define expect_true(expr) expect ((expr) != 0, 1)
36 112
37#define IN_RANGE_INC(type,val,beg,end) \ 113#define IN_RANGE_INC(type,val,beg,end) \
38 ((unsigned type)((unsigned type)(val) - (unsigned type)(beg)) \ 114 ((unsigned type)((unsigned type)(val) - (unsigned type)(beg)) \
39 <= (unsigned type)((unsigned type)(end) - (unsigned type)(beg))) 115 <= (unsigned type)((unsigned type)(end) - (unsigned type)(beg)))
40 116
46#else 122#else
47# define CBOR_SLOW 0 123# define CBOR_SLOW 0
48# define CBOR_STASH cbor_stash 124# define CBOR_STASH cbor_stash
49#endif 125#endif
50 126
51static HV *cbor_stash, *cbor_boolean_stash; // CBOR::XS:: 127static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS::
52static SV *cbor_true, *cbor_false; 128static SV *types_true, *types_false, *types_error, *sv_cbor, *default_filter;
53 129
54typedef struct { 130typedef struct {
55 U32 flags; 131 U32 flags;
56 U32 max_depth; 132 U32 max_depth;
57 STRLEN max_size; 133 STRLEN max_size;
134 SV *filter;
58 135
59 SV *cb_object; 136 // for the incremental parser
60 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.
61} CBOR; 140} CBOR;
62 141
63INLINE void 142ecb_inline void
64cbor_init (CBOR *cbor) 143cbor_init (CBOR *cbor)
65{ 144{
66 Zero (cbor, 1, CBOR); 145 Zero (cbor, 1, CBOR);
67 cbor->max_depth = 512; 146 cbor->max_depth = 512;
68} 147}
69 148
149ecb_inline void
150cbor_free (CBOR *cbor)
151{
152 SvREFCNT_dec (cbor->filter);
153 SvREFCNT_dec (cbor->incr_count);
154}
155
70///////////////////////////////////////////////////////////////////////////// 156/////////////////////////////////////////////////////////////////////////////
71// utility functions 157// utility functions
72 158
73INLINE SV * 159ecb_inline SV *
74get_bool (const char *name) 160get_bool (const char *name)
75{ 161{
76 SV *sv = get_sv (name, 1); 162 SV *sv = get_sv (name, 1);
77 163
78 SvREADONLY_on (sv); 164 SvREADONLY_on (sv);
79 SvREADONLY_on (SvRV (sv)); 165 SvREADONLY_on (SvRV (sv));
80 166
81 return sv; 167 return sv;
82} 168}
83 169
84INLINE void 170ecb_inline void
85shrink (SV *sv) 171shrink (SV *sv)
86{ 172{
87 sv_utf8_downgrade (sv, 1); 173 sv_utf8_downgrade (sv, 1);
88 174
89 if (SvLEN (sv) > SvCUR (sv) + 1) 175 if (SvLEN (sv) > SvCUR (sv) + 1)
94 SvPV_renew (sv, SvCUR (sv) + 1); 180 SvPV_renew (sv, SvCUR (sv) + 1);
95#endif 181#endif
96 } 182 }
97} 183}
98 184
99///////////////////////////////////////////////////////////////////////////// 185// minimum length of a string to be registered for stringref
100// fp hell 186ecb_inline int
101 187minimum_string_length (UV idx)
102//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}
103 199
104///////////////////////////////////////////////////////////////////////////// 200/////////////////////////////////////////////////////////////////////////////
105// encoder 201// encoder
106 202
107// structure used for encoding CBOR 203// structure used for encoding CBOR
110 char *cur; // SvPVX (sv) + current output position 206 char *cur; // SvPVX (sv) + current output position
111 char *end; // SvEND (sv) 207 char *end; // SvEND (sv)
112 SV *sv; // result scalar 208 SV *sv; // result scalar
113 CBOR cbor; 209 CBOR cbor;
114 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;
115} enc_t; 215} enc_t;
116 216
117INLINE void 217ecb_inline void
118need (enc_t *enc, STRLEN len) 218need (enc_t *enc, STRLEN len)
119{ 219{
120 if (expect_false (enc->cur + len >= enc->end)) 220 if (ecb_expect_false (enc->cur + len >= enc->end))
121 { 221 {
122 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv); 222 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv);
123 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1); 223 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1);
124 enc->cur = SvPVX (enc->sv) + cur; 224 enc->cur = SvPVX (enc->sv) + cur;
125 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1; 225 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;
126 } 226 }
127} 227}
128 228
129INLINE void 229ecb_inline void
130encode_ch (enc_t *enc, char ch) 230encode_ch (enc_t *enc, char ch)
131{ 231{
132 need (enc, 1); 232 need (enc, 1);
133 *enc->cur++ = ch; 233 *enc->cur++ = ch;
134} 234}
136static void 236static void
137encode_uint (enc_t *enc, int major, UV len) 237encode_uint (enc_t *enc, int major, UV len)
138{ 238{
139 need (enc, 9); 239 need (enc, 9);
140 240
141 if (len < 24) 241 if (ecb_expect_true (len < LENGTH_EXT1))
142 *enc->cur++ = major | len; 242 *enc->cur++ = major | len;
143 else if (len < 0x100) 243 else if (ecb_expect_true (len <= 0xffU))
144 { 244 {
145 *enc->cur++ = major | 24; 245 *enc->cur++ = major | LENGTH_EXT1;
146 *enc->cur++ = len; 246 *enc->cur++ = len;
147 } 247 }
148 else if (len < 0x10000) 248 else if (len <= 0xffffU)
149 { 249 {
150 *enc->cur++ = major | 25; 250 *enc->cur++ = major | LENGTH_EXT2;
151 *enc->cur++ = len >> 8; 251 *enc->cur++ = len >> 8;
152 *enc->cur++ = len; 252 *enc->cur++ = len;
153 } 253 }
154 else if (len < 0x100000000) 254 else if (len <= 0xffffffffU)
155 { 255 {
156 *enc->cur++ = major | 26; 256 *enc->cur++ = major | LENGTH_EXT4;
157 *enc->cur++ = len >> 24; 257 *enc->cur++ = len >> 24;
158 *enc->cur++ = len >> 16; 258 *enc->cur++ = len >> 16;
159 *enc->cur++ = len >> 8; 259 *enc->cur++ = len >> 8;
160 *enc->cur++ = len; 260 *enc->cur++ = len;
161 } 261 }
162 else if (len) 262 else
163 { 263 {
164 *enc->cur++ = major | 27; 264 *enc->cur++ = major | LENGTH_EXT8;
165 *enc->cur++ = len >> 56; 265 *enc->cur++ = len >> 56;
166 *enc->cur++ = len >> 48; 266 *enc->cur++ = len >> 48;
167 *enc->cur++ = len >> 40; 267 *enc->cur++ = len >> 40;
168 *enc->cur++ = len >> 32; 268 *enc->cur++ = len >> 32;
169 *enc->cur++ = len >> 24; 269 *enc->cur++ = len >> 24;
171 *enc->cur++ = len >> 8; 271 *enc->cur++ = len >> 8;
172 *enc->cur++ = len; 272 *enc->cur++ = len;
173 } 273 }
174} 274}
175 275
176static void 276ecb_inline void
277encode_tag (enc_t *enc, UV tag)
278{
279 encode_uint (enc, MAJOR_TAG, tag);
280}
281
282ecb_inline void
177encode_str (enc_t *enc, int utf8, char *str, STRLEN len) 283encode_str (enc_t *enc, int utf8, char *str, STRLEN len)
178{ 284{
179 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);
180 need (enc, len); 311 need (enc, len);
181 memcpy (enc->cur, str, len); 312 memcpy (enc->cur, str, len);
182 enc->cur += len; 313 enc->cur += len;
183} 314}
184 315
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
185static void encode_sv (enc_t *enc, SV *sv); 341static void encode_sv (enc_t *enc, SV *sv);
186 342
187static void 343static void
188encode_av (enc_t *enc, AV *av) 344encode_av (enc_t *enc, AV *av)
189{ 345{
192 if (enc->depth >= enc->cbor.max_depth) 348 if (enc->depth >= enc->cbor.max_depth)
193 croak (ERR_NESTING_EXCEEDED); 349 croak (ERR_NESTING_EXCEEDED);
194 350
195 ++enc->depth; 351 ++enc->depth;
196 352
197 encode_uint (enc, 0x80, len + 1); 353 encode_uint (enc, MAJOR_ARRAY, len + 1);
198 354
355 if (SvMAGICAL (av))
199 for (i = 0; i <= len; ++i) 356 for (i = 0; i <= len; ++i)
200 { 357 {
201 SV **svp = av_fetch (av, i, 0); 358 SV **svp = av_fetch (av, i, 0);
202 encode_sv (enc, svp ? *svp : &PL_sv_undef); 359 encode_sv (enc, svp ? *svp : &PL_sv_undef);
203 } 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 }
204 367
205 --enc->depth; 368 --enc->depth;
206} 369}
207 370
208static void 371static void
217 380
218 int pairs = hv_iterinit (hv); 381 int pairs = hv_iterinit (hv);
219 int mg = SvMAGICAL (hv); 382 int mg = SvMAGICAL (hv);
220 383
221 if (mg) 384 if (mg)
222 encode_ch (enc, 0xa0 | 31); 385 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
223 else 386 else
224 encode_uint (enc, 0xa0, pairs); 387 encode_uint (enc, MAJOR_MAP, pairs);
225 388
226 while ((he = hv_iternext (hv))) 389 while ((he = hv_iternext (hv)))
227 { 390 {
228 if (HeKLEN (he) == HEf_SVKEY) 391 if (HeKLEN (he) == HEf_SVKEY)
229 encode_sv (enc, HeSVKEY (he)); 392 encode_sv (enc, HeSVKEY (he));
230 else 393 else
231 encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); 394 encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
232 395
233 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));
234 } 397 }
235 398
236 if (mg) 399 if (mg)
237 encode_ch (enc, 0xe0 | 31); 400 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
238 401
239 --enc->depth; 402 --enc->depth;
240} 403}
241 404
242// encode objects, arrays and special \0=false and \1=true values. 405// encode objects, arrays and special \0=false and \1=true values.
243static void 406static void
244encode_rv (enc_t *enc, SV *sv) 407encode_rv (enc_t *enc, SV *sv)
245{ 408{
246 svtype svt;
247
248 SvGETMAGIC (sv); 409 SvGETMAGIC (sv);
410
249 svt = SvTYPE (sv); 411 svtype svt = SvTYPE (sv);
250 412
251 if (expect_false (SvOBJECT (sv))) 413 if (ecb_expect_false (SvOBJECT (sv)))
252 { 414 {
253 HV *stash = !CBOR_SLOW || cbor_boolean_stash 415 HV *boolean_stash = !CBOR_SLOW || types_boolean_stash
254 ? 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
255 : gv_stashpv ("CBOR::XS::Boolean", 1); 423 : gv_stashpv ("CBOR::XS::Tagged" , 1);
256 424
257 if (SvSTASH (sv) == stash) 425 HV *stash = SvSTASH (sv);
258 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 }
259 else 463 else
260 { 464 {
261#if 0 //TODO 465 sv_setuv (*svp, enc->shareable_idx);
262 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)))
263 { 477 {
478 dSP;
479
480 ENTER; SAVETMPS;
481 PUSHMARK (SP);
264 // 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
265 GV *to_cbor = gv_fetchmethod_autoload (SvSTASH (sv), "TO_CBOR", 0);
266
267 if (to_cbor)
268 {
269 dSP;
270
271 ENTER; SAVETMPS; PUSHMARK (SP);
272 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), SvSTASH (sv))); 483 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
273 484
274 // calling with G_SCALAR ensures that we always get a 1 return value
275 PUTBACK; 485 PUTBACK;
486 // G_SCALAR ensures that return value is 1
276 call_sv ((SV *)GvCV (to_cbor), G_SCALAR); 487 call_sv ((SV *)GvCV (method), G_SCALAR);
277 SPAGAIN; 488 SPAGAIN;
278 489
279 // catch this surprisingly common error 490 // catch this surprisingly common error
280 if (SvROK (TOPs) && SvRV (TOPs) == sv) 491 if (SvROK (TOPs) && SvRV (TOPs) == sv)
281 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));
282 493
283 sv = POPs;
284 PUTBACK;
285
286 encode_sv (enc, sv); 494 encode_sv (enc, POPs);
287 495
496 PUTBACK;
497
288 FREETMPS; LEAVE; 498 FREETMPS; LEAVE;
289 }
290 else if (enc->cbor.flags & F_ALLOW_BLESSED)
291 encode_str (enc, "null", 4, 0);
292 else
293 croak ("encountered object '%s', but neither allow_blessed enabled nor TO_CBOR method available on it",
294 SvPV_nolen (sv_2mortal (newRV_inc (sv))));
295 } 499 }
296 else if (enc->cbor.flags & F_ALLOW_BLESSED) 500 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
297 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 }
298 else 531 else
299 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",
300 SvPV_nolen (sv_2mortal (newRV_inc (sv)))); 533 SvPV_nolen (sv_2mortal (newRV_inc (sv))));
301#endif
302 }
303 } 534 }
304 else if (svt == SVt_PVHV) 535 else if (svt == SVt_PVHV)
305 encode_hv (enc, (HV *)sv); 536 encode_hv (enc, (HV *)sv);
306 else if (svt == SVt_PVAV) 537 else if (svt == SVt_PVAV)
307 encode_av (enc, (AV *)sv); 538 encode_av (enc, (AV *)sv);
308 else if (svt < SVt_PVAV)
309 {
310 STRLEN len = 0;
311 char *pv = svt ? SvPV (sv, len) : 0;
312
313 if (len == 1 && *pv == '1')
314 encode_ch (enc, 0xe0 | 21);
315 else if (len == 1 && *pv == '0')
316 encode_ch (enc, 0xe0 | 20);
317 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
318 encode_ch (enc, 0xe0 | 23);
319 else
320 croak ("cannot encode reference to scalar '%s' unless the scalar is 0 or 1",
321 SvPV_nolen (sv_2mortal (newRV_inc (sv))));
322 }
323 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
324 encode_ch (enc, 0xe0 | 23);
325 else 539 else
326 croak ("encountered %s, but CBOR can only represent references to arrays or hashes", 540 {
327 SvPV_nolen (sv_2mortal (newRV_inc (sv)))); 541 encode_tag (enc, CBOR_TAG_INDIRECTION);
542 encode_sv (enc, sv);
543 }
328} 544}
329 545
330static void 546static void
331encode_nv (enc_t *enc, SV *sv) 547encode_nv (enc_t *enc, SV *sv)
332{ 548{
333 double nv = SvNVX (sv); 549 double nv = SvNVX (sv);
334 550
335 need (enc, 9); 551 need (enc, 9);
336 552
337 if (expect_false (nv == (U32)nv)) 553 if (ecb_expect_false (nv == (NV)(U32)nv))
338 encode_uint (enc, 0x00, (U32)nv); 554 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
339 //TODO: maybe I32? 555 //TODO: maybe I32?
340 else if (expect_false (nv == (float)nv)) 556 else if (ecb_expect_false (nv == (float)nv))
341 { 557 {
342 uint32_t fp = ecb_float_to_binary32 (nv); 558 uint32_t fp = ecb_float_to_binary32 (nv);
343 559
344 *enc->cur++ = 0xe0 | 26; 560 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
345 561
346 if (!ecb_big_endian ()) 562 if (!ecb_big_endian ())
347 fp = ecb_bswap32 (fp); 563 fp = ecb_bswap32 (fp);
348 564
349 memcpy (enc->cur, &fp, 4); 565 memcpy (enc->cur, &fp, 4);
351 } 567 }
352 else 568 else
353 { 569 {
354 uint64_t fp = ecb_double_to_binary64 (nv); 570 uint64_t fp = ecb_double_to_binary64 (nv);
355 571
356 *enc->cur++ = 0xe0 | 27; 572 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
357 573
358 if (!ecb_big_endian ()) 574 if (!ecb_big_endian ())
359 fp = ecb_bswap64 (fp); 575 fp = ecb_bswap64 (fp);
360 576
361 memcpy (enc->cur, &fp, 8); 577 memcpy (enc->cur, &fp, 8);
370 586
371 if (SvPOKp (sv)) 587 if (SvPOKp (sv))
372 { 588 {
373 STRLEN len; 589 STRLEN len;
374 char *str = SvPV (sv, len); 590 char *str = SvPV (sv, len);
375 encode_str (enc, SvUTF8 (sv), str, len); 591 encode_strref (enc, SvUTF8 (sv), str, len);
376 } 592 }
377 else if (SvNOKp (sv)) 593 else if (SvNOKp (sv))
378 encode_nv (enc, sv); 594 encode_nv (enc, sv);
379 else if (SvIOKp (sv)) 595 else if (SvIOKp (sv))
380 { 596 {
381 if (SvIsUV (sv)) 597 if (SvIsUV (sv))
382 encode_uint (enc, 0x00, SvUVX (sv)); 598 encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
383 else if (SvIVX (sv) >= 0) 599 else if (SvIVX (sv) >= 0)
384 encode_uint (enc, 0x00, SvIVX (sv)); 600 encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
385 else 601 else
386 encode_uint (enc, 0x20, -(SvIVX (sv) + 1)); 602 encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
387 } 603 }
388 else if (SvROK (sv)) 604 else if (SvROK (sv))
389 encode_rv (enc, SvRV (sv)); 605 encode_rv (enc, SvRV (sv));
390 else if (!SvOK (sv)) 606 else if (!SvOK (sv))
391 encode_ch (enc, 0xe0 | 22); 607 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
392 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 608 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
393 encode_ch (enc, 0xe0 | 23); 609 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
394 else 610 else
395 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",
396 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 612 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
397} 613}
398 614
399static SV * 615static SV *
400encode_cbor (SV *scalar, CBOR *cbor) 616encode_cbor (SV *scalar, CBOR *cbor)
401{ 617{
402 enc_t enc; 618 enc_t enc = { 0 };
403 619
404 enc.cbor = *cbor; 620 enc.cbor = *cbor;
405 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 621 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
406 enc.cur = SvPVX (enc.sv); 622 enc.cur = SvPVX (enc.sv);
407 enc.end = SvEND (enc.sv); 623 enc.end = SvEND (enc.sv);
408 enc.depth = 0;
409 624
410 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
411 encode_sv (&enc, scalar); 634 encode_sv (&enc, scalar);
412 635
413 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv)); 636 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv));
414 *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
415 638
429 U8 *end; // end of input string 652 U8 *end; // end of input string
430 const char *err; // parse error, if != 0 653 const char *err; // parse error, if != 0
431 CBOR cbor; 654 CBOR cbor;
432 U32 depth; // recursion depth 655 U32 depth; // recursion depth
433 U32 maxdepth; // recursion depth limit 656 U32 maxdepth; // recursion depth limit
657 AV *shareable;
658 AV *stringref;
659 SV *decode_tagged;
434} dec_t; 660} dec_t;
435 661
436#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
437 663
438#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")
439 665
440#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)
441#define DEC_DEC_DEPTH --dec->depth 667#define DEC_DEC_DEPTH --dec->depth
442 668
443static UV 669static UV
444decode_uint (dec_t *dec) 670decode_uint (dec_t *dec)
445{ 671{
446 switch (*dec->cur & 31) 672 U8 m = *dec->cur & MINOR_MASK;
447 { 673 ++dec->cur;
448 case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
449 case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
450 case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
451 return *dec->cur++ & 31;
452 674
453 case 24: 675 if (ecb_expect_true (m < LENGTH_EXT1))
676 return m;
677 else if (ecb_expect_true (m == LENGTH_EXT1))
678 {
454 WANT (2); 679 WANT (1);
455 dec->cur += 2; 680 dec->cur += 1;
456 return dec->cur[-1]; 681 return dec->cur[-1];
457 682 }
458 case 25: 683 else if (ecb_expect_true (m == LENGTH_EXT2))
684 {
459 WANT (3); 685 WANT (2);
460 dec->cur += 3; 686 dec->cur += 2;
461 return (((UV)dec->cur[-2]) << 8) 687 return (((UV)dec->cur[-2]) << 8)
462 | ((UV)dec->cur[-1]); 688 | ((UV)dec->cur[-1]);
463 689 }
464 case 26: 690 else if (ecb_expect_true (m == LENGTH_EXT4))
691 {
465 WANT (5); 692 WANT (4);
466 dec->cur += 5; 693 dec->cur += 4;
467 return (((UV)dec->cur[-4]) << 24) 694 return (((UV)dec->cur[-4]) << 24)
468 | (((UV)dec->cur[-3]) << 16) 695 | (((UV)dec->cur[-3]) << 16)
469 | (((UV)dec->cur[-2]) << 8) 696 | (((UV)dec->cur[-2]) << 8)
470 | ((UV)dec->cur[-1]); 697 | ((UV)dec->cur[-1]);
471 698 }
472 case 27: 699 else if (ecb_expect_true (m == LENGTH_EXT8))
700 {
473 WANT (9); 701 WANT (8);
474 dec->cur += 9; 702 dec->cur += 8;
703
704 return
705#if UVSIZE < 8
706 0
707#else
475 return (((UV)dec->cur[-8]) << 56) 708 (((UV)dec->cur[-8]) << 56)
476 | (((UV)dec->cur[-7]) << 48) 709 | (((UV)dec->cur[-7]) << 48)
477 | (((UV)dec->cur[-6]) << 40) 710 | (((UV)dec->cur[-6]) << 40)
478 | (((UV)dec->cur[-5]) << 32) 711 | (((UV)dec->cur[-5]) << 32)
712#endif
479 | (((UV)dec->cur[-4]) << 24) 713 | (((UV)dec->cur[-4]) << 24)
480 | (((UV)dec->cur[-3]) << 16) 714 | (((UV)dec->cur[-3]) << 16)
481 | (((UV)dec->cur[-2]) << 8) 715 | (((UV)dec->cur[-2]) << 8)
482 | ((UV)dec->cur[-1]); 716 | ((UV)dec->cur[-1]);
483 717 }
484 default: 718 else
485 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 719 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
486 }
487 720
488fail: 721fail:
489 return 0; 722 return 0;
490} 723}
491 724
496{ 729{
497 AV *av = newAV (); 730 AV *av = newAV ();
498 731
499 DEC_INC_DEPTH; 732 DEC_INC_DEPTH;
500 733
501 if ((*dec->cur & 31) == 31) 734 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
502 { 735 {
503 ++dec->cur; 736 ++dec->cur;
504 737
505 for (;;) 738 for (;;)
506 { 739 {
507 WANT (1); 740 WANT (1);
508 741
509 if (*dec->cur == 0xe0 | 31) 742 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
510 { 743 {
511 ++dec->cur; 744 ++dec->cur;
512 break; 745 break;
513 } 746 }
514 747
517 } 750 }
518 else 751 else
519 { 752 {
520 int i, len = decode_uint (dec); 753 int i, len = decode_uint (dec);
521 754
755 WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
522 av_fill (av, len - 1); 756 av_fill (av, len - 1);
523 757
524 for (i = 0; i < len; ++i) 758 for (i = 0; i < len; ++i)
525 AvARRAY (av)[i] = decode_sv (dec); 759 AvARRAY (av)[i] = decode_sv (dec);
526 } 760 }
532 SvREFCNT_dec (av); 766 SvREFCNT_dec (av);
533 DEC_DEC_DEPTH; 767 DEC_DEC_DEPTH;
534 return &PL_sv_undef; 768 return &PL_sv_undef;
535} 769}
536 770
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
537static SV * 817static SV *
538decode_hv (dec_t *dec) 818decode_hv (dec_t *dec)
539{ 819{
540 HV *hv = newHV (); 820 HV *hv = newHV ();
541 821
542 DEC_INC_DEPTH; 822 DEC_INC_DEPTH;
543 823
544 if ((*dec->cur & 31) == 31) 824 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
545 { 825 {
546 ++dec->cur; 826 ++dec->cur;
547 827
548 for (;;) 828 for (;;)
549 { 829 {
550 WANT (1); 830 WANT (1);
551 831
552 if (*dec->cur == 0xe0 | 31) 832 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
553 { 833 {
554 ++dec->cur; 834 ++dec->cur;
555 break; 835 break;
556 } 836 }
557 837
558 SV *k = decode_sv (dec); 838 decode_he (dec, hv);
559 SV *v = decode_sv (dec);
560
561 hv_store_ent (hv, k, v, 0);
562 } 839 }
563 } 840 }
564 else 841 else
565 { 842 {
566 int len = decode_uint (dec); 843 int pairs = decode_uint (dec);
567 844
568 while (len--) 845 while (pairs--)
569 { 846 decode_he (dec, hv);
570 SV *k = decode_sv (dec);
571 SV *v = decode_sv (dec);
572
573 hv_store_ent (hv, k, v, 0);
574 }
575 } 847 }
576 848
577 DEC_DEC_DEPTH; 849 DEC_DEC_DEPTH;
578 return newRV_noinc ((SV *)hv); 850 return newRV_noinc ((SV *)hv);
579
580#if 0
581 SV *sv;
582 HV *hv = newHV ();
583
584 DEC_INC_DEPTH;
585 decode_ws (dec);
586
587 for (;;)
588 {
589 // heuristic: assume that
590 // a) decode_str + hv_store_ent are abysmally slow.
591 // b) most hash keys are short, simple ascii text.
592 // => try to "fast-match" such strings to avoid
593 // the overhead of decode_str + hv_store_ent.
594 {
595 SV *value;
596 char *p = dec->cur;
597 char *e = p + 24; // only try up to 24 bytes
598
599 for (;;)
600 {
601 // the >= 0x80 is false on most architectures
602 if (p == e || *p < 0x20 || *p >= 0x80 || *p == '\\')
603 {
604 // slow path, back up and use decode_str
605 SV *key = decode_str (dec);
606 if (!key)
607 goto fail;
608
609 decode_ws (dec); EXPECT_CH (':');
610
611 decode_ws (dec);
612 value = decode_sv (dec);
613 if (!value)
614 {
615 SvREFCNT_dec (key);
616 goto fail;
617 }
618
619 hv_store_ent (hv, key, value, 0);
620 SvREFCNT_dec (key);
621
622 break;
623 }
624 else if (*p == '"')
625 {
626 // fast path, got a simple key
627 char *key = dec->cur;
628 int len = p - key;
629 dec->cur = p + 1;
630
631 decode_ws (dec); EXPECT_CH (':');
632
633 decode_ws (dec);
634 value = decode_sv (dec);
635 if (!value)
636 goto fail;
637
638 hv_store (hv, key, len, value, 0);
639
640 break;
641 }
642
643 ++p;
644 }
645 }
646
647 decode_ws (dec);
648
649 if (*dec->cur == '}')
650 {
651 ++dec->cur;
652 break;
653 }
654
655 if (*dec->cur != ',')
656 ERR (", or } expected while parsing object/hash");
657
658 ++dec->cur;
659
660 decode_ws (dec);
661
662 if (*dec->cur == '}' && dec->cbor.flags & F_RELAXED)
663 {
664 ++dec->cur;
665 break;
666 }
667 }
668
669 DEC_DEC_DEPTH;
670 sv = newRV_noinc ((SV *)hv);
671
672 // check filter callbacks
673 if (dec->cbor.flags & F_HOOK)
674 {
675 if (dec->cbor.cb_sk_object && HvKEYS (hv) == 1)
676 {
677 HE *cb, *he;
678
679 hv_iterinit (hv);
680 he = hv_iternext (hv);
681 hv_iterinit (hv);
682
683 // the next line creates a mortal sv each time its called.
684 // might want to optimise this for common cases.
685 cb = hv_fetch_ent (dec->cbor.cb_sk_object, hv_iterkeysv (he), 0, 0);
686
687 if (cb)
688 {
689 dSP;
690 int count;
691
692 ENTER; SAVETMPS; PUSHMARK (SP);
693 XPUSHs (HeVAL (he));
694 sv_2mortal (sv);
695
696 PUTBACK; count = call_sv (HeVAL (cb), G_ARRAY); SPAGAIN;
697
698 if (count == 1)
699 {
700 sv = newSVsv (POPs);
701 FREETMPS; LEAVE;
702 return sv;
703 }
704
705 SvREFCNT_inc (sv);
706 FREETMPS; LEAVE;
707 }
708 }
709
710 if (dec->cbor.cb_object)
711 {
712 dSP;
713 int count;
714
715 ENTER; SAVETMPS; PUSHMARK (SP);
716 XPUSHs (sv_2mortal (sv));
717
718 PUTBACK; count = call_sv (dec->cbor.cb_object, G_ARRAY); SPAGAIN;
719
720 if (count == 1)
721 {
722 sv = newSVsv (POPs);
723 FREETMPS; LEAVE;
724 return sv;
725 }
726
727 SvREFCNT_inc (sv);
728 FREETMPS; LEAVE;
729 }
730 }
731
732 return sv;
733#endif
734 851
735fail: 852fail:
736 SvREFCNT_dec (hv); 853 SvREFCNT_dec (hv);
737 DEC_DEC_DEPTH; 854 DEC_DEC_DEPTH;
738 return &PL_sv_undef; 855 return &PL_sv_undef;
739} 856}
740 857
741static SV * 858static SV *
742decode_str (dec_t *dec, int utf8) 859decode_str (dec_t *dec, int utf8)
743{ 860{
744 SV *sv; 861 SV *sv = 0;
745 862
746 if ((*dec->cur & 31) == 31) 863 if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
747 { 864 {
865 // indefinite length strings
748 ++dec->cur; 866 ++dec->cur;
749 867
868 U8 major = *dec->cur & MAJOR_MISC;
869
750 sv = newSVpvn ("", 0); 870 sv = newSVpvn ("", 0);
751 871
752 // not very fast, and certainly not robust against illegal input
753 for (;;) 872 for (;;)
754 { 873 {
755 WANT (1); 874 WANT (1);
756 875
757 if (*dec->cur == 0xe0 | 31) 876 if ((*dec->cur - major) > LENGTH_EXT8)
877 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
758 { 878 {
759 ++dec->cur; 879 ++dec->cur;
760 break; 880 break;
761 } 881 }
882 else
883 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
762 884
763 SV *sv2 = decode_sv (dec); 885 STRLEN len = decode_uint (dec);
764 sv_catsv (sv, sv2); 886
887 WANT (len);
888 sv_catpvn (sv, dec->cur, len);
889 dec->cur += len;
765 } 890 }
766 } 891 }
767 else 892 else
768 { 893 {
769 STRLEN len = decode_uint (dec); 894 STRLEN len = decode_uint (dec);
770 895
771 WANT (len); 896 WANT (len);
772 sv = newSVpvn (dec->cur, len); 897 sv = newSVpvn (dec->cur, len);
773 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));
774 } 903 }
775 904
776 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
777 SvUTF8_on (sv); 911 SvUTF8_on (sv);
912 }
778 913
779 return sv; 914 return sv;
780 915
781fail: 916fail:
917 SvREFCNT_dec (sv);
918 return &PL_sv_undef;
919}
920
921static SV *
922decode_tagged (dec_t *dec)
923{
924 SV *sv = 0;
925 UV tag = decode_uint (dec);
926
927 WANT (1);
928
929 switch (tag)
930 {
931 case CBOR_TAG_MAGIC:
932 sv = decode_sv (dec);
933 break;
934
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
1105 return sv;
1106
1107fail:
1108 SvREFCNT_dec (sv);
782 return &PL_sv_undef; 1109 return &PL_sv_undef;
783} 1110}
784 1111
785static SV * 1112static SV *
786decode_sv (dec_t *dec) 1113decode_sv (dec_t *dec)
787{ 1114{
788 WANT (1); 1115 WANT (1);
789 1116
790 switch (*dec->cur >> 5) 1117 switch (*dec->cur >> MAJOR_SHIFT)
791 { 1118 {
792 case 0: // unsigned int 1119 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
793 //TODO: 64 bit values on 3 2bit perls 1120 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
794 return newSVuv (decode_uint (dec)); 1121 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
795 case 1: // negative int 1122 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
796 return newSViv (-1 - (IV)decode_uint (dec)); 1123 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
797 case 2: // octet string 1124 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
798 return decode_str (dec, 0); 1125 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
799 case 3: // utf-8 string 1126
800 return decode_str (dec, 1); 1127 case MAJOR_MISC >> MAJOR_SHIFT:
801 case 4: // array
802 return decode_av (dec);
803 case 5: // map
804 return decode_hv (dec);
805 case 6: // tag
806 abort ();
807 break;
808 case 7: // misc
809 switch (*dec->cur++ & 31) 1128 switch (*dec->cur++ & MINOR_MASK)
810 { 1129 {
811 case 20: 1130 case SIMPLE_FALSE:
812#if CBOR_SLOW 1131#if CBOR_SLOW
813 cbor_false = get_bool ("CBOR::XS::false"); 1132 types_false = get_bool ("Types::Serialiser::false");
814#endif 1133#endif
815 return newSVsv (cbor_false); 1134 return newSVsv (types_false);
816 case 21: 1135 case SIMPLE_TRUE:
817#if CBOR_SLOW 1136#if CBOR_SLOW
818 cbor_true = get_bool ("CBOR::XS::true"); 1137 types_true = get_bool ("Types::Serialiser::true");
819#endif 1138#endif
820 return newSVsv (cbor_true); 1139 return newSVsv (types_true);
821 case 22: 1140 case SIMPLE_NULL:
822 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);
823 1147
824 case 25: 1148 case MISC_FLOAT16:
825 // half float
826 abort ();
827 break; 1149 {
1150 WANT (2);
828 1151
1152 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
1153 dec->cur += 2;
1154
1155 return newSVnv (ecb_binary16_to_float (fp));
1156 }
1157
829 case 26: 1158 case MISC_FLOAT32:
830 { 1159 {
831 uint32_t fp; 1160 uint32_t fp;
832 WANT (4); 1161 WANT (4);
833 memcpy (&fp, dec->cur, 4); 1162 memcpy (&fp, dec->cur, 4);
834 dec->cur += 4; 1163 dec->cur += 4;
837 fp = ecb_bswap32 (fp); 1166 fp = ecb_bswap32 (fp);
838 1167
839 return newSVnv (ecb_binary32_to_float (fp)); 1168 return newSVnv (ecb_binary32_to_float (fp));
840 } 1169 }
841 1170
842 case 27: 1171 case MISC_FLOAT64:
843 { 1172 {
844 uint64_t fp; 1173 uint64_t fp;
845 WANT (8); 1174 WANT (8);
846 memcpy (&fp, dec->cur, 8); 1175 memcpy (&fp, dec->cur, 8);
847 dec->cur += 8; 1176 dec->cur += 8;
850 fp = ecb_bswap64 (fp); 1179 fp = ecb_bswap64 (fp);
851 1180
852 return newSVnv (ecb_binary64_to_double (fp)); 1181 return newSVnv (ecb_binary64_to_double (fp));
853 } 1182 }
854 1183
855 // 0..19 unassigned 1184 // 0..19 unassigned simple
856 // 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
857 default: 1188 default:
858 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1189 ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
859 } 1190 }
860 1191
861 break; 1192 break;
862 } 1193 }
863#if 0
864 switch (*dec->cur)
865 {
866 //case '"': ++dec->cur; return decode_str (dec);
867 case '[': ++dec->cur; return decode_av (dec);
868 case '{': ++dec->cur; return decode_hv (dec);
869
870 case '-':
871 case '0': case '1': case '2': case '3': case '4':
872 case '5': case '6': case '7': case '8': case '9':
873 //TODO return decode_num (dec);
874
875 case 't':
876 if (dec->end - dec->cur >= 4 && !memcmp (dec->cur, "true", 4))
877 {
878 dec->cur += 4;
879#if CBOR_SLOW
880 cbor_true = get_bool ("CBOR::XS::true");
881#endif
882 return newSVsv (cbor_true);
883 }
884 else
885 ERR ("'true' expected");
886
887 break;
888
889 case 'f':
890 if (dec->end - dec->cur >= 5 && !memcmp (dec->cur, "false", 5))
891 {
892 dec->cur += 5;
893#if CBOR_SLOW
894 cbor_false = get_bool ("CBOR::XS::false");
895#endif
896 return newSVsv (cbor_false);
897 }
898 else
899 ERR ("'false' expected");
900
901 break;
902
903 case 'n':
904 if (dec->end - dec->cur >= 4 && !memcmp (dec->cur, "null", 4))
905 {
906 dec->cur += 4;
907 return newSVsv (&PL_sv_undef);
908 }
909 else
910 ERR ("'null' expected");
911
912 break;
913
914 default:
915 ERR ("malformed CBOR string, neither array, object, number, string or atom");
916 break;
917 }
918#endif
919 1194
920fail: 1195fail:
921 return &PL_sv_undef; 1196 return &PL_sv_undef;
922} 1197}
923 1198
924static SV * 1199static SV *
925decode_cbor (SV *string, CBOR *cbor, char **offset_return) 1200decode_cbor (SV *string, CBOR *cbor, char **offset_return)
926{ 1201{
927 dec_t dec; 1202 dec_t dec = { 0 };
928 SV *sv; 1203 SV *sv;
1204 STRLEN len;
1205 char *data = SvPVbyte (string, len);
929 1206
930 /* work around bugs in 5.10 where manipulating magic values
931 * makes perl ignore the magic in subsequent accesses.
932 * also make a copy of non-PV values, to get them into a clean
933 * state (SvPV should do that, but it's buggy, see below).
934 */
935 /*SvGETMAGIC (string);*/
936 if (SvMAGICAL (string) || !SvPOK (string))
937 string = sv_2mortal (newSVsv (string));
938
939 SvUPGRADE (string, SVt_PV);
940
941 /* work around a bug in perl 5.10, which causes SvCUR to fail an
942 * assertion with -DDEBUGGING, although SvCUR is documented to
943 * return the xpv_cur field which certainly exists after upgrading.
944 * according to nicholas clark, calling SvPOK fixes this.
945 * But it doesn't fix it, so try another workaround, call SvPV_nolen
946 * and hope for the best.
947 * Damnit, SvPV_nolen still trips over yet another assertion. This
948 * assertion business is seriously broken, try yet another workaround
949 * for the broken -DDEBUGGING.
950 */
951 {
952#ifdef DEBUGGING
953 STRLEN offset = SvOK (string) ? sv_len (string) : 0;
954#else
955 STRLEN offset = SvCUR (string);
956#endif
957
958 if (offset > cbor->max_size && cbor->max_size) 1207 if (len > cbor->max_size && cbor->max_size)
959 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",
960 (unsigned long)SvCUR (string), (unsigned long)cbor->max_size); 1209 (unsigned long)len, (unsigned long)cbor->max_size);
961 }
962
963 sv_utf8_downgrade (string, 0);
964 1210
965 dec.cbor = *cbor; 1211 dec.cbor = *cbor;
966 dec.cur = (U8 *)SvPVX (string); 1212 dec.cur = (U8 *)data;
967 dec.end = (U8 *)SvEND (string); 1213 dec.end = (U8 *)data + len;
968 dec.err = 0;
969 dec.depth = 0;
970
971 if (dec.cbor.cb_object || dec.cbor.cb_sk_object)
972 ;//TODO dec.cbor.flags |= F_HOOK;
973 1214
974 sv = decode_sv (&dec); 1215 sv = decode_sv (&dec);
975 1216
976 if (offset_return) 1217 if (offset_return)
977 *offset_return = dec.cur; 1218 *offset_return = dec.cur;
978 1219
979 if (!(offset_return || !sv)) 1220 if (!(offset_return || !sv))
980 {
981 if (*dec.cur && !dec.err) 1221 if (dec.cur != dec.end && !dec.err)
982 {
983 dec.err = "garbage after CBOR object"; 1222 dec.err = "garbage after CBOR object";
1223
1224 if (dec.err)
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
984 SvREFCNT_dec (sv); 1237 SvREFCNT_dec (sv);
985 sv = 0;
986 }
987 }
988
989 if (!sv)
990 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);
1239 }
991 1240
992 sv = sv_2mortal (sv); 1241 sv = sv_2mortal (sv);
993 1242
994 return sv; 1243 return sv;
995} 1244}
996 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
997///////////////////////////////////////////////////////////////////////////// 1370/////////////////////////////////////////////////////////////////////////////
998// XS interface functions 1371// XS interface functions
999 1372
1000MODULE = CBOR::XS PACKAGE = CBOR::XS 1373MODULE = CBOR::XS PACKAGE = CBOR::XS
1001 1374
1002BOOT: 1375BOOT:
1003{ 1376{
1004 cbor_stash = gv_stashpv ("CBOR::XS" , 1); 1377 cbor_stash = gv_stashpv ("CBOR::XS" , 1);
1005 cbor_boolean_stash = gv_stashpv ("CBOR::XS::Boolean", 1); 1378 cbor_tagged_stash = gv_stashpv ("CBOR::XS::Tagged" , 1);
1006 1379
1007 cbor_true = get_bool ("CBOR::XS::true"); 1380 types_boolean_stash = gv_stashpv ("Types::Serialiser::Boolean", 1);
1008 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);
1009} 1391}
1010 1392
1011PROTOTYPES: DISABLE 1393PROTOTYPES: DISABLE
1012 1394
1013void CLONE (...) 1395void CLONE (...)
1014 CODE: 1396 CODE:
1015 cbor_stash = 0; 1397 cbor_stash = 0;
1398 cbor_tagged_stash = 0;
1399 types_error_stash = 0;
1016 cbor_boolean_stash = 0; 1400 types_boolean_stash = 0;
1017 1401
1018void new (char *klass) 1402void new (char *klass)
1019 PPCODE: 1403 PPCODE:
1020{ 1404{
1021 SV *pv = NEWSV (0, sizeof (CBOR)); 1405 SV *pv = NEWSV (0, sizeof (CBOR));
1029 1413
1030void shrink (CBOR *self, int enable = 1) 1414void shrink (CBOR *self, int enable = 1)
1031 ALIAS: 1415 ALIAS:
1032 shrink = F_SHRINK 1416 shrink = F_SHRINK
1033 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
1034 PPCODE: 1423 PPCODE:
1035{ 1424{
1036 if (enable) 1425 if (enable)
1037 self->flags |= ix; 1426 self->flags |= ix;
1038 else 1427 else
1043 1432
1044void get_shrink (CBOR *self) 1433void get_shrink (CBOR *self)
1045 ALIAS: 1434 ALIAS:
1046 get_shrink = F_SHRINK 1435 get_shrink = F_SHRINK
1047 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
1048 PPCODE: 1441 PPCODE:
1049 XPUSHs (boolSV (self->flags & ix)); 1442 XPUSHs (boolSV (self->flags & ix));
1050 1443
1051void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1444void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1052 PPCODE: 1445 PPCODE:
1068 CODE: 1461 CODE:
1069 RETVAL = self->max_size; 1462 RETVAL = self->max_size;
1070 OUTPUT: 1463 OUTPUT:
1071 RETVAL 1464 RETVAL
1072 1465
1073#if 0 //TODO 1466void filter (CBOR *self, SV *filter = 0)
1074
1075void filter_cbor_object (CBOR *self, SV *cb = &PL_sv_undef)
1076 PPCODE: 1467 PPCODE:
1077{
1078 SvREFCNT_dec (self->cb_object); 1468 SvREFCNT_dec (self->filter);
1079 self->cb_object = SvOK (cb) ? newSVsv (cb) : 0; 1469 self->filter = filter ? newSVsv (filter) : filter;
1080
1081 XPUSHs (ST (0)); 1470 XPUSHs (ST (0));
1082}
1083 1471
1084void filter_cbor_single_key_object (CBOR *self, SV *key, SV *cb = &PL_sv_undef) 1472SV *get_filter (CBOR *self)
1085 PPCODE: 1473 CODE:
1086{ 1474 RETVAL = self->filter ? self->filter : NEWSV (0, 0);
1087 if (!self->cb_sk_object) 1475 OUTPUT:
1088 self->cb_sk_object = newHV (); 1476 RETVAL
1089
1090 if (SvOK (cb))
1091 hv_store_ent (self->cb_sk_object, key, newSVsv (cb), 0);
1092 else
1093 {
1094 hv_delete_ent (self->cb_sk_object, key, G_DISCARD, 0);
1095
1096 if (!HvKEYS (self->cb_sk_object))
1097 {
1098 SvREFCNT_dec (self->cb_sk_object);
1099 self->cb_sk_object = 0;
1100 }
1101 }
1102
1103 XPUSHs (ST (0));
1104}
1105
1106#endif
1107 1477
1108void encode (CBOR *self, SV *scalar) 1478void encode (CBOR *self, SV *scalar)
1109 PPCODE: 1479 PPCODE:
1110 PUTBACK; scalar = encode_cbor (scalar, self); SPAGAIN; 1480 PUTBACK; scalar = encode_cbor (scalar, self); SPAGAIN;
1111 XPUSHs (scalar); 1481 XPUSHs (scalar);
1124 EXTEND (SP, 2); 1494 EXTEND (SP, 2);
1125 PUSHs (sv); 1495 PUSHs (sv);
1126 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1496 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1127} 1497}
1128 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
1129void DESTROY (CBOR *self) 1551void DESTROY (CBOR *self)
1130 CODE: 1552 PPCODE:
1131 SvREFCNT_dec (self->cb_sk_object); 1553 cbor_free (self);
1132 SvREFCNT_dec (self->cb_object);
1133 1554
1134PROTOTYPES: ENABLE 1555PROTOTYPES: ENABLE
1135 1556
1136void encode_cbor (SV *scalar) 1557void encode_cbor (SV *scalar)
1558 ALIAS:
1559 encode_cbor = 0
1560 encode_cbor_sharing = F_ALLOW_SHARING
1137 PPCODE: 1561 PPCODE:
1138{ 1562{
1139 CBOR cbor; 1563 CBOR cbor;
1140 cbor_init (&cbor); 1564 cbor_init (&cbor);
1565 cbor.flags |= ix;
1141 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1566 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1142 XPUSHs (scalar); 1567 XPUSHs (scalar);
1143} 1568}
1144 1569
1145void decode_cbor (SV *cborstr) 1570void decode_cbor (SV *cborstr)

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines