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 1.48 by root, Sun Sep 20 23:25:53 2015 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_VALIDATE_UTF8 0x00000020UL
16 106
17#define INIT_SIZE 32 // initial scalar size to be allocated 107#define INIT_SIZE 32 // initial scalar size to be allocated
18 108
19#define SB do { 109#define SB do {
20#define SE } while (0) 110#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)
32 111
33#define IN_RANGE_INC(type,val,beg,end) \ 112#define IN_RANGE_INC(type,val,beg,end) \
34 ((unsigned type)((unsigned type)(val) - (unsigned type)(beg)) \ 113 ((unsigned type)((unsigned type)(val) - (unsigned type)(beg)) \
35 <= (unsigned type)((unsigned type)(end) - (unsigned type)(beg))) 114 <= (unsigned type)((unsigned type)(end) - (unsigned type)(beg)))
36 115
42#else 121#else
43# define CBOR_SLOW 0 122# define CBOR_SLOW 0
44# define CBOR_STASH cbor_stash 123# define CBOR_STASH cbor_stash
45#endif 124#endif
46 125
47static HV *cbor_stash, *cbor_boolean_stash; // CBOR::XS:: 126static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS::
48static SV *cbor_true, *cbor_false; 127static SV *types_true, *types_false, *types_error, *sv_cbor, *default_filter;
49 128
50typedef struct { 129typedef struct {
51 U32 flags; 130 U32 flags;
52 U32 max_depth; 131 U32 max_depth;
53 STRLEN max_size; 132 STRLEN max_size;
133 SV *filter;
54 134
55 SV *cb_object; 135 // for the incremental parser
56 HV *cb_sk_object; 136 STRLEN incr_pos; // the current offset into the text
137 STRLEN incr_need; // minimum bytes needed to decode
138 AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef.
57} CBOR; 139} CBOR;
58 140
59INLINE void 141ecb_inline void
60cbor_init (CBOR *cbor) 142cbor_init (CBOR *cbor)
61{ 143{
62 Zero (cbor, 1, CBOR); 144 Zero (cbor, 1, CBOR);
63 cbor->max_depth = 512; 145 cbor->max_depth = 512;
64} 146}
65 147
148ecb_inline void
149cbor_free (CBOR *cbor)
150{
151 SvREFCNT_dec (cbor->filter);
152 SvREFCNT_dec (cbor->incr_count);
153}
154
66///////////////////////////////////////////////////////////////////////////// 155/////////////////////////////////////////////////////////////////////////////
67// utility functions 156// utility functions
68 157
69INLINE SV * 158ecb_inline SV *
70get_bool (const char *name) 159get_bool (const char *name)
71{ 160{
72 SV *sv = get_sv (name, 1); 161 SV *sv = get_sv (name, 1);
73 162
74 SvREADONLY_on (sv); 163 SvREADONLY_on (sv);
75 SvREADONLY_on (SvRV (sv)); 164 SvREADONLY_on (SvRV (sv));
76 165
77 return sv; 166 return sv;
78} 167}
79 168
80INLINE void 169ecb_inline void
81shrink (SV *sv) 170shrink (SV *sv)
82{ 171{
83 sv_utf8_downgrade (sv, 1); 172 sv_utf8_downgrade (sv, 1);
84 173
85 if (SvLEN (sv) > SvCUR (sv) + 1) 174 if (SvLEN (sv) > SvCUR (sv) + 1)
90 SvPV_renew (sv, SvCUR (sv) + 1); 179 SvPV_renew (sv, SvCUR (sv) + 1);
91#endif 180#endif
92 } 181 }
93} 182}
94 183
95///////////////////////////////////////////////////////////////////////////// 184// minimum length of a string to be registered for stringref
96// fp hell 185ecb_inline int
97 186minimum_string_length (UV idx)
98//TODO 187{
188 return idx > 23
189 ? idx > 0xffU
190 ? idx > 0xffffU
191 ? idx > 0xffffffffU
192 ? 11
193 : 7
194 : 5
195 : 4
196 : 3;
197}
99 198
100///////////////////////////////////////////////////////////////////////////// 199/////////////////////////////////////////////////////////////////////////////
101// encoder 200// encoder
102 201
103// structure used for encoding CBOR 202// structure used for encoding CBOR
106 char *cur; // SvPVX (sv) + current output position 205 char *cur; // SvPVX (sv) + current output position
107 char *end; // SvEND (sv) 206 char *end; // SvEND (sv)
108 SV *sv; // result scalar 207 SV *sv; // result scalar
109 CBOR cbor; 208 CBOR cbor;
110 U32 depth; // recursion level 209 U32 depth; // recursion level
210 HV *stringref[2]; // string => index, or 0 ([0] = bytes, [1] = utf-8)
211 UV stringref_idx;
212 HV *shareable; // ptr => index, or 0
213 UV shareable_idx;
111} enc_t; 214} enc_t;
112 215
113INLINE void 216ecb_inline void
114need (enc_t *enc, STRLEN len) 217need (enc_t *enc, STRLEN len)
115{ 218{
116 if (expect_false (enc->cur + len >= enc->end)) 219 if (ecb_expect_false (enc->cur + len >= enc->end))
117 { 220 {
118 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv); 221 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv);
119 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1); 222 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1);
120 enc->cur = SvPVX (enc->sv) + cur; 223 enc->cur = SvPVX (enc->sv) + cur;
121 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1; 224 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;
122 } 225 }
123} 226}
124 227
125INLINE void 228ecb_inline void
126encode_ch (enc_t *enc, char ch) 229encode_ch (enc_t *enc, char ch)
127{ 230{
128 need (enc, 1); 231 need (enc, 1);
129 *enc->cur++ = ch; 232 *enc->cur++ = ch;
130} 233}
132static void 235static void
133encode_uint (enc_t *enc, int major, UV len) 236encode_uint (enc_t *enc, int major, UV len)
134{ 237{
135 need (enc, 9); 238 need (enc, 9);
136 239
137 if (len < 24) 240 if (ecb_expect_true (len < LENGTH_EXT1))
138 *enc->cur++ = major | len; 241 *enc->cur++ = major | len;
139 else if (len < 0x100) 242 else if (ecb_expect_true (len <= 0xffU))
140 { 243 {
141 *enc->cur++ = major | 24; 244 *enc->cur++ = major | LENGTH_EXT1;
142 *enc->cur++ = len; 245 *enc->cur++ = len;
143 } 246 }
144 else if (len < 0x10000) 247 else if (len <= 0xffffU)
145 { 248 {
146 *enc->cur++ = major | 25; 249 *enc->cur++ = major | LENGTH_EXT2;
147 *enc->cur++ = len >> 8; 250 *enc->cur++ = len >> 8;
148 *enc->cur++ = len; 251 *enc->cur++ = len;
149 } 252 }
150 else if (len < 0x100000000) 253 else if (len <= 0xffffffffU)
151 { 254 {
152 *enc->cur++ = major | 26; 255 *enc->cur++ = major | LENGTH_EXT4;
153 *enc->cur++ = len >> 24; 256 *enc->cur++ = len >> 24;
154 *enc->cur++ = len >> 16; 257 *enc->cur++ = len >> 16;
155 *enc->cur++ = len >> 8; 258 *enc->cur++ = len >> 8;
156 *enc->cur++ = len; 259 *enc->cur++ = len;
157 } 260 }
158 else if (len) 261 else
159 { 262 {
160 *enc->cur++ = major | 27; 263 *enc->cur++ = major | LENGTH_EXT8;
161 *enc->cur++ = len >> 56; 264 *enc->cur++ = len >> 56;
162 *enc->cur++ = len >> 48; 265 *enc->cur++ = len >> 48;
163 *enc->cur++ = len >> 40; 266 *enc->cur++ = len >> 40;
164 *enc->cur++ = len >> 32; 267 *enc->cur++ = len >> 32;
165 *enc->cur++ = len >> 24; 268 *enc->cur++ = len >> 24;
167 *enc->cur++ = len >> 8; 270 *enc->cur++ = len >> 8;
168 *enc->cur++ = len; 271 *enc->cur++ = len;
169 } 272 }
170} 273}
171 274
172static void 275ecb_inline void
276encode_tag (enc_t *enc, UV tag)
277{
278 encode_uint (enc, MAJOR_TAG, tag);
279}
280
281ecb_inline void
173encode_str (enc_t *enc, int utf8, char *str, STRLEN len) 282encode_str (enc_t *enc, int utf8, char *str, STRLEN len)
174{ 283{
175 encode_uint (enc, utf8 ? 0x60 : 0x40, len); 284 encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
176 need (enc, len); 285 need (enc, len);
177 memcpy (enc->cur, str, len); 286 memcpy (enc->cur, str, len);
178 enc->cur += len; 287 enc->cur += len;
179} 288}
180 289
290static void
291encode_strref (enc_t *enc, int utf8, char *str, STRLEN len)
292{
293 if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
294 {
295 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
296
297 if (SvOK (*svp))
298 {
299 // already registered, use stringref
300 encode_tag (enc, CBOR_TAG_STRINGREF);
301 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
302 return;
303 }
304 else if (len >= minimum_string_length (enc->stringref_idx))
305 {
306 // register only
307 sv_setuv (*svp, enc->stringref_idx);
308 ++enc->stringref_idx;
309 }
310 }
311
312 encode_str (enc, utf8, str, len);
313}
314
181static void encode_sv (enc_t *enc, SV *sv); 315static void encode_sv (enc_t *enc, SV *sv);
182 316
183static void 317static void
184encode_av (enc_t *enc, AV *av) 318encode_av (enc_t *enc, AV *av)
185{ 319{
188 if (enc->depth >= enc->cbor.max_depth) 322 if (enc->depth >= enc->cbor.max_depth)
189 croak (ERR_NESTING_EXCEEDED); 323 croak (ERR_NESTING_EXCEEDED);
190 324
191 ++enc->depth; 325 ++enc->depth;
192 326
193 encode_uint (enc, 0x80, len + 1); 327 encode_uint (enc, MAJOR_ARRAY, len + 1);
194 328
329 if (SvMAGICAL (av))
195 for (i = 0; i <= len; ++i) 330 for (i = 0; i <= len; ++i)
196 { 331 {
197 SV **svp = av_fetch (av, i, 0); 332 SV **svp = av_fetch (av, i, 0);
198 encode_sv (enc, svp ? *svp : &PL_sv_undef); 333 encode_sv (enc, svp ? *svp : &PL_sv_undef);
199 } 334 }
335 else
336 for (i = 0; i <= len; ++i)
337 {
338 SV *sv = AvARRAY (av)[i];
339 encode_sv (enc, sv ? sv : &PL_sv_undef);
340 }
200 341
201 --enc->depth; 342 --enc->depth;
202} 343}
203 344
204static void 345static void
213 354
214 int pairs = hv_iterinit (hv); 355 int pairs = hv_iterinit (hv);
215 int mg = SvMAGICAL (hv); 356 int mg = SvMAGICAL (hv);
216 357
217 if (mg) 358 if (mg)
218 encode_ch (enc, 0xa0 | 31); 359 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
219 else 360 else
220 encode_uint (enc, 0xa0, pairs); 361 encode_uint (enc, MAJOR_MAP, pairs);
221 362
222 while ((he = hv_iternext (hv))) 363 while ((he = hv_iternext (hv)))
223 { 364 {
224 if (HeKLEN (he) == HEf_SVKEY) 365 if (HeKLEN (he) == HEf_SVKEY)
225 encode_sv (enc, HeSVKEY (he)); 366 encode_sv (enc, HeSVKEY (he));
226 else 367 else
227 encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); 368 encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
228 369
229 encode_sv (enc, expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); 370 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
230 } 371 }
231 372
232 if (mg) 373 if (mg)
233 encode_ch (enc, 0xe0 | 31); 374 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
234 375
235 --enc->depth; 376 --enc->depth;
236} 377}
237 378
238// encode objects, arrays and special \0=false and \1=true values. 379// encode objects, arrays and special \0=false and \1=true values.
239static void 380static void
240encode_rv (enc_t *enc, SV *sv) 381encode_rv (enc_t *enc, SV *sv)
241{ 382{
242 svtype svt;
243
244 SvGETMAGIC (sv); 383 SvGETMAGIC (sv);
384
245 svt = SvTYPE (sv); 385 svtype svt = SvTYPE (sv);
246 386
247 if (expect_false (SvOBJECT (sv))) 387 if (ecb_expect_false (SvOBJECT (sv)))
248 { 388 {
249 HV *stash = !CBOR_SLOW || cbor_boolean_stash 389 HV *boolean_stash = !CBOR_SLOW || types_boolean_stash
250 ? cbor_boolean_stash 390 ? types_boolean_stash
391 : gv_stashpv ("Types::Serialiser::Boolean", 1);
392 HV *error_stash = !CBOR_SLOW || types_error_stash
393 ? types_error_stash
394 : gv_stashpv ("Types::Serialiser::Error", 1);
395 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
396 ? cbor_tagged_stash
251 : gv_stashpv ("CBOR::XS::Boolean", 1); 397 : gv_stashpv ("CBOR::XS::Tagged" , 1);
252 398
253 if (SvSTASH (sv) == stash) 399 HV *stash = SvSTASH (sv);
254 encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20); 400
401 if (stash == boolean_stash)
402 {
403 encode_ch (enc, SvIV (sv) ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE);
404 return;
405 }
406 else if (stash == error_stash)
407 {
408 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
409 return;
410 }
411 else if (stash == tagged_stash)
412 {
413 if (svt != SVt_PVAV)
414 croak ("encountered CBOR::XS::Tagged object that isn't an array");
415
416 encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
417 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1));
418
419 return;
420 }
421 }
422
423 if (ecb_expect_false (SvREFCNT (sv) > 1)
424 && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
425 {
426 if (!enc->shareable)
427 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
428
429 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
430
431 if (SvOK (*svp))
432 {
433 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
434 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
435 return;
436 }
255 else 437 else
256 { 438 {
257#if 0 //TODO 439 sv_setuv (*svp, enc->shareable_idx);
258 if (enc->cbor.flags & F_CONV_BLESSED) 440 ++enc->shareable_idx;
441 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
442 }
443 }
444
445 if (ecb_expect_false (SvOBJECT (sv)))
446 {
447 HV *stash = SvSTASH (sv);
448 GV *method;
449
450 if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
259 { 451 {
452 dSP;
453
454 ENTER; SAVETMPS; PUSHMARK (SP);
260 // we re-bless the reference to get overload and other niceties right 455 // 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))); 456 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
269 457
270 // calling with G_SCALAR ensures that we always get a 1 return value
271 PUTBACK; 458 PUTBACK;
459 // G_SCALAR ensures that return value is 1
272 call_sv ((SV *)GvCV (to_cbor), G_SCALAR); 460 call_sv ((SV *)GvCV (method), G_SCALAR);
273 SPAGAIN; 461 SPAGAIN;
274 462
275 // catch this surprisingly common error 463 // catch this surprisingly common error
276 if (SvROK (TOPs) && SvRV (TOPs) == sv) 464 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))); 465 croak ("%s::TO_CBOR method returned same object as was passed instead of a new one", HvNAME (stash));
278 466
279 sv = POPs;
280 PUTBACK;
281
282 encode_sv (enc, sv); 467 encode_sv (enc, POPs);
283 468
469 PUTBACK;
470
284 FREETMPS; LEAVE; 471 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 } 472 }
292 else if (enc->cbor.flags & F_ALLOW_BLESSED) 473 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
293 encode_str (enc, "null", 4, 0); 474 {
475 dSP;
476
477 ENTER; SAVETMPS; PUSHMARK (SP);
478 EXTEND (SP, 2);
479 // we re-bless the reference to get overload and other niceties right
480 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
481 PUSHs (sv_cbor);
482
483 PUTBACK;
484 int count = call_sv ((SV *)GvCV (method), G_ARRAY);
485 SPAGAIN;
486
487 // catch this surprisingly common error
488 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
489 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
490
491 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
492 encode_uint (enc, MAJOR_ARRAY, count + 1);
493 encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
494
495 while (count)
496 encode_sv (enc, SP[1 - count--]);
497
498 PUTBACK;
499
500 FREETMPS; LEAVE;
501 }
294 else 502 else
295 croak ("encountered object '%s', but neither allow_blessed nor convert_blessed settings are enabled", 503 croak ("encountered object '%s', but no TO_CBOR or FREEZE methods available on it",
296 SvPV_nolen (sv_2mortal (newRV_inc (sv)))); 504 SvPV_nolen (sv_2mortal (newRV_inc (sv))));
297#endif
298 }
299 } 505 }
300 else if (svt == SVt_PVHV) 506 else if (svt == SVt_PVHV)
301 encode_hv (enc, (HV *)sv); 507 encode_hv (enc, (HV *)sv);
302 else if (svt == SVt_PVAV) 508 else if (svt == SVt_PVAV)
303 encode_av (enc, (AV *)sv); 509 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 510 else
322 croak ("encountered %s, but CBOR can only represent references to arrays or hashes", 511 {
323 SvPV_nolen (sv_2mortal (newRV_inc (sv)))); 512 encode_tag (enc, CBOR_TAG_INDIRECTION);
513 encode_sv (enc, sv);
514 }
324} 515}
325 516
326static void 517static void
327encode_nv (enc_t *enc, SV *sv) 518encode_nv (enc_t *enc, SV *sv)
328{ 519{
329 double nv = SvNVX (sv); 520 double nv = SvNVX (sv);
330 521
331 need (enc, 9); 522 need (enc, 9);
332 523
333 if (expect_false (nv == (U32)nv)) 524 if (ecb_expect_false (nv == (NV)(U32)nv))
334 encode_uint (enc, 0x00, (U32)nv); 525 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
335 //TODO: maybe I32? 526 //TODO: maybe I32?
336 else if (expect_false (nv == (float)nv)) 527 else if (ecb_expect_false (nv == (float)nv))
337 { 528 {
338 uint32_t fp = ecb_float_to_binary32 (nv); 529 uint32_t fp = ecb_float_to_binary32 (nv);
339 530
340 *enc->cur++ = 0xe0 | 26; 531 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
341 532
342 if (!ecb_big_endian ()) 533 if (!ecb_big_endian ())
343 fp = ecb_bswap32 (fp); 534 fp = ecb_bswap32 (fp);
344 535
345 memcpy (enc->cur, &fp, 4); 536 memcpy (enc->cur, &fp, 4);
347 } 538 }
348 else 539 else
349 { 540 {
350 uint64_t fp = ecb_double_to_binary64 (nv); 541 uint64_t fp = ecb_double_to_binary64 (nv);
351 542
352 *enc->cur++ = 0xe0 | 27; 543 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
353 544
354 if (!ecb_big_endian ()) 545 if (!ecb_big_endian ())
355 fp = ecb_bswap64 (fp); 546 fp = ecb_bswap64 (fp);
356 547
357 memcpy (enc->cur, &fp, 8); 548 memcpy (enc->cur, &fp, 8);
366 557
367 if (SvPOKp (sv)) 558 if (SvPOKp (sv))
368 { 559 {
369 STRLEN len; 560 STRLEN len;
370 char *str = SvPV (sv, len); 561 char *str = SvPV (sv, len);
371 encode_str (enc, SvUTF8 (sv), str, len); 562 encode_strref (enc, SvUTF8 (sv), str, len);
372 } 563 }
373 else if (SvNOKp (sv)) 564 else if (SvNOKp (sv))
374 encode_nv (enc, sv); 565 encode_nv (enc, sv);
375 else if (SvIOKp (sv)) 566 else if (SvIOKp (sv))
376 { 567 {
377 if (SvIsUV (sv)) 568 if (SvIsUV (sv))
378 encode_uint (enc, 0x00, SvUVX (sv)); 569 encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
379 else if (SvIVX (sv) >= 0) 570 else if (SvIVX (sv) >= 0)
380 encode_uint (enc, 0x00, SvIVX (sv)); 571 encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
381 else 572 else
382 encode_uint (enc, 0x20, -(SvIVX (sv) + 1)); 573 encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
383 } 574 }
384 else if (SvROK (sv)) 575 else if (SvROK (sv))
385 encode_rv (enc, SvRV (sv)); 576 encode_rv (enc, SvRV (sv));
386 else if (!SvOK (sv)) 577 else if (!SvOK (sv))
387 encode_ch (enc, 0xe0 | 22); 578 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
388 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 579 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
389 encode_ch (enc, 0xe0 | 23); 580 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
390 else 581 else
391 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data", 582 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
392 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 583 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
393} 584}
394 585
395static SV * 586static SV *
396encode_cbor (SV *scalar, CBOR *cbor) 587encode_cbor (SV *scalar, CBOR *cbor)
397{ 588{
398 enc_t enc; 589 enc_t enc = { 0 };
399 590
400 enc.cbor = *cbor; 591 enc.cbor = *cbor;
401 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 592 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
402 enc.cur = SvPVX (enc.sv); 593 enc.cur = SvPVX (enc.sv);
403 enc.end = SvEND (enc.sv); 594 enc.end = SvEND (enc.sv);
404 enc.depth = 0;
405 595
406 SvPOK_only (enc.sv); 596 SvPOK_only (enc.sv);
597
598 if (cbor->flags & F_PACK_STRINGS)
599 {
600 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
601 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
602 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
603 }
604
407 encode_sv (&enc, scalar); 605 encode_sv (&enc, scalar);
408 606
409 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv)); 607 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv));
410 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings 608 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings
411 609
425 U8 *end; // end of input string 623 U8 *end; // end of input string
426 const char *err; // parse error, if != 0 624 const char *err; // parse error, if != 0
427 CBOR cbor; 625 CBOR cbor;
428 U32 depth; // recursion depth 626 U32 depth; // recursion depth
429 U32 maxdepth; // recursion depth limit 627 U32 maxdepth; // recursion depth limit
628 AV *shareable;
629 AV *stringref;
630 SV *decode_tagged;
430} dec_t; 631} dec_t;
431 632
432#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE 633#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE
433 634
434#define WANT(len) if (expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data"); 635#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data")
435 636
436#define DEC_INC_DEPTH if (++dec->depth > dec->cbor.max_depth) ERR (ERR_NESTING_EXCEEDED) 637#define DEC_INC_DEPTH if (++dec->depth > dec->cbor.max_depth) ERR (ERR_NESTING_EXCEEDED)
437#define DEC_DEC_DEPTH --dec->depth 638#define DEC_DEC_DEPTH --dec->depth
438 639
439static UV 640static UV
440decode_uint (dec_t *dec) 641decode_uint (dec_t *dec)
441{ 642{
442 switch (*dec->cur & 31) 643 U8 m = *dec->cur & MINOR_MASK;
443 { 644 ++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;
448 645
449 case 24: 646 if (ecb_expect_true (m < LENGTH_EXT1))
647 return m;
648 else if (ecb_expect_true (m == LENGTH_EXT1))
649 {
450 WANT (2); 650 WANT (1);
451 dec->cur += 2; 651 dec->cur += 1;
452 return dec->cur[-1]; 652 return dec->cur[-1];
453 653 }
454 case 25: 654 else if (ecb_expect_true (m == LENGTH_EXT2))
655 {
455 WANT (3); 656 WANT (2);
456 dec->cur += 3; 657 dec->cur += 2;
457 return (((UV)dec->cur[-2]) << 8) 658 return (((UV)dec->cur[-2]) << 8)
458 | ((UV)dec->cur[-1]); 659 | ((UV)dec->cur[-1]);
459 660 }
460 case 26: 661 else if (ecb_expect_true (m == LENGTH_EXT4))
662 {
461 WANT (5); 663 WANT (4);
462 dec->cur += 5; 664 dec->cur += 4;
463 return (((UV)dec->cur[-4]) << 24) 665 return (((UV)dec->cur[-4]) << 24)
464 | (((UV)dec->cur[-3]) << 16) 666 | (((UV)dec->cur[-3]) << 16)
465 | (((UV)dec->cur[-2]) << 8) 667 | (((UV)dec->cur[-2]) << 8)
466 | ((UV)dec->cur[-1]); 668 | ((UV)dec->cur[-1]);
467 669 }
468 case 27: 670 else if (ecb_expect_true (m == LENGTH_EXT8))
671 {
469 WANT (9); 672 WANT (8);
470 dec->cur += 9; 673 dec->cur += 8;
674
675 return
676#if UVSIZE < 8
677 0
678#else
471 return (((UV)dec->cur[-8]) << 56) 679 (((UV)dec->cur[-8]) << 56)
472 | (((UV)dec->cur[-7]) << 48) 680 | (((UV)dec->cur[-7]) << 48)
473 | (((UV)dec->cur[-6]) << 40) 681 | (((UV)dec->cur[-6]) << 40)
474 | (((UV)dec->cur[-5]) << 32) 682 | (((UV)dec->cur[-5]) << 32)
683#endif
475 | (((UV)dec->cur[-4]) << 24) 684 | (((UV)dec->cur[-4]) << 24)
476 | (((UV)dec->cur[-3]) << 16) 685 | (((UV)dec->cur[-3]) << 16)
477 | (((UV)dec->cur[-2]) << 8) 686 | (((UV)dec->cur[-2]) << 8)
478 | ((UV)dec->cur[-1]); 687 | ((UV)dec->cur[-1]);
479 688 }
480 default: 689 else
481 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 690 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
482 }
483 691
484fail: 692fail:
485 return 0; 693 return 0;
486} 694}
487 695
492{ 700{
493 AV *av = newAV (); 701 AV *av = newAV ();
494 702
495 DEC_INC_DEPTH; 703 DEC_INC_DEPTH;
496 704
497 if ((*dec->cur & 31) == 31) 705 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
498 { 706 {
499 ++dec->cur; 707 ++dec->cur;
500 708
501 for (;;) 709 for (;;)
502 { 710 {
503 WANT (1); 711 WANT (1);
504 712
505 if (*dec->cur == (0xe0 | 31)) 713 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
506 { 714 {
507 ++dec->cur; 715 ++dec->cur;
508 break; 716 break;
509 } 717 }
510 718
513 } 721 }
514 else 722 else
515 { 723 {
516 int i, len = decode_uint (dec); 724 int i, len = decode_uint (dec);
517 725
726 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); 727 av_fill (av, len - 1);
519 728
520 for (i = 0; i < len; ++i) 729 for (i = 0; i < len; ++i)
521 AvARRAY (av)[i] = decode_sv (dec); 730 AvARRAY (av)[i] = decode_sv (dec);
522 } 731 }
528 SvREFCNT_dec (av); 737 SvREFCNT_dec (av);
529 DEC_DEC_DEPTH; 738 DEC_DEC_DEPTH;
530 return &PL_sv_undef; 739 return &PL_sv_undef;
531} 740}
532 741
742static void
743decode_he (dec_t *dec, HV *hv)
744{
745 // for speed reasons, we specialcase single-string
746 // byte or utf-8 strings as keys, but only when !stringref
747
748 if (ecb_expect_true (!dec->stringref))
749 if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
750 {
751 I32 len = decode_uint (dec);
752 char *key = (char *)dec->cur;
753
754 dec->cur += len;
755
756 hv_store (hv, key, len, decode_sv (dec), 0);
757
758 return;
759 }
760 else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
761 {
762 I32 len = decode_uint (dec);
763 char *key = (char *)dec->cur;
764
765 dec->cur += len;
766
767 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
768 if (!is_utf8_string (key, len))
769 ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
770
771 hv_store (hv, key, -len, decode_sv (dec), 0);
772
773 return;
774 }
775
776 SV *k = decode_sv (dec);
777 SV *v = decode_sv (dec);
778
779 hv_store_ent (hv, k, v, 0);
780 SvREFCNT_dec (k);
781
782fail:
783 ;
784}
785
533static SV * 786static SV *
534decode_hv (dec_t *dec) 787decode_hv (dec_t *dec)
535{ 788{
536 HV *hv = newHV (); 789 HV *hv = newHV ();
537 790
538 DEC_INC_DEPTH; 791 DEC_INC_DEPTH;
539 792
540 if ((*dec->cur & 31) == 31) 793 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
541 { 794 {
542 ++dec->cur; 795 ++dec->cur;
543 796
544 for (;;) 797 for (;;)
545 { 798 {
546 WANT (1); 799 WANT (1);
547 800
548 if (*dec->cur == (0xe0 | 31)) 801 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
549 { 802 {
550 ++dec->cur; 803 ++dec->cur;
551 break; 804 break;
552 } 805 }
553 806
554 SV *k = decode_sv (dec); 807 decode_he (dec, hv);
555 SV *v = decode_sv (dec);
556
557 hv_store_ent (hv, k, v, 0);
558 } 808 }
559 } 809 }
560 else 810 else
561 { 811 {
562 int len = decode_uint (dec); 812 int pairs = decode_uint (dec);
563 813
564 while (len--) 814 while (pairs--)
565 { 815 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 } 816 }
572 817
573 DEC_DEC_DEPTH; 818 DEC_DEC_DEPTH;
574 return newRV_noinc ((SV *)hv); 819 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
730 820
731fail: 821fail:
732 SvREFCNT_dec (hv); 822 SvREFCNT_dec (hv);
733 DEC_DEC_DEPTH; 823 DEC_DEC_DEPTH;
734 return &PL_sv_undef; 824 return &PL_sv_undef;
735} 825}
736 826
737static SV * 827static SV *
738decode_str (dec_t *dec, int utf8) 828decode_str (dec_t *dec, int utf8)
739{ 829{
740 SV *sv; 830 SV *sv = 0;
741 831
742 if ((*dec->cur & 31) == 31) 832 if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
743 { 833 {
834 // indefinite length strings
744 ++dec->cur; 835 ++dec->cur;
745 836
837 U8 major = *dec->cur & MAJOR_MISC;
838
746 sv = newSVpvn ("", 0); 839 sv = newSVpvn ("", 0);
747 840
748 // not very fast, and certainly not robust against illegal input
749 for (;;) 841 for (;;)
750 { 842 {
751 WANT (1); 843 WANT (1);
752 844
753 if (*dec->cur == (0xe0 | 31)) 845 if ((*dec->cur - major) > LENGTH_EXT8)
846 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
754 { 847 {
755 ++dec->cur; 848 ++dec->cur;
756 break; 849 break;
757 } 850 }
851 else
852 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
758 853
759 SV *sv2 = decode_sv (dec); 854 STRLEN len = decode_uint (dec);
760 sv_catsv (sv, sv2); 855
856 WANT (len);
857 sv_catpvn (sv, dec->cur, len);
858 dec->cur += len;
761 } 859 }
762 } 860 }
763 else 861 else
764 { 862 {
765 STRLEN len = decode_uint (dec); 863 STRLEN len = decode_uint (dec);
766 864
767 WANT (len); 865 WANT (len);
768 sv = newSVpvn (dec->cur, len); 866 sv = newSVpvn (dec->cur, len);
769 dec->cur += len; 867 dec->cur += len;
868
869 if (ecb_expect_false (dec->stringref)
870 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
871 av_push (dec->stringref, SvREFCNT_inc_NN (sv));
770 } 872 }
771 873
772 if (utf8) 874 if (utf8)
875 {
876 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
877 if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
878 ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
879
773 SvUTF8_on (sv); 880 SvUTF8_on (sv);
881 }
774 882
775 return sv; 883 return sv;
776 884
777fail: 885fail:
886 SvREFCNT_dec (sv);
778 return &PL_sv_undef; 887 return &PL_sv_undef;
779} 888}
780 889
781static SV * 890static SV *
782decode_tagged (dec_t *dec) 891decode_tagged (dec_t *dec)
783{ 892{
893 SV *sv = 0;
784 UV tag = decode_uint (dec); 894 UV tag = decode_uint (dec);
895
896 WANT (1);
897
898 switch (tag)
899 {
900 case CBOR_TAG_MAGIC:
785 SV *sv = decode_sv (dec); 901 sv = decode_sv (dec);
902 break;
786 903
787 if (tag == 55799) // 2.4.5 Self-Describe CBOR 904 case CBOR_TAG_INDIRECTION:
905 sv = newRV_noinc (decode_sv (dec));
906 break;
907
908 case CBOR_TAG_STRINGREF_NAMESPACE:
909 {
910 ENTER; SAVETMPS;
911
912 SAVESPTR (dec->stringref);
913 dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());
914
915 sv = decode_sv (dec);
916
917 FREETMPS; LEAVE;
918 }
919 break;
920
921 case CBOR_TAG_STRINGREF:
922 {
923 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
924 ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
925
926 UV idx = decode_uint (dec);
927
928 if (!dec->stringref || (int)idx > AvFILLp (dec->stringref))
929 ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)");
930
931 sv = newSVsv (AvARRAY (dec->stringref)[idx]);
932 }
933 break;
934
935 case CBOR_TAG_VALUE_SHAREABLE:
936 {
937 if (ecb_expect_false (!dec->shareable))
938 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
939
940 if (dec->cbor.flags & F_ALLOW_CYCLES)
941 {
942 sv = newSV (0);
943 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
944
945 SV *osv = decode_sv (dec);
946 sv_setsv (sv, osv);
947 SvREFCNT_dec_NN (osv);
948 }
949 else
950 {
951 av_push (dec->shareable, &PL_sv_undef);
952 int idx = AvFILLp (dec->shareable);
953 sv = decode_sv (dec);
954 av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
955 }
956 }
957 break;
958
959 case CBOR_TAG_VALUE_SHAREDREF:
960 {
961 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
962 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
963
964 UV idx = decode_uint (dec);
965
966 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable))
967 ERR ("corrupted CBOR data (sharedref index out of bounds)");
968
969 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
970
971 if (sv == &PL_sv_undef)
972 ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
973 }
974 break;
975
976 case CBOR_TAG_PERL_OBJECT:
977 {
978 sv = decode_sv (dec);
979
980 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
981 ERR ("corrupted CBOR data (non-array perl object)");
982
983 AV *av = (AV *)SvRV (sv);
984 int len = av_len (av) + 1;
985 HV *stash = gv_stashsv (*av_fetch (av, 0, 1), 0);
986
987 if (!stash)
988 ERR ("cannot decode perl-object (package does not exist)");
989
990 GV *method = gv_fetchmethod_autoload (stash, "THAW", 0);
991
992 if (!method)
993 ERR ("cannot decode perl-object (package does not have a THAW method)");
994
995 dSP;
996
997 ENTER; SAVETMPS; PUSHMARK (SP);
998 EXTEND (SP, len + 1);
999 // we re-bless the reference to get overload and other niceties right
1000 PUSHs (*av_fetch (av, 0, 1));
1001 PUSHs (sv_cbor);
1002
1003 int i;
1004
1005 for (i = 1; i < len; ++i)
1006 PUSHs (*av_fetch (av, i, 1));
1007
1008 PUTBACK;
1009 call_sv ((SV *)GvCV (method), G_SCALAR | G_EVAL);
1010 SPAGAIN;
1011
1012 if (SvTRUE (ERRSV))
1013 {
1014 FREETMPS; LEAVE;
1015 ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));
1016 }
1017
1018 SvREFCNT_dec (sv);
1019 sv = SvREFCNT_inc (POPs);
1020
1021 PUTBACK;
1022
1023 FREETMPS; LEAVE;
1024 }
1025 break;
1026
1027 default:
1028 {
1029 sv = decode_sv (dec);
1030
1031 dSP;
1032 ENTER; SAVETMPS; PUSHMARK (SP);
1033 EXTEND (SP, 2);
1034 PUSHs (newSVuv (tag));
1035 PUSHs (sv);
1036
1037 PUTBACK;
1038 int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL);
1039 SPAGAIN;
1040
1041 if (SvTRUE (ERRSV))
1042 {
1043 FREETMPS; LEAVE;
1044 ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));
1045 }
1046
1047 if (count)
1048 {
1049 SvREFCNT_dec (sv);
1050 sv = SvREFCNT_inc (POPs);
1051 }
1052 else
1053 {
1054 AV *av = newAV ();
1055 av_push (av, newSVuv (tag));
1056 av_push (av, sv);
1057
1058 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
1059 ? cbor_tagged_stash
1060 : gv_stashpv ("CBOR::XS::Tagged" , 1);
1061 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
1062 }
1063
1064 PUTBACK;
1065
1066 FREETMPS; LEAVE;
1067 }
1068 break;
1069 }
1070
788 return sv; 1071 return sv;
789 1072
790 AV *av = newAV (); 1073fail:
791 av_push (av, newSVuv (tag)); 1074 SvREFCNT_dec (sv);
792 av_push (av, sv); 1075 return &PL_sv_undef;
793 return newRV_noinc ((SV *)av);
794} 1076}
795 1077
796static SV * 1078static SV *
797decode_sv (dec_t *dec) 1079decode_sv (dec_t *dec)
798{ 1080{
799 WANT (1); 1081 WANT (1);
800 1082
801 switch (*dec->cur >> 5) 1083 switch (*dec->cur >> MAJOR_SHIFT)
802 { 1084 {
803 case 0: // unsigned int 1085 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
804 //TODO: 64 bit values on 3 2bit perls 1086 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
805 return newSVuv (decode_uint (dec)); 1087 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
806 case 1: // negative int 1088 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
807 return newSViv (-1 - (IV)decode_uint (dec)); 1089 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
808 case 2: // octet string 1090 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
809 return decode_str (dec, 0); 1091 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
810 case 3: // utf-8 string 1092
811 return decode_str (dec, 1); 1093 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) 1094 switch (*dec->cur++ & MINOR_MASK)
820 { 1095 {
821 case 20: 1096 case SIMPLE_FALSE:
822#if CBOR_SLOW 1097#if CBOR_SLOW
823 cbor_false = get_bool ("CBOR::XS::false"); 1098 types_false = get_bool ("Types::Serialiser::false");
824#endif 1099#endif
825 return newSVsv (cbor_false); 1100 return newSVsv (types_false);
826 case 21: 1101 case SIMPLE_TRUE:
827#if CBOR_SLOW 1102#if CBOR_SLOW
828 cbor_true = get_bool ("CBOR::XS::true"); 1103 types_true = get_bool ("Types::Serialiser::true");
829#endif 1104#endif
830 return newSVsv (cbor_true); 1105 return newSVsv (types_true);
831 case 22: 1106 case SIMPLE_NULL:
832 return newSVsv (&PL_sv_undef); 1107 return newSVsv (&PL_sv_undef);
1108 case SIMPLE_UNDEF:
1109#if CBOR_SLOW
1110 types_error = get_bool ("Types::Serialiser::error");
1111#endif
1112 return newSVsv (types_error);
833 1113
834 case 25: 1114 case MISC_FLOAT16:
835 { 1115 {
836 WANT (2); 1116 WANT (2);
837 1117
838 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1118 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
839 dec->cur += 2; 1119 dec->cur += 2;
840 1120
841 return newSVnv (ecb_binary16_to_float (fp)); 1121 return newSVnv (ecb_binary16_to_float (fp));
842 } 1122 }
843 1123
844 case 26: 1124 case MISC_FLOAT32:
845 { 1125 {
846 uint32_t fp; 1126 uint32_t fp;
847 WANT (4); 1127 WANT (4);
848 memcpy (&fp, dec->cur, 4); 1128 memcpy (&fp, dec->cur, 4);
849 dec->cur += 4; 1129 dec->cur += 4;
852 fp = ecb_bswap32 (fp); 1132 fp = ecb_bswap32 (fp);
853 1133
854 return newSVnv (ecb_binary32_to_float (fp)); 1134 return newSVnv (ecb_binary32_to_float (fp));
855 } 1135 }
856 1136
857 case 27: 1137 case MISC_FLOAT64:
858 { 1138 {
859 uint64_t fp; 1139 uint64_t fp;
860 WANT (8); 1140 WANT (8);
861 memcpy (&fp, dec->cur, 8); 1141 memcpy (&fp, dec->cur, 8);
862 dec->cur += 8; 1142 dec->cur += 8;
865 fp = ecb_bswap64 (fp); 1145 fp = ecb_bswap64 (fp);
866 1146
867 return newSVnv (ecb_binary64_to_double (fp)); 1147 return newSVnv (ecb_binary64_to_double (fp));
868 } 1148 }
869 1149
870 // 0..19 unassigned 1150 // 0..19 unassigned simple
871 // 24 reserved + unassigned (reserved values are not encodable) 1151 // 24 reserved + unassigned simple (reserved values are not encodable)
1152 // 28-30 unassigned misc
1153 // 31 break code
872 default: 1154 default:
873 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1155 ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
874 } 1156 }
875 1157
876 break; 1158 break;
877 } 1159 }
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 1160
935fail: 1161fail:
936 return &PL_sv_undef; 1162 return &PL_sv_undef;
937} 1163}
938 1164
939static SV * 1165static SV *
940decode_cbor (SV *string, CBOR *cbor, char **offset_return) 1166decode_cbor (SV *string, CBOR *cbor, char **offset_return)
941{ 1167{
942 dec_t dec; 1168 dec_t dec = { 0 };
943 SV *sv; 1169 SV *sv;
1170 STRLEN len;
1171 char *data = SvPVbyte (string, len);
944 1172
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) 1173 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", 1174 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); 1175 (unsigned long)len, (unsigned long)cbor->max_size);
976 }
977
978 sv_utf8_downgrade (string, 0);
979 1176
980 dec.cbor = *cbor; 1177 dec.cbor = *cbor;
981 dec.cur = (U8 *)SvPVX (string); 1178 dec.cur = (U8 *)data;
982 dec.end = (U8 *)SvEND (string); 1179 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 1180
989 sv = decode_sv (&dec); 1181 sv = decode_sv (&dec);
990 1182
991 if (offset_return) 1183 if (offset_return)
992 *offset_return = dec.cur; 1184 *offset_return = dec.cur;
995 if (dec.cur != dec.end && !dec.err) 1187 if (dec.cur != dec.end && !dec.err)
996 dec.err = "garbage after CBOR object"; 1188 dec.err = "garbage after CBOR object";
997 1189
998 if (dec.err) 1190 if (dec.err)
999 { 1191 {
1192 if (dec.shareable)
1193 {
1194 // need to break cyclic links, which whould all be in shareable
1195 int i;
1196 SV **svp;
1197
1198 for (i = av_len (dec.shareable) + 1; i--; )
1199 if ((svp = av_fetch (dec.shareable, i, 0)))
1200 sv_setsv (*svp, &PL_sv_undef);
1201 }
1202
1000 SvREFCNT_dec (sv); 1203 SvREFCNT_dec (sv);
1001 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)SvPVX (string), (int)(uint8_t)*dec.cur); 1204 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur);
1002 } 1205 }
1003 1206
1004 sv = sv_2mortal (sv); 1207 sv = sv_2mortal (sv);
1005 1208
1006 return sv; 1209 return sv;
1007} 1210}
1008 1211
1212/////////////////////////////////////////////////////////////////////////////
1213// incremental parser
1214
1215#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
1216
1217// returns 0 for notyet, 1 for success or error
1218static int
1219incr_parse (CBOR *self, SV *cborstr)
1220{
1221 STRLEN cur;
1222 SvPV (cborstr, cur);
1223
1224 while (ecb_expect_true (self->incr_need <= cur))
1225 {
1226 // table of integer count bytes
1227 static I8 incr_len[MINOR_MASK + 1] = {
1228 0, 0, 0, 0, 0, 0, 0, 0,
1229 0, 0, 0, 0, 0, 0, 0, 0,
1230 0, 0, 0, 0, 0, 0, 0, 0,
1231 1, 2, 4, 8,-1,-1,-1,-2
1232 };
1233
1234 const U8 *p = SvPVX (cborstr) + self->incr_pos;
1235 U8 m = *p & MINOR_MASK;
1236 IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1237 I8 ilen = incr_len[m];
1238
1239 self->incr_need = self->incr_pos + 1;
1240
1241 if (ecb_expect_false (ilen < 0))
1242 {
1243 if (m != MINOR_INDEF)
1244 return 1; // error
1245
1246 if (*p == (MAJOR_MISC | MINOR_INDEF))
1247 {
1248 if (count >= 0)
1249 return 1; // error
1250
1251 count = 1;
1252 }
1253 else
1254 {
1255 av_push (self->incr_count, newSViv (-1)); //TODO: nest
1256 count = -1;
1257 }
1258 }
1259 else
1260 {
1261 self->incr_need += ilen;
1262 if (ecb_expect_false (self->incr_need > cur))
1263 return 0;
1264
1265 int major = *p >> MAJOR_SHIFT;
1266
1267 switch (major)
1268 {
1269 case MAJOR_TAG >> MAJOR_SHIFT:
1270 ++count; // tags merely prefix another value
1271 break;
1272
1273 case MAJOR_BYTES >> MAJOR_SHIFT:
1274 case MAJOR_TEXT >> MAJOR_SHIFT:
1275 case MAJOR_ARRAY >> MAJOR_SHIFT:
1276 case MAJOR_MAP >> MAJOR_SHIFT:
1277 {
1278 UV len;
1279
1280 if (ecb_expect_false (ilen))
1281 {
1282 len = 0;
1283
1284 do {
1285 len = (len << 8) | *++p;
1286 } while (--ilen);
1287 }
1288 else
1289 len = m;
1290
1291 switch (major)
1292 {
1293 case MAJOR_BYTES >> MAJOR_SHIFT:
1294 case MAJOR_TEXT >> MAJOR_SHIFT:
1295 self->incr_need += len;
1296 if (ecb_expect_false (self->incr_need > cur))
1297 return 0;
1298
1299 break;
1300
1301 case MAJOR_MAP >> MAJOR_SHIFT:
1302 len <<= 1;
1303 case MAJOR_ARRAY >> MAJOR_SHIFT:
1304 if (len)
1305 {
1306 av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
1307 count = len + 1;
1308 }
1309 break;
1310 }
1311 }
1312 }
1313 }
1314
1315 self->incr_pos = self->incr_need;
1316
1317 if (count > 0)
1318 {
1319 while (!--count)
1320 {
1321 if (!AvFILLp (self->incr_count))
1322 return 1; // done
1323
1324 SvREFCNT_dec_NN (av_pop (self->incr_count));
1325 count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1326 }
1327
1328 SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
1329 }
1330 }
1331
1332 return 0;
1333}
1334
1335
1009///////////////////////////////////////////////////////////////////////////// 1336/////////////////////////////////////////////////////////////////////////////
1010// XS interface functions 1337// XS interface functions
1011 1338
1012MODULE = CBOR::XS PACKAGE = CBOR::XS 1339MODULE = CBOR::XS PACKAGE = CBOR::XS
1013 1340
1014BOOT: 1341BOOT:
1015{ 1342{
1016 cbor_stash = gv_stashpv ("CBOR::XS" , 1); 1343 cbor_stash = gv_stashpv ("CBOR::XS" , 1);
1017 cbor_boolean_stash = gv_stashpv ("CBOR::XS::Boolean", 1); 1344 cbor_tagged_stash = gv_stashpv ("CBOR::XS::Tagged" , 1);
1018 1345
1019 cbor_true = get_bool ("CBOR::XS::true"); 1346 types_boolean_stash = gv_stashpv ("Types::Serialiser::Boolean", 1);
1020 cbor_false = get_bool ("CBOR::XS::false"); 1347 types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1);
1348
1349 types_true = get_bool ("Types::Serialiser::true" );
1350 types_false = get_bool ("Types::Serialiser::false");
1351 types_error = get_bool ("Types::Serialiser::error");
1352
1353 default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1354
1355 sv_cbor = newSVpv ("CBOR", 0);
1356 SvREADONLY_on (sv_cbor);
1021} 1357}
1022 1358
1023PROTOTYPES: DISABLE 1359PROTOTYPES: DISABLE
1024 1360
1025void CLONE (...) 1361void CLONE (...)
1026 CODE: 1362 CODE:
1027 cbor_stash = 0; 1363 cbor_stash = 0;
1364 cbor_tagged_stash = 0;
1365 types_error_stash = 0;
1028 cbor_boolean_stash = 0; 1366 types_boolean_stash = 0;
1029 1367
1030void new (char *klass) 1368void new (char *klass)
1031 PPCODE: 1369 PPCODE:
1032{ 1370{
1033 SV *pv = NEWSV (0, sizeof (CBOR)); 1371 SV *pv = NEWSV (0, sizeof (CBOR));
1041 1379
1042void shrink (CBOR *self, int enable = 1) 1380void shrink (CBOR *self, int enable = 1)
1043 ALIAS: 1381 ALIAS:
1044 shrink = F_SHRINK 1382 shrink = F_SHRINK
1045 allow_unknown = F_ALLOW_UNKNOWN 1383 allow_unknown = F_ALLOW_UNKNOWN
1384 allow_sharing = F_ALLOW_SHARING
1385 allow_cycles = F_ALLOW_CYCLES
1386 pack_strings = F_PACK_STRINGS
1387 validate_utf8 = F_VALIDATE_UTF8
1046 PPCODE: 1388 PPCODE:
1047{ 1389{
1048 if (enable) 1390 if (enable)
1049 self->flags |= ix; 1391 self->flags |= ix;
1050 else 1392 else
1055 1397
1056void get_shrink (CBOR *self) 1398void get_shrink (CBOR *self)
1057 ALIAS: 1399 ALIAS:
1058 get_shrink = F_SHRINK 1400 get_shrink = F_SHRINK
1059 get_allow_unknown = F_ALLOW_UNKNOWN 1401 get_allow_unknown = F_ALLOW_UNKNOWN
1402 get_allow_sharing = F_ALLOW_SHARING
1403 get_allow_cycles = F_ALLOW_CYCLES
1404 get_pack_strings = F_PACK_STRINGS
1405 get_validate_utf8 = F_VALIDATE_UTF8
1060 PPCODE: 1406 PPCODE:
1061 XPUSHs (boolSV (self->flags & ix)); 1407 XPUSHs (boolSV (self->flags & ix));
1062 1408
1063void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1409void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1064 PPCODE: 1410 PPCODE:
1080 CODE: 1426 CODE:
1081 RETVAL = self->max_size; 1427 RETVAL = self->max_size;
1082 OUTPUT: 1428 OUTPUT:
1083 RETVAL 1429 RETVAL
1084 1430
1085#if 0 //TODO 1431void filter (CBOR *self, SV *filter = 0)
1086
1087void filter_cbor_object (CBOR *self, SV *cb = &PL_sv_undef)
1088 PPCODE: 1432 PPCODE:
1089{
1090 SvREFCNT_dec (self->cb_object); 1433 SvREFCNT_dec (self->filter);
1091 self->cb_object = SvOK (cb) ? newSVsv (cb) : 0; 1434 self->filter = filter ? newSVsv (filter) : filter;
1092
1093 XPUSHs (ST (0)); 1435 XPUSHs (ST (0));
1094}
1095 1436
1096void filter_cbor_single_key_object (CBOR *self, SV *key, SV *cb = &PL_sv_undef) 1437SV *get_filter (CBOR *self)
1097 PPCODE: 1438 CODE:
1098{ 1439 RETVAL = self->filter ? self->filter : NEWSV (0, 0);
1099 if (!self->cb_sk_object) 1440 OUTPUT:
1100 self->cb_sk_object = newHV (); 1441 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 1442
1120void encode (CBOR *self, SV *scalar) 1443void encode (CBOR *self, SV *scalar)
1121 PPCODE: 1444 PPCODE:
1122 PUTBACK; scalar = encode_cbor (scalar, self); SPAGAIN; 1445 PUTBACK; scalar = encode_cbor (scalar, self); SPAGAIN;
1123 XPUSHs (scalar); 1446 XPUSHs (scalar);
1136 EXTEND (SP, 2); 1459 EXTEND (SP, 2);
1137 PUSHs (sv); 1460 PUSHs (sv);
1138 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1461 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1139} 1462}
1140 1463
1464void incr_parse (CBOR *self, SV *cborstr)
1465 ALIAS:
1466 incr_parse_multiple = 1
1467 PPCODE:
1468{
1469 if (SvUTF8 (cborstr))
1470 sv_utf8_downgrade (cborstr, 0);
1471
1472 if (!self->incr_count)
1473 {
1474 self->incr_count = newAV ();
1475 self->incr_pos = 0;
1476 self->incr_need = 1;
1477
1478 av_push (self->incr_count, newSViv (1));
1479 }
1480
1481 do
1482 {
1483 if (!incr_parse (self, cborstr))
1484 {
1485 if (self->incr_need > self->max_size && self->max_size)
1486 croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
1487 (unsigned long)self->incr_need, (unsigned long)self->max_size);
1488
1489 break;
1490 }
1491
1492 SV *sv;
1493 char *offset;
1494
1495 PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
1496 XPUSHs (sv);
1497
1498 sv_chop (cborstr, offset);
1499
1500 av_clear (self->incr_count);
1501 av_push (self->incr_count, newSViv (1));
1502
1503 self->incr_pos = 0;
1504 self->incr_need = self->incr_pos + 1;
1505 }
1506 while (ix);
1507}
1508
1509void incr_reset (CBOR *self)
1510 CODE:
1511{
1512 SvREFCNT_dec (self->incr_count);
1513 self->incr_count = 0;
1514}
1515
1141void DESTROY (CBOR *self) 1516void DESTROY (CBOR *self)
1142 CODE: 1517 PPCODE:
1143 SvREFCNT_dec (self->cb_sk_object); 1518 cbor_free (self);
1144 SvREFCNT_dec (self->cb_object);
1145 1519
1146PROTOTYPES: ENABLE 1520PROTOTYPES: ENABLE
1147 1521
1148void encode_cbor (SV *scalar) 1522void encode_cbor (SV *scalar)
1523 ALIAS:
1524 encode_cbor = 0
1525 encode_cbor_sharing = F_ALLOW_SHARING
1149 PPCODE: 1526 PPCODE:
1150{ 1527{
1151 CBOR cbor; 1528 CBOR cbor;
1152 cbor_init (&cbor); 1529 cbor_init (&cbor);
1530 cbor.flags |= ix;
1153 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1531 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1154 XPUSHs (scalar); 1532 XPUSHs (scalar);
1155} 1533}
1156 1534
1157void decode_cbor (SV *cborstr) 1535void decode_cbor (SV *cborstr)

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines