ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/CBOR-XS/XS.xs
(Generate patch)

Comparing CBOR-XS/XS.xs (file contents):
Revision 1.13 by root, Mon Oct 28 22:50:50 2013 UTC vs.
Revision 1.35 by root, Sat Nov 30 17:19:34 2013 UTC

9#include <limits.h> 9#include <limits.h>
10#include <float.h> 10#include <float.h>
11 11
12#include "ecb.h" 12#include "ecb.h"
13 13
14// compatibility with perl <5.18
15#ifndef HvNAMELEN_get
16# define HvNAMELEN_get(hv) strlen (HvNAME (hv))
17#endif
18#ifndef HvNAMELEN
19# define HvNAMELEN(hv) HvNAMELEN_get (hv)
20#endif
21#ifndef HvNAMEUTF8
22# define HvNAMEUTF8(hv) 0
23#endif
24#ifndef SvREFCNT_dec_NN
25# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
26#endif
27
28// known major and minor types
29enum cbor_type
30{
31 MAJOR_SHIFT = 5,
32 MINOR_MASK = 0x1f,
33
34 MAJOR_POS_INT = 0 << MAJOR_SHIFT,
35 MAJOR_NEG_INT = 1 << MAJOR_SHIFT,
36 MAJOR_BYTES = 2 << MAJOR_SHIFT,
37 MAJOR_TEXT = 3 << MAJOR_SHIFT,
38 MAJOR_ARRAY = 4 << MAJOR_SHIFT,
39 MAJOR_MAP = 5 << MAJOR_SHIFT,
40 MAJOR_TAG = 6 << MAJOR_SHIFT,
41 MAJOR_MISC = 7 << MAJOR_SHIFT,
42
43 // INT/STRING/ARRAY/MAP subtypes
44 LENGTH_EXT1 = 24,
45 LENGTH_EXT2 = 25,
46 LENGTH_EXT4 = 26,
47 LENGTH_EXT8 = 27,
48
49 // SIMPLE types (effectively MISC subtypes)
50 SIMPLE_FALSE = 20,
51 SIMPLE_TRUE = 21,
52 SIMPLE_NULL = 22,
53 SIMPLE_UNDEF = 23,
54
55 // MISC subtype (unused)
56 MISC_EXT1 = 24,
57 MISC_FLOAT16 = 25,
58 MISC_FLOAT32 = 26,
59 MISC_FLOAT64 = 27,
60
61 // BYTES/TEXT/ARRAY/MAP
62 MINOR_INDEF = 31,
63};
64
14// known tags 65// known tags
15enum cbor_tag 66enum cbor_tag
16{ 67{
17 // inofficial extensions (pending iana registration) 68 // extensions
18 CBOR_TAG_PERL_OBJECT = 256, 69 CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
19 CBOR_TAG_GENERIC_OBJECT = 257, 70 CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
71 CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
72 CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
73 CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
74 CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref
75 CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
20 76
21 // rfc7049 77 // rfc7049
22 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8 78 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
23 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any 79 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
24 CBOR_TAG_POS_BIGNUM = 2, // byte string 80 CBOR_TAG_POS_BIGNUM = 2, // byte string
25 CBOR_TAG_NEG_BIGNUM = 3, // byte string 81 CBOR_TAG_NEG_BIGNUM = 3, // byte string
26 CBOR_TAG_DECIMAL = 4, // decimal fraction, array 82 CBOR_TAG_DECIMAL = 4, // decimal fraction, array
27 CBOR_TAG_BIGFLOAT = 5, // array 83 CBOR_TAG_BIGFLOAT = 5, // array
28 84
29 CBOR_TAG_CONV_B64U = 21, // base64url, any 85 CBOR_TAG_CONV_B64U = 21, // base64url, any
30 CBOR_TAG_CONV_B64 = 22, // base64, any 86 CBOR_TAG_CONV_B64 = 22, // base64, any
31 CBOR_TAG_CONV_HEX = 23, // base16, any 87 CBOR_TAG_CONV_HEX = 23, // base16, any
32 CBOR_TAG_CBOR = 24, // embedded cbor, byte string 88 CBOR_TAG_CBOR = 24, // embedded cbor, byte string
33 89
34 CBOR_TAG_URI = 32, // URI rfc3986, utf-8 90 CBOR_TAG_URI = 32, // URI rfc3986, utf-8
35 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8 91 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
36 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8 92 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
37 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8 93 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
38 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8 94 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
39 95
40 CBOR_TAG_MAGIC = 55799 // self-describe cbor 96 CBOR_TAG_MAGIC = 55799, // self-describe cbor
41}; 97};
42 98
43#define F_SHRINK 0x00000200UL 99#define F_SHRINK 0x00000001UL
44#define F_ALLOW_UNKNOWN 0x00002000UL 100#define F_ALLOW_UNKNOWN 0x00000002UL
101#define F_ALLOW_SHARING 0x00000004UL
102#define F_PACK_STRINGS 0x00000008UL
45 103
46#define INIT_SIZE 32 // initial scalar size to be allocated 104#define INIT_SIZE 32 // initial scalar size to be allocated
47 105
48#define SB do { 106#define SB do {
49#define SE } while (0) 107#define SE } while (0)
61# define CBOR_SLOW 0 119# define CBOR_SLOW 0
62# define CBOR_STASH cbor_stash 120# define CBOR_STASH cbor_stash
63#endif 121#endif
64 122
65static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS:: 123static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS::
66static SV *types_true, *types_false, *types_error, *sv_cbor; 124static SV *types_true, *types_false, *types_error, *sv_cbor, *default_filter;
67 125
68typedef struct { 126typedef struct {
69 U32 flags; 127 U32 flags;
70 U32 max_depth; 128 U32 max_depth;
71 STRLEN max_size; 129 STRLEN max_size;
130 SV *filter;
72} CBOR; 131} CBOR;
73 132
74ecb_inline void 133ecb_inline void
75cbor_init (CBOR *cbor) 134cbor_init (CBOR *cbor)
76{ 135{
77 Zero (cbor, 1, CBOR); 136 Zero (cbor, 1, CBOR);
78 cbor->max_depth = 512; 137 cbor->max_depth = 512;
138}
139
140ecb_inline void
141cbor_free (CBOR *cbor)
142{
143 SvREFCNT_dec (cbor->filter);
79} 144}
80 145
81///////////////////////////////////////////////////////////////////////////// 146/////////////////////////////////////////////////////////////////////////////
82// utility functions 147// utility functions
83 148
105 SvPV_renew (sv, SvCUR (sv) + 1); 170 SvPV_renew (sv, SvCUR (sv) + 1);
106#endif 171#endif
107 } 172 }
108} 173}
109 174
110///////////////////////////////////////////////////////////////////////////// 175// minimum length of a string to be registered for stringref
111// fp hell 176ecb_inline int
112 177minimum_string_length (UV idx)
113//TODO 178{
179 return idx > 23
180 ? idx > 0xffU
181 ? idx > 0xffffU
182 ? idx > 0xffffffffU
183 ? 11
184 : 7
185 : 5
186 : 4
187 : 3;
188}
114 189
115///////////////////////////////////////////////////////////////////////////// 190/////////////////////////////////////////////////////////////////////////////
116// encoder 191// encoder
117 192
118// structure used for encoding CBOR 193// structure used for encoding CBOR
121 char *cur; // SvPVX (sv) + current output position 196 char *cur; // SvPVX (sv) + current output position
122 char *end; // SvEND (sv) 197 char *end; // SvEND (sv)
123 SV *sv; // result scalar 198 SV *sv; // result scalar
124 CBOR cbor; 199 CBOR cbor;
125 U32 depth; // recursion level 200 U32 depth; // recursion level
201 HV *stringref[2]; // string => index, or 0 ([0] = bytes, [1] = utf-8)
202 UV stringref_idx;
203 HV *shareable; // ptr => index, or 0
204 UV shareable_idx;
126} enc_t; 205} enc_t;
127 206
128ecb_inline void 207ecb_inline void
129need (enc_t *enc, STRLEN len) 208need (enc_t *enc, STRLEN len)
130{ 209{
147static void 226static void
148encode_uint (enc_t *enc, int major, UV len) 227encode_uint (enc_t *enc, int major, UV len)
149{ 228{
150 need (enc, 9); 229 need (enc, 9);
151 230
152 if (len < 24) 231 if (ecb_expect_true (len < LENGTH_EXT1))
153 *enc->cur++ = major | len; 232 *enc->cur++ = major | len;
154 else if (len <= 0xff) 233 else if (ecb_expect_true (len <= 0xff))
155 { 234 {
156 *enc->cur++ = major | 24; 235 *enc->cur++ = major | LENGTH_EXT1;
157 *enc->cur++ = len; 236 *enc->cur++ = len;
158 } 237 }
159 else if (len <= 0xffff) 238 else if (len <= 0xffff)
160 { 239 {
161 *enc->cur++ = major | 25; 240 *enc->cur++ = major | LENGTH_EXT2;
162 *enc->cur++ = len >> 8; 241 *enc->cur++ = len >> 8;
163 *enc->cur++ = len; 242 *enc->cur++ = len;
164 } 243 }
165 else if (len <= 0xffffffff) 244 else if (len <= 0xffffffff)
166 { 245 {
167 *enc->cur++ = major | 26; 246 *enc->cur++ = major | LENGTH_EXT4;
168 *enc->cur++ = len >> 24; 247 *enc->cur++ = len >> 24;
169 *enc->cur++ = len >> 16; 248 *enc->cur++ = len >> 16;
170 *enc->cur++ = len >> 8; 249 *enc->cur++ = len >> 8;
171 *enc->cur++ = len; 250 *enc->cur++ = len;
172 } 251 }
173 else 252 else
174 { 253 {
175 *enc->cur++ = major | 27; 254 *enc->cur++ = major | LENGTH_EXT8;
176 *enc->cur++ = len >> 56; 255 *enc->cur++ = len >> 56;
177 *enc->cur++ = len >> 48; 256 *enc->cur++ = len >> 48;
178 *enc->cur++ = len >> 40; 257 *enc->cur++ = len >> 40;
179 *enc->cur++ = len >> 32; 258 *enc->cur++ = len >> 32;
180 *enc->cur++ = len >> 24; 259 *enc->cur++ = len >> 24;
182 *enc->cur++ = len >> 8; 261 *enc->cur++ = len >> 8;
183 *enc->cur++ = len; 262 *enc->cur++ = len;
184 } 263 }
185} 264}
186 265
187static void 266ecb_inline void
267encode_tag (enc_t *enc, UV tag)
268{
269 encode_uint (enc, MAJOR_TAG, tag);
270}
271
272ecb_inline void
188encode_str (enc_t *enc, int utf8, char *str, STRLEN len) 273encode_str (enc_t *enc, int utf8, char *str, STRLEN len)
189{ 274{
190 encode_uint (enc, utf8 ? 0x60 : 0x40, len); 275 encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
191 need (enc, len); 276 need (enc, len);
192 memcpy (enc->cur, str, len); 277 memcpy (enc->cur, str, len);
193 enc->cur += len; 278 enc->cur += len;
194} 279}
195 280
281static void
282encode_strref (enc_t *enc, int utf8, char *str, STRLEN len)
283{
284 if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
285 {
286 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
287
288 if (SvOK (*svp))
289 {
290 // already registered, use stringref
291 encode_tag (enc, CBOR_TAG_STRINGREF);
292 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
293 return;
294 }
295 else if (len >= minimum_string_length (enc->stringref_idx))
296 {
297 // register only
298 sv_setuv (*svp, enc->stringref_idx);
299 ++enc->stringref_idx;
300 }
301 }
302
303 encode_str (enc, utf8, str, len);
304}
305
196static void encode_sv (enc_t *enc, SV *sv); 306static void encode_sv (enc_t *enc, SV *sv);
197 307
198static void 308static void
199encode_av (enc_t *enc, AV *av) 309encode_av (enc_t *enc, AV *av)
200{ 310{
203 if (enc->depth >= enc->cbor.max_depth) 313 if (enc->depth >= enc->cbor.max_depth)
204 croak (ERR_NESTING_EXCEEDED); 314 croak (ERR_NESTING_EXCEEDED);
205 315
206 ++enc->depth; 316 ++enc->depth;
207 317
208 encode_uint (enc, 0x80, len + 1); 318 encode_uint (enc, MAJOR_ARRAY, len + 1);
209 319
210 for (i = 0; i <= len; ++i) 320 for (i = 0; i <= len; ++i)
211 { 321 {
212 SV **svp = av_fetch (av, i, 0); 322 SV **svp = av_fetch (av, i, 0);
213 encode_sv (enc, svp ? *svp : &PL_sv_undef); 323 encode_sv (enc, svp ? *svp : &PL_sv_undef);
228 338
229 int pairs = hv_iterinit (hv); 339 int pairs = hv_iterinit (hv);
230 int mg = SvMAGICAL (hv); 340 int mg = SvMAGICAL (hv);
231 341
232 if (mg) 342 if (mg)
233 encode_ch (enc, 0xa0 | 31); 343 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
234 else 344 else
235 encode_uint (enc, 0xa0, pairs); 345 encode_uint (enc, MAJOR_MAP, pairs);
236 346
237 while ((he = hv_iternext (hv))) 347 while ((he = hv_iternext (hv)))
238 { 348 {
239 if (HeKLEN (he) == HEf_SVKEY) 349 if (HeKLEN (he) == HEf_SVKEY)
240 encode_sv (enc, HeSVKEY (he)); 350 encode_sv (enc, HeSVKEY (he));
241 else 351 else
242 encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); 352 encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
243 353
244 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); 354 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
245 } 355 }
246 356
247 if (mg) 357 if (mg)
248 encode_ch (enc, 0xe0 | 31); 358 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
249 359
250 --enc->depth; 360 --enc->depth;
251} 361}
252 362
253// encode objects, arrays and special \0=false and \1=true values. 363// encode objects, arrays and special \0=false and \1=true values.
254static void 364static void
255encode_rv (enc_t *enc, SV *sv) 365encode_rv (enc_t *enc, SV *sv)
256{ 366{
257 svtype svt;
258
259 SvGETMAGIC (sv); 367 SvGETMAGIC (sv);
368
260 svt = SvTYPE (sv); 369 svtype svt = SvTYPE (sv);
261 370
262 if (ecb_expect_false (SvOBJECT (sv))) 371 if (ecb_expect_false (SvOBJECT (sv)))
263 { 372 {
264 HV *boolean_stash = !CBOR_SLOW || types_boolean_stash 373 HV *boolean_stash = !CBOR_SLOW || types_boolean_stash
265 ? types_boolean_stash 374 ? types_boolean_stash
270 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 379 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
271 ? cbor_tagged_stash 380 ? cbor_tagged_stash
272 : gv_stashpv ("CBOR::XS::Tagged" , 1); 381 : gv_stashpv ("CBOR::XS::Tagged" , 1);
273 382
274 HV *stash = SvSTASH (sv); 383 HV *stash = SvSTASH (sv);
275 GV *method;
276 384
277 if (stash == boolean_stash) 385 if (stash == boolean_stash)
278 encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20); 386 {
387 encode_ch (enc, SvIV (sv) ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE);
388 return;
389 }
279 else if (stash == error_stash) 390 else if (stash == error_stash)
280 encode_ch (enc, 0xe0 | 23); 391 {
392 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
393 return;
394 }
281 else if (stash == tagged_stash) 395 else if (stash == tagged_stash)
282 { 396 {
283 if (svt != SVt_PVAV) 397 if (svt != SVt_PVAV)
284 croak ("encountered CBOR::XS::Tagged object that isn't an array"); 398 croak ("encountered CBOR::XS::Tagged object that isn't an array");
285 399
286 encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); 400 encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
287 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); 401 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1));
402
403 return;
404 }
405 }
406
407 if (ecb_expect_false (SvREFCNT (sv) > 1)
408 && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
409 {
410 if (!enc->shareable)
411 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
412
413 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
414
415 if (SvOK (*svp))
288 } 416 {
417 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
418 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
419 return;
420 }
421 else
422 {
423 sv_setuv (*svp, enc->shareable_idx);
424 ++enc->shareable_idx;
425 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
426 }
427 }
428
429 if (ecb_expect_false (SvOBJECT (sv)))
430 {
431 HV *stash = SvSTASH (sv);
432 GV *method;
433
289 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) 434 if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
290 { 435 {
291 dSP; 436 dSP;
292 437
293 ENTER; SAVETMPS; PUSHMARK (SP); 438 ENTER; SAVETMPS; PUSHMARK (SP);
294 // we re-bless the reference to get overload and other niceties right 439 // we re-bless the reference to get overload and other niceties right
325 470
326 // catch this surprisingly common error 471 // catch this surprisingly common error
327 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) 472 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
328 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash)); 473 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
329 474
330 encode_uint (enc, 0xc0, CBOR_TAG_PERL_OBJECT); 475 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
331 encode_uint (enc, 0x80, count + 1); 476 encode_uint (enc, MAJOR_ARRAY, count + 1);
332 encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); 477 encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
333 478
334 while (count) 479 while (count)
335 encode_sv (enc, SP[1 - count--]); 480 encode_sv (enc, SP[1 - count--]);
336 481
337 PUTBACK; 482 PUTBACK;
344 } 489 }
345 else if (svt == SVt_PVHV) 490 else if (svt == SVt_PVHV)
346 encode_hv (enc, (HV *)sv); 491 encode_hv (enc, (HV *)sv);
347 else if (svt == SVt_PVAV) 492 else if (svt == SVt_PVAV)
348 encode_av (enc, (AV *)sv); 493 encode_av (enc, (AV *)sv);
349 else if (svt < SVt_PVAV)
350 {
351 STRLEN len = 0;
352 char *pv = svt ? SvPV (sv, len) : 0;
353
354 if (len == 1 && *pv == '1')
355 encode_ch (enc, 0xe0 | 21);
356 else if (len == 1 && *pv == '0')
357 encode_ch (enc, 0xe0 | 20);
358 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
359 encode_ch (enc, 0xe0 | 23);
360 else
361 croak ("cannot encode reference to scalar '%s' unless the scalar is 0 or 1",
362 SvPV_nolen (sv_2mortal (newRV_inc (sv))));
363 }
364 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
365 encode_ch (enc, 0xe0 | 23);
366 else 494 else
367 croak ("encountered %s, but CBOR can only represent references to arrays or hashes", 495 {
368 SvPV_nolen (sv_2mortal (newRV_inc (sv)))); 496 encode_tag (enc, CBOR_TAG_INDIRECTION);
497 encode_sv (enc, sv);
498 }
369} 499}
370 500
371static void 501static void
372encode_nv (enc_t *enc, SV *sv) 502encode_nv (enc_t *enc, SV *sv)
373{ 503{
374 double nv = SvNVX (sv); 504 double nv = SvNVX (sv);
375 505
376 need (enc, 9); 506 need (enc, 9);
377 507
378 if (ecb_expect_false (nv == (U32)nv)) 508 if (ecb_expect_false (nv == (NV)(U32)nv))
379 encode_uint (enc, 0x00, (U32)nv); 509 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
380 //TODO: maybe I32? 510 //TODO: maybe I32?
381 else if (ecb_expect_false (nv == (float)nv)) 511 else if (ecb_expect_false (nv == (float)nv))
382 { 512 {
383 uint32_t fp = ecb_float_to_binary32 (nv); 513 uint32_t fp = ecb_float_to_binary32 (nv);
384 514
385 *enc->cur++ = 0xe0 | 26; 515 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
386 516
387 if (!ecb_big_endian ()) 517 if (!ecb_big_endian ())
388 fp = ecb_bswap32 (fp); 518 fp = ecb_bswap32 (fp);
389 519
390 memcpy (enc->cur, &fp, 4); 520 memcpy (enc->cur, &fp, 4);
392 } 522 }
393 else 523 else
394 { 524 {
395 uint64_t fp = ecb_double_to_binary64 (nv); 525 uint64_t fp = ecb_double_to_binary64 (nv);
396 526
397 *enc->cur++ = 0xe0 | 27; 527 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
398 528
399 if (!ecb_big_endian ()) 529 if (!ecb_big_endian ())
400 fp = ecb_bswap64 (fp); 530 fp = ecb_bswap64 (fp);
401 531
402 memcpy (enc->cur, &fp, 8); 532 memcpy (enc->cur, &fp, 8);
411 541
412 if (SvPOKp (sv)) 542 if (SvPOKp (sv))
413 { 543 {
414 STRLEN len; 544 STRLEN len;
415 char *str = SvPV (sv, len); 545 char *str = SvPV (sv, len);
416 encode_str (enc, SvUTF8 (sv), str, len); 546 encode_strref (enc, SvUTF8 (sv), str, len);
417 } 547 }
418 else if (SvNOKp (sv)) 548 else if (SvNOKp (sv))
419 encode_nv (enc, sv); 549 encode_nv (enc, sv);
420 else if (SvIOKp (sv)) 550 else if (SvIOKp (sv))
421 { 551 {
422 if (SvIsUV (sv)) 552 if (SvIsUV (sv))
423 encode_uint (enc, 0x00, SvUVX (sv)); 553 encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
424 else if (SvIVX (sv) >= 0) 554 else if (SvIVX (sv) >= 0)
425 encode_uint (enc, 0x00, SvIVX (sv)); 555 encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
426 else 556 else
427 encode_uint (enc, 0x20, -(SvIVX (sv) + 1)); 557 encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
428 } 558 }
429 else if (SvROK (sv)) 559 else if (SvROK (sv))
430 encode_rv (enc, SvRV (sv)); 560 encode_rv (enc, SvRV (sv));
431 else if (!SvOK (sv)) 561 else if (!SvOK (sv))
432 encode_ch (enc, 0xe0 | 22); 562 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
433 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 563 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
434 encode_ch (enc, 0xe0 | 23); 564 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
435 else 565 else
436 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data", 566 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
437 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 567 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
438} 568}
439 569
440static SV * 570static SV *
441encode_cbor (SV *scalar, CBOR *cbor) 571encode_cbor (SV *scalar, CBOR *cbor)
442{ 572{
443 enc_t enc; 573 enc_t enc = { };
444 574
445 enc.cbor = *cbor; 575 enc.cbor = *cbor;
446 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 576 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
447 enc.cur = SvPVX (enc.sv); 577 enc.cur = SvPVX (enc.sv);
448 enc.end = SvEND (enc.sv); 578 enc.end = SvEND (enc.sv);
449 enc.depth = 0;
450 579
451 SvPOK_only (enc.sv); 580 SvPOK_only (enc.sv);
581
582 if (cbor->flags & F_PACK_STRINGS)
583 {
584 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
585 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
586 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
587 }
588
452 encode_sv (&enc, scalar); 589 encode_sv (&enc, scalar);
453 590
454 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv)); 591 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv));
455 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings 592 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings
456 593
470 U8 *end; // end of input string 607 U8 *end; // end of input string
471 const char *err; // parse error, if != 0 608 const char *err; // parse error, if != 0
472 CBOR cbor; 609 CBOR cbor;
473 U32 depth; // recursion depth 610 U32 depth; // recursion depth
474 U32 maxdepth; // recursion depth limit 611 U32 maxdepth; // recursion depth limit
612 AV *shareable;
613 AV *stringref;
614 SV *decode_tagged;
475} dec_t; 615} dec_t;
476 616
477#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE 617#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE
478 618
479#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data") 619#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data")
482#define DEC_DEC_DEPTH --dec->depth 622#define DEC_DEC_DEPTH --dec->depth
483 623
484static UV 624static UV
485decode_uint (dec_t *dec) 625decode_uint (dec_t *dec)
486{ 626{
487 switch (*dec->cur & 31) 627 U8 m = *dec->cur & MINOR_MASK;
628 ++dec->cur;
629
630 if (ecb_expect_true (m < LENGTH_EXT1))
631 return m;
632
633 switch (m)
488 { 634 {
489 case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7: 635 case LENGTH_EXT1:
490 case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15: 636 WANT (1);
491 case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23: 637 dec->cur += 1;
492 return *dec->cur++ & 31; 638 return dec->cur[-1];
493 639
494 case 24: 640 case LENGTH_EXT2:
495 WANT (2); 641 WANT (2);
496 dec->cur += 2; 642 dec->cur += 2;
497 return dec->cur[-1];
498
499 case 25:
500 WANT (3);
501 dec->cur += 3;
502 return (((UV)dec->cur[-2]) << 8) 643 return (((UV)dec->cur[-2]) << 8)
503 | ((UV)dec->cur[-1]); 644 | ((UV)dec->cur[-1]);
504 645
505 case 26: 646 case LENGTH_EXT4:
506 WANT (5); 647 WANT (4);
507 dec->cur += 5; 648 dec->cur += 4;
508 return (((UV)dec->cur[-4]) << 24) 649 return (((UV)dec->cur[-4]) << 24)
509 | (((UV)dec->cur[-3]) << 16) 650 | (((UV)dec->cur[-3]) << 16)
510 | (((UV)dec->cur[-2]) << 8) 651 | (((UV)dec->cur[-2]) << 8)
511 | ((UV)dec->cur[-1]); 652 | ((UV)dec->cur[-1]);
512 653
513 case 27: 654 case LENGTH_EXT8:
514 WANT (9); 655 WANT (8);
515 dec->cur += 9; 656 dec->cur += 8;
657
658 return
659#if UVSIZE < 8
660 0
661#else
516 return (((UV)dec->cur[-8]) << 56) 662 (((UV)dec->cur[-8]) << 56)
517 | (((UV)dec->cur[-7]) << 48) 663 | (((UV)dec->cur[-7]) << 48)
518 | (((UV)dec->cur[-6]) << 40) 664 | (((UV)dec->cur[-6]) << 40)
519 | (((UV)dec->cur[-5]) << 32) 665 | (((UV)dec->cur[-5]) << 32)
666#endif
520 | (((UV)dec->cur[-4]) << 24) 667 | (((UV)dec->cur[-4]) << 24)
521 | (((UV)dec->cur[-3]) << 16) 668 | (((UV)dec->cur[-3]) << 16)
522 | (((UV)dec->cur[-2]) << 8) 669 | (((UV)dec->cur[-2]) << 8)
523 | ((UV)dec->cur[-1]); 670 | ((UV)dec->cur[-1]);
524 671
537{ 684{
538 AV *av = newAV (); 685 AV *av = newAV ();
539 686
540 DEC_INC_DEPTH; 687 DEC_INC_DEPTH;
541 688
542 if ((*dec->cur & 31) == 31) 689 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
543 { 690 {
544 ++dec->cur; 691 ++dec->cur;
545 692
546 for (;;) 693 for (;;)
547 { 694 {
548 WANT (1); 695 WANT (1);
549 696
550 if (*dec->cur == (0xe0 | 31)) 697 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
551 { 698 {
552 ++dec->cur; 699 ++dec->cur;
553 break; 700 break;
554 } 701 }
555 702
573 SvREFCNT_dec (av); 720 SvREFCNT_dec (av);
574 DEC_DEC_DEPTH; 721 DEC_DEC_DEPTH;
575 return &PL_sv_undef; 722 return &PL_sv_undef;
576} 723}
577 724
725static void
726decode_he (dec_t *dec, HV *hv)
727{
728 // for speed reasons, we specialcase single-string
729 // byte or utf-8 strings as keys, but only when !stringref
730
731 if (ecb_expect_true (!dec->stringref))
732 if ((*dec->cur - MAJOR_BYTES) <= 27)
733 {
734 I32 len = decode_uint (dec);
735 char *key = (char *)dec->cur;
736
737 dec->cur += len;
738
739 if (ecb_expect_false (dec->stringref))
740 av_push (dec->stringref, newSVpvn (key, len));
741
742 hv_store (hv, key, len, decode_sv (dec), 0);
743
744 return;
745 }
746 else if ((*dec->cur - MAJOR_TEXT) <= 27)
747 {
748 I32 len = decode_uint (dec);
749 char *key = (char *)dec->cur;
750
751 dec->cur += len;
752
753 if (ecb_expect_false (dec->stringref))
754 av_push (dec->stringref, newSVpvn_utf8 (key, len, 1));
755
756 hv_store (hv, key, -len, decode_sv (dec), 0);
757
758 return;
759 }
760
761 SV *k = decode_sv (dec);
762 SV *v = decode_sv (dec);
763
764 hv_store_ent (hv, k, v, 0);
765 SvREFCNT_dec (k);
766}
767
578static SV * 768static SV *
579decode_hv (dec_t *dec) 769decode_hv (dec_t *dec)
580{ 770{
581 HV *hv = newHV (); 771 HV *hv = newHV ();
582 772
583 DEC_INC_DEPTH; 773 DEC_INC_DEPTH;
584 774
585 if ((*dec->cur & 31) == 31) 775 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
586 { 776 {
587 ++dec->cur; 777 ++dec->cur;
588 778
589 for (;;) 779 for (;;)
590 { 780 {
591 WANT (1); 781 WANT (1);
592 782
593 if (*dec->cur == (0xe0 | 31)) 783 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
594 { 784 {
595 ++dec->cur; 785 ++dec->cur;
596 break; 786 break;
597 } 787 }
598 788
599 SV *k = decode_sv (dec); 789 decode_he (dec, hv);
600 SV *v = decode_sv (dec);
601
602 hv_store_ent (hv, k, v, 0);
603 SvREFCNT_dec (k);
604 } 790 }
605 } 791 }
606 else 792 else
607 { 793 {
608 int len = decode_uint (dec); 794 int pairs = decode_uint (dec);
609 795
610 while (len--) 796 while (pairs--)
611 { 797 decode_he (dec, hv);
612 SV *k = decode_sv (dec);
613 SV *v = decode_sv (dec);
614
615 hv_store_ent (hv, k, v, 0);
616 SvREFCNT_dec (k);
617 }
618 } 798 }
619 799
620 DEC_DEC_DEPTH; 800 DEC_DEC_DEPTH;
621 return newRV_noinc ((SV *)hv); 801 return newRV_noinc ((SV *)hv);
622 802
629static SV * 809static SV *
630decode_str (dec_t *dec, int utf8) 810decode_str (dec_t *dec, int utf8)
631{ 811{
632 SV *sv = 0; 812 SV *sv = 0;
633 813
634 if ((*dec->cur & 31) == 31) 814 if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
635 { 815 {
816 // indefinite length strings
636 ++dec->cur; 817 ++dec->cur;
637 818
819 U8 major = *dec->cur & MAJOR_MISC;
820
638 sv = newSVpvn ("", 0); 821 sv = newSVpvn ("", 0);
639 822
640 // not very fast, and certainly not robust against illegal input
641 for (;;) 823 for (;;)
642 { 824 {
643 WANT (1); 825 WANT (1);
644 826
645 if (*dec->cur == (0xe0 | 31)) 827 if ((*dec->cur - major) > LENGTH_EXT8)
828 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
646 { 829 {
647 ++dec->cur; 830 ++dec->cur;
648 break; 831 break;
649 } 832 }
833 else
834 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
650 835
651 sv_catsv (sv, decode_sv (dec)); 836 STRLEN len = decode_uint (dec);
837
838 WANT (len);
839 sv_catpvn (sv, dec->cur, len);
840 dec->cur += len;
652 } 841 }
653 } 842 }
654 else 843 else
655 { 844 {
656 STRLEN len = decode_uint (dec); 845 STRLEN len = decode_uint (dec);
657 846
658 WANT (len); 847 WANT (len);
659 sv = newSVpvn (dec->cur, len); 848 sv = newSVpvn (dec->cur, len);
660 dec->cur += len; 849 dec->cur += len;
850
851 if (ecb_expect_false (dec->stringref)
852 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
853 av_push (dec->stringref, SvREFCNT_inc_NN (sv));
661 } 854 }
662 855
663 if (utf8) 856 if (utf8)
664 SvUTF8_on (sv); 857 SvUTF8_on (sv);
665 858
671} 864}
672 865
673static SV * 866static SV *
674decode_tagged (dec_t *dec) 867decode_tagged (dec_t *dec)
675{ 868{
869 SV *sv = 0;
676 UV tag = decode_uint (dec); 870 UV tag = decode_uint (dec);
871
872 WANT (1);
873
874 switch (tag)
875 {
876 case CBOR_TAG_MAGIC:
877 sv = decode_sv (dec);
878 break;
879
880 case CBOR_TAG_INDIRECTION:
881 sv = newRV_noinc (decode_sv (dec));
882 break;
883
884 case CBOR_TAG_STRINGREF_NAMESPACE:
885 {
886 ENTER; SAVETMPS;
887
888 SAVESPTR (dec->stringref);
889 dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());
890
891 sv = decode_sv (dec);
892
893 FREETMPS; LEAVE;
894 }
895 break;
896
897 case CBOR_TAG_STRINGREF:
898 {
899 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
900 ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
901
902 UV idx = decode_uint (dec);
903
904 if (!dec->stringref || (int)idx > AvFILLp (dec->stringref))
905 ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)");
906
907 sv = newSVsv (AvARRAY (dec->stringref)[idx]);
908 }
909 break;
910
911 case CBOR_TAG_VALUE_SHAREABLE:
912 {
913 if (ecb_expect_false (!dec->shareable))
914 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
915
916 sv = newSV (0);
917 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
918
677 SV *sv = decode_sv (dec); 919 SV *osv = decode_sv (dec);
920 sv_setsv (sv, osv);
921 SvREFCNT_dec_NN (osv);
922 }
923 break;
678 924
679 if (tag == CBOR_TAG_MAGIC) 925 case CBOR_TAG_VALUE_SHAREDREF:
680 return sv; 926 {
681 else if (tag == CBOR_TAG_PERL_OBJECT) 927 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
682 { 928 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
929
930 UV idx = decode_uint (dec);
931
932 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable))
933 ERR ("corrupted CBOR data (sharedref index out of bounds)");
934
935 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
936 }
937 break;
938
939 case CBOR_TAG_PERL_OBJECT:
940 {
941 sv = decode_sv (dec);
942
683 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) 943 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
684 ERR ("corrupted CBOR data (non-array perl object)"); 944 ERR ("corrupted CBOR data (non-array perl object)");
685 945
686 AV *av = (AV *)SvRV (sv); 946 AV *av = (AV *)SvRV (sv);
687 int len = av_len (av) + 1; 947 int len = av_len (av) + 1;
688 HV *stash = gv_stashsv (*av_fetch (av, 0, 1), 0); 948 HV *stash = gv_stashsv (*av_fetch (av, 0, 1), 0);
689 949
690 if (!stash) 950 if (!stash)
691 ERR ("cannot decode perl-object (package does not exist)"); 951 ERR ("cannot decode perl-object (package does not exist)");
692 952
693 GV *method = gv_fetchmethod_autoload (stash, "THAW", 0); 953 GV *method = gv_fetchmethod_autoload (stash, "THAW", 0);
694 954
695 if (!method) 955 if (!method)
696 ERR ("cannot decode perl-object (package does not have a THAW method)"); 956 ERR ("cannot decode perl-object (package does not have a THAW method)");
697 957
698 dSP; 958 dSP;
699 959
700 ENTER; SAVETMPS; PUSHMARK (SP); 960 ENTER; SAVETMPS; PUSHMARK (SP);
701 EXTEND (SP, len + 1); 961 EXTEND (SP, len + 1);
702 // we re-bless the reference to get overload and other niceties right 962 // we re-bless the reference to get overload and other niceties right
703 PUSHs (*av_fetch (av, 0, 1)); 963 PUSHs (*av_fetch (av, 0, 1));
704 PUSHs (sv_cbor); 964 PUSHs (sv_cbor);
705 965
706 int i; 966 int i;
707 967
708 for (i = 1; i < len; ++i) 968 for (i = 1; i < len; ++i)
709 PUSHs (*av_fetch (av, i, 1)); 969 PUSHs (*av_fetch (av, i, 1));
710 970
711 PUTBACK; 971 PUTBACK;
712 call_sv ((SV *)GvCV (method), G_SCALAR); 972 call_sv ((SV *)GvCV (method), G_SCALAR | G_EVAL);
713 SPAGAIN; 973 SPAGAIN;
714 974
975 if (SvTRUE (ERRSV))
976 {
977 FREETMPS; LEAVE;
978 ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));
979 }
980
715 SvREFCNT_dec (sv); 981 SvREFCNT_dec (sv);
716 sv = SvREFCNT_inc (POPs); 982 sv = SvREFCNT_inc (POPs);
717 983
718 PUTBACK; 984 PUTBACK;
719 985
720 FREETMPS; LEAVE; 986 FREETMPS; LEAVE;
987 }
988 break;
721 989
722 return sv; 990 default:
723 } 991 {
724 else 992 sv = decode_sv (dec);
725 { 993
994 dSP;
995 ENTER; SAVETMPS; PUSHMARK (SP);
996 EXTEND (SP, 2);
997 PUSHs (newSVuv (tag));
998 PUSHs (sv);
999
1000 PUTBACK;
1001 int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL);
1002 SPAGAIN;
1003
1004 if (SvTRUE (ERRSV))
1005 {
1006 FREETMPS; LEAVE;
1007 ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV))));
1008 }
1009
1010 if (count)
1011 {
1012 SvREFCNT_dec (sv);
1013 sv = SvREFCNT_inc (POPs);
1014 }
1015 else
1016 {
726 AV *av = newAV (); 1017 AV *av = newAV ();
727 av_push (av, newSVuv (tag)); 1018 av_push (av, newSVuv (tag));
728 av_push (av, sv); 1019 av_push (av, sv);
729 1020
730 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 1021 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
731 ? cbor_tagged_stash 1022 ? cbor_tagged_stash
732 : gv_stashpv ("CBOR::XS::Tagged" , 1); 1023 : gv_stashpv ("CBOR::XS::Tagged" , 1);
733
734 return sv_bless (newRV_noinc ((SV *)av), tagged_stash); 1024 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
1025 }
1026
1027 PUTBACK;
1028
1029 FREETMPS; LEAVE;
1030 }
1031 break;
735 } 1032 }
1033
1034 return sv;
736 1035
737fail: 1036fail:
738 SvREFCNT_dec (sv); 1037 SvREFCNT_dec (sv);
739 return &PL_sv_undef; 1038 return &PL_sv_undef;
740} 1039}
742static SV * 1041static SV *
743decode_sv (dec_t *dec) 1042decode_sv (dec_t *dec)
744{ 1043{
745 WANT (1); 1044 WANT (1);
746 1045
747 switch (*dec->cur >> 5) 1046 switch (*dec->cur >> MAJOR_SHIFT)
748 { 1047 {
749 case 0: // unsigned int 1048 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
750 return newSVuv (decode_uint (dec)); 1049 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
751 case 1: // negative int 1050 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
752 return newSViv (-1 - (IV)decode_uint (dec)); 1051 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
753 case 2: // octet string 1052 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
754 return decode_str (dec, 0); 1053 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
755 case 3: // utf-8 string 1054 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
756 return decode_str (dec, 1); 1055
757 case 4: // array 1056 case MAJOR_MISC >> MAJOR_SHIFT:
758 return decode_av (dec);
759 case 5: // map
760 return decode_hv (dec);
761 case 6: // tag
762 return decode_tagged (dec);
763 case 7: // misc
764 switch (*dec->cur++ & 31) 1057 switch (*dec->cur++ & MINOR_MASK)
765 { 1058 {
766 case 20: 1059 case SIMPLE_FALSE:
767#if CBOR_SLOW 1060#if CBOR_SLOW
768 types_false = get_bool ("Types::Serialiser::false"); 1061 types_false = get_bool ("Types::Serialiser::false");
769#endif 1062#endif
770 return newSVsv (types_false); 1063 return newSVsv (types_false);
771 case 21: 1064 case SIMPLE_TRUE:
772#if CBOR_SLOW 1065#if CBOR_SLOW
773 types_true = get_bool ("Types::Serialiser::true"); 1066 types_true = get_bool ("Types::Serialiser::true");
774#endif 1067#endif
775 return newSVsv (types_true); 1068 return newSVsv (types_true);
776 case 22: 1069 case SIMPLE_NULL:
777 return newSVsv (&PL_sv_undef); 1070 return newSVsv (&PL_sv_undef);
778 case 23: 1071 case SIMPLE_UNDEF:
779#if CBOR_SLOW 1072#if CBOR_SLOW
780 types_error = get_bool ("Types::Serialiser::error"); 1073 types_error = get_bool ("Types::Serialiser::error");
781#endif 1074#endif
782 return newSVsv (types_error); 1075 return newSVsv (types_error);
783 1076
784 case 25: 1077 case MISC_FLOAT16:
785 { 1078 {
786 WANT (2); 1079 WANT (2);
787 1080
788 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1081 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
789 dec->cur += 2; 1082 dec->cur += 2;
790 1083
791 return newSVnv (ecb_binary16_to_float (fp)); 1084 return newSVnv (ecb_binary16_to_float (fp));
792 } 1085 }
793 1086
794 case 26: 1087 case MISC_FLOAT32:
795 { 1088 {
796 uint32_t fp; 1089 uint32_t fp;
797 WANT (4); 1090 WANT (4);
798 memcpy (&fp, dec->cur, 4); 1091 memcpy (&fp, dec->cur, 4);
799 dec->cur += 4; 1092 dec->cur += 4;
802 fp = ecb_bswap32 (fp); 1095 fp = ecb_bswap32 (fp);
803 1096
804 return newSVnv (ecb_binary32_to_float (fp)); 1097 return newSVnv (ecb_binary32_to_float (fp));
805 } 1098 }
806 1099
807 case 27: 1100 case MISC_FLOAT64:
808 { 1101 {
809 uint64_t fp; 1102 uint64_t fp;
810 WANT (8); 1103 WANT (8);
811 memcpy (&fp, dec->cur, 8); 1104 memcpy (&fp, dec->cur, 8);
812 dec->cur += 8; 1105 dec->cur += 8;
815 fp = ecb_bswap64 (fp); 1108 fp = ecb_bswap64 (fp);
816 1109
817 return newSVnv (ecb_binary64_to_double (fp)); 1110 return newSVnv (ecb_binary64_to_double (fp));
818 } 1111 }
819 1112
820 // 0..19 unassigned 1113 // 0..19 unassigned simple
821 // 24 reserved + unassigned (reserved values are not encodable) 1114 // 24 reserved + unassigned (reserved values are not encodable)
822 default: 1115 default:
823 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1116 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)");
824 } 1117 }
825 1118
831} 1124}
832 1125
833static SV * 1126static SV *
834decode_cbor (SV *string, CBOR *cbor, char **offset_return) 1127decode_cbor (SV *string, CBOR *cbor, char **offset_return)
835{ 1128{
836 dec_t dec; 1129 dec_t dec = { };
837 SV *sv; 1130 SV *sv;
1131 STRLEN len;
1132 char *data = SvPVbyte (string, len);
838 1133
839 /* work around bugs in 5.10 where manipulating magic values
840 * makes perl ignore the magic in subsequent accesses.
841 * also make a copy of non-PV values, to get them into a clean
842 * state (SvPV should do that, but it's buggy, see below).
843 */
844 /*SvGETMAGIC (string);*/
845 if (SvMAGICAL (string) || !SvPOK (string))
846 string = sv_2mortal (newSVsv (string));
847
848 SvUPGRADE (string, SVt_PV);
849
850 /* work around a bug in perl 5.10, which causes SvCUR to fail an
851 * assertion with -DDEBUGGING, although SvCUR is documented to
852 * return the xpv_cur field which certainly exists after upgrading.
853 * according to nicholas clark, calling SvPOK fixes this.
854 * But it doesn't fix it, so try another workaround, call SvPV_nolen
855 * and hope for the best.
856 * Damnit, SvPV_nolen still trips over yet another assertion. This
857 * assertion business is seriously broken, try yet another workaround
858 * for the broken -DDEBUGGING.
859 */
860 {
861#ifdef DEBUGGING
862 STRLEN offset = SvOK (string) ? sv_len (string) : 0;
863#else
864 STRLEN offset = SvCUR (string);
865#endif
866
867 if (offset > cbor->max_size && cbor->max_size) 1134 if (len > cbor->max_size && cbor->max_size)
868 croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu", 1135 croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
869 (unsigned long)SvCUR (string), (unsigned long)cbor->max_size); 1136 (unsigned long)len, (unsigned long)cbor->max_size);
870 }
871
872 sv_utf8_downgrade (string, 0);
873 1137
874 dec.cbor = *cbor; 1138 dec.cbor = *cbor;
875 dec.cur = (U8 *)SvPVX (string); 1139 dec.cur = (U8 *)data;
876 dec.end = (U8 *)SvEND (string); 1140 dec.end = (U8 *)data + len;
877 dec.err = 0;
878 dec.depth = 0;
879 1141
880 sv = decode_sv (&dec); 1142 sv = decode_sv (&dec);
881 1143
882 if (offset_return) 1144 if (offset_return)
883 *offset_return = dec.cur; 1145 *offset_return = dec.cur;
887 dec.err = "garbage after CBOR object"; 1149 dec.err = "garbage after CBOR object";
888 1150
889 if (dec.err) 1151 if (dec.err)
890 { 1152 {
891 SvREFCNT_dec (sv); 1153 SvREFCNT_dec (sv);
892 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)SvPVX (string), (int)(uint8_t)*dec.cur); 1154 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur);
893 } 1155 }
894 1156
895 sv = sv_2mortal (sv); 1157 sv = sv_2mortal (sv);
896 1158
897 return sv; 1159 return sv;
911 types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1); 1173 types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1);
912 1174
913 types_true = get_bool ("Types::Serialiser::true" ); 1175 types_true = get_bool ("Types::Serialiser::true" );
914 types_false = get_bool ("Types::Serialiser::false"); 1176 types_false = get_bool ("Types::Serialiser::false");
915 types_error = get_bool ("Types::Serialiser::error"); 1177 types_error = get_bool ("Types::Serialiser::error");
1178
1179 default_filter = newSVpv ("CBOR::XS::default_filter", 0);
916 1180
917 sv_cbor = newSVpv ("CBOR", 0); 1181 sv_cbor = newSVpv ("CBOR", 0);
918 SvREADONLY_on (sv_cbor); 1182 SvREADONLY_on (sv_cbor);
919} 1183}
920 1184
941 1205
942void shrink (CBOR *self, int enable = 1) 1206void shrink (CBOR *self, int enable = 1)
943 ALIAS: 1207 ALIAS:
944 shrink = F_SHRINK 1208 shrink = F_SHRINK
945 allow_unknown = F_ALLOW_UNKNOWN 1209 allow_unknown = F_ALLOW_UNKNOWN
1210 allow_sharing = F_ALLOW_SHARING
1211 pack_strings = F_PACK_STRINGS
946 PPCODE: 1212 PPCODE:
947{ 1213{
948 if (enable) 1214 if (enable)
949 self->flags |= ix; 1215 self->flags |= ix;
950 else 1216 else
955 1221
956void get_shrink (CBOR *self) 1222void get_shrink (CBOR *self)
957 ALIAS: 1223 ALIAS:
958 get_shrink = F_SHRINK 1224 get_shrink = F_SHRINK
959 get_allow_unknown = F_ALLOW_UNKNOWN 1225 get_allow_unknown = F_ALLOW_UNKNOWN
1226 get_allow_sharing = F_ALLOW_SHARING
1227 get_pack_strings = F_PACK_STRINGS
960 PPCODE: 1228 PPCODE:
961 XPUSHs (boolSV (self->flags & ix)); 1229 XPUSHs (boolSV (self->flags & ix));
962 1230
963void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1231void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
964 PPCODE: 1232 PPCODE:
980 CODE: 1248 CODE:
981 RETVAL = self->max_size; 1249 RETVAL = self->max_size;
982 OUTPUT: 1250 OUTPUT:
983 RETVAL 1251 RETVAL
984 1252
985#if 0 //TODO 1253void filter (CBOR *self, SV *filter = 0)
986
987void filter_cbor_object (CBOR *self, SV *cb = &PL_sv_undef)
988 PPCODE: 1254 PPCODE:
989{
990 SvREFCNT_dec (self->cb_object); 1255 SvREFCNT_dec (self->filter);
991 self->cb_object = SvOK (cb) ? newSVsv (cb) : 0; 1256 self->filter = filter ? newSVsv (filter) : filter;
992
993 XPUSHs (ST (0)); 1257 XPUSHs (ST (0));
994}
995 1258
996void filter_cbor_single_key_object (CBOR *self, SV *key, SV *cb = &PL_sv_undef) 1259SV *get_filter (CBOR *self)
997 PPCODE: 1260 CODE:
998{ 1261 RETVAL = self->filter ? self->filter : NEWSV (0, 0);
999 if (!self->cb_sk_object) 1262 OUTPUT:
1000 self->cb_sk_object = newHV (); 1263 RETVAL
1001
1002 if (SvOK (cb))
1003 hv_store_ent (self->cb_sk_object, key, newSVsv (cb), 0);
1004 else
1005 {
1006 hv_delete_ent (self->cb_sk_object, key, G_DISCARD, 0);
1007
1008 if (!HvKEYS (self->cb_sk_object))
1009 {
1010 SvREFCNT_dec (self->cb_sk_object);
1011 self->cb_sk_object = 0;
1012 }
1013 }
1014
1015 XPUSHs (ST (0));
1016}
1017
1018#endif
1019 1264
1020void encode (CBOR *self, SV *scalar) 1265void encode (CBOR *self, SV *scalar)
1021 PPCODE: 1266 PPCODE:
1022 PUTBACK; scalar = encode_cbor (scalar, self); SPAGAIN; 1267 PUTBACK; scalar = encode_cbor (scalar, self); SPAGAIN;
1023 XPUSHs (scalar); 1268 XPUSHs (scalar);
1036 EXTEND (SP, 2); 1281 EXTEND (SP, 2);
1037 PUSHs (sv); 1282 PUSHs (sv);
1038 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1283 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1039} 1284}
1040 1285
1041#if 0
1042
1043void DESTROY (CBOR *self) 1286void DESTROY (CBOR *self)
1044 CODE: 1287 PPCODE:
1045 SvREFCNT_dec (self->cb_sk_object); 1288 cbor_free (self);
1046 SvREFCNT_dec (self->cb_object);
1047
1048#endif
1049 1289
1050PROTOTYPES: ENABLE 1290PROTOTYPES: ENABLE
1051 1291
1052void encode_cbor (SV *scalar) 1292void encode_cbor (SV *scalar)
1053 PPCODE: 1293 PPCODE:

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines