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 1.20 by root, Wed Nov 20 11:06:42 2013 UTC vs.
Revision 1.47 by root, Mon Apr 27 20:21:53 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
14// compatibility with perl <5.18 15// compatibility with perl <5.18
15#ifndef HvNAMELEN_get 16#ifndef HvNAMELEN_get
16# define HvNAMELEN_get(hv) strlen (HvNAME (hv)) 17# define HvNAMELEN_get(hv) strlen (HvNAME (hv))
19# define HvNAMELEN(hv) HvNAMELEN_get (hv) 20# define HvNAMELEN(hv) HvNAMELEN_get (hv)
20#endif 21#endif
21#ifndef HvNAMEUTF8 22#ifndef HvNAMEUTF8
22# define HvNAMEUTF8(hv) 0 23# define HvNAMEUTF8(hv) 0
23#endif 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};
24 65
25// known tags 66// known tags
26enum cbor_tag 67enum cbor_tag
27{ 68{
28 // inofficial extensions (pending iana registration) 69 // extensions
70 CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
29 CBOR_TAG_PERL_OBJECT = 24, // http://cbor.schmorp.de/perl-object 71 CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
30 CBOR_TAG_GENERIC_OBJECT = 25, // http://cbor.schmorp.de/generic-object 72 CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
31 CBOR_TAG_VALUE_SHAREABLE = 26, // http://cbor.schmorp.de/value-sharing 73 CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
32 CBOR_TAG_VALUE_SHAREDREF = 27, // http://cbor.schmorp.de/value-sharing 74 CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
33 CBOR_TAG_STRINGREF_NAMESPACE = 65537, // http://cbor.schmorp.de/stringref 75 CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref
34 CBOR_TAG_STRINGREF = 28, // http://cbor.schmorp.de/stringref
35 CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection 76 CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
36 77
37 // rfc7049 78 // rfc7049
38 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8 79 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
39 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any 80 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
40 CBOR_TAG_POS_BIGNUM = 2, // byte string 81 CBOR_TAG_POS_BIGNUM = 2, // byte string
41 CBOR_TAG_NEG_BIGNUM = 3, // byte string 82 CBOR_TAG_NEG_BIGNUM = 3, // byte string
42 CBOR_TAG_DECIMAL = 4, // decimal fraction, array 83 CBOR_TAG_DECIMAL = 4, // decimal fraction, array
43 CBOR_TAG_BIGFLOAT = 5, // array 84 CBOR_TAG_BIGFLOAT = 5, // array
44 85
45 CBOR_TAG_CONV_B64U = 21, // base64url, any 86 CBOR_TAG_CONV_B64U = 21, // base64url, any
46 CBOR_TAG_CONV_B64 = 22, // base64, any 87 CBOR_TAG_CONV_B64 = 22, // base64, any
47 CBOR_TAG_CONV_HEX = 23, // base16, any 88 CBOR_TAG_CONV_HEX = 23, // base16, any
48 CBOR_TAG_CBOR = 24, // embedded cbor, byte string 89 CBOR_TAG_CBOR = 24, // embedded cbor, byte string
49 90
50 CBOR_TAG_URI = 32, // URI rfc3986, utf-8 91 CBOR_TAG_URI = 32, // URI rfc3986, utf-8
51 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8 92 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
52 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8 93 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
53 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8 94 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
54 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8 95 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
55 96
56 CBOR_TAG_MAGIC = 55799 // self-describe cbor 97 CBOR_TAG_MAGIC = 55799, // self-describe cbor
57}; 98};
58 99
59#define F_SHRINK 0x00000001UL 100#define F_SHRINK 0x00000001UL
60#define F_ALLOW_UNKNOWN 0x00000002UL 101#define F_ALLOW_UNKNOWN 0x00000002UL
61#define F_ALLOW_SHARING 0x00000004UL //TODO 102#define F_ALLOW_SHARING 0x00000004UL
103#define F_ALLOW_CYCLES 0x00000008UL
62#define F_DEDUP_STRINGS 0x00000008UL //TODO 104#define F_PACK_STRINGS 0x00000010UL
63#define F_DEDUP_KEYS 0x00000010UL //TODO 105#define F_VALIDATE_UTF8 0x00000020UL
64 106
65#define INIT_SIZE 32 // initial scalar size to be allocated 107#define INIT_SIZE 32 // initial scalar size to be allocated
66 108
67#define SB do { 109#define SB do {
68#define SE } while (0) 110#define SE } while (0)
80# define CBOR_SLOW 0 122# define CBOR_SLOW 0
81# define CBOR_STASH cbor_stash 123# define CBOR_STASH cbor_stash
82#endif 124#endif
83 125
84static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS:: 126static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS::
85static SV *types_true, *types_false, *types_error, *sv_cbor; 127static SV *types_true, *types_false, *types_error, *sv_cbor, *default_filter;
86 128
87typedef struct { 129typedef struct {
88 U32 flags; 130 U32 flags;
89 U32 max_depth; 131 U32 max_depth;
90 STRLEN max_size; 132 STRLEN max_size;
133 SV *filter;
134
135 // for the incremental parser
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.
91} CBOR; 139} CBOR;
92 140
93ecb_inline void 141ecb_inline void
94cbor_init (CBOR *cbor) 142cbor_init (CBOR *cbor)
95{ 143{
96 Zero (cbor, 1, CBOR); 144 Zero (cbor, 1, CBOR);
97 cbor->max_depth = 512; 145 cbor->max_depth = 512;
146}
147
148ecb_inline void
149cbor_free (CBOR *cbor)
150{
151 SvREFCNT_dec (cbor->filter);
152 SvREFCNT_dec (cbor->incr_count);
98} 153}
99 154
100///////////////////////////////////////////////////////////////////////////// 155/////////////////////////////////////////////////////////////////////////////
101// utility functions 156// utility functions
102 157
122 SvPV_shrink_to_cur (sv); 177 SvPV_shrink_to_cur (sv);
123#elif defined (SvPV_renew) 178#elif defined (SvPV_renew)
124 SvPV_renew (sv, SvCUR (sv) + 1); 179 SvPV_renew (sv, SvCUR (sv) + 1);
125#endif 180#endif
126 } 181 }
182}
183
184// minimum length of a string to be registered for stringref
185ecb_inline int
186minimum_string_length (UV idx)
187{
188 return idx > 23
189 ? idx > 0xffU
190 ? idx > 0xffffU
191 ? idx > 0xffffffffU
192 ? 11
193 : 7
194 : 5
195 : 4
196 : 3;
127} 197}
128 198
129///////////////////////////////////////////////////////////////////////////// 199/////////////////////////////////////////////////////////////////////////////
130// encoder 200// encoder
131 201
165static void 235static void
166encode_uint (enc_t *enc, int major, UV len) 236encode_uint (enc_t *enc, int major, UV len)
167{ 237{
168 need (enc, 9); 238 need (enc, 9);
169 239
170 if (len < 24) 240 if (ecb_expect_true (len < LENGTH_EXT1))
171 *enc->cur++ = major | len; 241 *enc->cur++ = major | len;
172 else if (len <= 0xff) 242 else if (ecb_expect_true (len <= 0xffU))
173 { 243 {
174 *enc->cur++ = major | 24; 244 *enc->cur++ = major | LENGTH_EXT1;
175 *enc->cur++ = len; 245 *enc->cur++ = len;
176 } 246 }
177 else if (len <= 0xffff) 247 else if (len <= 0xffffU)
178 { 248 {
179 *enc->cur++ = major | 25; 249 *enc->cur++ = major | LENGTH_EXT2;
180 *enc->cur++ = len >> 8; 250 *enc->cur++ = len >> 8;
181 *enc->cur++ = len; 251 *enc->cur++ = len;
182 } 252 }
183 else if (len <= 0xffffffff) 253 else if (len <= 0xffffffffU)
184 { 254 {
185 *enc->cur++ = major | 26; 255 *enc->cur++ = major | LENGTH_EXT4;
186 *enc->cur++ = len >> 24; 256 *enc->cur++ = len >> 24;
187 *enc->cur++ = len >> 16; 257 *enc->cur++ = len >> 16;
188 *enc->cur++ = len >> 8; 258 *enc->cur++ = len >> 8;
189 *enc->cur++ = len; 259 *enc->cur++ = len;
190 } 260 }
191 else 261 else
192 { 262 {
193 *enc->cur++ = major | 27; 263 *enc->cur++ = major | LENGTH_EXT8;
194 *enc->cur++ = len >> 56; 264 *enc->cur++ = len >> 56;
195 *enc->cur++ = len >> 48; 265 *enc->cur++ = len >> 48;
196 *enc->cur++ = len >> 40; 266 *enc->cur++ = len >> 40;
197 *enc->cur++ = len >> 32; 267 *enc->cur++ = len >> 32;
198 *enc->cur++ = len >> 24; 268 *enc->cur++ = len >> 24;
200 *enc->cur++ = len >> 8; 270 *enc->cur++ = len >> 8;
201 *enc->cur++ = len; 271 *enc->cur++ = len;
202 } 272 }
203} 273}
204 274
205static void 275ecb_inline void
276encode_tag (enc_t *enc, UV tag)
277{
278 encode_uint (enc, MAJOR_TAG, tag);
279}
280
281ecb_inline void
206encode_str (enc_t *enc, int utf8, char *str, STRLEN len) 282encode_str (enc_t *enc, int utf8, char *str, STRLEN len)
207{ 283{
208 encode_uint (enc, utf8 ? 0x60 : 0x40, len); 284 encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
209 need (enc, len); 285 need (enc, len);
210 memcpy (enc->cur, str, len); 286 memcpy (enc->cur, str, len);
211 enc->cur += len; 287 enc->cur += len;
212} 288}
213 289
214ecb_inline void 290static void
215encode_tag (enc_t *enc, UV tag) 291encode_strref (enc_t *enc, int utf8, char *str, STRLEN len)
216{ 292{
217 encode_uint (enc, 0xc0, tag); 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);
218} 313}
219 314
220static void encode_sv (enc_t *enc, SV *sv); 315static void encode_sv (enc_t *enc, SV *sv);
221 316
222static void 317static void
227 if (enc->depth >= enc->cbor.max_depth) 322 if (enc->depth >= enc->cbor.max_depth)
228 croak (ERR_NESTING_EXCEEDED); 323 croak (ERR_NESTING_EXCEEDED);
229 324
230 ++enc->depth; 325 ++enc->depth;
231 326
232 encode_uint (enc, 0x80, len + 1); 327 encode_uint (enc, MAJOR_ARRAY, len + 1);
233 328
329 if (SvMAGICAL (av))
234 for (i = 0; i <= len; ++i) 330 for (i = 0; i <= len; ++i)
235 { 331 {
236 SV **svp = av_fetch (av, i, 0); 332 SV **svp = av_fetch (av, i, 0);
237 encode_sv (enc, svp ? *svp : &PL_sv_undef); 333 encode_sv (enc, svp ? *svp : &PL_sv_undef);
238 } 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 }
239 341
240 --enc->depth; 342 --enc->depth;
241}
242
243ecb_inline void
244encode_he (enc_t *enc, HE *he)
245{
246 if (HeKLEN (he) == HEf_SVKEY)
247 encode_sv (enc, HeSVKEY (he));
248 else
249 encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
250} 343}
251 344
252static void 345static void
253encode_hv (enc_t *enc, HV *hv) 346encode_hv (enc_t *enc, HV *hv)
254{ 347{
261 354
262 int pairs = hv_iterinit (hv); 355 int pairs = hv_iterinit (hv);
263 int mg = SvMAGICAL (hv); 356 int mg = SvMAGICAL (hv);
264 357
265 if (mg) 358 if (mg)
266 encode_ch (enc, 0xa0 | 31); 359 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
267 else 360 else
268 encode_uint (enc, 0xa0, pairs); 361 encode_uint (enc, MAJOR_MAP, pairs);
269 362
270 while ((he = hv_iternext (hv))) 363 while ((he = hv_iternext (hv)))
271 { 364 {
272 if (ecb_expect_false (enc->cbor.flags & (F_DEDUP_STRINGS | F_DEDUP_KEYS)))
273 {
274 SV **svp;
275
276 if (HeKLEN (he) == HEf_SVKEY) 365 if (HeKLEN (he) == HEf_SVKEY)
277 svp = hv_fetch_ent (enc->stringref[!! SvUTF8 (HeSVKEY (he))], HeSVKEY (he) , 1, 0);//TODO return HE :/ 366 encode_sv (enc, HeSVKEY (he));
278 else
279 svp = hv_fetch (enc->stringref[!! HeKUTF8 (he) ], HeKEY (he), HeKLEN (he), 1);
280
281 if (SvOK (*svp))
282 {
283 encode_tag (enc, CBOR_TAG_STRINGREF);
284 encode_uint (enc, 0x00, SvUV (*svp));
285 }
286 else
287 {
288 sv_setuv (*svp, enc->stringref_idx);
289 ++enc->stringref_idx;
290 encode_he (enc, he);
291 }
292 }
293 else 367 else
294 encode_he (enc, he); 368 encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he));
295 369
296 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); 370 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
297 } 371 }
298 372
299 if (mg) 373 if (mg)
300 encode_ch (enc, 0xe0 | 31); 374 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
301 375
302 --enc->depth; 376 --enc->depth;
303} 377}
304 378
305// encode objects, arrays and special \0=false and \1=true values. 379// encode objects, arrays and special \0=false and \1=true values.
306static void 380static void
307encode_rv (enc_t *enc, SV *sv) 381encode_rv (enc_t *enc, SV *sv)
308{ 382{
309 SvGETMAGIC (sv); 383 SvGETMAGIC (sv);
310
311 if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)
312 && ecb_expect_false (SvREFCNT (sv) > 1))
313 {
314 if (!enc->shareable)
315 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
316
317 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
318
319 if (SvOK (*svp))
320 {
321 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
322 encode_uint (enc, 0x00, SvUV (*svp));
323 return;
324 }
325 else
326 {
327 sv_setuv (*svp, enc->shareable_idx);
328 ++enc->shareable_idx;
329 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
330 }
331 }
332 384
333 svtype svt = SvTYPE (sv); 385 svtype svt = SvTYPE (sv);
334 386
335 if (ecb_expect_false (SvOBJECT (sv))) 387 if (ecb_expect_false (SvOBJECT (sv)))
336 { 388 {
343 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 395 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
344 ? cbor_tagged_stash 396 ? cbor_tagged_stash
345 : gv_stashpv ("CBOR::XS::Tagged" , 1); 397 : gv_stashpv ("CBOR::XS::Tagged" , 1);
346 398
347 HV *stash = SvSTASH (sv); 399 HV *stash = SvSTASH (sv);
348 GV *method;
349 400
350 if (stash == boolean_stash) 401 if (stash == boolean_stash)
351 encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20); 402 {
403 encode_ch (enc, SvIV (sv) ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE);
404 return;
405 }
352 else if (stash == error_stash) 406 else if (stash == error_stash)
353 encode_ch (enc, 0xe0 | 23); 407 {
408 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
409 return;
410 }
354 else if (stash == tagged_stash) 411 else if (stash == tagged_stash)
355 { 412 {
356 if (svt != SVt_PVAV) 413 if (svt != SVt_PVAV)
357 croak ("encountered CBOR::XS::Tagged object that isn't an array"); 414 croak ("encountered CBOR::XS::Tagged object that isn't an array");
358 415
359 encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); 416 encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
360 encode_sv (enc, *av_fetch ((AV *)sv, 1, 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))
361 } 432 {
433 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
434 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
435 return;
436 }
437 else
438 {
439 sv_setuv (*svp, enc->shareable_idx);
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
362 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) 450 if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
363 { 451 {
364 dSP; 452 dSP;
365 453
366 ENTER; SAVETMPS; PUSHMARK (SP); 454 ENTER; SAVETMPS; PUSHMARK (SP);
367 // 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
399 // catch this surprisingly common error 487 // catch this surprisingly common error
400 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) 488 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
401 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash)); 489 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
402 490
403 encode_tag (enc, CBOR_TAG_PERL_OBJECT); 491 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
404 encode_uint (enc, 0x80, count + 1); 492 encode_uint (enc, MAJOR_ARRAY, count + 1);
405 encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); 493 encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
406 494
407 while (count) 495 while (count)
408 encode_sv (enc, SP[1 - count--]); 496 encode_sv (enc, SP[1 - count--]);
409 497
410 PUTBACK; 498 PUTBACK;
431{ 519{
432 double nv = SvNVX (sv); 520 double nv = SvNVX (sv);
433 521
434 need (enc, 9); 522 need (enc, 9);
435 523
436 if (ecb_expect_false (nv == (U32)nv)) 524 if (ecb_expect_false (nv == (NV)(U32)nv))
437 encode_uint (enc, 0x00, (U32)nv); 525 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
438 //TODO: maybe I32? 526 //TODO: maybe I32?
439 else if (ecb_expect_false (nv == (float)nv)) 527 else if (ecb_expect_false (nv == (float)nv))
440 { 528 {
441 uint32_t fp = ecb_float_to_binary32 (nv); 529 uint32_t fp = ecb_float_to_binary32 (nv);
442 530
443 *enc->cur++ = 0xe0 | 26; 531 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
444 532
445 if (!ecb_big_endian ()) 533 if (!ecb_big_endian ())
446 fp = ecb_bswap32 (fp); 534 fp = ecb_bswap32 (fp);
447 535
448 memcpy (enc->cur, &fp, 4); 536 memcpy (enc->cur, &fp, 4);
450 } 538 }
451 else 539 else
452 { 540 {
453 uint64_t fp = ecb_double_to_binary64 (nv); 541 uint64_t fp = ecb_double_to_binary64 (nv);
454 542
455 *enc->cur++ = 0xe0 | 27; 543 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
456 544
457 if (!ecb_big_endian ()) 545 if (!ecb_big_endian ())
458 fp = ecb_bswap64 (fp); 546 fp = ecb_bswap64 (fp);
459 547
460 memcpy (enc->cur, &fp, 8); 548 memcpy (enc->cur, &fp, 8);
469 557
470 if (SvPOKp (sv)) 558 if (SvPOKp (sv))
471 { 559 {
472 STRLEN len; 560 STRLEN len;
473 char *str = SvPV (sv, len); 561 char *str = SvPV (sv, len);
474 encode_str (enc, SvUTF8 (sv), str, len); 562 encode_strref (enc, SvUTF8 (sv), str, len);
475 } 563 }
476 else if (SvNOKp (sv)) 564 else if (SvNOKp (sv))
477 encode_nv (enc, sv); 565 encode_nv (enc, sv);
478 else if (SvIOKp (sv)) 566 else if (SvIOKp (sv))
479 { 567 {
480 if (SvIsUV (sv)) 568 if (SvIsUV (sv))
481 encode_uint (enc, 0x00, SvUVX (sv)); 569 encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
482 else if (SvIVX (sv) >= 0) 570 else if (SvIVX (sv) >= 0)
483 encode_uint (enc, 0x00, SvIVX (sv)); 571 encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
484 else 572 else
485 encode_uint (enc, 0x20, -(SvIVX (sv) + 1)); 573 encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
486 } 574 }
487 else if (SvROK (sv)) 575 else if (SvROK (sv))
488 encode_rv (enc, SvRV (sv)); 576 encode_rv (enc, SvRV (sv));
489 else if (!SvOK (sv)) 577 else if (!SvOK (sv))
490 encode_ch (enc, 0xe0 | 22); 578 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
491 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 579 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
492 encode_ch (enc, 0xe0 | 23); 580 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
493 else 581 else
494 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",
495 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 583 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
496} 584}
497 585
505 enc.cur = SvPVX (enc.sv); 593 enc.cur = SvPVX (enc.sv);
506 enc.end = SvEND (enc.sv); 594 enc.end = SvEND (enc.sv);
507 595
508 SvPOK_only (enc.sv); 596 SvPOK_only (enc.sv);
509 597
510 if (cbor->flags & (F_DEDUP_STRINGS | F_DEDUP_KEYS)) 598 if (cbor->flags & F_PACK_STRINGS)
511 { 599 {
512 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE); 600 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
513 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ()); 601 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
514 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ()); 602 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
515 } 603 }
537 CBOR cbor; 625 CBOR cbor;
538 U32 depth; // recursion depth 626 U32 depth; // recursion depth
539 U32 maxdepth; // recursion depth limit 627 U32 maxdepth; // recursion depth limit
540 AV *shareable; 628 AV *shareable;
541 AV *stringref; 629 AV *stringref;
630 SV *decode_tagged;
542} dec_t; 631} dec_t;
543 632
544#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
545 634
546#define WANT(len) if (ecb_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")
549#define DEC_DEC_DEPTH --dec->depth 638#define DEC_DEC_DEPTH --dec->depth
550 639
551static UV 640static UV
552decode_uint (dec_t *dec) 641decode_uint (dec_t *dec)
553{ 642{
554 switch (*dec->cur & 31) 643 U8 m = *dec->cur & MINOR_MASK;
555 { 644 ++dec->cur;
556 case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
557 case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
558 case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
559 return *dec->cur++ & 31;
560 645
561 case 24: 646 if (ecb_expect_true (m < LENGTH_EXT1))
647 return m;
648 else if (ecb_expect_true (m == LENGTH_EXT1))
649 {
562 WANT (2); 650 WANT (1);
563 dec->cur += 2; 651 dec->cur += 1;
564 return dec->cur[-1]; 652 return dec->cur[-1];
565 653 }
566 case 25: 654 else if (ecb_expect_true (m == LENGTH_EXT2))
655 {
567 WANT (3); 656 WANT (2);
568 dec->cur += 3; 657 dec->cur += 2;
569 return (((UV)dec->cur[-2]) << 8) 658 return (((UV)dec->cur[-2]) << 8)
570 | ((UV)dec->cur[-1]); 659 | ((UV)dec->cur[-1]);
571 660 }
572 case 26: 661 else if (ecb_expect_true (m == LENGTH_EXT4))
662 {
573 WANT (5); 663 WANT (4);
574 dec->cur += 5; 664 dec->cur += 4;
575 return (((UV)dec->cur[-4]) << 24) 665 return (((UV)dec->cur[-4]) << 24)
576 | (((UV)dec->cur[-3]) << 16) 666 | (((UV)dec->cur[-3]) << 16)
577 | (((UV)dec->cur[-2]) << 8) 667 | (((UV)dec->cur[-2]) << 8)
578 | ((UV)dec->cur[-1]); 668 | ((UV)dec->cur[-1]);
579 669 }
580 case 27: 670 else if (ecb_expect_true (m == LENGTH_EXT8))
671 {
581 WANT (9); 672 WANT (8);
582 dec->cur += 9; 673 dec->cur += 8;
674
675 return
676#if UVSIZE < 8
677 0
678#else
583 return (((UV)dec->cur[-8]) << 56) 679 (((UV)dec->cur[-8]) << 56)
584 | (((UV)dec->cur[-7]) << 48) 680 | (((UV)dec->cur[-7]) << 48)
585 | (((UV)dec->cur[-6]) << 40) 681 | (((UV)dec->cur[-6]) << 40)
586 | (((UV)dec->cur[-5]) << 32) 682 | (((UV)dec->cur[-5]) << 32)
683#endif
587 | (((UV)dec->cur[-4]) << 24) 684 | (((UV)dec->cur[-4]) << 24)
588 | (((UV)dec->cur[-3]) << 16) 685 | (((UV)dec->cur[-3]) << 16)
589 | (((UV)dec->cur[-2]) << 8) 686 | (((UV)dec->cur[-2]) << 8)
590 | ((UV)dec->cur[-1]); 687 | ((UV)dec->cur[-1]);
591 688 }
592 default: 689 else
593 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 690 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
594 }
595 691
596fail: 692fail:
597 return 0; 693 return 0;
598} 694}
599 695
604{ 700{
605 AV *av = newAV (); 701 AV *av = newAV ();
606 702
607 DEC_INC_DEPTH; 703 DEC_INC_DEPTH;
608 704
609 if ((*dec->cur & 31) == 31) 705 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
610 { 706 {
611 ++dec->cur; 707 ++dec->cur;
612 708
613 for (;;) 709 for (;;)
614 { 710 {
615 WANT (1); 711 WANT (1);
616 712
617 if (*dec->cur == (0xe0 | 31)) 713 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
618 { 714 {
619 ++dec->cur; 715 ++dec->cur;
620 break; 716 break;
621 } 717 }
622 718
625 } 721 }
626 else 722 else
627 { 723 {
628 int i, len = decode_uint (dec); 724 int i, len = decode_uint (dec);
629 725
726 WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
630 av_fill (av, len - 1); 727 av_fill (av, len - 1);
631 728
632 for (i = 0; i < len; ++i) 729 for (i = 0; i < len; ++i)
633 AvARRAY (av)[i] = decode_sv (dec); 730 AvARRAY (av)[i] = decode_sv (dec);
634 } 731 }
644 741
645static void 742static void
646decode_he (dec_t *dec, HV *hv) 743decode_he (dec_t *dec, HV *hv)
647{ 744{
648 // for speed reasons, we specialcase single-string 745 // for speed reasons, we specialcase single-string
649 // byte or utf-8 strings as keys. 746 // byte or utf-8 strings as keys, but only when !stringref
650 747
651 if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27) 748 if (ecb_expect_true (!dec->stringref))
749 if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
652 { 750 {
653 I32 len = decode_uint (dec); 751 I32 len = decode_uint (dec);
654 char *key = (char *)dec->cur; 752 char *key = (char *)dec->cur;
655 753
656 dec->cur += len; 754 dec->cur += len;
657 755
658 if (ecb_expect_false (dec->stringref))
659 av_push (dec->stringref, newSVpvn (key, len));
660
661 hv_store (hv, key, len, decode_sv (dec), 0); 756 hv_store (hv, key, len, decode_sv (dec), 0);
757
758 return;
662 } 759 }
663 else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27) 760 else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
664 { 761 {
665 I32 len = decode_uint (dec); 762 I32 len = decode_uint (dec);
666 char *key = (char *)dec->cur; 763 char *key = (char *)dec->cur;
667 764
668 dec->cur += len; 765 dec->cur += len;
669 766
670 if (ecb_expect_false (dec->stringref)) 767 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
671 av_push (dec->stringref, newSVpvn_utf8 (key, len, 1)); 768 if (!is_utf8_string (key, len))
769 ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
672 770
673 hv_store (hv, key, -len, decode_sv (dec), 0); 771 hv_store (hv, key, -len, decode_sv (dec), 0);
772
773 return;
674 } 774 }
675 else 775
676 {
677 SV *k = decode_sv (dec); 776 SV *k = decode_sv (dec);
678 SV *v = decode_sv (dec); 777 SV *v = decode_sv (dec);
679 778
680 hv_store_ent (hv, k, v, 0); 779 hv_store_ent (hv, k, v, 0);
681 SvREFCNT_dec (k); 780 SvREFCNT_dec (k);
682 } 781
782fail:
783 ;
683} 784}
684 785
685static SV * 786static SV *
686decode_hv (dec_t *dec) 787decode_hv (dec_t *dec)
687{ 788{
688 HV *hv = newHV (); 789 HV *hv = newHV ();
689 790
690 DEC_INC_DEPTH; 791 DEC_INC_DEPTH;
691 792
692 if ((*dec->cur & 31) == 31) 793 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
693 { 794 {
694 ++dec->cur; 795 ++dec->cur;
695 796
696 for (;;) 797 for (;;)
697 { 798 {
698 WANT (1); 799 WANT (1);
699 800
700 if (*dec->cur == (0xe0 | 31)) 801 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
701 { 802 {
702 ++dec->cur; 803 ++dec->cur;
703 break; 804 break;
704 } 805 }
705 806
726static SV * 827static SV *
727decode_str (dec_t *dec, int utf8) 828decode_str (dec_t *dec, int utf8)
728{ 829{
729 SV *sv = 0; 830 SV *sv = 0;
730 831
731 if ((*dec->cur & 31) == 31) 832 if ((*dec->cur & MINOR_MASK) == MINOR_INDEF)
732 { 833 {
834 // indefinite length strings
733 ++dec->cur; 835 ++dec->cur;
734 836
837 U8 major = *dec->cur & MAJOR_MISC;
838
735 sv = newSVpvn ("", 0); 839 sv = newSVpvn ("", 0);
736 840
737 // not very fast, and certainly not robust against illegal input
738 for (;;) 841 for (;;)
739 { 842 {
740 WANT (1); 843 WANT (1);
741 844
742 if (*dec->cur == (0xe0 | 31)) 845 if ((*dec->cur - major) > LENGTH_EXT8)
846 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
743 { 847 {
744 ++dec->cur; 848 ++dec->cur;
745 break; 849 break;
746 } 850 }
851 else
852 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
747 853
748 sv_catsv (sv, decode_sv (dec)); 854 STRLEN len = decode_uint (dec);
855
856 WANT (len);
857 sv_catpvn (sv, dec->cur, len);
858 dec->cur += len;
749 } 859 }
750 } 860 }
751 else 861 else
752 { 862 {
753 STRLEN len = decode_uint (dec); 863 STRLEN len = decode_uint (dec);
754 864
755 WANT (len); 865 WANT (len);
756 sv = newSVpvn (dec->cur, len); 866 sv = newSVpvn (dec->cur, len);
757 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));
758 } 872 }
759 873
760 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
761 SvUTF8_on (sv); 880 SvUTF8_on (sv);
762 881 }
763 if (ecb_expect_false (dec->stringref))
764 av_push (dec->stringref, SvREFCNT_inc_NN (sv));
765 882
766 return sv; 883 return sv;
767 884
768fail: 885fail:
769 SvREFCNT_dec (sv); 886 SvREFCNT_dec (sv);
801 } 918 }
802 break; 919 break;
803 920
804 case CBOR_TAG_STRINGREF: 921 case CBOR_TAG_STRINGREF:
805 { 922 {
806 if ((*dec->cur >> 5) != 0) 923 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
807 ERR ("corrupted CBOR data (stringref index not an unsigned integer)"); 924 ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
808 925
809 UV idx = decode_uint (dec); 926 UV idx = decode_uint (dec);
810 927
811 if (!dec->stringref || (int)idx > AvFILLp (dec->stringref)) 928 if (!dec->stringref || (int)idx > AvFILLp (dec->stringref))
818 case CBOR_TAG_VALUE_SHAREABLE: 935 case CBOR_TAG_VALUE_SHAREABLE:
819 { 936 {
820 if (ecb_expect_false (!dec->shareable)) 937 if (ecb_expect_false (!dec->shareable))
821 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ()); 938 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
822 939
940 if (dec->cbor.flags & F_ALLOW_CYCLES)
941 {
823 sv = newSV (0); 942 sv = newSV (0);
824 av_push (dec->shareable, SvREFCNT_inc_NN (sv)); 943 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
825 944
826 SV *osv = decode_sv (dec); 945 SV *osv = decode_sv (dec);
827 sv_setsv (sv, osv); 946 sv_setsv (sv, osv);
828 SvREFCNT_dec_NN (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 }
829 } 956 }
830 break; 957 break;
831 958
832 case CBOR_TAG_VALUE_SHAREDREF: 959 case CBOR_TAG_VALUE_SHAREDREF:
833 { 960 {
834 if ((*dec->cur >> 5) != 0) 961 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
835 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)"); 962 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
836 963
837 UV idx = decode_uint (dec); 964 UV idx = decode_uint (dec);
838 965
839 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable)) 966 if (!dec->shareable || (int)idx > AvFILLp (dec->shareable))
840 ERR ("corrupted CBOR data (sharedref index out of bounds)"); 967 ERR ("corrupted CBOR data (sharedref index out of bounds)");
841 968
842 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]); 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");
843 } 973 }
844 break; 974 break;
845 975
846 case CBOR_TAG_PERL_OBJECT: 976 case CBOR_TAG_PERL_OBJECT:
847 { 977 {
896 1026
897 default: 1027 default:
898 { 1028 {
899 sv = decode_sv (dec); 1029 sv = decode_sv (dec);
900 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 {
901 AV *av = newAV (); 1054 AV *av = newAV ();
902 av_push (av, newSVuv (tag)); 1055 av_push (av, newSVuv (tag));
903 av_push (av, sv); 1056 av_push (av, sv);
904 1057
905 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 1058 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
906 ? cbor_tagged_stash 1059 ? cbor_tagged_stash
907 : gv_stashpv ("CBOR::XS::Tagged" , 1); 1060 : gv_stashpv ("CBOR::XS::Tagged" , 1);
908
909 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash); 1061 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
1062 }
1063
1064 PUTBACK;
1065
1066 FREETMPS; LEAVE;
910 } 1067 }
911 break; 1068 break;
912 } 1069 }
913 1070
914 return sv; 1071 return sv;
921static SV * 1078static SV *
922decode_sv (dec_t *dec) 1079decode_sv (dec_t *dec)
923{ 1080{
924 WANT (1); 1081 WANT (1);
925 1082
926 switch (*dec->cur >> 5) 1083 switch (*dec->cur >> MAJOR_SHIFT)
927 { 1084 {
928 case 0: // unsigned int 1085 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
929 return newSVuv (decode_uint (dec)); 1086 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
930 case 1: // negative int 1087 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
931 return newSViv (-1 - (IV)decode_uint (dec)); 1088 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
932 case 2: // octet string 1089 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
933 return decode_str (dec, 0); 1090 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
934 case 3: // utf-8 string 1091 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
935 return decode_str (dec, 1); 1092
936 case 4: // array 1093 case MAJOR_MISC >> MAJOR_SHIFT:
937 return decode_av (dec);
938 case 5: // map
939 return decode_hv (dec);
940 case 6: // tag
941 return decode_tagged (dec);
942 case 7: // misc
943 switch (*dec->cur++ & 31) 1094 switch (*dec->cur++ & MINOR_MASK)
944 { 1095 {
945 case 20: 1096 case SIMPLE_FALSE:
946#if CBOR_SLOW 1097#if CBOR_SLOW
947 types_false = get_bool ("Types::Serialiser::false"); 1098 types_false = get_bool ("Types::Serialiser::false");
948#endif 1099#endif
949 return newSVsv (types_false); 1100 return newSVsv (types_false);
950 case 21: 1101 case SIMPLE_TRUE:
951#if CBOR_SLOW 1102#if CBOR_SLOW
952 types_true = get_bool ("Types::Serialiser::true"); 1103 types_true = get_bool ("Types::Serialiser::true");
953#endif 1104#endif
954 return newSVsv (types_true); 1105 return newSVsv (types_true);
955 case 22: 1106 case SIMPLE_NULL:
956 return newSVsv (&PL_sv_undef); 1107 return newSVsv (&PL_sv_undef);
957 case 23: 1108 case SIMPLE_UNDEF:
958#if CBOR_SLOW 1109#if CBOR_SLOW
959 types_error = get_bool ("Types::Serialiser::error"); 1110 types_error = get_bool ("Types::Serialiser::error");
960#endif 1111#endif
961 return newSVsv (types_error); 1112 return newSVsv (types_error);
962 1113
963 case 25: 1114 case MISC_FLOAT16:
964 { 1115 {
965 WANT (2); 1116 WANT (2);
966 1117
967 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1118 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
968 dec->cur += 2; 1119 dec->cur += 2;
969 1120
970 return newSVnv (ecb_binary16_to_float (fp)); 1121 return newSVnv (ecb_binary16_to_float (fp));
971 } 1122 }
972 1123
973 case 26: 1124 case MISC_FLOAT32:
974 { 1125 {
975 uint32_t fp; 1126 uint32_t fp;
976 WANT (4); 1127 WANT (4);
977 memcpy (&fp, dec->cur, 4); 1128 memcpy (&fp, dec->cur, 4);
978 dec->cur += 4; 1129 dec->cur += 4;
981 fp = ecb_bswap32 (fp); 1132 fp = ecb_bswap32 (fp);
982 1133
983 return newSVnv (ecb_binary32_to_float (fp)); 1134 return newSVnv (ecb_binary32_to_float (fp));
984 } 1135 }
985 1136
986 case 27: 1137 case MISC_FLOAT64:
987 { 1138 {
988 uint64_t fp; 1139 uint64_t fp;
989 WANT (8); 1140 WANT (8);
990 memcpy (&fp, dec->cur, 8); 1141 memcpy (&fp, dec->cur, 8);
991 dec->cur += 8; 1142 dec->cur += 8;
994 fp = ecb_bswap64 (fp); 1145 fp = ecb_bswap64 (fp);
995 1146
996 return newSVnv (ecb_binary64_to_double (fp)); 1147 return newSVnv (ecb_binary64_to_double (fp));
997 } 1148 }
998 1149
999 // 0..19 unassigned 1150 // 0..19 unassigned simple
1000 // 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
1001 default: 1154 default:
1002 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1155 ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
1003 } 1156 }
1004 1157
1005 break; 1158 break;
1006 } 1159 }
1007 1160
1034 if (dec.cur != dec.end && !dec.err) 1187 if (dec.cur != dec.end && !dec.err)
1035 dec.err = "garbage after CBOR object"; 1188 dec.err = "garbage after CBOR object";
1036 1189
1037 if (dec.err) 1190 if (dec.err)
1038 { 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
1039 SvREFCNT_dec (sv); 1203 SvREFCNT_dec (sv);
1040 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (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);
1041 } 1205 }
1042 1206
1043 sv = sv_2mortal (sv); 1207 sv = sv_2mortal (sv);
1044 1208
1045 return sv; 1209 return sv;
1046} 1210}
1047 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
1048///////////////////////////////////////////////////////////////////////////// 1336/////////////////////////////////////////////////////////////////////////////
1049// XS interface functions 1337// XS interface functions
1050 1338
1051MODULE = CBOR::XS PACKAGE = CBOR::XS 1339MODULE = CBOR::XS PACKAGE = CBOR::XS
1052 1340
1059 types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1); 1347 types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1);
1060 1348
1061 types_true = get_bool ("Types::Serialiser::true" ); 1349 types_true = get_bool ("Types::Serialiser::true" );
1062 types_false = get_bool ("Types::Serialiser::false"); 1350 types_false = get_bool ("Types::Serialiser::false");
1063 types_error = get_bool ("Types::Serialiser::error"); 1351 types_error = get_bool ("Types::Serialiser::error");
1352
1353 default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1064 1354
1065 sv_cbor = newSVpv ("CBOR", 0); 1355 sv_cbor = newSVpv ("CBOR", 0);
1066 SvREADONLY_on (sv_cbor); 1356 SvREADONLY_on (sv_cbor);
1067} 1357}
1068 1358
1090void shrink (CBOR *self, int enable = 1) 1380void shrink (CBOR *self, int enable = 1)
1091 ALIAS: 1381 ALIAS:
1092 shrink = F_SHRINK 1382 shrink = F_SHRINK
1093 allow_unknown = F_ALLOW_UNKNOWN 1383 allow_unknown = F_ALLOW_UNKNOWN
1094 allow_sharing = F_ALLOW_SHARING 1384 allow_sharing = F_ALLOW_SHARING
1095 dedup_keys = F_DEDUP_KEYS 1385 allow_cycles = F_ALLOW_CYCLES
1096 dedup_strings = F_DEDUP_STRINGS 1386 pack_strings = F_PACK_STRINGS
1387 validate_utf8 = F_VALIDATE_UTF8
1097 PPCODE: 1388 PPCODE:
1098{ 1389{
1099 if (enable) 1390 if (enable)
1100 self->flags |= ix; 1391 self->flags |= ix;
1101 else 1392 else
1107void get_shrink (CBOR *self) 1398void get_shrink (CBOR *self)
1108 ALIAS: 1399 ALIAS:
1109 get_shrink = F_SHRINK 1400 get_shrink = F_SHRINK
1110 get_allow_unknown = F_ALLOW_UNKNOWN 1401 get_allow_unknown = F_ALLOW_UNKNOWN
1111 get_allow_sharing = F_ALLOW_SHARING 1402 get_allow_sharing = F_ALLOW_SHARING
1112 get_dedup_keys = F_DEDUP_KEYS 1403 get_allow_cycles = F_ALLOW_CYCLES
1113 get_dedup_strings = F_DEDUP_STRINGS 1404 get_pack_strings = F_PACK_STRINGS
1405 get_validate_utf8 = F_VALIDATE_UTF8
1114 PPCODE: 1406 PPCODE:
1115 XPUSHs (boolSV (self->flags & ix)); 1407 XPUSHs (boolSV (self->flags & ix));
1116 1408
1117void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1409void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1118 PPCODE: 1410 PPCODE:
1131 XPUSHs (ST (0)); 1423 XPUSHs (ST (0));
1132 1424
1133int get_max_size (CBOR *self) 1425int get_max_size (CBOR *self)
1134 CODE: 1426 CODE:
1135 RETVAL = self->max_size; 1427 RETVAL = self->max_size;
1428 OUTPUT:
1429 RETVAL
1430
1431void filter (CBOR *self, SV *filter = 0)
1432 PPCODE:
1433 SvREFCNT_dec (self->filter);
1434 self->filter = filter ? newSVsv (filter) : filter;
1435 XPUSHs (ST (0));
1436
1437SV *get_filter (CBOR *self)
1438 CODE:
1439 RETVAL = self->filter ? self->filter : NEWSV (0, 0);
1136 OUTPUT: 1440 OUTPUT:
1137 RETVAL 1441 RETVAL
1138 1442
1139void encode (CBOR *self, SV *scalar) 1443void encode (CBOR *self, SV *scalar)
1140 PPCODE: 1444 PPCODE:
1155 EXTEND (SP, 2); 1459 EXTEND (SP, 2);
1156 PUSHs (sv); 1460 PUSHs (sv);
1157 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1461 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1158} 1462}
1159 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
1516void DESTROY (CBOR *self)
1517 PPCODE:
1518 cbor_free (self);
1519
1160PROTOTYPES: ENABLE 1520PROTOTYPES: ENABLE
1161 1521
1162void encode_cbor (SV *scalar) 1522void encode_cbor (SV *scalar)
1523 ALIAS:
1524 encode_cbor = 0
1525 encode_cbor_sharing = F_ALLOW_SHARING
1163 PPCODE: 1526 PPCODE:
1164{ 1527{
1165 CBOR cbor; 1528 CBOR cbor;
1166 cbor_init (&cbor); 1529 cbor_init (&cbor);
1530 cbor.flags |= ix;
1167 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1531 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1168 XPUSHs (scalar); 1532 XPUSHs (scalar);
1169} 1533}
1170 1534
1171void decode_cbor (SV *cborstr) 1535void decode_cbor (SV *cborstr)

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines