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

Comparing cvsroot/CBOR-XS/XS.xs (file contents):
Revision 1.18 by root, Wed Nov 20 01:09:46 2013 UTC vs.
Revision 1.66 by root, Sun Nov 29 21:32:01 2020 UTC

6#include <string.h> 6#include <string.h>
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#include <inttypes.h>
11 12
13#define ECB_NO_THREADS 1
12#include "ecb.h" 14#include "ecb.h"
13 15
14// compatibility with perl <5.18 16// compatibility with perl <5.18
15#ifndef HvNAMELEN_get 17#ifndef HvNAMELEN_get
16# define HvNAMELEN_get(hv) strlen (HvNAME (hv)) 18# define HvNAMELEN_get(hv) strlen (HvNAME (hv))
19# define HvNAMELEN(hv) HvNAMELEN_get (hv) 21# define HvNAMELEN(hv) HvNAMELEN_get (hv)
20#endif 22#endif
21#ifndef HvNAMEUTF8 23#ifndef HvNAMEUTF8
22# define HvNAMEUTF8(hv) 0 24# define HvNAMEUTF8(hv) 0
23#endif 25#endif
26#ifndef SvREFCNT_inc_NN
27# define SvREFCNT_inc_NN(sv) SvREFCNT_inc (sv)
28#endif
29#ifndef SvREFCNT_dec_NN
30# define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv)
31#endif
32
33// known major and minor types
34enum cbor_type
35{
36 MAJOR_SHIFT = 5,
37 MINOR_MASK = 0x1f,
38
39 MAJOR_POS_INT = 0 << MAJOR_SHIFT,
40 MAJOR_NEG_INT = 1 << MAJOR_SHIFT,
41 MAJOR_BYTES = 2 << MAJOR_SHIFT,
42 MAJOR_TEXT = 3 << MAJOR_SHIFT,
43 MAJOR_ARRAY = 4 << MAJOR_SHIFT,
44 MAJOR_MAP = 5 << MAJOR_SHIFT,
45 MAJOR_TAG = 6 << MAJOR_SHIFT,
46 MAJOR_MISC = 7 << MAJOR_SHIFT,
47
48 // INT/STRING/ARRAY/MAP subtypes
49 LENGTH_EXT1 = 24,
50 LENGTH_EXT2 = 25,
51 LENGTH_EXT4 = 26,
52 LENGTH_EXT8 = 27,
53
54 // SIMPLE types (effectively MISC subtypes)
55 SIMPLE_FALSE = 20,
56 SIMPLE_TRUE = 21,
57 SIMPLE_NULL = 22,
58 SIMPLE_UNDEF = 23,
59
60 // MISC subtype (unused)
61 MISC_EXT1 = 24,
62 MISC_FLOAT16 = 25,
63 MISC_FLOAT32 = 26,
64 MISC_FLOAT64 = 27,
65
66 // BYTES/TEXT/ARRAY/MAP
67 MINOR_INDEF = 31,
68};
24 69
25// known tags 70// known tags
26enum cbor_tag 71enum cbor_tag
27{ 72{
28 // inofficial extensions (pending iana registration) 73 // extensions
74 CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref
29 CBOR_TAG_PERL_OBJECT = 24, // http://cbor.schmorp.de/perl-object 75 CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object
30 CBOR_TAG_GENERIC_OBJECT = 25, // http://cbor.schmorp.de/generic-object 76 CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object
31 CBOR_TAG_VALUE_SHARABLE = 26, // http://cbor.schmorp.de/value-sharing 77 CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing
32 CBOR_TAG_VALUE_SHAREDREF = 27, // http://cbor.schmorp.de/value-sharing 78 CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing
33 CBOR_TAG_STRINGREF_NAMESPACE = 65537, // http://cbor.schmorp.de/stringref 79 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 80 CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection
36 81
37 // rfc7049 82 // rfc7049
38 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8 83 CBOR_TAG_DATETIME = 0, // rfc4287, utf-8
39 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any 84 CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any
40 CBOR_TAG_POS_BIGNUM = 2, // byte string 85 CBOR_TAG_POS_BIGNUM = 2, // byte string
41 CBOR_TAG_NEG_BIGNUM = 3, // byte string 86 CBOR_TAG_NEG_BIGNUM = 3, // byte string
42 CBOR_TAG_DECIMAL = 4, // decimal fraction, array 87 CBOR_TAG_DECIMAL = 4, // decimal fraction, array
43 CBOR_TAG_BIGFLOAT = 5, // array 88 CBOR_TAG_BIGFLOAT = 5, // array
44 89
45 CBOR_TAG_CONV_B64U = 21, // base64url, any 90 CBOR_TAG_CONV_B64U = 21, // base64url, any
46 CBOR_TAG_CONV_B64 = 22, // base64, any 91 CBOR_TAG_CONV_B64 = 22, // base64, any
47 CBOR_TAG_CONV_HEX = 23, // base16, any 92 CBOR_TAG_CONV_HEX = 23, // base16, any
48 CBOR_TAG_CBOR = 24, // embedded cbor, byte string 93 CBOR_TAG_CBOR = 24, // embedded cbor, byte string
49 94
50 CBOR_TAG_URI = 32, // URI rfc3986, utf-8 95 CBOR_TAG_URI = 32, // URI rfc3986, utf-8
51 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8 96 CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8
52 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8 97 CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8
53 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8 98 CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8
54 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8 99 CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8
55 100
56 CBOR_TAG_MAGIC = 55799 // self-describe cbor 101 CBOR_TAG_MAGIC = 55799, // self-describe cbor
57}; 102};
58 103
104// known forced types, also hardcoded in CBOR.pm
105enum
106{
107 AS_CBOR = 0,
108 AS_BYTES = 1,
109 AS_TEXT = 2,
110 AS_FLOAT16 = 3,
111 AS_FLOAT32 = 4,
112 AS_FLOAT64 = 5,
113 // possibly future enhancements: float, integer
114};
115
59#define F_SHRINK 0x00000001UL 116#define F_SHRINK 0x00000001UL
60#define F_ALLOW_UNKNOWN 0x00000002UL 117#define F_ALLOW_UNKNOWN 0x00000002UL
61#define F_ALLOW_SHARING 0x00000004UL //TODO 118#define F_ALLOW_SHARING 0x00000004UL
119#define F_ALLOW_CYCLES 0x00000008UL
120#define F_FORBID_OBJECTS 0x00000010UL
62#define F_DEDUP_STRINGS 0x00000008UL //TODO 121#define F_PACK_STRINGS 0x00000020UL
63#define F_DEDUP_KEYS 0x00000010UL //TODO 122#define F_TEXT_KEYS 0x00000040UL
123#define F_TEXT_STRINGS 0x00000080UL
124#define F_VALIDATE_UTF8 0x00000100UL
64 125
65#define INIT_SIZE 32 // initial scalar size to be allocated 126#define INIT_SIZE 32 // initial scalar size to be allocated
66 127
67#define SB do { 128#define SB do {
68#define SE } while (0) 129#define SE } while (0)
80# define CBOR_SLOW 0 141# define CBOR_SLOW 0
81# define CBOR_STASH cbor_stash 142# define CBOR_STASH cbor_stash
82#endif 143#endif
83 144
84static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS:: 145static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS::
85static SV *types_true, *types_false, *types_error, *sv_cbor; 146static SV *types_true, *types_false, *types_error, *sv_cbor, *default_filter;
86 147
87typedef struct { 148typedef struct {
88 U32 flags; 149 U32 flags;
89 U32 max_depth; 150 U32 max_depth;
90 STRLEN max_size; 151 STRLEN max_size;
152 SV *filter;
153
154 // for the incremental parser
155 STRLEN incr_pos; // the current offset into the text
156 STRLEN incr_need; // minimum bytes needed to decode
157 AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef.
91} CBOR; 158} CBOR;
92 159
93ecb_inline void 160ecb_inline void
94cbor_init (CBOR *cbor) 161cbor_init (CBOR *cbor)
95{ 162{
96 Zero (cbor, 1, CBOR); 163 Zero (cbor, 1, CBOR);
97 cbor->max_depth = 512; 164 cbor->max_depth = 512;
165}
166
167ecb_inline void
168cbor_free (CBOR *cbor)
169{
170 SvREFCNT_dec (cbor->filter);
171 SvREFCNT_dec (cbor->incr_count);
98} 172}
99 173
100///////////////////////////////////////////////////////////////////////////// 174/////////////////////////////////////////////////////////////////////////////
101// utility functions 175// utility functions
102 176
124 SvPV_renew (sv, SvCUR (sv) + 1); 198 SvPV_renew (sv, SvCUR (sv) + 1);
125#endif 199#endif
126 } 200 }
127} 201}
128 202
203// minimum length of a string to be registered for stringref
204ecb_inline int
205minimum_string_length (UV idx)
206{
207 return idx <= 23 ? 3
208 : idx <= 0xffU ? 4
209 : idx <= 0xffffU ? 5
210 : idx <= 0xffffffffU ? 7
211 : 11;
212}
213
129///////////////////////////////////////////////////////////////////////////// 214/////////////////////////////////////////////////////////////////////////////
130// encoder 215// encoder
131 216
132// structure used for encoding CBOR 217// structure used for encoding CBOR
133typedef struct 218typedef struct
135 char *cur; // SvPVX (sv) + current output position 220 char *cur; // SvPVX (sv) + current output position
136 char *end; // SvEND (sv) 221 char *end; // SvEND (sv)
137 SV *sv; // result scalar 222 SV *sv; // result scalar
138 CBOR cbor; 223 CBOR cbor;
139 U32 depth; // recursion level 224 U32 depth; // recursion level
140 HV *stringref; // string => index, or 0 225 HV *stringref[2]; // string => index, or 0 ([0] = bytes, [1] = utf-8)
226 UV stringref_idx;
141 HV *sharable; // ptr => index, or 0 227 HV *shareable; // ptr => index, or 0
142 HV *sharable_idx; 228 UV shareable_idx;
143} enc_t; 229} enc_t;
144 230
145ecb_inline void 231ecb_inline void
146need (enc_t *enc, STRLEN len) 232need (enc_t *enc, STRLEN len)
147{ 233{
148 if (ecb_expect_false (enc->cur + len >= enc->end)) 234 if (ecb_expect_false ((uintptr_t)(enc->end - enc->cur) < len))
149 { 235 {
150 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv); 236 STRLEN cur = enc->cur - (char *)SvPVX (enc->sv);
151 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1); 237 SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1);
152 enc->cur = SvPVX (enc->sv) + cur; 238 enc->cur = SvPVX (enc->sv) + cur;
153 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1; 239 enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1;
159{ 245{
160 need (enc, 1); 246 need (enc, 1);
161 *enc->cur++ = ch; 247 *enc->cur++ = ch;
162} 248}
163 249
250// used for tags, intregers, element counts and so on
164static void 251static void
165encode_uint (enc_t *enc, int major, UV len) 252encode_uint (enc_t *enc, int major, UV len)
166{ 253{
167 need (enc, 9); 254 need (enc, 9);
168 255
169 if (len < 24) 256 if (ecb_expect_true (len < LENGTH_EXT1))
170 *enc->cur++ = major | len; 257 *enc->cur++ = major | len;
171 else if (len <= 0xff) 258 else if (ecb_expect_true (len <= 0xffU))
172 { 259 {
173 *enc->cur++ = major | 24; 260 *enc->cur++ = major | LENGTH_EXT1;
174 *enc->cur++ = len; 261 *enc->cur++ = len;
175 } 262 }
176 else if (len <= 0xffff) 263 else if (len <= 0xffffU)
177 { 264 {
178 *enc->cur++ = major | 25; 265 *enc->cur++ = major | LENGTH_EXT2;
179 *enc->cur++ = len >> 8; 266 *enc->cur++ = len >> 8;
180 *enc->cur++ = len; 267 *enc->cur++ = len;
181 } 268 }
182 else if (len <= 0xffffffff) 269 else if (len <= 0xffffffffU)
183 { 270 {
184 *enc->cur++ = major | 26; 271 *enc->cur++ = major | LENGTH_EXT4;
185 *enc->cur++ = len >> 24; 272 *enc->cur++ = len >> 24;
186 *enc->cur++ = len >> 16; 273 *enc->cur++ = len >> 16;
187 *enc->cur++ = len >> 8; 274 *enc->cur++ = len >> 8;
188 *enc->cur++ = len; 275 *enc->cur++ = len;
189 } 276 }
190 else 277 else
191 { 278 {
192 *enc->cur++ = major | 27; 279 *enc->cur++ = major | LENGTH_EXT8;
193 *enc->cur++ = len >> 56; 280 *enc->cur++ = len >> 56;
194 *enc->cur++ = len >> 48; 281 *enc->cur++ = len >> 48;
195 *enc->cur++ = len >> 40; 282 *enc->cur++ = len >> 40;
196 *enc->cur++ = len >> 32; 283 *enc->cur++ = len >> 32;
197 *enc->cur++ = len >> 24; 284 *enc->cur++ = len >> 24;
199 *enc->cur++ = len >> 8; 286 *enc->cur++ = len >> 8;
200 *enc->cur++ = len; 287 *enc->cur++ = len;
201 } 288 }
202} 289}
203 290
204static void 291// encodes a perl value into a CBOR integer
292ecb_inline
293encode_int (enc_t *enc, SV *sv)
294{
295 if (SvIsUV (sv))
296 encode_uint (enc, MAJOR_POS_INT, SvUVX (sv));
297 else if (SvIVX (sv) >= 0)
298 encode_uint (enc, MAJOR_POS_INT, SvIVX (sv));
299 else
300 encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1));
301}
302
303ecb_inline void
304encode_tag (enc_t *enc, UV tag)
305{
306 encode_uint (enc, MAJOR_TAG, tag);
307}
308
309// exceptional (hopefully) slow path for byte strings that need to be utf8-encoded
310ecb_noinline static void
205encode_str (enc_t *enc, int utf8, char *str, STRLEN len) 311encode_str_utf8 (enc_t *enc, int utf8, char *str, STRLEN len)
206{ 312{
207 encode_uint (enc, utf8 ? 0x60 : 0x40, len); 313 STRLEN ulen = len;
314 U8 *p, *pend = (U8 *)str + len;
315
316 for (p = (U8 *)str; p < pend; ++p)
317 ulen += *p >> 7; // count set high bits
318
319 encode_uint (enc, MAJOR_TEXT, ulen);
320
321 need (enc, ulen);
322 for (p = (U8 *)str; p < pend; ++p)
323 if (*p < 0x80)
324 *enc->cur++ = *p;
325 else
326 {
327 *enc->cur++ = 0xc0 + (*p >> 6);
328 *enc->cur++ = 0x80 + (*p & 63);
329 }
330}
331
332ecb_inline void
333encode_str (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
334{
335 if (ecb_expect_false (upgrade_utf8))
336 if (!utf8)
337 {
338 encode_str_utf8 (enc, utf8, str, len);
339 return;
340 }
341
342 encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len);
208 need (enc, len); 343 need (enc, len);
209 memcpy (enc->cur, str, len); 344 memcpy (enc->cur, str, len);
210 enc->cur += len; 345 enc->cur += len;
211} 346}
212 347
213ecb_inline void 348ecb_inline void
214encode_tag (enc_t *enc, UV tag) 349encode_strref (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len)
215{ 350{
216 encode_uint (enc, 0xc0, tag); 351 if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS))
217} 352 {
353 SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1);
218 354
219static int
220encode_sharable2 (enc_t *enc, SV *sv)
221{
222 if (!enc->sharable)
223 enc->sharable = (HV *)sv_2mortal ((SV *)newHV ());
224
225 SV **svp = hv_fetch (enc->sharable, &sv, sizeof (sv), 1);
226
227 if (SvOK (*svp)) 355 if (SvOK (*svp))
228 { 356 {
357 // already registered, use stringref
229 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF); 358 encode_tag (enc, CBOR_TAG_STRINGREF);
230 encode_uint (enc, 0x00, SvUV (*svp)); 359 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
231
232 return 1; 360 return;
233 } 361 }
234 else 362 else if (len >= minimum_string_length (enc->stringref_idx))
235 { 363 {
364 // register only
236 sv_setuv (*svp, enc->sharable_idx++); 365 sv_setuv (*svp, enc->stringref_idx);
237 encode_tag (enc, CBOR_TAG_VALUE_SHARABLE); 366 ++enc->stringref_idx;
238 367 }
239 return 0;
240 } 368 }
241}
242 369
370 encode_str (enc, upgrade_utf8, utf8, str, len);
371}
372
243ecb_inline int 373ecb_inline void
374encode_float16 (enc_t *enc, NV nv)
375{
376 need (enc, 1+2);
377
378 *enc->cur++ = MAJOR_MISC | MISC_FLOAT16;
379
380 uint16_t fp = ecb_float_to_binary16 (nv);
381
382 if (!ecb_big_endian ())
383 fp = ecb_bswap16 (fp);
384
385 memcpy (enc->cur, &fp, 2);
386 enc->cur += 2;
387}
388
389ecb_inline void
390encode_float32 (enc_t *enc, NV nv)
391{
392 need (enc, 1+4);
393
394 *enc->cur++ = MAJOR_MISC | MISC_FLOAT32;
395
396 uint32_t fp = ecb_float_to_binary32 (nv);
397
398 if (!ecb_big_endian ())
399 fp = ecb_bswap32 (fp);
400
401 memcpy (enc->cur, &fp, 4);
402 enc->cur += 4;
403}
404
405ecb_inline void
406encode_float64 (enc_t *enc, NV nv)
407{
408 need (enc, 1+8);
409
410 *enc->cur++ = MAJOR_MISC | MISC_FLOAT64;
411
412 uint64_t fp = ecb_double_to_binary64 (nv);
413
414 if (!ecb_big_endian ())
415 fp = ecb_bswap64 (fp);
416
417 memcpy (enc->cur, &fp, 8);
418 enc->cur += 8;
419}
420
421ecb_inline void
244encode_sharable (enc_t *enc, SV *sv) 422encode_forced (enc_t *enc, UV type, SV *sv)
245{ 423{
246 if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING) 424 switch (type)
247 && ecb_expect_false (SvREFCNT (sv) > 1)) 425 {
248 return encode_sharable2 (enc, sv); 426 case AS_CBOR:
427 {
428 STRLEN len;
429 char *str = SvPVbyte (sv, len);
249 430
250 return 0; 431 need (enc, len);
432 memcpy (enc->cur, str, len);
433 enc->cur += len;
434 }
435 break;
436
437 case AS_BYTES:
438 {
439 STRLEN len;
440 char *str = SvPVbyte (sv, len);
441 encode_strref (enc, 0, 0, str, len);
442 }
443 break;
444
445 case AS_TEXT:
446 {
447 STRLEN len;
448 char *str = SvPVutf8 (sv, len);
449 encode_strref (enc, 1, 1, str, len);
450 }
451 break;
452
453 case AS_FLOAT16: encode_float16 (enc, SvNV (sv)); break;
454 case AS_FLOAT32: encode_float32 (enc, SvNV (sv)); break;
455 case AS_FLOAT64: encode_float64 (enc, SvNV (sv)); break;
456
457 default:
458 croak ("encountered malformed CBOR::XS::Tagged object");
459 }
251} 460}
252 461
253static void encode_sv (enc_t *enc, SV *sv); 462static void encode_sv (enc_t *enc, SV *sv);
254 463
255static void 464static void
260 if (enc->depth >= enc->cbor.max_depth) 469 if (enc->depth >= enc->cbor.max_depth)
261 croak (ERR_NESTING_EXCEEDED); 470 croak (ERR_NESTING_EXCEEDED);
262 471
263 ++enc->depth; 472 ++enc->depth;
264 473
265 encode_uint (enc, 0x80, len + 1); 474 encode_uint (enc, MAJOR_ARRAY, len + 1);
266 475
476 if (ecb_expect_false (SvMAGICAL (av)))
267 for (i = 0; i <= len; ++i) 477 for (i = 0; i <= len; ++i)
268 { 478 {
269 SV **svp = av_fetch (av, i, 0); 479 SV **svp = av_fetch (av, i, 0);
270 encode_sv (enc, svp ? *svp : &PL_sv_undef); 480 encode_sv (enc, svp ? *svp : &PL_sv_undef);
271 } 481 }
482 else
483 for (i = 0; i <= len; ++i)
484 {
485 SV *sv = AvARRAY (av)[i];
486 encode_sv (enc, sv ? sv : &PL_sv_undef);
487 }
272 488
273 --enc->depth; 489 --enc->depth;
274} 490}
275 491
276static void 492static void
284 ++enc->depth; 500 ++enc->depth;
285 501
286 int pairs = hv_iterinit (hv); 502 int pairs = hv_iterinit (hv);
287 int mg = SvMAGICAL (hv); 503 int mg = SvMAGICAL (hv);
288 504
289 if (mg) 505 if (ecb_expect_false (mg))
290 encode_ch (enc, 0xa0 | 31); 506 encode_ch (enc, MAJOR_MAP | MINOR_INDEF);
291 else 507 else
292 encode_uint (enc, 0xa0, pairs); 508 encode_uint (enc, MAJOR_MAP, pairs);
293 509
294 while ((he = hv_iternext (hv))) 510 while ((he = hv_iternext (hv)))
295 { 511 {
296 if (HeKLEN (he) == HEf_SVKEY) 512 if (HeKLEN (he) == HEf_SVKEY)
297 encode_sv (enc, HeSVKEY (he)); 513 encode_sv (enc, HeSVKEY (he));
298 else 514 else
299 encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); 515 encode_strref (enc, enc->cbor.flags & (F_TEXT_KEYS | F_TEXT_STRINGS), HeKUTF8 (he), HeKEY (he), HeKLEN (he));
300 516
301 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); 517 encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he));
302 } 518 }
303 519
304 if (mg) 520 if (ecb_expect_false (mg))
305 encode_ch (enc, 0xe0 | 31); 521 encode_ch (enc, MAJOR_MISC | MINOR_INDEF);
306 522
307 --enc->depth; 523 --enc->depth;
308} 524}
309 525
310// encode objects, arrays and special \0=false and \1=true values. 526// encode objects, arrays and special \0=false and \1=true values.
311static void 527static void
312encode_rv (enc_t *enc, SV *sv) 528encode_rv (enc_t *enc, SV *sv)
313{ 529{
314 svtype svt;
315
316 SvGETMAGIC (sv); 530 SvGETMAGIC (sv);
531
317 svt = SvTYPE (sv); 532 svtype svt = SvTYPE (sv);
318
319 if (encode_sharable (enc, sv))
320 return;
321 533
322 if (ecb_expect_false (SvOBJECT (sv))) 534 if (ecb_expect_false (SvOBJECT (sv)))
323 { 535 {
324 HV *boolean_stash = !CBOR_SLOW || types_boolean_stash 536 HV *boolean_stash = !CBOR_SLOW || types_boolean_stash
325 ? types_boolean_stash 537 ? types_boolean_stash
330 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 542 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
331 ? cbor_tagged_stash 543 ? cbor_tagged_stash
332 : gv_stashpv ("CBOR::XS::Tagged" , 1); 544 : gv_stashpv ("CBOR::XS::Tagged" , 1);
333 545
334 HV *stash = SvSTASH (sv); 546 HV *stash = SvSTASH (sv);
335 GV *method;
336 547
337 if (stash == boolean_stash) 548 if (stash == boolean_stash)
338 encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20); 549 {
550 encode_ch (enc, SvIV (sv) ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE);
551 return;
552 }
339 else if (stash == error_stash) 553 else if (stash == error_stash)
340 encode_ch (enc, 0xe0 | 23); 554 {
555 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
556 return;
557 }
341 else if (stash == tagged_stash) 558 else if (stash == tagged_stash)
342 { 559 {
343 if (svt != SVt_PVAV) 560 if (svt != SVt_PVAV)
344 croak ("encountered CBOR::XS::Tagged object that isn't an array"); 561 croak ("encountered CBOR::XS::Tagged object that isn't an array");
345 562
563 switch (av_len ((AV *)sv))
564 {
565 case 2-1:
566 // actually a tagged value
346 encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); 567 encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1)));
347 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); 568 encode_sv (enc, *av_fetch ((AV *)sv, 1, 1));
569 break;
570
571 case 3-1:
572 // a forced type [value, type, undef]
573 encode_forced (enc, SvUV (*av_fetch ((AV *)sv, 1, 1)), *av_fetch ((AV *)sv, 0, 1));
574 break;
575
576 default:
577 croak ("encountered malformed CBOR::XS::Tagged object");
578 }
579
580 return;
581 }
582 }
583
584 if (ecb_expect_false (SvREFCNT (sv) > 1)
585 && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING))
586 {
587 if (ecb_expect_false (!enc->shareable))
588 enc->shareable = (HV *)sv_2mortal ((SV *)newHV ());
589
590 SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1);
591
592 if (SvOK (*svp))
348 } 593 {
594 encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF);
595 encode_uint (enc, MAJOR_POS_INT, SvUV (*svp));
596 return;
597 }
598 else
599 {
600 sv_setuv (*svp, enc->shareable_idx);
601 ++enc->shareable_idx;
602 encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE);
603 }
604 }
605
606 if (ecb_expect_false (SvOBJECT (sv)))
607 {
608 HV *stash = SvSTASH (sv);
609 GV *method;
610
611 if (enc->cbor.flags & F_FORBID_OBJECTS)
612 croak ("encountered object '%s', but forbid_objects is enabled",
613 SvPV_nolen (sv_2mortal (newRV_inc (sv))));
349 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) 614 else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0)))
350 { 615 {
351 dSP; 616 dSP;
352 617
353 ENTER; SAVETMPS; PUSHMARK (SP); 618 ENTER; SAVETMPS;
619 PUSHMARK (SP);
354 // we re-bless the reference to get overload and other niceties right 620 // we re-bless the reference to get overload and other niceties right
355 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 621 XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
356 622
357 PUTBACK; 623 PUTBACK;
358 // G_SCALAR ensures that return value is 1 624 // G_SCALAR ensures that return value is 1
371 } 637 }
372 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0) 638 else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0)
373 { 639 {
374 dSP; 640 dSP;
375 641
376 ENTER; SAVETMPS; PUSHMARK (SP); 642 ENTER; SAVETMPS;
643 PUSHMARK (SP);
377 EXTEND (SP, 2); 644 EXTEND (SP, 2);
378 // we re-bless the reference to get overload and other niceties right 645 // we re-bless the reference to get overload and other niceties right
379 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); 646 PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash));
380 PUSHs (sv_cbor); 647 PUSHs (sv_cbor);
381 648
386 // catch this surprisingly common error 653 // catch this surprisingly common error
387 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) 654 if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv)
388 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash)); 655 croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash));
389 656
390 encode_tag (enc, CBOR_TAG_PERL_OBJECT); 657 encode_tag (enc, CBOR_TAG_PERL_OBJECT);
391 encode_uint (enc, 0x80, count + 1); 658 encode_uint (enc, MAJOR_ARRAY, count + 1);
392 encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); 659 encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash));
393 660
394 while (count) 661 {
662 int i;
663
664 for (i = 0; i < count; ++i)
395 encode_sv (enc, SP[1 - count--]); 665 encode_sv (enc, SP[i + 1 - count]);
666
667 SP -= count;
668 }
396 669
397 PUTBACK; 670 PUTBACK;
398 671
399 FREETMPS; LEAVE; 672 FREETMPS; LEAVE;
400 } 673 }
418{ 691{
419 double nv = SvNVX (sv); 692 double nv = SvNVX (sv);
420 693
421 need (enc, 9); 694 need (enc, 9);
422 695
423 if (ecb_expect_false (nv == (U32)nv)) 696 if (ecb_expect_false (nv == (NV)(U32)nv))
424 encode_uint (enc, 0x00, (U32)nv); 697 encode_uint (enc, MAJOR_POS_INT, (U32)nv);
425 //TODO: maybe I32? 698 //TODO: maybe I32?
426 else if (ecb_expect_false (nv == (float)nv)) 699 else if (ecb_expect_false (nv == (float)nv))
427 { 700 encode_float32 (enc, nv);
428 uint32_t fp = ecb_float_to_binary32 (nv);
429
430 *enc->cur++ = 0xe0 | 26;
431
432 if (!ecb_big_endian ())
433 fp = ecb_bswap32 (fp);
434
435 memcpy (enc->cur, &fp, 4);
436 enc->cur += 4;
437 }
438 else 701 else
439 { 702 encode_float64 (enc, nv);
440 uint64_t fp = ecb_double_to_binary64 (nv);
441
442 *enc->cur++ = 0xe0 | 27;
443
444 if (!ecb_big_endian ())
445 fp = ecb_bswap64 (fp);
446
447 memcpy (enc->cur, &fp, 8);
448 enc->cur += 8;
449 }
450} 703}
451 704
452static void 705static void
453encode_sv (enc_t *enc, SV *sv) 706encode_sv (enc_t *enc, SV *sv)
454{ 707{
455 SvGETMAGIC (sv); 708 SvGETMAGIC (sv);
456 709
457 if (encode_sharable (enc, sv))
458 return;
459
460 if (SvPOKp (sv)) 710 if (SvPOKp (sv))
461 { 711 {
462 STRLEN len; 712 STRLEN len;
463 char *str = SvPV (sv, len); 713 char *str = SvPV (sv, len);
464 encode_str (enc, SvUTF8 (sv), str, len); 714 encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len);
465 } 715 }
466 else if (SvNOKp (sv)) 716 else if (SvNOKp (sv))
467 encode_nv (enc, sv); 717 encode_nv (enc, sv);
468 else if (SvIOKp (sv)) 718 else if (SvIOKp (sv))
469 { 719 encode_int (e,v sv);
470 if (SvIsUV (sv))
471 encode_uint (enc, 0x00, SvUVX (sv));
472 else if (SvIVX (sv) >= 0)
473 encode_uint (enc, 0x00, SvIVX (sv));
474 else
475 encode_uint (enc, 0x20, -(SvIVX (sv) + 1));
476 }
477 else if (SvROK (sv)) 720 else if (SvROK (sv))
478 encode_rv (enc, SvRV (sv)); 721 encode_rv (enc, SvRV (sv));
479 else if (!SvOK (sv)) 722 else if (!SvOK (sv))
480 encode_ch (enc, 0xe0 | 22); 723 encode_ch (enc, MAJOR_MISC | SIMPLE_NULL);
481 else if (enc->cbor.flags & F_ALLOW_UNKNOWN) 724 else if (enc->cbor.flags & F_ALLOW_UNKNOWN)
482 encode_ch (enc, 0xe0 | 23); 725 encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF);
483 else 726 else
484 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data", 727 croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data",
485 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); 728 SvPV_nolen (sv), (unsigned int)SvFLAGS (sv));
486} 729}
487 730
488static SV * 731static SV *
489encode_cbor (SV *scalar, CBOR *cbor) 732encode_cbor (SV *scalar, CBOR *cbor)
490{ 733{
491 enc_t enc = { }; 734 enc_t enc = { 0 };
492 735
493 enc.cbor = *cbor; 736 enc.cbor = *cbor;
494 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); 737 enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE));
495 enc.cur = SvPVX (enc.sv); 738 enc.cur = SvPVX (enc.sv);
496 enc.end = SvEND (enc.sv); 739 enc.end = SvEND (enc.sv);
497 740
498 SvPOK_only (enc.sv); 741 SvPOK_only (enc.sv);
742
743 if (cbor->flags & F_PACK_STRINGS)
744 {
745 encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE);
746 enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ());
747 enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ());
748 }
749
499 encode_sv (&enc, scalar); 750 encode_sv (&enc, scalar);
500 751
501 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv)); 752 SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv));
502 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings 753 *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings
503 754
517 U8 *end; // end of input string 768 U8 *end; // end of input string
518 const char *err; // parse error, if != 0 769 const char *err; // parse error, if != 0
519 CBOR cbor; 770 CBOR cbor;
520 U32 depth; // recursion depth 771 U32 depth; // recursion depth
521 U32 maxdepth; // recursion depth limit 772 U32 maxdepth; // recursion depth limit
522 AV *sharable; 773 AV *shareable;
774 AV *stringref;
775 SV *decode_tagged;
776 SV *err_sv; // optional sv for error, needs to be freed
523} dec_t; 777} dec_t;
524 778
525#define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE 779// set dec->err to ERRSV
780ecb_cold static void
781err_errsv (dec_t *dec)
782{
783 if (!dec->err)
784 {
785 dec->err_sv = newSVsv (ERRSV);
526 786
527#define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data") 787 // chop off the trailing \n
788 SvCUR_set (dec->err_sv, SvCUR (dec->err_sv) - 1);
789 *SvEND (dec->err_sv) = 0;
528 790
791 dec->err = SvPVutf8_nolen (dec->err_sv);
792 }
793}
794
795// the following functions are used to reduce code size and help the compiler to optimise
796ecb_cold static void
797err_set (dec_t *dec, const char *reason)
798{
799 if (!dec->err)
800 dec->err = reason;
801}
802
803ecb_cold static void
804err_unexpected_end (dec_t *dec)
805{
806 err_set (dec, "unexpected end of CBOR data");
807}
808
809#define ERR_DO(do) SB do; goto fail; SE
810#define ERR(reason) ERR_DO (err_set (dec, reason))
811#define ERR_ERRSV ERR_DO (err_errsv (dec))
812
813#define WANT(len) if (ecb_expect_false ((uintptr_t)(dec->end - dec->cur) < (STRLEN)len)) ERR_DO (err_unexpected_end (dec))
814
529#define DEC_INC_DEPTH if (++dec->depth > dec->cbor.max_depth) ERR (ERR_NESTING_EXCEEDED) 815#define DEC_INC_DEPTH if (ecb_expect_false (++dec->depth > dec->cbor.max_depth)) ERR (ERR_NESTING_EXCEEDED)
530#define DEC_DEC_DEPTH --dec->depth 816#define DEC_DEC_DEPTH --dec->depth
531 817
532static UV 818static UV
533decode_uint (dec_t *dec) 819decode_uint (dec_t *dec)
534{ 820{
535 switch (*dec->cur & 31) 821 U8 m = *dec->cur & MINOR_MASK;
536 { 822 ++dec->cur;
537 case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7:
538 case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15:
539 case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23:
540 return *dec->cur++ & 31;
541 823
542 case 24: 824 if (ecb_expect_true (m < LENGTH_EXT1))
825 return m;
826 else if (ecb_expect_true (m == LENGTH_EXT1))
827 {
543 WANT (2); 828 WANT (1);
544 dec->cur += 2; 829 dec->cur += 1;
545 return dec->cur[-1]; 830 return dec->cur[-1];
546 831 }
547 case 25: 832 else if (ecb_expect_true (m == LENGTH_EXT2))
833 {
548 WANT (3); 834 WANT (2);
549 dec->cur += 3; 835 dec->cur += 2;
550 return (((UV)dec->cur[-2]) << 8) 836 return (((UV)dec->cur[-2]) << 8)
551 | ((UV)dec->cur[-1]); 837 | ((UV)dec->cur[-1]);
552 838 }
553 case 26: 839 else if (ecb_expect_true (m == LENGTH_EXT4))
840 {
554 WANT (5); 841 WANT (4);
555 dec->cur += 5; 842 dec->cur += 4;
556 return (((UV)dec->cur[-4]) << 24) 843 return (((UV)dec->cur[-4]) << 24)
557 | (((UV)dec->cur[-3]) << 16) 844 | (((UV)dec->cur[-3]) << 16)
558 | (((UV)dec->cur[-2]) << 8) 845 | (((UV)dec->cur[-2]) << 8)
559 | ((UV)dec->cur[-1]); 846 | ((UV)dec->cur[-1]);
560 847 }
561 case 27: 848 else if (ecb_expect_true (m == LENGTH_EXT8))
849 {
562 WANT (9); 850 WANT (8);
563 dec->cur += 9; 851 dec->cur += 8;
852
853 return
854#if UVSIZE < 8
855 0
856#else
564 return (((UV)dec->cur[-8]) << 56) 857 (((UV)dec->cur[-8]) << 56)
565 | (((UV)dec->cur[-7]) << 48) 858 | (((UV)dec->cur[-7]) << 48)
566 | (((UV)dec->cur[-6]) << 40) 859 | (((UV)dec->cur[-6]) << 40)
567 | (((UV)dec->cur[-5]) << 32) 860 | (((UV)dec->cur[-5]) << 32)
861#endif
568 | (((UV)dec->cur[-4]) << 24) 862 | (((UV)dec->cur[-4]) << 24)
569 | (((UV)dec->cur[-3]) << 16) 863 | (((UV)dec->cur[-3]) << 16)
570 | (((UV)dec->cur[-2]) << 8) 864 | (((UV)dec->cur[-2]) << 8)
571 | ((UV)dec->cur[-1]); 865 | ((UV)dec->cur[-1]);
572 866 }
573 default: 867 else
574 ERR ("corrupted CBOR data (unsupported integer minor encoding)"); 868 ERR ("corrupted CBOR data (unsupported integer minor encoding)");
575 }
576 869
577fail: 870fail:
578 return 0; 871 return 0;
579} 872}
580 873
585{ 878{
586 AV *av = newAV (); 879 AV *av = newAV ();
587 880
588 DEC_INC_DEPTH; 881 DEC_INC_DEPTH;
589 882
590 if ((*dec->cur & 31) == 31) 883 if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF))
591 { 884 {
592 ++dec->cur; 885 ++dec->cur;
593 886
594 for (;;) 887 for (;;)
595 { 888 {
596 WANT (1); 889 WANT (1);
597 890
598 if (*dec->cur == (0xe0 | 31)) 891 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
599 { 892 {
600 ++dec->cur; 893 ++dec->cur;
601 break; 894 break;
602 } 895 }
603 896
604 av_push (av, decode_sv (dec)); 897 av_push (av, decode_sv (dec));
605 } 898 }
606 } 899 }
607 else 900 else
608 { 901 {
609 int i, len = decode_uint (dec); 902 UV i, len = decode_uint (dec);
610 903
904 WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays
611 av_fill (av, len - 1); 905 av_fill (av, len - 1);
612 906
613 for (i = 0; i < len; ++i) 907 for (i = 0; i < len; ++i)
614 AvARRAY (av)[i] = decode_sv (dec); 908 AvARRAY (av)[i] = decode_sv (dec);
615 } 909 }
616 910
617 DEC_DEC_DEPTH; 911 DEC_DEC_DEPTH;
618 return newRV_noinc ((SV *)av); 912 return newRV_noinc ((SV *)av);
619 913
620fail: 914fail:
621 SvREFCNT_dec (av); 915 SvREFCNT_dec_NN (av);
622 DEC_DEC_DEPTH; 916 DEC_DEC_DEPTH;
623 return &PL_sv_undef; 917 return &PL_sv_undef;
624} 918}
625 919
626static void 920static void
627decode_he (dec_t *dec, HV *hv) 921decode_he (dec_t *dec, HV *hv)
628{ 922{
629 // for speed reasons, we specialcase single-string 923 // for speed reasons, we specialcase single-string
630 // byte or utf-8 strings as keys. 924 // byte or utf-8 strings as keys, but only when !stringref
631 925
632 if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27) 926 if (ecb_expect_true (!dec->stringref))
927 if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8))
633 { 928 {
634 I32 len = decode_uint (dec); 929 STRLEN len = decode_uint (dec);
635 char *key = (char *)dec->cur; 930 char *key = (char *)dec->cur;
636 931
932 WANT (len);
637 dec->cur += len; 933 dec->cur += len;
638 934
639 hv_store (hv, key, len, decode_sv (dec), 0); 935 hv_store (hv, key, len, decode_sv (dec), 0);
936
937 return;
640 } 938 }
641 else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27) 939 else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8))
642 { 940 {
643 I32 len = decode_uint (dec); 941 STRLEN len = decode_uint (dec);
644 char *key = (char *)dec->cur; 942 char *key = (char *)dec->cur;
645 943
944 WANT (len);
646 dec->cur += len; 945 dec->cur += len;
647 946
947 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
948 if (!is_utf8_string (key, len))
949 ERR ("corrupted CBOR data (invalid UTF-8 in map key)");
950
648 hv_store (hv, key, -len, decode_sv (dec), 0); 951 hv_store (hv, key, -len, decode_sv (dec), 0);
952
953 return;
649 } 954 }
650 else 955
651 {
652 SV *k = decode_sv (dec); 956 SV *k = decode_sv (dec);
653 SV *v = decode_sv (dec); 957 SV *v = decode_sv (dec);
654 958
959 // we leak memory if uncaught exceptions are thrown by random magical
960 // methods, and this is hopefully the only place where it can happen,
961 // so if there is a chance of an exception, take the very slow path.
962 // since catching exceptions is "undocumented/internal/forbidden" by
963 // the new p5p powers, we need to call out to a perl function :/
964 if (ecb_expect_false (SvAMAGIC (k)))
965 {
966 dSP;
967
968 ENTER; SAVETMPS;
969 PUSHMARK (SP);
970 EXTEND (SP, 3);
971 PUSHs (sv_2mortal (newRV_inc ((SV *)hv)));
972 PUSHs (sv_2mortal (k));
973 PUSHs (sv_2mortal (v));
974
975 PUTBACK;
976 call_pv ("CBOR::XS::_hv_store", G_VOID | G_DISCARD | G_EVAL);
977 SPAGAIN;
978
979 FREETMPS; LEAVE;
980
981 if (SvTRUE (ERRSV))
982 ERR_ERRSV;
983
984 return;
985 }
986
655 hv_store_ent (hv, k, v, 0); 987 hv_store_ent (hv, k, v, 0);
656 SvREFCNT_dec (k); 988 SvREFCNT_dec_NN (k);
657 } 989
990fail:
991 ;
658} 992}
659 993
660static SV * 994static SV *
661decode_hv (dec_t *dec) 995decode_hv (dec_t *dec)
662{ 996{
663 HV *hv = newHV (); 997 HV *hv = newHV ();
664 998
665 DEC_INC_DEPTH; 999 DEC_INC_DEPTH;
666 1000
667 if ((*dec->cur & 31) == 31) 1001 if (*dec->cur == (MAJOR_MAP | MINOR_INDEF))
668 { 1002 {
669 ++dec->cur; 1003 ++dec->cur;
670 1004
671 for (;;) 1005 for (;;)
672 { 1006 {
673 WANT (1); 1007 WANT (1);
674 1008
675 if (*dec->cur == (0xe0 | 31)) 1009 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err)
676 { 1010 {
677 ++dec->cur; 1011 ++dec->cur;
678 break; 1012 break;
679 } 1013 }
680 1014
681 decode_he (dec, hv); 1015 decode_he (dec, hv);
682 } 1016 }
683 } 1017 }
684 else 1018 else
685 { 1019 {
686 int pairs = decode_uint (dec); 1020 UV pairs = decode_uint (dec);
1021
1022 WANT (pairs); // complexity check - need at least one byte per value, do not allow supersize hashes
687 1023
688 while (pairs--) 1024 while (pairs--)
689 decode_he (dec, hv); 1025 decode_he (dec, hv);
690 } 1026 }
691 1027
692 DEC_DEC_DEPTH; 1028 DEC_DEC_DEPTH;
693 return newRV_noinc ((SV *)hv); 1029 return newRV_noinc ((SV *)hv);
694 1030
695fail: 1031fail:
696 SvREFCNT_dec (hv); 1032 SvREFCNT_dec_NN (hv);
697 DEC_DEC_DEPTH; 1033 DEC_DEC_DEPTH;
698 return &PL_sv_undef; 1034 return &PL_sv_undef;
699} 1035}
700 1036
701static SV * 1037static SV *
702decode_str (dec_t *dec, int utf8) 1038decode_str (dec_t *dec, int utf8)
703{ 1039{
704 SV *sv = 0; 1040 SV *sv = 0;
705 1041
706 if ((*dec->cur & 31) == 31) 1042 if (ecb_expect_false ((*dec->cur & MINOR_MASK) == MINOR_INDEF))
707 { 1043 {
1044 // indefinite length strings
708 ++dec->cur; 1045 ++dec->cur;
709 1046
1047 U8 major = *dec->cur & MAJOR_MISC;
1048
710 sv = newSVpvn ("", 0); 1049 sv = newSVpvn ("", 0);
711 1050
712 // not very fast, and certainly not robust against illegal input
713 for (;;) 1051 for (;;)
714 { 1052 {
715 WANT (1); 1053 WANT (1);
716 1054
717 if (*dec->cur == (0xe0 | 31)) 1055 if ((*dec->cur - major) > LENGTH_EXT8)
1056 if (*dec->cur == (MAJOR_MISC | MINOR_INDEF))
718 { 1057 {
719 ++dec->cur; 1058 ++dec->cur;
720 break; 1059 break;
721 } 1060 }
1061 else
1062 ERR ("corrupted CBOR data (invalid chunks in indefinite length string)");
722 1063
723 sv_catsv (sv, decode_sv (dec)); 1064 STRLEN len = decode_uint (dec);
1065
1066 WANT (len);
1067 sv_catpvn (sv, dec->cur, len);
1068 dec->cur += len;
724 } 1069 }
725 } 1070 }
726 else 1071 else
727 { 1072 {
728 STRLEN len = decode_uint (dec); 1073 STRLEN len = decode_uint (dec);
729 1074
730 WANT (len); 1075 WANT (len);
731 sv = newSVpvn (dec->cur, len); 1076 sv = newSVpvn (dec->cur, len);
732 dec->cur += len; 1077 dec->cur += len;
1078
1079 if (ecb_expect_false (dec->stringref)
1080 && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1))
1081 av_push (dec->stringref, SvREFCNT_inc_NN (sv));
733 } 1082 }
734 1083
735 if (utf8) 1084 if (utf8)
1085 {
1086 if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8))
1087 if (!is_utf8_string (SvPVX (sv), SvCUR (sv)))
1088 ERR ("corrupted CBOR data (invalid UTF-8 in text string)");
1089
736 SvUTF8_on (sv); 1090 SvUTF8_on (sv);
1091 }
737 1092
738 return sv; 1093 return sv;
739 1094
740fail: 1095fail:
741 SvREFCNT_dec (sv); 1096 SvREFCNT_dec (sv);
743} 1098}
744 1099
745static SV * 1100static SV *
746decode_tagged (dec_t *dec) 1101decode_tagged (dec_t *dec)
747{ 1102{
1103 SV *sv = 0;
748 UV tag = decode_uint (dec); 1104 UV tag = decode_uint (dec);
749 SV *sv = decode_sv (dec); 1105
1106 WANT (1);
750 1107
751 switch (tag) 1108 switch (tag)
752 { 1109 {
753 case CBOR_TAG_MAGIC: 1110 case CBOR_TAG_MAGIC:
754 return sv; 1111 sv = decode_sv (dec);
1112 break;
755 1113
756 case CBOR_TAG_INDIRECTION: 1114 case CBOR_TAG_INDIRECTION:
757 return newRV_noinc (sv); 1115 sv = newRV_noinc (decode_sv (dec));
1116 break;
758 1117
1118 case CBOR_TAG_STRINGREF_NAMESPACE:
1119 {
1120 // do not use SAVETMPS/FREETMPS, as these will
1121 // erase mortalised caches, e.g. "shareable"
1122 ENTER;
1123
1124 SAVESPTR (dec->stringref);
1125 dec->stringref = (AV *)sv_2mortal ((SV *)newAV ());
1126
1127 sv = decode_sv (dec);
1128
1129 LEAVE;
1130 }
1131 break;
1132
1133 case CBOR_TAG_STRINGREF:
1134 {
1135 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
1136 ERR ("corrupted CBOR data (stringref index not an unsigned integer)");
1137
1138 UV idx = decode_uint (dec);
1139
1140 if (!dec->stringref || idx >= (UV)(1 + AvFILLp (dec->stringref)))
1141 ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)");
1142
1143 sv = newSVsv (AvARRAY (dec->stringref)[idx]);
1144 }
1145 break;
1146
759 case CBOR_TAG_VALUE_SHARABLE: 1147 case CBOR_TAG_VALUE_SHAREABLE:
1148 {
760 if (ecb_expect_false (!dec->sharable)) 1149 if (ecb_expect_false (!dec->shareable))
761 dec->sharable = (AV *)sv_2mortal ((SV *)newAV ()); 1150 dec->shareable = (AV *)sv_2mortal ((SV *)newAV ());
762 1151
1152 if (dec->cbor.flags & F_ALLOW_CYCLES)
1153 {
1154 sv = newSV (0);
763 av_push (dec->sharable, SvREFCNT_inc_NN (sv)); 1155 av_push (dec->shareable, SvREFCNT_inc_NN (sv));
764 1156
765 return sv; 1157 SV *osv = decode_sv (dec);
1158 sv_setsv (sv, osv);
1159 SvREFCNT_dec_NN (osv);
1160 }
1161 else
1162 {
1163 av_push (dec->shareable, &PL_sv_undef);
1164 int idx = AvFILLp (dec->shareable);
1165 sv = decode_sv (dec);
1166 av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv));
1167 }
1168 }
1169 break;
766 1170
767 case CBOR_TAG_VALUE_SHAREDREF: 1171 case CBOR_TAG_VALUE_SHAREDREF:
768 { 1172 {
769 // TODO: should verify that the sv atcually was a CBOR unsigned integer 1173 if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT))
770 UV idx = SvUV (sv); 1174 ERR ("corrupted CBOR data (sharedref index not an unsigned integer)");
771 1175
1176 UV idx = decode_uint (dec);
1177
772 if (!dec->sharable || idx > AvFILLp (dec->sharable)) 1178 if (!dec->shareable || idx >= (UV)(1 + AvFILLp (dec->shareable)))
773 ERR ("corrupted CBOR data (sharedref index out of bounds)"); 1179 ERR ("corrupted CBOR data (sharedref index out of bounds)");
774 1180
775 SvREFCNT_dec (sv);
776
777 return SvREFCNT_inc_NN (AvARRAY (dec->sharable)[idx]); 1181 sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]);
1182
1183 if (sv == &PL_sv_undef)
1184 ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled");
778 } 1185 }
1186 break;
779 1187
780 case CBOR_TAG_PERL_OBJECT: 1188 case CBOR_TAG_PERL_OBJECT:
781 { 1189 {
1190 if (dec->cbor.flags & F_FORBID_OBJECTS)
1191 goto filter;
1192
1193 sv = decode_sv (dec);
1194
782 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) 1195 if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
783 ERR ("corrupted CBOR data (non-array perl object)"); 1196 ERR ("corrupted CBOR data (non-array perl object)");
784 1197
785 AV *av = (AV *)SvRV (sv); 1198 AV *av = (AV *)SvRV (sv);
786 int len = av_len (av) + 1; 1199 int len = av_len (av) + 1;
794 if (!method) 1207 if (!method)
795 ERR ("cannot decode perl-object (package does not have a THAW method)"); 1208 ERR ("cannot decode perl-object (package does not have a THAW method)");
796 1209
797 dSP; 1210 dSP;
798 1211
799 ENTER; SAVETMPS; PUSHMARK (SP); 1212 ENTER; SAVETMPS;
1213 PUSHMARK (SP);
800 EXTEND (SP, len + 1); 1214 EXTEND (SP, len + 1);
801 // we re-bless the reference to get overload and other niceties right 1215 // we re-bless the reference to get overload and other niceties right
802 PUSHs (*av_fetch (av, 0, 1)); 1216 PUSHs (*av_fetch (av, 0, 1));
803 PUSHs (sv_cbor); 1217 PUSHs (sv_cbor);
804 1218
812 SPAGAIN; 1226 SPAGAIN;
813 1227
814 if (SvTRUE (ERRSV)) 1228 if (SvTRUE (ERRSV))
815 { 1229 {
816 FREETMPS; LEAVE; 1230 FREETMPS; LEAVE;
817 ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV)))); 1231 ERR_ERRSV;
818 } 1232 }
819 1233
820 SvREFCNT_dec (sv); 1234 SvREFCNT_dec_NN (sv);
821 sv = SvREFCNT_inc (POPs); 1235 sv = SvREFCNT_inc (POPs);
822 1236
823 PUTBACK; 1237 PUTBACK;
824 1238
825 FREETMPS; LEAVE; 1239 FREETMPS; LEAVE;
826
827 return sv;
828 } 1240 }
1241 break;
829 1242
830 default: 1243 default:
1244 filter:
1245 {
1246 SV *tag_sv = newSVuv (tag);
1247
1248 sv = decode_sv (dec);
1249
1250 dSP;
1251 ENTER; SAVETMPS;
1252 PUSHMARK (SP);
1253 EXTEND (SP, 2);
1254 PUSHs (tag_sv);
1255 PUSHs (sv);
1256
1257 PUTBACK;
1258 int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL);
1259 SPAGAIN;
1260
1261 if (SvTRUE (ERRSV))
831 { 1262 {
1263 SvREFCNT_dec_NN (tag_sv);
1264 FREETMPS; LEAVE;
1265 ERR_ERRSV;
1266 }
1267
1268 if (count)
1269 {
1270 SvREFCNT_dec_NN (tag_sv);
1271 SvREFCNT_dec_NN (sv);
1272 sv = SvREFCNT_inc_NN (TOPs);
1273 SP -= count;
1274 }
1275 else
1276 {
832 AV *av = newAV (); 1277 AV *av = newAV ();
833 av_push (av, newSVuv (tag)); 1278 av_push (av, tag_sv);
834 av_push (av, sv); 1279 av_push (av, sv);
835 1280
836 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash 1281 HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash
837 ? cbor_tagged_stash 1282 ? cbor_tagged_stash
838 : gv_stashpv ("CBOR::XS::Tagged" , 1); 1283 : gv_stashpv ("CBOR::XS::Tagged" , 1);
839
840 return sv_bless (newRV_noinc ((SV *)av), tagged_stash); 1284 sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash);
841 } 1285 }
1286
1287 PUTBACK;
1288
1289 FREETMPS; LEAVE;
1290 }
1291 break;
842 } 1292 }
1293
1294 return sv;
843 1295
844fail: 1296fail:
845 SvREFCNT_dec (sv); 1297 SvREFCNT_dec (sv);
846 return &PL_sv_undef; 1298 return &PL_sv_undef;
847} 1299}
849static SV * 1301static SV *
850decode_sv (dec_t *dec) 1302decode_sv (dec_t *dec)
851{ 1303{
852 WANT (1); 1304 WANT (1);
853 1305
854 switch (*dec->cur >> 5) 1306 switch (*dec->cur >> MAJOR_SHIFT)
855 { 1307 {
856 case 0: // unsigned int 1308 case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec));
857 return newSVuv (decode_uint (dec)); 1309 case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec));
858 case 1: // negative int 1310 case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0);
859 return newSViv (-1 - (IV)decode_uint (dec)); 1311 case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1);
860 case 2: // octet string 1312 case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec);
861 return decode_str (dec, 0); 1313 case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec);
862 case 3: // utf-8 string 1314 case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec);
863 return decode_str (dec, 1); 1315
864 case 4: // array 1316 case MAJOR_MISC >> MAJOR_SHIFT:
865 return decode_av (dec);
866 case 5: // map
867 return decode_hv (dec);
868 case 6: // tag
869 return decode_tagged (dec);
870 case 7: // misc
871 switch (*dec->cur++ & 31) 1317 switch (*dec->cur++ & MINOR_MASK)
872 { 1318 {
873 case 20: 1319 case SIMPLE_FALSE:
874#if CBOR_SLOW 1320#if CBOR_SLOW
875 types_false = get_bool ("Types::Serialiser::false"); 1321 types_false = get_bool ("Types::Serialiser::false");
876#endif 1322#endif
877 return newSVsv (types_false); 1323 return newSVsv (types_false);
878 case 21: 1324 case SIMPLE_TRUE:
879#if CBOR_SLOW 1325#if CBOR_SLOW
880 types_true = get_bool ("Types::Serialiser::true"); 1326 types_true = get_bool ("Types::Serialiser::true");
881#endif 1327#endif
882 return newSVsv (types_true); 1328 return newSVsv (types_true);
883 case 22: 1329 case SIMPLE_NULL:
884 return newSVsv (&PL_sv_undef); 1330 return newSVsv (&PL_sv_undef);
885 case 23: 1331 case SIMPLE_UNDEF:
886#if CBOR_SLOW 1332#if CBOR_SLOW
887 types_error = get_bool ("Types::Serialiser::error"); 1333 types_error = get_bool ("Types::Serialiser::error");
888#endif 1334#endif
889 return newSVsv (types_error); 1335 return newSVsv (types_error);
890 1336
891 case 25: 1337 case MISC_FLOAT16:
892 { 1338 {
893 WANT (2); 1339 WANT (2);
894 1340
895 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; 1341 uint16_t fp = (dec->cur[0] << 8) | dec->cur[1];
896 dec->cur += 2; 1342 dec->cur += 2;
897 1343
898 return newSVnv (ecb_binary16_to_float (fp)); 1344 return newSVnv (ecb_binary16_to_float (fp));
899 } 1345 }
900 1346
901 case 26: 1347 case MISC_FLOAT32:
902 { 1348 {
903 uint32_t fp; 1349 uint32_t fp;
904 WANT (4); 1350 WANT (4);
905 memcpy (&fp, dec->cur, 4); 1351 memcpy (&fp, dec->cur, 4);
906 dec->cur += 4; 1352 dec->cur += 4;
909 fp = ecb_bswap32 (fp); 1355 fp = ecb_bswap32 (fp);
910 1356
911 return newSVnv (ecb_binary32_to_float (fp)); 1357 return newSVnv (ecb_binary32_to_float (fp));
912 } 1358 }
913 1359
914 case 27: 1360 case MISC_FLOAT64:
915 { 1361 {
916 uint64_t fp; 1362 uint64_t fp;
917 WANT (8); 1363 WANT (8);
918 memcpy (&fp, dec->cur, 8); 1364 memcpy (&fp, dec->cur, 8);
919 dec->cur += 8; 1365 dec->cur += 8;
922 fp = ecb_bswap64 (fp); 1368 fp = ecb_bswap64 (fp);
923 1369
924 return newSVnv (ecb_binary64_to_double (fp)); 1370 return newSVnv (ecb_binary64_to_double (fp));
925 } 1371 }
926 1372
927 // 0..19 unassigned 1373 // 0..19 unassigned simple
928 // 24 reserved + unassigned (reserved values are not encodable) 1374 // 24 reserved + unassigned simple (reserved values are not encodable)
1375 // 28-30 unassigned misc
1376 // 31 break code
929 default: 1377 default:
930 ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); 1378 ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)");
931 } 1379 }
932 1380
933 break; 1381 break;
934 } 1382 }
935 1383
938} 1386}
939 1387
940static SV * 1388static SV *
941decode_cbor (SV *string, CBOR *cbor, char **offset_return) 1389decode_cbor (SV *string, CBOR *cbor, char **offset_return)
942{ 1390{
943 dec_t dec = { }; 1391 dec_t dec = { 0 };
944 SV *sv; 1392 SV *sv;
945 STRLEN len; 1393 STRLEN len;
946 char *data = SvPVbyte (string, len); 1394 char *data = SvPVbyte (string, len);
947 1395
948 if (len > cbor->max_size && cbor->max_size) 1396 if (len > cbor->max_size && cbor->max_size)
962 if (dec.cur != dec.end && !dec.err) 1410 if (dec.cur != dec.end && !dec.err)
963 dec.err = "garbage after CBOR object"; 1411 dec.err = "garbage after CBOR object";
964 1412
965 if (dec.err) 1413 if (dec.err)
966 { 1414 {
1415 if (dec.shareable)
1416 {
1417 // need to break cyclic links, which would all be in shareable
1418 int i;
1419 SV **svp;
1420
1421 for (i = av_len (dec.shareable) + 1; i--; )
1422 if ((svp = av_fetch (dec.shareable, i, 0)))
1423 sv_setsv (*svp, &PL_sv_undef);
1424 }
1425
967 SvREFCNT_dec (sv); 1426 SvREFCNT_dec_NN (sv);
1427
1428 if (dec.err_sv)
1429 sv_2mortal (dec.err_sv);
1430
968 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur); 1431 croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur);
969 } 1432 }
970 1433
971 sv = sv_2mortal (sv); 1434 sv = sv_2mortal (sv);
972 1435
973 return sv; 1436 return sv;
974} 1437}
975 1438
1439/////////////////////////////////////////////////////////////////////////////
1440// incremental parser
1441
1442#define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0)
1443
1444// returns 0 for notyet, 1 for success or error
1445static int
1446incr_parse (CBOR *self, SV *cborstr)
1447{
1448 STRLEN cur;
1449 SvPV (cborstr, cur);
1450
1451 while (ecb_expect_true (self->incr_need <= cur))
1452 {
1453 // table of integer count bytes
1454 static I8 incr_len[MINOR_MASK + 1] = {
1455 0, 0, 0, 0, 0, 0, 0, 0,
1456 0, 0, 0, 0, 0, 0, 0, 0,
1457 0, 0, 0, 0, 0, 0, 0, 0,
1458 1, 2, 4, 8,-1,-1,-1,-2
1459 };
1460
1461 const U8 *p = SvPVX (cborstr) + self->incr_pos;
1462 U8 m = *p & MINOR_MASK;
1463 IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1464 I8 ilen = incr_len[m];
1465
1466 self->incr_need = self->incr_pos + 1;
1467
1468 if (ecb_expect_false (ilen < 0))
1469 {
1470 if (m != MINOR_INDEF)
1471 return 1; // error
1472
1473 if (*p == (MAJOR_MISC | MINOR_INDEF))
1474 {
1475 if (count >= 0)
1476 return 1; // error
1477
1478 count = 1;
1479 }
1480 else
1481 {
1482 av_push (self->incr_count, newSViv (-1)); //TODO: nest
1483 count = -1;
1484 }
1485 }
1486 else
1487 {
1488 self->incr_need += ilen;
1489 if (ecb_expect_false (self->incr_need > cur))
1490 return 0;
1491
1492 int major = *p >> MAJOR_SHIFT;
1493
1494 switch (major)
1495 {
1496 case MAJOR_TAG >> MAJOR_SHIFT:
1497 ++count; // tags merely prefix another value
1498 break;
1499
1500 case MAJOR_BYTES >> MAJOR_SHIFT:
1501 case MAJOR_TEXT >> MAJOR_SHIFT:
1502 case MAJOR_ARRAY >> MAJOR_SHIFT:
1503 case MAJOR_MAP >> MAJOR_SHIFT:
1504 {
1505 UV len;
1506
1507 if (ecb_expect_false (ilen))
1508 {
1509 len = 0;
1510
1511 do {
1512 len = (len << 8) | *++p;
1513 } while (--ilen);
1514 }
1515 else
1516 len = m;
1517
1518 switch (major)
1519 {
1520 case MAJOR_BYTES >> MAJOR_SHIFT:
1521 case MAJOR_TEXT >> MAJOR_SHIFT:
1522 self->incr_need += len;
1523 if (ecb_expect_false (self->incr_need > cur))
1524 return 0;
1525
1526 break;
1527
1528 case MAJOR_MAP >> MAJOR_SHIFT:
1529 len <<= 1;
1530 case MAJOR_ARRAY >> MAJOR_SHIFT:
1531 if (len)
1532 {
1533 av_push (self->incr_count, newSViv (len + 1)); //TODO: nest
1534 count = len + 1;
1535 }
1536 break;
1537 }
1538 }
1539 }
1540 }
1541
1542 self->incr_pos = self->incr_need;
1543
1544 if (count > 0)
1545 {
1546 while (!--count)
1547 {
1548 if (!AvFILLp (self->incr_count))
1549 return 1; // done
1550
1551 SvREFCNT_dec_NN (av_pop (self->incr_count));
1552 count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]);
1553 }
1554
1555 SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count;
1556 }
1557 }
1558
1559 return 0;
1560}
1561
1562
976///////////////////////////////////////////////////////////////////////////// 1563/////////////////////////////////////////////////////////////////////////////
977// XS interface functions 1564// XS interface functions
978 1565
979MODULE = CBOR::XS PACKAGE = CBOR::XS 1566MODULE = CBOR::XS PACKAGE = CBOR::XS
980 1567
988 1575
989 types_true = get_bool ("Types::Serialiser::true" ); 1576 types_true = get_bool ("Types::Serialiser::true" );
990 types_false = get_bool ("Types::Serialiser::false"); 1577 types_false = get_bool ("Types::Serialiser::false");
991 types_error = get_bool ("Types::Serialiser::error"); 1578 types_error = get_bool ("Types::Serialiser::error");
992 1579
1580 default_filter = newSVpv ("CBOR::XS::default_filter", 0);
1581
993 sv_cbor = newSVpv ("CBOR", 0); 1582 sv_cbor = newSVpv ("CBOR", 0);
994 SvREADONLY_on (sv_cbor); 1583 SvREADONLY_on (sv_cbor);
1584
1585 assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1));
995} 1586}
996 1587
997PROTOTYPES: DISABLE 1588PROTOTYPES: DISABLE
998 1589
999void CLONE (...) 1590void CLONE (...)
1018void shrink (CBOR *self, int enable = 1) 1609void shrink (CBOR *self, int enable = 1)
1019 ALIAS: 1610 ALIAS:
1020 shrink = F_SHRINK 1611 shrink = F_SHRINK
1021 allow_unknown = F_ALLOW_UNKNOWN 1612 allow_unknown = F_ALLOW_UNKNOWN
1022 allow_sharing = F_ALLOW_SHARING 1613 allow_sharing = F_ALLOW_SHARING
1023 dedup_keys = F_DEDUP_KEYS 1614 allow_cycles = F_ALLOW_CYCLES
1615 forbid_objects = F_FORBID_OBJECTS
1024 dedup_strings = F_DEDUP_STRINGS 1616 pack_strings = F_PACK_STRINGS
1617 text_keys = F_TEXT_KEYS
1618 text_strings = F_TEXT_STRINGS
1619 validate_utf8 = F_VALIDATE_UTF8
1025 PPCODE: 1620 PPCODE:
1026{ 1621{
1027 if (enable) 1622 if (enable)
1028 self->flags |= ix; 1623 self->flags |= ix;
1029 else 1624 else
1035void get_shrink (CBOR *self) 1630void get_shrink (CBOR *self)
1036 ALIAS: 1631 ALIAS:
1037 get_shrink = F_SHRINK 1632 get_shrink = F_SHRINK
1038 get_allow_unknown = F_ALLOW_UNKNOWN 1633 get_allow_unknown = F_ALLOW_UNKNOWN
1039 get_allow_sharing = F_ALLOW_SHARING 1634 get_allow_sharing = F_ALLOW_SHARING
1040 get_dedup_keys = F_DEDUP_KEYS 1635 get_allow_cycles = F_ALLOW_CYCLES
1636 get_forbid_objects = F_FORBID_OBJECTS
1041 get_dedup_strings = F_DEDUP_STRINGS 1637 get_pack_strings = F_PACK_STRINGS
1638 get_text_keys = F_TEXT_KEYS
1639 get_text_strings = F_TEXT_STRINGS
1640 get_validate_utf8 = F_VALIDATE_UTF8
1042 PPCODE: 1641 PPCODE:
1043 XPUSHs (boolSV (self->flags & ix)); 1642 XPUSHs (boolSV (self->flags & ix));
1044 1643
1045void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) 1644void max_depth (CBOR *self, U32 max_depth = 0x80000000UL)
1046 PPCODE: 1645 PPCODE:
1059 XPUSHs (ST (0)); 1658 XPUSHs (ST (0));
1060 1659
1061int get_max_size (CBOR *self) 1660int get_max_size (CBOR *self)
1062 CODE: 1661 CODE:
1063 RETVAL = self->max_size; 1662 RETVAL = self->max_size;
1663 OUTPUT:
1664 RETVAL
1665
1666void filter (CBOR *self, SV *filter = 0)
1667 PPCODE:
1668 SvREFCNT_dec (self->filter);
1669 self->filter = filter ? newSVsv (filter) : filter;
1670 XPUSHs (ST (0));
1671
1672SV *get_filter (CBOR *self)
1673 CODE:
1674 RETVAL = self->filter ? self->filter : NEWSV (0, 0);
1064 OUTPUT: 1675 OUTPUT:
1065 RETVAL 1676 RETVAL
1066 1677
1067void encode (CBOR *self, SV *scalar) 1678void encode (CBOR *self, SV *scalar)
1068 PPCODE: 1679 PPCODE:
1083 EXTEND (SP, 2); 1694 EXTEND (SP, 2);
1084 PUSHs (sv); 1695 PUSHs (sv);
1085 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); 1696 PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr))));
1086} 1697}
1087 1698
1699void incr_parse (CBOR *self, SV *cborstr)
1700 ALIAS:
1701 incr_parse_multiple = 1
1702 PPCODE:
1703{
1704 if (SvUTF8 (cborstr))
1705 sv_utf8_downgrade (cborstr, 0);
1706
1707 if (!self->incr_count)
1708 {
1709 self->incr_count = newAV ();
1710 self->incr_pos = 0;
1711 self->incr_need = 1;
1712
1713 av_push (self->incr_count, newSViv (1));
1714 }
1715
1716 do
1717 {
1718 if (!incr_parse (self, cborstr))
1719 {
1720 if (self->incr_need > self->max_size && self->max_size)
1721 croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu",
1722 (unsigned long)self->incr_need, (unsigned long)self->max_size);
1723
1724 break;
1725 }
1726
1727 SV *sv;
1728 char *offset;
1729
1730 PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN;
1731 XPUSHs (sv);
1732
1733 sv_chop (cborstr, offset);
1734
1735 av_clear (self->incr_count);
1736 av_push (self->incr_count, newSViv (1));
1737
1738 self->incr_pos = 0;
1739 self->incr_need = self->incr_pos + 1;
1740 }
1741 while (ix);
1742}
1743
1744void incr_reset (CBOR *self)
1745 CODE:
1746{
1747 SvREFCNT_dec (self->incr_count);
1748 self->incr_count = 0;
1749}
1750
1751void DESTROY (CBOR *self)
1752 PPCODE:
1753 cbor_free (self);
1754
1088PROTOTYPES: ENABLE 1755PROTOTYPES: ENABLE
1089 1756
1090void encode_cbor (SV *scalar) 1757void encode_cbor (SV *scalar)
1758 ALIAS:
1759 encode_cbor = 0
1760 encode_cbor_sharing = F_ALLOW_SHARING
1091 PPCODE: 1761 PPCODE:
1092{ 1762{
1093 CBOR cbor; 1763 CBOR cbor;
1094 cbor_init (&cbor); 1764 cbor_init (&cbor);
1765 cbor.flags |= ix;
1095 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; 1766 PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN;
1096 XPUSHs (scalar); 1767 XPUSHs (scalar);
1097} 1768}
1098 1769
1099void decode_cbor (SV *cborstr) 1770void decode_cbor (SV *cborstr)
1103 cbor_init (&cbor); 1774 cbor_init (&cbor);
1104 PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN; 1775 PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN;
1105 XPUSHs (cborstr); 1776 XPUSHs (cborstr);
1106} 1777}
1107 1778
1779#ifdef __AFL_COMPILER
1780
1781void
1782afl_init ()
1783 CODE:
1784 __AFL_INIT ();
1785
1786int
1787afl_loop (unsigned int count = 10000)
1788 CODE:
1789 RETVAL = __AFL_LOOP (count);
1790 OUTPUT:
1791 RETVAL
1792
1793#endif
1794

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines