… | |
… | |
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)) |
… | |
… | |
23 | #endif |
24 | #endif |
24 | #ifndef SvREFCNT_dec_NN |
25 | #ifndef SvREFCNT_dec_NN |
25 | # define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv) |
26 | # define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv) |
26 | #endif |
27 | #endif |
27 | |
28 | |
|
|
29 | // known major and minor types |
|
|
30 | enum cbor_type |
|
|
31 | { |
|
|
32 | MAJOR_SHIFT = 5, |
|
|
33 | MINOR_MASK = 0x1f, |
|
|
34 | |
|
|
35 | MAJOR_POS_INT = 0 << MAJOR_SHIFT, |
|
|
36 | MAJOR_NEG_INT = 1 << MAJOR_SHIFT, |
|
|
37 | MAJOR_BYTES = 2 << MAJOR_SHIFT, |
|
|
38 | MAJOR_TEXT = 3 << MAJOR_SHIFT, |
|
|
39 | MAJOR_ARRAY = 4 << MAJOR_SHIFT, |
|
|
40 | MAJOR_MAP = 5 << MAJOR_SHIFT, |
|
|
41 | MAJOR_TAG = 6 << MAJOR_SHIFT, |
|
|
42 | MAJOR_MISC = 7 << MAJOR_SHIFT, |
|
|
43 | |
|
|
44 | // INT/STRING/ARRAY/MAP subtypes |
|
|
45 | LENGTH_EXT1 = 24, |
|
|
46 | LENGTH_EXT2 = 25, |
|
|
47 | LENGTH_EXT4 = 26, |
|
|
48 | LENGTH_EXT8 = 27, |
|
|
49 | |
|
|
50 | // SIMPLE types (effectively MISC subtypes) |
|
|
51 | SIMPLE_FALSE = 20, |
|
|
52 | SIMPLE_TRUE = 21, |
|
|
53 | SIMPLE_NULL = 22, |
|
|
54 | SIMPLE_UNDEF = 23, |
|
|
55 | |
|
|
56 | // MISC subtype (unused) |
|
|
57 | MISC_EXT1 = 24, |
|
|
58 | MISC_FLOAT16 = 25, |
|
|
59 | MISC_FLOAT32 = 26, |
|
|
60 | MISC_FLOAT64 = 27, |
|
|
61 | |
|
|
62 | // BYTES/TEXT/ARRAY/MAP |
|
|
63 | MINOR_INDEF = 31, |
|
|
64 | }; |
|
|
65 | |
28 | // known tags |
66 | // known tags |
29 | enum cbor_tag |
67 | enum cbor_tag |
30 | { |
68 | { |
31 | // extensions |
69 | // extensions |
32 | CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref |
70 | CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref |
33 | CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object |
71 | CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object |
34 | CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object |
72 | CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object |
35 | CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing |
73 | CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing |
36 | CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing |
74 | CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing |
37 | CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref |
75 | CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref |
38 | CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection |
76 | CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection |
39 | |
77 | |
40 | // rfc7049 |
78 | // rfc7049 |
41 | CBOR_TAG_DATETIME = 0, // rfc4287, utf-8 |
79 | CBOR_TAG_DATETIME = 0, // rfc4287, utf-8 |
42 | CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any |
80 | CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any |
43 | CBOR_TAG_POS_BIGNUM = 2, // byte string |
81 | CBOR_TAG_POS_BIGNUM = 2, // byte string |
44 | CBOR_TAG_NEG_BIGNUM = 3, // byte string |
82 | CBOR_TAG_NEG_BIGNUM = 3, // byte string |
45 | CBOR_TAG_DECIMAL = 4, // decimal fraction, array |
83 | CBOR_TAG_DECIMAL = 4, // decimal fraction, array |
46 | CBOR_TAG_BIGFLOAT = 5, // array |
84 | CBOR_TAG_BIGFLOAT = 5, // array |
47 | |
85 | |
48 | CBOR_TAG_CONV_B64U = 21, // base64url, any |
86 | CBOR_TAG_CONV_B64U = 21, // base64url, any |
49 | CBOR_TAG_CONV_B64 = 22, // base64, any |
87 | CBOR_TAG_CONV_B64 = 22, // base64, any |
50 | CBOR_TAG_CONV_HEX = 23, // base16, any |
88 | CBOR_TAG_CONV_HEX = 23, // base16, any |
51 | CBOR_TAG_CBOR = 24, // embedded cbor, byte string |
89 | CBOR_TAG_CBOR = 24, // embedded cbor, byte string |
52 | |
90 | |
53 | CBOR_TAG_URI = 32, // URI rfc3986, utf-8 |
91 | CBOR_TAG_URI = 32, // URI rfc3986, utf-8 |
54 | CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8 |
92 | CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8 |
55 | CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8 |
93 | CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8 |
56 | CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8 |
94 | CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8 |
57 | CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8 |
95 | CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8 |
58 | |
96 | |
59 | CBOR_TAG_MAGIC = 55799 // self-describe cbor |
97 | CBOR_TAG_MAGIC = 55799, // self-describe cbor |
60 | }; |
98 | }; |
61 | |
99 | |
62 | #define F_SHRINK 0x00000001UL |
100 | #define F_SHRINK 0x00000001UL |
63 | #define F_ALLOW_UNKNOWN 0x00000002UL |
101 | #define F_ALLOW_UNKNOWN 0x00000002UL |
64 | #define F_ALLOW_SHARING 0x00000004UL |
102 | #define F_ALLOW_SHARING 0x00000004UL |
|
|
103 | #define F_ALLOW_CYCLES 0x00000008UL |
65 | #define F_PACK_STRINGS 0x00000008UL |
104 | #define F_PACK_STRINGS 0x00000010UL |
|
|
105 | #define F_UTF8_STRINGS 0x00000020UL |
|
|
106 | #define F_VALIDATE_UTF8 0x00000040UL |
66 | |
107 | |
67 | #define INIT_SIZE 32 // initial scalar size to be allocated |
108 | #define INIT_SIZE 32 // initial scalar size to be allocated |
68 | |
109 | |
69 | #define SB do { |
110 | #define SB do { |
70 | #define SE } while (0) |
111 | #define SE } while (0) |
… | |
… | |
89 | typedef struct { |
130 | typedef struct { |
90 | U32 flags; |
131 | U32 flags; |
91 | U32 max_depth; |
132 | U32 max_depth; |
92 | STRLEN max_size; |
133 | STRLEN max_size; |
93 | SV *filter; |
134 | SV *filter; |
|
|
135 | |
|
|
136 | // for the incremental parser |
|
|
137 | STRLEN incr_pos; // the current offset into the text |
|
|
138 | STRLEN incr_need; // minimum bytes needed to decode |
|
|
139 | AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef. |
94 | } CBOR; |
140 | } CBOR; |
95 | |
141 | |
96 | ecb_inline void |
142 | ecb_inline void |
97 | cbor_init (CBOR *cbor) |
143 | cbor_init (CBOR *cbor) |
98 | { |
144 | { |
… | |
… | |
102 | |
148 | |
103 | ecb_inline void |
149 | ecb_inline void |
104 | cbor_free (CBOR *cbor) |
150 | cbor_free (CBOR *cbor) |
105 | { |
151 | { |
106 | SvREFCNT_dec (cbor->filter); |
152 | SvREFCNT_dec (cbor->filter); |
|
|
153 | SvREFCNT_dec (cbor->incr_count); |
107 | } |
154 | } |
108 | |
155 | |
109 | ///////////////////////////////////////////////////////////////////////////// |
156 | ///////////////////////////////////////////////////////////////////////////// |
110 | // utility functions |
157 | // utility functions |
111 | |
158 | |
… | |
… | |
189 | static void |
236 | static void |
190 | encode_uint (enc_t *enc, int major, UV len) |
237 | encode_uint (enc_t *enc, int major, UV len) |
191 | { |
238 | { |
192 | need (enc, 9); |
239 | need (enc, 9); |
193 | |
240 | |
194 | if (ecb_expect_true (len < 24)) |
241 | if (ecb_expect_true (len < LENGTH_EXT1)) |
195 | *enc->cur++ = major | len; |
242 | *enc->cur++ = major | len; |
196 | else if (ecb_expect_true (len <= 0xff)) |
243 | else if (ecb_expect_true (len <= 0xffU)) |
197 | { |
244 | { |
198 | *enc->cur++ = major | 24; |
245 | *enc->cur++ = major | LENGTH_EXT1; |
199 | *enc->cur++ = len; |
246 | *enc->cur++ = len; |
200 | } |
247 | } |
201 | else if (len <= 0xffff) |
248 | else if (len <= 0xffffU) |
202 | { |
249 | { |
203 | *enc->cur++ = major | 25; |
250 | *enc->cur++ = major | LENGTH_EXT2; |
204 | *enc->cur++ = len >> 8; |
251 | *enc->cur++ = len >> 8; |
205 | *enc->cur++ = len; |
252 | *enc->cur++ = len; |
206 | } |
253 | } |
207 | else if (len <= 0xffffffff) |
254 | else if (len <= 0xffffffffU) |
208 | { |
255 | { |
209 | *enc->cur++ = major | 26; |
256 | *enc->cur++ = major | LENGTH_EXT4; |
210 | *enc->cur++ = len >> 24; |
257 | *enc->cur++ = len >> 24; |
211 | *enc->cur++ = len >> 16; |
258 | *enc->cur++ = len >> 16; |
212 | *enc->cur++ = len >> 8; |
259 | *enc->cur++ = len >> 8; |
213 | *enc->cur++ = len; |
260 | *enc->cur++ = len; |
214 | } |
261 | } |
215 | else |
262 | else |
216 | { |
263 | { |
217 | *enc->cur++ = major | 27; |
264 | *enc->cur++ = major | LENGTH_EXT8; |
218 | *enc->cur++ = len >> 56; |
265 | *enc->cur++ = len >> 56; |
219 | *enc->cur++ = len >> 48; |
266 | *enc->cur++ = len >> 48; |
220 | *enc->cur++ = len >> 40; |
267 | *enc->cur++ = len >> 40; |
221 | *enc->cur++ = len >> 32; |
268 | *enc->cur++ = len >> 32; |
222 | *enc->cur++ = len >> 24; |
269 | *enc->cur++ = len >> 24; |
… | |
… | |
227 | } |
274 | } |
228 | |
275 | |
229 | ecb_inline void |
276 | ecb_inline void |
230 | encode_tag (enc_t *enc, UV tag) |
277 | encode_tag (enc_t *enc, UV tag) |
231 | { |
278 | { |
232 | encode_uint (enc, 0xc0, tag); |
279 | encode_uint (enc, MAJOR_TAG, tag); |
233 | } |
280 | } |
234 | |
281 | |
235 | ecb_inline void |
282 | ecb_inline void |
236 | encode_str (enc_t *enc, int utf8, char *str, STRLEN len) |
283 | encode_str (enc_t *enc, int utf8, char *str, STRLEN len) |
237 | { |
284 | { |
238 | encode_uint (enc, utf8 ? 0x60 : 0x40, len); |
285 | if (ecb_expect_false (enc->cbor.flags & F_UTF8_STRINGS)) |
|
|
286 | if (!utf8) |
|
|
287 | { |
|
|
288 | // exceptional path for bytze strings that need to be utf8-encoded |
|
|
289 | STRLEN ulen = len; |
|
|
290 | U8 *p, *pend = (U8 *)str + len; |
|
|
291 | |
|
|
292 | for (p = (U8 *)str; p < pend; ++p) |
|
|
293 | ulen += *p >> 7; // count set high bits |
|
|
294 | |
|
|
295 | encode_uint (enc, MAJOR_TEXT, ulen); |
|
|
296 | |
|
|
297 | need (enc, ulen); |
|
|
298 | for (p = (U8 *)str; p < pend; ++p) |
|
|
299 | if (*p < 0x80) |
|
|
300 | *enc->cur++ = *p; |
|
|
301 | else |
|
|
302 | { |
|
|
303 | *enc->cur++ = 0xc0 + (*p >> 6); |
|
|
304 | *enc->cur++ = 0x80 + (*p & 63); |
|
|
305 | } |
|
|
306 | |
|
|
307 | return; |
|
|
308 | } |
|
|
309 | |
|
|
310 | encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len); |
239 | need (enc, len); |
311 | need (enc, len); |
240 | memcpy (enc->cur, str, len); |
312 | memcpy (enc->cur, str, len); |
241 | enc->cur += len; |
313 | enc->cur += len; |
242 | } |
314 | } |
243 | |
315 | |
… | |
… | |
250 | |
322 | |
251 | if (SvOK (*svp)) |
323 | if (SvOK (*svp)) |
252 | { |
324 | { |
253 | // already registered, use stringref |
325 | // already registered, use stringref |
254 | encode_tag (enc, CBOR_TAG_STRINGREF); |
326 | encode_tag (enc, CBOR_TAG_STRINGREF); |
255 | encode_uint (enc, 0x00, SvUV (*svp)); |
327 | encode_uint (enc, MAJOR_POS_INT, SvUV (*svp)); |
256 | return; |
328 | return; |
257 | } |
329 | } |
258 | else if (len >= minimum_string_length (enc->stringref_idx)) |
330 | else if (len >= minimum_string_length (enc->stringref_idx)) |
259 | { |
331 | { |
260 | // register only |
332 | // register only |
… | |
… | |
276 | if (enc->depth >= enc->cbor.max_depth) |
348 | if (enc->depth >= enc->cbor.max_depth) |
277 | croak (ERR_NESTING_EXCEEDED); |
349 | croak (ERR_NESTING_EXCEEDED); |
278 | |
350 | |
279 | ++enc->depth; |
351 | ++enc->depth; |
280 | |
352 | |
281 | encode_uint (enc, 0x80, len + 1); |
353 | encode_uint (enc, MAJOR_ARRAY, len + 1); |
282 | |
354 | |
|
|
355 | if (SvMAGICAL (av)) |
283 | for (i = 0; i <= len; ++i) |
356 | for (i = 0; i <= len; ++i) |
284 | { |
357 | { |
285 | SV **svp = av_fetch (av, i, 0); |
358 | SV **svp = av_fetch (av, i, 0); |
286 | encode_sv (enc, svp ? *svp : &PL_sv_undef); |
359 | encode_sv (enc, svp ? *svp : &PL_sv_undef); |
287 | } |
360 | } |
|
|
361 | else |
|
|
362 | for (i = 0; i <= len; ++i) |
|
|
363 | { |
|
|
364 | SV *sv = AvARRAY (av)[i]; |
|
|
365 | encode_sv (enc, sv ? sv : &PL_sv_undef); |
|
|
366 | } |
288 | |
367 | |
289 | --enc->depth; |
368 | --enc->depth; |
290 | } |
369 | } |
291 | |
370 | |
292 | static void |
371 | static void |
… | |
… | |
301 | |
380 | |
302 | int pairs = hv_iterinit (hv); |
381 | int pairs = hv_iterinit (hv); |
303 | int mg = SvMAGICAL (hv); |
382 | int mg = SvMAGICAL (hv); |
304 | |
383 | |
305 | if (mg) |
384 | if (mg) |
306 | encode_ch (enc, 0xa0 | 31); |
385 | encode_ch (enc, MAJOR_MAP | MINOR_INDEF); |
307 | else |
386 | else |
308 | encode_uint (enc, 0xa0, pairs); |
387 | encode_uint (enc, MAJOR_MAP, pairs); |
309 | |
388 | |
310 | while ((he = hv_iternext (hv))) |
389 | while ((he = hv_iternext (hv))) |
311 | { |
390 | { |
312 | if (HeKLEN (he) == HEf_SVKEY) |
391 | if (HeKLEN (he) == HEf_SVKEY) |
313 | encode_sv (enc, HeSVKEY (he)); |
392 | encode_sv (enc, HeSVKEY (he)); |
… | |
… | |
316 | |
395 | |
317 | encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); |
396 | encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); |
318 | } |
397 | } |
319 | |
398 | |
320 | if (mg) |
399 | if (mg) |
321 | encode_ch (enc, 0xe0 | 31); |
400 | encode_ch (enc, MAJOR_MISC | MINOR_INDEF); |
322 | |
401 | |
323 | --enc->depth; |
402 | --enc->depth; |
324 | } |
403 | } |
325 | |
404 | |
326 | // encode objects, arrays and special \0=false and \1=true values. |
405 | // encode objects, arrays and special \0=false and \1=true values. |
… | |
… | |
342 | HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash |
421 | HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash |
343 | ? cbor_tagged_stash |
422 | ? cbor_tagged_stash |
344 | : gv_stashpv ("CBOR::XS::Tagged" , 1); |
423 | : gv_stashpv ("CBOR::XS::Tagged" , 1); |
345 | |
424 | |
346 | HV *stash = SvSTASH (sv); |
425 | HV *stash = SvSTASH (sv); |
347 | GV *method; |
|
|
348 | |
426 | |
349 | if (stash == boolean_stash) |
427 | if (stash == boolean_stash) |
350 | encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20); |
428 | { |
|
|
429 | encode_ch (enc, SvIV (sv) ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE); |
|
|
430 | return; |
|
|
431 | } |
351 | else if (stash == error_stash) |
432 | else if (stash == error_stash) |
352 | encode_ch (enc, 0xe0 | 23); |
433 | { |
|
|
434 | encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF); |
|
|
435 | return; |
|
|
436 | } |
353 | else if (stash == tagged_stash) |
437 | else if (stash == tagged_stash) |
354 | { |
438 | { |
355 | if (svt != SVt_PVAV) |
439 | if (svt != SVt_PVAV) |
356 | croak ("encountered CBOR::XS::Tagged object that isn't an array"); |
440 | croak ("encountered CBOR::XS::Tagged object that isn't an array"); |
357 | |
441 | |
358 | encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); |
442 | encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1))); |
359 | encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); |
443 | encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); |
|
|
444 | |
|
|
445 | return; |
|
|
446 | } |
|
|
447 | } |
|
|
448 | |
|
|
449 | if (ecb_expect_false (SvREFCNT (sv) > 1) |
|
|
450 | && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)) |
|
|
451 | { |
|
|
452 | if (!enc->shareable) |
|
|
453 | enc->shareable = (HV *)sv_2mortal ((SV *)newHV ()); |
|
|
454 | |
|
|
455 | SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1); |
|
|
456 | |
|
|
457 | if (SvOK (*svp)) |
360 | } |
458 | { |
|
|
459 | encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF); |
|
|
460 | encode_uint (enc, MAJOR_POS_INT, SvUV (*svp)); |
|
|
461 | return; |
|
|
462 | } |
|
|
463 | else |
|
|
464 | { |
|
|
465 | sv_setuv (*svp, enc->shareable_idx); |
|
|
466 | ++enc->shareable_idx; |
|
|
467 | encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE); |
|
|
468 | } |
|
|
469 | } |
|
|
470 | |
|
|
471 | if (ecb_expect_false (SvOBJECT (sv))) |
|
|
472 | { |
|
|
473 | HV *stash = SvSTASH (sv); |
|
|
474 | GV *method; |
|
|
475 | |
361 | else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) |
476 | if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) |
362 | { |
477 | { |
363 | dSP; |
478 | dSP; |
364 | |
479 | |
365 | ENTER; SAVETMPS; PUSHMARK (SP); |
480 | ENTER; SAVETMPS; |
|
|
481 | PUSHMARK (SP); |
366 | // we re-bless the reference to get overload and other niceties right |
482 | // we re-bless the reference to get overload and other niceties right |
367 | XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); |
483 | XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); |
368 | |
484 | |
369 | PUTBACK; |
485 | PUTBACK; |
370 | // G_SCALAR ensures that return value is 1 |
486 | // G_SCALAR ensures that return value is 1 |
… | |
… | |
383 | } |
499 | } |
384 | else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0) |
500 | else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0) |
385 | { |
501 | { |
386 | dSP; |
502 | dSP; |
387 | |
503 | |
388 | ENTER; SAVETMPS; PUSHMARK (SP); |
504 | ENTER; SAVETMPS; |
|
|
505 | SAVESTACK_POS (); |
|
|
506 | PUSHMARK (SP); |
389 | EXTEND (SP, 2); |
507 | EXTEND (SP, 2); |
390 | // we re-bless the reference to get overload and other niceties right |
508 | // we re-bless the reference to get overload and other niceties right |
391 | PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); |
509 | PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); |
392 | PUSHs (sv_cbor); |
510 | PUSHs (sv_cbor); |
393 | |
511 | |
… | |
… | |
398 | // catch this surprisingly common error |
516 | // catch this surprisingly common error |
399 | if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) |
517 | if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) |
400 | croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash)); |
518 | croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash)); |
401 | |
519 | |
402 | encode_tag (enc, CBOR_TAG_PERL_OBJECT); |
520 | encode_tag (enc, CBOR_TAG_PERL_OBJECT); |
403 | encode_uint (enc, 0x80, count + 1); |
521 | encode_uint (enc, MAJOR_ARRAY, count + 1); |
404 | encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); |
522 | encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); |
405 | |
523 | |
406 | while (count) |
524 | while (count) |
407 | encode_sv (enc, SP[1 - count--]); |
525 | encode_sv (enc, SP[1 - count--]); |
408 | |
526 | |
… | |
… | |
418 | encode_hv (enc, (HV *)sv); |
536 | encode_hv (enc, (HV *)sv); |
419 | else if (svt == SVt_PVAV) |
537 | else if (svt == SVt_PVAV) |
420 | encode_av (enc, (AV *)sv); |
538 | encode_av (enc, (AV *)sv); |
421 | else |
539 | else |
422 | { |
540 | { |
423 | if (ecb_expect_false (SvREFCNT (sv) > 1) |
|
|
424 | && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)) |
|
|
425 | { |
|
|
426 | if (!enc->shareable) |
|
|
427 | enc->shareable = (HV *)sv_2mortal ((SV *)newHV ()); |
|
|
428 | |
|
|
429 | SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1); |
|
|
430 | |
|
|
431 | if (SvOK (*svp)) |
|
|
432 | { |
|
|
433 | encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF); |
|
|
434 | encode_uint (enc, 0x00, 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 | encode_tag (enc, CBOR_TAG_INDIRECTION); |
541 | encode_tag (enc, CBOR_TAG_INDIRECTION); |
446 | encode_sv (enc, sv); |
542 | encode_sv (enc, sv); |
447 | } |
543 | } |
448 | } |
544 | } |
449 | |
545 | |
… | |
… | |
452 | { |
548 | { |
453 | double nv = SvNVX (sv); |
549 | double nv = SvNVX (sv); |
454 | |
550 | |
455 | need (enc, 9); |
551 | need (enc, 9); |
456 | |
552 | |
457 | if (ecb_expect_false (nv == (U32)nv)) |
553 | if (ecb_expect_false (nv == (NV)(U32)nv)) |
458 | encode_uint (enc, 0x00, (U32)nv); |
554 | encode_uint (enc, MAJOR_POS_INT, (U32)nv); |
459 | //TODO: maybe I32? |
555 | //TODO: maybe I32? |
460 | else if (ecb_expect_false (nv == (float)nv)) |
556 | else if (ecb_expect_false (nv == (float)nv)) |
461 | { |
557 | { |
462 | uint32_t fp = ecb_float_to_binary32 (nv); |
558 | uint32_t fp = ecb_float_to_binary32 (nv); |
463 | |
559 | |
464 | *enc->cur++ = 0xe0 | 26; |
560 | *enc->cur++ = MAJOR_MISC | MISC_FLOAT32; |
465 | |
561 | |
466 | if (!ecb_big_endian ()) |
562 | if (!ecb_big_endian ()) |
467 | fp = ecb_bswap32 (fp); |
563 | fp = ecb_bswap32 (fp); |
468 | |
564 | |
469 | memcpy (enc->cur, &fp, 4); |
565 | memcpy (enc->cur, &fp, 4); |
… | |
… | |
471 | } |
567 | } |
472 | else |
568 | else |
473 | { |
569 | { |
474 | uint64_t fp = ecb_double_to_binary64 (nv); |
570 | uint64_t fp = ecb_double_to_binary64 (nv); |
475 | |
571 | |
476 | *enc->cur++ = 0xe0 | 27; |
572 | *enc->cur++ = MAJOR_MISC | MISC_FLOAT64; |
477 | |
573 | |
478 | if (!ecb_big_endian ()) |
574 | if (!ecb_big_endian ()) |
479 | fp = ecb_bswap64 (fp); |
575 | fp = ecb_bswap64 (fp); |
480 | |
576 | |
481 | memcpy (enc->cur, &fp, 8); |
577 | memcpy (enc->cur, &fp, 8); |
… | |
… | |
497 | else if (SvNOKp (sv)) |
593 | else if (SvNOKp (sv)) |
498 | encode_nv (enc, sv); |
594 | encode_nv (enc, sv); |
499 | else if (SvIOKp (sv)) |
595 | else if (SvIOKp (sv)) |
500 | { |
596 | { |
501 | if (SvIsUV (sv)) |
597 | if (SvIsUV (sv)) |
502 | encode_uint (enc, 0x00, SvUVX (sv)); |
598 | encode_uint (enc, MAJOR_POS_INT, SvUVX (sv)); |
503 | else if (SvIVX (sv) >= 0) |
599 | else if (SvIVX (sv) >= 0) |
504 | encode_uint (enc, 0x00, SvIVX (sv)); |
600 | encode_uint (enc, MAJOR_POS_INT, SvIVX (sv)); |
505 | else |
601 | else |
506 | encode_uint (enc, 0x20, -(SvIVX (sv) + 1)); |
602 | encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1)); |
507 | } |
603 | } |
508 | else if (SvROK (sv)) |
604 | else if (SvROK (sv)) |
509 | encode_rv (enc, SvRV (sv)); |
605 | encode_rv (enc, SvRV (sv)); |
510 | else if (!SvOK (sv)) |
606 | else if (!SvOK (sv)) |
511 | encode_ch (enc, 0xe0 | 22); |
607 | encode_ch (enc, MAJOR_MISC | SIMPLE_NULL); |
512 | else if (enc->cbor.flags & F_ALLOW_UNKNOWN) |
608 | else if (enc->cbor.flags & F_ALLOW_UNKNOWN) |
513 | encode_ch (enc, 0xe0 | 23); |
609 | encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF); |
514 | else |
610 | else |
515 | croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data", |
611 | croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data", |
516 | SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); |
612 | SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); |
517 | } |
613 | } |
518 | |
614 | |
519 | static SV * |
615 | static SV * |
520 | encode_cbor (SV *scalar, CBOR *cbor) |
616 | encode_cbor (SV *scalar, CBOR *cbor) |
521 | { |
617 | { |
522 | enc_t enc = { }; |
618 | enc_t enc = { 0 }; |
523 | |
619 | |
524 | enc.cbor = *cbor; |
620 | enc.cbor = *cbor; |
525 | enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); |
621 | enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); |
526 | enc.cur = SvPVX (enc.sv); |
622 | enc.cur = SvPVX (enc.sv); |
527 | enc.end = SvEND (enc.sv); |
623 | enc.end = SvEND (enc.sv); |
… | |
… | |
571 | #define DEC_DEC_DEPTH --dec->depth |
667 | #define DEC_DEC_DEPTH --dec->depth |
572 | |
668 | |
573 | static UV |
669 | static UV |
574 | decode_uint (dec_t *dec) |
670 | decode_uint (dec_t *dec) |
575 | { |
671 | { |
576 | switch (*dec->cur & 31) |
672 | U8 m = *dec->cur & MINOR_MASK; |
577 | { |
673 | ++dec->cur; |
578 | case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7: |
|
|
579 | case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15: |
|
|
580 | case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23: |
|
|
581 | return *dec->cur++ & 31; |
|
|
582 | |
674 | |
583 | case 24: |
675 | if (ecb_expect_true (m < LENGTH_EXT1)) |
|
|
676 | return m; |
|
|
677 | else if (ecb_expect_true (m == LENGTH_EXT1)) |
|
|
678 | { |
584 | WANT (2); |
679 | WANT (1); |
585 | dec->cur += 2; |
680 | dec->cur += 1; |
586 | return dec->cur[-1]; |
681 | return dec->cur[-1]; |
587 | |
682 | } |
588 | case 25: |
683 | else if (ecb_expect_true (m == LENGTH_EXT2)) |
|
|
684 | { |
589 | WANT (3); |
685 | WANT (2); |
590 | dec->cur += 3; |
686 | dec->cur += 2; |
591 | return (((UV)dec->cur[-2]) << 8) |
687 | return (((UV)dec->cur[-2]) << 8) |
592 | | ((UV)dec->cur[-1]); |
688 | | ((UV)dec->cur[-1]); |
593 | |
689 | } |
594 | case 26: |
690 | else if (ecb_expect_true (m == LENGTH_EXT4)) |
|
|
691 | { |
595 | WANT (5); |
692 | WANT (4); |
596 | dec->cur += 5; |
693 | dec->cur += 4; |
597 | return (((UV)dec->cur[-4]) << 24) |
694 | return (((UV)dec->cur[-4]) << 24) |
598 | | (((UV)dec->cur[-3]) << 16) |
695 | | (((UV)dec->cur[-3]) << 16) |
599 | | (((UV)dec->cur[-2]) << 8) |
696 | | (((UV)dec->cur[-2]) << 8) |
600 | | ((UV)dec->cur[-1]); |
697 | | ((UV)dec->cur[-1]); |
601 | |
698 | } |
602 | case 27: |
699 | else if (ecb_expect_true (m == LENGTH_EXT8)) |
|
|
700 | { |
603 | WANT (9); |
701 | WANT (8); |
604 | dec->cur += 9; |
702 | dec->cur += 8; |
|
|
703 | |
|
|
704 | return |
|
|
705 | #if UVSIZE < 8 |
|
|
706 | 0 |
|
|
707 | #else |
605 | return (((UV)dec->cur[-8]) << 56) |
708 | (((UV)dec->cur[-8]) << 56) |
606 | | (((UV)dec->cur[-7]) << 48) |
709 | | (((UV)dec->cur[-7]) << 48) |
607 | | (((UV)dec->cur[-6]) << 40) |
710 | | (((UV)dec->cur[-6]) << 40) |
608 | | (((UV)dec->cur[-5]) << 32) |
711 | | (((UV)dec->cur[-5]) << 32) |
|
|
712 | #endif |
609 | | (((UV)dec->cur[-4]) << 24) |
713 | | (((UV)dec->cur[-4]) << 24) |
610 | | (((UV)dec->cur[-3]) << 16) |
714 | | (((UV)dec->cur[-3]) << 16) |
611 | | (((UV)dec->cur[-2]) << 8) |
715 | | (((UV)dec->cur[-2]) << 8) |
612 | | ((UV)dec->cur[-1]); |
716 | | ((UV)dec->cur[-1]); |
613 | |
717 | } |
614 | default: |
718 | else |
615 | ERR ("corrupted CBOR data (unsupported integer minor encoding)"); |
719 | ERR ("corrupted CBOR data (unsupported integer minor encoding)"); |
616 | } |
|
|
617 | |
720 | |
618 | fail: |
721 | fail: |
619 | return 0; |
722 | return 0; |
620 | } |
723 | } |
621 | |
724 | |
… | |
… | |
626 | { |
729 | { |
627 | AV *av = newAV (); |
730 | AV *av = newAV (); |
628 | |
731 | |
629 | DEC_INC_DEPTH; |
732 | DEC_INC_DEPTH; |
630 | |
733 | |
631 | if ((*dec->cur & 31) == 31) |
734 | if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF)) |
632 | { |
735 | { |
633 | ++dec->cur; |
736 | ++dec->cur; |
634 | |
737 | |
635 | for (;;) |
738 | for (;;) |
636 | { |
739 | { |
637 | WANT (1); |
740 | WANT (1); |
638 | |
741 | |
639 | if (*dec->cur == (0xe0 | 31)) |
742 | if (*dec->cur == (MAJOR_MISC | MINOR_INDEF)) |
640 | { |
743 | { |
641 | ++dec->cur; |
744 | ++dec->cur; |
642 | break; |
745 | break; |
643 | } |
746 | } |
644 | |
747 | |
… | |
… | |
647 | } |
750 | } |
648 | else |
751 | else |
649 | { |
752 | { |
650 | int i, len = decode_uint (dec); |
753 | int i, len = decode_uint (dec); |
651 | |
754 | |
|
|
755 | WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays |
652 | av_fill (av, len - 1); |
756 | av_fill (av, len - 1); |
653 | |
757 | |
654 | for (i = 0; i < len; ++i) |
758 | for (i = 0; i < len; ++i) |
655 | AvARRAY (av)[i] = decode_sv (dec); |
759 | AvARRAY (av)[i] = decode_sv (dec); |
656 | } |
760 | } |
… | |
… | |
669 | { |
773 | { |
670 | // for speed reasons, we specialcase single-string |
774 | // for speed reasons, we specialcase single-string |
671 | // byte or utf-8 strings as keys, but only when !stringref |
775 | // byte or utf-8 strings as keys, but only when !stringref |
672 | |
776 | |
673 | if (ecb_expect_true (!dec->stringref)) |
777 | if (ecb_expect_true (!dec->stringref)) |
674 | if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27) |
778 | if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8)) |
675 | { |
779 | { |
676 | I32 len = decode_uint (dec); |
780 | I32 len = decode_uint (dec); |
677 | char *key = (char *)dec->cur; |
781 | char *key = (char *)dec->cur; |
678 | |
782 | |
|
|
783 | WANT (len); |
679 | dec->cur += len; |
784 | dec->cur += len; |
680 | |
785 | |
681 | if (ecb_expect_false (dec->stringref)) |
|
|
682 | av_push (dec->stringref, newSVpvn (key, len)); |
|
|
683 | |
|
|
684 | hv_store (hv, key, len, decode_sv (dec), 0); |
786 | hv_store (hv, key, len, decode_sv (dec), 0); |
685 | |
787 | |
686 | return; |
788 | return; |
687 | } |
789 | } |
688 | else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27) |
790 | else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8)) |
689 | { |
791 | { |
690 | I32 len = decode_uint (dec); |
792 | I32 len = decode_uint (dec); |
691 | char *key = (char *)dec->cur; |
793 | char *key = (char *)dec->cur; |
692 | |
794 | |
|
|
795 | WANT (len); |
693 | dec->cur += len; |
796 | dec->cur += len; |
694 | |
797 | |
695 | if (ecb_expect_false (dec->stringref)) |
798 | if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8)) |
696 | av_push (dec->stringref, newSVpvn_utf8 (key, len, 1)); |
799 | if (!is_utf8_string (key, len)) |
|
|
800 | ERR ("corrupted CBOR data (invalid UTF-8 in map key)"); |
697 | |
801 | |
698 | hv_store (hv, key, -len, decode_sv (dec), 0); |
802 | hv_store (hv, key, -len, decode_sv (dec), 0); |
699 | |
803 | |
700 | return; |
804 | return; |
701 | } |
805 | } |
… | |
… | |
703 | SV *k = decode_sv (dec); |
807 | SV *k = decode_sv (dec); |
704 | SV *v = decode_sv (dec); |
808 | SV *v = decode_sv (dec); |
705 | |
809 | |
706 | hv_store_ent (hv, k, v, 0); |
810 | hv_store_ent (hv, k, v, 0); |
707 | SvREFCNT_dec (k); |
811 | SvREFCNT_dec (k); |
|
|
812 | |
|
|
813 | fail: |
|
|
814 | ; |
708 | } |
815 | } |
709 | |
816 | |
710 | static SV * |
817 | static SV * |
711 | decode_hv (dec_t *dec) |
818 | decode_hv (dec_t *dec) |
712 | { |
819 | { |
713 | HV *hv = newHV (); |
820 | HV *hv = newHV (); |
714 | |
821 | |
715 | DEC_INC_DEPTH; |
822 | DEC_INC_DEPTH; |
716 | |
823 | |
717 | if ((*dec->cur & 31) == 31) |
824 | if (*dec->cur == (MAJOR_MAP | MINOR_INDEF)) |
718 | { |
825 | { |
719 | ++dec->cur; |
826 | ++dec->cur; |
720 | |
827 | |
721 | for (;;) |
828 | for (;;) |
722 | { |
829 | { |
723 | WANT (1); |
830 | WANT (1); |
724 | |
831 | |
725 | if (*dec->cur == (0xe0 | 31)) |
832 | if (*dec->cur == (MAJOR_MISC | MINOR_INDEF)) |
726 | { |
833 | { |
727 | ++dec->cur; |
834 | ++dec->cur; |
728 | break; |
835 | break; |
729 | } |
836 | } |
730 | |
837 | |
… | |
… | |
751 | static SV * |
858 | static SV * |
752 | decode_str (dec_t *dec, int utf8) |
859 | decode_str (dec_t *dec, int utf8) |
753 | { |
860 | { |
754 | SV *sv = 0; |
861 | SV *sv = 0; |
755 | |
862 | |
756 | if ((*dec->cur & 31) == 31) |
863 | if ((*dec->cur & MINOR_MASK) == MINOR_INDEF) |
757 | { |
864 | { |
758 | // indefinite length strings |
865 | // indefinite length strings |
759 | ++dec->cur; |
866 | ++dec->cur; |
760 | |
867 | |
761 | unsigned char major = *dec->cur & 0xe0; |
868 | U8 major = *dec->cur & MAJOR_MISC; |
762 | |
869 | |
763 | sv = newSVpvn ("", 0); |
870 | sv = newSVpvn ("", 0); |
764 | |
871 | |
765 | for (;;) |
872 | for (;;) |
766 | { |
873 | { |
767 | WANT (1); |
874 | WANT (1); |
768 | |
875 | |
769 | if ((*dec->cur ^ major) >= 31) |
876 | if ((*dec->cur - major) > LENGTH_EXT8) |
770 | if (*dec->cur == (0xe0 | 31)) |
877 | if (*dec->cur == (MAJOR_MISC | MINOR_INDEF)) |
771 | { |
878 | { |
772 | ++dec->cur; |
879 | ++dec->cur; |
773 | break; |
880 | break; |
774 | } |
881 | } |
775 | else |
882 | else |
… | |
… | |
794 | && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1)) |
901 | && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1)) |
795 | av_push (dec->stringref, SvREFCNT_inc_NN (sv)); |
902 | av_push (dec->stringref, SvREFCNT_inc_NN (sv)); |
796 | } |
903 | } |
797 | |
904 | |
798 | if (utf8) |
905 | if (utf8) |
|
|
906 | { |
|
|
907 | if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8)) |
|
|
908 | if (!is_utf8_string (SvPVX (sv), SvCUR (sv))) |
|
|
909 | ERR ("corrupted CBOR data (invalid UTF-8 in text string)"); |
|
|
910 | |
799 | SvUTF8_on (sv); |
911 | SvUTF8_on (sv); |
|
|
912 | } |
800 | |
913 | |
801 | return sv; |
914 | return sv; |
802 | |
915 | |
803 | fail: |
916 | fail: |
804 | SvREFCNT_dec (sv); |
917 | SvREFCNT_dec (sv); |
… | |
… | |
836 | } |
949 | } |
837 | break; |
950 | break; |
838 | |
951 | |
839 | case CBOR_TAG_STRINGREF: |
952 | case CBOR_TAG_STRINGREF: |
840 | { |
953 | { |
841 | if ((*dec->cur >> 5) != 0) |
954 | if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT)) |
842 | ERR ("corrupted CBOR data (stringref index not an unsigned integer)"); |
955 | ERR ("corrupted CBOR data (stringref index not an unsigned integer)"); |
843 | |
956 | |
844 | UV idx = decode_uint (dec); |
957 | UV idx = decode_uint (dec); |
845 | |
958 | |
846 | if (!dec->stringref || (int)idx > AvFILLp (dec->stringref)) |
959 | if (!dec->stringref || (int)idx > AvFILLp (dec->stringref)) |
… | |
… | |
853 | case CBOR_TAG_VALUE_SHAREABLE: |
966 | case CBOR_TAG_VALUE_SHAREABLE: |
854 | { |
967 | { |
855 | if (ecb_expect_false (!dec->shareable)) |
968 | if (ecb_expect_false (!dec->shareable)) |
856 | dec->shareable = (AV *)sv_2mortal ((SV *)newAV ()); |
969 | dec->shareable = (AV *)sv_2mortal ((SV *)newAV ()); |
857 | |
970 | |
|
|
971 | if (dec->cbor.flags & F_ALLOW_CYCLES) |
|
|
972 | { |
858 | sv = newSV (0); |
973 | sv = newSV (0); |
859 | av_push (dec->shareable, SvREFCNT_inc_NN (sv)); |
974 | av_push (dec->shareable, SvREFCNT_inc_NN (sv)); |
860 | |
975 | |
861 | SV *osv = decode_sv (dec); |
976 | SV *osv = decode_sv (dec); |
862 | sv_setsv (sv, osv); |
977 | sv_setsv (sv, osv); |
863 | SvREFCNT_dec_NN (osv); |
978 | SvREFCNT_dec_NN (osv); |
|
|
979 | } |
|
|
980 | else |
|
|
981 | { |
|
|
982 | av_push (dec->shareable, &PL_sv_undef); |
|
|
983 | int idx = AvFILLp (dec->shareable); |
|
|
984 | sv = decode_sv (dec); |
|
|
985 | av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv)); |
|
|
986 | } |
864 | } |
987 | } |
865 | break; |
988 | break; |
866 | |
989 | |
867 | case CBOR_TAG_VALUE_SHAREDREF: |
990 | case CBOR_TAG_VALUE_SHAREDREF: |
868 | { |
991 | { |
869 | if ((*dec->cur >> 5) != 0) |
992 | if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT)) |
870 | ERR ("corrupted CBOR data (sharedref index not an unsigned integer)"); |
993 | ERR ("corrupted CBOR data (sharedref index not an unsigned integer)"); |
871 | |
994 | |
872 | UV idx = decode_uint (dec); |
995 | UV idx = decode_uint (dec); |
873 | |
996 | |
874 | if (!dec->shareable || (int)idx > AvFILLp (dec->shareable)) |
997 | if (!dec->shareable || (int)idx > AvFILLp (dec->shareable)) |
875 | ERR ("corrupted CBOR data (sharedref index out of bounds)"); |
998 | ERR ("corrupted CBOR data (sharedref index out of bounds)"); |
876 | |
999 | |
877 | sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]); |
1000 | sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]); |
|
|
1001 | |
|
|
1002 | if (sv == &PL_sv_undef) |
|
|
1003 | ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled"); |
878 | } |
1004 | } |
879 | break; |
1005 | break; |
880 | |
1006 | |
881 | case CBOR_TAG_PERL_OBJECT: |
1007 | case CBOR_TAG_PERL_OBJECT: |
882 | { |
1008 | { |
… | |
… | |
897 | if (!method) |
1023 | if (!method) |
898 | ERR ("cannot decode perl-object (package does not have a THAW method)"); |
1024 | ERR ("cannot decode perl-object (package does not have a THAW method)"); |
899 | |
1025 | |
900 | dSP; |
1026 | dSP; |
901 | |
1027 | |
902 | ENTER; SAVETMPS; PUSHMARK (SP); |
1028 | ENTER; SAVETMPS; |
|
|
1029 | PUSHMARK (SP); |
903 | EXTEND (SP, len + 1); |
1030 | EXTEND (SP, len + 1); |
904 | // we re-bless the reference to get overload and other niceties right |
1031 | // we re-bless the reference to get overload and other niceties right |
905 | PUSHs (*av_fetch (av, 0, 1)); |
1032 | PUSHs (*av_fetch (av, 0, 1)); |
906 | PUSHs (sv_cbor); |
1033 | PUSHs (sv_cbor); |
907 | |
1034 | |
… | |
… | |
932 | default: |
1059 | default: |
933 | { |
1060 | { |
934 | sv = decode_sv (dec); |
1061 | sv = decode_sv (dec); |
935 | |
1062 | |
936 | dSP; |
1063 | dSP; |
937 | ENTER; SAVETMPS; PUSHMARK (SP); |
1064 | ENTER; SAVETMPS; |
|
|
1065 | SAVESTACK_POS (); |
|
|
1066 | PUSHMARK (SP); |
938 | EXTEND (SP, 2); |
1067 | EXTEND (SP, 2); |
939 | PUSHs (newSVuv (tag)); |
1068 | PUSHs (newSVuv (tag)); |
940 | PUSHs (sv); |
1069 | PUSHs (sv); |
941 | |
1070 | |
942 | PUTBACK; |
1071 | PUTBACK; |
… | |
… | |
983 | static SV * |
1112 | static SV * |
984 | decode_sv (dec_t *dec) |
1113 | decode_sv (dec_t *dec) |
985 | { |
1114 | { |
986 | WANT (1); |
1115 | WANT (1); |
987 | |
1116 | |
988 | switch (*dec->cur >> 5) |
1117 | switch (*dec->cur >> MAJOR_SHIFT) |
989 | { |
1118 | { |
990 | case 0: // unsigned int |
1119 | case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec)); |
991 | return newSVuv (decode_uint (dec)); |
1120 | case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec)); |
992 | case 1: // negative int |
1121 | case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0); |
993 | return newSViv (-1 - (IV)decode_uint (dec)); |
1122 | case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1); |
994 | case 2: // octet string |
1123 | case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec); |
995 | return decode_str (dec, 0); |
1124 | case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec); |
996 | case 3: // utf-8 string |
1125 | case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec); |
997 | return decode_str (dec, 1); |
1126 | |
998 | case 4: // array |
1127 | case MAJOR_MISC >> MAJOR_SHIFT: |
999 | return decode_av (dec); |
|
|
1000 | case 5: // map |
|
|
1001 | return decode_hv (dec); |
|
|
1002 | case 6: // tag |
|
|
1003 | return decode_tagged (dec); |
|
|
1004 | case 7: // misc |
|
|
1005 | switch (*dec->cur++ & 31) |
1128 | switch (*dec->cur++ & MINOR_MASK) |
1006 | { |
1129 | { |
1007 | case 20: |
1130 | case SIMPLE_FALSE: |
1008 | #if CBOR_SLOW |
1131 | #if CBOR_SLOW |
1009 | types_false = get_bool ("Types::Serialiser::false"); |
1132 | types_false = get_bool ("Types::Serialiser::false"); |
1010 | #endif |
1133 | #endif |
1011 | return newSVsv (types_false); |
1134 | return newSVsv (types_false); |
1012 | case 21: |
1135 | case SIMPLE_TRUE: |
1013 | #if CBOR_SLOW |
1136 | #if CBOR_SLOW |
1014 | types_true = get_bool ("Types::Serialiser::true"); |
1137 | types_true = get_bool ("Types::Serialiser::true"); |
1015 | #endif |
1138 | #endif |
1016 | return newSVsv (types_true); |
1139 | return newSVsv (types_true); |
1017 | case 22: |
1140 | case SIMPLE_NULL: |
1018 | return newSVsv (&PL_sv_undef); |
1141 | return newSVsv (&PL_sv_undef); |
1019 | case 23: |
1142 | case SIMPLE_UNDEF: |
1020 | #if CBOR_SLOW |
1143 | #if CBOR_SLOW |
1021 | types_error = get_bool ("Types::Serialiser::error"); |
1144 | types_error = get_bool ("Types::Serialiser::error"); |
1022 | #endif |
1145 | #endif |
1023 | return newSVsv (types_error); |
1146 | return newSVsv (types_error); |
1024 | |
1147 | |
1025 | case 25: |
1148 | case MISC_FLOAT16: |
1026 | { |
1149 | { |
1027 | WANT (2); |
1150 | WANT (2); |
1028 | |
1151 | |
1029 | uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; |
1152 | uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; |
1030 | dec->cur += 2; |
1153 | dec->cur += 2; |
1031 | |
1154 | |
1032 | return newSVnv (ecb_binary16_to_float (fp)); |
1155 | return newSVnv (ecb_binary16_to_float (fp)); |
1033 | } |
1156 | } |
1034 | |
1157 | |
1035 | case 26: |
1158 | case MISC_FLOAT32: |
1036 | { |
1159 | { |
1037 | uint32_t fp; |
1160 | uint32_t fp; |
1038 | WANT (4); |
1161 | WANT (4); |
1039 | memcpy (&fp, dec->cur, 4); |
1162 | memcpy (&fp, dec->cur, 4); |
1040 | dec->cur += 4; |
1163 | dec->cur += 4; |
… | |
… | |
1043 | fp = ecb_bswap32 (fp); |
1166 | fp = ecb_bswap32 (fp); |
1044 | |
1167 | |
1045 | return newSVnv (ecb_binary32_to_float (fp)); |
1168 | return newSVnv (ecb_binary32_to_float (fp)); |
1046 | } |
1169 | } |
1047 | |
1170 | |
1048 | case 27: |
1171 | case MISC_FLOAT64: |
1049 | { |
1172 | { |
1050 | uint64_t fp; |
1173 | uint64_t fp; |
1051 | WANT (8); |
1174 | WANT (8); |
1052 | memcpy (&fp, dec->cur, 8); |
1175 | memcpy (&fp, dec->cur, 8); |
1053 | dec->cur += 8; |
1176 | dec->cur += 8; |
… | |
… | |
1056 | fp = ecb_bswap64 (fp); |
1179 | fp = ecb_bswap64 (fp); |
1057 | |
1180 | |
1058 | return newSVnv (ecb_binary64_to_double (fp)); |
1181 | return newSVnv (ecb_binary64_to_double (fp)); |
1059 | } |
1182 | } |
1060 | |
1183 | |
1061 | // 0..19 unassigned |
1184 | // 0..19 unassigned simple |
1062 | // 24 reserved + unassigned (reserved values are not encodable) |
1185 | // 24 reserved + unassigned simple (reserved values are not encodable) |
|
|
1186 | // 28-30 unassigned misc |
|
|
1187 | // 31 break code |
1063 | default: |
1188 | default: |
1064 | ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); |
1189 | ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)"); |
1065 | } |
1190 | } |
1066 | |
1191 | |
1067 | break; |
1192 | break; |
1068 | } |
1193 | } |
1069 | |
1194 | |
… | |
… | |
1072 | } |
1197 | } |
1073 | |
1198 | |
1074 | static SV * |
1199 | static SV * |
1075 | decode_cbor (SV *string, CBOR *cbor, char **offset_return) |
1200 | decode_cbor (SV *string, CBOR *cbor, char **offset_return) |
1076 | { |
1201 | { |
1077 | dec_t dec = { }; |
1202 | dec_t dec = { 0 }; |
1078 | SV *sv; |
1203 | SV *sv; |
1079 | STRLEN len; |
1204 | STRLEN len; |
1080 | char *data = SvPVbyte (string, len); |
1205 | char *data = SvPVbyte (string, len); |
1081 | |
1206 | |
1082 | if (len > cbor->max_size && cbor->max_size) |
1207 | if (len > cbor->max_size && cbor->max_size) |
… | |
… | |
1096 | if (dec.cur != dec.end && !dec.err) |
1221 | if (dec.cur != dec.end && !dec.err) |
1097 | dec.err = "garbage after CBOR object"; |
1222 | dec.err = "garbage after CBOR object"; |
1098 | |
1223 | |
1099 | if (dec.err) |
1224 | if (dec.err) |
1100 | { |
1225 | { |
|
|
1226 | if (dec.shareable) |
|
|
1227 | { |
|
|
1228 | // need to break cyclic links, which whould all be in shareable |
|
|
1229 | int i; |
|
|
1230 | SV **svp; |
|
|
1231 | |
|
|
1232 | for (i = av_len (dec.shareable) + 1; i--; ) |
|
|
1233 | if ((svp = av_fetch (dec.shareable, i, 0))) |
|
|
1234 | sv_setsv (*svp, &PL_sv_undef); |
|
|
1235 | } |
|
|
1236 | |
1101 | SvREFCNT_dec (sv); |
1237 | SvREFCNT_dec (sv); |
1102 | croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur); |
1238 | croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur); |
1103 | } |
1239 | } |
1104 | |
1240 | |
1105 | sv = sv_2mortal (sv); |
1241 | sv = sv_2mortal (sv); |
1106 | |
1242 | |
1107 | return sv; |
1243 | return sv; |
1108 | } |
1244 | } |
1109 | |
1245 | |
|
|
1246 | ///////////////////////////////////////////////////////////////////////////// |
|
|
1247 | // incremental parser |
|
|
1248 | |
|
|
1249 | #define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0) |
|
|
1250 | |
|
|
1251 | // returns 0 for notyet, 1 for success or error |
|
|
1252 | static int |
|
|
1253 | incr_parse (CBOR *self, SV *cborstr) |
|
|
1254 | { |
|
|
1255 | STRLEN cur; |
|
|
1256 | SvPV (cborstr, cur); |
|
|
1257 | |
|
|
1258 | while (ecb_expect_true (self->incr_need <= cur)) |
|
|
1259 | { |
|
|
1260 | // table of integer count bytes |
|
|
1261 | static I8 incr_len[MINOR_MASK + 1] = { |
|
|
1262 | 0, 0, 0, 0, 0, 0, 0, 0, |
|
|
1263 | 0, 0, 0, 0, 0, 0, 0, 0, |
|
|
1264 | 0, 0, 0, 0, 0, 0, 0, 0, |
|
|
1265 | 1, 2, 4, 8,-1,-1,-1,-2 |
|
|
1266 | }; |
|
|
1267 | |
|
|
1268 | const U8 *p = SvPVX (cborstr) + self->incr_pos; |
|
|
1269 | U8 m = *p & MINOR_MASK; |
|
|
1270 | IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]); |
|
|
1271 | I8 ilen = incr_len[m]; |
|
|
1272 | |
|
|
1273 | self->incr_need = self->incr_pos + 1; |
|
|
1274 | |
|
|
1275 | if (ecb_expect_false (ilen < 0)) |
|
|
1276 | { |
|
|
1277 | if (m != MINOR_INDEF) |
|
|
1278 | return 1; // error |
|
|
1279 | |
|
|
1280 | if (*p == (MAJOR_MISC | MINOR_INDEF)) |
|
|
1281 | { |
|
|
1282 | if (count >= 0) |
|
|
1283 | return 1; // error |
|
|
1284 | |
|
|
1285 | count = 1; |
|
|
1286 | } |
|
|
1287 | else |
|
|
1288 | { |
|
|
1289 | av_push (self->incr_count, newSViv (-1)); //TODO: nest |
|
|
1290 | count = -1; |
|
|
1291 | } |
|
|
1292 | } |
|
|
1293 | else |
|
|
1294 | { |
|
|
1295 | self->incr_need += ilen; |
|
|
1296 | if (ecb_expect_false (self->incr_need > cur)) |
|
|
1297 | return 0; |
|
|
1298 | |
|
|
1299 | int major = *p >> MAJOR_SHIFT; |
|
|
1300 | |
|
|
1301 | switch (major) |
|
|
1302 | { |
|
|
1303 | case MAJOR_TAG >> MAJOR_SHIFT: |
|
|
1304 | ++count; // tags merely prefix another value |
|
|
1305 | break; |
|
|
1306 | |
|
|
1307 | case MAJOR_BYTES >> MAJOR_SHIFT: |
|
|
1308 | case MAJOR_TEXT >> MAJOR_SHIFT: |
|
|
1309 | case MAJOR_ARRAY >> MAJOR_SHIFT: |
|
|
1310 | case MAJOR_MAP >> MAJOR_SHIFT: |
|
|
1311 | { |
|
|
1312 | UV len; |
|
|
1313 | |
|
|
1314 | if (ecb_expect_false (ilen)) |
|
|
1315 | { |
|
|
1316 | len = 0; |
|
|
1317 | |
|
|
1318 | do { |
|
|
1319 | len = (len << 8) | *++p; |
|
|
1320 | } while (--ilen); |
|
|
1321 | } |
|
|
1322 | else |
|
|
1323 | len = m; |
|
|
1324 | |
|
|
1325 | switch (major) |
|
|
1326 | { |
|
|
1327 | case MAJOR_BYTES >> MAJOR_SHIFT: |
|
|
1328 | case MAJOR_TEXT >> MAJOR_SHIFT: |
|
|
1329 | self->incr_need += len; |
|
|
1330 | if (ecb_expect_false (self->incr_need > cur)) |
|
|
1331 | return 0; |
|
|
1332 | |
|
|
1333 | break; |
|
|
1334 | |
|
|
1335 | case MAJOR_MAP >> MAJOR_SHIFT: |
|
|
1336 | len <<= 1; |
|
|
1337 | case MAJOR_ARRAY >> MAJOR_SHIFT: |
|
|
1338 | if (len) |
|
|
1339 | { |
|
|
1340 | av_push (self->incr_count, newSViv (len + 1)); //TODO: nest |
|
|
1341 | count = len + 1; |
|
|
1342 | } |
|
|
1343 | break; |
|
|
1344 | } |
|
|
1345 | } |
|
|
1346 | } |
|
|
1347 | } |
|
|
1348 | |
|
|
1349 | self->incr_pos = self->incr_need; |
|
|
1350 | |
|
|
1351 | if (count > 0) |
|
|
1352 | { |
|
|
1353 | while (!--count) |
|
|
1354 | { |
|
|
1355 | if (!AvFILLp (self->incr_count)) |
|
|
1356 | return 1; // done |
|
|
1357 | |
|
|
1358 | SvREFCNT_dec_NN (av_pop (self->incr_count)); |
|
|
1359 | count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]); |
|
|
1360 | } |
|
|
1361 | |
|
|
1362 | SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count; |
|
|
1363 | } |
|
|
1364 | } |
|
|
1365 | |
|
|
1366 | return 0; |
|
|
1367 | } |
|
|
1368 | |
|
|
1369 | |
1110 | ///////////////////////////////////////////////////////////////////////////// |
1370 | ///////////////////////////////////////////////////////////////////////////// |
1111 | // XS interface functions |
1371 | // XS interface functions |
1112 | |
1372 | |
1113 | MODULE = CBOR::XS PACKAGE = CBOR::XS |
1373 | MODULE = CBOR::XS PACKAGE = CBOR::XS |
1114 | |
1374 | |
… | |
… | |
1154 | void shrink (CBOR *self, int enable = 1) |
1414 | void shrink (CBOR *self, int enable = 1) |
1155 | ALIAS: |
1415 | ALIAS: |
1156 | shrink = F_SHRINK |
1416 | shrink = F_SHRINK |
1157 | allow_unknown = F_ALLOW_UNKNOWN |
1417 | allow_unknown = F_ALLOW_UNKNOWN |
1158 | allow_sharing = F_ALLOW_SHARING |
1418 | allow_sharing = F_ALLOW_SHARING |
|
|
1419 | allow_cycles = F_ALLOW_CYCLES |
1159 | pack_strings = F_PACK_STRINGS |
1420 | pack_strings = F_PACK_STRINGS |
|
|
1421 | utf8_strings = F_UTF8_STRINGS |
|
|
1422 | validate_utf8 = F_VALIDATE_UTF8 |
1160 | PPCODE: |
1423 | PPCODE: |
1161 | { |
1424 | { |
1162 | if (enable) |
1425 | if (enable) |
1163 | self->flags |= ix; |
1426 | self->flags |= ix; |
1164 | else |
1427 | else |
… | |
… | |
1170 | void get_shrink (CBOR *self) |
1433 | void get_shrink (CBOR *self) |
1171 | ALIAS: |
1434 | ALIAS: |
1172 | get_shrink = F_SHRINK |
1435 | get_shrink = F_SHRINK |
1173 | get_allow_unknown = F_ALLOW_UNKNOWN |
1436 | get_allow_unknown = F_ALLOW_UNKNOWN |
1174 | get_allow_sharing = F_ALLOW_SHARING |
1437 | get_allow_sharing = F_ALLOW_SHARING |
|
|
1438 | get_allow_cycles = F_ALLOW_CYCLES |
1175 | get_pack_strings = F_PACK_STRINGS |
1439 | get_pack_strings = F_PACK_STRINGS |
|
|
1440 | get_validate_utf8 = F_VALIDATE_UTF8 |
1176 | PPCODE: |
1441 | PPCODE: |
1177 | XPUSHs (boolSV (self->flags & ix)); |
1442 | XPUSHs (boolSV (self->flags & ix)); |
1178 | |
1443 | |
1179 | void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) |
1444 | void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) |
1180 | PPCODE: |
1445 | PPCODE: |
… | |
… | |
1229 | EXTEND (SP, 2); |
1494 | EXTEND (SP, 2); |
1230 | PUSHs (sv); |
1495 | PUSHs (sv); |
1231 | PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); |
1496 | PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); |
1232 | } |
1497 | } |
1233 | |
1498 | |
|
|
1499 | void incr_parse (CBOR *self, SV *cborstr) |
|
|
1500 | ALIAS: |
|
|
1501 | incr_parse_multiple = 1 |
|
|
1502 | PPCODE: |
|
|
1503 | { |
|
|
1504 | if (SvUTF8 (cborstr)) |
|
|
1505 | sv_utf8_downgrade (cborstr, 0); |
|
|
1506 | |
|
|
1507 | if (!self->incr_count) |
|
|
1508 | { |
|
|
1509 | self->incr_count = newAV (); |
|
|
1510 | self->incr_pos = 0; |
|
|
1511 | self->incr_need = 1; |
|
|
1512 | |
|
|
1513 | av_push (self->incr_count, newSViv (1)); |
|
|
1514 | } |
|
|
1515 | |
|
|
1516 | do |
|
|
1517 | { |
|
|
1518 | if (!incr_parse (self, cborstr)) |
|
|
1519 | { |
|
|
1520 | if (self->incr_need > self->max_size && self->max_size) |
|
|
1521 | croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu", |
|
|
1522 | (unsigned long)self->incr_need, (unsigned long)self->max_size); |
|
|
1523 | |
|
|
1524 | break; |
|
|
1525 | } |
|
|
1526 | |
|
|
1527 | SV *sv; |
|
|
1528 | char *offset; |
|
|
1529 | |
|
|
1530 | PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN; |
|
|
1531 | XPUSHs (sv); |
|
|
1532 | |
|
|
1533 | sv_chop (cborstr, offset); |
|
|
1534 | |
|
|
1535 | av_clear (self->incr_count); |
|
|
1536 | av_push (self->incr_count, newSViv (1)); |
|
|
1537 | |
|
|
1538 | self->incr_pos = 0; |
|
|
1539 | self->incr_need = self->incr_pos + 1; |
|
|
1540 | } |
|
|
1541 | while (ix); |
|
|
1542 | } |
|
|
1543 | |
|
|
1544 | void incr_reset (CBOR *self) |
|
|
1545 | CODE: |
|
|
1546 | { |
|
|
1547 | SvREFCNT_dec (self->incr_count); |
|
|
1548 | self->incr_count = 0; |
|
|
1549 | } |
|
|
1550 | |
1234 | void DESTROY (CBOR *self) |
1551 | void DESTROY (CBOR *self) |
1235 | PPCODE: |
1552 | PPCODE: |
1236 | cbor_free (self); |
1553 | cbor_free (self); |
1237 | |
1554 | |
1238 | PROTOTYPES: ENABLE |
1555 | PROTOTYPES: ENABLE |
1239 | |
1556 | |
1240 | void encode_cbor (SV *scalar) |
1557 | void encode_cbor (SV *scalar) |
|
|
1558 | ALIAS: |
|
|
1559 | encode_cbor = 0 |
|
|
1560 | encode_cbor_sharing = F_ALLOW_SHARING |
1241 | PPCODE: |
1561 | PPCODE: |
1242 | { |
1562 | { |
1243 | CBOR cbor; |
1563 | CBOR cbor; |
1244 | cbor_init (&cbor); |
1564 | cbor_init (&cbor); |
|
|
1565 | cbor.flags |= ix; |
1245 | PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; |
1566 | PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; |
1246 | XPUSHs (scalar); |
1567 | XPUSHs (scalar); |
1247 | } |
1568 | } |
1248 | |
1569 | |
1249 | void decode_cbor (SV *cborstr) |
1570 | void decode_cbor (SV *cborstr) |