… | |
… | |
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 //TODO |
102 | #define F_ALLOW_SHARING 0x00000004UL |
65 | #define F_ALLOW_STRINGREF 0x00000008UL //TODO |
103 | #define F_ALLOW_CYCLES 0x00000008UL |
|
|
104 | #define F_PACK_STRINGS 0x00000010UL |
|
|
105 | #define F_TEXT_KEYS 0x00000020UL |
|
|
106 | #define F_TEXT_STRINGS 0x00000040UL |
|
|
107 | #define F_VALIDATE_UTF8 0x00000080UL |
66 | |
108 | |
67 | #define INIT_SIZE 32 // initial scalar size to be allocated |
109 | #define INIT_SIZE 32 // initial scalar size to be allocated |
68 | |
110 | |
69 | #define SB do { |
111 | #define SB do { |
70 | #define SE } while (0) |
112 | #define SE } while (0) |
… | |
… | |
89 | typedef struct { |
131 | typedef struct { |
90 | U32 flags; |
132 | U32 flags; |
91 | U32 max_depth; |
133 | U32 max_depth; |
92 | STRLEN max_size; |
134 | STRLEN max_size; |
93 | SV *filter; |
135 | SV *filter; |
|
|
136 | |
|
|
137 | // for the incremental parser |
|
|
138 | STRLEN incr_pos; // the current offset into the text |
|
|
139 | STRLEN incr_need; // minimum bytes needed to decode |
|
|
140 | AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef. |
94 | } CBOR; |
141 | } CBOR; |
95 | |
142 | |
96 | ecb_inline void |
143 | ecb_inline void |
97 | cbor_init (CBOR *cbor) |
144 | cbor_init (CBOR *cbor) |
98 | { |
145 | { |
… | |
… | |
102 | |
149 | |
103 | ecb_inline void |
150 | ecb_inline void |
104 | cbor_free (CBOR *cbor) |
151 | cbor_free (CBOR *cbor) |
105 | { |
152 | { |
106 | SvREFCNT_dec (cbor->filter); |
153 | SvREFCNT_dec (cbor->filter); |
|
|
154 | SvREFCNT_dec (cbor->incr_count); |
107 | } |
155 | } |
108 | |
156 | |
109 | ///////////////////////////////////////////////////////////////////////////// |
157 | ///////////////////////////////////////////////////////////////////////////// |
110 | // utility functions |
158 | // utility functions |
111 | |
159 | |
… | |
… | |
189 | static void |
237 | static void |
190 | encode_uint (enc_t *enc, int major, UV len) |
238 | encode_uint (enc_t *enc, int major, UV len) |
191 | { |
239 | { |
192 | need (enc, 9); |
240 | need (enc, 9); |
193 | |
241 | |
194 | if (len < 24) |
242 | if (ecb_expect_true (len < LENGTH_EXT1)) |
195 | *enc->cur++ = major | len; |
243 | *enc->cur++ = major | len; |
196 | else if (len <= 0xff) |
244 | else if (ecb_expect_true (len <= 0xffU)) |
197 | { |
245 | { |
198 | *enc->cur++ = major | 24; |
246 | *enc->cur++ = major | LENGTH_EXT1; |
199 | *enc->cur++ = len; |
247 | *enc->cur++ = len; |
200 | } |
248 | } |
201 | else if (len <= 0xffff) |
249 | else if (len <= 0xffffU) |
202 | { |
250 | { |
203 | *enc->cur++ = major | 25; |
251 | *enc->cur++ = major | LENGTH_EXT2; |
204 | *enc->cur++ = len >> 8; |
252 | *enc->cur++ = len >> 8; |
205 | *enc->cur++ = len; |
253 | *enc->cur++ = len; |
206 | } |
254 | } |
207 | else if (len <= 0xffffffff) |
255 | else if (len <= 0xffffffffU) |
208 | { |
256 | { |
209 | *enc->cur++ = major | 26; |
257 | *enc->cur++ = major | LENGTH_EXT4; |
210 | *enc->cur++ = len >> 24; |
258 | *enc->cur++ = len >> 24; |
211 | *enc->cur++ = len >> 16; |
259 | *enc->cur++ = len >> 16; |
212 | *enc->cur++ = len >> 8; |
260 | *enc->cur++ = len >> 8; |
213 | *enc->cur++ = len; |
261 | *enc->cur++ = len; |
214 | } |
262 | } |
215 | else |
263 | else |
216 | { |
264 | { |
217 | *enc->cur++ = major | 27; |
265 | *enc->cur++ = major | LENGTH_EXT8; |
218 | *enc->cur++ = len >> 56; |
266 | *enc->cur++ = len >> 56; |
219 | *enc->cur++ = len >> 48; |
267 | *enc->cur++ = len >> 48; |
220 | *enc->cur++ = len >> 40; |
268 | *enc->cur++ = len >> 40; |
221 | *enc->cur++ = len >> 32; |
269 | *enc->cur++ = len >> 32; |
222 | *enc->cur++ = len >> 24; |
270 | *enc->cur++ = len >> 24; |
… | |
… | |
227 | } |
275 | } |
228 | |
276 | |
229 | ecb_inline void |
277 | ecb_inline void |
230 | encode_tag (enc_t *enc, UV tag) |
278 | encode_tag (enc_t *enc, UV tag) |
231 | { |
279 | { |
232 | encode_uint (enc, 0xc0, tag); |
280 | encode_uint (enc, MAJOR_TAG, tag); |
233 | } |
281 | } |
234 | |
282 | |
235 | static void |
283 | // exceptional (hopefully) slow path for byte strings that need to be utf8-encoded |
|
|
284 | ecb_noinline static void |
236 | encode_str (enc_t *enc, int utf8, char *str, STRLEN len) |
285 | encode_str_utf8 (enc_t *enc, int utf8, char *str, STRLEN len) |
237 | { |
286 | { |
|
|
287 | STRLEN ulen = len; |
|
|
288 | U8 *p, *pend = (U8 *)str + len; |
|
|
289 | |
|
|
290 | for (p = (U8 *)str; p < pend; ++p) |
|
|
291 | ulen += *p >> 7; // count set high bits |
|
|
292 | |
|
|
293 | encode_uint (enc, MAJOR_TEXT, ulen); |
|
|
294 | |
|
|
295 | need (enc, ulen); |
|
|
296 | for (p = (U8 *)str; p < pend; ++p) |
|
|
297 | if (*p < 0x80) |
|
|
298 | *enc->cur++ = *p; |
|
|
299 | else |
|
|
300 | { |
|
|
301 | *enc->cur++ = 0xc0 + (*p >> 6); |
|
|
302 | *enc->cur++ = 0x80 + (*p & 63); |
|
|
303 | } |
|
|
304 | } |
|
|
305 | |
|
|
306 | ecb_inline void |
|
|
307 | encode_str (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len) |
|
|
308 | { |
|
|
309 | if (ecb_expect_false (upgrade_utf8)) |
|
|
310 | if (!utf8) |
|
|
311 | { |
|
|
312 | encode_str_utf8 (enc, utf8, str, len); |
|
|
313 | return; |
|
|
314 | } |
|
|
315 | |
|
|
316 | encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len); |
|
|
317 | need (enc, len); |
|
|
318 | memcpy (enc->cur, str, len); |
|
|
319 | enc->cur += len; |
|
|
320 | } |
|
|
321 | |
|
|
322 | ecb_inline void |
|
|
323 | encode_strref (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len) |
|
|
324 | { |
238 | if (ecb_expect_false (enc->cbor.flags & F_ALLOW_STRINGREF)) |
325 | if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS)) |
239 | { |
326 | { |
240 | SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1); |
327 | SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1); |
241 | |
328 | |
242 | if (SvOK (*svp)) |
329 | if (SvOK (*svp)) |
243 | { |
330 | { |
244 | // already registered, use stringref |
331 | // already registered, use stringref |
245 | encode_tag (enc, CBOR_TAG_STRINGREF); |
332 | encode_tag (enc, CBOR_TAG_STRINGREF); |
246 | encode_uint (enc, 0x00, SvUV (*svp)); |
333 | encode_uint (enc, MAJOR_POS_INT, SvUV (*svp)); |
247 | return; |
334 | return; |
248 | } |
335 | } |
249 | else if (len >= minimum_string_length (enc->stringref_idx)) |
336 | else if (len >= minimum_string_length (enc->stringref_idx)) |
250 | { |
337 | { |
251 | // register only |
338 | // register only |
252 | sv_setuv (*svp, enc->stringref_idx); |
339 | sv_setuv (*svp, enc->stringref_idx); |
253 | ++enc->stringref_idx; |
340 | ++enc->stringref_idx; |
254 | } |
341 | } |
255 | } |
342 | } |
256 | |
343 | |
257 | encode_uint (enc, utf8 ? 0x60 : 0x40, len); |
344 | encode_str (enc, upgrade_utf8, utf8, str, len); |
258 | need (enc, len); |
|
|
259 | memcpy (enc->cur, str, len); |
|
|
260 | enc->cur += len; |
|
|
261 | } |
345 | } |
262 | |
346 | |
263 | static void encode_sv (enc_t *enc, SV *sv); |
347 | static void encode_sv (enc_t *enc, SV *sv); |
264 | |
348 | |
265 | static void |
349 | static void |
… | |
… | |
270 | if (enc->depth >= enc->cbor.max_depth) |
354 | if (enc->depth >= enc->cbor.max_depth) |
271 | croak (ERR_NESTING_EXCEEDED); |
355 | croak (ERR_NESTING_EXCEEDED); |
272 | |
356 | |
273 | ++enc->depth; |
357 | ++enc->depth; |
274 | |
358 | |
275 | encode_uint (enc, 0x80, len + 1); |
359 | encode_uint (enc, MAJOR_ARRAY, len + 1); |
276 | |
360 | |
|
|
361 | if (SvMAGICAL (av)) |
277 | for (i = 0; i <= len; ++i) |
362 | for (i = 0; i <= len; ++i) |
278 | { |
363 | { |
279 | SV **svp = av_fetch (av, i, 0); |
364 | SV **svp = av_fetch (av, i, 0); |
280 | encode_sv (enc, svp ? *svp : &PL_sv_undef); |
365 | encode_sv (enc, svp ? *svp : &PL_sv_undef); |
281 | } |
366 | } |
|
|
367 | else |
|
|
368 | for (i = 0; i <= len; ++i) |
|
|
369 | { |
|
|
370 | SV *sv = AvARRAY (av)[i]; |
|
|
371 | encode_sv (enc, sv ? sv : &PL_sv_undef); |
|
|
372 | } |
282 | |
373 | |
283 | --enc->depth; |
374 | --enc->depth; |
284 | } |
375 | } |
285 | |
376 | |
286 | static void |
377 | static void |
… | |
… | |
295 | |
386 | |
296 | int pairs = hv_iterinit (hv); |
387 | int pairs = hv_iterinit (hv); |
297 | int mg = SvMAGICAL (hv); |
388 | int mg = SvMAGICAL (hv); |
298 | |
389 | |
299 | if (mg) |
390 | if (mg) |
300 | encode_ch (enc, 0xa0 | 31); |
391 | encode_ch (enc, MAJOR_MAP | MINOR_INDEF); |
301 | else |
392 | else |
302 | encode_uint (enc, 0xa0, pairs); |
393 | encode_uint (enc, MAJOR_MAP, pairs); |
303 | |
394 | |
304 | while ((he = hv_iternext (hv))) |
395 | while ((he = hv_iternext (hv))) |
305 | { |
396 | { |
306 | if (HeKLEN (he) == HEf_SVKEY) |
397 | if (HeKLEN (he) == HEf_SVKEY) |
307 | encode_sv (enc, HeSVKEY (he)); |
398 | encode_sv (enc, HeSVKEY (he)); |
308 | else |
399 | else |
309 | encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); |
400 | encode_strref (enc, enc->cbor.flags & (F_TEXT_KEYS | F_TEXT_STRINGS), HeKUTF8 (he), HeKEY (he), HeKLEN (he)); |
310 | |
401 | |
311 | encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); |
402 | encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); |
312 | } |
403 | } |
313 | |
404 | |
314 | if (mg) |
405 | if (mg) |
315 | encode_ch (enc, 0xe0 | 31); |
406 | encode_ch (enc, MAJOR_MISC | MINOR_INDEF); |
316 | |
407 | |
317 | --enc->depth; |
408 | --enc->depth; |
318 | } |
409 | } |
319 | |
410 | |
320 | // encode objects, arrays and special \0=false and \1=true values. |
411 | // encode objects, arrays and special \0=false and \1=true values. |
321 | static void |
412 | static void |
322 | encode_rv (enc_t *enc, SV *sv) |
413 | encode_rv (enc_t *enc, SV *sv) |
323 | { |
414 | { |
324 | SvGETMAGIC (sv); |
415 | SvGETMAGIC (sv); |
325 | |
|
|
326 | if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING) |
|
|
327 | && ecb_expect_false (SvREFCNT (sv) > 1)) |
|
|
328 | { |
|
|
329 | if (!enc->shareable) |
|
|
330 | enc->shareable = (HV *)sv_2mortal ((SV *)newHV ()); |
|
|
331 | |
|
|
332 | SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1); |
|
|
333 | |
|
|
334 | if (SvOK (*svp)) |
|
|
335 | { |
|
|
336 | encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF); |
|
|
337 | encode_uint (enc, 0x00, SvUV (*svp)); |
|
|
338 | return; |
|
|
339 | } |
|
|
340 | else |
|
|
341 | { |
|
|
342 | sv_setuv (*svp, enc->shareable_idx); |
|
|
343 | ++enc->shareable_idx; |
|
|
344 | encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE); |
|
|
345 | } |
|
|
346 | } |
|
|
347 | |
416 | |
348 | svtype svt = SvTYPE (sv); |
417 | svtype svt = SvTYPE (sv); |
349 | |
418 | |
350 | if (ecb_expect_false (SvOBJECT (sv))) |
419 | if (ecb_expect_false (SvOBJECT (sv))) |
351 | { |
420 | { |
… | |
… | |
358 | HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash |
427 | HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash |
359 | ? cbor_tagged_stash |
428 | ? cbor_tagged_stash |
360 | : gv_stashpv ("CBOR::XS::Tagged" , 1); |
429 | : gv_stashpv ("CBOR::XS::Tagged" , 1); |
361 | |
430 | |
362 | HV *stash = SvSTASH (sv); |
431 | HV *stash = SvSTASH (sv); |
363 | GV *method; |
|
|
364 | |
432 | |
365 | if (stash == boolean_stash) |
433 | if (stash == boolean_stash) |
366 | encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20); |
434 | { |
|
|
435 | encode_ch (enc, SvIV (sv) ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE); |
|
|
436 | return; |
|
|
437 | } |
367 | else if (stash == error_stash) |
438 | else if (stash == error_stash) |
368 | encode_ch (enc, 0xe0 | 23); |
439 | { |
|
|
440 | encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF); |
|
|
441 | return; |
|
|
442 | } |
369 | else if (stash == tagged_stash) |
443 | else if (stash == tagged_stash) |
370 | { |
444 | { |
371 | if (svt != SVt_PVAV) |
445 | if (svt != SVt_PVAV) |
372 | croak ("encountered CBOR::XS::Tagged object that isn't an array"); |
446 | croak ("encountered CBOR::XS::Tagged object that isn't an array"); |
373 | |
447 | |
374 | encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); |
448 | encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1))); |
375 | encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); |
449 | encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); |
|
|
450 | |
|
|
451 | return; |
|
|
452 | } |
|
|
453 | } |
|
|
454 | |
|
|
455 | if (ecb_expect_false (SvREFCNT (sv) > 1) |
|
|
456 | && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)) |
|
|
457 | { |
|
|
458 | if (!enc->shareable) |
|
|
459 | enc->shareable = (HV *)sv_2mortal ((SV *)newHV ()); |
|
|
460 | |
|
|
461 | SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1); |
|
|
462 | |
|
|
463 | if (SvOK (*svp)) |
376 | } |
464 | { |
|
|
465 | encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF); |
|
|
466 | encode_uint (enc, MAJOR_POS_INT, SvUV (*svp)); |
|
|
467 | return; |
|
|
468 | } |
|
|
469 | else |
|
|
470 | { |
|
|
471 | sv_setuv (*svp, enc->shareable_idx); |
|
|
472 | ++enc->shareable_idx; |
|
|
473 | encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE); |
|
|
474 | } |
|
|
475 | } |
|
|
476 | |
|
|
477 | if (ecb_expect_false (SvOBJECT (sv))) |
|
|
478 | { |
|
|
479 | HV *stash = SvSTASH (sv); |
|
|
480 | GV *method; |
|
|
481 | |
377 | else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) |
482 | if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) |
378 | { |
483 | { |
379 | dSP; |
484 | dSP; |
380 | |
485 | |
381 | ENTER; SAVETMPS; PUSHMARK (SP); |
486 | ENTER; SAVETMPS; |
|
|
487 | PUSHMARK (SP); |
382 | // we re-bless the reference to get overload and other niceties right |
488 | // we re-bless the reference to get overload and other niceties right |
383 | XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); |
489 | XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); |
384 | |
490 | |
385 | PUTBACK; |
491 | PUTBACK; |
386 | // G_SCALAR ensures that return value is 1 |
492 | // G_SCALAR ensures that return value is 1 |
… | |
… | |
399 | } |
505 | } |
400 | else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0) |
506 | else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0) |
401 | { |
507 | { |
402 | dSP; |
508 | dSP; |
403 | |
509 | |
404 | ENTER; SAVETMPS; PUSHMARK (SP); |
510 | ENTER; SAVETMPS; |
|
|
511 | SAVESTACK_POS (); |
|
|
512 | PUSHMARK (SP); |
405 | EXTEND (SP, 2); |
513 | EXTEND (SP, 2); |
406 | // we re-bless the reference to get overload and other niceties right |
514 | // we re-bless the reference to get overload and other niceties right |
407 | PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); |
515 | PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); |
408 | PUSHs (sv_cbor); |
516 | PUSHs (sv_cbor); |
409 | |
517 | |
… | |
… | |
414 | // catch this surprisingly common error |
522 | // catch this surprisingly common error |
415 | if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) |
523 | if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) |
416 | croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash)); |
524 | croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash)); |
417 | |
525 | |
418 | encode_tag (enc, CBOR_TAG_PERL_OBJECT); |
526 | encode_tag (enc, CBOR_TAG_PERL_OBJECT); |
419 | encode_uint (enc, 0x80, count + 1); |
527 | encode_uint (enc, MAJOR_ARRAY, count + 1); |
420 | encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); |
528 | encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); |
421 | |
529 | |
422 | while (count) |
530 | while (count) |
423 | encode_sv (enc, SP[1 - count--]); |
531 | encode_sv (enc, SP[1 - count--]); |
424 | |
532 | |
425 | PUTBACK; |
533 | PUTBACK; |
… | |
… | |
446 | { |
554 | { |
447 | double nv = SvNVX (sv); |
555 | double nv = SvNVX (sv); |
448 | |
556 | |
449 | need (enc, 9); |
557 | need (enc, 9); |
450 | |
558 | |
451 | if (ecb_expect_false (nv == (U32)nv)) |
559 | if (ecb_expect_false (nv == (NV)(U32)nv)) |
452 | encode_uint (enc, 0x00, (U32)nv); |
560 | encode_uint (enc, MAJOR_POS_INT, (U32)nv); |
453 | //TODO: maybe I32? |
561 | //TODO: maybe I32? |
454 | else if (ecb_expect_false (nv == (float)nv)) |
562 | else if (ecb_expect_false (nv == (float)nv)) |
455 | { |
563 | { |
456 | uint32_t fp = ecb_float_to_binary32 (nv); |
564 | uint32_t fp = ecb_float_to_binary32 (nv); |
457 | |
565 | |
458 | *enc->cur++ = 0xe0 | 26; |
566 | *enc->cur++ = MAJOR_MISC | MISC_FLOAT32; |
459 | |
567 | |
460 | if (!ecb_big_endian ()) |
568 | if (!ecb_big_endian ()) |
461 | fp = ecb_bswap32 (fp); |
569 | fp = ecb_bswap32 (fp); |
462 | |
570 | |
463 | memcpy (enc->cur, &fp, 4); |
571 | memcpy (enc->cur, &fp, 4); |
… | |
… | |
465 | } |
573 | } |
466 | else |
574 | else |
467 | { |
575 | { |
468 | uint64_t fp = ecb_double_to_binary64 (nv); |
576 | uint64_t fp = ecb_double_to_binary64 (nv); |
469 | |
577 | |
470 | *enc->cur++ = 0xe0 | 27; |
578 | *enc->cur++ = MAJOR_MISC | MISC_FLOAT64; |
471 | |
579 | |
472 | if (!ecb_big_endian ()) |
580 | if (!ecb_big_endian ()) |
473 | fp = ecb_bswap64 (fp); |
581 | fp = ecb_bswap64 (fp); |
474 | |
582 | |
475 | memcpy (enc->cur, &fp, 8); |
583 | memcpy (enc->cur, &fp, 8); |
… | |
… | |
484 | |
592 | |
485 | if (SvPOKp (sv)) |
593 | if (SvPOKp (sv)) |
486 | { |
594 | { |
487 | STRLEN len; |
595 | STRLEN len; |
488 | char *str = SvPV (sv, len); |
596 | char *str = SvPV (sv, len); |
489 | encode_str (enc, SvUTF8 (sv), str, len); |
597 | encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len); |
490 | } |
598 | } |
491 | else if (SvNOKp (sv)) |
599 | else if (SvNOKp (sv)) |
492 | encode_nv (enc, sv); |
600 | encode_nv (enc, sv); |
493 | else if (SvIOKp (sv)) |
601 | else if (SvIOKp (sv)) |
494 | { |
602 | { |
495 | if (SvIsUV (sv)) |
603 | if (SvIsUV (sv)) |
496 | encode_uint (enc, 0x00, SvUVX (sv)); |
604 | encode_uint (enc, MAJOR_POS_INT, SvUVX (sv)); |
497 | else if (SvIVX (sv) >= 0) |
605 | else if (SvIVX (sv) >= 0) |
498 | encode_uint (enc, 0x00, SvIVX (sv)); |
606 | encode_uint (enc, MAJOR_POS_INT, SvIVX (sv)); |
499 | else |
607 | else |
500 | encode_uint (enc, 0x20, -(SvIVX (sv) + 1)); |
608 | encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1)); |
501 | } |
609 | } |
502 | else if (SvROK (sv)) |
610 | else if (SvROK (sv)) |
503 | encode_rv (enc, SvRV (sv)); |
611 | encode_rv (enc, SvRV (sv)); |
504 | else if (!SvOK (sv)) |
612 | else if (!SvOK (sv)) |
505 | encode_ch (enc, 0xe0 | 22); |
613 | encode_ch (enc, MAJOR_MISC | SIMPLE_NULL); |
506 | else if (enc->cbor.flags & F_ALLOW_UNKNOWN) |
614 | else if (enc->cbor.flags & F_ALLOW_UNKNOWN) |
507 | encode_ch (enc, 0xe0 | 23); |
615 | encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF); |
508 | else |
616 | else |
509 | croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data", |
617 | croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data", |
510 | SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); |
618 | SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); |
511 | } |
619 | } |
512 | |
620 | |
513 | static SV * |
621 | static SV * |
514 | encode_cbor (SV *scalar, CBOR *cbor) |
622 | encode_cbor (SV *scalar, CBOR *cbor) |
515 | { |
623 | { |
516 | enc_t enc = { }; |
624 | enc_t enc = { 0 }; |
517 | |
625 | |
518 | enc.cbor = *cbor; |
626 | enc.cbor = *cbor; |
519 | enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); |
627 | enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); |
520 | enc.cur = SvPVX (enc.sv); |
628 | enc.cur = SvPVX (enc.sv); |
521 | enc.end = SvEND (enc.sv); |
629 | enc.end = SvEND (enc.sv); |
522 | |
630 | |
523 | SvPOK_only (enc.sv); |
631 | SvPOK_only (enc.sv); |
524 | |
632 | |
525 | if (cbor->flags & F_ALLOW_STRINGREF) |
633 | if (cbor->flags & F_PACK_STRINGS) |
526 | { |
634 | { |
527 | encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE); |
635 | encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE); |
528 | enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ()); |
636 | enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ()); |
529 | enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ()); |
637 | enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ()); |
530 | } |
638 | } |
… | |
… | |
553 | U32 depth; // recursion depth |
661 | U32 depth; // recursion depth |
554 | U32 maxdepth; // recursion depth limit |
662 | U32 maxdepth; // recursion depth limit |
555 | AV *shareable; |
663 | AV *shareable; |
556 | AV *stringref; |
664 | AV *stringref; |
557 | SV *decode_tagged; |
665 | SV *decode_tagged; |
|
|
666 | SV *err_sv; // optional sv for error, needs to be freed |
558 | } dec_t; |
667 | } dec_t; |
559 | |
668 | |
560 | #define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE |
669 | // set dec->err to ERRSV |
|
|
670 | ecb_cold static void |
|
|
671 | err_errsv (dec_t *dec) |
|
|
672 | { |
|
|
673 | if (!dec->err) |
|
|
674 | { |
|
|
675 | dec->err_sv = newSVsv (ERRSV); |
561 | |
676 | |
562 | #define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data") |
677 | // chop off the trailing \n |
|
|
678 | SvCUR_set (dec->err_sv, SvCUR (dec->err_sv) - 1); |
|
|
679 | *SvEND (dec->err_sv) = 0; |
563 | |
680 | |
|
|
681 | dec->err = SvPVutf8_nolen (dec->err_sv); |
|
|
682 | } |
|
|
683 | } |
|
|
684 | |
|
|
685 | // the following functions are used to reduce code size and help the compiler to optimise |
|
|
686 | ecb_cold static void |
|
|
687 | err_set (dec_t *dec, const char *reason) |
|
|
688 | { |
|
|
689 | if (!dec->err) |
|
|
690 | dec->err = reason; |
|
|
691 | } |
|
|
692 | |
|
|
693 | ecb_cold static void |
|
|
694 | err_unexpected_end (dec_t *dec) |
|
|
695 | { |
|
|
696 | err_set (dec, "unexpected end of CBOR data"); |
|
|
697 | } |
|
|
698 | |
|
|
699 | ecb_cold static void |
|
|
700 | err_nesting_exceeded (dec_t *dec) |
|
|
701 | { |
|
|
702 | err_set (dec, ERR_NESTING_EXCEEDED); |
|
|
703 | } |
|
|
704 | |
|
|
705 | #define ERR_DO(do) SB do; goto fail; SE |
|
|
706 | #define ERR(reason) ERR_DO (err_set (dec, reason)) |
|
|
707 | #define ERR_ERRSV ERR_DO (err_errsv (dec)) |
|
|
708 | |
|
|
709 | #define WANT(len) if (ecb_expect_false ((UV)(dec->end - dec->cur) < (UV)len)) ERR_DO (err_unexpected_end (dec)) |
|
|
710 | |
564 | #define DEC_INC_DEPTH if (++dec->depth > dec->cbor.max_depth) ERR (ERR_NESTING_EXCEEDED) |
711 | #define DEC_INC_DEPTH if (ecb_expect_false (++dec->depth > dec->cbor.max_depth)) ERR (ERR_NESTING_EXCEEDED) |
565 | #define DEC_DEC_DEPTH --dec->depth |
712 | #define DEC_DEC_DEPTH --dec->depth |
566 | |
713 | |
567 | static UV |
714 | static UV |
568 | decode_uint (dec_t *dec) |
715 | decode_uint (dec_t *dec) |
569 | { |
716 | { |
570 | switch (*dec->cur & 31) |
717 | U8 m = *dec->cur & MINOR_MASK; |
571 | { |
718 | ++dec->cur; |
572 | case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7: |
|
|
573 | case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15: |
|
|
574 | case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23: |
|
|
575 | return *dec->cur++ & 31; |
|
|
576 | |
719 | |
577 | case 24: |
720 | if (ecb_expect_true (m < LENGTH_EXT1)) |
|
|
721 | return m; |
|
|
722 | else if (ecb_expect_true (m == LENGTH_EXT1)) |
|
|
723 | { |
578 | WANT (2); |
724 | WANT (1); |
579 | dec->cur += 2; |
725 | dec->cur += 1; |
580 | return dec->cur[-1]; |
726 | return dec->cur[-1]; |
581 | |
727 | } |
582 | case 25: |
728 | else if (ecb_expect_true (m == LENGTH_EXT2)) |
|
|
729 | { |
583 | WANT (3); |
730 | WANT (2); |
584 | dec->cur += 3; |
731 | dec->cur += 2; |
585 | return (((UV)dec->cur[-2]) << 8) |
732 | return (((UV)dec->cur[-2]) << 8) |
586 | | ((UV)dec->cur[-1]); |
733 | | ((UV)dec->cur[-1]); |
587 | |
734 | } |
588 | case 26: |
735 | else if (ecb_expect_true (m == LENGTH_EXT4)) |
|
|
736 | { |
589 | WANT (5); |
737 | WANT (4); |
590 | dec->cur += 5; |
738 | dec->cur += 4; |
591 | return (((UV)dec->cur[-4]) << 24) |
739 | return (((UV)dec->cur[-4]) << 24) |
592 | | (((UV)dec->cur[-3]) << 16) |
740 | | (((UV)dec->cur[-3]) << 16) |
593 | | (((UV)dec->cur[-2]) << 8) |
741 | | (((UV)dec->cur[-2]) << 8) |
594 | | ((UV)dec->cur[-1]); |
742 | | ((UV)dec->cur[-1]); |
595 | |
743 | } |
596 | case 27: |
744 | else if (ecb_expect_true (m == LENGTH_EXT8)) |
|
|
745 | { |
597 | WANT (9); |
746 | WANT (8); |
598 | dec->cur += 9; |
747 | dec->cur += 8; |
|
|
748 | |
|
|
749 | return |
|
|
750 | #if UVSIZE < 8 |
|
|
751 | 0 |
|
|
752 | #else |
599 | return (((UV)dec->cur[-8]) << 56) |
753 | (((UV)dec->cur[-8]) << 56) |
600 | | (((UV)dec->cur[-7]) << 48) |
754 | | (((UV)dec->cur[-7]) << 48) |
601 | | (((UV)dec->cur[-6]) << 40) |
755 | | (((UV)dec->cur[-6]) << 40) |
602 | | (((UV)dec->cur[-5]) << 32) |
756 | | (((UV)dec->cur[-5]) << 32) |
|
|
757 | #endif |
603 | | (((UV)dec->cur[-4]) << 24) |
758 | | (((UV)dec->cur[-4]) << 24) |
604 | | (((UV)dec->cur[-3]) << 16) |
759 | | (((UV)dec->cur[-3]) << 16) |
605 | | (((UV)dec->cur[-2]) << 8) |
760 | | (((UV)dec->cur[-2]) << 8) |
606 | | ((UV)dec->cur[-1]); |
761 | | ((UV)dec->cur[-1]); |
607 | |
762 | } |
608 | default: |
763 | else |
609 | ERR ("corrupted CBOR data (unsupported integer minor encoding)"); |
764 | ERR ("corrupted CBOR data (unsupported integer minor encoding)"); |
610 | } |
|
|
611 | |
765 | |
612 | fail: |
766 | fail: |
613 | return 0; |
767 | return 0; |
614 | } |
768 | } |
615 | |
769 | |
… | |
… | |
620 | { |
774 | { |
621 | AV *av = newAV (); |
775 | AV *av = newAV (); |
622 | |
776 | |
623 | DEC_INC_DEPTH; |
777 | DEC_INC_DEPTH; |
624 | |
778 | |
625 | if ((*dec->cur & 31) == 31) |
779 | if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF)) |
626 | { |
780 | { |
627 | ++dec->cur; |
781 | ++dec->cur; |
628 | |
782 | |
629 | for (;;) |
783 | for (;;) |
630 | { |
784 | { |
631 | WANT (1); |
785 | WANT (1); |
632 | |
786 | |
633 | if (*dec->cur == (0xe0 | 31)) |
787 | if (*dec->cur == (MAJOR_MISC | MINOR_INDEF)) |
634 | { |
788 | { |
635 | ++dec->cur; |
789 | ++dec->cur; |
636 | break; |
790 | break; |
637 | } |
791 | } |
638 | |
792 | |
639 | av_push (av, decode_sv (dec)); |
793 | av_push (av, decode_sv (dec)); |
640 | } |
794 | } |
641 | } |
795 | } |
642 | else |
796 | else |
643 | { |
797 | { |
644 | int i, len = decode_uint (dec); |
798 | UV i, len = decode_uint (dec); |
645 | |
799 | |
|
|
800 | WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays |
646 | av_fill (av, len - 1); |
801 | av_fill (av, len - 1); |
647 | |
802 | |
648 | for (i = 0; i < len; ++i) |
803 | for (i = 0; i < len; ++i) |
649 | AvARRAY (av)[i] = decode_sv (dec); |
804 | AvARRAY (av)[i] = decode_sv (dec); |
650 | } |
805 | } |
… | |
… | |
663 | { |
818 | { |
664 | // for speed reasons, we specialcase single-string |
819 | // for speed reasons, we specialcase single-string |
665 | // byte or utf-8 strings as keys, but only when !stringref |
820 | // byte or utf-8 strings as keys, but only when !stringref |
666 | |
821 | |
667 | if (ecb_expect_true (!dec->stringref)) |
822 | if (ecb_expect_true (!dec->stringref)) |
668 | if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27) |
823 | if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8)) |
669 | { |
824 | { |
670 | I32 len = decode_uint (dec); |
825 | STRLEN len = decode_uint (dec); |
671 | char *key = (char *)dec->cur; |
826 | char *key = (char *)dec->cur; |
672 | |
827 | |
|
|
828 | WANT (len); |
673 | dec->cur += len; |
829 | dec->cur += len; |
674 | |
830 | |
675 | if (ecb_expect_false (dec->stringref)) |
|
|
676 | av_push (dec->stringref, newSVpvn (key, len)); |
|
|
677 | |
|
|
678 | hv_store (hv, key, len, decode_sv (dec), 0); |
831 | hv_store (hv, key, len, decode_sv (dec), 0); |
679 | |
832 | |
680 | return; |
833 | return; |
681 | } |
834 | } |
682 | else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27) |
835 | else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8)) |
683 | { |
836 | { |
684 | I32 len = decode_uint (dec); |
837 | STRLEN len = decode_uint (dec); |
685 | char *key = (char *)dec->cur; |
838 | char *key = (char *)dec->cur; |
686 | |
839 | |
|
|
840 | WANT (len); |
687 | dec->cur += len; |
841 | dec->cur += len; |
688 | |
842 | |
689 | if (ecb_expect_false (dec->stringref)) |
843 | if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8)) |
690 | av_push (dec->stringref, newSVpvn_utf8 (key, len, 1)); |
844 | if (!is_utf8_string (key, len)) |
|
|
845 | ERR ("corrupted CBOR data (invalid UTF-8 in map key)"); |
691 | |
846 | |
692 | hv_store (hv, key, -len, decode_sv (dec), 0); |
847 | hv_store (hv, key, -len, decode_sv (dec), 0); |
693 | |
848 | |
694 | return; |
849 | return; |
695 | } |
850 | } |
696 | |
851 | |
697 | SV *k = decode_sv (dec); |
852 | SV *k = decode_sv (dec); |
698 | SV *v = decode_sv (dec); |
853 | SV *v = decode_sv (dec); |
699 | |
854 | |
|
|
855 | // we leak memory if uncaught exceptions are thrown by random magical |
|
|
856 | // methods, and this is hopefully the only place where it can happen, |
|
|
857 | // so if there is a chance of an exception, take the very slow path. |
|
|
858 | // since catching exceptions is "undocumented/internal/forbidden" by |
|
|
859 | // the new p5p powers, we need to call out to a perl function :/ |
|
|
860 | if (ecb_expect_false (SvAMAGIC (k))) |
|
|
861 | { |
|
|
862 | dSP; |
|
|
863 | |
|
|
864 | ENTER; SAVETMPS; |
|
|
865 | PUSHMARK (SP); |
|
|
866 | EXTEND (SP, 3); |
|
|
867 | PUSHs (sv_2mortal (newRV_inc ((SV *)hv))); |
|
|
868 | PUSHs (sv_2mortal (k)); |
|
|
869 | PUSHs (sv_2mortal (v)); |
|
|
870 | |
|
|
871 | PUTBACK; |
|
|
872 | call_pv ("CBOR::XS::_hv_store", G_VOID | G_DISCARD | G_EVAL); |
|
|
873 | SPAGAIN; |
|
|
874 | |
|
|
875 | FREETMPS; LEAVE; |
|
|
876 | |
|
|
877 | if (SvTRUE (ERRSV)) |
|
|
878 | ERR_ERRSV; |
|
|
879 | |
|
|
880 | return; |
|
|
881 | } |
|
|
882 | |
700 | hv_store_ent (hv, k, v, 0); |
883 | hv_store_ent (hv, k, v, 0); |
701 | SvREFCNT_dec (k); |
884 | SvREFCNT_dec (k); |
|
|
885 | |
|
|
886 | fail: |
|
|
887 | ; |
702 | } |
888 | } |
703 | |
889 | |
704 | static SV * |
890 | static SV * |
705 | decode_hv (dec_t *dec) |
891 | decode_hv (dec_t *dec) |
706 | { |
892 | { |
707 | HV *hv = newHV (); |
893 | HV *hv = newHV (); |
708 | |
894 | |
709 | DEC_INC_DEPTH; |
895 | DEC_INC_DEPTH; |
710 | |
896 | |
711 | if ((*dec->cur & 31) == 31) |
897 | if (*dec->cur == (MAJOR_MAP | MINOR_INDEF)) |
712 | { |
898 | { |
713 | ++dec->cur; |
899 | ++dec->cur; |
714 | |
900 | |
715 | for (;;) |
901 | for (;;) |
716 | { |
902 | { |
717 | WANT (1); |
903 | WANT (1); |
718 | |
904 | |
719 | if (*dec->cur == (0xe0 | 31)) |
905 | if (*dec->cur == (MAJOR_MISC | MINOR_INDEF)) |
720 | { |
906 | { |
721 | ++dec->cur; |
907 | ++dec->cur; |
722 | break; |
908 | break; |
723 | } |
909 | } |
724 | |
910 | |
725 | decode_he (dec, hv); |
911 | decode_he (dec, hv); |
726 | } |
912 | } |
727 | } |
913 | } |
728 | else |
914 | else |
729 | { |
915 | { |
730 | int pairs = decode_uint (dec); |
916 | UV pairs = decode_uint (dec); |
|
|
917 | |
|
|
918 | WANT (pairs); // complexity check - need at least one byte per value, do not allow supersize hashes |
731 | |
919 | |
732 | while (pairs--) |
920 | while (pairs--) |
733 | decode_he (dec, hv); |
921 | decode_he (dec, hv); |
734 | } |
922 | } |
735 | |
923 | |
… | |
… | |
745 | static SV * |
933 | static SV * |
746 | decode_str (dec_t *dec, int utf8) |
934 | decode_str (dec_t *dec, int utf8) |
747 | { |
935 | { |
748 | SV *sv = 0; |
936 | SV *sv = 0; |
749 | |
937 | |
750 | if ((*dec->cur & 31) == 31) |
938 | if ((*dec->cur & MINOR_MASK) == MINOR_INDEF) |
751 | { |
939 | { |
|
|
940 | // indefinite length strings |
752 | ++dec->cur; |
941 | ++dec->cur; |
753 | |
942 | |
|
|
943 | U8 major = *dec->cur & MAJOR_MISC; |
|
|
944 | |
754 | sv = newSVpvn ("", 0); |
945 | sv = newSVpvn ("", 0); |
755 | |
946 | |
756 | // not very fast, and certainly not robust against illegal input |
|
|
757 | for (;;) |
947 | for (;;) |
758 | { |
948 | { |
759 | WANT (1); |
949 | WANT (1); |
760 | |
950 | |
761 | if (*dec->cur == (0xe0 | 31)) |
951 | if ((*dec->cur - major) > LENGTH_EXT8) |
|
|
952 | if (*dec->cur == (MAJOR_MISC | MINOR_INDEF)) |
762 | { |
953 | { |
763 | ++dec->cur; |
954 | ++dec->cur; |
764 | break; |
955 | break; |
765 | } |
956 | } |
|
|
957 | else |
|
|
958 | ERR ("corrupted CBOR data (invalid chunks in indefinite length string)"); |
766 | |
959 | |
767 | sv_catsv (sv, decode_sv (dec)); |
960 | STRLEN len = decode_uint (dec); |
|
|
961 | |
|
|
962 | WANT (len); |
|
|
963 | sv_catpvn (sv, dec->cur, len); |
|
|
964 | dec->cur += len; |
768 | } |
965 | } |
769 | } |
966 | } |
770 | else |
967 | else |
771 | { |
968 | { |
772 | STRLEN len = decode_uint (dec); |
969 | STRLEN len = decode_uint (dec); |
… | |
… | |
779 | && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1)) |
976 | && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1)) |
780 | av_push (dec->stringref, SvREFCNT_inc_NN (sv)); |
977 | av_push (dec->stringref, SvREFCNT_inc_NN (sv)); |
781 | } |
978 | } |
782 | |
979 | |
783 | if (utf8) |
980 | if (utf8) |
|
|
981 | { |
|
|
982 | if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8)) |
|
|
983 | if (!is_utf8_string (SvPVX (sv), SvCUR (sv))) |
|
|
984 | ERR ("corrupted CBOR data (invalid UTF-8 in text string)"); |
|
|
985 | |
784 | SvUTF8_on (sv); |
986 | SvUTF8_on (sv); |
|
|
987 | } |
785 | |
988 | |
786 | return sv; |
989 | return sv; |
787 | |
990 | |
788 | fail: |
991 | fail: |
789 | SvREFCNT_dec (sv); |
992 | SvREFCNT_dec (sv); |
… | |
… | |
808 | sv = newRV_noinc (decode_sv (dec)); |
1011 | sv = newRV_noinc (decode_sv (dec)); |
809 | break; |
1012 | break; |
810 | |
1013 | |
811 | case CBOR_TAG_STRINGREF_NAMESPACE: |
1014 | case CBOR_TAG_STRINGREF_NAMESPACE: |
812 | { |
1015 | { |
|
|
1016 | // do nmot use SAVETMPS/FREETMPS, as these will |
|
|
1017 | // erase mortalised caches, e.g. "shareable" |
813 | ENTER; SAVETMPS; |
1018 | ENTER; |
814 | |
1019 | |
815 | SAVESPTR (dec->stringref); |
1020 | SAVESPTR (dec->stringref); |
816 | dec->stringref = (AV *)sv_2mortal ((SV *)newAV ()); |
1021 | dec->stringref = (AV *)sv_2mortal ((SV *)newAV ()); |
817 | |
1022 | |
818 | sv = decode_sv (dec); |
1023 | sv = decode_sv (dec); |
819 | |
1024 | |
820 | FREETMPS; LEAVE; |
1025 | LEAVE; |
821 | } |
1026 | } |
822 | break; |
1027 | break; |
823 | |
1028 | |
824 | case CBOR_TAG_STRINGREF: |
1029 | case CBOR_TAG_STRINGREF: |
825 | { |
1030 | { |
826 | if ((*dec->cur >> 5) != 0) |
1031 | if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT)) |
827 | ERR ("corrupted CBOR data (stringref index not an unsigned integer)"); |
1032 | ERR ("corrupted CBOR data (stringref index not an unsigned integer)"); |
828 | |
1033 | |
829 | UV idx = decode_uint (dec); |
1034 | UV idx = decode_uint (dec); |
830 | |
1035 | |
831 | if (!dec->stringref || (int)idx > AvFILLp (dec->stringref)) |
1036 | if (!dec->stringref || (int)idx > AvFILLp (dec->stringref)) |
… | |
… | |
838 | case CBOR_TAG_VALUE_SHAREABLE: |
1043 | case CBOR_TAG_VALUE_SHAREABLE: |
839 | { |
1044 | { |
840 | if (ecb_expect_false (!dec->shareable)) |
1045 | if (ecb_expect_false (!dec->shareable)) |
841 | dec->shareable = (AV *)sv_2mortal ((SV *)newAV ()); |
1046 | dec->shareable = (AV *)sv_2mortal ((SV *)newAV ()); |
842 | |
1047 | |
|
|
1048 | if (dec->cbor.flags & F_ALLOW_CYCLES) |
|
|
1049 | { |
843 | sv = newSV (0); |
1050 | sv = newSV (0); |
844 | av_push (dec->shareable, SvREFCNT_inc_NN (sv)); |
1051 | av_push (dec->shareable, SvREFCNT_inc_NN (sv)); |
845 | |
1052 | |
846 | SV *osv = decode_sv (dec); |
1053 | SV *osv = decode_sv (dec); |
847 | sv_setsv (sv, osv); |
1054 | sv_setsv (sv, osv); |
848 | SvREFCNT_dec_NN (osv); |
1055 | SvREFCNT_dec_NN (osv); |
|
|
1056 | } |
|
|
1057 | else |
|
|
1058 | { |
|
|
1059 | av_push (dec->shareable, &PL_sv_undef); |
|
|
1060 | int idx = AvFILLp (dec->shareable); |
|
|
1061 | sv = decode_sv (dec); |
|
|
1062 | av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv)); |
|
|
1063 | } |
849 | } |
1064 | } |
850 | break; |
1065 | break; |
851 | |
1066 | |
852 | case CBOR_TAG_VALUE_SHAREDREF: |
1067 | case CBOR_TAG_VALUE_SHAREDREF: |
853 | { |
1068 | { |
854 | if ((*dec->cur >> 5) != 0) |
1069 | if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT)) |
855 | ERR ("corrupted CBOR data (sharedref index not an unsigned integer)"); |
1070 | ERR ("corrupted CBOR data (sharedref index not an unsigned integer)"); |
856 | |
1071 | |
857 | UV idx = decode_uint (dec); |
1072 | UV idx = decode_uint (dec); |
858 | |
1073 | |
859 | if (!dec->shareable || (int)idx > AvFILLp (dec->shareable)) |
1074 | if (!dec->shareable || (int)idx > AvFILLp (dec->shareable)) |
860 | ERR ("corrupted CBOR data (sharedref index out of bounds)"); |
1075 | ERR ("corrupted CBOR data (sharedref index out of bounds)"); |
861 | |
1076 | |
862 | sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]); |
1077 | sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]); |
|
|
1078 | |
|
|
1079 | if (sv == &PL_sv_undef) |
|
|
1080 | ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled"); |
863 | } |
1081 | } |
864 | break; |
1082 | break; |
865 | |
1083 | |
866 | case CBOR_TAG_PERL_OBJECT: |
1084 | case CBOR_TAG_PERL_OBJECT: |
867 | { |
1085 | { |
… | |
… | |
882 | if (!method) |
1100 | if (!method) |
883 | ERR ("cannot decode perl-object (package does not have a THAW method)"); |
1101 | ERR ("cannot decode perl-object (package does not have a THAW method)"); |
884 | |
1102 | |
885 | dSP; |
1103 | dSP; |
886 | |
1104 | |
887 | ENTER; SAVETMPS; PUSHMARK (SP); |
1105 | ENTER; SAVETMPS; |
|
|
1106 | PUSHMARK (SP); |
888 | EXTEND (SP, len + 1); |
1107 | EXTEND (SP, len + 1); |
889 | // we re-bless the reference to get overload and other niceties right |
1108 | // we re-bless the reference to get overload and other niceties right |
890 | PUSHs (*av_fetch (av, 0, 1)); |
1109 | PUSHs (*av_fetch (av, 0, 1)); |
891 | PUSHs (sv_cbor); |
1110 | PUSHs (sv_cbor); |
892 | |
1111 | |
… | |
… | |
900 | SPAGAIN; |
1119 | SPAGAIN; |
901 | |
1120 | |
902 | if (SvTRUE (ERRSV)) |
1121 | if (SvTRUE (ERRSV)) |
903 | { |
1122 | { |
904 | FREETMPS; LEAVE; |
1123 | FREETMPS; LEAVE; |
905 | ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV)))); |
1124 | ERR_ERRSV; |
906 | } |
1125 | } |
907 | |
1126 | |
908 | SvREFCNT_dec (sv); |
1127 | SvREFCNT_dec (sv); |
909 | sv = SvREFCNT_inc (POPs); |
1128 | sv = SvREFCNT_inc (POPs); |
910 | |
1129 | |
… | |
… | |
914 | } |
1133 | } |
915 | break; |
1134 | break; |
916 | |
1135 | |
917 | default: |
1136 | default: |
918 | { |
1137 | { |
|
|
1138 | SV *tag_sv = newSVuv (tag); |
|
|
1139 | |
919 | sv = decode_sv (dec); |
1140 | sv = decode_sv (dec); |
920 | |
1141 | |
921 | dSP; |
1142 | dSP; |
922 | ENTER; SAVETMPS; PUSHMARK (SP); |
1143 | ENTER; SAVETMPS; |
|
|
1144 | SAVESTACK_POS (); |
|
|
1145 | PUSHMARK (SP); |
923 | EXTEND (SP, 2); |
1146 | EXTEND (SP, 2); |
924 | PUSHs (newSVuv (tag)); |
1147 | PUSHs (tag_sv); |
925 | PUSHs (sv); |
1148 | PUSHs (sv); |
926 | |
1149 | |
927 | PUTBACK; |
1150 | PUTBACK; |
928 | int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL); |
1151 | int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL); |
929 | SPAGAIN; |
1152 | SPAGAIN; |
930 | |
1153 | |
931 | if (SvTRUE (ERRSV)) |
1154 | if (SvTRUE (ERRSV)) |
932 | { |
1155 | { |
|
|
1156 | SvREFCNT_dec (tag_sv); |
933 | FREETMPS; LEAVE; |
1157 | FREETMPS; LEAVE; |
934 | ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV)))); |
1158 | ERR_ERRSV; |
935 | } |
1159 | } |
936 | |
1160 | |
937 | if (count) |
1161 | if (count) |
938 | { |
1162 | { |
|
|
1163 | SvREFCNT_dec (tag_sv); |
939 | SvREFCNT_dec (sv); |
1164 | SvREFCNT_dec (sv); |
940 | sv = SvREFCNT_inc (POPs); |
1165 | sv = SvREFCNT_inc (POPs); |
941 | } |
1166 | } |
942 | else |
1167 | else |
943 | { |
1168 | { |
944 | AV *av = newAV (); |
1169 | AV *av = newAV (); |
945 | av_push (av, newSVuv (tag)); |
1170 | av_push (av, tag_sv); |
946 | av_push (av, sv); |
1171 | av_push (av, sv); |
947 | |
1172 | |
948 | HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash |
1173 | HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash |
949 | ? cbor_tagged_stash |
1174 | ? cbor_tagged_stash |
950 | : gv_stashpv ("CBOR::XS::Tagged" , 1); |
1175 | : gv_stashpv ("CBOR::XS::Tagged" , 1); |
… | |
… | |
968 | static SV * |
1193 | static SV * |
969 | decode_sv (dec_t *dec) |
1194 | decode_sv (dec_t *dec) |
970 | { |
1195 | { |
971 | WANT (1); |
1196 | WANT (1); |
972 | |
1197 | |
973 | switch (*dec->cur >> 5) |
1198 | switch (*dec->cur >> MAJOR_SHIFT) |
974 | { |
1199 | { |
975 | case 0: // unsigned int |
1200 | case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec)); |
976 | return newSVuv (decode_uint (dec)); |
1201 | case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec)); |
977 | case 1: // negative int |
1202 | case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0); |
978 | return newSViv (-1 - (IV)decode_uint (dec)); |
1203 | case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1); |
979 | case 2: // octet string |
1204 | case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec); |
980 | return decode_str (dec, 0); |
1205 | case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec); |
981 | case 3: // utf-8 string |
1206 | case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec); |
982 | return decode_str (dec, 1); |
1207 | |
983 | case 4: // array |
1208 | case MAJOR_MISC >> MAJOR_SHIFT: |
984 | return decode_av (dec); |
|
|
985 | case 5: // map |
|
|
986 | return decode_hv (dec); |
|
|
987 | case 6: // tag |
|
|
988 | return decode_tagged (dec); |
|
|
989 | case 7: // misc |
|
|
990 | switch (*dec->cur++ & 31) |
1209 | switch (*dec->cur++ & MINOR_MASK) |
991 | { |
1210 | { |
992 | case 20: |
1211 | case SIMPLE_FALSE: |
993 | #if CBOR_SLOW |
1212 | #if CBOR_SLOW |
994 | types_false = get_bool ("Types::Serialiser::false"); |
1213 | types_false = get_bool ("Types::Serialiser::false"); |
995 | #endif |
1214 | #endif |
996 | return newSVsv (types_false); |
1215 | return newSVsv (types_false); |
997 | case 21: |
1216 | case SIMPLE_TRUE: |
998 | #if CBOR_SLOW |
1217 | #if CBOR_SLOW |
999 | types_true = get_bool ("Types::Serialiser::true"); |
1218 | types_true = get_bool ("Types::Serialiser::true"); |
1000 | #endif |
1219 | #endif |
1001 | return newSVsv (types_true); |
1220 | return newSVsv (types_true); |
1002 | case 22: |
1221 | case SIMPLE_NULL: |
1003 | return newSVsv (&PL_sv_undef); |
1222 | return newSVsv (&PL_sv_undef); |
1004 | case 23: |
1223 | case SIMPLE_UNDEF: |
1005 | #if CBOR_SLOW |
1224 | #if CBOR_SLOW |
1006 | types_error = get_bool ("Types::Serialiser::error"); |
1225 | types_error = get_bool ("Types::Serialiser::error"); |
1007 | #endif |
1226 | #endif |
1008 | return newSVsv (types_error); |
1227 | return newSVsv (types_error); |
1009 | |
1228 | |
1010 | case 25: |
1229 | case MISC_FLOAT16: |
1011 | { |
1230 | { |
1012 | WANT (2); |
1231 | WANT (2); |
1013 | |
1232 | |
1014 | uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; |
1233 | uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; |
1015 | dec->cur += 2; |
1234 | dec->cur += 2; |
1016 | |
1235 | |
1017 | return newSVnv (ecb_binary16_to_float (fp)); |
1236 | return newSVnv (ecb_binary16_to_float (fp)); |
1018 | } |
1237 | } |
1019 | |
1238 | |
1020 | case 26: |
1239 | case MISC_FLOAT32: |
1021 | { |
1240 | { |
1022 | uint32_t fp; |
1241 | uint32_t fp; |
1023 | WANT (4); |
1242 | WANT (4); |
1024 | memcpy (&fp, dec->cur, 4); |
1243 | memcpy (&fp, dec->cur, 4); |
1025 | dec->cur += 4; |
1244 | dec->cur += 4; |
… | |
… | |
1028 | fp = ecb_bswap32 (fp); |
1247 | fp = ecb_bswap32 (fp); |
1029 | |
1248 | |
1030 | return newSVnv (ecb_binary32_to_float (fp)); |
1249 | return newSVnv (ecb_binary32_to_float (fp)); |
1031 | } |
1250 | } |
1032 | |
1251 | |
1033 | case 27: |
1252 | case MISC_FLOAT64: |
1034 | { |
1253 | { |
1035 | uint64_t fp; |
1254 | uint64_t fp; |
1036 | WANT (8); |
1255 | WANT (8); |
1037 | memcpy (&fp, dec->cur, 8); |
1256 | memcpy (&fp, dec->cur, 8); |
1038 | dec->cur += 8; |
1257 | dec->cur += 8; |
… | |
… | |
1041 | fp = ecb_bswap64 (fp); |
1260 | fp = ecb_bswap64 (fp); |
1042 | |
1261 | |
1043 | return newSVnv (ecb_binary64_to_double (fp)); |
1262 | return newSVnv (ecb_binary64_to_double (fp)); |
1044 | } |
1263 | } |
1045 | |
1264 | |
1046 | // 0..19 unassigned |
1265 | // 0..19 unassigned simple |
1047 | // 24 reserved + unassigned (reserved values are not encodable) |
1266 | // 24 reserved + unassigned simple (reserved values are not encodable) |
|
|
1267 | // 28-30 unassigned misc |
|
|
1268 | // 31 break code |
1048 | default: |
1269 | default: |
1049 | ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); |
1270 | ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)"); |
1050 | } |
1271 | } |
1051 | |
1272 | |
1052 | break; |
1273 | break; |
1053 | } |
1274 | } |
1054 | |
1275 | |
… | |
… | |
1057 | } |
1278 | } |
1058 | |
1279 | |
1059 | static SV * |
1280 | static SV * |
1060 | decode_cbor (SV *string, CBOR *cbor, char **offset_return) |
1281 | decode_cbor (SV *string, CBOR *cbor, char **offset_return) |
1061 | { |
1282 | { |
1062 | dec_t dec = { }; |
1283 | dec_t dec = { 0 }; |
1063 | SV *sv; |
1284 | SV *sv; |
1064 | STRLEN len; |
1285 | STRLEN len; |
1065 | char *data = SvPVbyte (string, len); |
1286 | char *data = SvPVbyte (string, len); |
1066 | |
1287 | |
1067 | if (len > cbor->max_size && cbor->max_size) |
1288 | if (len > cbor->max_size && cbor->max_size) |
… | |
… | |
1081 | if (dec.cur != dec.end && !dec.err) |
1302 | if (dec.cur != dec.end && !dec.err) |
1082 | dec.err = "garbage after CBOR object"; |
1303 | dec.err = "garbage after CBOR object"; |
1083 | |
1304 | |
1084 | if (dec.err) |
1305 | if (dec.err) |
1085 | { |
1306 | { |
|
|
1307 | if (dec.shareable) |
|
|
1308 | { |
|
|
1309 | // need to break cyclic links, which would all be in shareable |
|
|
1310 | int i; |
|
|
1311 | SV **svp; |
|
|
1312 | |
|
|
1313 | for (i = av_len (dec.shareable) + 1; i--; ) |
|
|
1314 | if ((svp = av_fetch (dec.shareable, i, 0))) |
|
|
1315 | sv_setsv (*svp, &PL_sv_undef); |
|
|
1316 | } |
|
|
1317 | |
1086 | SvREFCNT_dec (sv); |
1318 | SvREFCNT_dec (sv); |
|
|
1319 | |
|
|
1320 | if (dec.err_sv) |
|
|
1321 | sv_2mortal (dec.err_sv); |
|
|
1322 | |
1087 | croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur); |
1323 | croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur); |
1088 | } |
1324 | } |
1089 | |
1325 | |
1090 | sv = sv_2mortal (sv); |
1326 | sv = sv_2mortal (sv); |
1091 | |
1327 | |
1092 | return sv; |
1328 | return sv; |
1093 | } |
1329 | } |
1094 | |
1330 | |
|
|
1331 | ///////////////////////////////////////////////////////////////////////////// |
|
|
1332 | // incremental parser |
|
|
1333 | |
|
|
1334 | #define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0) |
|
|
1335 | |
|
|
1336 | // returns 0 for notyet, 1 for success or error |
|
|
1337 | static int |
|
|
1338 | incr_parse (CBOR *self, SV *cborstr) |
|
|
1339 | { |
|
|
1340 | STRLEN cur; |
|
|
1341 | SvPV (cborstr, cur); |
|
|
1342 | |
|
|
1343 | while (ecb_expect_true (self->incr_need <= cur)) |
|
|
1344 | { |
|
|
1345 | // table of integer count bytes |
|
|
1346 | static I8 incr_len[MINOR_MASK + 1] = { |
|
|
1347 | 0, 0, 0, 0, 0, 0, 0, 0, |
|
|
1348 | 0, 0, 0, 0, 0, 0, 0, 0, |
|
|
1349 | 0, 0, 0, 0, 0, 0, 0, 0, |
|
|
1350 | 1, 2, 4, 8,-1,-1,-1,-2 |
|
|
1351 | }; |
|
|
1352 | |
|
|
1353 | const U8 *p = SvPVX (cborstr) + self->incr_pos; |
|
|
1354 | U8 m = *p & MINOR_MASK; |
|
|
1355 | IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]); |
|
|
1356 | I8 ilen = incr_len[m]; |
|
|
1357 | |
|
|
1358 | self->incr_need = self->incr_pos + 1; |
|
|
1359 | |
|
|
1360 | if (ecb_expect_false (ilen < 0)) |
|
|
1361 | { |
|
|
1362 | if (m != MINOR_INDEF) |
|
|
1363 | return 1; // error |
|
|
1364 | |
|
|
1365 | if (*p == (MAJOR_MISC | MINOR_INDEF)) |
|
|
1366 | { |
|
|
1367 | if (count >= 0) |
|
|
1368 | return 1; // error |
|
|
1369 | |
|
|
1370 | count = 1; |
|
|
1371 | } |
|
|
1372 | else |
|
|
1373 | { |
|
|
1374 | av_push (self->incr_count, newSViv (-1)); //TODO: nest |
|
|
1375 | count = -1; |
|
|
1376 | } |
|
|
1377 | } |
|
|
1378 | else |
|
|
1379 | { |
|
|
1380 | self->incr_need += ilen; |
|
|
1381 | if (ecb_expect_false (self->incr_need > cur)) |
|
|
1382 | return 0; |
|
|
1383 | |
|
|
1384 | int major = *p >> MAJOR_SHIFT; |
|
|
1385 | |
|
|
1386 | switch (major) |
|
|
1387 | { |
|
|
1388 | case MAJOR_TAG >> MAJOR_SHIFT: |
|
|
1389 | ++count; // tags merely prefix another value |
|
|
1390 | break; |
|
|
1391 | |
|
|
1392 | case MAJOR_BYTES >> MAJOR_SHIFT: |
|
|
1393 | case MAJOR_TEXT >> MAJOR_SHIFT: |
|
|
1394 | case MAJOR_ARRAY >> MAJOR_SHIFT: |
|
|
1395 | case MAJOR_MAP >> MAJOR_SHIFT: |
|
|
1396 | { |
|
|
1397 | UV len; |
|
|
1398 | |
|
|
1399 | if (ecb_expect_false (ilen)) |
|
|
1400 | { |
|
|
1401 | len = 0; |
|
|
1402 | |
|
|
1403 | do { |
|
|
1404 | len = (len << 8) | *++p; |
|
|
1405 | } while (--ilen); |
|
|
1406 | } |
|
|
1407 | else |
|
|
1408 | len = m; |
|
|
1409 | |
|
|
1410 | switch (major) |
|
|
1411 | { |
|
|
1412 | case MAJOR_BYTES >> MAJOR_SHIFT: |
|
|
1413 | case MAJOR_TEXT >> MAJOR_SHIFT: |
|
|
1414 | self->incr_need += len; |
|
|
1415 | if (ecb_expect_false (self->incr_need > cur)) |
|
|
1416 | return 0; |
|
|
1417 | |
|
|
1418 | break; |
|
|
1419 | |
|
|
1420 | case MAJOR_MAP >> MAJOR_SHIFT: |
|
|
1421 | len <<= 1; |
|
|
1422 | case MAJOR_ARRAY >> MAJOR_SHIFT: |
|
|
1423 | if (len) |
|
|
1424 | { |
|
|
1425 | av_push (self->incr_count, newSViv (len + 1)); //TODO: nest |
|
|
1426 | count = len + 1; |
|
|
1427 | } |
|
|
1428 | break; |
|
|
1429 | } |
|
|
1430 | } |
|
|
1431 | } |
|
|
1432 | } |
|
|
1433 | |
|
|
1434 | self->incr_pos = self->incr_need; |
|
|
1435 | |
|
|
1436 | if (count > 0) |
|
|
1437 | { |
|
|
1438 | while (!--count) |
|
|
1439 | { |
|
|
1440 | if (!AvFILLp (self->incr_count)) |
|
|
1441 | return 1; // done |
|
|
1442 | |
|
|
1443 | SvREFCNT_dec_NN (av_pop (self->incr_count)); |
|
|
1444 | count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]); |
|
|
1445 | } |
|
|
1446 | |
|
|
1447 | SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count; |
|
|
1448 | } |
|
|
1449 | } |
|
|
1450 | |
|
|
1451 | return 0; |
|
|
1452 | } |
|
|
1453 | |
|
|
1454 | |
1095 | ///////////////////////////////////////////////////////////////////////////// |
1455 | ///////////////////////////////////////////////////////////////////////////// |
1096 | // XS interface functions |
1456 | // XS interface functions |
1097 | |
1457 | |
1098 | MODULE = CBOR::XS PACKAGE = CBOR::XS |
1458 | MODULE = CBOR::XS PACKAGE = CBOR::XS |
1099 | |
1459 | |
… | |
… | |
1111 | |
1471 | |
1112 | default_filter = newSVpv ("CBOR::XS::default_filter", 0); |
1472 | default_filter = newSVpv ("CBOR::XS::default_filter", 0); |
1113 | |
1473 | |
1114 | sv_cbor = newSVpv ("CBOR", 0); |
1474 | sv_cbor = newSVpv ("CBOR", 0); |
1115 | SvREADONLY_on (sv_cbor); |
1475 | SvREADONLY_on (sv_cbor); |
|
|
1476 | |
|
|
1477 | assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1)); |
1116 | } |
1478 | } |
1117 | |
1479 | |
1118 | PROTOTYPES: DISABLE |
1480 | PROTOTYPES: DISABLE |
1119 | |
1481 | |
1120 | void CLONE (...) |
1482 | void CLONE (...) |
… | |
… | |
1139 | void shrink (CBOR *self, int enable = 1) |
1501 | void shrink (CBOR *self, int enable = 1) |
1140 | ALIAS: |
1502 | ALIAS: |
1141 | shrink = F_SHRINK |
1503 | shrink = F_SHRINK |
1142 | allow_unknown = F_ALLOW_UNKNOWN |
1504 | allow_unknown = F_ALLOW_UNKNOWN |
1143 | allow_sharing = F_ALLOW_SHARING |
1505 | allow_sharing = F_ALLOW_SHARING |
1144 | allow_stringref = F_ALLOW_STRINGREF |
1506 | allow_cycles = F_ALLOW_CYCLES |
|
|
1507 | pack_strings = F_PACK_STRINGS |
|
|
1508 | text_keys = F_TEXT_KEYS |
|
|
1509 | text_strings = F_TEXT_STRINGS |
|
|
1510 | validate_utf8 = F_VALIDATE_UTF8 |
1145 | PPCODE: |
1511 | PPCODE: |
1146 | { |
1512 | { |
1147 | if (enable) |
1513 | if (enable) |
1148 | self->flags |= ix; |
1514 | self->flags |= ix; |
1149 | else |
1515 | else |
… | |
… | |
1155 | void get_shrink (CBOR *self) |
1521 | void get_shrink (CBOR *self) |
1156 | ALIAS: |
1522 | ALIAS: |
1157 | get_shrink = F_SHRINK |
1523 | get_shrink = F_SHRINK |
1158 | get_allow_unknown = F_ALLOW_UNKNOWN |
1524 | get_allow_unknown = F_ALLOW_UNKNOWN |
1159 | get_allow_sharing = F_ALLOW_SHARING |
1525 | get_allow_sharing = F_ALLOW_SHARING |
1160 | get_allow_stringref = F_ALLOW_STRINGREF |
1526 | get_allow_cycles = F_ALLOW_CYCLES |
|
|
1527 | get_pack_strings = F_PACK_STRINGS |
|
|
1528 | get_text_keys = F_TEXT_KEYS |
|
|
1529 | get_text_strings = F_TEXT_STRINGS |
|
|
1530 | get_validate_utf8 = F_VALIDATE_UTF8 |
1161 | PPCODE: |
1531 | PPCODE: |
1162 | XPUSHs (boolSV (self->flags & ix)); |
1532 | XPUSHs (boolSV (self->flags & ix)); |
1163 | |
1533 | |
1164 | void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) |
1534 | void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) |
1165 | PPCODE: |
1535 | PPCODE: |
… | |
… | |
1214 | EXTEND (SP, 2); |
1584 | EXTEND (SP, 2); |
1215 | PUSHs (sv); |
1585 | PUSHs (sv); |
1216 | PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); |
1586 | PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); |
1217 | } |
1587 | } |
1218 | |
1588 | |
|
|
1589 | void incr_parse (CBOR *self, SV *cborstr) |
|
|
1590 | ALIAS: |
|
|
1591 | incr_parse_multiple = 1 |
|
|
1592 | PPCODE: |
|
|
1593 | { |
|
|
1594 | if (SvUTF8 (cborstr)) |
|
|
1595 | sv_utf8_downgrade (cborstr, 0); |
|
|
1596 | |
|
|
1597 | if (!self->incr_count) |
|
|
1598 | { |
|
|
1599 | self->incr_count = newAV (); |
|
|
1600 | self->incr_pos = 0; |
|
|
1601 | self->incr_need = 1; |
|
|
1602 | |
|
|
1603 | av_push (self->incr_count, newSViv (1)); |
|
|
1604 | } |
|
|
1605 | |
|
|
1606 | do |
|
|
1607 | { |
|
|
1608 | if (!incr_parse (self, cborstr)) |
|
|
1609 | { |
|
|
1610 | if (self->incr_need > self->max_size && self->max_size) |
|
|
1611 | croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu", |
|
|
1612 | (unsigned long)self->incr_need, (unsigned long)self->max_size); |
|
|
1613 | |
|
|
1614 | break; |
|
|
1615 | } |
|
|
1616 | |
|
|
1617 | SV *sv; |
|
|
1618 | char *offset; |
|
|
1619 | |
|
|
1620 | PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN; |
|
|
1621 | XPUSHs (sv); |
|
|
1622 | |
|
|
1623 | sv_chop (cborstr, offset); |
|
|
1624 | |
|
|
1625 | av_clear (self->incr_count); |
|
|
1626 | av_push (self->incr_count, newSViv (1)); |
|
|
1627 | |
|
|
1628 | self->incr_pos = 0; |
|
|
1629 | self->incr_need = self->incr_pos + 1; |
|
|
1630 | } |
|
|
1631 | while (ix); |
|
|
1632 | } |
|
|
1633 | |
|
|
1634 | void incr_reset (CBOR *self) |
|
|
1635 | CODE: |
|
|
1636 | { |
|
|
1637 | SvREFCNT_dec (self->incr_count); |
|
|
1638 | self->incr_count = 0; |
|
|
1639 | } |
|
|
1640 | |
1219 | void DESTROY (CBOR *self) |
1641 | void DESTROY (CBOR *self) |
1220 | PPCODE: |
1642 | PPCODE: |
1221 | cbor_free (self); |
1643 | cbor_free (self); |
1222 | |
1644 | |
1223 | PROTOTYPES: ENABLE |
1645 | PROTOTYPES: ENABLE |
1224 | |
1646 | |
1225 | void encode_cbor (SV *scalar) |
1647 | void encode_cbor (SV *scalar) |
|
|
1648 | ALIAS: |
|
|
1649 | encode_cbor = 0 |
|
|
1650 | encode_cbor_sharing = F_ALLOW_SHARING |
1226 | PPCODE: |
1651 | PPCODE: |
1227 | { |
1652 | { |
1228 | CBOR cbor; |
1653 | CBOR cbor; |
1229 | cbor_init (&cbor); |
1654 | cbor_init (&cbor); |
|
|
1655 | cbor.flags |= ix; |
1230 | PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; |
1656 | PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; |
1231 | XPUSHs (scalar); |
1657 | XPUSHs (scalar); |
1232 | } |
1658 | } |
1233 | |
1659 | |
1234 | void decode_cbor (SV *cborstr) |
1660 | void decode_cbor (SV *cborstr) |