… | |
… | |
6 | #include <string.h> |
6 | #include <string.h> |
7 | #include <stdlib.h> |
7 | #include <stdlib.h> |
8 | #include <stdio.h> |
8 | #include <stdio.h> |
9 | #include <limits.h> |
9 | #include <limits.h> |
10 | #include <float.h> |
10 | #include <float.h> |
|
|
11 | #include <inttypes.h> |
11 | |
12 | |
|
|
13 | #define ECB_NO_THREADS 1 |
12 | #include "ecb.h" |
14 | #include "ecb.h" |
13 | |
15 | |
14 | // compatibility with perl <5.18 |
16 | // compatibility with perl <5.18 |
15 | #ifndef HvNAMELEN_get |
17 | #ifndef HvNAMELEN_get |
16 | # define HvNAMELEN_get(hv) strlen (HvNAME (hv)) |
18 | # define HvNAMELEN_get(hv) strlen (HvNAME (hv)) |
… | |
… | |
19 | # define HvNAMELEN(hv) HvNAMELEN_get (hv) |
21 | # define HvNAMELEN(hv) HvNAMELEN_get (hv) |
20 | #endif |
22 | #endif |
21 | #ifndef HvNAMEUTF8 |
23 | #ifndef HvNAMEUTF8 |
22 | # define HvNAMEUTF8(hv) 0 |
24 | # define HvNAMEUTF8(hv) 0 |
23 | #endif |
25 | #endif |
|
|
26 | #ifndef SvREFCNT_inc_NN |
|
|
27 | # define SvREFCNT_inc_NN(sv) SvREFCNT_inc (sv) |
|
|
28 | #endif |
|
|
29 | #ifndef SvREFCNT_dec_NN |
|
|
30 | # define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv) |
|
|
31 | #endif |
|
|
32 | |
|
|
33 | // known major and minor types |
|
|
34 | enum cbor_type |
|
|
35 | { |
|
|
36 | MAJOR_SHIFT = 5, |
|
|
37 | MINOR_MASK = 0x1f, |
|
|
38 | |
|
|
39 | MAJOR_POS_INT = 0 << MAJOR_SHIFT, |
|
|
40 | MAJOR_NEG_INT = 1 << MAJOR_SHIFT, |
|
|
41 | MAJOR_BYTES = 2 << MAJOR_SHIFT, |
|
|
42 | MAJOR_TEXT = 3 << MAJOR_SHIFT, |
|
|
43 | MAJOR_ARRAY = 4 << MAJOR_SHIFT, |
|
|
44 | MAJOR_MAP = 5 << MAJOR_SHIFT, |
|
|
45 | MAJOR_TAG = 6 << MAJOR_SHIFT, |
|
|
46 | MAJOR_MISC = 7 << MAJOR_SHIFT, |
|
|
47 | |
|
|
48 | // INT/STRING/ARRAY/MAP subtypes |
|
|
49 | LENGTH_EXT1 = 24, |
|
|
50 | LENGTH_EXT2 = 25, |
|
|
51 | LENGTH_EXT4 = 26, |
|
|
52 | LENGTH_EXT8 = 27, |
|
|
53 | |
|
|
54 | // SIMPLE types (effectively MISC subtypes) |
|
|
55 | SIMPLE_FALSE = 20, |
|
|
56 | SIMPLE_TRUE = 21, |
|
|
57 | SIMPLE_NULL = 22, |
|
|
58 | SIMPLE_UNDEF = 23, |
|
|
59 | |
|
|
60 | // MISC subtype (unused) |
|
|
61 | MISC_EXT1 = 24, |
|
|
62 | MISC_FLOAT16 = 25, |
|
|
63 | MISC_FLOAT32 = 26, |
|
|
64 | MISC_FLOAT64 = 27, |
|
|
65 | |
|
|
66 | // BYTES/TEXT/ARRAY/MAP |
|
|
67 | MINOR_INDEF = 31, |
|
|
68 | }; |
24 | |
69 | |
25 | // known tags |
70 | // known tags |
26 | enum cbor_tag |
71 | enum cbor_tag |
27 | { |
72 | { |
28 | // inofficial extensions (pending iana registration) |
73 | // extensions |
|
|
74 | CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref |
29 | CBOR_TAG_PERL_OBJECT = 24, // http://cbor.schmorp.de/perl-object |
75 | CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object |
30 | CBOR_TAG_GENERIC_OBJECT = 25, // http://cbor.schmorp.de/generic-object |
76 | CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object |
31 | CBOR_TAG_VALUE_SHAREABLE = 26, // http://cbor.schmorp.de/value-sharing |
77 | CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing |
32 | CBOR_TAG_VALUE_SHAREDREF = 27, // http://cbor.schmorp.de/value-sharing |
78 | CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing |
33 | CBOR_TAG_STRINGREF_NAMESPACE = 65537, // http://cbor.schmorp.de/stringref |
79 | CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref |
34 | CBOR_TAG_STRINGREF = 28, // http://cbor.schmorp.de/stringref |
|
|
35 | CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection |
80 | CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection |
36 | |
81 | |
37 | // rfc7049 |
82 | // rfc7049 |
38 | CBOR_TAG_DATETIME = 0, // rfc4287, utf-8 |
83 | CBOR_TAG_DATETIME = 0, // rfc4287, utf-8 |
39 | CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any |
84 | CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any |
40 | CBOR_TAG_POS_BIGNUM = 2, // byte string |
85 | CBOR_TAG_POS_BIGNUM = 2, // byte string |
41 | CBOR_TAG_NEG_BIGNUM = 3, // byte string |
86 | CBOR_TAG_NEG_BIGNUM = 3, // byte string |
42 | CBOR_TAG_DECIMAL = 4, // decimal fraction, array |
87 | CBOR_TAG_DECIMAL = 4, // decimal fraction, array |
43 | CBOR_TAG_BIGFLOAT = 5, // array |
88 | CBOR_TAG_BIGFLOAT = 5, // array |
44 | |
89 | |
45 | CBOR_TAG_CONV_B64U = 21, // base64url, any |
90 | CBOR_TAG_CONV_B64U = 21, // base64url, any |
46 | CBOR_TAG_CONV_B64 = 22, // base64, any |
91 | CBOR_TAG_CONV_B64 = 22, // base64, any |
47 | CBOR_TAG_CONV_HEX = 23, // base16, any |
92 | CBOR_TAG_CONV_HEX = 23, // base16, any |
48 | CBOR_TAG_CBOR = 24, // embedded cbor, byte string |
93 | CBOR_TAG_CBOR = 24, // embedded cbor, byte string |
49 | |
94 | |
50 | CBOR_TAG_URI = 32, // URI rfc3986, utf-8 |
95 | CBOR_TAG_URI = 32, // URI rfc3986, utf-8 |
51 | CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8 |
96 | CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8 |
52 | CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8 |
97 | CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8 |
53 | CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8 |
98 | CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8 |
54 | CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8 |
99 | CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8 |
55 | |
100 | |
56 | CBOR_TAG_MAGIC = 55799 // self-describe cbor |
101 | CBOR_TAG_MAGIC = 55799, // self-describe cbor |
57 | }; |
102 | }; |
58 | |
103 | |
59 | #define F_SHRINK 0x00000001UL |
104 | #define F_SHRINK 0x00000001UL |
60 | #define F_ALLOW_UNKNOWN 0x00000002UL |
105 | #define F_ALLOW_UNKNOWN 0x00000002UL |
61 | #define F_ALLOW_SHARING 0x00000004UL //TODO |
106 | #define F_ALLOW_SHARING 0x00000004UL |
62 | #define F_ALLOW_STRINGREF 0x00000008UL //TODO |
107 | #define F_ALLOW_CYCLES 0x00000008UL |
|
|
108 | #define F_FORBID_OBJECTS 0x00000010UL |
|
|
109 | #define F_PACK_STRINGS 0x00000020UL |
|
|
110 | #define F_TEXT_KEYS 0x00000040UL |
|
|
111 | #define F_TEXT_STRINGS 0x00000080UL |
|
|
112 | #define F_VALIDATE_UTF8 0x00000100UL |
63 | |
113 | |
64 | #define INIT_SIZE 32 // initial scalar size to be allocated |
114 | #define INIT_SIZE 32 // initial scalar size to be allocated |
65 | |
115 | |
66 | #define SB do { |
116 | #define SB do { |
67 | #define SE } while (0) |
117 | #define SE } while (0) |
… | |
… | |
86 | typedef struct { |
136 | typedef struct { |
87 | U32 flags; |
137 | U32 flags; |
88 | U32 max_depth; |
138 | U32 max_depth; |
89 | STRLEN max_size; |
139 | STRLEN max_size; |
90 | SV *filter; |
140 | SV *filter; |
|
|
141 | |
|
|
142 | // for the incremental parser |
|
|
143 | STRLEN incr_pos; // the current offset into the text |
|
|
144 | STRLEN incr_need; // minimum bytes needed to decode |
|
|
145 | AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef. |
91 | } CBOR; |
146 | } CBOR; |
92 | |
147 | |
93 | ecb_inline void |
148 | ecb_inline void |
94 | cbor_init (CBOR *cbor) |
149 | cbor_init (CBOR *cbor) |
95 | { |
150 | { |
… | |
… | |
99 | |
154 | |
100 | ecb_inline void |
155 | ecb_inline void |
101 | cbor_free (CBOR *cbor) |
156 | cbor_free (CBOR *cbor) |
102 | { |
157 | { |
103 | SvREFCNT_dec (cbor->filter); |
158 | SvREFCNT_dec (cbor->filter); |
|
|
159 | SvREFCNT_dec (cbor->incr_count); |
104 | } |
160 | } |
105 | |
161 | |
106 | ///////////////////////////////////////////////////////////////////////////// |
162 | ///////////////////////////////////////////////////////////////////////////// |
107 | // utility functions |
163 | // utility functions |
108 | |
164 | |
… | |
… | |
134 | |
190 | |
135 | // minimum length of a string to be registered for stringref |
191 | // minimum length of a string to be registered for stringref |
136 | ecb_inline int |
192 | ecb_inline int |
137 | minimum_string_length (UV idx) |
193 | minimum_string_length (UV idx) |
138 | { |
194 | { |
139 | return idx > 23 |
195 | return idx <= 23 ? 3 |
140 | ? idx > 0xffU |
196 | : idx <= 0xffU ? 4 |
141 | ? idx > 0xffffU |
197 | : idx <= 0xffffU ? 5 |
142 | ? idx > 0xffffffffU |
198 | : idx <= 0xffffffffU ? 7 |
143 | ? 7 |
199 | : 11; |
144 | : 6 |
|
|
145 | : 5 |
|
|
146 | : 4 |
|
|
147 | : 3; |
|
|
148 | } |
200 | } |
149 | |
201 | |
150 | ///////////////////////////////////////////////////////////////////////////// |
202 | ///////////////////////////////////////////////////////////////////////////// |
151 | // encoder |
203 | // encoder |
152 | |
204 | |
… | |
… | |
165 | } enc_t; |
217 | } enc_t; |
166 | |
218 | |
167 | ecb_inline void |
219 | ecb_inline void |
168 | need (enc_t *enc, STRLEN len) |
220 | need (enc_t *enc, STRLEN len) |
169 | { |
221 | { |
170 | if (ecb_expect_false (enc->cur + len >= enc->end)) |
222 | if (ecb_expect_false ((uintptr_t)(enc->end - enc->cur) < len)) |
171 | { |
223 | { |
172 | STRLEN cur = enc->cur - (char *)SvPVX (enc->sv); |
224 | STRLEN cur = enc->cur - (char *)SvPVX (enc->sv); |
173 | SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1); |
225 | SvGROW (enc->sv, cur + (len < (cur >> 2) ? cur >> 2 : len) + 1); |
174 | enc->cur = SvPVX (enc->sv) + cur; |
226 | enc->cur = SvPVX (enc->sv) + cur; |
175 | enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1; |
227 | enc->end = SvPVX (enc->sv) + SvLEN (enc->sv) - 1; |
… | |
… | |
186 | static void |
238 | static void |
187 | encode_uint (enc_t *enc, int major, UV len) |
239 | encode_uint (enc_t *enc, int major, UV len) |
188 | { |
240 | { |
189 | need (enc, 9); |
241 | need (enc, 9); |
190 | |
242 | |
191 | if (len < 24) |
243 | if (ecb_expect_true (len < LENGTH_EXT1)) |
192 | *enc->cur++ = major | len; |
244 | *enc->cur++ = major | len; |
193 | else if (len <= 0xff) |
245 | else if (ecb_expect_true (len <= 0xffU)) |
194 | { |
246 | { |
195 | *enc->cur++ = major | 24; |
247 | *enc->cur++ = major | LENGTH_EXT1; |
196 | *enc->cur++ = len; |
248 | *enc->cur++ = len; |
197 | } |
249 | } |
198 | else if (len <= 0xffff) |
250 | else if (len <= 0xffffU) |
199 | { |
251 | { |
200 | *enc->cur++ = major | 25; |
252 | *enc->cur++ = major | LENGTH_EXT2; |
201 | *enc->cur++ = len >> 8; |
253 | *enc->cur++ = len >> 8; |
202 | *enc->cur++ = len; |
254 | *enc->cur++ = len; |
203 | } |
255 | } |
204 | else if (len <= 0xffffffff) |
256 | else if (len <= 0xffffffffU) |
205 | { |
257 | { |
206 | *enc->cur++ = major | 26; |
258 | *enc->cur++ = major | LENGTH_EXT4; |
207 | *enc->cur++ = len >> 24; |
259 | *enc->cur++ = len >> 24; |
208 | *enc->cur++ = len >> 16; |
260 | *enc->cur++ = len >> 16; |
209 | *enc->cur++ = len >> 8; |
261 | *enc->cur++ = len >> 8; |
210 | *enc->cur++ = len; |
262 | *enc->cur++ = len; |
211 | } |
263 | } |
212 | else |
264 | else |
213 | { |
265 | { |
214 | *enc->cur++ = major | 27; |
266 | *enc->cur++ = major | LENGTH_EXT8; |
215 | *enc->cur++ = len >> 56; |
267 | *enc->cur++ = len >> 56; |
216 | *enc->cur++ = len >> 48; |
268 | *enc->cur++ = len >> 48; |
217 | *enc->cur++ = len >> 40; |
269 | *enc->cur++ = len >> 40; |
218 | *enc->cur++ = len >> 32; |
270 | *enc->cur++ = len >> 32; |
219 | *enc->cur++ = len >> 24; |
271 | *enc->cur++ = len >> 24; |
… | |
… | |
224 | } |
276 | } |
225 | |
277 | |
226 | ecb_inline void |
278 | ecb_inline void |
227 | encode_tag (enc_t *enc, UV tag) |
279 | encode_tag (enc_t *enc, UV tag) |
228 | { |
280 | { |
229 | encode_uint (enc, 0xc0, tag); |
281 | encode_uint (enc, MAJOR_TAG, tag); |
230 | } |
282 | } |
231 | |
283 | |
232 | static void |
284 | // exceptional (hopefully) slow path for byte strings that need to be utf8-encoded |
|
|
285 | ecb_noinline static void |
233 | encode_str (enc_t *enc, int utf8, char *str, STRLEN len) |
286 | encode_str_utf8 (enc_t *enc, int utf8, char *str, STRLEN len) |
234 | { |
287 | { |
|
|
288 | STRLEN ulen = len; |
|
|
289 | U8 *p, *pend = (U8 *)str + len; |
|
|
290 | |
|
|
291 | for (p = (U8 *)str; p < pend; ++p) |
|
|
292 | ulen += *p >> 7; // count set high bits |
|
|
293 | |
|
|
294 | encode_uint (enc, MAJOR_TEXT, ulen); |
|
|
295 | |
|
|
296 | need (enc, ulen); |
|
|
297 | for (p = (U8 *)str; p < pend; ++p) |
|
|
298 | if (*p < 0x80) |
|
|
299 | *enc->cur++ = *p; |
|
|
300 | else |
|
|
301 | { |
|
|
302 | *enc->cur++ = 0xc0 + (*p >> 6); |
|
|
303 | *enc->cur++ = 0x80 + (*p & 63); |
|
|
304 | } |
|
|
305 | } |
|
|
306 | |
|
|
307 | ecb_inline void |
|
|
308 | encode_str (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len) |
|
|
309 | { |
|
|
310 | if (ecb_expect_false (upgrade_utf8)) |
|
|
311 | if (!utf8) |
|
|
312 | { |
|
|
313 | encode_str_utf8 (enc, utf8, str, len); |
|
|
314 | return; |
|
|
315 | } |
|
|
316 | |
|
|
317 | encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len); |
|
|
318 | need (enc, len); |
|
|
319 | memcpy (enc->cur, str, len); |
|
|
320 | enc->cur += len; |
|
|
321 | } |
|
|
322 | |
|
|
323 | ecb_inline void |
|
|
324 | encode_strref (enc_t *enc, int upgrade_utf8, int utf8, char *str, STRLEN len) |
|
|
325 | { |
235 | if (ecb_expect_false (enc->cbor.flags & F_ALLOW_STRINGREF)) |
326 | if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS)) |
236 | { |
327 | { |
237 | SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1); |
328 | SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1); |
238 | |
329 | |
239 | if (SvOK (*svp)) |
330 | if (SvOK (*svp)) |
240 | { |
331 | { |
241 | // already registered, use stringref |
332 | // already registered, use stringref |
242 | encode_tag (enc, CBOR_TAG_STRINGREF); |
333 | encode_tag (enc, CBOR_TAG_STRINGREF); |
243 | encode_uint (enc, 0x00, SvUV (*svp)); |
334 | encode_uint (enc, MAJOR_POS_INT, SvUV (*svp)); |
244 | return; |
335 | return; |
245 | } |
336 | } |
246 | else if (len >= minimum_string_length (enc->stringref_idx)) |
337 | else if (len >= minimum_string_length (enc->stringref_idx)) |
247 | { |
338 | { |
248 | // register only |
339 | // register only |
249 | sv_setuv (*svp, enc->stringref_idx); |
340 | sv_setuv (*svp, enc->stringref_idx); |
250 | ++enc->stringref_idx; |
341 | ++enc->stringref_idx; |
251 | } |
342 | } |
252 | } |
343 | } |
253 | |
344 | |
254 | encode_uint (enc, utf8 ? 0x60 : 0x40, len); |
345 | encode_str (enc, upgrade_utf8, utf8, str, len); |
255 | need (enc, len); |
|
|
256 | memcpy (enc->cur, str, len); |
|
|
257 | enc->cur += len; |
|
|
258 | } |
346 | } |
259 | |
347 | |
260 | static void encode_sv (enc_t *enc, SV *sv); |
348 | static void encode_sv (enc_t *enc, SV *sv); |
261 | |
349 | |
262 | static void |
350 | static void |
… | |
… | |
267 | if (enc->depth >= enc->cbor.max_depth) |
355 | if (enc->depth >= enc->cbor.max_depth) |
268 | croak (ERR_NESTING_EXCEEDED); |
356 | croak (ERR_NESTING_EXCEEDED); |
269 | |
357 | |
270 | ++enc->depth; |
358 | ++enc->depth; |
271 | |
359 | |
272 | encode_uint (enc, 0x80, len + 1); |
360 | encode_uint (enc, MAJOR_ARRAY, len + 1); |
273 | |
361 | |
|
|
362 | if (ecb_expect_false (SvMAGICAL (av))) |
274 | for (i = 0; i <= len; ++i) |
363 | for (i = 0; i <= len; ++i) |
275 | { |
364 | { |
276 | SV **svp = av_fetch (av, i, 0); |
365 | SV **svp = av_fetch (av, i, 0); |
277 | encode_sv (enc, svp ? *svp : &PL_sv_undef); |
366 | encode_sv (enc, svp ? *svp : &PL_sv_undef); |
278 | } |
367 | } |
|
|
368 | else |
|
|
369 | for (i = 0; i <= len; ++i) |
|
|
370 | { |
|
|
371 | SV *sv = AvARRAY (av)[i]; |
|
|
372 | encode_sv (enc, sv ? sv : &PL_sv_undef); |
|
|
373 | } |
279 | |
374 | |
280 | --enc->depth; |
375 | --enc->depth; |
281 | } |
376 | } |
282 | |
377 | |
283 | static void |
378 | static void |
… | |
… | |
291 | ++enc->depth; |
386 | ++enc->depth; |
292 | |
387 | |
293 | int pairs = hv_iterinit (hv); |
388 | int pairs = hv_iterinit (hv); |
294 | int mg = SvMAGICAL (hv); |
389 | int mg = SvMAGICAL (hv); |
295 | |
390 | |
296 | if (mg) |
391 | if (ecb_expect_false (mg)) |
297 | encode_ch (enc, 0xa0 | 31); |
392 | encode_ch (enc, MAJOR_MAP | MINOR_INDEF); |
298 | else |
393 | else |
299 | encode_uint (enc, 0xa0, pairs); |
394 | encode_uint (enc, MAJOR_MAP, pairs); |
300 | |
395 | |
301 | while ((he = hv_iternext (hv))) |
396 | while ((he = hv_iternext (hv))) |
302 | { |
397 | { |
303 | if (HeKLEN (he) == HEf_SVKEY) |
398 | if (HeKLEN (he) == HEf_SVKEY) |
304 | encode_sv (enc, HeSVKEY (he)); |
399 | encode_sv (enc, HeSVKEY (he)); |
305 | else |
400 | else |
306 | encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); |
401 | encode_strref (enc, enc->cbor.flags & (F_TEXT_KEYS | F_TEXT_STRINGS), HeKUTF8 (he), HeKEY (he), HeKLEN (he)); |
307 | |
402 | |
308 | encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); |
403 | encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); |
309 | } |
404 | } |
310 | |
405 | |
311 | if (mg) |
406 | if (ecb_expect_false (mg)) |
312 | encode_ch (enc, 0xe0 | 31); |
407 | encode_ch (enc, MAJOR_MISC | MINOR_INDEF); |
313 | |
408 | |
314 | --enc->depth; |
409 | --enc->depth; |
315 | } |
410 | } |
316 | |
411 | |
317 | // encode objects, arrays and special \0=false and \1=true values. |
412 | // encode objects, arrays and special \0=false and \1=true values. |
318 | static void |
413 | static void |
319 | encode_rv (enc_t *enc, SV *sv) |
414 | encode_rv (enc_t *enc, SV *sv) |
320 | { |
415 | { |
321 | SvGETMAGIC (sv); |
416 | SvGETMAGIC (sv); |
322 | |
|
|
323 | if (ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING) |
|
|
324 | && ecb_expect_false (SvREFCNT (sv) > 1)) |
|
|
325 | { |
|
|
326 | if (!enc->shareable) |
|
|
327 | enc->shareable = (HV *)sv_2mortal ((SV *)newHV ()); |
|
|
328 | |
|
|
329 | SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1); |
|
|
330 | |
|
|
331 | if (SvOK (*svp)) |
|
|
332 | { |
|
|
333 | encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF); |
|
|
334 | encode_uint (enc, 0x00, SvUV (*svp)); |
|
|
335 | return; |
|
|
336 | } |
|
|
337 | else |
|
|
338 | { |
|
|
339 | sv_setuv (*svp, enc->shareable_idx); |
|
|
340 | ++enc->shareable_idx; |
|
|
341 | encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE); |
|
|
342 | } |
|
|
343 | } |
|
|
344 | |
417 | |
345 | svtype svt = SvTYPE (sv); |
418 | svtype svt = SvTYPE (sv); |
346 | |
419 | |
347 | if (ecb_expect_false (SvOBJECT (sv))) |
420 | if (ecb_expect_false (SvOBJECT (sv))) |
348 | { |
421 | { |
… | |
… | |
355 | HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash |
428 | HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash |
356 | ? cbor_tagged_stash |
429 | ? cbor_tagged_stash |
357 | : gv_stashpv ("CBOR::XS::Tagged" , 1); |
430 | : gv_stashpv ("CBOR::XS::Tagged" , 1); |
358 | |
431 | |
359 | HV *stash = SvSTASH (sv); |
432 | HV *stash = SvSTASH (sv); |
360 | GV *method; |
|
|
361 | |
433 | |
362 | if (stash == boolean_stash) |
434 | if (stash == boolean_stash) |
363 | encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20); |
435 | { |
|
|
436 | encode_ch (enc, SvIV (sv) ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE); |
|
|
437 | return; |
|
|
438 | } |
364 | else if (stash == error_stash) |
439 | else if (stash == error_stash) |
365 | encode_ch (enc, 0xe0 | 23); |
440 | { |
|
|
441 | encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF); |
|
|
442 | return; |
|
|
443 | } |
366 | else if (stash == tagged_stash) |
444 | else if (stash == tagged_stash) |
367 | { |
445 | { |
368 | if (svt != SVt_PVAV) |
446 | if (svt != SVt_PVAV) |
369 | croak ("encountered CBOR::XS::Tagged object that isn't an array"); |
447 | croak ("encountered CBOR::XS::Tagged object that isn't an array"); |
370 | |
448 | |
371 | encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); |
449 | encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1))); |
372 | encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); |
450 | encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); |
|
|
451 | |
|
|
452 | return; |
|
|
453 | } |
|
|
454 | } |
|
|
455 | |
|
|
456 | if (ecb_expect_false (SvREFCNT (sv) > 1) |
|
|
457 | && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)) |
|
|
458 | { |
|
|
459 | if (ecb_expect_false (!enc->shareable)) |
|
|
460 | enc->shareable = (HV *)sv_2mortal ((SV *)newHV ()); |
|
|
461 | |
|
|
462 | SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1); |
|
|
463 | |
|
|
464 | if (SvOK (*svp)) |
373 | } |
465 | { |
|
|
466 | encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF); |
|
|
467 | encode_uint (enc, MAJOR_POS_INT, SvUV (*svp)); |
|
|
468 | return; |
|
|
469 | } |
|
|
470 | else |
|
|
471 | { |
|
|
472 | sv_setuv (*svp, enc->shareable_idx); |
|
|
473 | ++enc->shareable_idx; |
|
|
474 | encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE); |
|
|
475 | } |
|
|
476 | } |
|
|
477 | |
|
|
478 | if (ecb_expect_false (SvOBJECT (sv))) |
|
|
479 | { |
|
|
480 | HV *stash = SvSTASH (sv); |
|
|
481 | GV *method; |
|
|
482 | |
|
|
483 | if (enc->cbor.flags & F_FORBID_OBJECTS) |
|
|
484 | croak ("encountered object '%s', but forbid_objects is enabled", |
|
|
485 | SvPV_nolen (sv_2mortal (newRV_inc (sv)))); |
374 | else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) |
486 | else if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) |
375 | { |
487 | { |
376 | dSP; |
488 | dSP; |
377 | |
489 | |
378 | ENTER; SAVETMPS; PUSHMARK (SP); |
490 | ENTER; SAVETMPS; |
|
|
491 | PUSHMARK (SP); |
379 | // we re-bless the reference to get overload and other niceties right |
492 | // we re-bless the reference to get overload and other niceties right |
380 | XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); |
493 | XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); |
381 | |
494 | |
382 | PUTBACK; |
495 | PUTBACK; |
383 | // G_SCALAR ensures that return value is 1 |
496 | // G_SCALAR ensures that return value is 1 |
… | |
… | |
396 | } |
509 | } |
397 | else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0) |
510 | else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0) |
398 | { |
511 | { |
399 | dSP; |
512 | dSP; |
400 | |
513 | |
401 | ENTER; SAVETMPS; PUSHMARK (SP); |
514 | ENTER; SAVETMPS; |
|
|
515 | SAVESTACK_POS (); |
|
|
516 | PUSHMARK (SP); |
402 | EXTEND (SP, 2); |
517 | EXTEND (SP, 2); |
403 | // we re-bless the reference to get overload and other niceties right |
518 | // we re-bless the reference to get overload and other niceties right |
404 | PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); |
519 | PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); |
405 | PUSHs (sv_cbor); |
520 | PUSHs (sv_cbor); |
406 | |
521 | |
… | |
… | |
411 | // catch this surprisingly common error |
526 | // catch this surprisingly common error |
412 | if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) |
527 | if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) |
413 | croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash)); |
528 | croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash)); |
414 | |
529 | |
415 | encode_tag (enc, CBOR_TAG_PERL_OBJECT); |
530 | encode_tag (enc, CBOR_TAG_PERL_OBJECT); |
416 | encode_uint (enc, 0x80, count + 1); |
531 | encode_uint (enc, MAJOR_ARRAY, count + 1); |
417 | encode_str (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); |
532 | encode_strref (enc, 0, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); |
418 | |
533 | |
419 | while (count) |
534 | while (count) |
420 | encode_sv (enc, SP[1 - count--]); |
535 | encode_sv (enc, SP[1 - count--]); |
421 | |
536 | |
422 | PUTBACK; |
537 | PUTBACK; |
… | |
… | |
443 | { |
558 | { |
444 | double nv = SvNVX (sv); |
559 | double nv = SvNVX (sv); |
445 | |
560 | |
446 | need (enc, 9); |
561 | need (enc, 9); |
447 | |
562 | |
448 | if (ecb_expect_false (nv == (U32)nv)) |
563 | if (ecb_expect_false (nv == (NV)(U32)nv)) |
449 | encode_uint (enc, 0x00, (U32)nv); |
564 | encode_uint (enc, MAJOR_POS_INT, (U32)nv); |
450 | //TODO: maybe I32? |
565 | //TODO: maybe I32? |
451 | else if (ecb_expect_false (nv == (float)nv)) |
566 | else if (ecb_expect_false (nv == (float)nv)) |
452 | { |
567 | { |
|
|
568 | *enc->cur++ = MAJOR_MISC | MISC_FLOAT32; |
|
|
569 | |
453 | uint32_t fp = ecb_float_to_binary32 (nv); |
570 | uint32_t fp = ecb_float_to_binary32 (nv); |
454 | |
|
|
455 | *enc->cur++ = 0xe0 | 26; |
|
|
456 | |
571 | |
457 | if (!ecb_big_endian ()) |
572 | if (!ecb_big_endian ()) |
458 | fp = ecb_bswap32 (fp); |
573 | fp = ecb_bswap32 (fp); |
459 | |
574 | |
460 | memcpy (enc->cur, &fp, 4); |
575 | memcpy (enc->cur, &fp, 4); |
461 | enc->cur += 4; |
576 | enc->cur += 4; |
462 | } |
577 | } |
463 | else |
578 | else |
464 | { |
579 | { |
|
|
580 | *enc->cur++ = MAJOR_MISC | MISC_FLOAT64; |
|
|
581 | |
465 | uint64_t fp = ecb_double_to_binary64 (nv); |
582 | uint64_t fp = ecb_double_to_binary64 (nv); |
466 | |
|
|
467 | *enc->cur++ = 0xe0 | 27; |
|
|
468 | |
583 | |
469 | if (!ecb_big_endian ()) |
584 | if (!ecb_big_endian ()) |
470 | fp = ecb_bswap64 (fp); |
585 | fp = ecb_bswap64 (fp); |
471 | |
586 | |
472 | memcpy (enc->cur, &fp, 8); |
587 | memcpy (enc->cur, &fp, 8); |
… | |
… | |
481 | |
596 | |
482 | if (SvPOKp (sv)) |
597 | if (SvPOKp (sv)) |
483 | { |
598 | { |
484 | STRLEN len; |
599 | STRLEN len; |
485 | char *str = SvPV (sv, len); |
600 | char *str = SvPV (sv, len); |
486 | encode_str (enc, SvUTF8 (sv), str, len); |
601 | encode_strref (enc, enc->cbor.flags & F_TEXT_STRINGS, SvUTF8 (sv), str, len); |
487 | } |
602 | } |
488 | else if (SvNOKp (sv)) |
603 | else if (SvNOKp (sv)) |
489 | encode_nv (enc, sv); |
604 | encode_nv (enc, sv); |
490 | else if (SvIOKp (sv)) |
605 | else if (SvIOKp (sv)) |
491 | { |
606 | { |
492 | if (SvIsUV (sv)) |
607 | if (SvIsUV (sv)) |
493 | encode_uint (enc, 0x00, SvUVX (sv)); |
608 | encode_uint (enc, MAJOR_POS_INT, SvUVX (sv)); |
494 | else if (SvIVX (sv) >= 0) |
609 | else if (SvIVX (sv) >= 0) |
495 | encode_uint (enc, 0x00, SvIVX (sv)); |
610 | encode_uint (enc, MAJOR_POS_INT, SvIVX (sv)); |
496 | else |
611 | else |
497 | encode_uint (enc, 0x20, -(SvIVX (sv) + 1)); |
612 | encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1)); |
498 | } |
613 | } |
499 | else if (SvROK (sv)) |
614 | else if (SvROK (sv)) |
500 | encode_rv (enc, SvRV (sv)); |
615 | encode_rv (enc, SvRV (sv)); |
501 | else if (!SvOK (sv)) |
616 | else if (!SvOK (sv)) |
502 | encode_ch (enc, 0xe0 | 22); |
617 | encode_ch (enc, MAJOR_MISC | SIMPLE_NULL); |
503 | else if (enc->cbor.flags & F_ALLOW_UNKNOWN) |
618 | else if (enc->cbor.flags & F_ALLOW_UNKNOWN) |
504 | encode_ch (enc, 0xe0 | 23); |
619 | encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF); |
505 | else |
620 | else |
506 | croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data", |
621 | croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data", |
507 | SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); |
622 | SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); |
508 | } |
623 | } |
509 | |
624 | |
510 | static SV * |
625 | static SV * |
511 | encode_cbor (SV *scalar, CBOR *cbor) |
626 | encode_cbor (SV *scalar, CBOR *cbor) |
512 | { |
627 | { |
513 | enc_t enc = { }; |
628 | enc_t enc = { 0 }; |
514 | |
629 | |
515 | enc.cbor = *cbor; |
630 | enc.cbor = *cbor; |
516 | enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); |
631 | enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); |
517 | enc.cur = SvPVX (enc.sv); |
632 | enc.cur = SvPVX (enc.sv); |
518 | enc.end = SvEND (enc.sv); |
633 | enc.end = SvEND (enc.sv); |
519 | |
634 | |
520 | SvPOK_only (enc.sv); |
635 | SvPOK_only (enc.sv); |
521 | |
636 | |
522 | if (cbor->flags & F_ALLOW_STRINGREF) |
637 | if (cbor->flags & F_PACK_STRINGS) |
523 | { |
638 | { |
524 | encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE); |
639 | encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE); |
525 | enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ()); |
640 | enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ()); |
526 | enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ()); |
641 | enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ()); |
527 | } |
642 | } |
… | |
… | |
550 | U32 depth; // recursion depth |
665 | U32 depth; // recursion depth |
551 | U32 maxdepth; // recursion depth limit |
666 | U32 maxdepth; // recursion depth limit |
552 | AV *shareable; |
667 | AV *shareable; |
553 | AV *stringref; |
668 | AV *stringref; |
554 | SV *decode_tagged; |
669 | SV *decode_tagged; |
|
|
670 | SV *err_sv; // optional sv for error, needs to be freed |
555 | } dec_t; |
671 | } dec_t; |
556 | |
672 | |
557 | #define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE |
673 | // set dec->err to ERRSV |
|
|
674 | ecb_cold static void |
|
|
675 | err_errsv (dec_t *dec) |
|
|
676 | { |
|
|
677 | if (!dec->err) |
|
|
678 | { |
|
|
679 | dec->err_sv = newSVsv (ERRSV); |
558 | |
680 | |
559 | #define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data") |
681 | // chop off the trailing \n |
|
|
682 | SvCUR_set (dec->err_sv, SvCUR (dec->err_sv) - 1); |
|
|
683 | *SvEND (dec->err_sv) = 0; |
560 | |
684 | |
|
|
685 | dec->err = SvPVutf8_nolen (dec->err_sv); |
|
|
686 | } |
|
|
687 | } |
|
|
688 | |
|
|
689 | // the following functions are used to reduce code size and help the compiler to optimise |
|
|
690 | ecb_cold static void |
|
|
691 | err_set (dec_t *dec, const char *reason) |
|
|
692 | { |
|
|
693 | if (!dec->err) |
|
|
694 | dec->err = reason; |
|
|
695 | } |
|
|
696 | |
|
|
697 | ecb_cold static void |
|
|
698 | err_unexpected_end (dec_t *dec) |
|
|
699 | { |
|
|
700 | err_set (dec, "unexpected end of CBOR data"); |
|
|
701 | } |
|
|
702 | |
|
|
703 | #define ERR_DO(do) SB do; goto fail; SE |
|
|
704 | #define ERR(reason) ERR_DO (err_set (dec, reason)) |
|
|
705 | #define ERR_ERRSV ERR_DO (err_errsv (dec)) |
|
|
706 | |
|
|
707 | #define WANT(len) if (ecb_expect_false ((uintptr_t)(dec->end - dec->cur) < (STRLEN)len)) ERR_DO (err_unexpected_end (dec)) |
|
|
708 | |
561 | #define DEC_INC_DEPTH if (++dec->depth > dec->cbor.max_depth) ERR (ERR_NESTING_EXCEEDED) |
709 | #define DEC_INC_DEPTH if (ecb_expect_false (++dec->depth > dec->cbor.max_depth)) ERR (ERR_NESTING_EXCEEDED) |
562 | #define DEC_DEC_DEPTH --dec->depth |
710 | #define DEC_DEC_DEPTH --dec->depth |
563 | |
711 | |
564 | static UV |
712 | static UV |
565 | decode_uint (dec_t *dec) |
713 | decode_uint (dec_t *dec) |
566 | { |
714 | { |
567 | switch (*dec->cur & 31) |
715 | U8 m = *dec->cur & MINOR_MASK; |
568 | { |
716 | ++dec->cur; |
569 | case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7: |
|
|
570 | case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15: |
|
|
571 | case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23: |
|
|
572 | return *dec->cur++ & 31; |
|
|
573 | |
717 | |
574 | case 24: |
718 | if (ecb_expect_true (m < LENGTH_EXT1)) |
|
|
719 | return m; |
|
|
720 | else if (ecb_expect_true (m == LENGTH_EXT1)) |
|
|
721 | { |
575 | WANT (2); |
722 | WANT (1); |
576 | dec->cur += 2; |
723 | dec->cur += 1; |
577 | return dec->cur[-1]; |
724 | return dec->cur[-1]; |
578 | |
725 | } |
579 | case 25: |
726 | else if (ecb_expect_true (m == LENGTH_EXT2)) |
|
|
727 | { |
580 | WANT (3); |
728 | WANT (2); |
581 | dec->cur += 3; |
729 | dec->cur += 2; |
582 | return (((UV)dec->cur[-2]) << 8) |
730 | return (((UV)dec->cur[-2]) << 8) |
583 | | ((UV)dec->cur[-1]); |
731 | | ((UV)dec->cur[-1]); |
584 | |
732 | } |
585 | case 26: |
733 | else if (ecb_expect_true (m == LENGTH_EXT4)) |
|
|
734 | { |
586 | WANT (5); |
735 | WANT (4); |
587 | dec->cur += 5; |
736 | dec->cur += 4; |
588 | return (((UV)dec->cur[-4]) << 24) |
737 | return (((UV)dec->cur[-4]) << 24) |
589 | | (((UV)dec->cur[-3]) << 16) |
738 | | (((UV)dec->cur[-3]) << 16) |
590 | | (((UV)dec->cur[-2]) << 8) |
739 | | (((UV)dec->cur[-2]) << 8) |
591 | | ((UV)dec->cur[-1]); |
740 | | ((UV)dec->cur[-1]); |
592 | |
741 | } |
593 | case 27: |
742 | else if (ecb_expect_true (m == LENGTH_EXT8)) |
|
|
743 | { |
594 | WANT (9); |
744 | WANT (8); |
595 | dec->cur += 9; |
745 | dec->cur += 8; |
|
|
746 | |
|
|
747 | return |
|
|
748 | #if UVSIZE < 8 |
|
|
749 | 0 |
|
|
750 | #else |
596 | return (((UV)dec->cur[-8]) << 56) |
751 | (((UV)dec->cur[-8]) << 56) |
597 | | (((UV)dec->cur[-7]) << 48) |
752 | | (((UV)dec->cur[-7]) << 48) |
598 | | (((UV)dec->cur[-6]) << 40) |
753 | | (((UV)dec->cur[-6]) << 40) |
599 | | (((UV)dec->cur[-5]) << 32) |
754 | | (((UV)dec->cur[-5]) << 32) |
|
|
755 | #endif |
600 | | (((UV)dec->cur[-4]) << 24) |
756 | | (((UV)dec->cur[-4]) << 24) |
601 | | (((UV)dec->cur[-3]) << 16) |
757 | | (((UV)dec->cur[-3]) << 16) |
602 | | (((UV)dec->cur[-2]) << 8) |
758 | | (((UV)dec->cur[-2]) << 8) |
603 | | ((UV)dec->cur[-1]); |
759 | | ((UV)dec->cur[-1]); |
604 | |
760 | } |
605 | default: |
761 | else |
606 | ERR ("corrupted CBOR data (unsupported integer minor encoding)"); |
762 | ERR ("corrupted CBOR data (unsupported integer minor encoding)"); |
607 | } |
|
|
608 | |
763 | |
609 | fail: |
764 | fail: |
610 | return 0; |
765 | return 0; |
611 | } |
766 | } |
612 | |
767 | |
… | |
… | |
617 | { |
772 | { |
618 | AV *av = newAV (); |
773 | AV *av = newAV (); |
619 | |
774 | |
620 | DEC_INC_DEPTH; |
775 | DEC_INC_DEPTH; |
621 | |
776 | |
622 | if ((*dec->cur & 31) == 31) |
777 | if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF)) |
623 | { |
778 | { |
624 | ++dec->cur; |
779 | ++dec->cur; |
625 | |
780 | |
626 | for (;;) |
781 | for (;;) |
627 | { |
782 | { |
628 | WANT (1); |
783 | WANT (1); |
629 | |
784 | |
630 | if (*dec->cur == (0xe0 | 31)) |
785 | if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err) |
631 | { |
786 | { |
632 | ++dec->cur; |
787 | ++dec->cur; |
633 | break; |
788 | break; |
634 | } |
789 | } |
635 | |
790 | |
636 | av_push (av, decode_sv (dec)); |
791 | av_push (av, decode_sv (dec)); |
637 | } |
792 | } |
638 | } |
793 | } |
639 | else |
794 | else |
640 | { |
795 | { |
641 | int i, len = decode_uint (dec); |
796 | UV i, len = decode_uint (dec); |
642 | |
797 | |
|
|
798 | WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays |
643 | av_fill (av, len - 1); |
799 | av_fill (av, len - 1); |
644 | |
800 | |
645 | for (i = 0; i < len; ++i) |
801 | for (i = 0; i < len; ++i) |
646 | AvARRAY (av)[i] = decode_sv (dec); |
802 | AvARRAY (av)[i] = decode_sv (dec); |
647 | } |
803 | } |
648 | |
804 | |
649 | DEC_DEC_DEPTH; |
805 | DEC_DEC_DEPTH; |
650 | return newRV_noinc ((SV *)av); |
806 | return newRV_noinc ((SV *)av); |
651 | |
807 | |
652 | fail: |
808 | fail: |
653 | SvREFCNT_dec (av); |
809 | SvREFCNT_dec_NN (av); |
654 | DEC_DEC_DEPTH; |
810 | DEC_DEC_DEPTH; |
655 | return &PL_sv_undef; |
811 | return &PL_sv_undef; |
656 | } |
812 | } |
657 | |
813 | |
658 | static void |
814 | static void |
… | |
… | |
660 | { |
816 | { |
661 | // for speed reasons, we specialcase single-string |
817 | // for speed reasons, we specialcase single-string |
662 | // byte or utf-8 strings as keys, but only when !stringref |
818 | // byte or utf-8 strings as keys, but only when !stringref |
663 | |
819 | |
664 | if (ecb_expect_true (!dec->stringref)) |
820 | if (ecb_expect_true (!dec->stringref)) |
665 | if (*dec->cur >= 0x40 && *dec->cur <= 0x40 + 27) |
821 | if (ecb_expect_true ((U8)(*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8)) |
666 | { |
822 | { |
667 | I32 len = decode_uint (dec); |
823 | STRLEN len = decode_uint (dec); |
668 | char *key = (char *)dec->cur; |
824 | char *key = (char *)dec->cur; |
669 | |
825 | |
|
|
826 | WANT (len); |
670 | dec->cur += len; |
827 | dec->cur += len; |
671 | |
828 | |
672 | if (ecb_expect_false (dec->stringref)) |
|
|
673 | av_push (dec->stringref, newSVpvn (key, len)); |
|
|
674 | |
|
|
675 | hv_store (hv, key, len, decode_sv (dec), 0); |
829 | hv_store (hv, key, len, decode_sv (dec), 0); |
676 | |
830 | |
677 | return; |
831 | return; |
678 | } |
832 | } |
679 | else if (*dec->cur >= 0x60 && *dec->cur <= 0x60 + 27) |
833 | else if (ecb_expect_true ((U8)(*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8)) |
680 | { |
834 | { |
681 | I32 len = decode_uint (dec); |
835 | STRLEN len = decode_uint (dec); |
682 | char *key = (char *)dec->cur; |
836 | char *key = (char *)dec->cur; |
683 | |
837 | |
|
|
838 | WANT (len); |
684 | dec->cur += len; |
839 | dec->cur += len; |
685 | |
840 | |
686 | if (ecb_expect_false (dec->stringref)) |
841 | if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8)) |
687 | av_push (dec->stringref, newSVpvn_utf8 (key, len, 1)); |
842 | if (!is_utf8_string (key, len)) |
|
|
843 | ERR ("corrupted CBOR data (invalid UTF-8 in map key)"); |
688 | |
844 | |
689 | hv_store (hv, key, -len, decode_sv (dec), 0); |
845 | hv_store (hv, key, -len, decode_sv (dec), 0); |
690 | |
846 | |
691 | return; |
847 | return; |
692 | } |
848 | } |
693 | |
849 | |
694 | SV *k = decode_sv (dec); |
850 | SV *k = decode_sv (dec); |
695 | SV *v = decode_sv (dec); |
851 | SV *v = decode_sv (dec); |
696 | |
852 | |
|
|
853 | // we leak memory if uncaught exceptions are thrown by random magical |
|
|
854 | // methods, and this is hopefully the only place where it can happen, |
|
|
855 | // so if there is a chance of an exception, take the very slow path. |
|
|
856 | // since catching exceptions is "undocumented/internal/forbidden" by |
|
|
857 | // the new p5p powers, we need to call out to a perl function :/ |
|
|
858 | if (ecb_expect_false (SvAMAGIC (k))) |
|
|
859 | { |
|
|
860 | dSP; |
|
|
861 | |
|
|
862 | ENTER; SAVETMPS; |
|
|
863 | PUSHMARK (SP); |
|
|
864 | EXTEND (SP, 3); |
|
|
865 | PUSHs (sv_2mortal (newRV_inc ((SV *)hv))); |
|
|
866 | PUSHs (sv_2mortal (k)); |
|
|
867 | PUSHs (sv_2mortal (v)); |
|
|
868 | |
|
|
869 | PUTBACK; |
|
|
870 | call_pv ("CBOR::XS::_hv_store", G_VOID | G_DISCARD | G_EVAL); |
|
|
871 | SPAGAIN; |
|
|
872 | |
|
|
873 | FREETMPS; LEAVE; |
|
|
874 | |
|
|
875 | if (SvTRUE (ERRSV)) |
|
|
876 | ERR_ERRSV; |
|
|
877 | |
|
|
878 | return; |
|
|
879 | } |
|
|
880 | |
697 | hv_store_ent (hv, k, v, 0); |
881 | hv_store_ent (hv, k, v, 0); |
698 | SvREFCNT_dec (k); |
882 | SvREFCNT_dec_NN (k); |
|
|
883 | |
|
|
884 | fail: |
|
|
885 | ; |
699 | } |
886 | } |
700 | |
887 | |
701 | static SV * |
888 | static SV * |
702 | decode_hv (dec_t *dec) |
889 | decode_hv (dec_t *dec) |
703 | { |
890 | { |
704 | HV *hv = newHV (); |
891 | HV *hv = newHV (); |
705 | |
892 | |
706 | DEC_INC_DEPTH; |
893 | DEC_INC_DEPTH; |
707 | |
894 | |
708 | if ((*dec->cur & 31) == 31) |
895 | if (*dec->cur == (MAJOR_MAP | MINOR_INDEF)) |
709 | { |
896 | { |
710 | ++dec->cur; |
897 | ++dec->cur; |
711 | |
898 | |
712 | for (;;) |
899 | for (;;) |
713 | { |
900 | { |
714 | WANT (1); |
901 | WANT (1); |
715 | |
902 | |
716 | if (*dec->cur == (0xe0 | 31)) |
903 | if (*dec->cur == (MAJOR_MISC | MINOR_INDEF) || dec->err) |
717 | { |
904 | { |
718 | ++dec->cur; |
905 | ++dec->cur; |
719 | break; |
906 | break; |
720 | } |
907 | } |
721 | |
908 | |
722 | decode_he (dec, hv); |
909 | decode_he (dec, hv); |
723 | } |
910 | } |
724 | } |
911 | } |
725 | else |
912 | else |
726 | { |
913 | { |
727 | int pairs = decode_uint (dec); |
914 | UV pairs = decode_uint (dec); |
|
|
915 | |
|
|
916 | WANT (pairs); // complexity check - need at least one byte per value, do not allow supersize hashes |
728 | |
917 | |
729 | while (pairs--) |
918 | while (pairs--) |
730 | decode_he (dec, hv); |
919 | decode_he (dec, hv); |
731 | } |
920 | } |
732 | |
921 | |
733 | DEC_DEC_DEPTH; |
922 | DEC_DEC_DEPTH; |
734 | return newRV_noinc ((SV *)hv); |
923 | return newRV_noinc ((SV *)hv); |
735 | |
924 | |
736 | fail: |
925 | fail: |
737 | SvREFCNT_dec (hv); |
926 | SvREFCNT_dec_NN (hv); |
738 | DEC_DEC_DEPTH; |
927 | DEC_DEC_DEPTH; |
739 | return &PL_sv_undef; |
928 | return &PL_sv_undef; |
740 | } |
929 | } |
741 | |
930 | |
742 | static SV * |
931 | static SV * |
743 | decode_str (dec_t *dec, int utf8) |
932 | decode_str (dec_t *dec, int utf8) |
744 | { |
933 | { |
745 | SV *sv = 0; |
934 | SV *sv = 0; |
746 | |
935 | |
747 | if ((*dec->cur & 31) == 31) |
936 | if (ecb_expect_false ((*dec->cur & MINOR_MASK) == MINOR_INDEF)) |
748 | { |
937 | { |
|
|
938 | // indefinite length strings |
749 | ++dec->cur; |
939 | ++dec->cur; |
750 | |
940 | |
|
|
941 | U8 major = *dec->cur & MAJOR_MISC; |
|
|
942 | |
751 | sv = newSVpvn ("", 0); |
943 | sv = newSVpvn ("", 0); |
752 | |
944 | |
753 | // not very fast, and certainly not robust against illegal input |
|
|
754 | for (;;) |
945 | for (;;) |
755 | { |
946 | { |
756 | WANT (1); |
947 | WANT (1); |
757 | |
948 | |
758 | if (*dec->cur == (0xe0 | 31)) |
949 | if ((*dec->cur - major) > LENGTH_EXT8) |
|
|
950 | if (*dec->cur == (MAJOR_MISC | MINOR_INDEF)) |
759 | { |
951 | { |
760 | ++dec->cur; |
952 | ++dec->cur; |
761 | break; |
953 | break; |
762 | } |
954 | } |
|
|
955 | else |
|
|
956 | ERR ("corrupted CBOR data (invalid chunks in indefinite length string)"); |
763 | |
957 | |
764 | sv_catsv (sv, decode_sv (dec)); |
958 | STRLEN len = decode_uint (dec); |
|
|
959 | |
|
|
960 | WANT (len); |
|
|
961 | sv_catpvn (sv, dec->cur, len); |
|
|
962 | dec->cur += len; |
765 | } |
963 | } |
766 | } |
964 | } |
767 | else |
965 | else |
768 | { |
966 | { |
769 | STRLEN len = decode_uint (dec); |
967 | STRLEN len = decode_uint (dec); |
… | |
… | |
776 | && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1)) |
974 | && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1)) |
777 | av_push (dec->stringref, SvREFCNT_inc_NN (sv)); |
975 | av_push (dec->stringref, SvREFCNT_inc_NN (sv)); |
778 | } |
976 | } |
779 | |
977 | |
780 | if (utf8) |
978 | if (utf8) |
|
|
979 | { |
|
|
980 | if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8)) |
|
|
981 | if (!is_utf8_string (SvPVX (sv), SvCUR (sv))) |
|
|
982 | ERR ("corrupted CBOR data (invalid UTF-8 in text string)"); |
|
|
983 | |
781 | SvUTF8_on (sv); |
984 | SvUTF8_on (sv); |
|
|
985 | } |
782 | |
986 | |
783 | return sv; |
987 | return sv; |
784 | |
988 | |
785 | fail: |
989 | fail: |
786 | SvREFCNT_dec (sv); |
990 | SvREFCNT_dec (sv); |
… | |
… | |
805 | sv = newRV_noinc (decode_sv (dec)); |
1009 | sv = newRV_noinc (decode_sv (dec)); |
806 | break; |
1010 | break; |
807 | |
1011 | |
808 | case CBOR_TAG_STRINGREF_NAMESPACE: |
1012 | case CBOR_TAG_STRINGREF_NAMESPACE: |
809 | { |
1013 | { |
|
|
1014 | // do not use SAVETMPS/FREETMPS, as these will |
|
|
1015 | // erase mortalised caches, e.g. "shareable" |
810 | ENTER; SAVETMPS; |
1016 | ENTER; |
811 | |
1017 | |
812 | SAVESPTR (dec->stringref); |
1018 | SAVESPTR (dec->stringref); |
813 | dec->stringref = (AV *)sv_2mortal ((SV *)newAV ()); |
1019 | dec->stringref = (AV *)sv_2mortal ((SV *)newAV ()); |
814 | |
1020 | |
815 | sv = decode_sv (dec); |
1021 | sv = decode_sv (dec); |
816 | |
1022 | |
817 | FREETMPS; LEAVE; |
1023 | LEAVE; |
818 | } |
1024 | } |
819 | break; |
1025 | break; |
820 | |
1026 | |
821 | case CBOR_TAG_STRINGREF: |
1027 | case CBOR_TAG_STRINGREF: |
822 | { |
1028 | { |
823 | if ((*dec->cur >> 5) != 0) |
1029 | if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT)) |
824 | ERR ("corrupted CBOR data (stringref index not an unsigned integer)"); |
1030 | ERR ("corrupted CBOR data (stringref index not an unsigned integer)"); |
825 | |
1031 | |
826 | UV idx = decode_uint (dec); |
1032 | UV idx = decode_uint (dec); |
827 | |
1033 | |
828 | if (!dec->stringref || (int)idx > AvFILLp (dec->stringref)) |
1034 | if (!dec->stringref || idx >= (UV)(1 + AvFILLp (dec->stringref))) |
829 | ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)"); |
1035 | ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)"); |
830 | |
1036 | |
831 | sv = newSVsv (AvARRAY (dec->stringref)[idx]); |
1037 | sv = newSVsv (AvARRAY (dec->stringref)[idx]); |
832 | } |
1038 | } |
833 | break; |
1039 | break; |
… | |
… | |
835 | case CBOR_TAG_VALUE_SHAREABLE: |
1041 | case CBOR_TAG_VALUE_SHAREABLE: |
836 | { |
1042 | { |
837 | if (ecb_expect_false (!dec->shareable)) |
1043 | if (ecb_expect_false (!dec->shareable)) |
838 | dec->shareable = (AV *)sv_2mortal ((SV *)newAV ()); |
1044 | dec->shareable = (AV *)sv_2mortal ((SV *)newAV ()); |
839 | |
1045 | |
|
|
1046 | if (dec->cbor.flags & F_ALLOW_CYCLES) |
|
|
1047 | { |
840 | sv = newSV (0); |
1048 | sv = newSV (0); |
841 | av_push (dec->shareable, SvREFCNT_inc_NN (sv)); |
1049 | av_push (dec->shareable, SvREFCNT_inc_NN (sv)); |
842 | |
1050 | |
843 | SV *osv = decode_sv (dec); |
1051 | SV *osv = decode_sv (dec); |
844 | sv_setsv (sv, osv); |
1052 | sv_setsv (sv, osv); |
845 | SvREFCNT_dec_NN (osv); |
1053 | SvREFCNT_dec_NN (osv); |
|
|
1054 | } |
|
|
1055 | else |
|
|
1056 | { |
|
|
1057 | av_push (dec->shareable, &PL_sv_undef); |
|
|
1058 | int idx = AvFILLp (dec->shareable); |
|
|
1059 | sv = decode_sv (dec); |
|
|
1060 | av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv)); |
|
|
1061 | } |
846 | } |
1062 | } |
847 | break; |
1063 | break; |
848 | |
1064 | |
849 | case CBOR_TAG_VALUE_SHAREDREF: |
1065 | case CBOR_TAG_VALUE_SHAREDREF: |
850 | { |
1066 | { |
851 | if ((*dec->cur >> 5) != 0) |
1067 | if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT)) |
852 | ERR ("corrupted CBOR data (sharedref index not an unsigned integer)"); |
1068 | ERR ("corrupted CBOR data (sharedref index not an unsigned integer)"); |
853 | |
1069 | |
854 | UV idx = decode_uint (dec); |
1070 | UV idx = decode_uint (dec); |
855 | |
1071 | |
856 | if (!dec->shareable || (int)idx > AvFILLp (dec->shareable)) |
1072 | if (!dec->shareable || idx >= (UV)(1 + AvFILLp (dec->shareable))) |
857 | ERR ("corrupted CBOR data (sharedref index out of bounds)"); |
1073 | ERR ("corrupted CBOR data (sharedref index out of bounds)"); |
858 | |
1074 | |
859 | sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]); |
1075 | sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]); |
|
|
1076 | |
|
|
1077 | if (sv == &PL_sv_undef) |
|
|
1078 | ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled"); |
860 | } |
1079 | } |
861 | break; |
1080 | break; |
862 | |
1081 | |
863 | case CBOR_TAG_PERL_OBJECT: |
1082 | case CBOR_TAG_PERL_OBJECT: |
864 | { |
1083 | { |
|
|
1084 | if (dec->cbor.flags & F_FORBID_OBJECTS) |
|
|
1085 | goto filter; |
|
|
1086 | |
865 | sv = decode_sv (dec); |
1087 | sv = decode_sv (dec); |
866 | |
1088 | |
867 | if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) |
1089 | if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) |
868 | ERR ("corrupted CBOR data (non-array perl object)"); |
1090 | ERR ("corrupted CBOR data (non-array perl object)"); |
869 | |
1091 | |
… | |
… | |
879 | if (!method) |
1101 | if (!method) |
880 | ERR ("cannot decode perl-object (package does not have a THAW method)"); |
1102 | ERR ("cannot decode perl-object (package does not have a THAW method)"); |
881 | |
1103 | |
882 | dSP; |
1104 | dSP; |
883 | |
1105 | |
884 | ENTER; SAVETMPS; PUSHMARK (SP); |
1106 | ENTER; SAVETMPS; |
|
|
1107 | PUSHMARK (SP); |
885 | EXTEND (SP, len + 1); |
1108 | EXTEND (SP, len + 1); |
886 | // we re-bless the reference to get overload and other niceties right |
1109 | // we re-bless the reference to get overload and other niceties right |
887 | PUSHs (*av_fetch (av, 0, 1)); |
1110 | PUSHs (*av_fetch (av, 0, 1)); |
888 | PUSHs (sv_cbor); |
1111 | PUSHs (sv_cbor); |
889 | |
1112 | |
… | |
… | |
897 | SPAGAIN; |
1120 | SPAGAIN; |
898 | |
1121 | |
899 | if (SvTRUE (ERRSV)) |
1122 | if (SvTRUE (ERRSV)) |
900 | { |
1123 | { |
901 | FREETMPS; LEAVE; |
1124 | FREETMPS; LEAVE; |
902 | ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV)))); |
1125 | ERR_ERRSV; |
903 | } |
1126 | } |
904 | |
1127 | |
905 | SvREFCNT_dec (sv); |
1128 | SvREFCNT_dec_NN (sv); |
906 | sv = SvREFCNT_inc (POPs); |
1129 | sv = SvREFCNT_inc (POPs); |
907 | |
1130 | |
908 | PUTBACK; |
1131 | PUTBACK; |
909 | |
1132 | |
910 | FREETMPS; LEAVE; |
1133 | FREETMPS; LEAVE; |
911 | } |
1134 | } |
912 | break; |
1135 | break; |
913 | |
1136 | |
914 | default: |
1137 | default: |
|
|
1138 | filter: |
915 | { |
1139 | { |
|
|
1140 | SV *tag_sv = newSVuv (tag); |
|
|
1141 | |
916 | sv = decode_sv (dec); |
1142 | sv = decode_sv (dec); |
917 | |
1143 | |
918 | dSP; |
1144 | dSP; |
919 | ENTER; SAVETMPS; PUSHMARK (SP); |
1145 | ENTER; SAVETMPS; |
|
|
1146 | SAVESTACK_POS (); |
|
|
1147 | PUSHMARK (SP); |
920 | EXTEND (SP, 2); |
1148 | EXTEND (SP, 2); |
921 | PUSHs (newSVuv (tag)); |
1149 | PUSHs (tag_sv); |
922 | PUSHs (sv); |
1150 | PUSHs (sv); |
923 | |
1151 | |
924 | PUTBACK; |
1152 | PUTBACK; |
925 | int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL); |
1153 | int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL); |
926 | SPAGAIN; |
1154 | SPAGAIN; |
927 | |
1155 | |
928 | if (SvTRUE (ERRSV)) |
1156 | if (SvTRUE (ERRSV)) |
929 | { |
1157 | { |
|
|
1158 | SvREFCNT_dec_NN (tag_sv); |
930 | FREETMPS; LEAVE; |
1159 | FREETMPS; LEAVE; |
931 | ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV)))); |
1160 | ERR_ERRSV; |
932 | } |
1161 | } |
933 | |
1162 | |
934 | if (count) |
1163 | if (count) |
935 | { |
1164 | { |
|
|
1165 | SvREFCNT_dec_NN (tag_sv); |
936 | SvREFCNT_dec (sv); |
1166 | SvREFCNT_dec_NN (sv); |
937 | sv = SvREFCNT_inc (POPs); |
1167 | sv = SvREFCNT_inc_NN (POPs); |
938 | } |
1168 | } |
939 | else |
1169 | else |
940 | { |
1170 | { |
941 | AV *av = newAV (); |
1171 | AV *av = newAV (); |
942 | av_push (av, newSVuv (tag)); |
1172 | av_push (av, tag_sv); |
943 | av_push (av, sv); |
1173 | av_push (av, sv); |
944 | |
1174 | |
945 | HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash |
1175 | HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash |
946 | ? cbor_tagged_stash |
1176 | ? cbor_tagged_stash |
947 | : gv_stashpv ("CBOR::XS::Tagged" , 1); |
1177 | : gv_stashpv ("CBOR::XS::Tagged" , 1); |
… | |
… | |
965 | static SV * |
1195 | static SV * |
966 | decode_sv (dec_t *dec) |
1196 | decode_sv (dec_t *dec) |
967 | { |
1197 | { |
968 | WANT (1); |
1198 | WANT (1); |
969 | |
1199 | |
970 | switch (*dec->cur >> 5) |
1200 | switch (*dec->cur >> MAJOR_SHIFT) |
971 | { |
1201 | { |
972 | case 0: // unsigned int |
1202 | case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec)); |
973 | return newSVuv (decode_uint (dec)); |
1203 | case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec)); |
974 | case 1: // negative int |
1204 | case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0); |
975 | return newSViv (-1 - (IV)decode_uint (dec)); |
1205 | case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1); |
976 | case 2: // octet string |
1206 | case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec); |
977 | return decode_str (dec, 0); |
1207 | case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec); |
978 | case 3: // utf-8 string |
1208 | case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec); |
979 | return decode_str (dec, 1); |
1209 | |
980 | case 4: // array |
1210 | case MAJOR_MISC >> MAJOR_SHIFT: |
981 | return decode_av (dec); |
|
|
982 | case 5: // map |
|
|
983 | return decode_hv (dec); |
|
|
984 | case 6: // tag |
|
|
985 | return decode_tagged (dec); |
|
|
986 | case 7: // misc |
|
|
987 | switch (*dec->cur++ & 31) |
1211 | switch (*dec->cur++ & MINOR_MASK) |
988 | { |
1212 | { |
989 | case 20: |
1213 | case SIMPLE_FALSE: |
990 | #if CBOR_SLOW |
1214 | #if CBOR_SLOW |
991 | types_false = get_bool ("Types::Serialiser::false"); |
1215 | types_false = get_bool ("Types::Serialiser::false"); |
992 | #endif |
1216 | #endif |
993 | return newSVsv (types_false); |
1217 | return newSVsv (types_false); |
994 | case 21: |
1218 | case SIMPLE_TRUE: |
995 | #if CBOR_SLOW |
1219 | #if CBOR_SLOW |
996 | types_true = get_bool ("Types::Serialiser::true"); |
1220 | types_true = get_bool ("Types::Serialiser::true"); |
997 | #endif |
1221 | #endif |
998 | return newSVsv (types_true); |
1222 | return newSVsv (types_true); |
999 | case 22: |
1223 | case SIMPLE_NULL: |
1000 | return newSVsv (&PL_sv_undef); |
1224 | return newSVsv (&PL_sv_undef); |
1001 | case 23: |
1225 | case SIMPLE_UNDEF: |
1002 | #if CBOR_SLOW |
1226 | #if CBOR_SLOW |
1003 | types_error = get_bool ("Types::Serialiser::error"); |
1227 | types_error = get_bool ("Types::Serialiser::error"); |
1004 | #endif |
1228 | #endif |
1005 | return newSVsv (types_error); |
1229 | return newSVsv (types_error); |
1006 | |
1230 | |
1007 | case 25: |
1231 | case MISC_FLOAT16: |
1008 | { |
1232 | { |
1009 | WANT (2); |
1233 | WANT (2); |
1010 | |
1234 | |
1011 | uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; |
1235 | uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; |
1012 | dec->cur += 2; |
1236 | dec->cur += 2; |
1013 | |
1237 | |
1014 | return newSVnv (ecb_binary16_to_float (fp)); |
1238 | return newSVnv (ecb_binary16_to_float (fp)); |
1015 | } |
1239 | } |
1016 | |
1240 | |
1017 | case 26: |
1241 | case MISC_FLOAT32: |
1018 | { |
1242 | { |
1019 | uint32_t fp; |
1243 | uint32_t fp; |
1020 | WANT (4); |
1244 | WANT (4); |
1021 | memcpy (&fp, dec->cur, 4); |
1245 | memcpy (&fp, dec->cur, 4); |
1022 | dec->cur += 4; |
1246 | dec->cur += 4; |
… | |
… | |
1025 | fp = ecb_bswap32 (fp); |
1249 | fp = ecb_bswap32 (fp); |
1026 | |
1250 | |
1027 | return newSVnv (ecb_binary32_to_float (fp)); |
1251 | return newSVnv (ecb_binary32_to_float (fp)); |
1028 | } |
1252 | } |
1029 | |
1253 | |
1030 | case 27: |
1254 | case MISC_FLOAT64: |
1031 | { |
1255 | { |
1032 | uint64_t fp; |
1256 | uint64_t fp; |
1033 | WANT (8); |
1257 | WANT (8); |
1034 | memcpy (&fp, dec->cur, 8); |
1258 | memcpy (&fp, dec->cur, 8); |
1035 | dec->cur += 8; |
1259 | dec->cur += 8; |
… | |
… | |
1038 | fp = ecb_bswap64 (fp); |
1262 | fp = ecb_bswap64 (fp); |
1039 | |
1263 | |
1040 | return newSVnv (ecb_binary64_to_double (fp)); |
1264 | return newSVnv (ecb_binary64_to_double (fp)); |
1041 | } |
1265 | } |
1042 | |
1266 | |
1043 | // 0..19 unassigned |
1267 | // 0..19 unassigned simple |
1044 | // 24 reserved + unassigned (reserved values are not encodable) |
1268 | // 24 reserved + unassigned simple (reserved values are not encodable) |
|
|
1269 | // 28-30 unassigned misc |
|
|
1270 | // 31 break code |
1045 | default: |
1271 | default: |
1046 | ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); |
1272 | ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)"); |
1047 | } |
1273 | } |
1048 | |
1274 | |
1049 | break; |
1275 | break; |
1050 | } |
1276 | } |
1051 | |
1277 | |
… | |
… | |
1054 | } |
1280 | } |
1055 | |
1281 | |
1056 | static SV * |
1282 | static SV * |
1057 | decode_cbor (SV *string, CBOR *cbor, char **offset_return) |
1283 | decode_cbor (SV *string, CBOR *cbor, char **offset_return) |
1058 | { |
1284 | { |
1059 | dec_t dec = { }; |
1285 | dec_t dec = { 0 }; |
1060 | SV *sv; |
1286 | SV *sv; |
1061 | STRLEN len; |
1287 | STRLEN len; |
1062 | char *data = SvPVbyte (string, len); |
1288 | char *data = SvPVbyte (string, len); |
1063 | |
1289 | |
1064 | if (len > cbor->max_size && cbor->max_size) |
1290 | if (len > cbor->max_size && cbor->max_size) |
… | |
… | |
1078 | if (dec.cur != dec.end && !dec.err) |
1304 | if (dec.cur != dec.end && !dec.err) |
1079 | dec.err = "garbage after CBOR object"; |
1305 | dec.err = "garbage after CBOR object"; |
1080 | |
1306 | |
1081 | if (dec.err) |
1307 | if (dec.err) |
1082 | { |
1308 | { |
|
|
1309 | if (dec.shareable) |
|
|
1310 | { |
|
|
1311 | // need to break cyclic links, which would all be in shareable |
|
|
1312 | int i; |
|
|
1313 | SV **svp; |
|
|
1314 | |
|
|
1315 | for (i = av_len (dec.shareable) + 1; i--; ) |
|
|
1316 | if ((svp = av_fetch (dec.shareable, i, 0))) |
|
|
1317 | sv_setsv (*svp, &PL_sv_undef); |
|
|
1318 | } |
|
|
1319 | |
1083 | SvREFCNT_dec (sv); |
1320 | SvREFCNT_dec_NN (sv); |
|
|
1321 | |
|
|
1322 | if (dec.err_sv) |
|
|
1323 | sv_2mortal (dec.err_sv); |
|
|
1324 | |
1084 | croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur); |
1325 | croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur); |
1085 | } |
1326 | } |
1086 | |
1327 | |
1087 | sv = sv_2mortal (sv); |
1328 | sv = sv_2mortal (sv); |
1088 | |
1329 | |
1089 | return sv; |
1330 | return sv; |
1090 | } |
1331 | } |
1091 | |
1332 | |
|
|
1333 | ///////////////////////////////////////////////////////////////////////////// |
|
|
1334 | // incremental parser |
|
|
1335 | |
|
|
1336 | #define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0) |
|
|
1337 | |
|
|
1338 | // returns 0 for notyet, 1 for success or error |
|
|
1339 | static int |
|
|
1340 | incr_parse (CBOR *self, SV *cborstr) |
|
|
1341 | { |
|
|
1342 | STRLEN cur; |
|
|
1343 | SvPV (cborstr, cur); |
|
|
1344 | |
|
|
1345 | while (ecb_expect_true (self->incr_need <= cur)) |
|
|
1346 | { |
|
|
1347 | // table of integer count bytes |
|
|
1348 | static I8 incr_len[MINOR_MASK + 1] = { |
|
|
1349 | 0, 0, 0, 0, 0, 0, 0, 0, |
|
|
1350 | 0, 0, 0, 0, 0, 0, 0, 0, |
|
|
1351 | 0, 0, 0, 0, 0, 0, 0, 0, |
|
|
1352 | 1, 2, 4, 8,-1,-1,-1,-2 |
|
|
1353 | }; |
|
|
1354 | |
|
|
1355 | const U8 *p = SvPVX (cborstr) + self->incr_pos; |
|
|
1356 | U8 m = *p & MINOR_MASK; |
|
|
1357 | IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]); |
|
|
1358 | I8 ilen = incr_len[m]; |
|
|
1359 | |
|
|
1360 | self->incr_need = self->incr_pos + 1; |
|
|
1361 | |
|
|
1362 | if (ecb_expect_false (ilen < 0)) |
|
|
1363 | { |
|
|
1364 | if (m != MINOR_INDEF) |
|
|
1365 | return 1; // error |
|
|
1366 | |
|
|
1367 | if (*p == (MAJOR_MISC | MINOR_INDEF)) |
|
|
1368 | { |
|
|
1369 | if (count >= 0) |
|
|
1370 | return 1; // error |
|
|
1371 | |
|
|
1372 | count = 1; |
|
|
1373 | } |
|
|
1374 | else |
|
|
1375 | { |
|
|
1376 | av_push (self->incr_count, newSViv (-1)); //TODO: nest |
|
|
1377 | count = -1; |
|
|
1378 | } |
|
|
1379 | } |
|
|
1380 | else |
|
|
1381 | { |
|
|
1382 | self->incr_need += ilen; |
|
|
1383 | if (ecb_expect_false (self->incr_need > cur)) |
|
|
1384 | return 0; |
|
|
1385 | |
|
|
1386 | int major = *p >> MAJOR_SHIFT; |
|
|
1387 | |
|
|
1388 | switch (major) |
|
|
1389 | { |
|
|
1390 | case MAJOR_TAG >> MAJOR_SHIFT: |
|
|
1391 | ++count; // tags merely prefix another value |
|
|
1392 | break; |
|
|
1393 | |
|
|
1394 | case MAJOR_BYTES >> MAJOR_SHIFT: |
|
|
1395 | case MAJOR_TEXT >> MAJOR_SHIFT: |
|
|
1396 | case MAJOR_ARRAY >> MAJOR_SHIFT: |
|
|
1397 | case MAJOR_MAP >> MAJOR_SHIFT: |
|
|
1398 | { |
|
|
1399 | UV len; |
|
|
1400 | |
|
|
1401 | if (ecb_expect_false (ilen)) |
|
|
1402 | { |
|
|
1403 | len = 0; |
|
|
1404 | |
|
|
1405 | do { |
|
|
1406 | len = (len << 8) | *++p; |
|
|
1407 | } while (--ilen); |
|
|
1408 | } |
|
|
1409 | else |
|
|
1410 | len = m; |
|
|
1411 | |
|
|
1412 | switch (major) |
|
|
1413 | { |
|
|
1414 | case MAJOR_BYTES >> MAJOR_SHIFT: |
|
|
1415 | case MAJOR_TEXT >> MAJOR_SHIFT: |
|
|
1416 | self->incr_need += len; |
|
|
1417 | if (ecb_expect_false (self->incr_need > cur)) |
|
|
1418 | return 0; |
|
|
1419 | |
|
|
1420 | break; |
|
|
1421 | |
|
|
1422 | case MAJOR_MAP >> MAJOR_SHIFT: |
|
|
1423 | len <<= 1; |
|
|
1424 | case MAJOR_ARRAY >> MAJOR_SHIFT: |
|
|
1425 | if (len) |
|
|
1426 | { |
|
|
1427 | av_push (self->incr_count, newSViv (len + 1)); //TODO: nest |
|
|
1428 | count = len + 1; |
|
|
1429 | } |
|
|
1430 | break; |
|
|
1431 | } |
|
|
1432 | } |
|
|
1433 | } |
|
|
1434 | } |
|
|
1435 | |
|
|
1436 | self->incr_pos = self->incr_need; |
|
|
1437 | |
|
|
1438 | if (count > 0) |
|
|
1439 | { |
|
|
1440 | while (!--count) |
|
|
1441 | { |
|
|
1442 | if (!AvFILLp (self->incr_count)) |
|
|
1443 | return 1; // done |
|
|
1444 | |
|
|
1445 | SvREFCNT_dec_NN (av_pop (self->incr_count)); |
|
|
1446 | count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]); |
|
|
1447 | } |
|
|
1448 | |
|
|
1449 | SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count; |
|
|
1450 | } |
|
|
1451 | } |
|
|
1452 | |
|
|
1453 | return 0; |
|
|
1454 | } |
|
|
1455 | |
|
|
1456 | |
1092 | ///////////////////////////////////////////////////////////////////////////// |
1457 | ///////////////////////////////////////////////////////////////////////////// |
1093 | // XS interface functions |
1458 | // XS interface functions |
1094 | |
1459 | |
1095 | MODULE = CBOR::XS PACKAGE = CBOR::XS |
1460 | MODULE = CBOR::XS PACKAGE = CBOR::XS |
1096 | |
1461 | |
… | |
… | |
1108 | |
1473 | |
1109 | default_filter = newSVpv ("CBOR::XS::default_filter", 0); |
1474 | default_filter = newSVpv ("CBOR::XS::default_filter", 0); |
1110 | |
1475 | |
1111 | sv_cbor = newSVpv ("CBOR", 0); |
1476 | sv_cbor = newSVpv ("CBOR", 0); |
1112 | SvREADONLY_on (sv_cbor); |
1477 | SvREADONLY_on (sv_cbor); |
|
|
1478 | |
|
|
1479 | assert (("STRLEN must be an unsigned type", 0 <= (STRLEN)-1)); |
1113 | } |
1480 | } |
1114 | |
1481 | |
1115 | PROTOTYPES: DISABLE |
1482 | PROTOTYPES: DISABLE |
1116 | |
1483 | |
1117 | void CLONE (...) |
1484 | void CLONE (...) |
… | |
… | |
1136 | void shrink (CBOR *self, int enable = 1) |
1503 | void shrink (CBOR *self, int enable = 1) |
1137 | ALIAS: |
1504 | ALIAS: |
1138 | shrink = F_SHRINK |
1505 | shrink = F_SHRINK |
1139 | allow_unknown = F_ALLOW_UNKNOWN |
1506 | allow_unknown = F_ALLOW_UNKNOWN |
1140 | allow_sharing = F_ALLOW_SHARING |
1507 | allow_sharing = F_ALLOW_SHARING |
1141 | allow_stringref = F_ALLOW_STRINGREF |
1508 | allow_cycles = F_ALLOW_CYCLES |
|
|
1509 | forbid_objects = F_FORBID_OBJECTS |
|
|
1510 | pack_strings = F_PACK_STRINGS |
|
|
1511 | text_keys = F_TEXT_KEYS |
|
|
1512 | text_strings = F_TEXT_STRINGS |
|
|
1513 | validate_utf8 = F_VALIDATE_UTF8 |
1142 | PPCODE: |
1514 | PPCODE: |
1143 | { |
1515 | { |
1144 | if (enable) |
1516 | if (enable) |
1145 | self->flags |= ix; |
1517 | self->flags |= ix; |
1146 | else |
1518 | else |
… | |
… | |
1152 | void get_shrink (CBOR *self) |
1524 | void get_shrink (CBOR *self) |
1153 | ALIAS: |
1525 | ALIAS: |
1154 | get_shrink = F_SHRINK |
1526 | get_shrink = F_SHRINK |
1155 | get_allow_unknown = F_ALLOW_UNKNOWN |
1527 | get_allow_unknown = F_ALLOW_UNKNOWN |
1156 | get_allow_sharing = F_ALLOW_SHARING |
1528 | get_allow_sharing = F_ALLOW_SHARING |
1157 | get_allow_stringref = F_ALLOW_STRINGREF |
1529 | get_allow_cycles = F_ALLOW_CYCLES |
|
|
1530 | get_forbid_objects = F_FORBID_OBJECTS |
|
|
1531 | get_pack_strings = F_PACK_STRINGS |
|
|
1532 | get_text_keys = F_TEXT_KEYS |
|
|
1533 | get_text_strings = F_TEXT_STRINGS |
|
|
1534 | get_validate_utf8 = F_VALIDATE_UTF8 |
1158 | PPCODE: |
1535 | PPCODE: |
1159 | XPUSHs (boolSV (self->flags & ix)); |
1536 | XPUSHs (boolSV (self->flags & ix)); |
1160 | |
1537 | |
1161 | void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) |
1538 | void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) |
1162 | PPCODE: |
1539 | PPCODE: |
… | |
… | |
1211 | EXTEND (SP, 2); |
1588 | EXTEND (SP, 2); |
1212 | PUSHs (sv); |
1589 | PUSHs (sv); |
1213 | PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); |
1590 | PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); |
1214 | } |
1591 | } |
1215 | |
1592 | |
|
|
1593 | void incr_parse (CBOR *self, SV *cborstr) |
|
|
1594 | ALIAS: |
|
|
1595 | incr_parse_multiple = 1 |
|
|
1596 | PPCODE: |
|
|
1597 | { |
|
|
1598 | if (SvUTF8 (cborstr)) |
|
|
1599 | sv_utf8_downgrade (cborstr, 0); |
|
|
1600 | |
|
|
1601 | if (!self->incr_count) |
|
|
1602 | { |
|
|
1603 | self->incr_count = newAV (); |
|
|
1604 | self->incr_pos = 0; |
|
|
1605 | self->incr_need = 1; |
|
|
1606 | |
|
|
1607 | av_push (self->incr_count, newSViv (1)); |
|
|
1608 | } |
|
|
1609 | |
|
|
1610 | do |
|
|
1611 | { |
|
|
1612 | if (!incr_parse (self, cborstr)) |
|
|
1613 | { |
|
|
1614 | if (self->incr_need > self->max_size && self->max_size) |
|
|
1615 | croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu", |
|
|
1616 | (unsigned long)self->incr_need, (unsigned long)self->max_size); |
|
|
1617 | |
|
|
1618 | break; |
|
|
1619 | } |
|
|
1620 | |
|
|
1621 | SV *sv; |
|
|
1622 | char *offset; |
|
|
1623 | |
|
|
1624 | PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN; |
|
|
1625 | XPUSHs (sv); |
|
|
1626 | |
|
|
1627 | sv_chop (cborstr, offset); |
|
|
1628 | |
|
|
1629 | av_clear (self->incr_count); |
|
|
1630 | av_push (self->incr_count, newSViv (1)); |
|
|
1631 | |
|
|
1632 | self->incr_pos = 0; |
|
|
1633 | self->incr_need = self->incr_pos + 1; |
|
|
1634 | } |
|
|
1635 | while (ix); |
|
|
1636 | } |
|
|
1637 | |
|
|
1638 | void incr_reset (CBOR *self) |
|
|
1639 | CODE: |
|
|
1640 | { |
|
|
1641 | SvREFCNT_dec (self->incr_count); |
|
|
1642 | self->incr_count = 0; |
|
|
1643 | } |
|
|
1644 | |
1216 | void DESTROY (CBOR *self) |
1645 | void DESTROY (CBOR *self) |
1217 | PPCODE: |
1646 | PPCODE: |
1218 | cbor_free (self); |
1647 | cbor_free (self); |
1219 | |
1648 | |
1220 | PROTOTYPES: ENABLE |
1649 | PROTOTYPES: ENABLE |
1221 | |
1650 | |
1222 | void encode_cbor (SV *scalar) |
1651 | void encode_cbor (SV *scalar) |
|
|
1652 | ALIAS: |
|
|
1653 | encode_cbor = 0 |
|
|
1654 | encode_cbor_sharing = F_ALLOW_SHARING |
1223 | PPCODE: |
1655 | PPCODE: |
1224 | { |
1656 | { |
1225 | CBOR cbor; |
1657 | CBOR cbor; |
1226 | cbor_init (&cbor); |
1658 | cbor_init (&cbor); |
|
|
1659 | cbor.flags |= ix; |
1227 | PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; |
1660 | PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; |
1228 | XPUSHs (scalar); |
1661 | XPUSHs (scalar); |
1229 | } |
1662 | } |
1230 | |
1663 | |
1231 | void decode_cbor (SV *cborstr) |
1664 | void decode_cbor (SV *cborstr) |
… | |
… | |
1235 | cbor_init (&cbor); |
1668 | cbor_init (&cbor); |
1236 | PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN; |
1669 | PUTBACK; cborstr = decode_cbor (cborstr, &cbor, 0); SPAGAIN; |
1237 | XPUSHs (cborstr); |
1670 | XPUSHs (cborstr); |
1238 | } |
1671 | } |
1239 | |
1672 | |
|
|
1673 | #ifdef __AFL_COMPILER |
|
|
1674 | |
|
|
1675 | void |
|
|
1676 | afl_init () |
|
|
1677 | CODE: |
|
|
1678 | __AFL_INIT (); |
|
|
1679 | |
|
|
1680 | int |
|
|
1681 | afl_loop (unsigned int count = 10000) |
|
|
1682 | CODE: |
|
|
1683 | RETVAL = __AFL_LOOP (count); |
|
|
1684 | OUTPUT: |
|
|
1685 | RETVAL |
|
|
1686 | |
|
|
1687 | #endif |
|
|
1688 | |