… | |
… | |
9 | #include <limits.h> |
9 | #include <limits.h> |
10 | #include <float.h> |
10 | #include <float.h> |
11 | |
11 | |
12 | #include "ecb.h" |
12 | #include "ecb.h" |
13 | |
13 | |
14 | // known tags, rfc7049 |
14 | // compatibility with perl <5.18 |
|
|
15 | #ifndef HvNAMELEN_get |
|
|
16 | # define HvNAMELEN_get(hv) strlen (HvNAME (hv)) |
|
|
17 | #endif |
|
|
18 | #ifndef HvNAMELEN |
|
|
19 | # define HvNAMELEN(hv) HvNAMELEN_get (hv) |
|
|
20 | #endif |
|
|
21 | #ifndef HvNAMEUTF8 |
|
|
22 | # define HvNAMEUTF8(hv) 0 |
|
|
23 | #endif |
|
|
24 | #ifndef SvREFCNT_dec_NN |
|
|
25 | # define SvREFCNT_dec_NN(sv) SvREFCNT_dec (sv) |
|
|
26 | #endif |
|
|
27 | |
|
|
28 | // known major and minor types |
|
|
29 | enum cbor_type |
|
|
30 | { |
|
|
31 | MAJOR_SHIFT = 5, |
|
|
32 | MINOR_MASK = 0x1f, |
|
|
33 | |
|
|
34 | MAJOR_POS_INT = 0 << MAJOR_SHIFT, |
|
|
35 | MAJOR_NEG_INT = 1 << MAJOR_SHIFT, |
|
|
36 | MAJOR_BYTES = 2 << MAJOR_SHIFT, |
|
|
37 | MAJOR_TEXT = 3 << MAJOR_SHIFT, |
|
|
38 | MAJOR_ARRAY = 4 << MAJOR_SHIFT, |
|
|
39 | MAJOR_MAP = 5 << MAJOR_SHIFT, |
|
|
40 | MAJOR_TAG = 6 << MAJOR_SHIFT, |
|
|
41 | MAJOR_MISC = 7 << MAJOR_SHIFT, |
|
|
42 | |
|
|
43 | // INT/STRING/ARRAY/MAP subtypes |
|
|
44 | LENGTH_EXT1 = 24, |
|
|
45 | LENGTH_EXT2 = 25, |
|
|
46 | LENGTH_EXT4 = 26, |
|
|
47 | LENGTH_EXT8 = 27, |
|
|
48 | |
|
|
49 | // SIMPLE types (effectively MISC subtypes) |
|
|
50 | SIMPLE_FALSE = 20, |
|
|
51 | SIMPLE_TRUE = 21, |
|
|
52 | SIMPLE_NULL = 22, |
|
|
53 | SIMPLE_UNDEF = 23, |
|
|
54 | |
|
|
55 | // MISC subtype (unused) |
|
|
56 | MISC_EXT1 = 24, |
|
|
57 | MISC_FLOAT16 = 25, |
|
|
58 | MISC_FLOAT32 = 26, |
|
|
59 | MISC_FLOAT64 = 27, |
|
|
60 | |
|
|
61 | // BYTES/TEXT/ARRAY/MAP |
|
|
62 | MINOR_INDEF = 31, |
|
|
63 | }; |
|
|
64 | |
|
|
65 | // known tags |
15 | enum cbor_tag |
66 | enum cbor_tag |
16 | { |
67 | { |
|
|
68 | // extensions |
|
|
69 | CBOR_TAG_STRINGREF = 25, // http://cbor.schmorp.de/stringref |
|
|
70 | CBOR_TAG_PERL_OBJECT = 26, // http://cbor.schmorp.de/perl-object |
|
|
71 | CBOR_TAG_GENERIC_OBJECT = 27, // http://cbor.schmorp.de/generic-object |
|
|
72 | CBOR_TAG_VALUE_SHAREABLE = 28, // http://cbor.schmorp.de/value-sharing |
|
|
73 | CBOR_TAG_VALUE_SHAREDREF = 29, // http://cbor.schmorp.de/value-sharing |
|
|
74 | CBOR_TAG_STRINGREF_NAMESPACE = 256, // http://cbor.schmorp.de/stringref |
|
|
75 | CBOR_TAG_INDIRECTION = 22098, // http://cbor.schmorp.de/indirection |
|
|
76 | |
|
|
77 | // rfc7049 |
17 | CBOR_TAG_DATETIME = 0, // rfc4287, utf-8 |
78 | CBOR_TAG_DATETIME = 0, // rfc4287, utf-8 |
18 | CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any |
79 | CBOR_TAG_TIMESTAMP = 1, // unix timestamp, any |
19 | CBOR_TAG_POS_BIGNUM = 2, // byte string |
80 | CBOR_TAG_POS_BIGNUM = 2, // byte string |
20 | CBOR_TAG_NEG_BIGNUM = 3, // byte string |
81 | CBOR_TAG_NEG_BIGNUM = 3, // byte string |
21 | CBOR_TAG_DECIMAL = 4, // decimal fraction, array |
82 | CBOR_TAG_DECIMAL = 4, // decimal fraction, array |
22 | CBOR_TAG_BIGFLOAT = 5, // array |
83 | CBOR_TAG_BIGFLOAT = 5, // array |
23 | |
84 | |
24 | CBOR_TAG_CONV_B64U = 21, // base64url, any |
85 | CBOR_TAG_CONV_B64U = 21, // base64url, any |
25 | CBOR_TAG_CONV_B64 = 22, // base64, any |
86 | CBOR_TAG_CONV_B64 = 22, // base64, any |
26 | CBOR_TAG_CONV_HEX = 23, // base16, any |
87 | CBOR_TAG_CONV_HEX = 23, // base16, any |
27 | CBOR_TAG_CBOR = 24, // embedded cbor, byte string |
88 | CBOR_TAG_CBOR = 24, // embedded cbor, byte string |
28 | |
89 | |
29 | CBOR_TAG_URI = 32, // URI rfc3986, utf-8 |
90 | CBOR_TAG_URI = 32, // URI rfc3986, utf-8 |
30 | CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8 |
91 | CBOR_TAG_B64U = 33, // base64url rfc4648, utf-8 |
31 | CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8 |
92 | CBOR_TAG_B64 = 34, // base6 rfc46484, utf-8 |
32 | CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8 |
93 | CBOR_TAG_REGEX = 35, // regex pcre/ecma262, utf-8 |
33 | CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8 |
94 | CBOR_TAG_MIME = 36, // mime message rfc2045, utf-8 |
34 | |
95 | |
35 | CBOR_TAG_MAGIC = 55799 |
96 | CBOR_TAG_MAGIC = 55799, // self-describe cbor |
36 | }; |
97 | }; |
37 | |
98 | |
38 | #define F_SHRINK 0x00000200UL |
99 | #define F_SHRINK 0x00000001UL |
39 | #define F_ALLOW_UNKNOWN 0x00002000UL |
100 | #define F_ALLOW_UNKNOWN 0x00000002UL |
|
|
101 | #define F_ALLOW_SHARING 0x00000004UL |
|
|
102 | #define F_ALLOW_CYCLES 0x00000008UL |
|
|
103 | #define F_PACK_STRINGS 0x00000010UL |
|
|
104 | #define F_VALIDATE_UTF8 0x00000020UL |
40 | |
105 | |
41 | #define INIT_SIZE 32 // initial scalar size to be allocated |
106 | #define INIT_SIZE 32 // initial scalar size to be allocated |
42 | |
107 | |
43 | #define SB do { |
108 | #define SB do { |
44 | #define SE } while (0) |
109 | #define SE } while (0) |
… | |
… | |
55 | #else |
120 | #else |
56 | # define CBOR_SLOW 0 |
121 | # define CBOR_SLOW 0 |
57 | # define CBOR_STASH cbor_stash |
122 | # define CBOR_STASH cbor_stash |
58 | #endif |
123 | #endif |
59 | |
124 | |
60 | static HV *cbor_stash, *cbor_boolean_stash, *cbor_tagged_stash; // CBOR::XS:: |
125 | static HV *cbor_stash, *types_boolean_stash, *types_error_stash, *cbor_tagged_stash; // CBOR::XS:: |
61 | static SV *cbor_true, *cbor_false; |
126 | static SV *types_true, *types_false, *types_error, *sv_cbor, *default_filter; |
62 | |
127 | |
63 | typedef struct { |
128 | typedef struct { |
64 | U32 flags; |
129 | U32 flags; |
65 | U32 max_depth; |
130 | U32 max_depth; |
66 | STRLEN max_size; |
131 | STRLEN max_size; |
|
|
132 | SV *filter; |
|
|
133 | |
|
|
134 | // for the incremental parser |
|
|
135 | STRLEN incr_pos; // the current offset into the text |
|
|
136 | STRLEN incr_need; // minimum bytes needed to decode |
|
|
137 | AV *incr_count; // for every nesting level, the number of outstanding values, or -1 for indef. |
67 | } CBOR; |
138 | } CBOR; |
68 | |
139 | |
69 | ecb_inline void |
140 | ecb_inline void |
70 | cbor_init (CBOR *cbor) |
141 | cbor_init (CBOR *cbor) |
71 | { |
142 | { |
72 | Zero (cbor, 1, CBOR); |
143 | Zero (cbor, 1, CBOR); |
73 | cbor->max_depth = 512; |
144 | cbor->max_depth = 512; |
|
|
145 | } |
|
|
146 | |
|
|
147 | ecb_inline void |
|
|
148 | cbor_free (CBOR *cbor) |
|
|
149 | { |
|
|
150 | SvREFCNT_dec (cbor->filter); |
|
|
151 | SvREFCNT_dec (cbor->incr_count); |
74 | } |
152 | } |
75 | |
153 | |
76 | ///////////////////////////////////////////////////////////////////////////// |
154 | ///////////////////////////////////////////////////////////////////////////// |
77 | // utility functions |
155 | // utility functions |
78 | |
156 | |
… | |
… | |
100 | SvPV_renew (sv, SvCUR (sv) + 1); |
178 | SvPV_renew (sv, SvCUR (sv) + 1); |
101 | #endif |
179 | #endif |
102 | } |
180 | } |
103 | } |
181 | } |
104 | |
182 | |
105 | ///////////////////////////////////////////////////////////////////////////// |
183 | // minimum length of a string to be registered for stringref |
106 | // fp hell |
184 | ecb_inline int |
107 | |
185 | minimum_string_length (UV idx) |
108 | //TODO |
186 | { |
|
|
187 | return idx > 23 |
|
|
188 | ? idx > 0xffU |
|
|
189 | ? idx > 0xffffU |
|
|
190 | ? idx > 0xffffffffU |
|
|
191 | ? 11 |
|
|
192 | : 7 |
|
|
193 | : 5 |
|
|
194 | : 4 |
|
|
195 | : 3; |
|
|
196 | } |
109 | |
197 | |
110 | ///////////////////////////////////////////////////////////////////////////// |
198 | ///////////////////////////////////////////////////////////////////////////// |
111 | // encoder |
199 | // encoder |
112 | |
200 | |
113 | // structure used for encoding CBOR |
201 | // structure used for encoding CBOR |
… | |
… | |
116 | char *cur; // SvPVX (sv) + current output position |
204 | char *cur; // SvPVX (sv) + current output position |
117 | char *end; // SvEND (sv) |
205 | char *end; // SvEND (sv) |
118 | SV *sv; // result scalar |
206 | SV *sv; // result scalar |
119 | CBOR cbor; |
207 | CBOR cbor; |
120 | U32 depth; // recursion level |
208 | U32 depth; // recursion level |
|
|
209 | HV *stringref[2]; // string => index, or 0 ([0] = bytes, [1] = utf-8) |
|
|
210 | UV stringref_idx; |
|
|
211 | HV *shareable; // ptr => index, or 0 |
|
|
212 | UV shareable_idx; |
121 | } enc_t; |
213 | } enc_t; |
122 | |
214 | |
123 | ecb_inline void |
215 | ecb_inline void |
124 | need (enc_t *enc, STRLEN len) |
216 | need (enc_t *enc, STRLEN len) |
125 | { |
217 | { |
… | |
… | |
142 | static void |
234 | static void |
143 | encode_uint (enc_t *enc, int major, UV len) |
235 | encode_uint (enc_t *enc, int major, UV len) |
144 | { |
236 | { |
145 | need (enc, 9); |
237 | need (enc, 9); |
146 | |
238 | |
147 | if (len < 24) |
239 | if (ecb_expect_true (len < LENGTH_EXT1)) |
148 | *enc->cur++ = major | len; |
240 | *enc->cur++ = major | len; |
149 | else if (len <= 0xff) |
241 | else if (ecb_expect_true (len <= 0xffU)) |
150 | { |
242 | { |
151 | *enc->cur++ = major | 24; |
243 | *enc->cur++ = major | LENGTH_EXT1; |
152 | *enc->cur++ = len; |
244 | *enc->cur++ = len; |
153 | } |
245 | } |
154 | else if (len <= 0xffff) |
246 | else if (len <= 0xffffU) |
155 | { |
247 | { |
156 | *enc->cur++ = major | 25; |
248 | *enc->cur++ = major | LENGTH_EXT2; |
157 | *enc->cur++ = len >> 8; |
249 | *enc->cur++ = len >> 8; |
158 | *enc->cur++ = len; |
250 | *enc->cur++ = len; |
159 | } |
251 | } |
160 | else if (len <= 0xffffffff) |
252 | else if (len <= 0xffffffffU) |
161 | { |
253 | { |
162 | *enc->cur++ = major | 26; |
254 | *enc->cur++ = major | LENGTH_EXT4; |
163 | *enc->cur++ = len >> 24; |
255 | *enc->cur++ = len >> 24; |
164 | *enc->cur++ = len >> 16; |
256 | *enc->cur++ = len >> 16; |
165 | *enc->cur++ = len >> 8; |
257 | *enc->cur++ = len >> 8; |
166 | *enc->cur++ = len; |
258 | *enc->cur++ = len; |
167 | } |
259 | } |
168 | else |
260 | else |
169 | { |
261 | { |
170 | *enc->cur++ = major | 27; |
262 | *enc->cur++ = major | LENGTH_EXT8; |
171 | *enc->cur++ = len >> 56; |
263 | *enc->cur++ = len >> 56; |
172 | *enc->cur++ = len >> 48; |
264 | *enc->cur++ = len >> 48; |
173 | *enc->cur++ = len >> 40; |
265 | *enc->cur++ = len >> 40; |
174 | *enc->cur++ = len >> 32; |
266 | *enc->cur++ = len >> 32; |
175 | *enc->cur++ = len >> 24; |
267 | *enc->cur++ = len >> 24; |
… | |
… | |
177 | *enc->cur++ = len >> 8; |
269 | *enc->cur++ = len >> 8; |
178 | *enc->cur++ = len; |
270 | *enc->cur++ = len; |
179 | } |
271 | } |
180 | } |
272 | } |
181 | |
273 | |
182 | static void |
274 | ecb_inline void |
|
|
275 | encode_tag (enc_t *enc, UV tag) |
|
|
276 | { |
|
|
277 | encode_uint (enc, MAJOR_TAG, tag); |
|
|
278 | } |
|
|
279 | |
|
|
280 | ecb_inline void |
183 | encode_str (enc_t *enc, int utf8, char *str, STRLEN len) |
281 | encode_str (enc_t *enc, int utf8, char *str, STRLEN len) |
184 | { |
282 | { |
185 | encode_uint (enc, utf8 ? 0x60 : 0x40, len); |
283 | encode_uint (enc, utf8 ? MAJOR_TEXT : MAJOR_BYTES, len); |
186 | need (enc, len); |
284 | need (enc, len); |
187 | memcpy (enc->cur, str, len); |
285 | memcpy (enc->cur, str, len); |
188 | enc->cur += len; |
286 | enc->cur += len; |
189 | } |
287 | } |
190 | |
288 | |
|
|
289 | static void |
|
|
290 | encode_strref (enc_t *enc, int utf8, char *str, STRLEN len) |
|
|
291 | { |
|
|
292 | if (ecb_expect_false (enc->cbor.flags & F_PACK_STRINGS)) |
|
|
293 | { |
|
|
294 | SV **svp = hv_fetch (enc->stringref[!!utf8], str, len, 1); |
|
|
295 | |
|
|
296 | if (SvOK (*svp)) |
|
|
297 | { |
|
|
298 | // already registered, use stringref |
|
|
299 | encode_tag (enc, CBOR_TAG_STRINGREF); |
|
|
300 | encode_uint (enc, MAJOR_POS_INT, SvUV (*svp)); |
|
|
301 | return; |
|
|
302 | } |
|
|
303 | else if (len >= minimum_string_length (enc->stringref_idx)) |
|
|
304 | { |
|
|
305 | // register only |
|
|
306 | sv_setuv (*svp, enc->stringref_idx); |
|
|
307 | ++enc->stringref_idx; |
|
|
308 | } |
|
|
309 | } |
|
|
310 | |
|
|
311 | encode_str (enc, utf8, str, len); |
|
|
312 | } |
|
|
313 | |
191 | static void encode_sv (enc_t *enc, SV *sv); |
314 | static void encode_sv (enc_t *enc, SV *sv); |
192 | |
315 | |
193 | static void |
316 | static void |
194 | encode_av (enc_t *enc, AV *av) |
317 | encode_av (enc_t *enc, AV *av) |
195 | { |
318 | { |
… | |
… | |
198 | if (enc->depth >= enc->cbor.max_depth) |
321 | if (enc->depth >= enc->cbor.max_depth) |
199 | croak (ERR_NESTING_EXCEEDED); |
322 | croak (ERR_NESTING_EXCEEDED); |
200 | |
323 | |
201 | ++enc->depth; |
324 | ++enc->depth; |
202 | |
325 | |
203 | encode_uint (enc, 0x80, len + 1); |
326 | encode_uint (enc, MAJOR_ARRAY, len + 1); |
204 | |
327 | |
205 | for (i = 0; i <= len; ++i) |
328 | for (i = 0; i <= len; ++i) |
206 | { |
329 | { |
207 | SV **svp = av_fetch (av, i, 0); |
330 | SV **svp = av_fetch (av, i, 0); |
208 | encode_sv (enc, svp ? *svp : &PL_sv_undef); |
331 | encode_sv (enc, svp ? *svp : &PL_sv_undef); |
… | |
… | |
223 | |
346 | |
224 | int pairs = hv_iterinit (hv); |
347 | int pairs = hv_iterinit (hv); |
225 | int mg = SvMAGICAL (hv); |
348 | int mg = SvMAGICAL (hv); |
226 | |
349 | |
227 | if (mg) |
350 | if (mg) |
228 | encode_ch (enc, 0xa0 | 31); |
351 | encode_ch (enc, MAJOR_MAP | MINOR_INDEF); |
229 | else |
352 | else |
230 | encode_uint (enc, 0xa0, pairs); |
353 | encode_uint (enc, MAJOR_MAP, pairs); |
231 | |
354 | |
232 | while ((he = hv_iternext (hv))) |
355 | while ((he = hv_iternext (hv))) |
233 | { |
356 | { |
234 | if (HeKLEN (he) == HEf_SVKEY) |
357 | if (HeKLEN (he) == HEf_SVKEY) |
235 | encode_sv (enc, HeSVKEY (he)); |
358 | encode_sv (enc, HeSVKEY (he)); |
236 | else |
359 | else |
237 | encode_str (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); |
360 | encode_strref (enc, HeKUTF8 (he), HeKEY (he), HeKLEN (he)); |
238 | |
361 | |
239 | encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); |
362 | encode_sv (enc, ecb_expect_false (mg) ? hv_iterval (hv, he) : HeVAL (he)); |
240 | } |
363 | } |
241 | |
364 | |
242 | if (mg) |
365 | if (mg) |
243 | encode_ch (enc, 0xe0 | 31); |
366 | encode_ch (enc, MAJOR_MISC | MINOR_INDEF); |
244 | |
367 | |
245 | --enc->depth; |
368 | --enc->depth; |
246 | } |
369 | } |
247 | |
370 | |
248 | // encode objects, arrays and special \0=false and \1=true values. |
371 | // encode objects, arrays and special \0=false and \1=true values. |
249 | static void |
372 | static void |
250 | encode_rv (enc_t *enc, SV *sv) |
373 | encode_rv (enc_t *enc, SV *sv) |
251 | { |
374 | { |
252 | svtype svt; |
|
|
253 | |
|
|
254 | SvGETMAGIC (sv); |
375 | SvGETMAGIC (sv); |
|
|
376 | |
255 | svt = SvTYPE (sv); |
377 | svtype svt = SvTYPE (sv); |
256 | |
378 | |
257 | if (ecb_expect_false (SvOBJECT (sv))) |
379 | if (ecb_expect_false (SvOBJECT (sv))) |
258 | { |
380 | { |
259 | HV *boolean_stash = !CBOR_SLOW || cbor_boolean_stash |
381 | HV *boolean_stash = !CBOR_SLOW || types_boolean_stash |
260 | ? cbor_boolean_stash |
382 | ? types_boolean_stash |
261 | : gv_stashpv ("CBOR::XS::Boolean", 1); |
383 | : gv_stashpv ("Types::Serialiser::Boolean", 1); |
|
|
384 | HV *error_stash = !CBOR_SLOW || types_error_stash |
|
|
385 | ? types_error_stash |
|
|
386 | : gv_stashpv ("Types::Serialiser::Error", 1); |
262 | HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash |
387 | HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash |
263 | ? cbor_tagged_stash |
388 | ? cbor_tagged_stash |
264 | : gv_stashpv ("CBOR::XS::Tagged" , 1); |
389 | : gv_stashpv ("CBOR::XS::Tagged" , 1); |
265 | |
390 | |
|
|
391 | HV *stash = SvSTASH (sv); |
|
|
392 | |
266 | if (SvSTASH (sv) == boolean_stash) |
393 | if (stash == boolean_stash) |
267 | encode_ch (enc, SvIV (sv) ? 0xe0 | 21 : 0xe0 | 20); |
394 | { |
|
|
395 | encode_ch (enc, SvIV (sv) ? MAJOR_MISC | SIMPLE_TRUE : MAJOR_MISC | SIMPLE_FALSE); |
|
|
396 | return; |
|
|
397 | } |
|
|
398 | else if (stash == error_stash) |
|
|
399 | { |
|
|
400 | encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF); |
|
|
401 | return; |
|
|
402 | } |
268 | else if (SvSTASH (sv) == tagged_stash) |
403 | else if (stash == tagged_stash) |
269 | { |
404 | { |
270 | if (svt != SVt_PVAV) |
405 | if (svt != SVt_PVAV) |
271 | croak ("encountered CBOR::XS::Tagged object that isn't an array"); |
406 | croak ("encountered CBOR::XS::Tagged object that isn't an array"); |
272 | |
407 | |
273 | encode_uint (enc, 0xc0, SvUV (*av_fetch ((AV *)sv, 0, 1))); |
408 | encode_uint (enc, MAJOR_TAG, SvUV (*av_fetch ((AV *)sv, 0, 1))); |
274 | encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); |
409 | encode_sv (enc, *av_fetch ((AV *)sv, 1, 1)); |
|
|
410 | |
|
|
411 | return; |
|
|
412 | } |
|
|
413 | } |
|
|
414 | |
|
|
415 | if (ecb_expect_false (SvREFCNT (sv) > 1) |
|
|
416 | && ecb_expect_false (enc->cbor.flags & F_ALLOW_SHARING)) |
|
|
417 | { |
|
|
418 | if (!enc->shareable) |
|
|
419 | enc->shareable = (HV *)sv_2mortal ((SV *)newHV ()); |
|
|
420 | |
|
|
421 | SV **svp = hv_fetch (enc->shareable, (char *)&sv, sizeof (sv), 1); |
|
|
422 | |
|
|
423 | if (SvOK (*svp)) |
|
|
424 | { |
|
|
425 | encode_tag (enc, CBOR_TAG_VALUE_SHAREDREF); |
|
|
426 | encode_uint (enc, MAJOR_POS_INT, SvUV (*svp)); |
|
|
427 | return; |
275 | } |
428 | } |
276 | else |
429 | else |
277 | { |
430 | { |
|
|
431 | sv_setuv (*svp, enc->shareable_idx); |
|
|
432 | ++enc->shareable_idx; |
|
|
433 | encode_tag (enc, CBOR_TAG_VALUE_SHAREABLE); |
|
|
434 | } |
|
|
435 | } |
|
|
436 | |
|
|
437 | if (ecb_expect_false (SvOBJECT (sv))) |
|
|
438 | { |
|
|
439 | HV *stash = SvSTASH (sv); |
|
|
440 | GV *method; |
|
|
441 | |
|
|
442 | if ((method = gv_fetchmethod_autoload (stash, "TO_CBOR", 0))) |
|
|
443 | { |
|
|
444 | dSP; |
|
|
445 | |
|
|
446 | ENTER; SAVETMPS; PUSHMARK (SP); |
278 | // we re-bless the reference to get overload and other niceties right |
447 | // we re-bless the reference to get overload and other niceties right |
279 | GV *to_cbor = gv_fetchmethod_autoload (SvSTASH (sv), "TO_CBOR", 0); |
|
|
280 | |
|
|
281 | if (to_cbor) |
|
|
282 | { |
|
|
283 | dSP; |
|
|
284 | |
|
|
285 | ENTER; SAVETMPS; PUSHMARK (SP); |
|
|
286 | XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), SvSTASH (sv))); |
448 | XPUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); |
287 | |
449 | |
288 | // calling with G_SCALAR ensures that we always get a 1 return value |
|
|
289 | PUTBACK; |
450 | PUTBACK; |
|
|
451 | // G_SCALAR ensures that return value is 1 |
290 | call_sv ((SV *)GvCV (to_cbor), G_SCALAR); |
452 | call_sv ((SV *)GvCV (method), G_SCALAR); |
291 | SPAGAIN; |
453 | SPAGAIN; |
292 | |
454 | |
293 | // catch this surprisingly common error |
455 | // catch this surprisingly common error |
294 | if (SvROK (TOPs) && SvRV (TOPs) == sv) |
456 | if (SvROK (TOPs) && SvRV (TOPs) == sv) |
295 | croak ("%s::TO_CBOR method returned same object as was passed instead of a new one", HvNAME (SvSTASH (sv))); |
457 | croak ("%s::TO_CBOR method returned same object as was passed instead of a new one", HvNAME (stash)); |
296 | |
458 | |
297 | sv = POPs; |
459 | encode_sv (enc, POPs); |
|
|
460 | |
298 | PUTBACK; |
461 | PUTBACK; |
299 | |
462 | |
300 | encode_sv (enc, sv); |
|
|
301 | |
|
|
302 | FREETMPS; LEAVE; |
463 | FREETMPS; LEAVE; |
303 | } |
464 | } |
|
|
465 | else if ((method = gv_fetchmethod_autoload (stash, "FREEZE", 0)) != 0) |
|
|
466 | { |
|
|
467 | dSP; |
|
|
468 | |
|
|
469 | ENTER; SAVETMPS; PUSHMARK (SP); |
|
|
470 | EXTEND (SP, 2); |
|
|
471 | // we re-bless the reference to get overload and other niceties right |
|
|
472 | PUSHs (sv_bless (sv_2mortal (newRV_inc (sv)), stash)); |
|
|
473 | PUSHs (sv_cbor); |
|
|
474 | |
|
|
475 | PUTBACK; |
|
|
476 | int count = call_sv ((SV *)GvCV (method), G_ARRAY); |
|
|
477 | SPAGAIN; |
|
|
478 | |
|
|
479 | // catch this surprisingly common error |
|
|
480 | if (count == 1 && SvROK (TOPs) && SvRV (TOPs) == sv) |
|
|
481 | croak ("%s::FREEZE(CBOR) method returned same object as was passed instead of a new one", HvNAME (stash)); |
|
|
482 | |
|
|
483 | encode_tag (enc, CBOR_TAG_PERL_OBJECT); |
|
|
484 | encode_uint (enc, MAJOR_ARRAY, count + 1); |
|
|
485 | encode_strref (enc, HvNAMEUTF8 (stash), HvNAME (stash), HvNAMELEN (stash)); |
|
|
486 | |
|
|
487 | while (count) |
|
|
488 | encode_sv (enc, SP[1 - count--]); |
|
|
489 | |
|
|
490 | PUTBACK; |
|
|
491 | |
|
|
492 | FREETMPS; LEAVE; |
|
|
493 | } |
304 | else |
494 | else |
305 | croak ("encountered object '%s', but no TO_CBOR method available on it", |
495 | croak ("encountered object '%s', but no TO_CBOR or FREEZE methods available on it", |
306 | SvPV_nolen (sv_2mortal (newRV_inc (sv)))); |
496 | SvPV_nolen (sv_2mortal (newRV_inc (sv)))); |
307 | } |
|
|
308 | } |
497 | } |
309 | else if (svt == SVt_PVHV) |
498 | else if (svt == SVt_PVHV) |
310 | encode_hv (enc, (HV *)sv); |
499 | encode_hv (enc, (HV *)sv); |
311 | else if (svt == SVt_PVAV) |
500 | else if (svt == SVt_PVAV) |
312 | encode_av (enc, (AV *)sv); |
501 | encode_av (enc, (AV *)sv); |
313 | else if (svt < SVt_PVAV) |
|
|
314 | { |
|
|
315 | STRLEN len = 0; |
|
|
316 | char *pv = svt ? SvPV (sv, len) : 0; |
|
|
317 | |
|
|
318 | if (len == 1 && *pv == '1') |
|
|
319 | encode_ch (enc, 0xe0 | 21); |
|
|
320 | else if (len == 1 && *pv == '0') |
|
|
321 | encode_ch (enc, 0xe0 | 20); |
|
|
322 | else if (enc->cbor.flags & F_ALLOW_UNKNOWN) |
|
|
323 | encode_ch (enc, 0xe0 | 23); |
|
|
324 | else |
|
|
325 | croak ("cannot encode reference to scalar '%s' unless the scalar is 0 or 1", |
|
|
326 | SvPV_nolen (sv_2mortal (newRV_inc (sv)))); |
|
|
327 | } |
|
|
328 | else if (enc->cbor.flags & F_ALLOW_UNKNOWN) |
|
|
329 | encode_ch (enc, 0xe0 | 23); |
|
|
330 | else |
502 | else |
331 | croak ("encountered %s, but CBOR can only represent references to arrays or hashes", |
503 | { |
332 | SvPV_nolen (sv_2mortal (newRV_inc (sv)))); |
504 | encode_tag (enc, CBOR_TAG_INDIRECTION); |
|
|
505 | encode_sv (enc, sv); |
|
|
506 | } |
333 | } |
507 | } |
334 | |
508 | |
335 | static void |
509 | static void |
336 | encode_nv (enc_t *enc, SV *sv) |
510 | encode_nv (enc_t *enc, SV *sv) |
337 | { |
511 | { |
338 | double nv = SvNVX (sv); |
512 | double nv = SvNVX (sv); |
339 | |
513 | |
340 | need (enc, 9); |
514 | need (enc, 9); |
341 | |
515 | |
342 | if (ecb_expect_false (nv == (U32)nv)) |
516 | if (ecb_expect_false (nv == (NV)(U32)nv)) |
343 | encode_uint (enc, 0x00, (U32)nv); |
517 | encode_uint (enc, MAJOR_POS_INT, (U32)nv); |
344 | //TODO: maybe I32? |
518 | //TODO: maybe I32? |
345 | else if (ecb_expect_false (nv == (float)nv)) |
519 | else if (ecb_expect_false (nv == (float)nv)) |
346 | { |
520 | { |
347 | uint32_t fp = ecb_float_to_binary32 (nv); |
521 | uint32_t fp = ecb_float_to_binary32 (nv); |
348 | |
522 | |
349 | *enc->cur++ = 0xe0 | 26; |
523 | *enc->cur++ = MAJOR_MISC | MISC_FLOAT32; |
350 | |
524 | |
351 | if (!ecb_big_endian ()) |
525 | if (!ecb_big_endian ()) |
352 | fp = ecb_bswap32 (fp); |
526 | fp = ecb_bswap32 (fp); |
353 | |
527 | |
354 | memcpy (enc->cur, &fp, 4); |
528 | memcpy (enc->cur, &fp, 4); |
… | |
… | |
356 | } |
530 | } |
357 | else |
531 | else |
358 | { |
532 | { |
359 | uint64_t fp = ecb_double_to_binary64 (nv); |
533 | uint64_t fp = ecb_double_to_binary64 (nv); |
360 | |
534 | |
361 | *enc->cur++ = 0xe0 | 27; |
535 | *enc->cur++ = MAJOR_MISC | MISC_FLOAT64; |
362 | |
536 | |
363 | if (!ecb_big_endian ()) |
537 | if (!ecb_big_endian ()) |
364 | fp = ecb_bswap64 (fp); |
538 | fp = ecb_bswap64 (fp); |
365 | |
539 | |
366 | memcpy (enc->cur, &fp, 8); |
540 | memcpy (enc->cur, &fp, 8); |
… | |
… | |
375 | |
549 | |
376 | if (SvPOKp (sv)) |
550 | if (SvPOKp (sv)) |
377 | { |
551 | { |
378 | STRLEN len; |
552 | STRLEN len; |
379 | char *str = SvPV (sv, len); |
553 | char *str = SvPV (sv, len); |
380 | encode_str (enc, SvUTF8 (sv), str, len); |
554 | encode_strref (enc, SvUTF8 (sv), str, len); |
381 | } |
555 | } |
382 | else if (SvNOKp (sv)) |
556 | else if (SvNOKp (sv)) |
383 | encode_nv (enc, sv); |
557 | encode_nv (enc, sv); |
384 | else if (SvIOKp (sv)) |
558 | else if (SvIOKp (sv)) |
385 | { |
559 | { |
386 | if (SvIsUV (sv)) |
560 | if (SvIsUV (sv)) |
387 | encode_uint (enc, 0x00, SvUVX (sv)); |
561 | encode_uint (enc, MAJOR_POS_INT, SvUVX (sv)); |
388 | else if (SvIVX (sv) >= 0) |
562 | else if (SvIVX (sv) >= 0) |
389 | encode_uint (enc, 0x00, SvIVX (sv)); |
563 | encode_uint (enc, MAJOR_POS_INT, SvIVX (sv)); |
390 | else |
564 | else |
391 | encode_uint (enc, 0x20, -(SvIVX (sv) + 1)); |
565 | encode_uint (enc, MAJOR_NEG_INT, -(SvIVX (sv) + 1)); |
392 | } |
566 | } |
393 | else if (SvROK (sv)) |
567 | else if (SvROK (sv)) |
394 | encode_rv (enc, SvRV (sv)); |
568 | encode_rv (enc, SvRV (sv)); |
395 | else if (!SvOK (sv)) |
569 | else if (!SvOK (sv)) |
396 | encode_ch (enc, 0xe0 | 22); |
570 | encode_ch (enc, MAJOR_MISC | SIMPLE_NULL); |
397 | else if (enc->cbor.flags & F_ALLOW_UNKNOWN) |
571 | else if (enc->cbor.flags & F_ALLOW_UNKNOWN) |
398 | encode_ch (enc, 0xe0 | 23); |
572 | encode_ch (enc, MAJOR_MISC | SIMPLE_UNDEF); |
399 | else |
573 | else |
400 | croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data", |
574 | croak ("encountered perl type (%s,0x%x) that CBOR cannot handle, check your input data", |
401 | SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); |
575 | SvPV_nolen (sv), (unsigned int)SvFLAGS (sv)); |
402 | } |
576 | } |
403 | |
577 | |
404 | static SV * |
578 | static SV * |
405 | encode_cbor (SV *scalar, CBOR *cbor) |
579 | encode_cbor (SV *scalar, CBOR *cbor) |
406 | { |
580 | { |
407 | enc_t enc; |
581 | enc_t enc = { }; |
408 | |
582 | |
409 | enc.cbor = *cbor; |
583 | enc.cbor = *cbor; |
410 | enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); |
584 | enc.sv = sv_2mortal (NEWSV (0, INIT_SIZE)); |
411 | enc.cur = SvPVX (enc.sv); |
585 | enc.cur = SvPVX (enc.sv); |
412 | enc.end = SvEND (enc.sv); |
586 | enc.end = SvEND (enc.sv); |
413 | enc.depth = 0; |
|
|
414 | |
587 | |
415 | SvPOK_only (enc.sv); |
588 | SvPOK_only (enc.sv); |
|
|
589 | |
|
|
590 | if (cbor->flags & F_PACK_STRINGS) |
|
|
591 | { |
|
|
592 | encode_tag (&enc, CBOR_TAG_STRINGREF_NAMESPACE); |
|
|
593 | enc.stringref[0]= (HV *)sv_2mortal ((SV *)newHV ()); |
|
|
594 | enc.stringref[1]= (HV *)sv_2mortal ((SV *)newHV ()); |
|
|
595 | } |
|
|
596 | |
416 | encode_sv (&enc, scalar); |
597 | encode_sv (&enc, scalar); |
417 | |
598 | |
418 | SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv)); |
599 | SvCUR_set (enc.sv, enc.cur - SvPVX (enc.sv)); |
419 | *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings |
600 | *SvEND (enc.sv) = 0; // many xs functions expect a trailing 0 for text strings |
420 | |
601 | |
… | |
… | |
434 | U8 *end; // end of input string |
615 | U8 *end; // end of input string |
435 | const char *err; // parse error, if != 0 |
616 | const char *err; // parse error, if != 0 |
436 | CBOR cbor; |
617 | CBOR cbor; |
437 | U32 depth; // recursion depth |
618 | U32 depth; // recursion depth |
438 | U32 maxdepth; // recursion depth limit |
619 | U32 maxdepth; // recursion depth limit |
|
|
620 | AV *shareable; |
|
|
621 | AV *stringref; |
|
|
622 | SV *decode_tagged; |
439 | } dec_t; |
623 | } dec_t; |
440 | |
624 | |
441 | #define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE |
625 | #define ERR(reason) SB if (!dec->err) dec->err = reason; goto fail; SE |
442 | |
626 | |
443 | #define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data") |
627 | #define WANT(len) if (ecb_expect_false (dec->cur + len > dec->end)) ERR ("unexpected end of CBOR data") |
… | |
… | |
446 | #define DEC_DEC_DEPTH --dec->depth |
630 | #define DEC_DEC_DEPTH --dec->depth |
447 | |
631 | |
448 | static UV |
632 | static UV |
449 | decode_uint (dec_t *dec) |
633 | decode_uint (dec_t *dec) |
450 | { |
634 | { |
451 | switch (*dec->cur & 31) |
635 | U8 m = *dec->cur & MINOR_MASK; |
452 | { |
636 | ++dec->cur; |
453 | case 0: case 1: case 2: case 3: case 4: case 5: case 6: case 7: |
|
|
454 | case 8: case 9: case 10: case 11: case 12: case 13: case 14: case 15: |
|
|
455 | case 16: case 17: case 18: case 19: case 20: case 21: case 22: case 23: |
|
|
456 | return *dec->cur++ & 31; |
|
|
457 | |
637 | |
458 | case 24: |
638 | if (ecb_expect_true (m < LENGTH_EXT1)) |
|
|
639 | return m; |
|
|
640 | else if (ecb_expect_true (m == LENGTH_EXT1)) |
|
|
641 | { |
459 | WANT (2); |
642 | WANT (1); |
460 | dec->cur += 2; |
643 | dec->cur += 1; |
461 | return dec->cur[-1]; |
644 | return dec->cur[-1]; |
462 | |
645 | } |
463 | case 25: |
646 | else if (ecb_expect_true (m == LENGTH_EXT2)) |
|
|
647 | { |
464 | WANT (3); |
648 | WANT (2); |
465 | dec->cur += 3; |
649 | dec->cur += 2; |
466 | return (((UV)dec->cur[-2]) << 8) |
650 | return (((UV)dec->cur[-2]) << 8) |
467 | | ((UV)dec->cur[-1]); |
651 | | ((UV)dec->cur[-1]); |
468 | |
652 | } |
469 | case 26: |
653 | else if (ecb_expect_true (m == LENGTH_EXT4)) |
|
|
654 | { |
470 | WANT (5); |
655 | WANT (4); |
471 | dec->cur += 5; |
656 | dec->cur += 4; |
472 | return (((UV)dec->cur[-4]) << 24) |
657 | return (((UV)dec->cur[-4]) << 24) |
473 | | (((UV)dec->cur[-3]) << 16) |
658 | | (((UV)dec->cur[-3]) << 16) |
474 | | (((UV)dec->cur[-2]) << 8) |
659 | | (((UV)dec->cur[-2]) << 8) |
475 | | ((UV)dec->cur[-1]); |
660 | | ((UV)dec->cur[-1]); |
476 | |
661 | } |
477 | case 27: |
662 | else if (ecb_expect_true (m == LENGTH_EXT8)) |
|
|
663 | { |
478 | WANT (9); |
664 | WANT (8); |
479 | dec->cur += 9; |
665 | dec->cur += 8; |
|
|
666 | |
|
|
667 | return |
|
|
668 | #if UVSIZE < 8 |
|
|
669 | 0 |
|
|
670 | #else |
480 | return (((UV)dec->cur[-8]) << 56) |
671 | (((UV)dec->cur[-8]) << 56) |
481 | | (((UV)dec->cur[-7]) << 48) |
672 | | (((UV)dec->cur[-7]) << 48) |
482 | | (((UV)dec->cur[-6]) << 40) |
673 | | (((UV)dec->cur[-6]) << 40) |
483 | | (((UV)dec->cur[-5]) << 32) |
674 | | (((UV)dec->cur[-5]) << 32) |
|
|
675 | #endif |
484 | | (((UV)dec->cur[-4]) << 24) |
676 | | (((UV)dec->cur[-4]) << 24) |
485 | | (((UV)dec->cur[-3]) << 16) |
677 | | (((UV)dec->cur[-3]) << 16) |
486 | | (((UV)dec->cur[-2]) << 8) |
678 | | (((UV)dec->cur[-2]) << 8) |
487 | | ((UV)dec->cur[-1]); |
679 | | ((UV)dec->cur[-1]); |
488 | |
680 | } |
489 | default: |
681 | else |
490 | ERR ("corrupted CBOR data (unsupported integer minor encoding)"); |
682 | ERR ("corrupted CBOR data (unsupported integer minor encoding)"); |
491 | } |
|
|
492 | |
683 | |
493 | fail: |
684 | fail: |
494 | return 0; |
685 | return 0; |
495 | } |
686 | } |
496 | |
687 | |
… | |
… | |
501 | { |
692 | { |
502 | AV *av = newAV (); |
693 | AV *av = newAV (); |
503 | |
694 | |
504 | DEC_INC_DEPTH; |
695 | DEC_INC_DEPTH; |
505 | |
696 | |
506 | if ((*dec->cur & 31) == 31) |
697 | if (*dec->cur == (MAJOR_ARRAY | MINOR_INDEF)) |
507 | { |
698 | { |
508 | ++dec->cur; |
699 | ++dec->cur; |
509 | |
700 | |
510 | for (;;) |
701 | for (;;) |
511 | { |
702 | { |
512 | WANT (1); |
703 | WANT (1); |
513 | |
704 | |
514 | if (*dec->cur == (0xe0 | 31)) |
705 | if (*dec->cur == (MAJOR_MISC | MINOR_INDEF)) |
515 | { |
706 | { |
516 | ++dec->cur; |
707 | ++dec->cur; |
517 | break; |
708 | break; |
518 | } |
709 | } |
519 | |
710 | |
… | |
… | |
522 | } |
713 | } |
523 | else |
714 | else |
524 | { |
715 | { |
525 | int i, len = decode_uint (dec); |
716 | int i, len = decode_uint (dec); |
526 | |
717 | |
|
|
718 | WANT (len); // complexity check for av_fill - need at least one byte per value, do not allow supersize arrays |
527 | av_fill (av, len - 1); |
719 | av_fill (av, len - 1); |
528 | |
720 | |
529 | for (i = 0; i < len; ++i) |
721 | for (i = 0; i < len; ++i) |
530 | AvARRAY (av)[i] = decode_sv (dec); |
722 | AvARRAY (av)[i] = decode_sv (dec); |
531 | } |
723 | } |
… | |
… | |
537 | SvREFCNT_dec (av); |
729 | SvREFCNT_dec (av); |
538 | DEC_DEC_DEPTH; |
730 | DEC_DEC_DEPTH; |
539 | return &PL_sv_undef; |
731 | return &PL_sv_undef; |
540 | } |
732 | } |
541 | |
733 | |
|
|
734 | static void |
|
|
735 | decode_he (dec_t *dec, HV *hv) |
|
|
736 | { |
|
|
737 | // for speed reasons, we specialcase single-string |
|
|
738 | // byte or utf-8 strings as keys, but only when !stringref |
|
|
739 | |
|
|
740 | if (ecb_expect_true (!dec->stringref)) |
|
|
741 | if (ecb_expect_true ((*dec->cur - MAJOR_BYTES) <= LENGTH_EXT8)) |
|
|
742 | { |
|
|
743 | I32 len = decode_uint (dec); |
|
|
744 | char *key = (char *)dec->cur; |
|
|
745 | |
|
|
746 | dec->cur += len; |
|
|
747 | |
|
|
748 | hv_store (hv, key, len, decode_sv (dec), 0); |
|
|
749 | |
|
|
750 | return; |
|
|
751 | } |
|
|
752 | else if (ecb_expect_true ((*dec->cur - MAJOR_TEXT) <= LENGTH_EXT8)) |
|
|
753 | { |
|
|
754 | I32 len = decode_uint (dec); |
|
|
755 | char *key = (char *)dec->cur; |
|
|
756 | |
|
|
757 | dec->cur += len; |
|
|
758 | |
|
|
759 | if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8)) |
|
|
760 | if (!is_utf8_string (key, len)) |
|
|
761 | ERR ("corrupted CBOR data (invalid UTF-8 in map key)"); |
|
|
762 | |
|
|
763 | hv_store (hv, key, -len, decode_sv (dec), 0); |
|
|
764 | |
|
|
765 | return; |
|
|
766 | } |
|
|
767 | |
|
|
768 | SV *k = decode_sv (dec); |
|
|
769 | SV *v = decode_sv (dec); |
|
|
770 | |
|
|
771 | hv_store_ent (hv, k, v, 0); |
|
|
772 | SvREFCNT_dec (k); |
|
|
773 | |
|
|
774 | fail: |
|
|
775 | ; |
|
|
776 | } |
|
|
777 | |
542 | static SV * |
778 | static SV * |
543 | decode_hv (dec_t *dec) |
779 | decode_hv (dec_t *dec) |
544 | { |
780 | { |
545 | HV *hv = newHV (); |
781 | HV *hv = newHV (); |
546 | |
782 | |
547 | DEC_INC_DEPTH; |
783 | DEC_INC_DEPTH; |
548 | |
784 | |
549 | if ((*dec->cur & 31) == 31) |
785 | if (*dec->cur == (MAJOR_MAP | MINOR_INDEF)) |
550 | { |
786 | { |
551 | ++dec->cur; |
787 | ++dec->cur; |
552 | |
788 | |
553 | for (;;) |
789 | for (;;) |
554 | { |
790 | { |
555 | WANT (1); |
791 | WANT (1); |
556 | |
792 | |
557 | if (*dec->cur == (0xe0 | 31)) |
793 | if (*dec->cur == (MAJOR_MISC | MINOR_INDEF)) |
558 | { |
794 | { |
559 | ++dec->cur; |
795 | ++dec->cur; |
560 | break; |
796 | break; |
561 | } |
797 | } |
562 | |
798 | |
563 | SV *k = decode_sv (dec); |
799 | decode_he (dec, hv); |
564 | SV *v = decode_sv (dec); |
|
|
565 | |
|
|
566 | hv_store_ent (hv, k, v, 0); |
|
|
567 | } |
800 | } |
568 | } |
801 | } |
569 | else |
802 | else |
570 | { |
803 | { |
571 | int len = decode_uint (dec); |
804 | int pairs = decode_uint (dec); |
572 | |
805 | |
573 | while (len--) |
806 | while (pairs--) |
574 | { |
807 | decode_he (dec, hv); |
575 | SV *k = decode_sv (dec); |
|
|
576 | SV *v = decode_sv (dec); |
|
|
577 | |
|
|
578 | hv_store_ent (hv, k, v, 0); |
|
|
579 | } |
|
|
580 | } |
808 | } |
581 | |
809 | |
582 | DEC_DEC_DEPTH; |
810 | DEC_DEC_DEPTH; |
583 | return newRV_noinc ((SV *)hv); |
811 | return newRV_noinc ((SV *)hv); |
584 | |
812 | |
… | |
… | |
591 | static SV * |
819 | static SV * |
592 | decode_str (dec_t *dec, int utf8) |
820 | decode_str (dec_t *dec, int utf8) |
593 | { |
821 | { |
594 | SV *sv = 0; |
822 | SV *sv = 0; |
595 | |
823 | |
596 | if ((*dec->cur & 31) == 31) |
824 | if ((*dec->cur & MINOR_MASK) == MINOR_INDEF) |
597 | { |
825 | { |
|
|
826 | // indefinite length strings |
598 | ++dec->cur; |
827 | ++dec->cur; |
599 | |
828 | |
|
|
829 | U8 major = *dec->cur & MAJOR_MISC; |
|
|
830 | |
600 | sv = newSVpvn ("", 0); |
831 | sv = newSVpvn ("", 0); |
601 | |
832 | |
602 | // not very fast, and certainly not robust against illegal input |
|
|
603 | for (;;) |
833 | for (;;) |
604 | { |
834 | { |
605 | WANT (1); |
835 | WANT (1); |
606 | |
836 | |
607 | if (*dec->cur == (0xe0 | 31)) |
837 | if ((*dec->cur - major) > LENGTH_EXT8) |
|
|
838 | if (*dec->cur == (MAJOR_MISC | MINOR_INDEF)) |
608 | { |
839 | { |
609 | ++dec->cur; |
840 | ++dec->cur; |
610 | break; |
841 | break; |
611 | } |
842 | } |
|
|
843 | else |
|
|
844 | ERR ("corrupted CBOR data (invalid chunks in indefinite length string)"); |
612 | |
845 | |
613 | sv_catsv (sv, decode_sv (dec)); |
846 | STRLEN len = decode_uint (dec); |
|
|
847 | |
|
|
848 | WANT (len); |
|
|
849 | sv_catpvn (sv, dec->cur, len); |
|
|
850 | dec->cur += len; |
614 | } |
851 | } |
615 | } |
852 | } |
616 | else |
853 | else |
617 | { |
854 | { |
618 | STRLEN len = decode_uint (dec); |
855 | STRLEN len = decode_uint (dec); |
619 | |
856 | |
620 | WANT (len); |
857 | WANT (len); |
621 | sv = newSVpvn (dec->cur, len); |
858 | sv = newSVpvn (dec->cur, len); |
622 | dec->cur += len; |
859 | dec->cur += len; |
|
|
860 | |
|
|
861 | if (ecb_expect_false (dec->stringref) |
|
|
862 | && SvCUR (sv) >= minimum_string_length (AvFILLp (dec->stringref) + 1)) |
|
|
863 | av_push (dec->stringref, SvREFCNT_inc_NN (sv)); |
623 | } |
864 | } |
624 | |
865 | |
625 | if (utf8) |
866 | if (utf8) |
|
|
867 | { |
|
|
868 | if (ecb_expect_false (dec->cbor.flags & F_VALIDATE_UTF8)) |
|
|
869 | if (!is_utf8_string (SvPVX (sv), SvCUR (sv))) |
|
|
870 | ERR ("corrupted CBOR data (invalid UTF-8 in text string)"); |
|
|
871 | |
626 | SvUTF8_on (sv); |
872 | SvUTF8_on (sv); |
|
|
873 | } |
627 | |
874 | |
628 | return sv; |
875 | return sv; |
629 | |
876 | |
630 | fail: |
877 | fail: |
631 | SvREFCNT_dec (sv); |
878 | SvREFCNT_dec (sv); |
… | |
… | |
633 | } |
880 | } |
634 | |
881 | |
635 | static SV * |
882 | static SV * |
636 | decode_tagged (dec_t *dec) |
883 | decode_tagged (dec_t *dec) |
637 | { |
884 | { |
|
|
885 | SV *sv = 0; |
638 | UV tag = decode_uint (dec); |
886 | UV tag = decode_uint (dec); |
|
|
887 | |
|
|
888 | WANT (1); |
|
|
889 | |
|
|
890 | switch (tag) |
|
|
891 | { |
|
|
892 | case CBOR_TAG_MAGIC: |
639 | SV *sv = decode_sv (dec); |
893 | sv = decode_sv (dec); |
|
|
894 | break; |
640 | |
895 | |
641 | if (tag == CBOR_TAG_MAGIC) // 2.4.5 Self-Describe CBOR |
896 | case CBOR_TAG_INDIRECTION: |
|
|
897 | sv = newRV_noinc (decode_sv (dec)); |
|
|
898 | break; |
|
|
899 | |
|
|
900 | case CBOR_TAG_STRINGREF_NAMESPACE: |
|
|
901 | { |
|
|
902 | ENTER; SAVETMPS; |
|
|
903 | |
|
|
904 | SAVESPTR (dec->stringref); |
|
|
905 | dec->stringref = (AV *)sv_2mortal ((SV *)newAV ()); |
|
|
906 | |
|
|
907 | sv = decode_sv (dec); |
|
|
908 | |
|
|
909 | FREETMPS; LEAVE; |
|
|
910 | } |
|
|
911 | break; |
|
|
912 | |
|
|
913 | case CBOR_TAG_STRINGREF: |
|
|
914 | { |
|
|
915 | if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT)) |
|
|
916 | ERR ("corrupted CBOR data (stringref index not an unsigned integer)"); |
|
|
917 | |
|
|
918 | UV idx = decode_uint (dec); |
|
|
919 | |
|
|
920 | if (!dec->stringref || (int)idx > AvFILLp (dec->stringref)) |
|
|
921 | ERR ("corrupted CBOR data (stringref index out of bounds or outside namespace)"); |
|
|
922 | |
|
|
923 | sv = newSVsv (AvARRAY (dec->stringref)[idx]); |
|
|
924 | } |
|
|
925 | break; |
|
|
926 | |
|
|
927 | case CBOR_TAG_VALUE_SHAREABLE: |
|
|
928 | { |
|
|
929 | if (ecb_expect_false (!dec->shareable)) |
|
|
930 | dec->shareable = (AV *)sv_2mortal ((SV *)newAV ()); |
|
|
931 | |
|
|
932 | if (dec->cbor.flags & F_ALLOW_CYCLES) |
|
|
933 | { |
|
|
934 | sv = newSV (0); |
|
|
935 | av_push (dec->shareable, SvREFCNT_inc_NN (sv)); |
|
|
936 | |
|
|
937 | SV *osv = decode_sv (dec); |
|
|
938 | sv_setsv (sv, osv); |
|
|
939 | SvREFCNT_dec_NN (osv); |
|
|
940 | } |
|
|
941 | else |
|
|
942 | { |
|
|
943 | av_push (dec->shareable, &PL_sv_undef); |
|
|
944 | int idx = AvFILLp (dec->shareable); |
|
|
945 | sv = decode_sv (dec); |
|
|
946 | av_store (dec->shareable, idx, SvREFCNT_inc_NN (sv)); |
|
|
947 | } |
|
|
948 | } |
|
|
949 | break; |
|
|
950 | |
|
|
951 | case CBOR_TAG_VALUE_SHAREDREF: |
|
|
952 | { |
|
|
953 | if ((*dec->cur >> MAJOR_SHIFT) != (MAJOR_POS_INT >> MAJOR_SHIFT)) |
|
|
954 | ERR ("corrupted CBOR data (sharedref index not an unsigned integer)"); |
|
|
955 | |
|
|
956 | UV idx = decode_uint (dec); |
|
|
957 | |
|
|
958 | if (!dec->shareable || (int)idx > AvFILLp (dec->shareable)) |
|
|
959 | ERR ("corrupted CBOR data (sharedref index out of bounds)"); |
|
|
960 | |
|
|
961 | sv = SvREFCNT_inc_NN (AvARRAY (dec->shareable)[idx]); |
|
|
962 | |
|
|
963 | if (sv == &PL_sv_undef) |
|
|
964 | ERR ("cyclic CBOR data structure found, but allow_cycles is not enabled"); |
|
|
965 | } |
|
|
966 | break; |
|
|
967 | |
|
|
968 | case CBOR_TAG_PERL_OBJECT: |
|
|
969 | { |
|
|
970 | sv = decode_sv (dec); |
|
|
971 | |
|
|
972 | if (!SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV) |
|
|
973 | ERR ("corrupted CBOR data (non-array perl object)"); |
|
|
974 | |
|
|
975 | AV *av = (AV *)SvRV (sv); |
|
|
976 | int len = av_len (av) + 1; |
|
|
977 | HV *stash = gv_stashsv (*av_fetch (av, 0, 1), 0); |
|
|
978 | |
|
|
979 | if (!stash) |
|
|
980 | ERR ("cannot decode perl-object (package does not exist)"); |
|
|
981 | |
|
|
982 | GV *method = gv_fetchmethod_autoload (stash, "THAW", 0); |
|
|
983 | |
|
|
984 | if (!method) |
|
|
985 | ERR ("cannot decode perl-object (package does not have a THAW method)"); |
|
|
986 | |
|
|
987 | dSP; |
|
|
988 | |
|
|
989 | ENTER; SAVETMPS; PUSHMARK (SP); |
|
|
990 | EXTEND (SP, len + 1); |
|
|
991 | // we re-bless the reference to get overload and other niceties right |
|
|
992 | PUSHs (*av_fetch (av, 0, 1)); |
|
|
993 | PUSHs (sv_cbor); |
|
|
994 | |
|
|
995 | int i; |
|
|
996 | |
|
|
997 | for (i = 1; i < len; ++i) |
|
|
998 | PUSHs (*av_fetch (av, i, 1)); |
|
|
999 | |
|
|
1000 | PUTBACK; |
|
|
1001 | call_sv ((SV *)GvCV (method), G_SCALAR | G_EVAL); |
|
|
1002 | SPAGAIN; |
|
|
1003 | |
|
|
1004 | if (SvTRUE (ERRSV)) |
|
|
1005 | { |
|
|
1006 | FREETMPS; LEAVE; |
|
|
1007 | ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV)))); |
|
|
1008 | } |
|
|
1009 | |
|
|
1010 | SvREFCNT_dec (sv); |
|
|
1011 | sv = SvREFCNT_inc (POPs); |
|
|
1012 | |
|
|
1013 | PUTBACK; |
|
|
1014 | |
|
|
1015 | FREETMPS; LEAVE; |
|
|
1016 | } |
|
|
1017 | break; |
|
|
1018 | |
|
|
1019 | default: |
|
|
1020 | { |
|
|
1021 | sv = decode_sv (dec); |
|
|
1022 | |
|
|
1023 | dSP; |
|
|
1024 | ENTER; SAVETMPS; PUSHMARK (SP); |
|
|
1025 | EXTEND (SP, 2); |
|
|
1026 | PUSHs (newSVuv (tag)); |
|
|
1027 | PUSHs (sv); |
|
|
1028 | |
|
|
1029 | PUTBACK; |
|
|
1030 | int count = call_sv (dec->cbor.filter ? dec->cbor.filter : default_filter, G_ARRAY | G_EVAL); |
|
|
1031 | SPAGAIN; |
|
|
1032 | |
|
|
1033 | if (SvTRUE (ERRSV)) |
|
|
1034 | { |
|
|
1035 | FREETMPS; LEAVE; |
|
|
1036 | ERR (SvPVutf8_nolen (sv_2mortal (SvREFCNT_inc (ERRSV)))); |
|
|
1037 | } |
|
|
1038 | |
|
|
1039 | if (count) |
|
|
1040 | { |
|
|
1041 | SvREFCNT_dec (sv); |
|
|
1042 | sv = SvREFCNT_inc (POPs); |
|
|
1043 | } |
|
|
1044 | else |
|
|
1045 | { |
|
|
1046 | AV *av = newAV (); |
|
|
1047 | av_push (av, newSVuv (tag)); |
|
|
1048 | av_push (av, sv); |
|
|
1049 | |
|
|
1050 | HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash |
|
|
1051 | ? cbor_tagged_stash |
|
|
1052 | : gv_stashpv ("CBOR::XS::Tagged" , 1); |
|
|
1053 | sv = sv_bless (newRV_noinc ((SV *)av), tagged_stash); |
|
|
1054 | } |
|
|
1055 | |
|
|
1056 | PUTBACK; |
|
|
1057 | |
|
|
1058 | FREETMPS; LEAVE; |
|
|
1059 | } |
|
|
1060 | break; |
|
|
1061 | } |
|
|
1062 | |
642 | return sv; |
1063 | return sv; |
643 | |
1064 | |
644 | AV *av = newAV (); |
1065 | fail: |
645 | av_push (av, newSVuv (tag)); |
1066 | SvREFCNT_dec (sv); |
646 | av_push (av, sv); |
1067 | return &PL_sv_undef; |
647 | |
|
|
648 | HV *tagged_stash = !CBOR_SLOW || cbor_tagged_stash |
|
|
649 | ? cbor_tagged_stash |
|
|
650 | : gv_stashpv ("CBOR::XS::Tagged" , 1); |
|
|
651 | |
|
|
652 | return sv_bless (newRV_noinc ((SV *)av), tagged_stash); |
|
|
653 | } |
1068 | } |
654 | |
1069 | |
655 | static SV * |
1070 | static SV * |
656 | decode_sv (dec_t *dec) |
1071 | decode_sv (dec_t *dec) |
657 | { |
1072 | { |
658 | WANT (1); |
1073 | WANT (1); |
659 | |
1074 | |
660 | switch (*dec->cur >> 5) |
1075 | switch (*dec->cur >> MAJOR_SHIFT) |
661 | { |
1076 | { |
662 | case 0: // unsigned int |
1077 | case MAJOR_POS_INT >> MAJOR_SHIFT: return newSVuv (decode_uint (dec)); |
663 | return newSVuv (decode_uint (dec)); |
1078 | case MAJOR_NEG_INT >> MAJOR_SHIFT: return newSViv (-1 - (IV)decode_uint (dec)); |
664 | case 1: // negative int |
1079 | case MAJOR_BYTES >> MAJOR_SHIFT: return decode_str (dec, 0); |
665 | return newSViv (-1 - (IV)decode_uint (dec)); |
1080 | case MAJOR_TEXT >> MAJOR_SHIFT: return decode_str (dec, 1); |
666 | case 2: // octet string |
1081 | case MAJOR_ARRAY >> MAJOR_SHIFT: return decode_av (dec); |
667 | return decode_str (dec, 0); |
1082 | case MAJOR_MAP >> MAJOR_SHIFT: return decode_hv (dec); |
668 | case 3: // utf-8 string |
1083 | case MAJOR_TAG >> MAJOR_SHIFT: return decode_tagged (dec); |
669 | return decode_str (dec, 1); |
1084 | |
670 | case 4: // array |
1085 | case MAJOR_MISC >> MAJOR_SHIFT: |
671 | return decode_av (dec); |
|
|
672 | case 5: // map |
|
|
673 | return decode_hv (dec); |
|
|
674 | case 6: // tag |
|
|
675 | return decode_tagged (dec); |
|
|
676 | case 7: // misc |
|
|
677 | switch (*dec->cur++ & 31) |
1086 | switch (*dec->cur++ & MINOR_MASK) |
678 | { |
1087 | { |
679 | case 20: |
1088 | case SIMPLE_FALSE: |
680 | #if CBOR_SLOW |
1089 | #if CBOR_SLOW |
681 | cbor_false = get_bool ("CBOR::XS::false"); |
1090 | types_false = get_bool ("Types::Serialiser::false"); |
682 | #endif |
1091 | #endif |
683 | return newSVsv (cbor_false); |
1092 | return newSVsv (types_false); |
684 | case 21: |
1093 | case SIMPLE_TRUE: |
685 | #if CBOR_SLOW |
1094 | #if CBOR_SLOW |
686 | cbor_true = get_bool ("CBOR::XS::true"); |
1095 | types_true = get_bool ("Types::Serialiser::true"); |
687 | #endif |
1096 | #endif |
688 | return newSVsv (cbor_true); |
1097 | return newSVsv (types_true); |
689 | case 22: |
1098 | case SIMPLE_NULL: |
690 | return newSVsv (&PL_sv_undef); |
1099 | return newSVsv (&PL_sv_undef); |
|
|
1100 | case SIMPLE_UNDEF: |
|
|
1101 | #if CBOR_SLOW |
|
|
1102 | types_error = get_bool ("Types::Serialiser::error"); |
|
|
1103 | #endif |
|
|
1104 | return newSVsv (types_error); |
691 | |
1105 | |
692 | case 25: |
1106 | case MISC_FLOAT16: |
693 | { |
1107 | { |
694 | WANT (2); |
1108 | WANT (2); |
695 | |
1109 | |
696 | uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; |
1110 | uint16_t fp = (dec->cur[0] << 8) | dec->cur[1]; |
697 | dec->cur += 2; |
1111 | dec->cur += 2; |
698 | |
1112 | |
699 | return newSVnv (ecb_binary16_to_float (fp)); |
1113 | return newSVnv (ecb_binary16_to_float (fp)); |
700 | } |
1114 | } |
701 | |
1115 | |
702 | case 26: |
1116 | case MISC_FLOAT32: |
703 | { |
1117 | { |
704 | uint32_t fp; |
1118 | uint32_t fp; |
705 | WANT (4); |
1119 | WANT (4); |
706 | memcpy (&fp, dec->cur, 4); |
1120 | memcpy (&fp, dec->cur, 4); |
707 | dec->cur += 4; |
1121 | dec->cur += 4; |
… | |
… | |
710 | fp = ecb_bswap32 (fp); |
1124 | fp = ecb_bswap32 (fp); |
711 | |
1125 | |
712 | return newSVnv (ecb_binary32_to_float (fp)); |
1126 | return newSVnv (ecb_binary32_to_float (fp)); |
713 | } |
1127 | } |
714 | |
1128 | |
715 | case 27: |
1129 | case MISC_FLOAT64: |
716 | { |
1130 | { |
717 | uint64_t fp; |
1131 | uint64_t fp; |
718 | WANT (8); |
1132 | WANT (8); |
719 | memcpy (&fp, dec->cur, 8); |
1133 | memcpy (&fp, dec->cur, 8); |
720 | dec->cur += 8; |
1134 | dec->cur += 8; |
… | |
… | |
723 | fp = ecb_bswap64 (fp); |
1137 | fp = ecb_bswap64 (fp); |
724 | |
1138 | |
725 | return newSVnv (ecb_binary64_to_double (fp)); |
1139 | return newSVnv (ecb_binary64_to_double (fp)); |
726 | } |
1140 | } |
727 | |
1141 | |
728 | // 0..19 unassigned |
1142 | // 0..19 unassigned simple |
729 | // 24 reserved + unassigned (reserved values are not encodable) |
1143 | // 24 reserved + unassigned simple (reserved values are not encodable) |
|
|
1144 | // 28-30 unassigned misc |
|
|
1145 | // 31 break code |
730 | default: |
1146 | default: |
731 | ERR ("corrupted CBOR data (reserved/unassigned major 7 value)"); |
1147 | ERR ("corrupted CBOR data (reserved/unassigned/unexpected major 7 value)"); |
732 | } |
1148 | } |
733 | |
1149 | |
734 | break; |
1150 | break; |
735 | } |
1151 | } |
736 | |
1152 | |
… | |
… | |
739 | } |
1155 | } |
740 | |
1156 | |
741 | static SV * |
1157 | static SV * |
742 | decode_cbor (SV *string, CBOR *cbor, char **offset_return) |
1158 | decode_cbor (SV *string, CBOR *cbor, char **offset_return) |
743 | { |
1159 | { |
744 | dec_t dec; |
1160 | dec_t dec = { }; |
745 | SV *sv; |
1161 | SV *sv; |
|
|
1162 | STRLEN len; |
|
|
1163 | char *data = SvPVbyte (string, len); |
746 | |
1164 | |
747 | /* work around bugs in 5.10 where manipulating magic values |
|
|
748 | * makes perl ignore the magic in subsequent accesses. |
|
|
749 | * also make a copy of non-PV values, to get them into a clean |
|
|
750 | * state (SvPV should do that, but it's buggy, see below). |
|
|
751 | */ |
|
|
752 | /*SvGETMAGIC (string);*/ |
|
|
753 | if (SvMAGICAL (string) || !SvPOK (string)) |
|
|
754 | string = sv_2mortal (newSVsv (string)); |
|
|
755 | |
|
|
756 | SvUPGRADE (string, SVt_PV); |
|
|
757 | |
|
|
758 | /* work around a bug in perl 5.10, which causes SvCUR to fail an |
|
|
759 | * assertion with -DDEBUGGING, although SvCUR is documented to |
|
|
760 | * return the xpv_cur field which certainly exists after upgrading. |
|
|
761 | * according to nicholas clark, calling SvPOK fixes this. |
|
|
762 | * But it doesn't fix it, so try another workaround, call SvPV_nolen |
|
|
763 | * and hope for the best. |
|
|
764 | * Damnit, SvPV_nolen still trips over yet another assertion. This |
|
|
765 | * assertion business is seriously broken, try yet another workaround |
|
|
766 | * for the broken -DDEBUGGING. |
|
|
767 | */ |
|
|
768 | { |
|
|
769 | #ifdef DEBUGGING |
|
|
770 | STRLEN offset = SvOK (string) ? sv_len (string) : 0; |
|
|
771 | #else |
|
|
772 | STRLEN offset = SvCUR (string); |
|
|
773 | #endif |
|
|
774 | |
|
|
775 | if (offset > cbor->max_size && cbor->max_size) |
1165 | if (len > cbor->max_size && cbor->max_size) |
776 | croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu", |
1166 | croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu", |
777 | (unsigned long)SvCUR (string), (unsigned long)cbor->max_size); |
1167 | (unsigned long)len, (unsigned long)cbor->max_size); |
778 | } |
|
|
779 | |
|
|
780 | sv_utf8_downgrade (string, 0); |
|
|
781 | |
1168 | |
782 | dec.cbor = *cbor; |
1169 | dec.cbor = *cbor; |
783 | dec.cur = (U8 *)SvPVX (string); |
1170 | dec.cur = (U8 *)data; |
784 | dec.end = (U8 *)SvEND (string); |
1171 | dec.end = (U8 *)data + len; |
785 | dec.err = 0; |
|
|
786 | dec.depth = 0; |
|
|
787 | |
1172 | |
788 | sv = decode_sv (&dec); |
1173 | sv = decode_sv (&dec); |
789 | |
1174 | |
790 | if (offset_return) |
1175 | if (offset_return) |
791 | *offset_return = dec.cur; |
1176 | *offset_return = dec.cur; |
… | |
… | |
794 | if (dec.cur != dec.end && !dec.err) |
1179 | if (dec.cur != dec.end && !dec.err) |
795 | dec.err = "garbage after CBOR object"; |
1180 | dec.err = "garbage after CBOR object"; |
796 | |
1181 | |
797 | if (dec.err) |
1182 | if (dec.err) |
798 | { |
1183 | { |
|
|
1184 | if (dec.shareable) |
|
|
1185 | { |
|
|
1186 | // need to break cyclic links, which whould all be in shareable |
|
|
1187 | int i; |
|
|
1188 | SV **svp; |
|
|
1189 | |
|
|
1190 | for (i = av_len (dec.shareable) + 1; i--; ) |
|
|
1191 | if ((svp = av_fetch (dec.shareable, i, 0))) |
|
|
1192 | sv_setsv (*svp, &PL_sv_undef); |
|
|
1193 | } |
|
|
1194 | |
799 | SvREFCNT_dec (sv); |
1195 | SvREFCNT_dec (sv); |
800 | croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)SvPVX (string), (int)(uint8_t)*dec.cur); |
1196 | croak ("%s, at offset %d (octet 0x%02x)", dec.err, dec.cur - (U8 *)data, (int)(uint8_t)*dec.cur); |
801 | } |
1197 | } |
802 | |
1198 | |
803 | sv = sv_2mortal (sv); |
1199 | sv = sv_2mortal (sv); |
804 | |
1200 | |
805 | return sv; |
1201 | return sv; |
806 | } |
1202 | } |
807 | |
1203 | |
|
|
1204 | ///////////////////////////////////////////////////////////////////////////// |
|
|
1205 | // incremental parser |
|
|
1206 | |
|
|
1207 | #define INCR_DONE(cbor) (AvFILLp (cbor->incr_count) < 0) |
|
|
1208 | |
|
|
1209 | // returns 0 for notyet, 1 for success or error |
|
|
1210 | static int |
|
|
1211 | incr_parse (CBOR *self, SV *cborstr) |
|
|
1212 | { |
|
|
1213 | STRLEN cur; |
|
|
1214 | SvPV (cborstr, cur); |
|
|
1215 | |
|
|
1216 | while (ecb_expect_true (self->incr_need <= cur)) |
|
|
1217 | { |
|
|
1218 | // table of integer count bytes |
|
|
1219 | static I8 incr_len[MINOR_MASK + 1] = { |
|
|
1220 | 0, 0, 0, 0, 0, 0, 0, 0, |
|
|
1221 | 0, 0, 0, 0, 0, 0, 0, 0, |
|
|
1222 | 0, 0, 0, 0, 0, 0, 0, 0, |
|
|
1223 | 1, 2, 4, 8,-1,-1,-1,-2 |
|
|
1224 | }; |
|
|
1225 | |
|
|
1226 | const U8 *p = SvPVX (cborstr) + self->incr_pos; |
|
|
1227 | U8 m = *p & MINOR_MASK; |
|
|
1228 | IV count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]); |
|
|
1229 | I8 ilen = incr_len[m]; |
|
|
1230 | |
|
|
1231 | self->incr_need = self->incr_pos + 1; |
|
|
1232 | |
|
|
1233 | if (ecb_expect_false (ilen < 0)) |
|
|
1234 | { |
|
|
1235 | if (m != MINOR_INDEF) |
|
|
1236 | return 1; // error |
|
|
1237 | |
|
|
1238 | if (*p == (MAJOR_MISC | MINOR_INDEF)) |
|
|
1239 | { |
|
|
1240 | if (count >= 0) |
|
|
1241 | return 1; // error |
|
|
1242 | |
|
|
1243 | count = 1; |
|
|
1244 | } |
|
|
1245 | else |
|
|
1246 | { |
|
|
1247 | av_push (self->incr_count, newSViv (-1)); //TODO: nest |
|
|
1248 | count = -1; |
|
|
1249 | } |
|
|
1250 | } |
|
|
1251 | else |
|
|
1252 | { |
|
|
1253 | self->incr_need += ilen; |
|
|
1254 | if (ecb_expect_false (self->incr_need > cur)) |
|
|
1255 | return 0; |
|
|
1256 | |
|
|
1257 | int major = *p >> MAJOR_SHIFT; |
|
|
1258 | |
|
|
1259 | switch (major) |
|
|
1260 | { |
|
|
1261 | case MAJOR_BYTES >> MAJOR_SHIFT: |
|
|
1262 | case MAJOR_TEXT >> MAJOR_SHIFT: |
|
|
1263 | case MAJOR_ARRAY >> MAJOR_SHIFT: |
|
|
1264 | case MAJOR_MAP >> MAJOR_SHIFT: |
|
|
1265 | { |
|
|
1266 | UV len; |
|
|
1267 | |
|
|
1268 | if (ecb_expect_false (ilen)) |
|
|
1269 | { |
|
|
1270 | len = 0; |
|
|
1271 | |
|
|
1272 | do { |
|
|
1273 | len = (len << 8) | *++p; |
|
|
1274 | } while (--ilen); |
|
|
1275 | } |
|
|
1276 | else |
|
|
1277 | len = m; |
|
|
1278 | |
|
|
1279 | switch (major) |
|
|
1280 | { |
|
|
1281 | case MAJOR_BYTES >> MAJOR_SHIFT: |
|
|
1282 | case MAJOR_TEXT >> MAJOR_SHIFT: |
|
|
1283 | self->incr_need += len; |
|
|
1284 | if (ecb_expect_false (self->incr_need > cur)) |
|
|
1285 | return 0; |
|
|
1286 | |
|
|
1287 | break; |
|
|
1288 | |
|
|
1289 | case MAJOR_MAP >> MAJOR_SHIFT: |
|
|
1290 | len <<= 1; |
|
|
1291 | case MAJOR_ARRAY >> MAJOR_SHIFT: |
|
|
1292 | if (len) |
|
|
1293 | { |
|
|
1294 | av_push (self->incr_count, newSViv (len + 1)); //TODO: nest |
|
|
1295 | count = len + 1; |
|
|
1296 | } |
|
|
1297 | break; |
|
|
1298 | } |
|
|
1299 | } |
|
|
1300 | } |
|
|
1301 | } |
|
|
1302 | |
|
|
1303 | self->incr_pos = self->incr_need; |
|
|
1304 | |
|
|
1305 | if (count > 0) |
|
|
1306 | { |
|
|
1307 | while (!--count) |
|
|
1308 | { |
|
|
1309 | if (!AvFILLp (self->incr_count)) |
|
|
1310 | return 1; // done |
|
|
1311 | |
|
|
1312 | SvREFCNT_dec_NN (av_pop (self->incr_count)); |
|
|
1313 | count = SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]); |
|
|
1314 | } |
|
|
1315 | |
|
|
1316 | SvIVX (AvARRAY (self->incr_count)[AvFILLp (self->incr_count)]) = count; |
|
|
1317 | } |
|
|
1318 | } |
|
|
1319 | |
|
|
1320 | return 0; |
|
|
1321 | } |
|
|
1322 | |
|
|
1323 | |
808 | ///////////////////////////////////////////////////////////////////////////// |
1324 | ///////////////////////////////////////////////////////////////////////////// |
809 | // XS interface functions |
1325 | // XS interface functions |
810 | |
1326 | |
811 | MODULE = CBOR::XS PACKAGE = CBOR::XS |
1327 | MODULE = CBOR::XS PACKAGE = CBOR::XS |
812 | |
1328 | |
813 | BOOT: |
1329 | BOOT: |
814 | { |
1330 | { |
815 | cbor_stash = gv_stashpv ("CBOR::XS" , 1); |
1331 | cbor_stash = gv_stashpv ("CBOR::XS" , 1); |
816 | cbor_boolean_stash = gv_stashpv ("CBOR::XS::Boolean", 1); |
|
|
817 | cbor_tagged_stash = gv_stashpv ("CBOR::XS::Tagged" , 1); |
1332 | cbor_tagged_stash = gv_stashpv ("CBOR::XS::Tagged" , 1); |
818 | |
1333 | |
819 | cbor_true = get_bool ("CBOR::XS::true"); |
1334 | types_boolean_stash = gv_stashpv ("Types::Serialiser::Boolean", 1); |
820 | cbor_false = get_bool ("CBOR::XS::false"); |
1335 | types_error_stash = gv_stashpv ("Types::Serialiser::Error" , 1); |
|
|
1336 | |
|
|
1337 | types_true = get_bool ("Types::Serialiser::true" ); |
|
|
1338 | types_false = get_bool ("Types::Serialiser::false"); |
|
|
1339 | types_error = get_bool ("Types::Serialiser::error"); |
|
|
1340 | |
|
|
1341 | default_filter = newSVpv ("CBOR::XS::default_filter", 0); |
|
|
1342 | |
|
|
1343 | sv_cbor = newSVpv ("CBOR", 0); |
|
|
1344 | SvREADONLY_on (sv_cbor); |
821 | } |
1345 | } |
822 | |
1346 | |
823 | PROTOTYPES: DISABLE |
1347 | PROTOTYPES: DISABLE |
824 | |
1348 | |
825 | void CLONE (...) |
1349 | void CLONE (...) |
826 | CODE: |
1350 | CODE: |
827 | cbor_stash = 0; |
1351 | cbor_stash = 0; |
828 | cbor_boolean_stash = 0; |
|
|
829 | cbor_tagged_stash = 0; |
1352 | cbor_tagged_stash = 0; |
|
|
1353 | types_error_stash = 0; |
|
|
1354 | types_boolean_stash = 0; |
830 | |
1355 | |
831 | void new (char *klass) |
1356 | void new (char *klass) |
832 | PPCODE: |
1357 | PPCODE: |
833 | { |
1358 | { |
834 | SV *pv = NEWSV (0, sizeof (CBOR)); |
1359 | SV *pv = NEWSV (0, sizeof (CBOR)); |
… | |
… | |
842 | |
1367 | |
843 | void shrink (CBOR *self, int enable = 1) |
1368 | void shrink (CBOR *self, int enable = 1) |
844 | ALIAS: |
1369 | ALIAS: |
845 | shrink = F_SHRINK |
1370 | shrink = F_SHRINK |
846 | allow_unknown = F_ALLOW_UNKNOWN |
1371 | allow_unknown = F_ALLOW_UNKNOWN |
|
|
1372 | allow_sharing = F_ALLOW_SHARING |
|
|
1373 | allow_cycles = F_ALLOW_CYCLES |
|
|
1374 | pack_strings = F_PACK_STRINGS |
|
|
1375 | validate_utf8 = F_VALIDATE_UTF8 |
847 | PPCODE: |
1376 | PPCODE: |
848 | { |
1377 | { |
849 | if (enable) |
1378 | if (enable) |
850 | self->flags |= ix; |
1379 | self->flags |= ix; |
851 | else |
1380 | else |
… | |
… | |
856 | |
1385 | |
857 | void get_shrink (CBOR *self) |
1386 | void get_shrink (CBOR *self) |
858 | ALIAS: |
1387 | ALIAS: |
859 | get_shrink = F_SHRINK |
1388 | get_shrink = F_SHRINK |
860 | get_allow_unknown = F_ALLOW_UNKNOWN |
1389 | get_allow_unknown = F_ALLOW_UNKNOWN |
|
|
1390 | get_allow_sharing = F_ALLOW_SHARING |
|
|
1391 | get_allow_cycles = F_ALLOW_CYCLES |
|
|
1392 | get_pack_strings = F_PACK_STRINGS |
|
|
1393 | get_validate_utf8 = F_VALIDATE_UTF8 |
861 | PPCODE: |
1394 | PPCODE: |
862 | XPUSHs (boolSV (self->flags & ix)); |
1395 | XPUSHs (boolSV (self->flags & ix)); |
863 | |
1396 | |
864 | void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) |
1397 | void max_depth (CBOR *self, U32 max_depth = 0x80000000UL) |
865 | PPCODE: |
1398 | PPCODE: |
… | |
… | |
881 | CODE: |
1414 | CODE: |
882 | RETVAL = self->max_size; |
1415 | RETVAL = self->max_size; |
883 | OUTPUT: |
1416 | OUTPUT: |
884 | RETVAL |
1417 | RETVAL |
885 | |
1418 | |
886 | #if 0 //TODO |
1419 | void filter (CBOR *self, SV *filter = 0) |
887 | |
|
|
888 | void filter_cbor_object (CBOR *self, SV *cb = &PL_sv_undef) |
|
|
889 | PPCODE: |
1420 | PPCODE: |
890 | { |
|
|
891 | SvREFCNT_dec (self->cb_object); |
1421 | SvREFCNT_dec (self->filter); |
892 | self->cb_object = SvOK (cb) ? newSVsv (cb) : 0; |
1422 | self->filter = filter ? newSVsv (filter) : filter; |
893 | |
|
|
894 | XPUSHs (ST (0)); |
1423 | XPUSHs (ST (0)); |
895 | } |
|
|
896 | |
1424 | |
897 | void filter_cbor_single_key_object (CBOR *self, SV *key, SV *cb = &PL_sv_undef) |
1425 | SV *get_filter (CBOR *self) |
898 | PPCODE: |
1426 | CODE: |
899 | { |
1427 | RETVAL = self->filter ? self->filter : NEWSV (0, 0); |
900 | if (!self->cb_sk_object) |
1428 | OUTPUT: |
901 | self->cb_sk_object = newHV (); |
1429 | RETVAL |
902 | |
|
|
903 | if (SvOK (cb)) |
|
|
904 | hv_store_ent (self->cb_sk_object, key, newSVsv (cb), 0); |
|
|
905 | else |
|
|
906 | { |
|
|
907 | hv_delete_ent (self->cb_sk_object, key, G_DISCARD, 0); |
|
|
908 | |
|
|
909 | if (!HvKEYS (self->cb_sk_object)) |
|
|
910 | { |
|
|
911 | SvREFCNT_dec (self->cb_sk_object); |
|
|
912 | self->cb_sk_object = 0; |
|
|
913 | } |
|
|
914 | } |
|
|
915 | |
|
|
916 | XPUSHs (ST (0)); |
|
|
917 | } |
|
|
918 | |
|
|
919 | #endif |
|
|
920 | |
1430 | |
921 | void encode (CBOR *self, SV *scalar) |
1431 | void encode (CBOR *self, SV *scalar) |
922 | PPCODE: |
1432 | PPCODE: |
923 | PUTBACK; scalar = encode_cbor (scalar, self); SPAGAIN; |
1433 | PUTBACK; scalar = encode_cbor (scalar, self); SPAGAIN; |
924 | XPUSHs (scalar); |
1434 | XPUSHs (scalar); |
… | |
… | |
937 | EXTEND (SP, 2); |
1447 | EXTEND (SP, 2); |
938 | PUSHs (sv); |
1448 | PUSHs (sv); |
939 | PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); |
1449 | PUSHs (sv_2mortal (newSVuv (offset - SvPVX (cborstr)))); |
940 | } |
1450 | } |
941 | |
1451 | |
942 | #if 0 |
1452 | void incr_parse (CBOR *self, SV *cborstr) |
|
|
1453 | ALIAS: |
|
|
1454 | incr_parse_multiple = 1 |
|
|
1455 | PPCODE: |
|
|
1456 | { |
|
|
1457 | if (SvUTF8 (cborstr)) |
|
|
1458 | sv_utf8_downgrade (cborstr, 0); |
|
|
1459 | |
|
|
1460 | if (!self->incr_count) |
|
|
1461 | { |
|
|
1462 | self->incr_count = newAV (); |
|
|
1463 | self->incr_pos = 0; |
|
|
1464 | self->incr_need = 1; |
|
|
1465 | |
|
|
1466 | av_push (self->incr_count, newSViv (1)); |
|
|
1467 | } |
|
|
1468 | |
|
|
1469 | do |
|
|
1470 | { |
|
|
1471 | if (!incr_parse (self, cborstr)) |
|
|
1472 | { |
|
|
1473 | if (self->incr_need > self->max_size && self->max_size) |
|
|
1474 | croak ("attempted decode of CBOR text of %lu bytes size, but max_size is set to %lu", |
|
|
1475 | (unsigned long)self->incr_need, (unsigned long)self->max_size); |
|
|
1476 | |
|
|
1477 | break; |
|
|
1478 | } |
|
|
1479 | |
|
|
1480 | SV *sv; |
|
|
1481 | char *offset; |
|
|
1482 | |
|
|
1483 | PUTBACK; sv = decode_cbor (cborstr, self, &offset); SPAGAIN; |
|
|
1484 | XPUSHs (sv); |
|
|
1485 | |
|
|
1486 | sv_chop (cborstr, offset); |
|
|
1487 | |
|
|
1488 | av_clear (self->incr_count); |
|
|
1489 | av_push (self->incr_count, newSViv (1)); |
|
|
1490 | |
|
|
1491 | self->incr_pos = 0; |
|
|
1492 | self->incr_need = self->incr_pos + 1; |
|
|
1493 | } |
|
|
1494 | while (ix); |
|
|
1495 | } |
|
|
1496 | |
|
|
1497 | void incr_reset (CBOR *self) |
|
|
1498 | CODE: |
|
|
1499 | { |
|
|
1500 | SvREFCNT_dec (self->incr_count); |
|
|
1501 | self->incr_count = 0; |
|
|
1502 | } |
943 | |
1503 | |
944 | void DESTROY (CBOR *self) |
1504 | void DESTROY (CBOR *self) |
945 | CODE: |
1505 | PPCODE: |
946 | SvREFCNT_dec (self->cb_sk_object); |
1506 | cbor_free (self); |
947 | SvREFCNT_dec (self->cb_object); |
|
|
948 | |
|
|
949 | #endif |
|
|
950 | |
1507 | |
951 | PROTOTYPES: ENABLE |
1508 | PROTOTYPES: ENABLE |
952 | |
1509 | |
953 | void encode_cbor (SV *scalar) |
1510 | void encode_cbor (SV *scalar) |
|
|
1511 | ALIAS: |
|
|
1512 | encode_cbor = 0 |
|
|
1513 | encode_cbor_sharing = F_ALLOW_SHARING |
954 | PPCODE: |
1514 | PPCODE: |
955 | { |
1515 | { |
956 | CBOR cbor; |
1516 | CBOR cbor; |
957 | cbor_init (&cbor); |
1517 | cbor_init (&cbor); |
|
|
1518 | cbor.flags |= ix; |
958 | PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; |
1519 | PUTBACK; scalar = encode_cbor (scalar, &cbor); SPAGAIN; |
959 | XPUSHs (scalar); |
1520 | XPUSHs (scalar); |
960 | } |
1521 | } |
961 | |
1522 | |
962 | void decode_cbor (SV *cborstr) |
1523 | void decode_cbor (SV *cborstr) |