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

Comparing CBOR-XS/README (file contents):
Revision 1.9 by root, Fri Nov 22 16:18:59 2013 UTC vs.
Revision 1.18 by root, Wed Dec 7 14:14:30 2016 UTC

… …
21 # data was decoded 21 # data was decoded
22 substr $many_cbor_strings, 0, $length, ""; # remove decoded cbor string 22 substr $many_cbor_strings, 0, $length, ""; # remove decoded cbor string
23 } 23 }
24 24
25DESCRIPTION 25DESCRIPTION
26 WARNING! This module is very new, and not very well tested (that's up to
27 you to do). Furthermore, details of the implementation might change
28 freely before version 1.0. And lastly, most extensions depend on an IANA
29 assignment, and until that assignment is official, this implementation
30 is not interoperable with other implementations (even future versions of
31 this module) until the assignment is done.
32
33 You are still invited to try out CBOR, and this module.
34
35 This module converts Perl data structures to the Concise Binary Object 26 This module converts Perl data structures to the Concise Binary Object
36 Representation (CBOR) and vice versa. CBOR is a fast binary 27 Representation (CBOR) and vice versa. CBOR is a fast binary
37 serialisation format that aims to use a superset of the JSON data model, 28 serialisation format that aims to use an (almost) superset of the JSON
38 i.e. when you can represent something in JSON, you should be able to 29 data model, i.e. when you can represent something useful in JSON, you
39 represent it in CBOR. 30 should be able to represent it in CBOR.
40 31
41 In short, CBOR is a faster and very compact binary alternative to JSON, 32 In short, CBOR is a faster and quite compact binary alternative to JSON,
42 with the added ability of supporting serialisation of Perl objects. 33 with the added ability of supporting serialisation of Perl objects.
43 (JSON often compresses better than CBOR though, so if you plan to 34 (JSON often compresses better than CBOR though, so if you plan to
44 compress the data later you might want to compare both formats first). 35 compress the data later and speed is less important you might want to
36 compare both formats first).
45 37
46 To give you a general idea about speed, with texts in the megabyte 38 To give you a general idea about speed, with texts in the megabyte
47 range, "CBOR::XS" usually encodes roughly twice as fast as Storable or 39 range, "CBOR::XS" usually encodes roughly twice as fast as Storable or
48 JSON::XS and decodes about 15%-30% faster than those. The shorter the 40 JSON::XS and decodes about 15%-30% faster than those. The shorter the
49 data, the worse Storable performs in comparison. 41 data, the worse Storable performs in comparison.
50 42
51 As for compactness, "CBOR::XS" encoded data structures are usually about 43 Regarding compactness, "CBOR::XS"-encoded data structures are usually
52 20% smaller than the same data encoded as (compact) JSON or Storable. 44 about 20% smaller than the same data encoded as (compact) JSON or
45 Storable.
53 46
54 In addition to the core CBOR data format, this module implements a 47 In addition to the core CBOR data format, this module implements a
55 number of extensions, to support cyclic and self-referencing data 48 number of extensions, to support cyclic and shared data structures (see
56 structures (see "allow_sharing"), string deduplication (see 49 "allow_sharing" and "allow_cycles"), string deduplication (see
57 "allow_stringref") and scalar references (always enabled). 50 "pack_strings") and scalar references (always enabled).
58 51
59 The primary goal of this module is to be *correct* and the secondary 52 The primary goal of this module is to be *correct* and the secondary
60 goal is to be *fast*. To reach the latter goal it was written in C. 53 goal is to be *fast*. To reach the latter goal it was written in C.
61 54
62 See MAPPING, below, on how CBOR::XS maps perl values to CBOR values and 55 See MAPPING, below, on how CBOR::XS maps perl values to CBOR values and
… …
85 78
86 The mutators for flags all return the CBOR object again and thus 79 The mutators for flags all return the CBOR object again and thus
87 calls can be chained: 80 calls can be chained:
88 81
89 my $cbor = CBOR::XS->new->encode ({a => [1,2]}); 82 my $cbor = CBOR::XS->new->encode ({a => [1,2]});
83
84 $cbor = new_safe CBOR::XS
85 Create a new, safe/secure CBOR::XS object. This is similar to "new",
86 but configures the coder object to be safe to use with untrusted
87 data. Currently, this is equivalent to:
88
89 my $cbor = CBOR::XS
90 ->new
91 ->forbid_objects
92 ->filter (\&CBOR::XS::safe_filter)
93 ->max_size (1e8);
94
95 But is more future proof (it is better to crash because of a change
96 than to be exploited in other ways).
90 97
91 $cbor = $cbor->max_depth ([$maximum_nesting_depth]) 98 $cbor = $cbor->max_depth ([$maximum_nesting_depth])
92 $max_depth = $cbor->get_max_depth 99 $max_depth = $cbor->get_max_depth
93 Sets the maximum nesting level (default 512) accepted while encoding 100 Sets the maximum nesting level (default 512) accepted while encoding
94 or decoding. If a higher nesting level is detected in CBOR data or a 101 or decoding. If a higher nesting level is detected in CBOR data or a
… …
108 115
109 Note that nesting is implemented by recursion in C. The default 116 Note that nesting is implemented by recursion in C. The default
110 value has been chosen to be as large as typical operating systems 117 value has been chosen to be as large as typical operating systems
111 allow without crashing. 118 allow without crashing.
112 119
113 See SECURITY CONSIDERATIONS, below, for more info on why this is 120 See "SECURITY CONSIDERATIONS", below, for more info on why this is
114 useful. 121 useful.
115 122
116 $cbor = $cbor->max_size ([$maximum_string_size]) 123 $cbor = $cbor->max_size ([$maximum_string_size])
117 $max_size = $cbor->get_max_size 124 $max_size = $cbor->get_max_size
118 Set the maximum length a CBOR string may have (in bytes) where 125 Set the maximum length a CBOR string may have (in bytes) where
… …
122 exception. This setting has no effect on "encode" (yet). 129 exception. This setting has no effect on "encode" (yet).
123 130
124 If no argument is given, the limit check will be deactivated (same 131 If no argument is given, the limit check will be deactivated (same
125 as when 0 is specified). 132 as when 0 is specified).
126 133
127 See SECURITY CONSIDERATIONS, below, for more info on why this is 134 See "SECURITY CONSIDERATIONS", below, for more info on why this is
128 useful. 135 useful.
129 136
130 $cbor = $cbor->allow_unknown ([$enable]) 137 $cbor = $cbor->allow_unknown ([$enable])
131 $enabled = $cbor->get_allow_unknown 138 $enabled = $cbor->get_allow_unknown
132 If $enable is true (or missing), then "encode" will *not* throw an 139 If $enable is true (or missing), then "encode" will *not* throw an
… …
147 same object, such as an array, is referenced multiple times), but 154 same object, such as an array, is referenced multiple times), but
148 instead will emit a reference to the earlier value. 155 instead will emit a reference to the earlier value.
149 156
150 This means that such values will only be encoded once, and will not 157 This means that such values will only be encoded once, and will not
151 result in a deep cloning of the value on decode, in decoders 158 result in a deep cloning of the value on decode, in decoders
152 supporting the value sharing extension. 159 supporting the value sharing extension. This also makes it possible
160 to encode cyclic data structures (which need "allow_cycles" to be
161 enabled to be decoded by this module).
153 162
154 It is recommended to leave it off unless you know your communication 163 It is recommended to leave it off unless you know your communication
155 partner supports the value sharing extensions to CBOR 164 partner supports the value sharing extensions to CBOR
156 (http://cbor.schmorp.de/value-sharing). 165 (<http://cbor.schmorp.de/value-sharing>), as without decoder
166 support, the resulting data structure might be unusable.
157 167
158 Detecting shared values incurs a runtime overhead when values are 168 Detecting shared values incurs a runtime overhead when values are
159 encoded that have a reference counter large than one, and might 169 encoded that have a reference counter large than one, and might
160 unnecessarily increase the encoded size, as potentially shared 170 unnecessarily increase the encoded size, as potentially shared
161 values are encode as sharable whether or not they are actually 171 values are encode as shareable whether or not they are actually
162 shared. 172 shared.
163 173
164 At the moment, only targets of references can be shared (e.g. 174 At the moment, only targets of references can be shared (e.g.
165 scalars, arrays or hashes pointed to by a reference). Weirder 175 scalars, arrays or hashes pointed to by a reference). Weirder
166 constructs, such as an array with multiple "copies" of the *same* 176 constructs, such as an array with multiple "copies" of the *same*
167 string, which are hard but not impossible to create in Perl, are not 177 string, which are hard but not impossible to create in Perl, are not
168 supported (this is the same as for Storable). 178 supported (this is the same as with Storable).
169 179
170 If $enable is false (the default), then "encode" will encode 180 If $enable is false (the default), then "encode" will encode shared
171 exception when it encounters anything it cannot encode as CBOR. 181 data structures repeatedly, unsharing them in the process. Cyclic
182 data structures cannot be encoded in this mode.
172 183
173 This option does not affect "decode" in any way - shared values and 184 This option does not affect "decode" in any way - shared values and
174 references will always be decoded properly if present. 185 references will always be decoded properly if present.
175 186
187 $cbor = $cbor->allow_cycles ([$enable])
188 $enabled = $cbor->get_allow_cycles
189 If $enable is true (or missing), then "decode" will happily decode
190 self-referential (cyclic) data structures. By default these will not
191 be decoded, as they need manual cleanup to avoid memory leaks, so
192 code that isn't prepared for this will not leak memory.
193
194 If $enable is false (the default), then "decode" will throw an error
195 when it encounters a self-referential/cyclic data structure.
196
197 FUTURE DIRECTION: the motivation behind this option is to avoid
198 *real* cycles - future versions of this module might chose to decode
199 cyclic data structures using weak references when this option is
200 off, instead of throwing an error.
201
202 This option does not affect "encode" in any way - shared values and
203 references will always be encoded properly if present.
204
205 $cbor = $cbor->forbid_objects ([$enable])
206 $enabled = $cbor->get_forbid_objects
207 Disables the use of the object serialiser protocol.
208
209 If $enable is true (or missing), then "encode" will will throw an
210 exception when it encounters perl objects that would be encoded
211 using the perl-object tag (26). When "decode" encounters such tags,
212 it will fall back to the general filter/tagged logic as if this were
213 an unknown tag (by default resulting in a "CBOR::XC::Tagged"
214 object).
215
216 If $enable is false (the default), then "encode" will use the
217 Types::Serialiser object serialisation protocol to serialise objects
218 into perl-object tags, and "decode" will do the same to decode such
219 tags.
220
221 See "SECURITY CONSIDERATIONS", below, for more info on why
222 forbidding this protocol can be useful.
223
176 $cbor = $cbor->allow_stringref ([$enable]) 224 $cbor = $cbor->pack_strings ([$enable])
177 $enabled = $cbor->get_allow_stringref 225 $enabled = $cbor->get_pack_strings
178 If $enable is true (or missing), then "encode" will try not to 226 If $enable is true (or missing), then "encode" will try not to
179 encode the same string twice, but will instead encode a reference to 227 encode the same string twice, but will instead encode a reference to
180 the string instead. Depending on your data format. this can save a 228 the string instead. Depending on your data format, this can save a
181 lot of space, but also results in a very large runtime overhead 229 lot of space, but also results in a very large runtime overhead
182 (expect encoding times to be 2-4 times as high as without). 230 (expect encoding times to be 2-4 times as high as without).
183 231
184 It is recommended to leave it off unless you know your 232 It is recommended to leave it off unless you know your
185 communications partner supports the stringref extension to CBOR 233 communications partner supports the stringref extension to CBOR
186 (http://cbor.schmorp.de/stringref). 234 (<http://cbor.schmorp.de/stringref>), as without decoder support,
235 the resulting data structure might not be usable.
187 236
188 If $enable is false (the default), then "encode" will encode 237 If $enable is false (the default), then "encode" will encode strings
189 exception when it encounters anything it cannot encode as CBOR. 238 the standard CBOR way.
190 239
191 This option does not affect "decode" in any way - string references 240 This option does not affect "decode" in any way - string references
192 will always be decoded properly if present. 241 will always be decoded properly if present.
242
243 $cbor = $cbor->text_keys ([$enable])
244 $enabled = $cbor->get_text_keys
245 If $enabled is true (or missing), then "encode" will encode all perl
246 hash keys as CBOR text strings/UTF-8 string, upgrading them as
247 needed.
248
249 If $enable is false (the default), then "encode" will encode hash
250 keys normally - upgraded perl strings (strings internally encoded as
251 UTF-8) as CBOR text strings, and downgraded perl strings as CBOR
252 byte strings.
253
254 This option does not affect "decode" in any way.
255
256 This option is useful for interoperability with CBOR decoders that
257 don't treat byte strings as a form of text. It is especially useful
258 as Perl gives very little control over hash keys.
259
260 Enabling this option can be slow, as all downgraded hash keys that
261 are encoded need to be scanned and converted to UTF-8.
262
263 $cbor = $cbor->text_strings ([$enable])
264 $enabled = $cbor->get_text_strings
265 This option works similar to "text_keys", above, but works on all
266 strings (including hash keys), so "text_keys" has no further effect
267 after enabling "text_strings".
268
269 If $enabled is true (or missing), then "encode" will encode all perl
270 strings as CBOR text strings/UTF-8 strings, upgrading them as
271 needed.
272
273 If $enable is false (the default), then "encode" will encode strings
274 normally (but see "text_keys") - upgraded perl strings (strings
275 internally encoded as UTF-8) as CBOR text strings, and downgraded
276 perl strings as CBOR byte strings.
277
278 This option does not affect "decode" in any way.
279
280 This option has similar advantages and disadvantages as "text_keys".
281 In addition, this option effectively removes the ability to encode
282 byte strings, which might break some "FREEZE" and "TO_CBOR" methods
283 that rely on this, such as bignum encoding, so this option is mainly
284 useful for very simple data.
285
286 $cbor = $cbor->validate_utf8 ([$enable])
287 $enabled = $cbor->get_validate_utf8
288 If $enable is true (or missing), then "decode" will validate that
289 elements (text strings) containing UTF-8 data in fact contain valid
290 UTF-8 data (instead of blindly accepting it). This validation
291 obviously takes extra time during decoding.
292
293 The concept of "valid UTF-8" used is perl's concept, which is a
294 superset of the official UTF-8.
295
296 If $enable is false (the default), then "decode" will blindly accept
297 UTF-8 data, marking them as valid UTF-8 in the resulting data
298 structure regardless of whether that's true or not.
299
300 Perl isn't too happy about corrupted UTF-8 in strings, but should
301 generally not crash or do similarly evil things. Extensions might be
302 not so forgiving, so it's recommended to turn on this setting if you
303 receive untrusted CBOR.
304
305 This option does not affect "encode" in any way - strings that are
306 supposedly valid UTF-8 will simply be dumped into the resulting CBOR
307 string without checking whether that is, in fact, true or not.
193 308
194 $cbor = $cbor->filter ([$cb->($tag, $value)]) 309 $cbor = $cbor->filter ([$cb->($tag, $value)])
195 $cb_or_undef = $cbor->get_filter 310 $cb_or_undef = $cbor->get_filter
196 Sets or replaces the tagged value decoding filter (when $cb is 311 Sets or replaces the tagged value decoding filter (when $cb is
197 specified) or clears the filter (if no argument or "undef" is 312 specified) or clears the filter (if no argument or "undef" is
… …
215 When the filter is cleared (the default state), the default filter 330 When the filter is cleared (the default state), the default filter
216 function, "CBOR::XS::default_filter", is used. This function simply 331 function, "CBOR::XS::default_filter", is used. This function simply
217 looks up the tag in the %CBOR::XS::FILTER hash. If an entry exists 332 looks up the tag in the %CBOR::XS::FILTER hash. If an entry exists
218 it must be a code reference that is called with tag and value, and 333 it must be a code reference that is called with tag and value, and
219 is responsible for decoding the value. If no entry exists, it 334 is responsible for decoding the value. If no entry exists, it
220 returns no values. 335 returns no values. "CBOR::XS" provides a number of default filter
336 functions already, the the %CBOR::XS::FILTER hash can be freely
337 extended with more.
338
339 "CBOR::XS" additionally provides an alternative filter function that
340 is supposed to be safe to use with untrusted data (which the default
341 filter might not), called "CBOR::XS::safe_filter", which works the
342 same as the "default_filter" but uses the %CBOR::XS::SAFE_FILTER
343 variable instead. It is prepopulated with the tag decoding functions
344 that are deemed safe (basically the same as %CBOR::XS::FILTER
345 without all the bignum tags), and can be extended by user code as
346 wlel, although, obviously, one should be very careful about adding
347 decoding functions here, since the expectation is that they are safe
348 to use on untrusted data, after all.
221 349
222 Example: decode all tags not handled internally into 350 Example: decode all tags not handled internally into
223 CBOR::XS::Tagged objects, with no other special handling (useful 351 "CBOR::XS::Tagged" objects, with no other special handling (useful
224 when working with potentially "unsafe" CBOR data). 352 when working with potentially "unsafe" CBOR data).
225 353
226 CBOR::XS->new->filter (sub { })->decode ($cbor_data); 354 CBOR::XS->new->filter (sub { })->decode ($cbor_data);
227 355
228 Example: provide a global filter for tag 1347375694, converting the 356 Example: provide a global filter for tag 1347375694, converting the
… …
231 $CBOR::XS::FILTER{1347375694} = sub { 359 $CBOR::XS::FILTER{1347375694} = sub {
232 my ($tag, $value); 360 my ($tag, $value);
233 361
234 "tag 1347375694 value $value" 362 "tag 1347375694 value $value"
235 }; 363 };
364
365 Example: provide your own filter function that looks up tags in your
366 own hash:
367
368 my %my_filter = (
369 998347484 => sub {
370 my ($tag, $value);
371
372 "tag 998347484 value $value"
373 };
374 );
375
376 my $coder = CBOR::XS->new->filter (sub {
377 &{ $my_filter{$_[0]} or return }
378 });
379
380 Example: use the safe filter function (see "SECURITY CONSIDERATIONS"
381 for more considerations on security).
382
383 CBOR::XS->new->filter (\&CBOR::XS::safe_filter)->decode ($cbor_data);
236 384
237 $cbor_data = $cbor->encode ($perl_scalar) 385 $cbor_data = $cbor->encode ($perl_scalar)
238 Converts the given Perl data structure (a scalar value) to its CBOR 386 Converts the given Perl data structure (a scalar value) to its CBOR
239 representation. 387 representation.
240 388
… …
253 the next one starts. 401 the next one starts.
254 402
255 CBOR::XS->new->decode_prefix ("......") 403 CBOR::XS->new->decode_prefix ("......")
256 => ("...", 3) 404 => ("...", 3)
257 405
406 INCREMENTAL PARSING
407 In some cases, there is the need for incremental parsing of JSON texts.
408 While this module always has to keep both CBOR text and resulting Perl
409 data structure in memory at one time, it does allow you to parse a CBOR
410 stream incrementally, using a similar to using "decode_prefix" to see if
411 a full CBOR object is available, but is much more efficient.
412
413 It basically works by parsing as much of a CBOR string as possible - if
414 the CBOR data is not complete yet, the pasrer will remember where it
415 was, to be able to restart when more data has been accumulated. Once
416 enough data is available to either decode a complete CBOR value or raise
417 an error, a real decode will be attempted.
418
419 A typical use case would be a network protocol that consists of sending
420 and receiving CBOR-encoded messages. The solution that works with CBOR
421 and about anything else is by prepending a length to every CBOR value,
422 so the receiver knows how many octets to read. More compact (and
423 slightly slower) would be to just send CBOR values back-to-back, as
424 "CBOR::XS" knows where a CBOR value ends, and doesn't need an explicit
425 length.
426
427 The following methods help with this:
428
429 @decoded = $cbor->incr_parse ($buffer)
430 This method attempts to decode exactly one CBOR value from the
431 beginning of the given $buffer. The value is removed from the
432 $buffer on success. When $buffer doesn't contain a complete value
433 yet, it returns nothing. Finally, when the $buffer doesn't start
434 with something that could ever be a valid CBOR value, it raises an
435 exception, just as "decode" would. In the latter case the decoder
436 state is undefined and must be reset before being able to parse
437 further.
438
439 This method modifies the $buffer in place. When no CBOR value can be
440 decoded, the decoder stores the current string offset. On the next
441 call, continues decoding at the place where it stopped before. For
442 this to make sense, the $buffer must begin with the same octets as
443 on previous unsuccessful calls.
444
445 You can call this method in scalar context, in which case it either
446 returns a decoded value or "undef". This makes it impossible to
447 distinguish between CBOR null values (which decode to "undef") and
448 an unsuccessful decode, which is often acceptable.
449
450 @decoded = $cbor->incr_parse_multiple ($buffer)
451 Same as "incr_parse", but attempts to decode as many CBOR values as
452 possible in one go, instead of at most one. Calls to "incr_parse"
453 and "incr_parse_multiple" can be interleaved.
454
455 $cbor->incr_reset
456 Resets the incremental decoder. This throws away any saved state, so
457 that subsequent calls to "incr_parse" or "incr_parse_multiple" start
458 to parse a new CBOR value from the beginning of the $buffer again.
459
460 This method can be called at any time, but it *must* be called if
461 you want to change your $buffer or there was a decoding error and
462 you want to reuse the $cbor object for future incremental parsings.
463
258MAPPING 464MAPPING
259 This section describes how CBOR::XS maps Perl values to CBOR values and 465 This section describes how CBOR::XS maps Perl values to CBOR values and
260 vice versa. These mappings are designed to "do the right thing" in most 466 vice versa. These mappings are designed to "do the right thing" in most
261 circumstances automatically, preserving round-tripping characteristics 467 circumstances automatically, preserving round-tripping characteristics
262 (what you put in comes out as something equivalent). 468 (what you put in comes out as something equivalent).
… …
269 integers 475 integers
270 CBOR integers become (numeric) perl scalars. On perls without 64 bit 476 CBOR integers become (numeric) perl scalars. On perls without 64 bit
271 support, 64 bit integers will be truncated or otherwise corrupted. 477 support, 64 bit integers will be truncated or otherwise corrupted.
272 478
273 byte strings 479 byte strings
274 Byte strings will become octet strings in Perl (the byte values 480 Byte strings will become octet strings in Perl (the Byte values
275 0..255 will simply become characters of the same value in Perl). 481 0..255 will simply become characters of the same value in Perl).
276 482
277 UTF-8 strings 483 UTF-8 strings
278 UTF-8 strings in CBOR will be decoded, i.e. the UTF-8 octets will be 484 UTF-8 strings in CBOR will be decoded, i.e. the UTF-8 octets will be
279 decoded into proper Unicode code points. At the moment, the validity 485 decoded into proper Unicode code points. At the moment, the validity
… …
297 503
298 tagged values 504 tagged values
299 Tagged items consists of a numeric tag and another CBOR value. 505 Tagged items consists of a numeric tag and another CBOR value.
300 506
301 See "TAG HANDLING AND EXTENSIONS" and the description of "->filter" 507 See "TAG HANDLING AND EXTENSIONS" and the description of "->filter"
302 for details. 508 for details on which tags are handled how.
303 509
304 anything else 510 anything else
305 Anything else (e.g. unsupported simple values) will raise a decoding 511 Anything else (e.g. unsupported simple values) will raise a decoding
306 error. 512 error.
307 513
308 PERL -> CBOR 514 PERL -> CBOR
309 The mapping from Perl to CBOR is slightly more difficult, as Perl is a 515 The mapping from Perl to CBOR is slightly more difficult, as Perl is a
310 truly typeless language, so we can only guess which CBOR type is meant 516 typeless language. That means this module can only guess which CBOR type
311 by a Perl value. 517 is meant by a perl value.
312 518
313 hash references 519 hash references
314 Perl hash references become CBOR maps. As there is no inherent 520 Perl hash references become CBOR maps. As there is no inherent
315 ordering in hash keys (or CBOR maps), they will usually be encoded 521 ordering in hash keys (or CBOR maps), they will usually be encoded
316 in a pseudo-random order. 522 in a pseudo-random order. This order can be different each time a
523 hash is encoded.
317 524
318 Currently, tied hashes will use the indefinite-length format, while 525 Currently, tied hashes will use the indefinite-length format, while
319 normal hashes will use the fixed-length format. 526 normal hashes will use the fixed-length format.
320 527
321 array references 528 array references
322 Perl array references become fixed-length CBOR arrays. 529 Perl array references become fixed-length CBOR arrays.
323 530
324 other references 531 other references
325 Other unblessed references are generally not allowed and will cause 532 Other unblessed references will be represented using the indirection
326 an exception to be thrown, except for references to the integers 0 533 tag extension (tag value 22098,
327 and 1, which get turned into false and true in CBOR. 534 <http://cbor.schmorp.de/indirection>). CBOR decoders are guaranteed
535 to be able to decode these values somehow, by either "doing the
536 right thing", decoding into a generic tagged object, simply ignoring
537 the tag, or something else.
328 538
329 CBOR::XS::Tagged objects 539 CBOR::XS::Tagged objects
330 Objects of this type must be arrays consisting of a single "[tag, 540 Objects of this type must be arrays consisting of a single "[tag,
331 value]" pair. The (numerical) tag will be encoded as a CBOR tag, the 541 value]" pair. The (numerical) tag will be encoded as a CBOR tag, the
332 value will be encoded as appropriate for the value. You cna use 542 value will be encoded as appropriate for the value. You must use
333 "CBOR::XS::tag" to create such objects. 543 "CBOR::XS::tag" to create such objects.
334 544
335 Types::Serialiser::true, Types::Serialiser::false, 545 Types::Serialiser::true, Types::Serialiser::false,
336 Types::Serialiser::error 546 Types::Serialiser::error
337 These special values become CBOR true, CBOR false and CBOR undefined 547 These special values become CBOR true, CBOR false and CBOR undefined
… …
353 # dump as number 563 # dump as number
354 encode_cbor [2] # yields [2] 564 encode_cbor [2] # yields [2]
355 encode_cbor [-3.0e17] # yields [-3e+17] 565 encode_cbor [-3.0e17] # yields [-3e+17]
356 my $value = 5; encode_cbor [$value] # yields [5] 566 my $value = 5; encode_cbor [$value] # yields [5]
357 567
358 # used as string, so dump as string 568 # used as string, so dump as string (either byte or text)
359 print $value; 569 print $value;
360 encode_cbor [$value] # yields ["5"] 570 encode_cbor [$value] # yields ["5"]
361 571
362 # undef becomes null 572 # undef becomes null
363 encode_cbor [undef] # yields [null] 573 encode_cbor [undef] # yields [null]
… …
366 576
367 my $x = 3.1; # some variable containing a number 577 my $x = 3.1; # some variable containing a number
368 "$x"; # stringified 578 "$x"; # stringified
369 $x .= ""; # another, more awkward way to stringify 579 $x .= ""; # another, more awkward way to stringify
370 print $x; # perl does it for you, too, quite often 580 print $x; # perl does it for you, too, quite often
581
582 You can force whether a string is encoded as byte or text string by
583 using "utf8::upgrade" and "utf8::downgrade" (if "text_strings" is
584 disabled):
585
586 utf8::upgrade $x; # encode $x as text string
587 utf8::downgrade $x; # encode $x as byte string
588
589 Perl doesn't define what operations up- and downgrade strings, so if
590 the difference between byte and text is important, you should up- or
591 downgrade your string as late as possible before encoding. You can
592 also force the use of CBOR text strings by using "text_keys" or
593 "text_strings".
371 594
372 You can force the type to be a CBOR number by numifying it: 595 You can force the type to be a CBOR number by numifying it:
373 596
374 my $x = "3"; # some variable containing a string 597 my $x = "3"; # some variable containing a string
375 $x += 0; # numify it, ensuring it will be dumped as a number 598 $x += 0; # numify it, ensuring it will be dumped as a number
… …
385 the IEEE double format will be used. Perls that use formats other 608 the IEEE double format will be used. Perls that use formats other
386 than IEEE double to represent numerical values are supported, but 609 than IEEE double to represent numerical values are supported, but
387 might suffer loss of precision. 610 might suffer loss of precision.
388 611
389 OBJECT SERIALISATION 612 OBJECT SERIALISATION
613 This module implements both a CBOR-specific and the generic
614 Types::Serialier object serialisation protocol. The following
615 subsections explain both methods.
616
617 ENCODING
390 This module knows two way to serialise a Perl object: The CBOR-specific 618 This module knows two way to serialise a Perl object: The CBOR-specific
391 way, and the generic way. 619 way, and the generic way.
392 620
393 Whenever the encoder encounters a Perl object that it cnanot serialise 621 Whenever the encoder encounters a Perl object that it cannot serialise
394 directly (most of them), it will first look up the "TO_CBOR" method on 622 directly (most of them), it will first look up the "TO_CBOR" method on
395 it. 623 it.
396 624
397 If it has a "TO_CBOR" method, it will call it with the object as only 625 If it has a "TO_CBOR" method, it will call it with the object as only
398 argument, and expects exactly one return value, which it will then 626 argument, and expects exactly one return value, which it will then
… …
403 "CBOR" as the second argument, to distinguish it from other serialisers. 631 "CBOR" as the second argument, to distinguish it from other serialisers.
404 632
405 The "FREEZE" method can return any number of values (i.e. zero or more). 633 The "FREEZE" method can return any number of values (i.e. zero or more).
406 These will be encoded as CBOR perl object, together with the classname. 634 These will be encoded as CBOR perl object, together with the classname.
407 635
636 These methods *MUST NOT* change the data structure that is being
637 serialised. Failure to comply to this can result in memory corruption -
638 and worse.
639
408 If an object supports neither "TO_CBOR" nor "FREEZE", encoding will fail 640 If an object supports neither "TO_CBOR" nor "FREEZE", encoding will fail
409 with an error. 641 with an error.
410 642
643 DECODING
411 Objects encoded via "TO_CBOR" cannot be automatically decoded, but 644 Objects encoded via "TO_CBOR" cannot (normally) be automatically
412 objects encoded via "FREEZE" can be decoded using the following 645 decoded, but objects encoded via "FREEZE" can be decoded using the
413 protocol: 646 following protocol:
414 647
415 When an encoded CBOR perl object is encountered by the decoder, it will 648 When an encoded CBOR perl object is encountered by the decoder, it will
416 look up the "THAW" method, by using the stored classname, and will fail 649 look up the "THAW" method, by using the stored classname, and will fail
417 if the method cannot be found. 650 if the method cannot be found.
418 651
… …
439 672
440 sub URI::TO_CBOR { 673 sub URI::TO_CBOR {
441 my ($self) = @_; 674 my ($self) = @_;
442 my $uri = "$self"; # stringify uri 675 my $uri = "$self"; # stringify uri
443 utf8::upgrade $uri; # make sure it will be encoded as UTF-8 string 676 utf8::upgrade $uri; # make sure it will be encoded as UTF-8 string
444 CBOR::XS::tagged 32, "$_[0]" 677 CBOR::XS::tag 32, "$_[0]"
445 } 678 }
446 679
447 This will encode URIs as a UTF-8 string with tag 32, which indicates an 680 This will encode URIs as a UTF-8 string with tag 32, which indicates an
448 URI. 681 URI.
449 682
… …
460 "$self" # encode url string 693 "$self" # encode url string
461 } 694 }
462 695
463 sub URI::THAW { 696 sub URI::THAW {
464 my ($class, $serialiser, $uri) = @_; 697 my ($class, $serialiser, $uri) = @_;
465
466 $class->new ($uri) 698 $class->new ($uri)
467 } 699 }
468 700
469 Unlike "TO_CBOR", multiple values can be returned by "FREEZE". For 701 Unlike "TO_CBOR", multiple values can be returned by "FREEZE". For
470 example, a "FREEZE" method that returns "type", "id" and "variant" 702 example, a "FREEZE" method that returns "type", "id" and "variant"
… …
566 Future versions of this module reserve the right to special case 798 Future versions of this module reserve the right to special case
567 additional tags (such as base64url). 799 additional tags (such as base64url).
568 800
569 ENFORCED TAGS 801 ENFORCED TAGS
570 These tags are always handled when decoding, and their handling cannot 802 These tags are always handled when decoding, and their handling cannot
571 be overriden by the user. 803 be overridden by the user.
572 804
573 <unassigned> (perl-object, <http://cbor.schmorp.de/perl-object>) 805 26 (perl-object, <http://cbor.schmorp.de/perl-object>)
574 These tags are automatically created (and decoded) for serialisable 806 These tags are automatically created (and decoded) for serialisable
575 objects using the "FREEZE/THAW" methods (the Types::Serialier object 807 objects using the "FREEZE/THAW" methods (the Types::Serialier object
576 serialisation protocol). See "OBJECT SERIALISATION" for details. 808 serialisation protocol). See "OBJECT SERIALISATION" for details.
577 809
578 <unassigned>, <unassigned> (sharable, sharedref, L 810 28, 29 (shareable, sharedref, <http://cbor.schmorp.de/value-sharing>)
579 <http://cbor.schmorp.de/value-sharing>)
580 These tags are automatically decoded when encountered, resulting in 811 These tags are automatically decoded when encountered (and they do
812 not result in a cyclic data structure, see "allow_cycles"),
581 shared values in the decoded object. They are only encoded, however, 813 resulting in shared values in the decoded object. They are only
582 when "allow_sharable" is enabled. 814 encoded, however, when "allow_sharing" is enabled.
583 815
584 <unassigned>, <unassigned> (stringref-namespace, stringref, L 816 Not all shared values can be successfully decoded: values that
817 reference themselves will *currently* decode as "undef" (this is not
818 the same as a reference pointing to itself, which will be
819 represented as a value that contains an indirect reference to itself
820 - these will be decoded properly).
821
822 Note that considerably more shared value data structures can be
823 decoded than will be encoded - currently, only values pointed to by
824 references will be shared, others will not. While non-reference
825 shared values can be generated in Perl with some effort, they were
826 considered too unimportant to be supported in the encoder. The
827 decoder, however, will decode these values as shared values.
828
829 256, 25 (stringref-namespace, stringref,
585 <http://cbor.schmorp.de/stringref>) 830 <http://cbor.schmorp.de/stringref>)
586 These tags are automatically decoded when encountered. They are only 831 These tags are automatically decoded when encountered. They are only
587 encoded, however, when "allow_stringref" is enabled. 832 encoded, however, when "pack_strings" is enabled.
588 833
589 22098 (indirection, <http://cbor.schmorp.de/indirection>) 834 22098 (indirection, <http://cbor.schmorp.de/indirection>)
590 This tag is automatically generated when a reference are encountered 835 This tag is automatically generated when a reference are encountered
591 (with the exception of hash and array refernces). It is converted to 836 (with the exception of hash and array references). It is converted
592 a reference when decoding. 837 to a reference when decoding.
593 838
594 55799 (self-describe CBOR, RFC 7049) 839 55799 (self-describe CBOR, RFC 7049)
595 This value is not generated on encoding (unless explicitly requested 840 This value is not generated on encoding (unless explicitly requested
596 by the user), and is simply ignored when decoding. 841 by the user), and is simply ignored when decoding.
597 842
598 NON-ENFORCED TAGS 843 NON-ENFORCED TAGS
599 These tags have default filters provided when decoding. Their handling 844 These tags have default filters provided when decoding. Their handling
600 can be overriden by changing the %CBOR::XS::FILTER entry for the tag, or 845 can be overridden by changing the %CBOR::XS::FILTER entry for the tag,
601 by providing a custom "filter" callback when decoding. 846 or by providing a custom "filter" callback when decoding.
602 847
603 When they result in decoding into a specific Perl class, the module 848 When they result in decoding into a specific Perl class, the module
604 usually provides a corresponding "TO_CBOR" method as well. 849 usually provides a corresponding "TO_CBOR" method as well.
605 850
606 When any of these need to load additional modules that are not part of 851 When any of these need to load additional modules that are not part of
607 the perl core distribution (e.g. URI), it is (currently) up to the user 852 the perl core distribution (e.g. URI), it is (currently) up to the user
608 to provide these modules. The decoding usually fails with an exception 853 to provide these modules. The decoding usually fails with an exception
609 if the required module cannot be loaded. 854 if the required module cannot be loaded.
610 855
856 0, 1 (date/time string, seconds since the epoch)
857 These tags are decoded into Time::Piece objects. The corresponding
858 "Time::Piece::TO_CBOR" method always encodes into tag 1 values
859 currently.
860
861 The Time::Piece API is generally surprisingly bad, and fractional
862 seconds are only accidentally kept intact, so watch out. On the plus
863 side, the module comes with perl since 5.10, which has to count for
864 something.
865
611 2, 3 (positive/negative bignum) 866 2, 3 (positive/negative bignum)
612 These tags are decoded into Math::BigInt objects. The corresponding 867 These tags are decoded into Math::BigInt objects. The corresponding
613 "Math::BigInt::TO_CBOR" method encodes "small" bigints into normal 868 "Math::BigInt::TO_CBOR" method encodes "small" bigints into normal
614 CBOR integers, and others into positive/negative CBOR bignums. 869 CBOR integers, and others into positive/negative CBOR bignums.
615 870
616 4, 5 (decimal fraction/bigfloat) 871 4, 5, 264, 265 (decimal fraction/bigfloat)
617 Both decimal fractions and bigfloats are decoded into Math::BigFloat 872 Both decimal fractions and bigfloats are decoded into Math::BigFloat
618 objects. The corresponding "Math::BigFloat::TO_CBOR" method *always* 873 objects. The corresponding "Math::BigFloat::TO_CBOR" method *always*
619 encodes into a decimal fraction. 874 encodes into a decimal fraction (either tag 4 or 264).
620 875
621 CBOR cannot represent bigfloats with *very* large exponents -
622 conversion of such big float objects is undefined.
623
624 Also, NaN and infinities are not encoded properly. 876 NaN and infinities are not encoded properly, as they cannot be
877 represented in CBOR.
878
879 See "BIGNUM SECURITY CONSIDERATIONS" for more info.
880
881 30 (rational numbers)
882 These tags are decoded into Math::BigRat objects. The corresponding
883 "Math::BigRat::TO_CBOR" method encodes rational numbers with
884 denominator 1 via their numerator only, i.e., they become normal
885 integers or "bignums".
886
887 See "BIGNUM SECURITY CONSIDERATIONS" for more info.
625 888
626 21, 22, 23 (expected later JSON conversion) 889 21, 22, 23 (expected later JSON conversion)
627 CBOR::XS is not a CBOR-to-JSON converter, and will simply ignore 890 CBOR::XS is not a CBOR-to-JSON converter, and will simply ignore
628 these tags. 891 these tags.
629 892
… …
643 interoperability is improved in the future, then the goal will be to 906 interoperability is improved in the future, then the goal will be to
644 ensure that decoded JSON data will round-trip encoding and decoding to 907 ensure that decoded JSON data will round-trip encoding and decoding to
645 CBOR intact. 908 CBOR intact.
646 909
647SECURITY CONSIDERATIONS 910SECURITY CONSIDERATIONS
648 When you are using CBOR in a protocol, talking to untrusted potentially 911 Tl;dr... if you want to decode or encode CBOR from untrusted sources,
649 hostile creatures requires relatively few measures. 912 you should start with a coder object created via "new_safe":
650 913
914 my $coder = CBOR::XS->new_safe;
915
916 my $data = $coder->decode ($cbor_text);
917 my $cbor = $coder->encode ($data);
918
919 Longer version: When you are using CBOR in a protocol, talking to
920 untrusted potentially hostile creatures requires some thought:
921
922 Security of the CBOR decoder itself
651 First of all, your CBOR decoder should be secure, that is, should not 923 First and foremost, your CBOR decoder should be secure, that is,
652 have any buffer overflows. Obviously, this module should ensure that and 924 should not have any buffer overflows or similar bugs that could
925 potentially be exploited. Obviously, this module should ensure that
653 I am trying hard on making that true, but you never know. 926 and I am trying hard on making that true, but you never know.
654 927
928 CBOR::XS can invoke almost arbitrary callbacks during decoding
929 CBOR::XS supports object serialisation - decoding CBOR can cause
930 calls to *any* "THAW" method in *any* package that exists in your
931 process (that is, CBOR::XS will not try to load modules, but any
932 existing "THAW" method or function can be called, so they all have
933 to be secure).
934
935 Less obviously, it will also invoke "TO_CBOR" and "FREEZE" methods -
936 even if all your "THAW" methods are secure, encoding data structures
937 from untrusted sources can invoke those and trigger bugs in those.
938
939 So, if you are not sure about the security of all the modules you
940 have loaded (you shouldn't), you should disable this part using
941 "forbid_objects".
942
943 CBOR can be extended with tags that call library code
944 CBOR can be extended with tags, and "CBOR::XS" has a registry of
945 conversion functions for many existing tags that can be extended via
946 third-party modules (see the "filter" method).
947
948 If you don't trust these, you should configure the "safe" filter
949 function, "CBOR::XS::safe_filter", which by default only includes
950 conversion functions that are considered "safe" by the author (but
951 again, they can be extended by third party modules).
952
953 Depending on your level of paranoia, you can use the "safe" filter:
954
955 $cbor->filter (\&CBOR::XS::safe_filter);
956
957 ... your own filter...
958
959 $cbor->filter (sub { ... do your stuffs here ... });
960
961 ... or even no filter at all, disabling all tag decoding:
962
963 $cbor->filter (sub { });
964
965 This is never a problem for encoding, as the tag mechanism only
966 exists in CBOR texts.
967
968 Resource-starving attacks: object memory usage
655 Second, you need to avoid resource-starving attacks. That means you 969 You need to avoid resource-starving attacks. That means you should
656 should limit the size of CBOR data you accept, or make sure then when 970 limit the size of CBOR data you accept, or make sure then when your
657 your resources run out, that's just fine (e.g. by using a separate 971 resources run out, that's just fine (e.g. by using a separate
658 process that can crash safely). The size of a CBOR string in octets is 972 process that can crash safely). The size of a CBOR string in octets
659 usually a good indication of the size of the resources required to 973 is usually a good indication of the size of the resources required
660 decode it into a Perl structure. While CBOR::XS can check the size of 974 to decode it into a Perl structure. While CBOR::XS can check the
661 the CBOR text, it might be too late when you already have it in memory, 975 size of the CBOR text (using "max_size"), it might be too late when
662 so you might want to check the size before you accept the string. 976 you already have it in memory, so you might want to check the size
977 before you accept the string.
663 978
979 As for encoding, it is possible to construct data structures that
980 are relatively small but result in large CBOR texts (for example by
981 having an array full of references to the same big data structure,
982 which will all be deep-cloned during encoding by default). This is
983 rarely an actual issue (and the worst case is still just running out
984 of memory), but you can reduce this risk by using "allow_sharing".
985
986 Resource-starving attacks: stack overflows
664 Third, CBOR::XS recurses using the C stack when decoding objects and 987 CBOR::XS recurses using the C stack when decoding objects and
665 arrays. The C stack is a limited resource: for instance, on my amd64 988 arrays. The C stack is a limited resource: for instance, on my amd64
666 machine with 8MB of stack size I can decode around 180k nested arrays 989 machine with 8MB of stack size I can decode around 180k nested
667 but only 14k nested CBOR objects (due to perl itself recursing deeply on 990 arrays but only 14k nested CBOR objects (due to perl itself
668 croak to free the temporary). If that is exceeded, the program crashes. 991 recursing deeply on croak to free the temporary). If that is
669 To be conservative, the default nesting limit is set to 512. If your 992 exceeded, the program crashes. To be conservative, the default
670 process has a smaller stack, you should adjust this setting accordingly 993 nesting limit is set to 512. If your process has a smaller stack,
671 with the "max_depth" method. 994 you should adjust this setting accordingly with the "max_depth"
995 method.
672 996
997 Resource-starving attacks: CPU en-/decoding complexity
998 CBOR::XS will use the Math::BigInt, Math::BigFloat and Math::BigRat
999 libraries to represent encode/decode bignums. These can be very slow
1000 (as in, centuries of CPU time) and can even crash your program (and
1001 are generally not very trustworthy). See the next section for
1002 details.
1003
1004 Data breaches: leaking information in error messages
1005 CBOR::XS might leak contents of your Perl data structures in its
1006 error messages, so when you serialise sensitive information you
1007 might want to make sure that exceptions thrown by CBOR::XS will not
1008 end up in front of untrusted eyes.
1009
1010 Something else...
673 Something else could bomb you, too, that I forgot to think of. In that 1011 Something else could bomb you, too, that I forgot to think of. In
674 case, you get to keep the pieces. I am always open for hints, though... 1012 that case, you get to keep the pieces. I am always open for hints,
1013 though...
675 1014
676 Also keep in mind that CBOR::XS might leak contents of your Perl data 1015BIGNUM SECURITY CONSIDERATIONS
677 structures in its error messages, so when you serialise sensitive 1016 CBOR::XS provides a "TO_CBOR" method for both Math::BigInt and
678 information you might want to make sure that exceptions thrown by 1017 Math::BigFloat that tries to encode the number in the simplest possible
679 CBOR::XS will not end up in front of untrusted eyes. 1018 way, that is, either a CBOR integer, a CBOR bigint/decimal fraction (tag
1019 4) or an arbitrary-exponent decimal fraction (tag 264). Rational numbers
1020 (Math::BigRat, tag 30) can also contain bignums as members.
1021
1022 CBOR::XS will also understand base-2 bigfloat or arbitrary-exponent
1023 bigfloats (tags 5 and 265), but it will never generate these on its own.
1024
1025 Using the built-in Math::BigInt::Calc support, encoding and decoding
1026 decimal fractions is generally fast. Decoding bigints can be slow for
1027 very big numbers (tens of thousands of digits, something that could
1028 potentially be caught by limiting the size of CBOR texts), and decoding
1029 bigfloats or arbitrary-exponent bigfloats can be *extremely* slow
1030 (minutes, decades) for large exponents (roughly 40 bit and longer).
1031
1032 Additionally, Math::BigInt can take advantage of other bignum libraries,
1033 such as Math::GMP, which cannot handle big floats with large exponents,
1034 and might simply abort or crash your program, due to their code quality.
1035
1036 This can be a concern if you want to parse untrusted CBOR. If it is, you
1037 might want to disable decoding of tag 2 (bigint) and 3 (negative bigint)
1038 types. You should also disable types 5 and 265, as these can be slow
1039 even without bigints.
1040
1041 Disabling bigints will also partially or fully disable types that rely
1042 on them, e.g. rational numbers that use bignums.
680 1043
681CBOR IMPLEMENTATION NOTES 1044CBOR IMPLEMENTATION NOTES
682 This section contains some random implementation notes. They do not 1045 This section contains some random implementation notes. They do not
683 describe guaranteed behaviour, but merely behaviour as-is implemented 1046 describe guaranteed behaviour, but merely behaviour as-is implemented
684 right now. 1047 right now.
… …
692 Only the double data type is supported for NV data types - when Perl 1055 Only the double data type is supported for NV data types - when Perl
693 uses long double to represent floating point values, they might not be 1056 uses long double to represent floating point values, they might not be
694 encoded properly. Half precision types are accepted, but not encoded. 1057 encoded properly. Half precision types are accepted, but not encoded.
695 1058
696 Strict mode and canonical mode are not implemented. 1059 Strict mode and canonical mode are not implemented.
1060
1061LIMITATIONS ON PERLS WITHOUT 64-BIT INTEGER SUPPORT
1062 On perls that were built without 64 bit integer support (these are rare
1063 nowadays, even on 32 bit architectures, as all major Perl distributions
1064 are built with 64 bit integer support), support for any kind of 64 bit
1065 integer in CBOR is very limited - most likely, these 64 bit values will
1066 be truncated, corrupted, or otherwise not decoded correctly. This also
1067 includes string, array and map sizes that are stored as 64 bit integers.
697 1068
698THREADS 1069THREADS
699 This module is *not* guaranteed to be thread safe and there are no plans 1070 This module is *not* guaranteed to be thread safe and there are no plans
700 to change this until Perl gets thread support (as opposed to the 1071 to change this until Perl gets thread support (as opposed to the
701 horribly slow so-called "threads" which are simply slow and bloated 1072 horribly slow so-called "threads" which are simply slow and bloated

Diff Legend

– Removed lines
+ Added lines
< Changed lines
> Changed lines