1 |
=head1 NAME |
2 |
|
3 |
CBOR::XS - Concise Binary Object Representation (CBOR, RFC7049) |
4 |
|
5 |
=encoding utf-8 |
6 |
|
7 |
=head1 SYNOPSIS |
8 |
|
9 |
use CBOR::XS; |
10 |
|
11 |
$binary_cbor_data = encode_cbor $perl_value; |
12 |
$perl_value = decode_cbor $binary_cbor_data; |
13 |
|
14 |
# OO-interface |
15 |
|
16 |
$coder = CBOR::XS->new; |
17 |
$binary_cbor_data = $coder->encode ($perl_value); |
18 |
$perl_value = $coder->decode ($binary_cbor_data); |
19 |
|
20 |
# prefix decoding |
21 |
|
22 |
my $many_cbor_strings = ...; |
23 |
while (length $many_cbor_strings) { |
24 |
my ($data, $length) = $cbor->decode_prefix ($many_cbor_strings); |
25 |
# data was decoded |
26 |
substr $many_cbor_strings, 0, $length, ""; # remove decoded cbor string |
27 |
} |
28 |
|
29 |
=head1 DESCRIPTION |
30 |
|
31 |
This module converts Perl data structures to the Concise Binary Object |
32 |
Representation (CBOR) and vice versa. CBOR is a fast binary serialisation |
33 |
format that aims to use an (almost) superset of the JSON data model, i.e. |
34 |
when you can represent something useful in JSON, you should be able to |
35 |
represent it in CBOR. |
36 |
|
37 |
In short, CBOR is a faster and quite compact binary alternative to JSON, |
38 |
with the added ability of supporting serialisation of Perl objects. (JSON |
39 |
often compresses better than CBOR though, so if you plan to compress the |
40 |
data later and speed is less important you might want to compare both |
41 |
formats first). |
42 |
|
43 |
To give you a general idea about speed, with texts in the megabyte range, |
44 |
C<CBOR::XS> usually encodes roughly twice as fast as L<Storable> or |
45 |
L<JSON::XS> and decodes about 15%-30% faster than those. The shorter the |
46 |
data, the worse L<Storable> performs in comparison. |
47 |
|
48 |
Regarding compactness, C<CBOR::XS>-encoded data structures are usually |
49 |
about 20% smaller than the same data encoded as (compact) JSON or |
50 |
L<Storable>. |
51 |
|
52 |
In addition to the core CBOR data format, this module implements a |
53 |
number of extensions, to support cyclic and shared data structures |
54 |
(see C<allow_sharing> and C<allow_cycles>), string deduplication (see |
55 |
C<pack_strings>) and scalar references (always enabled). |
56 |
|
57 |
The primary goal of this module is to be I<correct> and the secondary goal |
58 |
is to be I<fast>. To reach the latter goal it was written in C. |
59 |
|
60 |
See MAPPING, below, on how CBOR::XS maps perl values to CBOR values and |
61 |
vice versa. |
62 |
|
63 |
=cut |
64 |
|
65 |
package CBOR::XS; |
66 |
|
67 |
use common::sense; |
68 |
|
69 |
our $VERSION = 1.11; |
70 |
our @ISA = qw(Exporter); |
71 |
|
72 |
our @EXPORT = qw(encode_cbor decode_cbor); |
73 |
|
74 |
use Exporter; |
75 |
use XSLoader; |
76 |
|
77 |
use Types::Serialiser; |
78 |
|
79 |
our $MAGIC = "\xd9\xd9\xf7"; |
80 |
|
81 |
=head1 FUNCTIONAL INTERFACE |
82 |
|
83 |
The following convenience methods are provided by this module. They are |
84 |
exported by default: |
85 |
|
86 |
=over 4 |
87 |
|
88 |
=item $cbor_data = encode_cbor $perl_scalar |
89 |
|
90 |
Converts the given Perl data structure to CBOR representation. Croaks on |
91 |
error. |
92 |
|
93 |
=item $perl_scalar = decode_cbor $cbor_data |
94 |
|
95 |
The opposite of C<encode_cbor>: expects a valid CBOR string to parse, |
96 |
returning the resulting perl scalar. Croaks on error. |
97 |
|
98 |
=back |
99 |
|
100 |
|
101 |
=head1 OBJECT-ORIENTED INTERFACE |
102 |
|
103 |
The object oriented interface lets you configure your own encoding or |
104 |
decoding style, within the limits of supported formats. |
105 |
|
106 |
=over 4 |
107 |
|
108 |
=item $cbor = new CBOR::XS |
109 |
|
110 |
Creates a new CBOR::XS object that can be used to de/encode CBOR |
111 |
strings. All boolean flags described below are by default I<disabled>. |
112 |
|
113 |
The mutators for flags all return the CBOR object again and thus calls can |
114 |
be chained: |
115 |
|
116 |
my $cbor = CBOR::XS->new->encode ({a => [1,2]}); |
117 |
|
118 |
=item $cbor = $cbor->max_depth ([$maximum_nesting_depth]) |
119 |
|
120 |
=item $max_depth = $cbor->get_max_depth |
121 |
|
122 |
Sets the maximum nesting level (default C<512>) accepted while encoding |
123 |
or decoding. If a higher nesting level is detected in CBOR data or a Perl |
124 |
data structure, then the encoder and decoder will stop and croak at that |
125 |
point. |
126 |
|
127 |
Nesting level is defined by number of hash- or arrayrefs that the encoder |
128 |
needs to traverse to reach a given point or the number of C<{> or C<[> |
129 |
characters without their matching closing parenthesis crossed to reach a |
130 |
given character in a string. |
131 |
|
132 |
Setting the maximum depth to one disallows any nesting, so that ensures |
133 |
that the object is only a single hash/object or array. |
134 |
|
135 |
If no argument is given, the highest possible setting will be used, which |
136 |
is rarely useful. |
137 |
|
138 |
Note that nesting is implemented by recursion in C. The default value has |
139 |
been chosen to be as large as typical operating systems allow without |
140 |
crashing. |
141 |
|
142 |
See SECURITY CONSIDERATIONS, below, for more info on why this is useful. |
143 |
|
144 |
=item $cbor = $cbor->max_size ([$maximum_string_size]) |
145 |
|
146 |
=item $max_size = $cbor->get_max_size |
147 |
|
148 |
Set the maximum length a CBOR string may have (in bytes) where decoding |
149 |
is being attempted. The default is C<0>, meaning no limit. When C<decode> |
150 |
is called on a string that is longer then this many bytes, it will not |
151 |
attempt to decode the string but throw an exception. This setting has no |
152 |
effect on C<encode> (yet). |
153 |
|
154 |
If no argument is given, the limit check will be deactivated (same as when |
155 |
C<0> is specified). |
156 |
|
157 |
See SECURITY CONSIDERATIONS, below, for more info on why this is useful. |
158 |
|
159 |
=item $cbor = $cbor->allow_unknown ([$enable]) |
160 |
|
161 |
=item $enabled = $cbor->get_allow_unknown |
162 |
|
163 |
If C<$enable> is true (or missing), then C<encode> will I<not> throw an |
164 |
exception when it encounters values it cannot represent in CBOR (for |
165 |
example, filehandles) but instead will encode a CBOR C<error> value. |
166 |
|
167 |
If C<$enable> is false (the default), then C<encode> will throw an |
168 |
exception when it encounters anything it cannot encode as CBOR. |
169 |
|
170 |
This option does not affect C<decode> in any way, and it is recommended to |
171 |
leave it off unless you know your communications partner. |
172 |
|
173 |
=item $cbor = $cbor->allow_sharing ([$enable]) |
174 |
|
175 |
=item $enabled = $cbor->get_allow_sharing |
176 |
|
177 |
If C<$enable> is true (or missing), then C<encode> will not double-encode |
178 |
values that have been referenced before (e.g. when the same object, such |
179 |
as an array, is referenced multiple times), but instead will emit a |
180 |
reference to the earlier value. |
181 |
|
182 |
This means that such values will only be encoded once, and will not result |
183 |
in a deep cloning of the value on decode, in decoders supporting the value |
184 |
sharing extension. This also makes it possible to encode cyclic data |
185 |
structures (which need C<allow_cycles> to ne enabled to be decoded by this |
186 |
module). |
187 |
|
188 |
It is recommended to leave it off unless you know your |
189 |
communication partner supports the value sharing extensions to CBOR |
190 |
(L<http://cbor.schmorp.de/value-sharing>), as without decoder support, the |
191 |
resulting data structure might be unusable. |
192 |
|
193 |
Detecting shared values incurs a runtime overhead when values are encoded |
194 |
that have a reference counter large than one, and might unnecessarily |
195 |
increase the encoded size, as potentially shared values are encode as |
196 |
shareable whether or not they are actually shared. |
197 |
|
198 |
At the moment, only targets of references can be shared (e.g. scalars, |
199 |
arrays or hashes pointed to by a reference). Weirder constructs, such as |
200 |
an array with multiple "copies" of the I<same> string, which are hard but |
201 |
not impossible to create in Perl, are not supported (this is the same as |
202 |
with L<Storable>). |
203 |
|
204 |
If C<$enable> is false (the default), then C<encode> will encode shared |
205 |
data structures repeatedly, unsharing them in the process. Cyclic data |
206 |
structures cannot be encoded in this mode. |
207 |
|
208 |
This option does not affect C<decode> in any way - shared values and |
209 |
references will always be decoded properly if present. |
210 |
|
211 |
=item $cbor = $cbor->allow_cycles ([$enable]) |
212 |
|
213 |
=item $enabled = $cbor->get_allow_cycles |
214 |
|
215 |
If C<$enable> is true (or missing), then C<decode> will happily decode |
216 |
self-referential (cyclic) data structures. By default these will not be |
217 |
decoded, as they need manual cleanup to avoid memory leaks, so code that |
218 |
isn't prepared for this will not leak memory. |
219 |
|
220 |
If C<$enable> is false (the default), then C<decode> will throw an error |
221 |
when it encounters a self-referential/cyclic data structure. |
222 |
|
223 |
This option does not affect C<encode> in any way - shared values and |
224 |
references will always be decoded properly if present. |
225 |
|
226 |
=item $cbor = $cbor->pack_strings ([$enable]) |
227 |
|
228 |
=item $enabled = $cbor->get_pack_strings |
229 |
|
230 |
If C<$enable> is true (or missing), then C<encode> will try not to encode |
231 |
the same string twice, but will instead encode a reference to the string |
232 |
instead. Depending on your data format, this can save a lot of space, but |
233 |
also results in a very large runtime overhead (expect encoding times to be |
234 |
2-4 times as high as without). |
235 |
|
236 |
It is recommended to leave it off unless you know your |
237 |
communications partner supports the stringref extension to CBOR |
238 |
(L<http://cbor.schmorp.de/stringref>), as without decoder support, the |
239 |
resulting data structure might not be usable. |
240 |
|
241 |
If C<$enable> is false (the default), then C<encode> will encode strings |
242 |
the standard CBOR way. |
243 |
|
244 |
This option does not affect C<decode> in any way - string references will |
245 |
always be decoded properly if present. |
246 |
|
247 |
=item $cbor = $cbor->validate_utf8 ([$enable]) |
248 |
|
249 |
=item $enabled = $cbor->get_validate_utf8 |
250 |
|
251 |
If C<$enable> is true (or missing), then C<decode> will validate that |
252 |
elements (text strings) containing UTF-8 data in fact contain valid UTF-8 |
253 |
data (instead of blindly accepting it). This validation obviously takes |
254 |
extra time during decoding. |
255 |
|
256 |
The concept of "valid UTF-8" used is perl's concept, which is a superset |
257 |
of the official UTF-8. |
258 |
|
259 |
If C<$enable> is false (the default), then C<decode> will blindly accept |
260 |
UTF-8 data, marking them as valid UTF-8 in the resulting data structure |
261 |
regardless of whether thats true or not. |
262 |
|
263 |
Perl isn't too happy about corrupted UTF-8 in strings, but should |
264 |
generally not crash or do similarly evil things. Extensions might be not |
265 |
so forgiving, so it's recommended to turn on this setting if you receive |
266 |
untrusted CBOR. |
267 |
|
268 |
This option does not affect C<encode> in any way - strings that are |
269 |
supposedly valid UTF-8 will simply be dumped into the resulting CBOR |
270 |
string without checking whether that is, in fact, true or not. |
271 |
|
272 |
=item $cbor = $cbor->filter ([$cb->($tag, $value)]) |
273 |
|
274 |
=item $cb_or_undef = $cbor->get_filter |
275 |
|
276 |
Sets or replaces the tagged value decoding filter (when C<$cb> is |
277 |
specified) or clears the filter (if no argument or C<undef> is provided). |
278 |
|
279 |
The filter callback is called only during decoding, when a non-enforced |
280 |
tagged value has been decoded (see L<TAG HANDLING AND EXTENSIONS> for a |
281 |
list of enforced tags). For specific tags, it's often better to provide a |
282 |
default converter using the C<%CBOR::XS::FILTER> hash (see below). |
283 |
|
284 |
The first argument is the numerical tag, the second is the (decoded) value |
285 |
that has been tagged. |
286 |
|
287 |
The filter function should return either exactly one value, which will |
288 |
replace the tagged value in the decoded data structure, or no values, |
289 |
which will result in default handling, which currently means the decoder |
290 |
creates a C<CBOR::XS::Tagged> object to hold the tag and the value. |
291 |
|
292 |
When the filter is cleared (the default state), the default filter |
293 |
function, C<CBOR::XS::default_filter>, is used. This function simply looks |
294 |
up the tag in the C<%CBOR::XS::FILTER> hash. If an entry exists it must be |
295 |
a code reference that is called with tag and value, and is responsible for |
296 |
decoding the value. If no entry exists, it returns no values. |
297 |
|
298 |
Example: decode all tags not handled internally into C<CBOR::XS::Tagged> |
299 |
objects, with no other special handling (useful when working with |
300 |
potentially "unsafe" CBOR data). |
301 |
|
302 |
CBOR::XS->new->filter (sub { })->decode ($cbor_data); |
303 |
|
304 |
Example: provide a global filter for tag 1347375694, converting the value |
305 |
into some string form. |
306 |
|
307 |
$CBOR::XS::FILTER{1347375694} = sub { |
308 |
my ($tag, $value); |
309 |
|
310 |
"tag 1347375694 value $value" |
311 |
}; |
312 |
|
313 |
=item $cbor_data = $cbor->encode ($perl_scalar) |
314 |
|
315 |
Converts the given Perl data structure (a scalar value) to its CBOR |
316 |
representation. |
317 |
|
318 |
=item $perl_scalar = $cbor->decode ($cbor_data) |
319 |
|
320 |
The opposite of C<encode>: expects CBOR data and tries to parse it, |
321 |
returning the resulting simple scalar or reference. Croaks on error. |
322 |
|
323 |
=item ($perl_scalar, $octets) = $cbor->decode_prefix ($cbor_data) |
324 |
|
325 |
This works like the C<decode> method, but instead of raising an exception |
326 |
when there is trailing garbage after the CBOR string, it will silently |
327 |
stop parsing there and return the number of characters consumed so far. |
328 |
|
329 |
This is useful if your CBOR texts are not delimited by an outer protocol |
330 |
and you need to know where the first CBOR string ends amd the next one |
331 |
starts. |
332 |
|
333 |
CBOR::XS->new->decode_prefix ("......") |
334 |
=> ("...", 3) |
335 |
|
336 |
=back |
337 |
|
338 |
|
339 |
=head1 MAPPING |
340 |
|
341 |
This section describes how CBOR::XS maps Perl values to CBOR values and |
342 |
vice versa. These mappings are designed to "do the right thing" in most |
343 |
circumstances automatically, preserving round-tripping characteristics |
344 |
(what you put in comes out as something equivalent). |
345 |
|
346 |
For the more enlightened: note that in the following descriptions, |
347 |
lowercase I<perl> refers to the Perl interpreter, while uppercase I<Perl> |
348 |
refers to the abstract Perl language itself. |
349 |
|
350 |
|
351 |
=head2 CBOR -> PERL |
352 |
|
353 |
=over 4 |
354 |
|
355 |
=item integers |
356 |
|
357 |
CBOR integers become (numeric) perl scalars. On perls without 64 bit |
358 |
support, 64 bit integers will be truncated or otherwise corrupted. |
359 |
|
360 |
=item byte strings |
361 |
|
362 |
Byte strings will become octet strings in Perl (the Byte values 0..255 |
363 |
will simply become characters of the same value in Perl). |
364 |
|
365 |
=item UTF-8 strings |
366 |
|
367 |
UTF-8 strings in CBOR will be decoded, i.e. the UTF-8 octets will be |
368 |
decoded into proper Unicode code points. At the moment, the validity of |
369 |
the UTF-8 octets will not be validated - corrupt input will result in |
370 |
corrupted Perl strings. |
371 |
|
372 |
=item arrays, maps |
373 |
|
374 |
CBOR arrays and CBOR maps will be converted into references to a Perl |
375 |
array or hash, respectively. The keys of the map will be stringified |
376 |
during this process. |
377 |
|
378 |
=item null |
379 |
|
380 |
CBOR null becomes C<undef> in Perl. |
381 |
|
382 |
=item true, false, undefined |
383 |
|
384 |
These CBOR values become C<Types:Serialiser::true>, |
385 |
C<Types:Serialiser::false> and C<Types::Serialiser::error>, |
386 |
respectively. They are overloaded to act almost exactly like the numbers |
387 |
C<1> and C<0> (for true and false) or to throw an exception on access (for |
388 |
error). See the L<Types::Serialiser> manpage for details. |
389 |
|
390 |
=item tagged values |
391 |
|
392 |
Tagged items consists of a numeric tag and another CBOR value. |
393 |
|
394 |
See L<TAG HANDLING AND EXTENSIONS> and the description of C<< ->filter >> |
395 |
for details on which tags are handled how. |
396 |
|
397 |
=item anything else |
398 |
|
399 |
Anything else (e.g. unsupported simple values) will raise a decoding |
400 |
error. |
401 |
|
402 |
=back |
403 |
|
404 |
|
405 |
=head2 PERL -> CBOR |
406 |
|
407 |
The mapping from Perl to CBOR is slightly more difficult, as Perl is a |
408 |
typeless language. That means this module can only guess which CBOR type |
409 |
is meant by a perl value. |
410 |
|
411 |
=over 4 |
412 |
|
413 |
=item hash references |
414 |
|
415 |
Perl hash references become CBOR maps. As there is no inherent ordering in |
416 |
hash keys (or CBOR maps), they will usually be encoded in a pseudo-random |
417 |
order. This order can be different each time a hahs is encoded. |
418 |
|
419 |
Currently, tied hashes will use the indefinite-length format, while normal |
420 |
hashes will use the fixed-length format. |
421 |
|
422 |
=item array references |
423 |
|
424 |
Perl array references become fixed-length CBOR arrays. |
425 |
|
426 |
=item other references |
427 |
|
428 |
Other unblessed references will be represented using |
429 |
the indirection tag extension (tag value C<22098>, |
430 |
L<http://cbor.schmorp.de/indirection>). CBOR decoders are guaranteed |
431 |
to be able to decode these values somehow, by either "doing the right |
432 |
thing", decoding into a generic tagged object, simply ignoring the tag, or |
433 |
something else. |
434 |
|
435 |
=item CBOR::XS::Tagged objects |
436 |
|
437 |
Objects of this type must be arrays consisting of a single C<[tag, value]> |
438 |
pair. The (numerical) tag will be encoded as a CBOR tag, the value will |
439 |
be encoded as appropriate for the value. You must use C<CBOR::XS::tag> to |
440 |
create such objects. |
441 |
|
442 |
=item Types::Serialiser::true, Types::Serialiser::false, Types::Serialiser::error |
443 |
|
444 |
These special values become CBOR true, CBOR false and CBOR undefined |
445 |
values, respectively. You can also use C<\1>, C<\0> and C<\undef> directly |
446 |
if you want. |
447 |
|
448 |
=item other blessed objects |
449 |
|
450 |
Other blessed objects are serialised via C<TO_CBOR> or C<FREEZE>. See |
451 |
L<TAG HANDLING AND EXTENSIONS> for specific classes handled by this |
452 |
module, and L<OBJECT SERIALISATION> for generic object serialisation. |
453 |
|
454 |
=item simple scalars |
455 |
|
456 |
Simple Perl scalars (any scalar that is not a reference) are the most |
457 |
difficult objects to encode: CBOR::XS will encode undefined scalars as |
458 |
CBOR null values, scalars that have last been used in a string context |
459 |
before encoding as CBOR strings, and anything else as number value: |
460 |
|
461 |
# dump as number |
462 |
encode_cbor [2] # yields [2] |
463 |
encode_cbor [-3.0e17] # yields [-3e+17] |
464 |
my $value = 5; encode_cbor [$value] # yields [5] |
465 |
|
466 |
# used as string, so dump as string (either byte or text) |
467 |
print $value; |
468 |
encode_cbor [$value] # yields ["5"] |
469 |
|
470 |
# undef becomes null |
471 |
encode_cbor [undef] # yields [null] |
472 |
|
473 |
You can force the type to be a CBOR string by stringifying it: |
474 |
|
475 |
my $x = 3.1; # some variable containing a number |
476 |
"$x"; # stringified |
477 |
$x .= ""; # another, more awkward way to stringify |
478 |
print $x; # perl does it for you, too, quite often |
479 |
|
480 |
You can force whether a string ie encoded as byte or text string by using |
481 |
C<utf8::upgrade> and C<utf8::downgrade>): |
482 |
|
483 |
utf8::upgrade $x; # encode $x as text string |
484 |
utf8::downgrade $x; # encode $x as byte string |
485 |
|
486 |
Perl doesn't define what operations up- and downgrade strings, so if the |
487 |
difference between byte and text is important, you should up- or downgrade |
488 |
your string as late as possible before encoding. |
489 |
|
490 |
You can force the type to be a CBOR number by numifying it: |
491 |
|
492 |
my $x = "3"; # some variable containing a string |
493 |
$x += 0; # numify it, ensuring it will be dumped as a number |
494 |
$x *= 1; # same thing, the choice is yours. |
495 |
|
496 |
You can not currently force the type in other, less obscure, ways. Tell me |
497 |
if you need this capability (but don't forget to explain why it's needed |
498 |
:). |
499 |
|
500 |
Perl values that seem to be integers generally use the shortest possible |
501 |
representation. Floating-point values will use either the IEEE single |
502 |
format if possible without loss of precision, otherwise the IEEE double |
503 |
format will be used. Perls that use formats other than IEEE double to |
504 |
represent numerical values are supported, but might suffer loss of |
505 |
precision. |
506 |
|
507 |
=back |
508 |
|
509 |
=head2 OBJECT SERIALISATION |
510 |
|
511 |
This module implements both a CBOR-specific and the generic |
512 |
L<Types::Serialier> object serialisation protocol. The following |
513 |
subsections explain both methods. |
514 |
|
515 |
=head3 ENCODING |
516 |
|
517 |
This module knows two way to serialise a Perl object: The CBOR-specific |
518 |
way, and the generic way. |
519 |
|
520 |
Whenever the encoder encounters a Perl object that it cannot serialise |
521 |
directly (most of them), it will first look up the C<TO_CBOR> method on |
522 |
it. |
523 |
|
524 |
If it has a C<TO_CBOR> method, it will call it with the object as only |
525 |
argument, and expects exactly one return value, which it will then |
526 |
substitute and encode it in the place of the object. |
527 |
|
528 |
Otherwise, it will look up the C<FREEZE> method. If it exists, it will |
529 |
call it with the object as first argument, and the constant string C<CBOR> |
530 |
as the second argument, to distinguish it from other serialisers. |
531 |
|
532 |
The C<FREEZE> method can return any number of values (i.e. zero or |
533 |
more). These will be encoded as CBOR perl object, together with the |
534 |
classname. |
535 |
|
536 |
These methods I<MUST NOT> change the data structure that is being |
537 |
serialised. Failure to comply to this can result in memory corruption - |
538 |
and worse. |
539 |
|
540 |
If an object supports neither C<TO_CBOR> nor C<FREEZE>, encoding will fail |
541 |
with an error. |
542 |
|
543 |
=head3 DECODING |
544 |
|
545 |
Objects encoded via C<TO_CBOR> cannot (normally) be automatically decoded, |
546 |
but objects encoded via C<FREEZE> can be decoded using the following |
547 |
protocol: |
548 |
|
549 |
When an encoded CBOR perl object is encountered by the decoder, it will |
550 |
look up the C<THAW> method, by using the stored classname, and will fail |
551 |
if the method cannot be found. |
552 |
|
553 |
After the lookup it will call the C<THAW> method with the stored classname |
554 |
as first argument, the constant string C<CBOR> as second argument, and all |
555 |
values returned by C<FREEZE> as remaining arguments. |
556 |
|
557 |
=head3 EXAMPLES |
558 |
|
559 |
Here is an example C<TO_CBOR> method: |
560 |
|
561 |
sub My::Object::TO_CBOR { |
562 |
my ($obj) = @_; |
563 |
|
564 |
["this is a serialised My::Object object", $obj->{id}] |
565 |
} |
566 |
|
567 |
When a C<My::Object> is encoded to CBOR, it will instead encode a simple |
568 |
array with two members: a string, and the "object id". Decoding this CBOR |
569 |
string will yield a normal perl array reference in place of the object. |
570 |
|
571 |
A more useful and practical example would be a serialisation method for |
572 |
the URI module. CBOR has a custom tag value for URIs, namely 32: |
573 |
|
574 |
sub URI::TO_CBOR { |
575 |
my ($self) = @_; |
576 |
my $uri = "$self"; # stringify uri |
577 |
utf8::upgrade $uri; # make sure it will be encoded as UTF-8 string |
578 |
CBOR::XS::tag 32, "$_[0]" |
579 |
} |
580 |
|
581 |
This will encode URIs as a UTF-8 string with tag 32, which indicates an |
582 |
URI. |
583 |
|
584 |
Decoding such an URI will not (currently) give you an URI object, but |
585 |
instead a CBOR::XS::Tagged object with tag number 32 and the string - |
586 |
exactly what was returned by C<TO_CBOR>. |
587 |
|
588 |
To serialise an object so it can automatically be deserialised, you need |
589 |
to use C<FREEZE> and C<THAW>. To take the URI module as example, this |
590 |
would be a possible implementation: |
591 |
|
592 |
sub URI::FREEZE { |
593 |
my ($self, $serialiser) = @_; |
594 |
"$self" # encode url string |
595 |
} |
596 |
|
597 |
sub URI::THAW { |
598 |
my ($class, $serialiser, $uri) = @_; |
599 |
|
600 |
$class->new ($uri) |
601 |
} |
602 |
|
603 |
Unlike C<TO_CBOR>, multiple values can be returned by C<FREEZE>. For |
604 |
example, a C<FREEZE> method that returns "type", "id" and "variant" values |
605 |
would cause an invocation of C<THAW> with 5 arguments: |
606 |
|
607 |
sub My::Object::FREEZE { |
608 |
my ($self, $serialiser) = @_; |
609 |
|
610 |
($self->{type}, $self->{id}, $self->{variant}) |
611 |
} |
612 |
|
613 |
sub My::Object::THAW { |
614 |
my ($class, $serialiser, $type, $id, $variant) = @_; |
615 |
|
616 |
$class-<new (type => $type, id => $id, variant => $variant) |
617 |
} |
618 |
|
619 |
|
620 |
=head1 MAGIC HEADER |
621 |
|
622 |
There is no way to distinguish CBOR from other formats |
623 |
programmatically. To make it easier to distinguish CBOR from other |
624 |
formats, the CBOR specification has a special "magic string" that can be |
625 |
prepended to any CBOR string without changing its meaning. |
626 |
|
627 |
This string is available as C<$CBOR::XS::MAGIC>. This module does not |
628 |
prepend this string to the CBOR data it generates, but it will ignore it |
629 |
if present, so users can prepend this string as a "file type" indicator as |
630 |
required. |
631 |
|
632 |
|
633 |
=head1 THE CBOR::XS::Tagged CLASS |
634 |
|
635 |
CBOR has the concept of tagged values - any CBOR value can be tagged with |
636 |
a numeric 64 bit number, which are centrally administered. |
637 |
|
638 |
C<CBOR::XS> handles a few tags internally when en- or decoding. You can |
639 |
also create tags yourself by encoding C<CBOR::XS::Tagged> objects, and the |
640 |
decoder will create C<CBOR::XS::Tagged> objects itself when it hits an |
641 |
unknown tag. |
642 |
|
643 |
These objects are simply blessed array references - the first member of |
644 |
the array being the numerical tag, the second being the value. |
645 |
|
646 |
You can interact with C<CBOR::XS::Tagged> objects in the following ways: |
647 |
|
648 |
=over 4 |
649 |
|
650 |
=item $tagged = CBOR::XS::tag $tag, $value |
651 |
|
652 |
This function(!) creates a new C<CBOR::XS::Tagged> object using the given |
653 |
C<$tag> (0..2**64-1) to tag the given C<$value> (which can be any Perl |
654 |
value that can be encoded in CBOR, including serialisable Perl objects and |
655 |
C<CBOR::XS::Tagged> objects). |
656 |
|
657 |
=item $tagged->[0] |
658 |
|
659 |
=item $tagged->[0] = $new_tag |
660 |
|
661 |
=item $tag = $tagged->tag |
662 |
|
663 |
=item $new_tag = $tagged->tag ($new_tag) |
664 |
|
665 |
Access/mutate the tag. |
666 |
|
667 |
=item $tagged->[1] |
668 |
|
669 |
=item $tagged->[1] = $new_value |
670 |
|
671 |
=item $value = $tagged->value |
672 |
|
673 |
=item $new_value = $tagged->value ($new_value) |
674 |
|
675 |
Access/mutate the tagged value. |
676 |
|
677 |
=back |
678 |
|
679 |
=cut |
680 |
|
681 |
sub tag($$) { |
682 |
bless [@_], CBOR::XS::Tagged::; |
683 |
} |
684 |
|
685 |
sub CBOR::XS::Tagged::tag { |
686 |
$_[0][0] = $_[1] if $#_; |
687 |
$_[0][0] |
688 |
} |
689 |
|
690 |
sub CBOR::XS::Tagged::value { |
691 |
$_[0][1] = $_[1] if $#_; |
692 |
$_[0][1] |
693 |
} |
694 |
|
695 |
=head2 EXAMPLES |
696 |
|
697 |
Here are some examples of C<CBOR::XS::Tagged> uses to tag objects. |
698 |
|
699 |
You can look up CBOR tag value and emanings in the IANA registry at |
700 |
L<http://www.iana.org/assignments/cbor-tags/cbor-tags.xhtml>. |
701 |
|
702 |
Prepend a magic header (C<$CBOR::XS::MAGIC>): |
703 |
|
704 |
my $cbor = encode_cbor CBOR::XS::tag 55799, $value; |
705 |
# same as: |
706 |
my $cbor = $CBOR::XS::MAGIC . encode_cbor $value; |
707 |
|
708 |
Serialise some URIs and a regex in an array: |
709 |
|
710 |
my $cbor = encode_cbor [ |
711 |
(CBOR::XS::tag 32, "http://www.nethype.de/"), |
712 |
(CBOR::XS::tag 32, "http://software.schmorp.de/"), |
713 |
(CBOR::XS::tag 35, "^[Pp][Ee][Rr][lL]\$"), |
714 |
]; |
715 |
|
716 |
Wrap CBOR data in CBOR: |
717 |
|
718 |
my $cbor_cbor = encode_cbor |
719 |
CBOR::XS::tag 24, |
720 |
encode_cbor [1, 2, 3]; |
721 |
|
722 |
=head1 TAG HANDLING AND EXTENSIONS |
723 |
|
724 |
This section describes how this module handles specific tagged values |
725 |
and extensions. If a tag is not mentioned here and no additional filters |
726 |
are provided for it, then the default handling applies (creating a |
727 |
CBOR::XS::Tagged object on decoding, and only encoding the tag when |
728 |
explicitly requested). |
729 |
|
730 |
Tags not handled specifically are currently converted into a |
731 |
L<CBOR::XS::Tagged> object, which is simply a blessed array reference |
732 |
consisting of the numeric tag value followed by the (decoded) CBOR value. |
733 |
|
734 |
Future versions of this module reserve the right to special case |
735 |
additional tags (such as base64url). |
736 |
|
737 |
=head2 ENFORCED TAGS |
738 |
|
739 |
These tags are always handled when decoding, and their handling cannot be |
740 |
overriden by the user. |
741 |
|
742 |
=over 4 |
743 |
|
744 |
=item 26 (perl-object, L<http://cbor.schmorp.de/perl-object>) |
745 |
|
746 |
These tags are automatically created (and decoded) for serialisable |
747 |
objects using the C<FREEZE/THAW> methods (the L<Types::Serialier> object |
748 |
serialisation protocol). See L<OBJECT SERIALISATION> for details. |
749 |
|
750 |
=item 28, 29 (shareable, sharedref, L <http://cbor.schmorp.de/value-sharing>) |
751 |
|
752 |
These tags are automatically decoded when encountered (and they do not |
753 |
result in a cyclic data structure, see C<allow_cycles>), resulting in |
754 |
shared values in the decoded object. They are only encoded, however, when |
755 |
C<allow_sharing> is enabled. |
756 |
|
757 |
Not all shared values can be successfully decoded: values that reference |
758 |
themselves will I<currently> decode as C<undef> (this is not the same |
759 |
as a reference pointing to itself, which will be represented as a value |
760 |
that contains an indirect reference to itself - these will be decoded |
761 |
properly). |
762 |
|
763 |
Note that considerably more shared value data structures can be decoded |
764 |
than will be encoded - currently, only values pointed to by references |
765 |
will be shared, others will not. While non-reference shared values can be |
766 |
generated in Perl with some effort, they were considered too unimportant |
767 |
to be supported in the encoder. The decoder, however, will decode these |
768 |
values as shared values. |
769 |
|
770 |
=item 256, 25 (stringref-namespace, stringref, L <http://cbor.schmorp.de/stringref>) |
771 |
|
772 |
These tags are automatically decoded when encountered. They are only |
773 |
encoded, however, when C<pack_strings> is enabled. |
774 |
|
775 |
=item 22098 (indirection, L<http://cbor.schmorp.de/indirection>) |
776 |
|
777 |
This tag is automatically generated when a reference are encountered (with |
778 |
the exception of hash and array refernces). It is converted to a reference |
779 |
when decoding. |
780 |
|
781 |
=item 55799 (self-describe CBOR, RFC 7049) |
782 |
|
783 |
This value is not generated on encoding (unless explicitly requested by |
784 |
the user), and is simply ignored when decoding. |
785 |
|
786 |
=back |
787 |
|
788 |
=head2 NON-ENFORCED TAGS |
789 |
|
790 |
These tags have default filters provided when decoding. Their handling can |
791 |
be overriden by changing the C<%CBOR::XS::FILTER> entry for the tag, or by |
792 |
providing a custom C<filter> callback when decoding. |
793 |
|
794 |
When they result in decoding into a specific Perl class, the module |
795 |
usually provides a corresponding C<TO_CBOR> method as well. |
796 |
|
797 |
When any of these need to load additional modules that are not part of the |
798 |
perl core distribution (e.g. L<URI>), it is (currently) up to the user to |
799 |
provide these modules. The decoding usually fails with an exception if the |
800 |
required module cannot be loaded. |
801 |
|
802 |
=over 4 |
803 |
|
804 |
=item 2, 3 (positive/negative bignum) |
805 |
|
806 |
These tags are decoded into L<Math::BigInt> objects. The corresponding |
807 |
C<Math::BigInt::TO_CBOR> method encodes "small" bigints into normal CBOR |
808 |
integers, and others into positive/negative CBOR bignums. |
809 |
|
810 |
=item 4, 5 (decimal fraction/bigfloat) |
811 |
|
812 |
Both decimal fractions and bigfloats are decoded into L<Math::BigFloat> |
813 |
objects. The corresponding C<Math::BigFloat::TO_CBOR> method I<always> |
814 |
encodes into a decimal fraction. |
815 |
|
816 |
CBOR cannot represent bigfloats with I<very> large exponents - conversion |
817 |
of such big float objects is undefined. |
818 |
|
819 |
Also, NaN and infinities are not encoded properly. |
820 |
|
821 |
=item 21, 22, 23 (expected later JSON conversion) |
822 |
|
823 |
CBOR::XS is not a CBOR-to-JSON converter, and will simply ignore these |
824 |
tags. |
825 |
|
826 |
=item 32 (URI) |
827 |
|
828 |
These objects decode into L<URI> objects. The corresponding |
829 |
C<URI::TO_CBOR> method again results in a CBOR URI value. |
830 |
|
831 |
=back |
832 |
|
833 |
=cut |
834 |
|
835 |
our %FILTER = ( |
836 |
# 0 # rfc4287 datetime, utf-8 |
837 |
# 1 # unix timestamp, any |
838 |
|
839 |
2 => sub { # pos bigint |
840 |
require Math::BigInt; |
841 |
Math::BigInt->new ("0x" . unpack "H*", pop) |
842 |
}, |
843 |
|
844 |
3 => sub { # neg bigint |
845 |
require Math::BigInt; |
846 |
-Math::BigInt->new ("0x" . unpack "H*", pop) |
847 |
}, |
848 |
|
849 |
4 => sub { # decimal fraction, array |
850 |
require Math::BigFloat; |
851 |
Math::BigFloat->new ($_[1][1] . "E" . $_[1][0]) |
852 |
}, |
853 |
|
854 |
5 => sub { # bigfloat, array |
855 |
require Math::BigFloat; |
856 |
scalar Math::BigFloat->new ($_[1][1])->blsft ($_[1][0], 2) |
857 |
}, |
858 |
|
859 |
21 => sub { pop }, # expected conversion to base64url encoding |
860 |
22 => sub { pop }, # expected conversion to base64 encoding |
861 |
23 => sub { pop }, # expected conversion to base16 encoding |
862 |
|
863 |
# 24 # embedded cbor, byte string |
864 |
|
865 |
32 => sub { |
866 |
require URI; |
867 |
URI->new (pop) |
868 |
}, |
869 |
|
870 |
# 33 # base64url rfc4648, utf-8 |
871 |
# 34 # base64 rfc46484, utf-8 |
872 |
# 35 # regex pcre/ecma262, utf-8 |
873 |
# 36 # mime message rfc2045, utf-8 |
874 |
); |
875 |
|
876 |
|
877 |
=head1 CBOR and JSON |
878 |
|
879 |
CBOR is supposed to implement a superset of the JSON data model, and is, |
880 |
with some coercion, able to represent all JSON texts (something that other |
881 |
"binary JSON" formats such as BSON generally do not support). |
882 |
|
883 |
CBOR implements some extra hints and support for JSON interoperability, |
884 |
and the spec offers further guidance for conversion between CBOR and |
885 |
JSON. None of this is currently implemented in CBOR, and the guidelines |
886 |
in the spec do not result in correct round-tripping of data. If JSON |
887 |
interoperability is improved in the future, then the goal will be to |
888 |
ensure that decoded JSON data will round-trip encoding and decoding to |
889 |
CBOR intact. |
890 |
|
891 |
|
892 |
=head1 SECURITY CONSIDERATIONS |
893 |
|
894 |
When you are using CBOR in a protocol, talking to untrusted potentially |
895 |
hostile creatures requires relatively few measures. |
896 |
|
897 |
First of all, your CBOR decoder should be secure, that is, should not have |
898 |
any buffer overflows. Obviously, this module should ensure that and I am |
899 |
trying hard on making that true, but you never know. |
900 |
|
901 |
Second, you need to avoid resource-starving attacks. That means you should |
902 |
limit the size of CBOR data you accept, or make sure then when your |
903 |
resources run out, that's just fine (e.g. by using a separate process that |
904 |
can crash safely). The size of a CBOR string in octets is usually a good |
905 |
indication of the size of the resources required to decode it into a Perl |
906 |
structure. While CBOR::XS can check the size of the CBOR text, it might be |
907 |
too late when you already have it in memory, so you might want to check |
908 |
the size before you accept the string. |
909 |
|
910 |
Third, CBOR::XS recurses using the C stack when decoding objects and |
911 |
arrays. The C stack is a limited resource: for instance, on my amd64 |
912 |
machine with 8MB of stack size I can decode around 180k nested arrays but |
913 |
only 14k nested CBOR objects (due to perl itself recursing deeply on croak |
914 |
to free the temporary). If that is exceeded, the program crashes. To be |
915 |
conservative, the default nesting limit is set to 512. If your process |
916 |
has a smaller stack, you should adjust this setting accordingly with the |
917 |
C<max_depth> method. |
918 |
|
919 |
Something else could bomb you, too, that I forgot to think of. In that |
920 |
case, you get to keep the pieces. I am always open for hints, though... |
921 |
|
922 |
Also keep in mind that CBOR::XS might leak contents of your Perl data |
923 |
structures in its error messages, so when you serialise sensitive |
924 |
information you might want to make sure that exceptions thrown by CBOR::XS |
925 |
will not end up in front of untrusted eyes. |
926 |
|
927 |
=head1 CBOR IMPLEMENTATION NOTES |
928 |
|
929 |
This section contains some random implementation notes. They do not |
930 |
describe guaranteed behaviour, but merely behaviour as-is implemented |
931 |
right now. |
932 |
|
933 |
64 bit integers are only properly decoded when Perl was built with 64 bit |
934 |
support. |
935 |
|
936 |
Strings and arrays are encoded with a definite length. Hashes as well, |
937 |
unless they are tied (or otherwise magical). |
938 |
|
939 |
Only the double data type is supported for NV data types - when Perl uses |
940 |
long double to represent floating point values, they might not be encoded |
941 |
properly. Half precision types are accepted, but not encoded. |
942 |
|
943 |
Strict mode and canonical mode are not implemented. |
944 |
|
945 |
|
946 |
=head1 LIMITATIONS ON PERLS WITHOUT 64-BIT INTEGER SUPPORT |
947 |
|
948 |
On perls that were built without 64 bit integer support (these are rare |
949 |
nowadays, even on 32 bit architectures), support for any kind of 64 bit |
950 |
integer in CBOR is very limited - most likely, these 64 bit values will |
951 |
be truncated, corrupted, or otherwise not decoded correctly. This also |
952 |
includes string, array and map sizes that are stored as 64 bit integers. |
953 |
|
954 |
|
955 |
=head1 THREADS |
956 |
|
957 |
This module is I<not> guaranteed to be thread safe and there are no |
958 |
plans to change this until Perl gets thread support (as opposed to the |
959 |
horribly slow so-called "threads" which are simply slow and bloated |
960 |
process simulations - use fork, it's I<much> faster, cheaper, better). |
961 |
|
962 |
(It might actually work, but you have been warned). |
963 |
|
964 |
|
965 |
=head1 BUGS |
966 |
|
967 |
While the goal of this module is to be correct, that unfortunately does |
968 |
not mean it's bug-free, only that I think its design is bug-free. If you |
969 |
keep reporting bugs they will be fixed swiftly, though. |
970 |
|
971 |
Please refrain from using rt.cpan.org or any other bug reporting |
972 |
service. I put the contact address into my modules for a reason. |
973 |
|
974 |
=cut |
975 |
|
976 |
our %FILTER = ( |
977 |
# 0 # rfc4287 datetime, utf-8 |
978 |
# 1 # unix timestamp, any |
979 |
|
980 |
2 => sub { # pos bigint |
981 |
require Math::BigInt; |
982 |
Math::BigInt->new ("0x" . unpack "H*", pop) |
983 |
}, |
984 |
|
985 |
3 => sub { # neg bigint |
986 |
require Math::BigInt; |
987 |
-Math::BigInt->new ("0x" . unpack "H*", pop) |
988 |
}, |
989 |
|
990 |
4 => sub { # decimal fraction, array |
991 |
require Math::BigFloat; |
992 |
Math::BigFloat->new ($_[1][1] . "E" . $_[1][0]) |
993 |
}, |
994 |
|
995 |
5 => sub { # bigfloat, array |
996 |
require Math::BigFloat; |
997 |
scalar Math::BigFloat->new ($_[1][1])->blsft ($_[1][0], 2) |
998 |
}, |
999 |
|
1000 |
21 => sub { pop }, # expected conversion to base64url encoding |
1001 |
22 => sub { pop }, # expected conversion to base64 encoding |
1002 |
23 => sub { pop }, # expected conversion to base16 encoding |
1003 |
|
1004 |
# 24 # embedded cbor, byte string |
1005 |
|
1006 |
32 => sub { |
1007 |
require URI; |
1008 |
URI->new (pop) |
1009 |
}, |
1010 |
|
1011 |
# 33 # base64url rfc4648, utf-8 |
1012 |
# 34 # base64 rfc46484, utf-8 |
1013 |
# 35 # regex pcre/ecma262, utf-8 |
1014 |
# 36 # mime message rfc2045, utf-8 |
1015 |
); |
1016 |
|
1017 |
sub CBOR::XS::default_filter { |
1018 |
&{ $FILTER{$_[0]} or return } |
1019 |
} |
1020 |
|
1021 |
sub URI::TO_CBOR { |
1022 |
my $uri = $_[0]->as_string; |
1023 |
utf8::upgrade $uri; |
1024 |
CBOR::XS::tag 32, $uri |
1025 |
} |
1026 |
|
1027 |
sub Math::BigInt::TO_CBOR { |
1028 |
if ($_[0] >= -2147483648 && $_[0] <= 2147483647) { |
1029 |
$_[0]->numify |
1030 |
} else { |
1031 |
my $hex = substr $_[0]->as_hex, 2; |
1032 |
$hex = "0$hex" if 1 & length $hex; # sigh |
1033 |
CBOR::XS::tag $_[0] >= 0 ? 2 : 3, pack "H*", $hex |
1034 |
} |
1035 |
} |
1036 |
|
1037 |
sub Math::BigFloat::TO_CBOR { |
1038 |
my ($m, $e) = $_[0]->parts; |
1039 |
CBOR::XS::tag 4, [$e->numify, $m] |
1040 |
} |
1041 |
|
1042 |
XSLoader::load "CBOR::XS", $VERSION; |
1043 |
|
1044 |
=head1 SEE ALSO |
1045 |
|
1046 |
The L<JSON> and L<JSON::XS> modules that do similar, but human-readable, |
1047 |
serialisation. |
1048 |
|
1049 |
The L<Types::Serialiser> module provides the data model for true, false |
1050 |
and error values. |
1051 |
|
1052 |
=head1 AUTHOR |
1053 |
|
1054 |
Marc Lehmann <schmorp@schmorp.de> |
1055 |
http://home.schmorp.de/ |
1056 |
|
1057 |
=cut |
1058 |
|
1059 |
1 |
1060 |
|