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 = 1.1; |
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 = $cbor->filter ([$cb->($tag, $value)]) |
248 |
|
249 |
=item $cb_or_undef = $cbor->get_filter |
250 |
|
251 |
Sets or replaces the tagged value decoding filter (when C<$cb> is |
252 |
specified) or clears the filter (if no argument or C<undef> is provided). |
253 |
|
254 |
The filter callback is called only during decoding, when a non-enforced |
255 |
tagged value has been decoded (see L<TAG HANDLING AND EXTENSIONS> for a |
256 |
list of enforced tags). For specific tags, it's often better to provide a |
257 |
default converter using the C<%CBOR::XS::FILTER> hash (see below). |
258 |
|
259 |
The first argument is the numerical tag, the second is the (decoded) value |
260 |
that has been tagged. |
261 |
|
262 |
The filter function should return either exactly one value, which will |
263 |
replace the tagged value in the decoded data structure, or no values, |
264 |
which will result in default handling, which currently means the decoder |
265 |
creates a C<CBOR::XS::Tagged> object to hold the tag and the value. |
266 |
|
267 |
When the filter is cleared (the default state), the default filter |
268 |
function, C<CBOR::XS::default_filter>, is used. This function simply looks |
269 |
up the tag in the C<%CBOR::XS::FILTER> hash. If an entry exists it must be |
270 |
a code reference that is called with tag and value, and is responsible for |
271 |
decoding the value. If no entry exists, it returns no values. |
272 |
|
273 |
Example: decode all tags not handled internally into C<CBOR::XS::Tagged> |
274 |
objects, with no other special handling (useful when working with |
275 |
potentially "unsafe" CBOR data). |
276 |
|
277 |
CBOR::XS->new->filter (sub { })->decode ($cbor_data); |
278 |
|
279 |
Example: provide a global filter for tag 1347375694, converting the value |
280 |
into some string form. |
281 |
|
282 |
$CBOR::XS::FILTER{1347375694} = sub { |
283 |
my ($tag, $value); |
284 |
|
285 |
"tag 1347375694 value $value" |
286 |
}; |
287 |
|
288 |
=item $cbor_data = $cbor->encode ($perl_scalar) |
289 |
|
290 |
Converts the given Perl data structure (a scalar value) to its CBOR |
291 |
representation. |
292 |
|
293 |
=item $perl_scalar = $cbor->decode ($cbor_data) |
294 |
|
295 |
The opposite of C<encode>: expects CBOR data and tries to parse it, |
296 |
returning the resulting simple scalar or reference. Croaks on error. |
297 |
|
298 |
=item ($perl_scalar, $octets) = $cbor->decode_prefix ($cbor_data) |
299 |
|
300 |
This works like the C<decode> method, but instead of raising an exception |
301 |
when there is trailing garbage after the CBOR string, it will silently |
302 |
stop parsing there and return the number of characters consumed so far. |
303 |
|
304 |
This is useful if your CBOR texts are not delimited by an outer protocol |
305 |
and you need to know where the first CBOR string ends amd the next one |
306 |
starts. |
307 |
|
308 |
CBOR::XS->new->decode_prefix ("......") |
309 |
=> ("...", 3) |
310 |
|
311 |
=back |
312 |
|
313 |
|
314 |
=head1 MAPPING |
315 |
|
316 |
This section describes how CBOR::XS maps Perl values to CBOR values and |
317 |
vice versa. These mappings are designed to "do the right thing" in most |
318 |
circumstances automatically, preserving round-tripping characteristics |
319 |
(what you put in comes out as something equivalent). |
320 |
|
321 |
For the more enlightened: note that in the following descriptions, |
322 |
lowercase I<perl> refers to the Perl interpreter, while uppercase I<Perl> |
323 |
refers to the abstract Perl language itself. |
324 |
|
325 |
|
326 |
=head2 CBOR -> PERL |
327 |
|
328 |
=over 4 |
329 |
|
330 |
=item integers |
331 |
|
332 |
CBOR integers become (numeric) perl scalars. On perls without 64 bit |
333 |
support, 64 bit integers will be truncated or otherwise corrupted. |
334 |
|
335 |
=item byte strings |
336 |
|
337 |
Byte strings will become octet strings in Perl (the Byte values 0..255 |
338 |
will simply become characters of the same value in Perl). |
339 |
|
340 |
=item UTF-8 strings |
341 |
|
342 |
UTF-8 strings in CBOR will be decoded, i.e. the UTF-8 octets will be |
343 |
decoded into proper Unicode code points. At the moment, the validity of |
344 |
the UTF-8 octets will not be validated - corrupt input will result in |
345 |
corrupted Perl strings. |
346 |
|
347 |
=item arrays, maps |
348 |
|
349 |
CBOR arrays and CBOR maps will be converted into references to a Perl |
350 |
array or hash, respectively. The keys of the map will be stringified |
351 |
during this process. |
352 |
|
353 |
=item null |
354 |
|
355 |
CBOR null becomes C<undef> in Perl. |
356 |
|
357 |
=item true, false, undefined |
358 |
|
359 |
These CBOR values become C<Types:Serialiser::true>, |
360 |
C<Types:Serialiser::false> and C<Types::Serialiser::error>, |
361 |
respectively. They are overloaded to act almost exactly like the numbers |
362 |
C<1> and C<0> (for true and false) or to throw an exception on access (for |
363 |
error). See the L<Types::Serialiser> manpage for details. |
364 |
|
365 |
=item tagged values |
366 |
|
367 |
Tagged items consists of a numeric tag and another CBOR value. |
368 |
|
369 |
See L<TAG HANDLING AND EXTENSIONS> and the description of C<< ->filter >> |
370 |
for details on which tags are handled how. |
371 |
|
372 |
=item anything else |
373 |
|
374 |
Anything else (e.g. unsupported simple values) will raise a decoding |
375 |
error. |
376 |
|
377 |
=back |
378 |
|
379 |
|
380 |
=head2 PERL -> CBOR |
381 |
|
382 |
The mapping from Perl to CBOR is slightly more difficult, as Perl is a |
383 |
typeless language. That means this module can only guess which CBOR type |
384 |
is meant by a perl value. |
385 |
|
386 |
=over 4 |
387 |
|
388 |
=item hash references |
389 |
|
390 |
Perl hash references become CBOR maps. As there is no inherent ordering in |
391 |
hash keys (or CBOR maps), they will usually be encoded in a pseudo-random |
392 |
order. This order can be different each time a hahs is encoded. |
393 |
|
394 |
Currently, tied hashes will use the indefinite-length format, while normal |
395 |
hashes will use the fixed-length format. |
396 |
|
397 |
=item array references |
398 |
|
399 |
Perl array references become fixed-length CBOR arrays. |
400 |
|
401 |
=item other references |
402 |
|
403 |
Other unblessed references will be represented using |
404 |
the indirection tag extension (tag value C<22098>, |
405 |
L<http://cbor.schmorp.de/indirection>). CBOR decoders are guaranteed |
406 |
to be able to decode these values somehow, by either "doing the right |
407 |
thing", decoding into a generic tagged object, simply ignoring the tag, or |
408 |
something else. |
409 |
|
410 |
=item CBOR::XS::Tagged objects |
411 |
|
412 |
Objects of this type must be arrays consisting of a single C<[tag, value]> |
413 |
pair. The (numerical) tag will be encoded as a CBOR tag, the value will |
414 |
be encoded as appropriate for the value. You must use C<CBOR::XS::tag> to |
415 |
create such objects. |
416 |
|
417 |
=item Types::Serialiser::true, Types::Serialiser::false, Types::Serialiser::error |
418 |
|
419 |
These special values become CBOR true, CBOR false and CBOR undefined |
420 |
values, respectively. You can also use C<\1>, C<\0> and C<\undef> directly |
421 |
if you want. |
422 |
|
423 |
=item other blessed objects |
424 |
|
425 |
Other blessed objects are serialised via C<TO_CBOR> or C<FREEZE>. See |
426 |
L<TAG HANDLING AND EXTENSIONS> for specific classes handled by this |
427 |
module, and L<OBJECT SERIALISATION> for generic object serialisation. |
428 |
|
429 |
=item simple scalars |
430 |
|
431 |
Simple Perl scalars (any scalar that is not a reference) are the most |
432 |
difficult objects to encode: CBOR::XS will encode undefined scalars as |
433 |
CBOR null values, scalars that have last been used in a string context |
434 |
before encoding as CBOR strings, and anything else as number value: |
435 |
|
436 |
# dump as number |
437 |
encode_cbor [2] # yields [2] |
438 |
encode_cbor [-3.0e17] # yields [-3e+17] |
439 |
my $value = 5; encode_cbor [$value] # yields [5] |
440 |
|
441 |
# used as string, so dump as string (either byte or text) |
442 |
print $value; |
443 |
encode_cbor [$value] # yields ["5"] |
444 |
|
445 |
# undef becomes null |
446 |
encode_cbor [undef] # yields [null] |
447 |
|
448 |
You can force the type to be a CBOR string by stringifying it: |
449 |
|
450 |
my $x = 3.1; # some variable containing a number |
451 |
"$x"; # stringified |
452 |
$x .= ""; # another, more awkward way to stringify |
453 |
print $x; # perl does it for you, too, quite often |
454 |
|
455 |
You can force whether a string ie encoded as byte or text string by using |
456 |
C<utf8::upgrade> and C<utf8::downgrade>): |
457 |
|
458 |
utf8::upgrade $x; # encode $x as text string |
459 |
utf8::downgrade $x; # encode $x as byte string |
460 |
|
461 |
Perl doesn't define what operations up- and downgrade strings, so if the |
462 |
difference between byte and text is important, you should up- or downgrade |
463 |
your string as late as possible before encoding. |
464 |
|
465 |
You can force the type to be a CBOR number by numifying it: |
466 |
|
467 |
my $x = "3"; # some variable containing a string |
468 |
$x += 0; # numify it, ensuring it will be dumped as a number |
469 |
$x *= 1; # same thing, the choice is yours. |
470 |
|
471 |
You can not currently force the type in other, less obscure, ways. Tell me |
472 |
if you need this capability (but don't forget to explain why it's needed |
473 |
:). |
474 |
|
475 |
Perl values that seem to be integers generally use the shortest possible |
476 |
representation. Floating-point values will use either the IEEE single |
477 |
format if possible without loss of precision, otherwise the IEEE double |
478 |
format will be used. Perls that use formats other than IEEE double to |
479 |
represent numerical values are supported, but might suffer loss of |
480 |
precision. |
481 |
|
482 |
=back |
483 |
|
484 |
=head2 OBJECT SERIALISATION |
485 |
|
486 |
This module implements both a CBOR-specific and the generic |
487 |
L<Types::Serialier> object serialisation protocol. The following |
488 |
subsections explain both methods. |
489 |
|
490 |
=head3 ENCODING |
491 |
|
492 |
This module knows two way to serialise a Perl object: The CBOR-specific |
493 |
way, and the generic way. |
494 |
|
495 |
Whenever the encoder encounters a Perl object that it cannot serialise |
496 |
directly (most of them), it will first look up the C<TO_CBOR> method on |
497 |
it. |
498 |
|
499 |
If it has a C<TO_CBOR> method, it will call it with the object as only |
500 |
argument, and expects exactly one return value, which it will then |
501 |
substitute and encode it in the place of the object. |
502 |
|
503 |
Otherwise, it will look up the C<FREEZE> method. If it exists, it will |
504 |
call it with the object as first argument, and the constant string C<CBOR> |
505 |
as the second argument, to distinguish it from other serialisers. |
506 |
|
507 |
The C<FREEZE> method can return any number of values (i.e. zero or |
508 |
more). These will be encoded as CBOR perl object, together with the |
509 |
classname. |
510 |
|
511 |
These methods I<MUST NOT> change the data structure that is being |
512 |
serialised. Failure to comply to this can result in memory corruption - |
513 |
and worse. |
514 |
|
515 |
If an object supports neither C<TO_CBOR> nor C<FREEZE>, encoding will fail |
516 |
with an error. |
517 |
|
518 |
=head3 DECODING |
519 |
|
520 |
Objects encoded via C<TO_CBOR> cannot (normally) be automatically decoded, |
521 |
but objects encoded via C<FREEZE> can be decoded using the following |
522 |
protocol: |
523 |
|
524 |
When an encoded CBOR perl object is encountered by the decoder, it will |
525 |
look up the C<THAW> method, by using the stored classname, and will fail |
526 |
if the method cannot be found. |
527 |
|
528 |
After the lookup it will call the C<THAW> method with the stored classname |
529 |
as first argument, the constant string C<CBOR> as second argument, and all |
530 |
values returned by C<FREEZE> as remaining arguments. |
531 |
|
532 |
=head3 EXAMPLES |
533 |
|
534 |
Here is an example C<TO_CBOR> method: |
535 |
|
536 |
sub My::Object::TO_CBOR { |
537 |
my ($obj) = @_; |
538 |
|
539 |
["this is a serialised My::Object object", $obj->{id}] |
540 |
} |
541 |
|
542 |
When a C<My::Object> is encoded to CBOR, it will instead encode a simple |
543 |
array with two members: a string, and the "object id". Decoding this CBOR |
544 |
string will yield a normal perl array reference in place of the object. |
545 |
|
546 |
A more useful and practical example would be a serialisation method for |
547 |
the URI module. CBOR has a custom tag value for URIs, namely 32: |
548 |
|
549 |
sub URI::TO_CBOR { |
550 |
my ($self) = @_; |
551 |
my $uri = "$self"; # stringify uri |
552 |
utf8::upgrade $uri; # make sure it will be encoded as UTF-8 string |
553 |
CBOR::XS::tag 32, "$_[0]" |
554 |
} |
555 |
|
556 |
This will encode URIs as a UTF-8 string with tag 32, which indicates an |
557 |
URI. |
558 |
|
559 |
Decoding such an URI will not (currently) give you an URI object, but |
560 |
instead a CBOR::XS::Tagged object with tag number 32 and the string - |
561 |
exactly what was returned by C<TO_CBOR>. |
562 |
|
563 |
To serialise an object so it can automatically be deserialised, you need |
564 |
to use C<FREEZE> and C<THAW>. To take the URI module as example, this |
565 |
would be a possible implementation: |
566 |
|
567 |
sub URI::FREEZE { |
568 |
my ($self, $serialiser) = @_; |
569 |
"$self" # encode url string |
570 |
} |
571 |
|
572 |
sub URI::THAW { |
573 |
my ($class, $serialiser, $uri) = @_; |
574 |
|
575 |
$class->new ($uri) |
576 |
} |
577 |
|
578 |
Unlike C<TO_CBOR>, multiple values can be returned by C<FREEZE>. For |
579 |
example, a C<FREEZE> method that returns "type", "id" and "variant" values |
580 |
would cause an invocation of C<THAW> with 5 arguments: |
581 |
|
582 |
sub My::Object::FREEZE { |
583 |
my ($self, $serialiser) = @_; |
584 |
|
585 |
($self->{type}, $self->{id}, $self->{variant}) |
586 |
} |
587 |
|
588 |
sub My::Object::THAW { |
589 |
my ($class, $serialiser, $type, $id, $variant) = @_; |
590 |
|
591 |
$class-<new (type => $type, id => $id, variant => $variant) |
592 |
} |
593 |
|
594 |
|
595 |
=head1 MAGIC HEADER |
596 |
|
597 |
There is no way to distinguish CBOR from other formats |
598 |
programmatically. To make it easier to distinguish CBOR from other |
599 |
formats, the CBOR specification has a special "magic string" that can be |
600 |
prepended to any CBOR string without changing its meaning. |
601 |
|
602 |
This string is available as C<$CBOR::XS::MAGIC>. This module does not |
603 |
prepend this string to the CBOR data it generates, but it will ignore it |
604 |
if present, so users can prepend this string as a "file type" indicator as |
605 |
required. |
606 |
|
607 |
|
608 |
=head1 THE CBOR::XS::Tagged CLASS |
609 |
|
610 |
CBOR has the concept of tagged values - any CBOR value can be tagged with |
611 |
a numeric 64 bit number, which are centrally administered. |
612 |
|
613 |
C<CBOR::XS> handles a few tags internally when en- or decoding. You can |
614 |
also create tags yourself by encoding C<CBOR::XS::Tagged> objects, and the |
615 |
decoder will create C<CBOR::XS::Tagged> objects itself when it hits an |
616 |
unknown tag. |
617 |
|
618 |
These objects are simply blessed array references - the first member of |
619 |
the array being the numerical tag, the second being the value. |
620 |
|
621 |
You can interact with C<CBOR::XS::Tagged> objects in the following ways: |
622 |
|
623 |
=over 4 |
624 |
|
625 |
=item $tagged = CBOR::XS::tag $tag, $value |
626 |
|
627 |
This function(!) creates a new C<CBOR::XS::Tagged> object using the given |
628 |
C<$tag> (0..2**64-1) to tag the given C<$value> (which can be any Perl |
629 |
value that can be encoded in CBOR, including serialisable Perl objects and |
630 |
C<CBOR::XS::Tagged> objects). |
631 |
|
632 |
=item $tagged->[0] |
633 |
|
634 |
=item $tagged->[0] = $new_tag |
635 |
|
636 |
=item $tag = $tagged->tag |
637 |
|
638 |
=item $new_tag = $tagged->tag ($new_tag) |
639 |
|
640 |
Access/mutate the tag. |
641 |
|
642 |
=item $tagged->[1] |
643 |
|
644 |
=item $tagged->[1] = $new_value |
645 |
|
646 |
=item $value = $tagged->value |
647 |
|
648 |
=item $new_value = $tagged->value ($new_value) |
649 |
|
650 |
Access/mutate the tagged value. |
651 |
|
652 |
=back |
653 |
|
654 |
=cut |
655 |
|
656 |
sub tag($$) { |
657 |
bless [@_], CBOR::XS::Tagged::; |
658 |
} |
659 |
|
660 |
sub CBOR::XS::Tagged::tag { |
661 |
$_[0][0] = $_[1] if $#_; |
662 |
$_[0][0] |
663 |
} |
664 |
|
665 |
sub CBOR::XS::Tagged::value { |
666 |
$_[0][1] = $_[1] if $#_; |
667 |
$_[0][1] |
668 |
} |
669 |
|
670 |
=head2 EXAMPLES |
671 |
|
672 |
Here are some examples of C<CBOR::XS::Tagged> uses to tag objects. |
673 |
|
674 |
You can look up CBOR tag value and emanings in the IANA registry at |
675 |
L<http://www.iana.org/assignments/cbor-tags/cbor-tags.xhtml>. |
676 |
|
677 |
Prepend a magic header (C<$CBOR::XS::MAGIC>): |
678 |
|
679 |
my $cbor = encode_cbor CBOR::XS::tag 55799, $value; |
680 |
# same as: |
681 |
my $cbor = $CBOR::XS::MAGIC . encode_cbor $value; |
682 |
|
683 |
Serialise some URIs and a regex in an array: |
684 |
|
685 |
my $cbor = encode_cbor [ |
686 |
(CBOR::XS::tag 32, "http://www.nethype.de/"), |
687 |
(CBOR::XS::tag 32, "http://software.schmorp.de/"), |
688 |
(CBOR::XS::tag 35, "^[Pp][Ee][Rr][lL]\$"), |
689 |
]; |
690 |
|
691 |
Wrap CBOR data in CBOR: |
692 |
|
693 |
my $cbor_cbor = encode_cbor |
694 |
CBOR::XS::tag 24, |
695 |
encode_cbor [1, 2, 3]; |
696 |
|
697 |
=head1 TAG HANDLING AND EXTENSIONS |
698 |
|
699 |
This section describes how this module handles specific tagged values |
700 |
and extensions. If a tag is not mentioned here and no additional filters |
701 |
are provided for it, then the default handling applies (creating a |
702 |
CBOR::XS::Tagged object on decoding, and only encoding the tag when |
703 |
explicitly requested). |
704 |
|
705 |
Tags not handled specifically are currently converted into a |
706 |
L<CBOR::XS::Tagged> object, which is simply a blessed array reference |
707 |
consisting of the numeric tag value followed by the (decoded) CBOR value. |
708 |
|
709 |
Future versions of this module reserve the right to special case |
710 |
additional tags (such as base64url). |
711 |
|
712 |
=head2 ENFORCED TAGS |
713 |
|
714 |
These tags are always handled when decoding, and their handling cannot be |
715 |
overriden by the user. |
716 |
|
717 |
=over 4 |
718 |
|
719 |
=item 26 (perl-object, L<http://cbor.schmorp.de/perl-object>) |
720 |
|
721 |
These tags are automatically created (and decoded) for serialisable |
722 |
objects using the C<FREEZE/THAW> methods (the L<Types::Serialier> object |
723 |
serialisation protocol). See L<OBJECT SERIALISATION> for details. |
724 |
|
725 |
=item 28, 29 (shareable, sharedref, L <http://cbor.schmorp.de/value-sharing>) |
726 |
|
727 |
These tags are automatically decoded when encountered (and they do not |
728 |
result in a cyclic data structure, see C<allow_cycles>), resulting in |
729 |
shared values in the decoded object. They are only encoded, however, when |
730 |
C<allow_sharing> is enabled. |
731 |
|
732 |
Not all shared values can be successfully decoded: values that reference |
733 |
themselves will I<currently> decode as C<undef> (this is not the same |
734 |
as a reference pointing to itself, which will be represented as a value |
735 |
that contains an indirect reference to itself - these will be decoded |
736 |
properly). |
737 |
|
738 |
Note that considerably more shared value data structures can be decoded |
739 |
than will be encoded - currently, only values pointed to by references |
740 |
will be shared, others will not. While non-reference shared values can be |
741 |
generated in Perl with some effort, they were considered too unimportant |
742 |
to be supported in the encoder. The decoder, however, will decode these |
743 |
values as shared values. |
744 |
|
745 |
=item 256, 25 (stringref-namespace, stringref, L <http://cbor.schmorp.de/stringref>) |
746 |
|
747 |
These tags are automatically decoded when encountered. They are only |
748 |
encoded, however, when C<pack_strings> is enabled. |
749 |
|
750 |
=item 22098 (indirection, L<http://cbor.schmorp.de/indirection>) |
751 |
|
752 |
This tag is automatically generated when a reference are encountered (with |
753 |
the exception of hash and array refernces). It is converted to a reference |
754 |
when decoding. |
755 |
|
756 |
=item 55799 (self-describe CBOR, RFC 7049) |
757 |
|
758 |
This value is not generated on encoding (unless explicitly requested by |
759 |
the user), and is simply ignored when decoding. |
760 |
|
761 |
=back |
762 |
|
763 |
=head2 NON-ENFORCED TAGS |
764 |
|
765 |
These tags have default filters provided when decoding. Their handling can |
766 |
be overriden by changing the C<%CBOR::XS::FILTER> entry for the tag, or by |
767 |
providing a custom C<filter> callback when decoding. |
768 |
|
769 |
When they result in decoding into a specific Perl class, the module |
770 |
usually provides a corresponding C<TO_CBOR> method as well. |
771 |
|
772 |
When any of these need to load additional modules that are not part of the |
773 |
perl core distribution (e.g. L<URI>), it is (currently) up to the user to |
774 |
provide these modules. The decoding usually fails with an exception if the |
775 |
required module cannot be loaded. |
776 |
|
777 |
=over 4 |
778 |
|
779 |
=item 2, 3 (positive/negative bignum) |
780 |
|
781 |
These tags are decoded into L<Math::BigInt> objects. The corresponding |
782 |
C<Math::BigInt::TO_CBOR> method encodes "small" bigints into normal CBOR |
783 |
integers, and others into positive/negative CBOR bignums. |
784 |
|
785 |
=item 4, 5 (decimal fraction/bigfloat) |
786 |
|
787 |
Both decimal fractions and bigfloats are decoded into L<Math::BigFloat> |
788 |
objects. The corresponding C<Math::BigFloat::TO_CBOR> method I<always> |
789 |
encodes into a decimal fraction. |
790 |
|
791 |
CBOR cannot represent bigfloats with I<very> large exponents - conversion |
792 |
of such big float objects is undefined. |
793 |
|
794 |
Also, NaN and infinities are not encoded properly. |
795 |
|
796 |
=item 21, 22, 23 (expected later JSON conversion) |
797 |
|
798 |
CBOR::XS is not a CBOR-to-JSON converter, and will simply ignore these |
799 |
tags. |
800 |
|
801 |
=item 32 (URI) |
802 |
|
803 |
These objects decode into L<URI> objects. The corresponding |
804 |
C<URI::TO_CBOR> method again results in a CBOR URI value. |
805 |
|
806 |
=back |
807 |
|
808 |
=cut |
809 |
|
810 |
our %FILTER = ( |
811 |
# 0 # rfc4287 datetime, utf-8 |
812 |
# 1 # unix timestamp, any |
813 |
|
814 |
2 => sub { # pos bigint |
815 |
require Math::BigInt; |
816 |
Math::BigInt->new ("0x" . unpack "H*", pop) |
817 |
}, |
818 |
|
819 |
3 => sub { # neg bigint |
820 |
require Math::BigInt; |
821 |
-Math::BigInt->new ("0x" . unpack "H*", pop) |
822 |
}, |
823 |
|
824 |
4 => sub { # decimal fraction, array |
825 |
require Math::BigFloat; |
826 |
Math::BigFloat->new ($_[1][1] . "E" . $_[1][0]) |
827 |
}, |
828 |
|
829 |
5 => sub { # bigfloat, array |
830 |
require Math::BigFloat; |
831 |
scalar Math::BigFloat->new ($_[1][1])->blsft ($_[1][0], 2) |
832 |
}, |
833 |
|
834 |
21 => sub { pop }, # expected conversion to base64url encoding |
835 |
22 => sub { pop }, # expected conversion to base64 encoding |
836 |
23 => sub { pop }, # expected conversion to base16 encoding |
837 |
|
838 |
# 24 # embedded cbor, byte string |
839 |
|
840 |
32 => sub { |
841 |
require URI; |
842 |
URI->new (pop) |
843 |
}, |
844 |
|
845 |
# 33 # base64url rfc4648, utf-8 |
846 |
# 34 # base64 rfc46484, utf-8 |
847 |
# 35 # regex pcre/ecma262, utf-8 |
848 |
# 36 # mime message rfc2045, utf-8 |
849 |
); |
850 |
|
851 |
|
852 |
=head1 CBOR and JSON |
853 |
|
854 |
CBOR is supposed to implement a superset of the JSON data model, and is, |
855 |
with some coercion, able to represent all JSON texts (something that other |
856 |
"binary JSON" formats such as BSON generally do not support). |
857 |
|
858 |
CBOR implements some extra hints and support for JSON interoperability, |
859 |
and the spec offers further guidance for conversion between CBOR and |
860 |
JSON. None of this is currently implemented in CBOR, and the guidelines |
861 |
in the spec do not result in correct round-tripping of data. If JSON |
862 |
interoperability is improved in the future, then the goal will be to |
863 |
ensure that decoded JSON data will round-trip encoding and decoding to |
864 |
CBOR intact. |
865 |
|
866 |
|
867 |
=head1 SECURITY CONSIDERATIONS |
868 |
|
869 |
When you are using CBOR in a protocol, talking to untrusted potentially |
870 |
hostile creatures requires relatively few measures. |
871 |
|
872 |
First of all, your CBOR decoder should be secure, that is, should not have |
873 |
any buffer overflows. Obviously, this module should ensure that and I am |
874 |
trying hard on making that true, but you never know. |
875 |
|
876 |
Second, you need to avoid resource-starving attacks. That means you should |
877 |
limit the size of CBOR data you accept, or make sure then when your |
878 |
resources run out, that's just fine (e.g. by using a separate process that |
879 |
can crash safely). The size of a CBOR string in octets is usually a good |
880 |
indication of the size of the resources required to decode it into a Perl |
881 |
structure. While CBOR::XS can check the size of the CBOR text, it might be |
882 |
too late when you already have it in memory, so you might want to check |
883 |
the size before you accept the string. |
884 |
|
885 |
Third, CBOR::XS recurses using the C stack when decoding objects and |
886 |
arrays. The C stack is a limited resource: for instance, on my amd64 |
887 |
machine with 8MB of stack size I can decode around 180k nested arrays but |
888 |
only 14k nested CBOR objects (due to perl itself recursing deeply on croak |
889 |
to free the temporary). If that is exceeded, the program crashes. To be |
890 |
conservative, the default nesting limit is set to 512. If your process |
891 |
has a smaller stack, you should adjust this setting accordingly with the |
892 |
C<max_depth> method. |
893 |
|
894 |
Something else could bomb you, too, that I forgot to think of. In that |
895 |
case, you get to keep the pieces. I am always open for hints, though... |
896 |
|
897 |
Also keep in mind that CBOR::XS might leak contents of your Perl data |
898 |
structures in its error messages, so when you serialise sensitive |
899 |
information you might want to make sure that exceptions thrown by CBOR::XS |
900 |
will not end up in front of untrusted eyes. |
901 |
|
902 |
=head1 CBOR IMPLEMENTATION NOTES |
903 |
|
904 |
This section contains some random implementation notes. They do not |
905 |
describe guaranteed behaviour, but merely behaviour as-is implemented |
906 |
right now. |
907 |
|
908 |
64 bit integers are only properly decoded when Perl was built with 64 bit |
909 |
support. |
910 |
|
911 |
Strings and arrays are encoded with a definite length. Hashes as well, |
912 |
unless they are tied (or otherwise magical). |
913 |
|
914 |
Only the double data type is supported for NV data types - when Perl uses |
915 |
long double to represent floating point values, they might not be encoded |
916 |
properly. Half precision types are accepted, but not encoded. |
917 |
|
918 |
Strict mode and canonical mode are not implemented. |
919 |
|
920 |
|
921 |
=head1 LIMITATIONS ON PERLS WITHOUT 64-BIT INTEGER SUPPORT |
922 |
|
923 |
On perls that were built without 64 bit integer support (these are rare |
924 |
nowadays, even on 32 bit architectures), support for any kind of 64 bit |
925 |
integer in CBOR is very limited - most likely, these 64 bit values will |
926 |
be truncated, corrupted, or otherwise not decoded correctly. This also |
927 |
includes string, array and map sizes that are stored as 64 bit integers. |
928 |
|
929 |
|
930 |
=head1 THREADS |
931 |
|
932 |
This module is I<not> guaranteed to be thread safe and there are no |
933 |
plans to change this until Perl gets thread support (as opposed to the |
934 |
horribly slow so-called "threads" which are simply slow and bloated |
935 |
process simulations - use fork, it's I<much> faster, cheaper, better). |
936 |
|
937 |
(It might actually work, but you have been warned). |
938 |
|
939 |
|
940 |
=head1 BUGS |
941 |
|
942 |
While the goal of this module is to be correct, that unfortunately does |
943 |
not mean it's bug-free, only that I think its design is bug-free. If you |
944 |
keep reporting bugs they will be fixed swiftly, though. |
945 |
|
946 |
Please refrain from using rt.cpan.org or any other bug reporting |
947 |
service. I put the contact address into my modules for a reason. |
948 |
|
949 |
=cut |
950 |
|
951 |
our %FILTER = ( |
952 |
# 0 # rfc4287 datetime, utf-8 |
953 |
# 1 # unix timestamp, any |
954 |
|
955 |
2 => sub { # pos bigint |
956 |
require Math::BigInt; |
957 |
Math::BigInt->new ("0x" . unpack "H*", pop) |
958 |
}, |
959 |
|
960 |
3 => sub { # neg bigint |
961 |
require Math::BigInt; |
962 |
-Math::BigInt->new ("0x" . unpack "H*", pop) |
963 |
}, |
964 |
|
965 |
4 => sub { # decimal fraction, array |
966 |
require Math::BigFloat; |
967 |
Math::BigFloat->new ($_[1][1] . "E" . $_[1][0]) |
968 |
}, |
969 |
|
970 |
5 => sub { # bigfloat, array |
971 |
require Math::BigFloat; |
972 |
scalar Math::BigFloat->new ($_[1][1])->blsft ($_[1][0], 2) |
973 |
}, |
974 |
|
975 |
21 => sub { pop }, # expected conversion to base64url encoding |
976 |
22 => sub { pop }, # expected conversion to base64 encoding |
977 |
23 => sub { pop }, # expected conversion to base16 encoding |
978 |
|
979 |
# 24 # embedded cbor, byte string |
980 |
|
981 |
32 => sub { |
982 |
require URI; |
983 |
URI->new (pop) |
984 |
}, |
985 |
|
986 |
# 33 # base64url rfc4648, utf-8 |
987 |
# 34 # base64 rfc46484, utf-8 |
988 |
# 35 # regex pcre/ecma262, utf-8 |
989 |
# 36 # mime message rfc2045, utf-8 |
990 |
); |
991 |
|
992 |
sub CBOR::XS::default_filter { |
993 |
&{ $FILTER{$_[0]} or return } |
994 |
} |
995 |
|
996 |
sub URI::TO_CBOR { |
997 |
my $uri = $_[0]->as_string; |
998 |
utf8::upgrade $uri; |
999 |
CBOR::XS::tag 32, $uri |
1000 |
} |
1001 |
|
1002 |
sub Math::BigInt::TO_CBOR { |
1003 |
if ($_[0] >= -2147483648 && $_[0] <= 2147483647) { |
1004 |
$_[0]->numify |
1005 |
} else { |
1006 |
my $hex = substr $_[0]->as_hex, 2; |
1007 |
$hex = "0$hex" if 1 & length $hex; # sigh |
1008 |
CBOR::XS::tag $_[0] >= 0 ? 2 : 3, pack "H*", $hex |
1009 |
} |
1010 |
} |
1011 |
|
1012 |
sub Math::BigFloat::TO_CBOR { |
1013 |
my ($m, $e) = $_[0]->parts; |
1014 |
CBOR::XS::tag 4, [$e->numify, $m] |
1015 |
} |
1016 |
|
1017 |
XSLoader::load "CBOR::XS", $VERSION; |
1018 |
|
1019 |
=head1 SEE ALSO |
1020 |
|
1021 |
The L<JSON> and L<JSON::XS> modules that do similar, but human-readable, |
1022 |
serialisation. |
1023 |
|
1024 |
The L<Types::Serialiser> module provides the data model for true, false |
1025 |
and error values. |
1026 |
|
1027 |
=head1 AUTHOR |
1028 |
|
1029 |
Marc Lehmann <schmorp@schmorp.de> |
1030 |
http://home.schmorp.de/ |
1031 |
|
1032 |
=cut |
1033 |
|
1034 |
1 |
1035 |
|