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 |
#TODO |
18 |
|
19 |
=head1 DESCRIPTION |
20 |
|
21 |
WARNING! THIS IS A PRE-ALPHA RELEASE! IT WILL CRASH, CORRUPT YOUR DATA AND |
22 |
EAT YOUR CHILDREN! |
23 |
|
24 |
This module converts Perl data structures to CBOR and vice versa. Its |
25 |
primary goal is to be I<correct> and its secondary goal is to be |
26 |
I<fast>. To reach the latter goal it was written in C. |
27 |
|
28 |
See MAPPING, below, on how CBOR::XS maps perl values to CBOR values and |
29 |
vice versa. |
30 |
|
31 |
=cut |
32 |
|
33 |
package CBOR::XS; |
34 |
|
35 |
use common::sense; |
36 |
|
37 |
our $VERSION = 0.02; |
38 |
our @ISA = qw(Exporter); |
39 |
|
40 |
our @EXPORT = qw(encode_cbor decode_cbor); |
41 |
|
42 |
use Exporter; |
43 |
use XSLoader; |
44 |
|
45 |
our $MAGIC = "\xd9\xd9\xf7"; |
46 |
|
47 |
=head1 FUNCTIONAL INTERFACE |
48 |
|
49 |
The following convenience methods are provided by this module. They are |
50 |
exported by default: |
51 |
|
52 |
=over 4 |
53 |
|
54 |
=item $cbor_data = encode_cbor $perl_scalar |
55 |
|
56 |
Converts the given Perl data structure to CBOR representation. Croaks on |
57 |
error. |
58 |
|
59 |
=item $perl_scalar = decode_cbor $cbor_data |
60 |
|
61 |
The opposite of C<encode_cbor>: expects a valid CBOR string to parse, |
62 |
returning the resulting perl scalar. Croaks on error. |
63 |
|
64 |
=back |
65 |
|
66 |
|
67 |
=head1 OBJECT-ORIENTED INTERFACE |
68 |
|
69 |
The object oriented interface lets you configure your own encoding or |
70 |
decoding style, within the limits of supported formats. |
71 |
|
72 |
=over 4 |
73 |
|
74 |
=item $cbor = new CBOR::XS |
75 |
|
76 |
Creates a new CBOR::XS object that can be used to de/encode CBOR |
77 |
strings. All boolean flags described below are by default I<disabled>. |
78 |
|
79 |
The mutators for flags all return the CBOR object again and thus calls can |
80 |
be chained: |
81 |
|
82 |
#TODO |
83 |
my $cbor = CBOR::XS->new->encode ({a => [1,2]}); |
84 |
|
85 |
=item $cbor = $cbor->max_depth ([$maximum_nesting_depth]) |
86 |
|
87 |
=item $max_depth = $cbor->get_max_depth |
88 |
|
89 |
Sets the maximum nesting level (default C<512>) accepted while encoding |
90 |
or decoding. If a higher nesting level is detected in CBOR data or a Perl |
91 |
data structure, then the encoder and decoder will stop and croak at that |
92 |
point. |
93 |
|
94 |
Nesting level is defined by number of hash- or arrayrefs that the encoder |
95 |
needs to traverse to reach a given point or the number of C<{> or C<[> |
96 |
characters without their matching closing parenthesis crossed to reach a |
97 |
given character in a string. |
98 |
|
99 |
Setting the maximum depth to one disallows any nesting, so that ensures |
100 |
that the object is only a single hash/object or array. |
101 |
|
102 |
If no argument is given, the highest possible setting will be used, which |
103 |
is rarely useful. |
104 |
|
105 |
Note that nesting is implemented by recursion in C. The default value has |
106 |
been chosen to be as large as typical operating systems allow without |
107 |
crashing. |
108 |
|
109 |
See SECURITY CONSIDERATIONS, below, for more info on why this is useful. |
110 |
|
111 |
=item $cbor = $cbor->max_size ([$maximum_string_size]) |
112 |
|
113 |
=item $max_size = $cbor->get_max_size |
114 |
|
115 |
Set the maximum length a CBOR string may have (in bytes) where decoding |
116 |
is being attempted. The default is C<0>, meaning no limit. When C<decode> |
117 |
is called on a string that is longer then this many bytes, it will not |
118 |
attempt to decode the string but throw an exception. This setting has no |
119 |
effect on C<encode> (yet). |
120 |
|
121 |
If no argument is given, the limit check will be deactivated (same as when |
122 |
C<0> is specified). |
123 |
|
124 |
See SECURITY CONSIDERATIONS, below, for more info on why this is useful. |
125 |
|
126 |
=item $cbor_data = $cbor->encode ($perl_scalar) |
127 |
|
128 |
Converts the given Perl data structure (a scalar value) to its CBOR |
129 |
representation. |
130 |
|
131 |
=item $perl_scalar = $cbor->decode ($cbor_data) |
132 |
|
133 |
The opposite of C<encode>: expects CBOR data and tries to parse it, |
134 |
returning the resulting simple scalar or reference. Croaks on error. |
135 |
|
136 |
=item ($perl_scalar, $octets) = $cbor->decode_prefix ($cbor_data) |
137 |
|
138 |
This works like the C<decode> method, but instead of raising an exception |
139 |
when there is trailing garbage after the CBOR string, it will silently |
140 |
stop parsing there and return the number of characters consumed so far. |
141 |
|
142 |
This is useful if your CBOR texts are not delimited by an outer protocol |
143 |
and you need to know where the first CBOR string ends amd the next one |
144 |
starts. |
145 |
|
146 |
CBOR::XS->new->decode_prefix ("......") |
147 |
=> ("...", 3) |
148 |
|
149 |
=back |
150 |
|
151 |
|
152 |
=head1 MAPPING |
153 |
|
154 |
This section describes how CBOR::XS maps Perl values to CBOR values and |
155 |
vice versa. These mappings are designed to "do the right thing" in most |
156 |
circumstances automatically, preserving round-tripping characteristics |
157 |
(what you put in comes out as something equivalent). |
158 |
|
159 |
For the more enlightened: note that in the following descriptions, |
160 |
lowercase I<perl> refers to the Perl interpreter, while uppercase I<Perl> |
161 |
refers to the abstract Perl language itself. |
162 |
|
163 |
|
164 |
=head2 CBOR -> PERL |
165 |
|
166 |
=over 4 |
167 |
|
168 |
=item integers |
169 |
|
170 |
CBOR integers become (numeric) perl scalars. On perls without 64 bit |
171 |
support, 64 bit integers will be truncated or otherwise corrupted. |
172 |
|
173 |
=item byte strings |
174 |
|
175 |
Byte strings will become octet strings in Perl (the byte values 0..255 |
176 |
will simply become characters of the same value in Perl). |
177 |
|
178 |
=item UTF-8 strings |
179 |
|
180 |
UTF-8 strings in CBOR will be decoded, i.e. the UTF-8 octets will be |
181 |
decoded into proper Unicode code points. At the moment, the validity of |
182 |
the UTF-8 octets will not be validated - corrupt input will result in |
183 |
corrupted Perl strings. |
184 |
|
185 |
=item arrays, maps |
186 |
|
187 |
CBOR arrays and CBOR maps will be converted into references to a Perl |
188 |
array or hash, respectively. The keys of the map will be stringified |
189 |
during this process. |
190 |
|
191 |
=item true, false |
192 |
|
193 |
These CBOR values become C<CBOR::XS::true> and C<CBOR::XS::false>, |
194 |
respectively. They are overloaded to act almost exactly like the numbers |
195 |
C<1> and C<0>. You can check whether a scalar is a CBOR boolean by using |
196 |
the C<CBOR::XS::is_bool> function. |
197 |
|
198 |
=item null, undefined |
199 |
|
200 |
CBOR null and undefined values becomes C<undef> in Perl (in the future, |
201 |
Undefined may raise an exception or something else). |
202 |
|
203 |
=item tags |
204 |
|
205 |
Tagged items consists of a numeric tag and another CBOR value. The tag |
206 |
55799 is ignored (this tag implements the magic header). |
207 |
|
208 |
All other tags are currently converted into a L<CBOR::XS::Tagged> object, |
209 |
which is simply a blessed array reference consistsing of the numeric tag |
210 |
value followed by the (decoded) BOR value. |
211 |
|
212 |
=item anything else |
213 |
|
214 |
Anything else (e.g. unsupported simple values) will raise a decoding |
215 |
error. |
216 |
|
217 |
=back |
218 |
|
219 |
|
220 |
=head2 PERL -> CBOR |
221 |
|
222 |
The mapping from Perl to CBOR is slightly more difficult, as Perl is a |
223 |
truly typeless language, so we can only guess which CBOR type is meant by |
224 |
a Perl value. |
225 |
|
226 |
=over 4 |
227 |
|
228 |
=item hash references |
229 |
|
230 |
Perl hash references become CBOR maps. As there is no inherent ordering in |
231 |
hash keys (or CBOR maps), they will usually be encoded in a pseudo-random |
232 |
order. |
233 |
|
234 |
Currently, tied hashes will use the indefinite-length format, while normal |
235 |
hashes will use the fixed-length format. |
236 |
|
237 |
=item array references |
238 |
|
239 |
Perl array references become fixed-length CBOR arrays. |
240 |
|
241 |
=item other references |
242 |
|
243 |
Other unblessed references are generally not allowed and will cause an |
244 |
exception to be thrown, except for references to the integers C<0> and |
245 |
C<1>, which get turned into false and true in CBOR. |
246 |
|
247 |
=item CBOR::XS::Tagged objects |
248 |
|
249 |
Objects of this type must be arrays consisting of a single C<[tag, value]> |
250 |
pair. The (numerical) tag will be encoded as a CBOR tag, the value will be |
251 |
encoded as appropriate for the value. |
252 |
|
253 |
=item CBOR::XS::true, CBOR::XS::false |
254 |
|
255 |
These special values become CBOR true and CBOR false values, |
256 |
respectively. You can also use C<\1> and C<\0> directly if you want. |
257 |
|
258 |
=item blessed objects |
259 |
|
260 |
Other blessed objects currently need to have a C<TO_CBOR> method. It |
261 |
will be called on every object that is being serialised, and must return |
262 |
something that can be encoded in CBOR. |
263 |
|
264 |
=item simple scalars |
265 |
|
266 |
TODO |
267 |
Simple Perl scalars (any scalar that is not a reference) are the most |
268 |
difficult objects to encode: CBOR::XS will encode undefined scalars as |
269 |
CBOR null values, scalars that have last been used in a string context |
270 |
before encoding as CBOR strings, and anything else as number value: |
271 |
|
272 |
# dump as number |
273 |
encode_cbor [2] # yields [2] |
274 |
encode_cbor [-3.0e17] # yields [-3e+17] |
275 |
my $value = 5; encode_cbor [$value] # yields [5] |
276 |
|
277 |
# used as string, so dump as string |
278 |
print $value; |
279 |
encode_cbor [$value] # yields ["5"] |
280 |
|
281 |
# undef becomes null |
282 |
encode_cbor [undef] # yields [null] |
283 |
|
284 |
You can force the type to be a CBOR string by stringifying it: |
285 |
|
286 |
my $x = 3.1; # some variable containing a number |
287 |
"$x"; # stringified |
288 |
$x .= ""; # another, more awkward way to stringify |
289 |
print $x; # perl does it for you, too, quite often |
290 |
|
291 |
You can force the type to be a CBOR number by numifying it: |
292 |
|
293 |
my $x = "3"; # some variable containing a string |
294 |
$x += 0; # numify it, ensuring it will be dumped as a number |
295 |
$x *= 1; # same thing, the choice is yours. |
296 |
|
297 |
You can not currently force the type in other, less obscure, ways. Tell me |
298 |
if you need this capability (but don't forget to explain why it's needed |
299 |
:). |
300 |
|
301 |
Perl values that seem to be integers generally use the shortest possible |
302 |
representation. Floating-point values will use either the IEEE single |
303 |
format if possible without loss of precision, otherwise the IEEE double |
304 |
format will be used. Perls that use formats other than IEEE double to |
305 |
represent numerical values are supported, but might suffer loss of |
306 |
precision. |
307 |
|
308 |
=back |
309 |
|
310 |
|
311 |
=head2 MAGIC HEADER |
312 |
|
313 |
There is no way to distinguish CBOR from other formats |
314 |
programmatically. To make it easier to distinguish CBOR from other |
315 |
formats, the CBOR specification has a special "magic string" that can be |
316 |
prepended to any CBOR string without changing it's meaning. |
317 |
|
318 |
This string is available as C<$CBOR::XS::MAGIC>. This module does not |
319 |
prepend this string tot he CBOR data it generates, but it will ignroe it |
320 |
if present, so users can prepend this string as a "file type" indicator as |
321 |
required. |
322 |
|
323 |
|
324 |
=head2 CBOR and JSON |
325 |
|
326 |
CBOR is supposed to implement a superset of the JSON data model, and is, |
327 |
with some coercion, able to represent all JSON texts (something that other |
328 |
"binary JSON" formats such as BSON generally do not support). |
329 |
|
330 |
CBOR implements some extra hints and support for JSON interoperability, |
331 |
and the spec offers further guidance for conversion between CBOR and |
332 |
JSON. None of this is currently implemented in CBOR, and the guidelines |
333 |
in the spec do not result in correct round-tripping of data. If JSON |
334 |
interoperability is improved in the future, then the goal will be to |
335 |
ensure that decoded JSON data will round-trip encoding and decoding to |
336 |
CBOR intact. |
337 |
|
338 |
|
339 |
=head1 SECURITY CONSIDERATIONS |
340 |
|
341 |
When you are using CBOR in a protocol, talking to untrusted potentially |
342 |
hostile creatures requires relatively few measures. |
343 |
|
344 |
First of all, your CBOR decoder should be secure, that is, should not have |
345 |
any buffer overflows. Obviously, this module should ensure that and I am |
346 |
trying hard on making that true, but you never know. |
347 |
|
348 |
Second, you need to avoid resource-starving attacks. That means you should |
349 |
limit the size of CBOR data you accept, or make sure then when your |
350 |
resources run out, that's just fine (e.g. by using a separate process that |
351 |
can crash safely). The size of a CBOR string in octets is usually a good |
352 |
indication of the size of the resources required to decode it into a Perl |
353 |
structure. While CBOR::XS can check the size of the CBOR text, it might be |
354 |
too late when you already have it in memory, so you might want to check |
355 |
the size before you accept the string. |
356 |
|
357 |
Third, CBOR::XS recurses using the C stack when decoding objects and |
358 |
arrays. The C stack is a limited resource: for instance, on my amd64 |
359 |
machine with 8MB of stack size I can decode around 180k nested arrays but |
360 |
only 14k nested CBOR objects (due to perl itself recursing deeply on croak |
361 |
to free the temporary). If that is exceeded, the program crashes. To be |
362 |
conservative, the default nesting limit is set to 512. If your process |
363 |
has a smaller stack, you should adjust this setting accordingly with the |
364 |
C<max_depth> method. |
365 |
|
366 |
Something else could bomb you, too, that I forgot to think of. In that |
367 |
case, you get to keep the pieces. I am always open for hints, though... |
368 |
|
369 |
Also keep in mind that CBOR::XS might leak contents of your Perl data |
370 |
structures in its error messages, so when you serialise sensitive |
371 |
information you might want to make sure that exceptions thrown by CBOR::XS |
372 |
will not end up in front of untrusted eyes. |
373 |
|
374 |
=head1 CBOR IMPLEMENTATION NOTES |
375 |
|
376 |
This section contains some random implementation notes. They do not |
377 |
describe guaranteed behaviour, but merely behaviour as-is implemented |
378 |
right now. |
379 |
|
380 |
64 bit integers are only properly decoded when Perl was built with 64 bit |
381 |
support. |
382 |
|
383 |
Strings and arrays are encoded with a definite length. Hashes as well, |
384 |
unless they are tied (or otherwise magical). |
385 |
|
386 |
Only the double data type is supported for NV data types - when Perl uses |
387 |
long double to represent floating point values, they might not be encoded |
388 |
properly. Half precision types are accepted, but not encoded. |
389 |
|
390 |
Strict mode and canonical mode are not implemented. |
391 |
|
392 |
|
393 |
=head1 THREADS |
394 |
|
395 |
This module is I<not> guaranteed to be thread safe and there are no |
396 |
plans to change this until Perl gets thread support (as opposed to the |
397 |
horribly slow so-called "threads" which are simply slow and bloated |
398 |
process simulations - use fork, it's I<much> faster, cheaper, better). |
399 |
|
400 |
(It might actually work, but you have been warned). |
401 |
|
402 |
|
403 |
=head1 BUGS |
404 |
|
405 |
While the goal of this module is to be correct, that unfortunately does |
406 |
not mean it's bug-free, only that I think its design is bug-free. If you |
407 |
keep reporting bugs they will be fixed swiftly, though. |
408 |
|
409 |
Please refrain from using rt.cpan.org or any other bug reporting |
410 |
service. I put the contact address into my modules for a reason. |
411 |
|
412 |
=cut |
413 |
|
414 |
our $true = do { bless \(my $dummy = 1), "CBOR::XS::Boolean" }; |
415 |
our $false = do { bless \(my $dummy = 0), "CBOR::XS::Boolean" }; |
416 |
|
417 |
sub true() { $true } |
418 |
sub false() { $false } |
419 |
|
420 |
sub is_bool($) { |
421 |
UNIVERSAL::isa $_[0], "CBOR::XS::Boolean" |
422 |
# or UNIVERSAL::isa $_[0], "CBOR::Literal" |
423 |
} |
424 |
|
425 |
XSLoader::load "CBOR::XS", $VERSION; |
426 |
|
427 |
package CBOR::XS::Boolean; |
428 |
|
429 |
use overload |
430 |
"0+" => sub { ${$_[0]} }, |
431 |
"++" => sub { $_[0] = ${$_[0]} + 1 }, |
432 |
"--" => sub { $_[0] = ${$_[0]} - 1 }, |
433 |
fallback => 1; |
434 |
|
435 |
1; |
436 |
|
437 |
=head1 SEE ALSO |
438 |
|
439 |
The L<JSON> and L<JSON::XS> modules that do similar, but human-readable, |
440 |
serialisation. |
441 |
|
442 |
=head1 AUTHOR |
443 |
|
444 |
Marc Lehmann <schmorp@schmorp.de> |
445 |
http://home.schmorp.de/ |
446 |
|
447 |
=cut |
448 |
|