ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/CBOR-XS/XS.pm

Comparing CBOR-XS/XS.pm (file contents):
Revision 1.3 by root, Sat Oct 26 11:08:34 2013 UTC vs.
Revision 1.88 by root, Thu Sep 7 23:52:24 2023 UTC

…		…
12	$perl_value = decode_cbor $binary_cbor_data;	12	$perl_value = decode_cbor $binary_cbor_data;
13		13
14	# OO-interface	14	# OO-interface
15		15
16	$coder = CBOR::XS->new;	16	$coder = CBOR::XS->new;
17	#TODO	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	}
18		28
19	=head1 DESCRIPTION	29	=head1 DESCRIPTION
20		30
21	WARNING! THIS IS A PRE-ALPHA RELEASE! IT WILL CRASH, CORRUPT YOUR DATA AND	31	This module converts Perl data structures to the Concise Binary Object
22	EAT YOUR CHILDREN!	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.
23		36
24	This module converts Perl data structures to CBOR and vice versa. Its	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
25	primary goal is to be I<correct> and its secondary goal is to be	43	The primary goal of this module is to be I<correct> and the secondary goal
26	I<fast>. To reach the latter goal it was written in C.	44	is to be I<fast>. To reach the latter goal it was written in C.
		45
		46	To give you a general idea about speed, with texts in the megabyte range,
		47	C<CBOR::XS> usually encodes roughly twice as fast as L<Storable> or
		48	L<JSON::XS> and decodes about 15%-30% faster than those. The shorter the
		49	data, the worse L<Storable> performs in comparison.
		50
		51	Regarding compactness, C<CBOR::XS>-encoded data structures are usually
		52	about 20% smaller than the same data encoded as (compact) JSON or
		53	L<Storable>.
		54
		55	In addition to the core CBOR data format, this module implements a
		56	number of extensions, to support cyclic and shared data structures
		57	(see C<allow_sharing> and C<allow_cycles>), string deduplication (see
		58	C<pack_strings>) and scalar references (always enabled).
27		59
28	See MAPPING, below, on how CBOR::XS maps perl values to CBOR values and	60	See MAPPING, below, on how CBOR::XS maps perl values to CBOR values and
29	vice versa.	61	vice versa.
30		62
31	=cut	63	=cut
32		64
33	package CBOR::XS;	65	package CBOR::XS;
34		66
35	use common::sense;	67	use common::sense;
36		68
37	our $VERSION = 0.02;	69	our $VERSION = 1.86;
38	our @ISA = qw(Exporter);	70	our @ISA = qw(Exporter);
39		71
40	our @EXPORT = qw(encode_cbor decode_cbor);	72	our @EXPORT = qw(encode_cbor decode_cbor);
41		73
42	use Exporter;	74	use Exporter;
43	use XSLoader;	75	use XSLoader;
44		76
		77	use Types::Serialiser;
		78
45	our $MAGIC = "\xd9\xd9\xf7";	79	our $MAGIC = "\xd9\xd9\xf7";
46		80
47	=head1 FUNCTIONAL INTERFACE	81	=head1 FUNCTIONAL INTERFACE
48		82
49	The following convenience methods are provided by this module. They are	83	The following convenience methods are provided by this module. They are
…		…
77	strings. All boolean flags described below are by default I<disabled>.	111	strings. All boolean flags described below are by default I<disabled>.
78		112
79	The mutators for flags all return the CBOR object again and thus calls can	113	The mutators for flags all return the CBOR object again and thus calls can
80	be chained:	114	be chained:
81		115
82	#TODO
83	my $cbor = CBOR::XS->new->encode ({a => [1,2]});	116	my $cbor = CBOR::XS->new->encode ({a => [1,2]});
		117
		118	=item $cbor = new_safe CBOR::XS
		119
		120	Create a new, safe/secure CBOR::XS object. This is similar to C<new>,
		121	but configures the coder object to be safe to use with untrusted
		122	data. Currently, this is equivalent to:
		123
		124	my $cbor = CBOR::XS
		125	->new
		126	->validate_utf8
		127	->forbid_objects
		128	->filter (\&CBOR::XS::safe_filter)
		129	->max_size (1e8);
		130
		131	But is more future proof (it is better to crash because of a change than
		132	to be exploited in other ways).
		133
		134	=cut
		135
		136	sub new_safe {
		137	CBOR::XS
		138	->new
		139	->validate_utf8
		140	->forbid_objects
		141	->filter (\&CBOR::XS::safe_filter)
		142	->max_size (1e8)
		143	}
84		144
85	=item $cbor = $cbor->max_depth ([$maximum_nesting_depth])	145	=item $cbor = $cbor->max_depth ([$maximum_nesting_depth])
86		146
87	=item $max_depth = $cbor->get_max_depth	147	=item $max_depth = $cbor->get_max_depth
88		148
…		…
104		164
105	Note that nesting is implemented by recursion in C. The default value has	165	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	166	been chosen to be as large as typical operating systems allow without
107	crashing.	167	crashing.
108		168
109	See SECURITY CONSIDERATIONS, below, for more info on why this is useful.	169	See L<SECURITY CONSIDERATIONS>, below, for more info on why this is useful.
110		170
111	=item $cbor = $cbor->max_size ([$maximum_string_size])	171	=item $cbor = $cbor->max_size ([$maximum_string_size])
112		172
113	=item $max_size = $cbor->get_max_size	173	=item $max_size = $cbor->get_max_size
114		174
…		…
119	effect on C<encode> (yet).	179	effect on C<encode> (yet).
120		180
121	If no argument is given, the limit check will be deactivated (same as when	181	If no argument is given, the limit check will be deactivated (same as when
122	C<0> is specified).	182	C<0> is specified).
123		183
124	See SECURITY CONSIDERATIONS, below, for more info on why this is useful.	184	See L<SECURITY CONSIDERATIONS>, below, for more info on why this is useful.
		185
		186	=item $cbor = $cbor->allow_unknown ([$enable])
		187
		188	=item $enabled = $cbor->get_allow_unknown
		189
		190	If C<$enable> is true (or missing), then C<encode> will I<not> throw an
		191	exception when it encounters values it cannot represent in CBOR (for
		192	example, filehandles) but instead will encode a CBOR C<error> value.
		193
		194	If C<$enable> is false (the default), then C<encode> will throw an
		195	exception when it encounters anything it cannot encode as CBOR.
		196
		197	This option does not affect C<decode> in any way, and it is recommended to
		198	leave it off unless you know your communications partner.
		199
		200	=item $cbor = $cbor->allow_sharing ([$enable])
		201
		202	=item $enabled = $cbor->get_allow_sharing
		203
		204	If C<$enable> is true (or missing), then C<encode> will not double-encode
		205	values that have been referenced before (e.g. when the same object, such
		206	as an array, is referenced multiple times), but instead will emit a
		207	reference to the earlier value.
		208
		209	This means that such values will only be encoded once, and will not result
		210	in a deep cloning of the value on decode, in decoders supporting the value
		211	sharing extension. This also makes it possible to encode cyclic data
		212	structures (which need C<allow_cycles> to be enabled to be decoded by this
		213	module).
		214
		215	It is recommended to leave it off unless you know your
		216	communication partner supports the value sharing extensions to CBOR
		217	(L<http://cbor.schmorp.de/value-sharing>), as without decoder support, the
		218	resulting data structure might be unusable.
		219
		220	Detecting shared values incurs a runtime overhead when values are encoded
		221	that have a reference counter larger than one, and might unnecessarily
		222	increase the encoded size, as potentially shared values are encoded as
		223	shareable whether or not they are actually shared.
		224
		225	At the moment, only targets of references can be shared (e.g. scalars,
		226	arrays or hashes pointed to by a reference). Weirder constructs, such as
		227	an array with multiple "copies" of the I<same> string, which are hard but
		228	not impossible to create in Perl, are not supported (this is the same as
		229	with L<Storable>).
		230
		231	If C<$enable> is false (the default), then C<encode> will encode shared
		232	data structures repeatedly, unsharing them in the process. Cyclic data
		233	structures cannot be encoded in this mode.
		234
		235	This option does not affect C<decode> in any way - shared values and
		236	references will always be decoded properly if present.
		237
		238	=item $cbor = $cbor->allow_cycles ([$enable])
		239
		240	=item $enabled = $cbor->get_allow_cycles
		241
		242	If C<$enable> is true (or missing), then C<decode> will happily decode
		243	self-referential (cyclic) data structures. By default these will not be
		244	decoded, as they need manual cleanup to avoid memory leaks, so code that
		245	isn't prepared for this will not leak memory.
		246
		247	If C<$enable> is false (the default), then C<decode> will throw an error
		248	when it encounters a self-referential/cyclic data structure.
		249
		250	FUTURE DIRECTION: the motivation behind this option is to avoid I<real>
		251	cycles - future versions of this module might chose to decode cyclic data
		252	structures using weak references when this option is off, instead of
		253	throwing an error.
		254
		255	This option does not affect C<encode> in any way - shared values and
		256	references will always be encoded properly if present.
		257
		258	=item $cbor = $cbor->forbid_objects ([$enable])
		259
		260	=item $enabled = $cbor->get_forbid_objects
		261
		262	Disables the use of the object serialiser protocol.
		263
		264	If C<$enable> is true (or missing), then C<encode> will will throw an
		265	exception when it encounters perl objects that would be encoded using the
		266	perl-object tag (26). When C<decode> encounters such tags, it will fall
		267	back to the general filter/tagged logic as if this were an unknown tag (by
		268	default resulting in a C<CBOR::XC::Tagged> object).
		269
		270	If C<$enable> is false (the default), then C<encode> will use the
		271	L<Types::Serialiser> object serialisation protocol to serialise objects
		272	into perl-object tags, and C<decode> will do the same to decode such tags.
		273
		274	See L<SECURITY CONSIDERATIONS>, below, for more info on why forbidding this
		275	protocol can be useful.
		276
		277	=item $cbor = $cbor->pack_strings ([$enable])
		278
		279	=item $enabled = $cbor->get_pack_strings
		280
		281	If C<$enable> is true (or missing), then C<encode> will try not to encode
		282	the same string twice, but will instead encode a reference to the string
		283	instead. Depending on your data format, this can save a lot of space, but
		284	also results in a very large runtime overhead (expect encoding times to be
		285	2-4 times as high as without).
		286
		287	It is recommended to leave it off unless you know your
		288	communications partner supports the stringref extension to CBOR
		289	(L<http://cbor.schmorp.de/stringref>), as without decoder support, the
		290	resulting data structure might not be usable.
		291
		292	If C<$enable> is false (the default), then C<encode> will encode strings
		293	the standard CBOR way.
		294
		295	This option does not affect C<decode> in any way - string references will
		296	always be decoded properly if present.
		297
		298	=item $cbor = $cbor->text_keys ([$enable])
		299
		300	=item $enabled = $cbor->get_text_keys
		301
		302	If C<$enabled> is true (or missing), then C<encode> will encode all
		303	perl hash keys as CBOR text strings/UTF-8 string, upgrading them as needed.
		304
		305	If C<$enable> is false (the default), then C<encode> will encode hash keys
		306	normally - upgraded perl strings (strings internally encoded as UTF-8) as
		307	CBOR text strings, and downgraded perl strings as CBOR byte strings.
		308
		309	This option does not affect C<decode> in any way.
		310
		311	This option is useful for interoperability with CBOR decoders that don't
		312	treat byte strings as a form of text. It is especially useful as Perl
		313	gives very little control over hash keys.
		314
		315	Enabling this option can be slow, as all downgraded hash keys that are
		316	encoded need to be scanned and converted to UTF-8.
		317
		318	=item $cbor = $cbor->text_strings ([$enable])
		319
		320	=item $enabled = $cbor->get_text_strings
		321
		322	This option works similar to C<text_keys>, above, but works on all strings
		323	(including hash keys), so C<text_keys> has no further effect after
		324	enabling C<text_strings>.
		325
		326	If C<$enabled> is true (or missing), then C<encode> will encode all perl
		327	strings as CBOR text strings/UTF-8 strings, upgrading them as needed.
		328
		329	If C<$enable> is false (the default), then C<encode> will encode strings
		330	normally (but see C<text_keys>) - upgraded perl strings (strings
		331	internally encoded as UTF-8) as CBOR text strings, and downgraded perl
		332	strings as CBOR byte strings.
		333
		334	This option does not affect C<decode> in any way.
		335
		336	This option has similar advantages and disadvantages as C<text_keys>. In
		337	addition, this option effectively removes the ability to automatically
		338	encode byte strings, which might break some C<FREEZE> and C<TO_CBOR>
		339	methods that rely on this.
		340
		341	A workaround is to use explicit type casts, which are unaffected by this option.
		342
		343	=item $cbor = $cbor->validate_utf8 ([$enable])
		344
		345	=item $enabled = $cbor->get_validate_utf8
		346
		347	If C<$enable> is true (or missing), then C<decode> will validate that
		348	elements (text strings) containing UTF-8 data in fact contain valid UTF-8
		349	data (instead of blindly accepting it). This validation obviously takes
		350	extra time during decoding.
		351
		352	The concept of "valid UTF-8" used is perl's concept, which is a superset
		353	of the official UTF-8.
		354
		355	If C<$enable> is false (the default), then C<decode> will blindly accept
		356	UTF-8 data, marking them as valid UTF-8 in the resulting data structure
		357	regardless of whether that's true or not.
		358
		359	Perl isn't too happy about corrupted UTF-8 in strings, but should
		360	generally not crash or do similarly evil things. Extensions might be not
		361	so forgiving, so it's recommended to turn on this setting if you receive
		362	untrusted CBOR.
		363
		364	This option does not affect C<encode> in any way - strings that are
		365	supposedly valid UTF-8 will simply be dumped into the resulting CBOR
		366	string without checking whether that is, in fact, true or not.
		367
		368	=item $cbor = $cbor->filter ([$cb->($tag, $value)])
		369
		370	=item $cb_or_undef = $cbor->get_filter
		371
		372	Sets or replaces the tagged value decoding filter (when C<$cb> is
		373	specified) or clears the filter (if no argument or C<undef> is provided).
		374
		375	The filter callback is called only during decoding, when a non-enforced
		376	tagged value has been decoded (see L<TAG HANDLING AND EXTENSIONS> for a
		377	list of enforced tags). For specific tags, it's often better to provide a
		378	default converter using the C<%CBOR::XS::FILTER> hash (see below).
		379
		380	The first argument is the numerical tag, the second is the (decoded) value
		381	that has been tagged.
		382
		383	The filter function should return either exactly one value, which will
		384	replace the tagged value in the decoded data structure, or no values,
		385	which will result in default handling, which currently means the decoder
		386	creates a C<CBOR::XS::Tagged> object to hold the tag and the value.
		387
		388	When the filter is cleared (the default state), the default filter
		389	function, C<CBOR::XS::default_filter>, is used. This function simply
		390	looks up the tag in the C<%CBOR::XS::FILTER> hash. If an entry exists
		391	it must be a code reference that is called with tag and value, and is
		392	responsible for decoding the value. If no entry exists, it returns no
		393	values. C<CBOR::XS> provides a number of default filter functions already,
		394	the the C<%CBOR::XS::FILTER> hash can be freely extended with more.
		395
		396	C<CBOR::XS> additionally provides an alternative filter function that is
		397	supposed to be safe to use with untrusted data (which the default filter
		398	might not), called C<CBOR::XS::safe_filter>, which works the same as
		399	the C<default_filter> but uses the C<%CBOR::XS::SAFE_FILTER> variable
		400	instead. It is prepopulated with the tag decoding functions that are
		401	deemed safe (basically the same as C<%CBOR::XS::FILTER> without all
		402	the bignum tags), and can be extended by user code as wlel, although,
		403	obviously, one should be very careful about adding decoding functions
		404	here, since the expectation is that they are safe to use on untrusted
		405	data, after all.
		406
		407	Example: decode all tags not handled internally into C<CBOR::XS::Tagged>
		408	objects, with no other special handling (useful when working with
		409	potentially "unsafe" CBOR data).
		410
		411	CBOR::XS->new->filter (sub { })->decode ($cbor_data);
		412
		413	Example: provide a global filter for tag 1347375694, converting the value
		414	into some string form.
		415
		416	$CBOR::XS::FILTER{1347375694} = sub {
		417	my ($tag, $value);
		418
		419	"tag 1347375694 value $value"
		420	};
		421
		422	Example: provide your own filter function that looks up tags in your own
		423	hash:
		424
		425	my %my_filter = (
		426	998347484 => sub {
		427	my ($tag, $value);
		428
		429	"tag 998347484 value $value"
		430	};
		431	);
		432
		433	my $coder = CBOR::XS->new->filter (sub {
		434	&{ $my_filter{$_[0]} or return }
		435	});
		436
		437
		438	Example: use the safe filter function (see L<SECURITY CONSIDERATIONS> for
		439	more considerations on security).
		440
		441	CBOR::XS->new->filter (\&CBOR::XS::safe_filter)->decode ($cbor_data);
125		442
126	=item $cbor_data = $cbor->encode ($perl_scalar)	443	=item $cbor_data = $cbor->encode ($perl_scalar)
127		444
128	Converts the given Perl data structure (a scalar value) to its CBOR	445	Converts the given Perl data structure (a scalar value) to its CBOR
129	representation.	446	representation.
…		…
139	when there is trailing garbage after the CBOR string, it will silently	456	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.	457	stop parsing there and return the number of characters consumed so far.
141		458
142	This is useful if your CBOR texts are not delimited by an outer protocol	459	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	460	and you need to know where the first CBOR string ends amd the next one
144	starts.	461	starts - CBOR strings are self-delimited, so it is possible to concatenate
		462	CBOR strings without any delimiters or size fields and recover their data.
145		463
146	CBOR::XS->new->decode_prefix ("......")	464	CBOR::XS->new->decode_prefix ("......")
147	=> ("...", 3)	465	=> ("...", 3)
		466
		467	=back
		468
		469	=head2 INCREMENTAL PARSING
		470
		471	In some cases, there is the need for incremental parsing of JSON
		472	texts. While this module always has to keep both CBOR text and resulting
		473	Perl data structure in memory at one time, it does allow you to parse a
		474	CBOR stream incrementally, using a similar to using "decode_prefix" to see
		475	if a full CBOR object is available, but is much more efficient.
		476
		477	It basically works by parsing as much of a CBOR string as possible - if
		478	the CBOR data is not complete yet, the parser will remember where it was,
		479	to be able to restart when more data has been accumulated. Once enough
		480	data is available to either decode a complete CBOR value or raise an
		481	error, a real decode will be attempted.
		482
		483	A typical use case would be a network protocol that consists of sending
		484	and receiving CBOR-encoded messages. The solution that works with CBOR and
		485	about anything else is by prepending a length to every CBOR value, so the
		486	receiver knows how many octets to read. More compact (and slightly slower)
		487	would be to just send CBOR values back-to-back, as C<CBOR::XS> knows where
		488	a CBOR value ends, and doesn't need an explicit length.
		489
		490	The following methods help with this:
		491
		492	=over 4
		493
		494	=item @decoded = $cbor->incr_parse ($buffer)
		495
		496	This method attempts to decode exactly one CBOR value from the beginning
		497	of the given C<$buffer>. The value is removed from the C<$buffer> on
		498	success. When C<$buffer> doesn't contain a complete value yet, it returns
		499	nothing. Finally, when the C<$buffer> doesn't start with something
		500	that could ever be a valid CBOR value, it raises an exception, just as
		501	C<decode> would. In the latter case the decoder state is undefined and
		502	must be reset before being able to parse further.
		503
		504	This method modifies the C<$buffer> in place. When no CBOR value can be
		505	decoded, the decoder stores the current string offset. On the next call,
		506	continues decoding at the place where it stopped before. For this to make
		507	sense, the C<$buffer> must begin with the same octets as on previous
		508	unsuccessful calls.
		509
		510	You can call this method in scalar context, in which case it either
		511	returns a decoded value or C<undef>. This makes it impossible to
		512	distinguish between CBOR null values (which decode to C<undef>) and an
		513	unsuccessful decode, which is often acceptable.
		514
		515	=item @decoded = $cbor->incr_parse_multiple ($buffer)
		516
		517	Same as C<incr_parse>, but attempts to decode as many CBOR values as
		518	possible in one go, instead of at most one. Calls to C<incr_parse> and
		519	C<incr_parse_multiple> can be interleaved.
		520
		521	=item $cbor->incr_reset
		522
		523	Resets the incremental decoder. This throws away any saved state, so that
		524	subsequent calls to C<incr_parse> or C<incr_parse_multiple> start to parse
		525	a new CBOR value from the beginning of the C<$buffer> again.
		526
		527	This method can be called at any time, but it I<must> be called if you want
		528	to change your C<$buffer> or there was a decoding error and you want to
		529	reuse the C<$cbor> object for future incremental parsings.
148		530
149	=back	531	=back
150		532
151		533
152	=head1 MAPPING	534	=head1 MAPPING
…		…
163		545
164	=head2 CBOR -> PERL	546	=head2 CBOR -> PERL
165		547
166	=over 4	548	=over 4
167		549
168	=item True, False	550	=item integers
169		551
170	These CBOR values become C<CBOR::XS::true> and C<CBOR::XS::false>,	552	CBOR integers become (numeric) perl scalars. On perls without 64 bit
		553	support, 64 bit integers will be truncated or otherwise corrupted.
		554
		555	=item byte strings
		556
		557	Byte strings will become octet strings in Perl (the Byte values 0..255
		558	will simply become characters of the same value in Perl).
		559
		560	=item UTF-8 strings
		561
		562	UTF-8 strings in CBOR will be decoded, i.e. the UTF-8 octets will be
		563	decoded into proper Unicode code points. At the moment, the validity of
		564	the UTF-8 octets will not be validated - corrupt input will result in
		565	corrupted Perl strings.
		566
		567	=item arrays, maps
		568
		569	CBOR arrays and CBOR maps will be converted into references to a Perl
		570	array or hash, respectively. The keys of the map will be stringified
		571	during this process.
		572
		573	=item null
		574
		575	CBOR null becomes C<undef> in Perl.
		576
		577	=item true, false, undefined
		578
		579	These CBOR values become C<Types:Serialiser::true>,
		580	C<Types:Serialiser::false> and C<Types::Serialiser::error>,
171	respectively. They are overloaded to act almost exactly like the numbers	581	respectively. They are overloaded to act almost exactly like the numbers
172	C<1> and C<0>. You can check whether a scalar is a CBOR boolean by using	582	C<1> and C<0> (for true and false) or to throw an exception on access (for
173	the C<CBOR::XS::is_bool> function.	583	error). See the L<Types::Serialiser> manpage for details.
174		584
175	=item Null, Undefined	585	=item tagged values
176		586
177	CBOR Null and Undefined values becomes C<undef> in Perl (in the future,	587	Tagged items consists of a numeric tag and another CBOR value.
178	Undefined may raise an exception).	588
		589	See L<TAG HANDLING AND EXTENSIONS> and the description of C<< ->filter >>
		590	for details on which tags are handled how.
		591
		592	=item anything else
		593
		594	Anything else (e.g. unsupported simple values) will raise a decoding
		595	error.
179		596
180	=back	597	=back
181		598
182		599
183	=head2 PERL -> CBOR	600	=head2 PERL -> CBOR
184		601
185	The mapping from Perl to CBOR is slightly more difficult, as Perl is a	602	The mapping from Perl to CBOR is slightly more difficult, as Perl is a
186	truly typeless language, so we can only guess which CBOR type is meant by	603	typeless language. That means this module can only guess which CBOR type
187	a Perl value.	604	is meant by a perl value.
188		605
189	=over 4	606	=over 4
190		607
191	=item hash references	608	=item hash references
192		609
193	Perl hash references become CBOR maps. As there is no inherent ordering	610	Perl hash references become CBOR maps. As there is no inherent ordering in
194	in hash keys (or CBOR maps), they will usually be encoded in a	611	hash keys (or CBOR maps), they will usually be encoded in a pseudo-random
195	pseudo-random order.	612	order. This order can be different each time a hash is encoded.
		613
		614	Currently, tied hashes will use the indefinite-length format, while normal
		615	hashes will use the fixed-length format.
196		616
197	=item array references	617	=item array references
198		618
199	Perl array references become CBOR arrays.	619	Perl array references become fixed-length CBOR arrays.
200		620
201	=item other references	621	=item other references
202		622
203	Other unblessed references are generally not allowed and will cause an	623	Other unblessed references will be represented using
204	exception to be thrown, except for references to the integers C<0> and	624	the indirection tag extension (tag value C<22098>,
205	C<1>, which get turned into C<False> and C<True> in CBOR.	625	L<http://cbor.schmorp.de/indirection>). CBOR decoders are guaranteed
		626	to be able to decode these values somehow, by either "doing the right
		627	thing", decoding into a generic tagged object, simply ignoring the tag, or
		628	something else.
206		629
207	=item CBOR::XS::true, CBOR::XS::false	630	=item CBOR::XS::Tagged objects
208		631
		632	Objects of this type must be arrays consisting of a single C<[tag, value]>
		633	pair. The (numerical) tag will be encoded as a CBOR tag, the value will
		634	be encoded as appropriate for the value. You must use C<CBOR::XS::tag> to
		635	create such objects.
		636
		637	=item Types::Serialiser::true, Types::Serialiser::false, Types::Serialiser::error
		638
209	These special values become CBOR True and CBOR False values,	639	These special values become CBOR true, CBOR false and CBOR undefined
210	respectively. You can also use C<\1> and C<\0> directly if you want.	640	values, respectively.
211		641
212	=item blessed objects	642	=item other blessed objects
213		643
214	Blessed objects are not directly representable in CBOR. TODO	644	Other blessed objects are serialised via C<TO_CBOR> or C<FREEZE>. See
215	See the	645	L<TAG HANDLING AND EXTENSIONS> for specific classes handled by this
216	C<allow_blessed> and C<convert_blessed> methods on various options on	646	module, and L<OBJECT SERIALISATION> for generic object serialisation.
217	how to deal with this: basically, you can choose between throwing an
218	exception, encoding the reference as if it weren't blessed, or provide
219	your own serialiser method.
220		647
221	=item simple scalars	648	=item simple scalars
222		649
223	TODO
224	Simple Perl scalars (any scalar that is not a reference) are the most	650	Simple Perl scalars (any scalar that is not a reference) are the most
225	difficult objects to encode: CBOR::XS will encode undefined scalars as	651	difficult objects to encode: CBOR::XS will encode undefined scalars as
226	CBOR C<Null> values, scalars that have last been used in a string context	652	CBOR null values, scalars that have last been used in a string context
227	before encoding as CBOR strings, and anything else as number value:	653	before encoding as CBOR strings, and anything else as number value:
228		654
229	# dump as number	655	# dump as number
230	encode_cbor [2] # yields [2]	656	encode_cbor [2] # yields [2]
231	encode_cbor [-3.0e17] # yields [-3e+17]	657	encode_cbor [-3.0e17] # yields [-3e+17]
232	my $value = 5; encode_cbor [$value] # yields [5]	658	my $value = 5; encode_cbor [$value] # yields [5]
233		659
234	# used as string, so dump as string	660	# used as string, so dump as string (either byte or text)
235	print $value;	661	print $value;
236	encode_cbor [$value] # yields ["5"]	662	encode_cbor [$value] # yields ["5"]
237		663
238	# undef becomes null	664	# undef becomes null
239	encode_cbor [undef] # yields [null]	665	encode_cbor [undef] # yields [null]
…		…
243	my $x = 3.1; # some variable containing a number	669	my $x = 3.1; # some variable containing a number
244	"$x"; # stringified	670	"$x"; # stringified
245	$x .= ""; # another, more awkward way to stringify	671	$x .= ""; # another, more awkward way to stringify
246	print $x; # perl does it for you, too, quite often	672	print $x; # perl does it for you, too, quite often
247		673
		674	You can force whether a string is encoded as byte or text string by using
		675	C<utf8::upgrade> and C<utf8::downgrade> (if C<text_strings> is disabled).
		676
		677	utf8::upgrade $x; # encode $x as text string
		678	utf8::downgrade $x; # encode $x as byte string
		679
		680	More options are available, see L<TYPE CASTS>, below, and the C<text_keys>
		681	and C<text_strings> options.
		682
		683	Perl doesn't define what operations up- and downgrade strings, so if the
		684	difference between byte and text is important, you should up- or downgrade
		685	your string as late as possible before encoding. You can also force the
		686	use of CBOR text strings by using C<text_keys> or C<text_strings>.
		687
248	You can force the type to be a CBOR number by numifying it:	688	You can force the type to be a CBOR number by numifying it:
249		689
250	my $x = "3"; # some variable containing a string	690	my $x = "3"; # some variable containing a string
251	$x += 0; # numify it, ensuring it will be dumped as a number	691	$x += 0; # numify it, ensuring it will be dumped as a number
252	$x *= 1; # same thing, the choice is yours.	692	$x *= 1; # same thing, the choice is yours.
253		693
254	You can not currently force the type in other, less obscure, ways. Tell me	694	You can not currently force the type in other, less obscure, ways. Tell me
255	if you need this capability (but don't forget to explain why it's needed	695	if you need this capability (but don't forget to explain why it's needed
256	:).	696	:).
257		697
258	Note that numerical precision has the same meaning as under Perl (so	698	Perl values that seem to be integers generally use the shortest possible
259	binary to decimal conversion follows the same rules as in Perl, which	699	representation. Floating-point values will use either the IEEE single
260	can differ to other languages). Also, your perl interpreter might expose	700	format if possible without loss of precision, otherwise the IEEE double
261	extensions to the floating point numbers of your platform, such as	701	format will be used. Perls that use formats other than IEEE double to
262	infinities or NaN's - these cannot be represented in CBOR, and it is an	702	represent numerical values are supported, but might suffer loss of
263	error to pass those in.	703	precision.
264		704
265	=back	705	=back
266		706
		707	=head2 TYPE CASTS
267		708
		709	B<EXPERIMENTAL>: As an experimental extension, C<CBOR::XS> allows you to
		710	force specific CBOR types to be used when encoding. That allows you to
		711	encode types not normally accessible (e.g. half floats) as well as force
		712	string types even when C<text_strings> is in effect.
		713
		714	Type forcing is done by calling a special "cast" function which keeps a
		715	copy of the value and returns a new value that can be handed over to any
		716	CBOR encoder function.
		717
		718	The following casts are currently available (all of which are unary
		719	operators, that is, have a prototype of C<$>):
		720
		721	=over
		722
		723	=item CBOR::XS::as_int $value
		724
		725	Forces the value to be encoded as some form of (basic, not bignum) integer
		726	type.
		727
		728	=item CBOR::XS::as_text $value
		729
		730	Forces the value to be encoded as (UTF-8) text values.
		731
		732	=item CBOR::XS::as_bytes $value
		733
		734	Forces the value to be encoded as a (binary) string value.
		735
		736	Example: encode a perl string as binary even though C<text_strings> is in
		737	effect.
		738
		739	CBOR::XS->new->text_strings->encode ([4, "text", CBOR::XS::bytes "bytevalue"]);
		740
		741	=item CBOR::XS::as_bool $value
		742
		743	Converts a Perl boolean (which can be any kind of scalar) into a CBOR
		744	boolean. Strictly the same, but shorter to write, than:
		745
		746	$value ? Types::Serialiser::true : Types::Serialiser::false
		747
		748	=item CBOR::XS::as_float16 $value
		749
		750	Forces half-float (IEEE 754 binary16) encoding of the given value.
		751
		752	=item CBOR::XS::as_float32 $value
		753
		754	Forces single-float (IEEE 754 binary32) encoding of the given value.
		755
		756	=item CBOR::XS::as_float64 $value
		757
		758	Forces double-float (IEEE 754 binary64) encoding of the given value.
		759
		760	=item CBOR::XS::as_cbor $cbor_text
		761
		762	Not a type cast per-se, this type cast forces the argument to be encoded
		763	as-is. This can be used to embed pre-encoded CBOR data.
		764
		765	Note that no checking on the validity of the C<$cbor_text> is done - it's
		766	the callers responsibility to correctly encode values.
		767
		768	=item CBOR::XS::as_map [key => value...]
		769
		770	Treat the array reference as key value pairs and output a CBOR map. This
		771	allows you to generate CBOR maps with arbitrary key types (or, if you
		772	don't care about semantics, duplicate keys or pairs in a custom order),
		773	which is otherwise hard to do with Perl.
		774
		775	The single argument must be an array reference with an even number of
		776	elements.
		777
		778	Note that only the reference to the array is copied, the array itself is
		779	not. Modifications done to the array before calling an encoding function
		780	will be reflected in the encoded output.
		781
		782	Example: encode a CBOR map with a string and an integer as keys.
		783
		784	encode_cbor CBOR::XS::as_map [string => "value", 5 => "value"]
		785
		786	=back
		787
		788	=cut
		789
		790	sub CBOR::XS::as_cbor ($) { bless [$_[0], 0, undef], CBOR::XS::Tagged:: }
		791	sub CBOR::XS::as_int ($) { bless [$_[0], 1, undef], CBOR::XS::Tagged:: }
		792	sub CBOR::XS::as_bytes ($) { bless [$_[0], 2, undef], CBOR::XS::Tagged:: }
		793	sub CBOR::XS::as_text ($) { bless [$_[0], 3, undef], CBOR::XS::Tagged:: }
		794	sub CBOR::XS::as_float16 ($) { bless [$_[0], 4, undef], CBOR::XS::Tagged:: }
		795	sub CBOR::XS::as_float32 ($) { bless [$_[0], 5, undef], CBOR::XS::Tagged:: }
		796	sub CBOR::XS::as_float64 ($) { bless [$_[0], 6, undef], CBOR::XS::Tagged:: }
		797
		798	sub CBOR::XS::as_bool ($) { $_[0] ? $Types::Serialiser::true : $Types::Serialiser::false }
		799
		800	sub CBOR::XS::as_map ($) {
		801	ARRAY:: eq ref $_[0]
		802	and $#{ $_[0] } & 1
		803	or do { require Carp; Carp::croak ("CBOR::XS::as_map only acepts array references with an even number of elements, caught") };
		804
		805	bless [$_[0], 7, undef], CBOR::XS::Tagged::
		806	}
		807
		808	=head2 OBJECT SERIALISATION
		809
		810	This module implements both a CBOR-specific and the generic
		811	L<Types::Serialier> object serialisation protocol. The following
		812	subsections explain both methods.
		813
		814	=head3 ENCODING
		815
		816	This module knows two way to serialise a Perl object: The CBOR-specific
		817	way, and the generic way.
		818
		819	Whenever the encoder encounters a Perl object that it cannot serialise
		820	directly (most of them), it will first look up the C<TO_CBOR> method on
		821	it.
		822
		823	If it has a C<TO_CBOR> method, it will call it with the object as only
		824	argument, and expects exactly one return value, which it will then
		825	substitute and encode it in the place of the object.
		826
		827	Otherwise, it will look up the C<FREEZE> method. If it exists, it will
		828	call it with the object as first argument, and the constant string C<CBOR>
		829	as the second argument, to distinguish it from other serialisers.
		830
		831	The C<FREEZE> method can return any number of values (i.e. zero or
		832	more). These will be encoded as CBOR perl object, together with the
		833	classname.
		834
		835	These methods I<MUST NOT> change the data structure that is being
		836	serialised. Failure to comply to this can result in memory corruption -
		837	and worse.
		838
		839	If an object supports neither C<TO_CBOR> nor C<FREEZE>, encoding will fail
		840	with an error.
		841
		842	=head3 DECODING
		843
		844	Objects encoded via C<TO_CBOR> cannot (normally) be automatically decoded,
		845	but objects encoded via C<FREEZE> can be decoded using the following
		846	protocol:
		847
		848	When an encoded CBOR perl object is encountered by the decoder, it will
		849	look up the C<THAW> method, by using the stored classname, and will fail
		850	if the method cannot be found.
		851
		852	After the lookup it will call the C<THAW> method with the stored classname
		853	as first argument, the constant string C<CBOR> as second argument, and all
		854	values returned by C<FREEZE> as remaining arguments.
		855
		856	=head3 EXAMPLES
		857
		858	Here is an example C<TO_CBOR> method:
		859
		860	sub My::Object::TO_CBOR {
		861	my ($obj) = @_;
		862
		863	["this is a serialised My::Object object", $obj->{id}]
		864	}
		865
		866	When a C<My::Object> is encoded to CBOR, it will instead encode a simple
		867	array with two members: a string, and the "object id". Decoding this CBOR
		868	string will yield a normal perl array reference in place of the object.
		869
		870	A more useful and practical example would be a serialisation method for
		871	the URI module. CBOR has a custom tag value for URIs, namely 32:
		872
		873	sub URI::TO_CBOR {
		874	my ($self) = @_;
		875	my $uri = "$self"; # stringify uri
		876	utf8::upgrade $uri; # make sure it will be encoded as UTF-8 string
		877	CBOR::XS::tag 32, "$_[0]"
		878	}
		879
		880	This will encode URIs as a UTF-8 string with tag 32, which indicates an
		881	URI.
		882
		883	Decoding such an URI will not (currently) give you an URI object, but
		884	instead a CBOR::XS::Tagged object with tag number 32 and the string -
		885	exactly what was returned by C<TO_CBOR>.
		886
		887	To serialise an object so it can automatically be deserialised, you need
		888	to use C<FREEZE> and C<THAW>. To take the URI module as example, this
		889	would be a possible implementation:
		890
		891	sub URI::FREEZE {
		892	my ($self, $serialiser) = @_;
		893	"$self" # encode url string
		894	}
		895
		896	sub URI::THAW {
		897	my ($class, $serialiser, $uri) = @_;
		898	$class->new ($uri)
		899	}
		900
		901	Unlike C<TO_CBOR>, multiple values can be returned by C<FREEZE>. For
		902	example, a C<FREEZE> method that returns "type", "id" and "variant" values
		903	would cause an invocation of C<THAW> with 5 arguments:
		904
		905	sub My::Object::FREEZE {
		906	my ($self, $serialiser) = @_;
		907
		908	($self->{type}, $self->{id}, $self->{variant})
		909	}
		910
		911	sub My::Object::THAW {
		912	my ($class, $serialiser, $type, $id, $variant) = @_;
		913
		914	$class-<new (type => $type, id => $id, variant => $variant)
		915	}
		916
		917
268	=head2 MAGIC HEADER	918	=head1 MAGIC HEADER
269		919
270	There is no way to distinguish CBOR from other formats	920	There is no way to distinguish CBOR from other formats
271	programmatically. To make it easier to distinguish CBOR from other	921	programmatically. To make it easier to distinguish CBOR from other
272	formats, the CBOR specification has a special "magic string" that can be	922	formats, the CBOR specification has a special "magic string" that can be
273	prepended to any CBOR string without changing it's meaning.	923	prepended to any CBOR string without changing its meaning.
274		924
275	This string is available as C<$CBOR::XS::MAGIC>. This module does not	925	This string is available as C<$CBOR::XS::MAGIC>. This module does not
276	prepend this string tot he CBOR data it generates, but it will ignroe it	926	prepend this string to the CBOR data it generates, but it will ignore it
277	if present, so users can prepend this string as a "file type" indicator as	927	if present, so users can prepend this string as a "file type" indicator as
278	required.	928	required.
279		929
280		930
		931	=head1 THE CBOR::XS::Tagged CLASS
		932
		933	CBOR has the concept of tagged values - any CBOR value can be tagged with
		934	a numeric 64 bit number, which are centrally administered.
		935
		936	C<CBOR::XS> handles a few tags internally when en- or decoding. You can
		937	also create tags yourself by encoding C<CBOR::XS::Tagged> objects, and the
		938	decoder will create C<CBOR::XS::Tagged> objects itself when it hits an
		939	unknown tag.
		940
		941	These objects are simply blessed array references - the first member of
		942	the array being the numerical tag, the second being the value.
		943
		944	You can interact with C<CBOR::XS::Tagged> objects in the following ways:
		945
		946	=over 4
		947
		948	=item $tagged = CBOR::XS::tag $tag, $value
		949
		950	This function(!) creates a new C<CBOR::XS::Tagged> object using the given
		951	C<$tag> (0..2**64-1) to tag the given C<$value> (which can be any Perl
		952	value that can be encoded in CBOR, including serialisable Perl objects and
		953	C<CBOR::XS::Tagged> objects).
		954
		955	=item $tagged->[0]
		956
		957	=item $tagged->[0] = $new_tag
		958
		959	=item $tag = $tagged->tag
		960
		961	=item $new_tag = $tagged->tag ($new_tag)
		962
		963	Access/mutate the tag.
		964
		965	=item $tagged->[1]
		966
		967	=item $tagged->[1] = $new_value
		968
		969	=item $value = $tagged->value
		970
		971	=item $new_value = $tagged->value ($new_value)
		972
		973	Access/mutate the tagged value.
		974
		975	=back
		976
		977	=cut
		978
		979	sub tag($$) {
		980	bless [@_], CBOR::XS::Tagged::;
		981	}
		982
		983	sub CBOR::XS::Tagged::tag {
		984	$_[0][0] = $_[1] if $#_;
		985	$_[0][0]
		986	}
		987
		988	sub CBOR::XS::Tagged::value {
		989	$_[0][1] = $_[1] if $#_;
		990	$_[0][1]
		991	}
		992
		993	=head2 EXAMPLES
		994
		995	Here are some examples of C<CBOR::XS::Tagged> uses to tag objects.
		996
		997	You can look up CBOR tag value and emanings in the IANA registry at
		998	L<http://www.iana.org/assignments/cbor-tags/cbor-tags.xhtml>.
		999
		1000	Prepend a magic header (C<$CBOR::XS::MAGIC>):
		1001
		1002	my $cbor = encode_cbor CBOR::XS::tag 55799, $value;
		1003	# same as:
		1004	my $cbor = $CBOR::XS::MAGIC . encode_cbor $value;
		1005
		1006	Serialise some URIs and a regex in an array:
		1007
		1008	my $cbor = encode_cbor [
		1009	(CBOR::XS::tag 32, "http://www.nethype.de/"),
		1010	(CBOR::XS::tag 32, "http://software.schmorp.de/"),
		1011	(CBOR::XS::tag 35, "^[Pp][Ee][Rr][lL]\$"),
		1012	];
		1013
		1014	Wrap CBOR data in CBOR:
		1015
		1016	my $cbor_cbor = encode_cbor
		1017	CBOR::XS::tag 24,
		1018	encode_cbor [1, 2, 3];
		1019
		1020	=head1 TAG HANDLING AND EXTENSIONS
		1021
		1022	This section describes how this module handles specific tagged values
		1023	and extensions. If a tag is not mentioned here and no additional filters
		1024	are provided for it, then the default handling applies (creating a
		1025	CBOR::XS::Tagged object on decoding, and only encoding the tag when
		1026	explicitly requested).
		1027
		1028	Tags not handled specifically are currently converted into a
		1029	L<CBOR::XS::Tagged> object, which is simply a blessed array reference
		1030	consisting of the numeric tag value followed by the (decoded) CBOR value.
		1031
		1032	Future versions of this module reserve the right to special case
		1033	additional tags (such as base64url).
		1034
		1035	=head2 ENFORCED TAGS
		1036
		1037	These tags are always handled when decoding, and their handling cannot be
		1038	overridden by the user.
		1039
		1040	=over 4
		1041
		1042	=item 26 (perl-object, L<http://cbor.schmorp.de/perl-object>)
		1043
		1044	These tags are automatically created (and decoded) for serialisable
		1045	objects using the C<FREEZE/THAW> methods (the L<Types::Serialier> object
		1046	serialisation protocol). See L<OBJECT SERIALISATION> for details.
		1047
		1048	=item 28, 29 (shareable, sharedref, L<http://cbor.schmorp.de/value-sharing>)
		1049
		1050	These tags are automatically decoded when encountered (and they do not
		1051	result in a cyclic data structure, see C<allow_cycles>), resulting in
		1052	shared values in the decoded object. They are only encoded, however, when
		1053	C<allow_sharing> is enabled.
		1054
		1055	Not all shared values can be successfully decoded: values that reference
		1056	themselves will I<currently> decode as C<undef> (this is not the same
		1057	as a reference pointing to itself, which will be represented as a value
		1058	that contains an indirect reference to itself - these will be decoded
		1059	properly).
		1060
		1061	Note that considerably more shared value data structures can be decoded
		1062	than will be encoded - currently, only values pointed to by references
		1063	will be shared, others will not. While non-reference shared values can be
		1064	generated in Perl with some effort, they were considered too unimportant
		1065	to be supported in the encoder. The decoder, however, will decode these
		1066	values as shared values.
		1067
		1068	=item 256, 25 (stringref-namespace, stringref, L<http://cbor.schmorp.de/stringref>)
		1069
		1070	These tags are automatically decoded when encountered. They are only
		1071	encoded, however, when C<pack_strings> is enabled.
		1072
		1073	=item 22098 (indirection, L<http://cbor.schmorp.de/indirection>)
		1074
		1075	This tag is automatically generated when a reference are encountered (with
		1076	the exception of hash and array references). It is converted to a reference
		1077	when decoding.
		1078
		1079	=item 55799 (self-describe CBOR, RFC 7049)
		1080
		1081	This value is not generated on encoding (unless explicitly requested by
		1082	the user), and is simply ignored when decoding.
		1083
		1084	=back
		1085
		1086	=head2 NON-ENFORCED TAGS
		1087
		1088	These tags have default filters provided when decoding. Their handling can
		1089	be overridden by changing the C<%CBOR::XS::FILTER> entry for the tag, or by
		1090	providing a custom C<filter> callback when decoding.
		1091
		1092	When they result in decoding into a specific Perl class, the module
		1093	usually provides a corresponding C<TO_CBOR> method as well.
		1094
		1095	When any of these need to load additional modules that are not part of the
		1096	perl core distribution (e.g. L<URI>), it is (currently) up to the user to
		1097	provide these modules. The decoding usually fails with an exception if the
		1098	required module cannot be loaded.
		1099
		1100	=over 4
		1101
		1102	=item 0, 1 (date/time string, seconds since the epoch)
		1103
		1104	These tags are decoded into L<Time::Piece> objects. The corresponding
		1105	C<Time::Piece::TO_CBOR> method always encodes into tag 1 values currently.
		1106
		1107	The L<Time::Piece> API is generally surprisingly bad, and fractional
		1108	seconds are only accidentally kept intact, so watch out. On the plus side,
		1109	the module comes with perl since 5.10, which has to count for something.
		1110
		1111	=item 2, 3 (positive/negative bignum)
		1112
		1113	These tags are decoded into L<Math::BigInt> objects. The corresponding
		1114	C<Math::BigInt::TO_CBOR> method encodes "small" bigints into normal CBOR
		1115	integers, and others into positive/negative CBOR bignums.
		1116
		1117	=item 4, 5, 264, 265 (decimal fraction/bigfloat)
		1118
		1119	Both decimal fractions and bigfloats are decoded into L<Math::BigFloat>
		1120	objects. The corresponding C<Math::BigFloat::TO_CBOR> method I<always>
		1121	encodes into a decimal fraction (either tag 4 or 264).
		1122
		1123	NaN and infinities are not encoded properly, as they cannot be represented
		1124	in CBOR.
		1125
		1126	See L<BIGNUM SECURITY CONSIDERATIONS> for more info.
		1127
		1128	=item 30 (rational numbers)
		1129
		1130	These tags are decoded into L<Math::BigRat> objects. The corresponding
		1131	C<Math::BigRat::TO_CBOR> method encodes rational numbers with denominator
		1132	C<1> via their numerator only, i.e., they become normal integers or
		1133	C<bignums>.
		1134
		1135	See L<BIGNUM SECURITY CONSIDERATIONS> for more info.
		1136
		1137	=item 21, 22, 23 (expected later JSON conversion)
		1138
		1139	CBOR::XS is not a CBOR-to-JSON converter, and will simply ignore these
		1140	tags.
		1141
		1142	=item 32 (URI)
		1143
		1144	These objects decode into L<URI> objects. The corresponding
		1145	C<URI::TO_CBOR> method again results in a CBOR URI value.
		1146
		1147	=back
		1148
		1149	=cut
		1150
281	=head2 CBOR and JSON	1151	=head1 CBOR and JSON
282		1152
283	TODO	1153	CBOR is supposed to implement a superset of the JSON data model, and is,
		1154	with some coercion, able to represent all JSON texts (something that other
		1155	"binary JSON" formats such as BSON generally do not support).
		1156
		1157	CBOR implements some extra hints and support for JSON interoperability,
		1158	and the spec offers further guidance for conversion between CBOR and
		1159	JSON. None of this is currently implemented in CBOR, and the guidelines
		1160	in the spec do not result in correct round-tripping of data. If JSON
		1161	interoperability is improved in the future, then the goal will be to
		1162	ensure that decoded JSON data will round-trip encoding and decoding to
		1163	CBOR intact.
284		1164
285		1165
286	=head1 SECURITY CONSIDERATIONS	1166	=head1 SECURITY CONSIDERATIONS
287		1167
288	When you are using CBOR in a protocol, talking to untrusted potentially	1168	Tl;dr... if you want to decode or encode CBOR from untrusted sources, you
289	hostile creatures requires relatively few measures.	1169	should start with a coder object created via C<new_safe> (which implements
		1170	the mitigations explained below):
290		1171
		1172	my $coder = CBOR::XS->new_safe;
		1173
		1174	my $data = $coder->decode ($cbor_text);
		1175	my $cbor = $coder->encode ($data);
		1176
		1177	Longer version: When you are using CBOR in a protocol, talking to
		1178	untrusted potentially hostile creatures requires some thought:
		1179
		1180	=over 4
		1181
		1182	=item Security of the CBOR decoder itself
		1183
291	First of all, your CBOR decoder should be secure, that is, should not have	1184	First and foremost, your CBOR decoder should be secure, that is, should
		1185	not have any buffer overflows or similar bugs that could potentially be
292	any buffer overflows. Obviously, this module should ensure that and I am	1186	exploited. Obviously, this module should ensure that and I am trying hard
293	trying hard on making that true, but you never know.	1187	on making that true, but you never know.
294		1188
		1189	=item CBOR::XS can invoke almost arbitrary callbacks during decoding
		1190
		1191	CBOR::XS supports object serialisation - decoding CBOR can cause calls
		1192	to I<any> C<THAW> method in I<any> package that exists in your process
		1193	(that is, CBOR::XS will not try to load modules, but any existing C<THAW>
		1194	method or function can be called, so they all have to be secure).
		1195
		1196	Less obviously, it will also invoke C<TO_CBOR> and C<FREEZE> methods -
		1197	even if all your C<THAW> methods are secure, encoding data structures from
		1198	untrusted sources can invoke those and trigger bugs in those.
		1199
		1200	So, if you are not sure about the security of all the modules you
		1201	have loaded (you shouldn't), you should disable this part using
		1202	C<forbid_objects> or using C<new_safe>.
		1203
		1204	=item CBOR can be extended with tags that call library code
		1205
		1206	CBOR can be extended with tags, and C<CBOR::XS> has a registry of
		1207	conversion functions for many existing tags that can be extended via
		1208	third-party modules (see the C<filter> method).
		1209
		1210	If you don't trust these, you should configure the "safe" filter function,
		1211	C<CBOR::XS::safe_filter> (C<new_safe> does this), which by default only
		1212	includes conversion functions that are considered "safe" by the author
		1213	(but again, they can be extended by third party modules).
		1214
		1215	Depending on your level of paranoia, you can use the "safe" filter:
		1216
		1217	$cbor->filter (\&CBOR::XS::safe_filter);
		1218
		1219	... your own filter...
		1220
		1221	$cbor->filter (sub { ... do your stuffs here ... });
		1222
		1223	... or even no filter at all, disabling all tag decoding:
		1224
		1225	$cbor->filter (sub { });
		1226
		1227	This is never a problem for encoding, as the tag mechanism only exists in
		1228	CBOR texts.
		1229
		1230	=item Resource-starving attacks: object memory usage
		1231
295	Second, you need to avoid resource-starving attacks. That means you should	1232	You need to avoid resource-starving attacks. That means you should limit
296	limit the size of CBOR data you accept, or make sure then when your	1233	the size of CBOR data you accept, or make sure then when your resources
297	resources run out, that's just fine (e.g. by using a separate process that	1234	run out, that's just fine (e.g. by using a separate process that can
298	can crash safely). The size of a CBOR string in octets is usually a good	1235	crash safely). The size of a CBOR string in octets is usually a good
299	indication of the size of the resources required to decode it into a Perl	1236	indication of the size of the resources required to decode it into a Perl
300	structure. While CBOR::XS can check the size of the CBOR text, it might be	1237	structure. While CBOR::XS can check the size of the CBOR text (using
301	too late when you already have it in memory, so you might want to check	1238	C<max_size> - done by C<new_safe>), it might be too late when you already
302	the size before you accept the string.	1239	have it in memory, so you might want to check the size before you accept
		1240	the string.
303		1241
		1242	As for encoding, it is possible to construct data structures that are
		1243	relatively small but result in large CBOR texts (for example by having an
		1244	array full of references to the same big data structure, which will all be
		1245	deep-cloned during encoding by default). This is rarely an actual issue
		1246	(and the worst case is still just running out of memory), but you can
		1247	reduce this risk by using C<allow_sharing>.
		1248
		1249	=item Resource-starving attacks: stack overflows
		1250
304	Third, CBOR::XS recurses using the C stack when decoding objects and	1251	CBOR::XS recurses using the C stack when decoding objects and arrays. The
305	arrays. The C stack is a limited resource: for instance, on my amd64	1252	C stack is a limited resource: for instance, on my amd64 machine with 8MB
306	machine with 8MB of stack size I can decode around 180k nested arrays but	1253	of stack size I can decode around 180k nested arrays but only 14k nested
307	only 14k nested CBOR objects (due to perl itself recursing deeply on croak	1254	CBOR objects (due to perl itself recursing deeply on croak to free the
308	to free the temporary). If that is exceeded, the program crashes. To be	1255	temporary). If that is exceeded, the program crashes. To be conservative,
309	conservative, the default nesting limit is set to 512. If your process	1256	the default nesting limit is set to 512. If your process has a smaller
310	has a smaller stack, you should adjust this setting accordingly with the	1257	stack, you should adjust this setting accordingly with the C<max_depth>
311	C<max_depth> method.	1258	method.
		1259
		1260	=item Resource-starving attacks: CPU en-/decoding complexity
		1261
		1262	CBOR::XS will use the L<Math::BigInt>, L<Math::BigFloat> and
		1263	L<Math::BigRat> libraries to represent encode/decode bignums. These can be
		1264	very slow (as in, centuries of CPU time) and can even crash your program
		1265	(and are generally not very trustworthy). See the next section on bignum
		1266	security for details.
		1267
		1268	=item Data breaches: leaking information in error messages
		1269
		1270	CBOR::XS might leak contents of your Perl data structures in its error
		1271	messages, so when you serialise sensitive information you might want to
		1272	make sure that exceptions thrown by CBOR::XS will not end up in front of
		1273	untrusted eyes.
		1274
		1275	=item Something else...
312		1276
313	Something else could bomb you, too, that I forgot to think of. In that	1277	Something else could bomb you, too, that I forgot to think of. In that
314	case, you get to keep the pieces. I am always open for hints, though...	1278	case, you get to keep the pieces. I am always open for hints, though...
315		1279
316	Also keep in mind that CBOR::XS might leak contents of your Perl data	1280	=back
317	structures in its error messages, so when you serialise sensitive	1281
318	information you might want to make sure that exceptions thrown by CBOR::XS	1282
319	will not end up in front of untrusted eyes.	1283	=head1 BIGNUM SECURITY CONSIDERATIONS
		1284
		1285	CBOR::XS provides a C<TO_CBOR> method for both L<Math::BigInt> and
		1286	L<Math::BigFloat> that tries to encode the number in the simplest possible
		1287	way, that is, either a CBOR integer, a CBOR bigint/decimal fraction (tag
		1288	4) or an arbitrary-exponent decimal fraction (tag 264). Rational numbers
		1289	(L<Math::BigRat>, tag 30) can also contain bignums as members.
		1290
		1291	CBOR::XS will also understand base-2 bigfloat or arbitrary-exponent
		1292	bigfloats (tags 5 and 265), but it will never generate these on its own.
		1293
		1294	Using the built-in L<Math::BigInt::Calc> support, encoding and decoding
		1295	decimal fractions is generally fast. Decoding bigints can be slow for very
		1296	big numbers (tens of thousands of digits, something that could potentially
		1297	be caught by limiting the size of CBOR texts), and decoding bigfloats or
		1298	arbitrary-exponent bigfloats can be I<extremely> slow (minutes, decades)
		1299	for large exponents (roughly 40 bit and longer).
		1300
		1301	Additionally, L<Math::BigInt> can take advantage of other bignum
		1302	libraries, such as L<Math::GMP>, which cannot handle big floats with large
		1303	exponents, and might simply abort or crash your program, due to their code
		1304	quality.
		1305
		1306	This can be a concern if you want to parse untrusted CBOR. If it is, you
		1307	might want to disable decoding of tag 2 (bigint) and 3 (negative bigint)
		1308	types. You should also disable types 5 and 265, as these can be slow even
		1309	without bigints.
		1310
		1311	Disabling bigints will also partially or fully disable types that rely on
		1312	them, e.g. rational numbers that use bignums.
		1313
320		1314
321	=head1 CBOR IMPLEMENTATION NOTES	1315	=head1 CBOR IMPLEMENTATION NOTES
322		1316
323	This section contains some random implementation notes. They do not	1317	This section contains some random implementation notes. They do not
324	describe guaranteed behaviour, but merely behaviour as-is implemented	1318	describe guaranteed behaviour, but merely behaviour as-is implemented
…		…
333	Only the double data type is supported for NV data types - when Perl uses	1327	Only the double data type is supported for NV data types - when Perl uses
334	long double to represent floating point values, they might not be encoded	1328	long double to represent floating point values, they might not be encoded
335	properly. Half precision types are accepted, but not encoded.	1329	properly. Half precision types are accepted, but not encoded.
336		1330
337	Strict mode and canonical mode are not implemented.	1331	Strict mode and canonical mode are not implemented.
		1332
		1333
		1334	=head1 LIMITATIONS ON PERLS WITHOUT 64-BIT INTEGER SUPPORT
		1335
		1336	On perls that were built without 64 bit integer support (these are rare
		1337	nowadays, even on 32 bit architectures, as all major Perl distributions
		1338	are built with 64 bit integer support), support for any kind of 64 bit
		1339	value in CBOR is very limited - most likely, these 64 bit values will
		1340	be truncated, corrupted, or otherwise not decoded correctly. This also
		1341	includes string, float, array and map sizes that are stored as 64 bit
		1342	integers.
338		1343
339		1344
340	=head1 THREADS	1345	=head1 THREADS
341		1346
342	This module is I<not> guaranteed to be thread safe and there are no	1347	This module is I<not> guaranteed to be thread safe and there are no
…		…
356	Please refrain from using rt.cpan.org or any other bug reporting	1361	Please refrain from using rt.cpan.org or any other bug reporting
357	service. I put the contact address into my modules for a reason.	1362	service. I put the contact address into my modules for a reason.
358		1363
359	=cut	1364	=cut
360		1365
361	our $true = do { bless \(my $dummy = 1), "CBOR::XS::Boolean" };	1366	# clumsy and slow hv_store-in-hash helper function
362	our $false = do { bless \(my $dummy = 0), "CBOR::XS::Boolean" };	1367	sub _hv_store {
363		1368	$_[0]{$_[1]} = $_[2];
364	sub true() { $true }
365	sub false() { $false }
366
367	sub is_bool($) {
368	UNIVERSAL::isa $_[0], "CBOR::XS::Boolean"
369	# or UNIVERSAL::isa $_[0], "CBOR::Literal"
370	}	1369	}
371		1370
		1371	our %FILTER = (
		1372	0 => sub { # rfc4287 datetime, utf-8
		1373	require Time::Piece;
		1374	# Time::Piece::Strptime uses the "incredibly flexible date parsing routine"
		1375	# from FreeBSD, which can't parse ISO 8601, RFC3339, RFC4287 or much of anything
		1376	# else either. Whats incredibe over standard strptime totally escapes me.
		1377	# doesn't do fractional times, either. sigh.
		1378	# In fact, it's all a lie, it uses whatever strptime it wants, and of course,
		1379	# they are all incompatible. The openbsd one simply ignores %z (but according to the
		1380	# docs, it would be much more incredibly flexible indeed. If it worked, that is.).
		1381	scalar eval {
		1382	my $s = $_[1];
		1383
		1384	$s =~ s/Z$/+00:00/;
		1385	$s =~ s/(\.[0-9]+)?([+-][0-9][0-9]):([0-9][0-9])$//
		1386	or die;
		1387
		1388	my $b = $1 - ($2 * 60 + $3) * 60; # fractional part + offset. hopefully
		1389	my $d = Time::Piece->strptime ($s, "%Y-%m-%dT%H:%M:%S");
		1390
		1391	Time::Piece::gmtime ($d->epoch + $b)
		1392	} || die "corrupted CBOR date/time string ($_[0])";
		1393	},
		1394
		1395	1 => sub { # seconds since the epoch, possibly fractional
		1396	require Time::Piece;
		1397	scalar Time::Piece::gmtime (pop)
		1398	},
		1399
		1400	2 => sub { # pos bigint
		1401	require Math::BigInt;
		1402	Math::BigInt->new ("0x" . unpack "H*", pop)
		1403	},
		1404
		1405	3 => sub { # neg bigint
		1406	require Math::BigInt;
		1407	-Math::BigInt->new ("0x" . unpack "H*", pop)
		1408	},
		1409
		1410	4 => sub { # decimal fraction, array
		1411	require Math::BigFloat;
		1412	Math::BigFloat->new ($_[1][1] . "E" . $_[1][0])
		1413	},
		1414
		1415	264 => sub { # decimal fraction with arbitrary exponent
		1416	require Math::BigFloat;
		1417	Math::BigFloat->new ($_[1][1] . "E" . $_[1][0])
		1418	},
		1419
		1420	5 => sub { # bigfloat, array
		1421	require Math::BigFloat;
		1422	scalar Math::BigFloat->new ($_[1][1]) * Math::BigFloat->new (2)->bpow ($_[1][0])
		1423	},
		1424
		1425	265 => sub { # bigfloat with arbitrary exponent
		1426	require Math::BigFloat;
		1427	scalar Math::BigFloat->new ($_[1][1]) * Math::BigFloat->new (2)->bpow ($_[1][0])
		1428	},
		1429
		1430	30 => sub { # rational number
		1431	require Math::BigRat;
		1432	Math::BigRat->new ("$_[1][0]/$_[1][1]") # separate parameters only work in recent versons
		1433	},
		1434
		1435	21 => sub { pop }, # expected conversion to base64url encoding
		1436	22 => sub { pop }, # expected conversion to base64 encoding
		1437	23 => sub { pop }, # expected conversion to base16 encoding
		1438
		1439	# 24 # embedded cbor, byte string
		1440
		1441	32 => sub {
		1442	require URI;
		1443	URI->new (pop)
		1444	},
		1445
		1446	# 33 # base64url rfc4648, utf-8
		1447	# 34 # base64 rfc46484, utf-8
		1448	# 35 # regex pcre/ecma262, utf-8
		1449	# 36 # mime message rfc2045, utf-8
		1450	);
		1451
		1452	sub default_filter {
		1453	&{ $FILTER{$_[0]} or return }
		1454	}
		1455
		1456	our %SAFE_FILTER = map { $_ => $FILTER{$_} } 0, 1, 21, 22, 23, 32;
		1457
		1458	sub safe_filter {
		1459	&{ $SAFE_FILTER{$_[0]} or return }
		1460	}
		1461
		1462	sub URI::TO_CBOR {
		1463	my $uri = $_[0]->as_string;
		1464	utf8::upgrade $uri;
		1465	tag 32, $uri
		1466	}
		1467
		1468	sub Math::BigInt::TO_CBOR {
		1469	if (-2147483648 <= $_[0] && $_[0] <= 2147483647) {
		1470	$_[0]->numify
		1471	} else {
		1472	my $hex = substr $_[0]->as_hex, 2;
		1473	$hex = "0$hex" if 1 & length $hex; # sigh
		1474	tag $_[0] >= 0 ? 2 : 3, pack "H*", $hex
		1475	}
		1476	}
		1477
		1478	sub Math::BigFloat::TO_CBOR {
		1479	my ($m, $e) = $_[0]->parts;
		1480
		1481	-9223372036854775808 <= $e && $e <= 18446744073709551615
		1482	? tag 4, [$e->numify, $m]
		1483	: tag 264, [$e, $m]
		1484	}
		1485
		1486	sub Math::BigRat::TO_CBOR {
		1487	my ($n, $d) = $_[0]->parts;
		1488
		1489	# older versions of BigRat need *1, as they not always return numbers
		1490
		1491	$d*1 == 1
		1492	? $n*1
		1493	: tag 30, [$n*1, $d*1]
		1494	}
		1495
		1496	sub Time::Piece::TO_CBOR {
		1497	tag 1, 0 + $_[0]->epoch
		1498	}
		1499
372	XSLoader::load "CBOR::XS", $VERSION;	1500	XSLoader::load "CBOR::XS", $VERSION;
373
374	package CBOR::XS::Boolean;
375
376	use overload
377	"0+" => sub { ${$_[0]} },
378	"++" => sub { $_[0] = ${$_[0]} + 1 },
379	"--" => sub { $_[0] = ${$_[0]} - 1 },
380	fallback => 1;
381
382	1;
383		1501
384	=head1 SEE ALSO	1502	=head1 SEE ALSO
385		1503
386	The L<JSON> and L<JSON::XS> modules that do similar, but human-readable,	1504	The L<JSON> and L<JSON::XS> modules that do similar, but human-readable,
387	serialisation.	1505	serialisation.
388		1506
		1507	The L<Types::Serialiser> module provides the data model for true, false
		1508	and error values.
		1509
389	=head1 AUTHOR	1510	=head1 AUTHOR
390		1511
391	Marc Lehmann <schmorp@schmorp.de>	1512	Marc Lehmann <schmorp@schmorp.de>
392	http://home.schmorp.de/	1513	http://home.schmorp.de/
393		1514
394	=cut	1515	=cut
395		1516
		1517	1
		1518

Diff Legend

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