ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/JSON-XS/XS.pm

Comparing JSON-XS/XS.pm (file contents):
Revision 1.72 by root, Sun Nov 25 19:11:07 2007 UTC vs.
Revision 1.168 by root, Thu Nov 15 20:13:03 2018 UTC

1	=head1 NAME	1	=head1 NAME
2		2
3	JSON::XS - JSON serialising/deserialising, done correctly and fast	3	JSON::XS - JSON serialising/deserialising, done correctly and fast
		4
		5	=encoding utf-8
4		6
5	JSON::XS - 正しくて高速な JSON シリアライザ/デシリアライザ	7	JSON::XS - 正しくて高速な JSON シリアライザ/デシリアライザ
6	(http://fleur.hio.jp/perldoc/mix/lib/JSON/XS.html)	8	(http://fleur.hio.jp/perldoc/mix/lib/JSON/XS.html)
7		9
8	=head1 SYNOPSIS	10	=head1 SYNOPSIS
…		…
10	use JSON::XS;	12	use JSON::XS;
11		13
12	# exported functions, they croak on error	14	# exported functions, they croak on error
13	# and expect/generate UTF-8	15	# and expect/generate UTF-8
14		16
15	$utf8_encoded_json_text = to_json $perl_hash_or_arrayref;	17	$utf8_encoded_json_text = encode_json $perl_hash_or_arrayref;
16	$perl_hash_or_arrayref = from_json $utf8_encoded_json_text;	18	$perl_hash_or_arrayref = decode_json $utf8_encoded_json_text;
17		19
18	# OO-interface	20	# OO-interface
19		21
20	$coder = JSON::XS->new->ascii->pretty->allow_nonref;	22	$coder = JSON::XS->new->ascii->pretty->allow_nonref;
21	$pretty_printed_unencoded = $coder->encode ($perl_scalar);	23	$pretty_printed_unencoded = $coder->encode ($perl_scalar);
22	$perl_scalar = $coder->decode ($unicode_json_text);	24	$perl_scalar = $coder->decode ($unicode_json_text);
23		25
		26	# Note that JSON version 2.0 and above will automatically use JSON::XS
		27	# if available, at virtually no speed overhead either, so you should
		28	# be able to just:
		29
		30	use JSON;
		31
		32	# and do the same things, except that you have a pure-perl fallback now.
		33
24	=head1 DESCRIPTION	34	=head1 DESCRIPTION
25		35
26	This module converts Perl data structures to JSON and vice versa. Its	36	This module converts Perl data structures to JSON and vice versa. Its
27	primary goal is to be I<correct> and its secondary goal is to be	37	primary goal is to be I<correct> and its secondary goal is to be
28	I<fast>. To reach the latter goal it was written in C.	38	I<fast>. To reach the latter goal it was written in C.
		39
		40	Beginning with version 2.0 of the JSON module, when both JSON and
		41	JSON::XS are installed, then JSON will fall back on JSON::XS (this can be
		42	overridden) with no overhead due to emulation (by inheriting constructor
		43	and methods). If JSON::XS is not available, it will fall back to the
		44	compatible JSON::PP module as backend, so using JSON instead of JSON::XS
		45	gives you a portable JSON API that can be fast when you need it and
		46	doesn't require a C compiler when that is a problem.
29		47
30	As this is the n-th-something JSON module on CPAN, what was the reason	48	As this is the n-th-something JSON module on CPAN, what was the reason
31	to write yet another JSON module? While it seems there are many JSON	49	to write yet another JSON module? While it seems there are many JSON
32	modules, none of them correctly handle all corner cases, and in most cases	50	modules, none of them correctly handle all corner cases, and in most cases
33	their maintainers are unresponsive, gone missing, or not listening to bug	51	their maintainers are unresponsive, gone missing, or not listening to bug
34	reports for other reasons.	52	reports for other reasons.
35		53
36	See COMPARISON, below, for a comparison to some other JSON modules.
37
38	See MAPPING, below, on how JSON::XS maps perl values to JSON values and	54	See MAPPING, below, on how JSON::XS maps perl values to JSON values and
39	vice versa.	55	vice versa.
40		56
41	=head2 FEATURES	57	=head2 FEATURES
42		58
43	=over 4	59	=over 4
44		60
45	=item * correct Unicode handling	61	=item * correct Unicode handling
46		62
47	This module knows how to handle Unicode, and even documents how and when	63	This module knows how to handle Unicode, documents how and when it does
48	it does so.	64	so, and even documents what "correct" means.
49		65
50	=item * round-trip integrity	66	=item * round-trip integrity
51		67
52	When you serialise a perl data structure using only datatypes supported	68	When you serialise a perl data structure using only data types supported
53	by JSON, the deserialised data structure is identical on the Perl level.	69	by JSON and Perl, the deserialised data structure is identical on the Perl
54	(e.g. the string "2.0" doesn't suddenly become "2" just because it looks	70	level. (e.g. the string "2.0" doesn't suddenly become "2" just because
55	like a number).	71	it looks like a number). There I<are> minor exceptions to this, read the
		72	MAPPING section below to learn about those.
56		73
57	=item * strict checking of JSON correctness	74	=item * strict checking of JSON correctness
58		75
59	There is no guessing, no generating of illegal JSON texts by default,	76	There is no guessing, no generating of illegal JSON texts by default,
60	and only JSON is accepted as input by default (the latter is a security	77	and only JSON is accepted as input by default (the latter is a security
61	feature).	78	feature).
62		79
63	=item * fast	80	=item * fast
64		81
65	Compared to other JSON modules, this module compares favourably in terms	82	Compared to other JSON modules and other serialisers such as Storable,
66	of speed, too.	83	this module usually compares favourably in terms of speed, too.
67		84
68	=item * simple to use	85	=item * simple to use
69		86
70	This module has both a simple functional interface as well as an OO	87	This module has both a simple functional interface as well as an object
71	interface.	88	oriented interface.
72		89
73	=item * reasonably versatile output formats	90	=item * reasonably versatile output formats
74		91
75	You can choose between the most compact guaranteed single-line format	92	You can choose between the most compact guaranteed-single-line format
76	possible (nice for simple line-based protocols), a pure-ascii format	93	possible (nice for simple line-based protocols), a pure-ASCII format
77	(for when your transport is not 8-bit clean, still supports the whole	94	(for when your transport is not 8-bit clean, still supports the whole
78	Unicode range), or a pretty-printed format (for when you want to read that	95	Unicode range), or a pretty-printed format (for when you want to read that
79	stuff). Or you can combine those features in whatever way you like.	96	stuff). Or you can combine those features in whatever way you like.
80		97
81	=back	98	=back
82		99
83	=cut	100	=cut
84		101
85	package JSON::XS;	102	package JSON::XS;
86		103
87	use strict;	104	use common::sense;
88		105
89	our $VERSION = '1.53';	106	our $VERSION = 3.04;
90	our @ISA = qw(Exporter);	107	our @ISA = qw(Exporter);
91		108
92	our @EXPORT = qw(to_json from_json);	109	our @EXPORT = qw(encode_json decode_json);
93		110
94	use Exporter;	111	use Exporter;
95	use XSLoader;	112	use XSLoader;
96		113
		114	use Types::Serialiser ();
		115
97	=head1 FUNCTIONAL INTERFACE	116	=head1 FUNCTIONAL INTERFACE
98		117
99	The following convenience methods are provided by this module. They are	118	The following convenience methods are provided by this module. They are
100	exported by default:	119	exported by default:
101		120
102	=over 4	121	=over 4
103		122
104	=item $json_text = to_json $perl_scalar	123	=item $json_text = encode_json $perl_scalar
105		124
106	Converts the given Perl data structure to a UTF-8 encoded, binary string	125	Converts the given Perl data structure to a UTF-8 encoded, binary string
107	(that is, the string contains octets only). Croaks on error.	126	(that is, the string contains octets only). Croaks on error.
108		127
109	This function call is functionally identical to:	128	This function call is functionally identical to:
110		129
111	$json_text = JSON::XS->new->utf8->encode ($perl_scalar)	130	$json_text = JSON::XS->new->utf8->encode ($perl_scalar)
112		131
113	except being faster.	132	Except being faster.
114		133
115	=item $perl_scalar = from_json $json_text	134	=item $perl_scalar = decode_json $json_text
116		135
117	The opposite of C<to_json>: expects an UTF-8 (binary) string and tries	136	The opposite of C<encode_json>: expects a UTF-8 (binary) string and tries
118	to parse that as an UTF-8 encoded JSON text, returning the resulting	137	to parse that as a UTF-8 encoded JSON text, returning the resulting
119	reference. Croaks on error.	138	reference. Croaks on error.
120		139
121	This function call is functionally identical to:	140	This function call is functionally identical to:
122		141
123	$perl_scalar = JSON::XS->new->utf8->decode ($json_text)	142	$perl_scalar = JSON::XS->new->utf8->decode ($json_text)
124		143
125	except being faster.	144	Except being faster.
126
127	=item $is_boolean = JSON::XS::is_bool $scalar
128
129	Returns true if the passed scalar represents either JSON::XS::true or
130	JSON::XS::false, two constants that act like C<1> and C<0>, respectively
131	and are used to represent JSON C<true> and C<false> values in Perl.
132
133	See MAPPING, below, for more information on how JSON values are mapped to
134	Perl.
135		145
136	=back	146	=back
137		147
138		148
139	=head1 A FEW NOTES ON UNICODE AND PERL	149	=head1 A FEW NOTES ON UNICODE AND PERL
…		…
148	This enables you to store Unicode characters as single characters in a	158	This enables you to store Unicode characters as single characters in a
149	Perl string - very natural.	159	Perl string - very natural.
150		160
151	=item 2. Perl does I<not> associate an encoding with your strings.	161	=item 2. Perl does I<not> associate an encoding with your strings.
152		162
153	Unless you force it to, e.g. when matching it against a regex, or printing	163	... until you force it to, e.g. when matching it against a regex, or
154	the scalar to a file, in which case Perl either interprets your string as	164	printing the scalar to a file, in which case Perl either interprets your
155	locale-encoded text, octets/binary, or as Unicode, depending on various	165	string as locale-encoded text, octets/binary, or as Unicode, depending
156	settings. In no case is an encoding stored together with your data, it is	166	on various settings. In no case is an encoding stored together with your
157	I<use> that decides encoding, not any magical metadata.	167	data, it is I<use> that decides encoding, not any magical meta data.
158		168
159	=item 3. The internal utf-8 flag has no meaning with regards to the	169	=item 3. The internal utf-8 flag has no meaning with regards to the
160	encoding of your string.	170	encoding of your string.
161		171
162	Just ignore that flag unless you debug a Perl bug, a module written in	172	Just ignore that flag unless you debug a Perl bug, a module written in
…		…
168		178
169	If you didn't know about that flag, just the better, pretend it doesn't	179	If you didn't know about that flag, just the better, pretend it doesn't
170	exist.	180	exist.
171		181
172	=item 4. A "Unicode String" is simply a string where each character can be	182	=item 4. A "Unicode String" is simply a string where each character can be
173	validly interpreted as a Unicode codepoint.	183	validly interpreted as a Unicode code point.
174		184
175	If you have UTF-8 encoded data, it is no longer a Unicode string, but a	185	If you have UTF-8 encoded data, it is no longer a Unicode string, but a
176	Unicode string encoded in UTF-8, giving you a binary string.	186	Unicode string encoded in UTF-8, giving you a binary string.
177		187
178	=item 5. A string containing "high" (> 255) character values is I<not> a UTF-8 string.	188	=item 5. A string containing "high" (> 255) character values is I<not> a UTF-8 string.
…		…
216		226
217	If C<$enable> is false, then the C<encode> method will not escape Unicode	227	If C<$enable> is false, then the C<encode> method will not escape Unicode
218	characters unless required by the JSON syntax or other flags. This results	228	characters unless required by the JSON syntax or other flags. This results
219	in a faster and more compact format.	229	in a faster and more compact format.
220		230
		231	See also the section I<ENCODING/CODESET FLAG NOTES> later in this
		232	document.
		233
221	The main use for this flag is to produce JSON texts that can be	234	The main use for this flag is to produce JSON texts that can be
222	transmitted over a 7-bit channel, as the encoded JSON texts will not	235	transmitted over a 7-bit channel, as the encoded JSON texts will not
223	contain any 8 bit characters.	236	contain any 8 bit characters.
224		237
225	JSON::XS->new->ascii (1)->encode ([chr 0x10401])	238	JSON::XS->new->ascii (1)->encode ([chr 0x10401])
…		…
236	will not be affected in any way by this flag, as C<decode> by default	249	will not be affected in any way by this flag, as C<decode> by default
237	expects Unicode, which is a strict superset of latin1.	250	expects Unicode, which is a strict superset of latin1.
238		251
239	If C<$enable> is false, then the C<encode> method will not escape Unicode	252	If C<$enable> is false, then the C<encode> method will not escape Unicode
240	characters unless required by the JSON syntax or other flags.	253	characters unless required by the JSON syntax or other flags.
		254
		255	See also the section I<ENCODING/CODESET FLAG NOTES> later in this
		256	document.
241		257
242	The main use for this flag is efficiently encoding binary data as JSON	258	The main use for this flag is efficiently encoding binary data as JSON
243	text, as most octets will not be escaped, resulting in a smaller encoded	259	text, as most octets will not be escaped, resulting in a smaller encoded
244	size. The disadvantage is that the resulting JSON text is encoded	260	size. The disadvantage is that the resulting JSON text is encoded
245	in latin1 (and must correctly be treated as such when storing and	261	in latin1 (and must correctly be treated as such when storing and
…		…
254		270
255	=item $enabled = $json->get_utf8	271	=item $enabled = $json->get_utf8
256		272
257	If C<$enable> is true (or missing), then the C<encode> method will encode	273	If C<$enable> is true (or missing), then the C<encode> method will encode
258	the JSON result into UTF-8, as required by many protocols, while the	274	the JSON result into UTF-8, as required by many protocols, while the
259	C<decode> method expects to be handled an UTF-8-encoded string. Please	275	C<decode> method expects to be handed a UTF-8-encoded string. Please
260	note that UTF-8-encoded strings do not contain any characters outside the	276	note that UTF-8-encoded strings do not contain any characters outside the
261	range C<0..255>, they are thus useful for bytewise/binary I/O. In future	277	range C<0..255>, they are thus useful for bytewise/binary I/O. In future
262	versions, enabling this option might enable autodetection of the UTF-16	278	versions, enabling this option might enable autodetection of the UTF-16
263	and UTF-32 encoding families, as described in RFC4627.	279	and UTF-32 encoding families, as described in RFC4627.
264		280
265	If C<$enable> is false, then the C<encode> method will return the JSON	281	If C<$enable> is false, then the C<encode> method will return the JSON
266	string as a (non-encoded) Unicode string, while C<decode> expects thus a	282	string as a (non-encoded) Unicode string, while C<decode> expects thus a
267	Unicode string. Any decoding or encoding (e.g. to UTF-8 or UTF-16) needs	283	Unicode string. Any decoding or encoding (e.g. to UTF-8 or UTF-16) needs
268	to be done yourself, e.g. using the Encode module.	284	to be done yourself, e.g. using the Encode module.
269		285
		286	See also the section I<ENCODING/CODESET FLAG NOTES> later in this
		287	document.
		288
270	Example, output UTF-16BE-encoded JSON:	289	Example, output UTF-16BE-encoded JSON:
271		290
272	use Encode;	291	use Encode;
273	$jsontext = encode "UTF-16BE", JSON::XS->new->encode ($object);	292	$jsontext = encode "UTF-16BE", JSON::XS->new->encode ($object);
274		293
…		…
276		295
277	use Encode;	296	use Encode;
278	$object = JSON::XS->new->decode (decode "UTF-32LE", $jsontext);	297	$object = JSON::XS->new->decode (decode "UTF-32LE", $jsontext);
279		298
280	=item $json = $json->pretty ([$enable])	299	=item $json = $json->pretty ([$enable])
281
282	=item $enabled = $json->get_pretty
283		300
284	This enables (or disables) all of the C<indent>, C<space_before> and	301	This enables (or disables) all of the C<indent>, C<space_before> and
285	C<space_after> (and in the future possibly more) flags in one call to	302	C<space_after> (and in the future possibly more) flags in one call to
286	generate the most readable (or most compact) form possible.	303	generate the most readable (or most compact) form possible.
287		304
…		…
348		365
349	=item $enabled = $json->get_relaxed	366	=item $enabled = $json->get_relaxed
350		367
351	If C<$enable> is true (or missing), then C<decode> will accept some	368	If C<$enable> is true (or missing), then C<decode> will accept some
352	extensions to normal JSON syntax (see below). C<encode> will not be	369	extensions to normal JSON syntax (see below). C<encode> will not be
353	affected in anyway. I<Be aware that this option makes you accept invalid	370	affected in any way. I<Be aware that this option makes you accept invalid
354	JSON texts as if they were valid!>. I suggest only to use this option to	371	JSON texts as if they were valid!>. I suggest only to use this option to
355	parse application-specific files written by humans (configuration files,	372	parse application-specific files written by humans (configuration files,
356	resource files etc.)	373	resource files etc.)
357		374
358	If C<$enable> is false (the default), then C<decode> will only accept	375	If C<$enable> is false (the default), then C<decode> will only accept
…		…
387	[	404	[
388	1, # this comment not allowed in JSON	405	1, # this comment not allowed in JSON
389	# neither this one...	406	# neither this one...
390	]	407	]
391		408
		409	=item * literal ASCII TAB characters in strings
		410
		411	Literal ASCII TAB characters are now allowed in strings (and treated as
		412	C<\t>).
		413
		414	[
		415	"Hello\tWorld",
		416	"Hello<TAB>World", # literal <TAB> would not normally be allowed
		417	]
		418
392	=back	419	=back
393		420
394	=item $json = $json->canonical ([$enable])	421	=item $json = $json->canonical ([$enable])
395		422
396	=item $enabled = $json->get_canonical	423	=item $enabled = $json->get_canonical
…		…
398	If C<$enable> is true (or missing), then the C<encode> method will output JSON objects	425	If C<$enable> is true (or missing), then the C<encode> method will output JSON objects
399	by sorting their keys. This is adding a comparatively high overhead.	426	by sorting their keys. This is adding a comparatively high overhead.
400		427
401	If C<$enable> is false, then the C<encode> method will output key-value	428	If C<$enable> is false, then the C<encode> method will output key-value
402	pairs in the order Perl stores them (which will likely change between runs	429	pairs in the order Perl stores them (which will likely change between runs
403	of the same script).	430	of the same script, and can change even within the same run from 5.18
		431	onwards).
404		432
405	This option is useful if you want the same data structure to be encoded as	433	This option is useful if you want the same data structure to be encoded as
406	the same JSON text (given the same overall settings). If it is disabled,	434	the same JSON text (given the same overall settings). If it is disabled,
407	the same hash might be encoded differently even if contains the same data,	435	the same hash might be encoded differently even if contains the same data,
408	as key-value pairs have no inherent ordering in Perl.	436	as key-value pairs have no inherent ordering in Perl.
409		437
410	This setting has no effect when decoding JSON texts.	438	This setting has no effect when decoding JSON texts.
411		439
		440	This setting has currently no effect on tied hashes.
		441
412	=item $json = $json->allow_nonref ([$enable])	442	=item $json = $json->allow_nonref ([$enable])
413		443
414	=item $enabled = $json->get_allow_nonref	444	=item $enabled = $json->get_allow_nonref
415		445
416	If C<$enable> is true (or missing), then the C<encode> method can convert a	446	If C<$enable> is true (or missing), then the C<encode> method can convert a
…		…
427	resulting in an invalid JSON text:	457	resulting in an invalid JSON text:
428		458
429	JSON::XS->new->allow_nonref->encode ("Hello, World!")	459	JSON::XS->new->allow_nonref->encode ("Hello, World!")
430	=> "Hello, World!"	460	=> "Hello, World!"
431		461
		462	=item $json = $json->allow_unknown ([$enable])
		463
		464	=item $enabled = $json->get_allow_unknown
		465
		466	If C<$enable> is true (or missing), then C<encode> will I<not> throw an
		467	exception when it encounters values it cannot represent in JSON (for
		468	example, filehandles) but instead will encode a JSON C<null> value. Note
		469	that blessed objects are not included here and are handled separately by
		470	c<allow_nonref>.
		471
		472	If C<$enable> is false (the default), then C<encode> will throw an
		473	exception when it encounters anything it cannot encode as JSON.
		474
		475	This option does not affect C<decode> in any way, and it is recommended to
		476	leave it off unless you know your communications partner.
		477
432	=item $json = $json->allow_blessed ([$enable])	478	=item $json = $json->allow_blessed ([$enable])
433		479
434	=item $enabled = $json->get_allow_bless	480	=item $enabled = $json->get_allow_blessed
		481
		482	See L<OBJECT SERIALISATION> for details.
435		483
436	If C<$enable> is true (or missing), then the C<encode> method will not	484	If C<$enable> is true (or missing), then the C<encode> method will not
437	barf when it encounters a blessed reference. Instead, the value of the	485	barf when it encounters a blessed reference that it cannot convert
438	B<convert_blessed> option will decide whether C<null> (C<convert_blessed>	486	otherwise. Instead, a JSON C<null> value is encoded instead of the object.
439	disabled or no C<to_json> method found) or a representation of the
440	object (C<convert_blessed> enabled and C<to_json> method found) is being
441	encoded. Has no effect on C<decode>.
442		487
443	If C<$enable> is false (the default), then C<encode> will throw an	488	If C<$enable> is false (the default), then C<encode> will throw an
444	exception when it encounters a blessed object.	489	exception when it encounters a blessed object that it cannot convert
		490	otherwise.
		491
		492	This setting has no effect on C<decode>.
445		493
446	=item $json = $json->convert_blessed ([$enable])	494	=item $json = $json->convert_blessed ([$enable])
447		495
448	=item $enabled = $json->get_convert_blessed	496	=item $enabled = $json->get_convert_blessed
		497
		498	See L<OBJECT SERIALISATION> for details.
449		499
450	If C<$enable> is true (or missing), then C<encode>, upon encountering a	500	If C<$enable> is true (or missing), then C<encode>, upon encountering a
451	blessed object, will check for the availability of the C<TO_JSON> method	501	blessed object, will check for the availability of the C<TO_JSON> method
452	on the object's class. If found, it will be called in scalar context	502	on the object's class. If found, it will be called in scalar context and
453	and the resulting scalar will be encoded instead of the object. If no	503	the resulting scalar will be encoded instead of the object.
454	C<TO_JSON> method is found, the value of C<allow_blessed> will decide what
455	to do.
456		504
457	The C<TO_JSON> method may safely call die if it wants. If C<TO_JSON>	505	The C<TO_JSON> method may safely call die if it wants. If C<TO_JSON>
458	returns other blessed objects, those will be handled in the same	506	returns other blessed objects, those will be handled in the same
459	way. C<TO_JSON> must take care of not causing an endless recursion cycle	507	way. C<TO_JSON> must take care of not causing an endless recursion cycle
460	(== crash) in this case. The name of C<TO_JSON> was chosen because other	508	(== crash) in this case. The name of C<TO_JSON> was chosen because other
461	methods called by the Perl core (== not by the user of the object) are	509	methods called by the Perl core (== not by the user of the object) are
462	usually in upper case letters and to avoid collisions with the C<to_json>	510	usually in upper case letters and to avoid collisions with any C<to_json>
463	function.	511	function or method.
464		512
465	This setting does not yet influence C<decode> in any way, but in the	513	If C<$enable> is false (the default), then C<encode> will not consider
466	future, global hooks might get installed that influence C<decode> and are	514	this type of conversion.
467	enabled by this setting.
468		515
469	If C<$enable> is false, then the C<allow_blessed> setting will decide what	516	This setting has no effect on C<decode>.
470	to do when a blessed object is found.	517
		518	=item $json = $json->allow_tags ([$enable])
		519
		520	=item $enabled = $json->get_allow_tags
		521
		522	See L<OBJECT SERIALISATION> for details.
		523
		524	If C<$enable> is true (or missing), then C<encode>, upon encountering a
		525	blessed object, will check for the availability of the C<FREEZE> method on
		526	the object's class. If found, it will be used to serialise the object into
		527	a nonstandard tagged JSON value (that JSON decoders cannot decode).
		528
		529	It also causes C<decode> to parse such tagged JSON values and deserialise
		530	them via a call to the C<THAW> method.
		531
		532	If C<$enable> is false (the default), then C<encode> will not consider
		533	this type of conversion, and tagged JSON values will cause a parse error
		534	in C<decode>, as if tags were not part of the grammar.
471		535
472	=item $json = $json->filter_json_object ([$coderef->($hashref)])	536	=item $json = $json->filter_json_object ([$coderef->($hashref)])
473		537
474	When C<$coderef> is specified, it will be called from C<decode> each	538	When C<$coderef> is specified, it will be called from C<decode> each
475	time it decodes a JSON object. The only argument is a reference to the	539	time it decodes a JSON object. The only argument is a reference to
476	newly-created hash. If the code references returns a single scalar (which	540	the newly-created hash. If the code reference returns a single scalar
477	need not be a reference), this value (i.e. a copy of that scalar to avoid	541	(which need not be a reference), this value (or rather a copy of it) is
478	aliasing) is inserted into the deserialised data structure. If it returns	542	inserted into the deserialised data structure. If it returns an empty
479	an empty list (NOTE: I<not> C<undef>, which is a valid scalar), the	543	list (NOTE: I<not> C<undef>, which is a valid scalar), the original
480	original deserialised hash will be inserted. This setting can slow down	544	deserialised hash will be inserted. This setting can slow down decoding
481	decoding considerably.	545	considerably.
482		546
483	When C<$coderef> is omitted or undefined, any existing callback will	547	When C<$coderef> is omitted or undefined, any existing callback will
484	be removed and C<decode> will not change the deserialised hash in any	548	be removed and C<decode> will not change the deserialised hash in any
485	way.	549	way.
486		550
…		…
576	=item $json = $json->max_depth ([$maximum_nesting_depth])	640	=item $json = $json->max_depth ([$maximum_nesting_depth])
577		641
578	=item $max_depth = $json->get_max_depth	642	=item $max_depth = $json->get_max_depth
579		643
580	Sets the maximum nesting level (default C<512>) accepted while encoding	644	Sets the maximum nesting level (default C<512>) accepted while encoding
581	or decoding. If the JSON text or Perl data structure has an equal or	645	or decoding. If a higher nesting level is detected in JSON text or a Perl
582	higher nesting level then this limit, then the encoder and decoder will	646	data structure, then the encoder and decoder will stop and croak at that
583	stop and croak at that point.	647	point.
584		648
585	Nesting level is defined by number of hash- or arrayrefs that the encoder	649	Nesting level is defined by number of hash- or arrayrefs that the encoder
586	needs to traverse to reach a given point or the number of C<{> or C<[>	650	needs to traverse to reach a given point or the number of C<{> or C<[>
587	characters without their matching closing parenthesis crossed to reach a	651	characters without their matching closing parenthesis crossed to reach a
588	given character in a string.	652	given character in a string.
589		653
590	Setting the maximum depth to one disallows any nesting, so that ensures	654	Setting the maximum depth to one disallows any nesting, so that ensures
591	that the object is only a single hash/object or array.	655	that the object is only a single hash/object or array.
592		656
593	The argument to C<max_depth> will be rounded up to the next highest power
594	of two. If no argument is given, the highest possible setting will be	657	If no argument is given, the highest possible setting will be used, which
595	used, which is rarely useful.	658	is rarely useful.
		659
		660	Note that nesting is implemented by recursion in C. The default value has
		661	been chosen to be as large as typical operating systems allow without
		662	crashing.
596		663
597	See SECURITY CONSIDERATIONS, below, for more info on why this is useful.	664	See SECURITY CONSIDERATIONS, below, for more info on why this is useful.
598		665
599	=item $json = $json->max_size ([$maximum_string_size])	666	=item $json = $json->max_size ([$maximum_string_size])
600		667
601	=item $max_size = $json->get_max_size	668	=item $max_size = $json->get_max_size
602		669
603	Set the maximum length a JSON text may have (in bytes) where decoding is	670	Set the maximum length a JSON text may have (in bytes) where decoding is
604	being attempted. The default is C<0>, meaning no limit. When C<decode>	671	being attempted. The default is C<0>, meaning no limit. When C<decode>
605	is called on a string longer then this number of characters it will not	672	is called on a string that is longer then this many bytes, it will not
606	attempt to decode the string but throw an exception. This setting has no	673	attempt to decode the string but throw an exception. This setting has no
607	effect on C<encode> (yet).	674	effect on C<encode> (yet).
608		675
609	The argument to C<max_size> will be rounded up to the next B<highest>	676	If no argument is given, the limit check will be deactivated (same as when
610	power of two (so may be more than requested). If no argument is given, the	677	C<0> is specified).
611	limit check will be deactivated (same as when C<0> is specified).
612		678
613	See SECURITY CONSIDERATIONS, below, for more info on why this is useful.	679	See SECURITY CONSIDERATIONS, below, for more info on why this is useful.
614		680
615	=item $json_text = $json->encode ($perl_scalar)	681	=item $json_text = $json->encode ($perl_scalar)
616		682
617	Converts the given Perl data structure (a simple scalar or a reference	683	Converts the given Perl value or data structure to its JSON
618	to a hash or array) to its JSON representation. Simple scalars will be	684	representation. Croaks on error.
619	converted into JSON string or number sequences, while references to arrays
620	become JSON arrays and references to hashes become JSON objects. Undefined
621	Perl values (e.g. C<undef>) become JSON C<null> values. Neither C<true>
622	nor C<false> values will be generated.
623		685
624	=item $perl_scalar = $json->decode ($json_text)	686	=item $perl_scalar = $json->decode ($json_text)
625		687
626	The opposite of C<encode>: expects a JSON text and tries to parse it,	688	The opposite of C<encode>: expects a JSON text and tries to parse it,
627	returning the resulting simple scalar or reference. Croaks on error.	689	returning the resulting simple scalar or reference. Croaks on error.
628
629	JSON numbers and strings become simple Perl scalars. JSON arrays become
630	Perl arrayrefs and JSON objects become Perl hashrefs. C<true> becomes
631	C<1>, C<false> becomes C<0> and C<null> becomes C<undef>.
632		690
633	=item ($perl_scalar, $characters) = $json->decode_prefix ($json_text)	691	=item ($perl_scalar, $characters) = $json->decode_prefix ($json_text)
634		692
635	This works like the C<decode> method, but instead of raising an exception	693	This works like the C<decode> method, but instead of raising an exception
636	when there is trailing garbage after the first JSON object, it will	694	when there is trailing garbage after the first JSON object, it will
637	silently stop parsing there and return the number of characters consumed	695	silently stop parsing there and return the number of characters consumed
638	so far.	696	so far.
639		697
640	This is useful if your JSON texts are not delimited by an outer protocol	698	This is useful if your JSON texts are not delimited by an outer protocol
641	(which is not the brightest thing to do in the first place) and you need
642	to know where the JSON text ends.	699	and you need to know where the JSON text ends.
643		700
644	JSON::XS->new->decode_prefix ("[1] the tail")	701	JSON::XS->new->decode_prefix ("[1] the tail")
645	=> ([], 3)	702	=> ([1], 3)
646		703
647	=back	704	=back
		705
		706
		707	=head1 INCREMENTAL PARSING
		708
		709	In some cases, there is the need for incremental parsing of JSON
		710	texts. While this module always has to keep both JSON text and resulting
		711	Perl data structure in memory at one time, it does allow you to parse a
		712	JSON stream incrementally. It does so by accumulating text until it has
		713	a full JSON object, which it then can decode. This process is similar to
		714	using C<decode_prefix> to see if a full JSON object is available, but
		715	is much more efficient (and can be implemented with a minimum of method
		716	calls).
		717
		718	JSON::XS will only attempt to parse the JSON text once it is sure it
		719	has enough text to get a decisive result, using a very simple but
		720	truly incremental parser. This means that it sometimes won't stop as
		721	early as the full parser, for example, it doesn't detect mismatched
		722	parentheses. The only thing it guarantees is that it starts decoding as
		723	soon as a syntactically valid JSON text has been seen. This means you need
		724	to set resource limits (e.g. C<max_size>) to ensure the parser will stop
		725	parsing in the presence if syntax errors.
		726
		727	The following methods implement this incremental parser.
		728
		729	=over 4
		730
		731	=item [void, scalar or list context] = $json->incr_parse ([$string])
		732
		733	This is the central parsing function. It can both append new text and
		734	extract objects from the stream accumulated so far (both of these
		735	functions are optional).
		736
		737	If C<$string> is given, then this string is appended to the already
		738	existing JSON fragment stored in the C<$json> object.
		739
		740	After that, if the function is called in void context, it will simply
		741	return without doing anything further. This can be used to add more text
		742	in as many chunks as you want.
		743
		744	If the method is called in scalar context, then it will try to extract
		745	exactly I<one> JSON object. If that is successful, it will return this
		746	object, otherwise it will return C<undef>. If there is a parse error,
		747	this method will croak just as C<decode> would do (one can then use
		748	C<incr_skip> to skip the erroneous part). This is the most common way of
		749	using the method.
		750
		751	And finally, in list context, it will try to extract as many objects
		752	from the stream as it can find and return them, or the empty list
		753	otherwise. For this to work, there must be no separators (other than
		754	whitespace) between the JSON objects or arrays, instead they must be
		755	concatenated back-to-back. If an error occurs, an exception will be
		756	raised as in the scalar context case. Note that in this case, any
		757	previously-parsed JSON texts will be lost.
		758
		759	Example: Parse some JSON arrays/objects in a given string and return
		760	them.
		761
		762	my @objs = JSON::XS->new->incr_parse ("[5][7][1,2]");
		763
		764	=item $lvalue_string = $json->incr_text
		765
		766	This method returns the currently stored JSON fragment as an lvalue, that
		767	is, you can manipulate it. This I<only> works when a preceding call to
		768	C<incr_parse> in I<scalar context> successfully returned an object. Under
		769	all other circumstances you must not call this function (I mean it.
		770	although in simple tests it might actually work, it I<will> fail under
		771	real world conditions). As a special exception, you can also call this
		772	method before having parsed anything.
		773
		774	That means you can only use this function to look at or manipulate text
		775	before or after complete JSON objects, not while the parser is in the
		776	middle of parsing a JSON object.
		777
		778	This function is useful in two cases: a) finding the trailing text after a
		779	JSON object or b) parsing multiple JSON objects separated by non-JSON text
		780	(such as commas).
		781
		782	=item $json->incr_skip
		783
		784	This will reset the state of the incremental parser and will remove
		785	the parsed text from the input buffer so far. This is useful after
		786	C<incr_parse> died, in which case the input buffer and incremental parser
		787	state is left unchanged, to skip the text parsed so far and to reset the
		788	parse state.
		789
		790	The difference to C<incr_reset> is that only text until the parse error
		791	occurred is removed.
		792
		793	=item $json->incr_reset
		794
		795	This completely resets the incremental parser, that is, after this call,
		796	it will be as if the parser had never parsed anything.
		797
		798	This is useful if you want to repeatedly parse JSON objects and want to
		799	ignore any trailing data, which means you have to reset the parser after
		800	each successful decode.
		801
		802	=back
		803
		804	=head2 LIMITATIONS
		805
		806	All options that affect decoding are supported, except
		807	C<allow_nonref>. The reason for this is that it cannot be made to work
		808	sensibly: JSON objects and arrays are self-delimited, i.e. you can
		809	concatenate them back to back and still decode them perfectly. This does
		810	not hold true for JSON numbers, however.
		811
		812	For example, is the string C<1> a single JSON number, or is it simply the
		813	start of C<12>? Or is C<12> a single JSON number, or the concatenation
		814	of C<1> and C<2>? In neither case you can tell, and this is why JSON::XS
		815	takes the conservative route and disallows this case.
		816
		817	=head2 EXAMPLES
		818
		819	Some examples will make all this clearer. First, a simple example that
		820	works similarly to C<decode_prefix>: We want to decode the JSON object at
		821	the start of a string and identify the portion after the JSON object:
		822
		823	my $text = "[1,2,3] hello";
		824
		825	my $json = new JSON::XS;
		826
		827	my $obj = $json->incr_parse ($text)
		828	or die "expected JSON object or array at beginning of string";
		829
		830	my $tail = $json->incr_text;
		831	# $tail now contains " hello"
		832
		833	Easy, isn't it?
		834
		835	Now for a more complicated example: Imagine a hypothetical protocol where
		836	you read some requests from a TCP stream, and each request is a JSON
		837	array, without any separation between them (in fact, it is often useful to
		838	use newlines as "separators", as these get interpreted as whitespace at
		839	the start of the JSON text, which makes it possible to test said protocol
		840	with C<telnet>...).
		841
		842	Here is how you'd do it (it is trivial to write this in an event-based
		843	manner):
		844
		845	my $json = new JSON::XS;
		846
		847	# read some data from the socket
		848	while (sysread $socket, my $buf, 4096) {
		849
		850	# split and decode as many requests as possible
		851	for my $request ($json->incr_parse ($buf)) {
		852	# act on the $request
		853	}
		854	}
		855
		856	Another complicated example: Assume you have a string with JSON objects
		857	or arrays, all separated by (optional) comma characters (e.g. C<[1],[2],
		858	[3]>). To parse them, we have to skip the commas between the JSON texts,
		859	and here is where the lvalue-ness of C<incr_text> comes in useful:
		860
		861	my $text = "[1],[2], [3]";
		862	my $json = new JSON::XS;
		863
		864	# void context, so no parsing done
		865	$json->incr_parse ($text);
		866
		867	# now extract as many objects as possible. note the
		868	# use of scalar context so incr_text can be called.
		869	while (my $obj = $json->incr_parse) {
		870	# do something with $obj
		871
		872	# now skip the optional comma
		873	$json->incr_text =~ s/^ \s* , //x;
		874	}
		875
		876	Now lets go for a very complex example: Assume that you have a gigantic
		877	JSON array-of-objects, many gigabytes in size, and you want to parse it,
		878	but you cannot load it into memory fully (this has actually happened in
		879	the real world :).
		880
		881	Well, you lost, you have to implement your own JSON parser. But JSON::XS
		882	can still help you: You implement a (very simple) array parser and let
		883	JSON decode the array elements, which are all full JSON objects on their
		884	own (this wouldn't work if the array elements could be JSON numbers, for
		885	example):
		886
		887	my $json = new JSON::XS;
		888
		889	# open the monster
		890	open my $fh, "<bigfile.json"
		891	or die "bigfile: $!";
		892
		893	# first parse the initial "["
		894	for (;;) {
		895	sysread $fh, my $buf, 65536
		896	or die "read error: $!";
		897	$json->incr_parse ($buf); # void context, so no parsing
		898
		899	# Exit the loop once we found and removed(!) the initial "[".
		900	# In essence, we are (ab-)using the $json object as a simple scalar
		901	# we append data to.
		902	last if $json->incr_text =~ s/^ \s* \[ //x;
		903	}
		904
		905	# now we have the skipped the initial "[", so continue
		906	# parsing all the elements.
		907	for (;;) {
		908	# in this loop we read data until we got a single JSON object
		909	for (;;) {
		910	if (my $obj = $json->incr_parse) {
		911	# do something with $obj
		912	last;
		913	}
		914
		915	# add more data
		916	sysread $fh, my $buf, 65536
		917	or die "read error: $!";
		918	$json->incr_parse ($buf); # void context, so no parsing
		919	}
		920
		921	# in this loop we read data until we either found and parsed the
		922	# separating "," between elements, or the final "]"
		923	for (;;) {
		924	# first skip whitespace
		925	$json->incr_text =~ s/^\s*//;
		926
		927	# if we find "]", we are done
		928	if ($json->incr_text =~ s/^\]//) {
		929	print "finished.\n";
		930	exit;
		931	}
		932
		933	# if we find ",", we can continue with the next element
		934	if ($json->incr_text =~ s/^,//) {
		935	last;
		936	}
		937
		938	# if we find anything else, we have a parse error!
		939	if (length $json->incr_text) {
		940	die "parse error near ", $json->incr_text;
		941	}
		942
		943	# else add more data
		944	sysread $fh, my $buf, 65536
		945	or die "read error: $!";
		946	$json->incr_parse ($buf); # void context, so no parsing
		947	}
		948
		949	This is a complex example, but most of the complexity comes from the fact
		950	that we are trying to be correct (bear with me if I am wrong, I never ran
		951	the above example :).
		952
648		953
649		954
650	=head1 MAPPING	955	=head1 MAPPING
651		956
652	This section describes how JSON::XS maps Perl values to JSON values and	957	This section describes how JSON::XS maps Perl values to JSON values and
…		…
682		987
683	A JSON number becomes either an integer, numeric (floating point) or	988	A JSON number becomes either an integer, numeric (floating point) or
684	string scalar in perl, depending on its range and any fractional parts. On	989	string scalar in perl, depending on its range and any fractional parts. On
685	the Perl level, there is no difference between those as Perl handles all	990	the Perl level, there is no difference between those as Perl handles all
686	the conversion details, but an integer may take slightly less memory and	991	the conversion details, but an integer may take slightly less memory and
687	might represent more values exactly than (floating point) numbers.	992	might represent more values exactly than floating point numbers.
688		993
689	If the number consists of digits only, JSON::XS will try to represent	994	If the number consists of digits only, JSON::XS will try to represent
690	it as an integer value. If that fails, it will try to represent it as	995	it as an integer value. If that fails, it will try to represent it as
691	a numeric (floating point) value if that is possible without loss of	996	a numeric (floating point) value if that is possible without loss of
692	precision. Otherwise it will preserve the number as a string value.	997	precision. Otherwise it will preserve the number as a string value (in
		998	which case you lose roundtripping ability, as the JSON number will be
		999	re-encoded to a JSON string).
693		1000
694	Numbers containing a fractional or exponential part will always be	1001	Numbers containing a fractional or exponential part will always be
695	represented as numeric (floating point) values, possibly at a loss of	1002	represented as numeric (floating point) values, possibly at a loss of
696	precision.	1003	precision (in which case you might lose perfect roundtripping ability, but
		1004	the JSON number will still be re-encoded as a JSON number).
697		1005
698	This might create round-tripping problems as numbers might become strings,	1006	Note that precision is not accuracy - binary floating point values cannot
699	but as Perl is typeless there is no other way to do it.	1007	represent most decimal fractions exactly, and when converting from and to
		1008	floating point, JSON::XS only guarantees precision up to but not including
		1009	the least significant bit.
700		1010
701	=item true, false	1011	=item true, false
702		1012
703	These JSON atoms become C<JSON::XS::true> and C<JSON::XS::false>,	1013	These JSON atoms become C<Types::Serialiser::true> and
704	respectively. They are overloaded to act almost exactly like the numbers	1014	C<Types::Serialiser::false>, respectively. They are overloaded to act
705	C<1> and C<0>. You can check whether a scalar is a JSON boolean by using	1015	almost exactly like the numbers C<1> and C<0>. You can check whether
706	the C<JSON::XS::is_bool> function.	1016	a scalar is a JSON boolean by using the C<Types::Serialiser::is_bool>
		1017	function (after C<use Types::Serialier>, of course).
707		1018
708	=item null	1019	=item null
709		1020
710	A JSON null atom becomes C<undef> in Perl.	1021	A JSON null atom becomes C<undef> in Perl.
		1022
		1023	=item shell-style comments (C<< # I<text> >>)
		1024
		1025	As a nonstandard extension to the JSON syntax that is enabled by the
		1026	C<relaxed> setting, shell-style comments are allowed. They can start
		1027	anywhere outside strings and go till the end of the line.
		1028
		1029	=item tagged values (C<< (I<tag>)I<value> >>).
		1030
		1031	Another nonstandard extension to the JSON syntax, enabled with the
		1032	C<allow_tags> setting, are tagged values. In this implementation, the
		1033	I<tag> must be a perl package/class name encoded as a JSON string, and the
		1034	I<value> must be a JSON array encoding optional constructor arguments.
		1035
		1036	See L<OBJECT SERIALISATION>, below, for details.
711		1037
712	=back	1038	=back
713		1039
714		1040
715	=head2 PERL -> JSON	1041	=head2 PERL -> JSON
…		…
720		1046
721	=over 4	1047	=over 4
722		1048
723	=item hash references	1049	=item hash references
724		1050
725	Perl hash references become JSON objects. As there is no inherent ordering	1051	Perl hash references become JSON objects. As there is no inherent
726	in hash keys (or JSON objects), they will usually be encoded in a	1052	ordering in hash keys (or JSON objects), they will usually be encoded
727	pseudo-random order that can change between runs of the same program but	1053	in a pseudo-random order. JSON::XS can optionally sort the hash keys
728	stays generally the same within a single run of a program. JSON::XS can	1054	(determined by the I<canonical> flag), so the same datastructure will
729	optionally sort the hash keys (determined by the I<canonical> flag), so	1055	serialise to the same JSON text (given same settings and version of
730	the same datastructure will serialise to the same JSON text (given same	1056	JSON::XS), but this incurs a runtime overhead and is only rarely useful,
731	settings and version of JSON::XS), but this incurs a runtime overhead	1057	e.g. when you want to compare some JSON text against another for equality.
732	and is only rarely useful, e.g. when you want to compare some JSON text
733	against another for equality.
734		1058
735	=item array references	1059	=item array references
736		1060
737	Perl array references become JSON arrays.	1061	Perl array references become JSON arrays.
738		1062
739	=item other references	1063	=item other references
740		1064
741	Other unblessed references are generally not allowed and will cause an	1065	Other unblessed references are generally not allowed and will cause an
742	exception to be thrown, except for references to the integers C<0> and	1066	exception to be thrown, except for references to the integers C<0> and
743	C<1>, which get turned into C<false> and C<true> atoms in JSON. You can	1067	C<1>, which get turned into C<false> and C<true> atoms in JSON.
744	also use C<JSON::XS::false> and C<JSON::XS::true> to improve readability.
745		1068
		1069	Since C<JSON::XS> uses the boolean model from L<Types::Serialiser>, you
		1070	can also C<use Types::Serialiser> and then use C<Types::Serialiser::false>
		1071	and C<Types::Serialiser::true> to improve readability.
		1072
		1073	use Types::Serialiser;
746	to_json [\0,JSON::XS::true] # yields [false,true]	1074	encode_json [\0, Types::Serialiser::true] # yields [false,true]
747		1075
748	=item JSON::XS::true, JSON::XS::false	1076	=item Types::Serialiser::true, Types::Serialiser::false
749		1077
750	These special values become JSON true and JSON false values,	1078	These special values from the L<Types::Serialiser> module become JSON true
751	respectively. You can also use C<\1> and C<\0> directly if you want.	1079	and JSON false values, respectively. You can also use C<\1> and C<\0>
		1080	directly if you want.
752		1081
753	=item blessed objects	1082	=item blessed objects
754		1083
755	Blessed objects are not allowed. JSON::XS currently tries to encode their	1084	Blessed objects are not directly representable in JSON, but C<JSON::XS>
756	underlying representation (hash- or arrayref), but this behaviour might	1085	allows various ways of handling objects. See L<OBJECT SERIALISATION>,
757	change in future versions.	1086	below, for details.
758		1087
759	=item simple scalars	1088	=item simple scalars
760		1089
761	Simple Perl scalars (any scalar that is not a reference) are the most	1090	Simple Perl scalars (any scalar that is not a reference) are the most
762	difficult objects to encode: JSON::XS will encode undefined scalars as	1091	difficult objects to encode: JSON::XS will encode undefined scalars as
763	JSON null value, scalars that have last been used in a string context	1092	JSON C<null> values, scalars that have last been used in a string context
764	before encoding as JSON strings and anything else as number value:	1093	before encoding as JSON strings, and anything else as number value:
765		1094
766	# dump as number	1095	# dump as number
767	to_json [2] # yields [2]	1096	encode_json [2] # yields [2]
768	to_json [-3.0e17] # yields [-3e+17]	1097	encode_json [-3.0e17] # yields [-3e+17]
769	my $value = 5; to_json [$value] # yields [5]	1098	my $value = 5; encode_json [$value] # yields [5]
770		1099
771	# used as string, so dump as string	1100	# used as string, so dump as string
772	print $value;	1101	print $value;
773	to_json [$value] # yields ["5"]	1102	encode_json [$value] # yields ["5"]
774		1103
775	# undef becomes null	1104	# undef becomes null
776	to_json [undef] # yields [null]	1105	encode_json [undef] # yields [null]
777		1106
778	You can force the type to be a JSON string by stringifying it:	1107	You can force the type to be a JSON string by stringifying it:
779		1108
780	my $x = 3.1; # some variable containing a number	1109	my $x = 3.1; # some variable containing a number
781	"$x"; # stringified	1110	"$x"; # stringified
…		…
787	my $x = "3"; # some variable containing a string	1116	my $x = "3"; # some variable containing a string
788	$x += 0; # numify it, ensuring it will be dumped as a number	1117	$x += 0; # numify it, ensuring it will be dumped as a number
789	$x *= 1; # same thing, the choice is yours.	1118	$x *= 1; # same thing, the choice is yours.
790		1119
791	You can not currently force the type in other, less obscure, ways. Tell me	1120	You can not currently force the type in other, less obscure, ways. Tell me
792	if you need this capability.	1121	if you need this capability (but don't forget to explain why it's needed
		1122	:).
		1123
		1124	Note that numerical precision has the same meaning as under Perl (so
		1125	binary to decimal conversion follows the same rules as in Perl, which
		1126	can differ to other languages). Also, your perl interpreter might expose
		1127	extensions to the floating point numbers of your platform, such as
		1128	infinities or NaN's - these cannot be represented in JSON, and it is an
		1129	error to pass those in.
793		1130
794	=back	1131	=back
795		1132
		1133	=head2 OBJECT SERIALISATION
796		1134
797	=head1 COMPARISON	1135	As JSON cannot directly represent Perl objects, you have to choose between
		1136	a pure JSON representation (without the ability to deserialise the object
		1137	automatically again), and a nonstandard extension to the JSON syntax,
		1138	tagged values.
798		1139
799	As already mentioned, this module was created because none of the existing	1140	=head3 SERIALISATION
800	JSON modules could be made to work correctly. First I will describe the	1141
801	problems (or pleasures) I encountered with various existing JSON modules,	1142	What happens when C<JSON::XS> encounters a Perl object depends on the
802	followed by some benchmark values. JSON::XS was designed not to suffer	1143	C<allow_blessed>, C<convert_blessed> and C<allow_tags> settings, which are
803	from any of these problems or limitations.	1144	used in this order:
804		1145
805	=over 4	1146	=over 4
806		1147
807	=item JSON 1.07	1148	=item 1. C<allow_tags> is enabled and the object has a C<FREEZE> method.
808		1149
809	Slow (but very portable, as it is written in pure Perl).	1150	In this case, C<JSON::XS> uses the L<Types::Serialiser> object
		1151	serialisation protocol to create a tagged JSON value, using a nonstandard
		1152	extension to the JSON syntax.
810		1153
811	Undocumented/buggy Unicode handling (how JSON handles Unicode values is	1154	This works by invoking the C<FREEZE> method on the object, with the first
812	undocumented. One can get far by feeding it Unicode strings and doing	1155	argument being the object to serialise, and the second argument being the
813	en-/decoding oneself, but Unicode escapes are not working properly).	1156	constant string C<JSON> to distinguish it from other serialisers.
814		1157
815	No round-tripping (strings get clobbered if they look like numbers, e.g.	1158	The C<FREEZE> method can return any number of values (i.e. zero or
816	the string C<2.0> will encode to C<2.0> instead of C<"2.0">, and that will	1159	more). These values and the paclkage/classname of the object will then be
817	decode into the number 2.	1160	encoded as a tagged JSON value in the following format:
818		1161
819	=item JSON::PC 0.01	1162	("classname")[FREEZE return values...]
820		1163
821	Very fast.	1164	e.g.:
822		1165
823	Undocumented/buggy Unicode handling.	1166	("URI")["http://www.google.com/"]
		1167	("MyDate")[2013,10,29]
		1168	("ImageData::JPEG")["Z3...VlCg=="]
824		1169
825	No round-tripping.	1170	For example, the hypothetical C<My::Object> C<FREEZE> method might use the
		1171	objects C<type> and C<id> members to encode the object:
826		1172
827	Has problems handling many Perl values (e.g. regex results and other magic	1173	sub My::Object::FREEZE {
828	values will make it croak).	1174	my ($self, $serialiser) = @_;
829		1175
830	Does not even generate valid JSON (C<{1,2}> gets converted to C<{1:2}>	1176	($self->{type}, $self->{id})
831	which is not a valid JSON text.	1177	}
832		1178
833	Unmaintained (maintainer unresponsive for many months, bugs are not	1179	=item 2. C<convert_blessed> is enabled and the object has a C<TO_JSON> method.
834	getting fixed).
835		1180
836	=item JSON::Syck 0.21	1181	In this case, the C<TO_JSON> method of the object is invoked in scalar
		1182	context. It must return a single scalar that can be directly encoded into
		1183	JSON. This scalar replaces the object in the JSON text.
837		1184
838	Very buggy (often crashes).	1185	For example, the following C<TO_JSON> method will convert all L<URI>
		1186	objects to JSON strings when serialised. The fatc that these values
		1187	originally were L<URI> objects is lost.
839		1188
840	Very inflexible (no human-readable format supported, format pretty much	1189	sub URI::TO_JSON {
841	undocumented. I need at least a format for easy reading by humans and a	1190	my ($uri) = @_;
842	single-line compact format for use in a protocol, and preferably a way to	1191	$uri->as_string
843	generate ASCII-only JSON texts).	1192	}
844		1193
845	Completely broken (and confusingly documented) Unicode handling (Unicode	1194	=item 3. C<allow_blessed> is enabled.
846	escapes are not working properly, you need to set ImplicitUnicode to
847	I<different> values on en- and decoding to get symmetric behaviour).
848		1195
849	No round-tripping (simple cases work, but this depends on whether the scalar	1196	The object will be serialised as a JSON null value.
850	value was used in a numeric context or not).
851		1197
852	Dumping hashes may skip hash values depending on iterator state.	1198	=item 4. none of the above
853		1199
854	Unmaintained (maintainer unresponsive for many months, bugs are not	1200	If none of the settings are enabled or the respective methods are missing,
855	getting fixed).	1201	C<JSON::XS> throws an exception.
856
857	Does not check input for validity (i.e. will accept non-JSON input and
858	return "something" instead of raising an exception. This is a security
859	issue: imagine two banks transferring money between each other using
860	JSON. One bank might parse a given non-JSON request and deduct money,
861	while the other might reject the transaction with a syntax error. While a
862	good protocol will at least recover, that is extra unnecessary work and
863	the transaction will still not succeed).
864
865	=item JSON::DWIW 0.04
866
867	Very fast. Very natural. Very nice.
868
869	Undocumented Unicode handling (but the best of the pack. Unicode escapes
870	still don't get parsed properly).
871
872	Very inflexible.
873
874	No round-tripping.
875
876	Does not generate valid JSON texts (key strings are often unquoted, empty keys
877	result in nothing being output)
878
879	Does not check input for validity.
880		1202
881	=back	1203	=back
882		1204
		1205	=head3 DESERIALISATION
		1206
		1207	For deserialisation there are only two cases to consider: either
		1208	nonstandard tagging was used, in which case C<allow_tags> decides,
		1209	or objects cannot be automatically be deserialised, in which
		1210	case you can use postprocessing or the C<filter_json_object> or
		1211	C<filter_json_single_key_object> callbacks to get some real objects our of
		1212	your JSON.
		1213
		1214	This section only considers the tagged value case: I a tagged JSON object
		1215	is encountered during decoding and C<allow_tags> is disabled, a parse
		1216	error will result (as if tagged values were not part of the grammar).
		1217
		1218	If C<allow_tags> is enabled, C<JSON::XS> will look up the C<THAW> method
		1219	of the package/classname used during serialisation (it will not attempt
		1220	to load the package as a Perl module). If there is no such method, the
		1221	decoding will fail with an error.
		1222
		1223	Otherwise, the C<THAW> method is invoked with the classname as first
		1224	argument, the constant string C<JSON> as second argument, and all the
		1225	values from the JSON array (the values originally returned by the
		1226	C<FREEZE> method) as remaining arguments.
		1227
		1228	The method must then return the object. While technically you can return
		1229	any Perl scalar, you might have to enable the C<enable_nonref> setting to
		1230	make that work in all cases, so better return an actual blessed reference.
		1231
		1232	As an example, let's implement a C<THAW> function that regenerates the
		1233	C<My::Object> from the C<FREEZE> example earlier:
		1234
		1235	sub My::Object::THAW {
		1236	my ($class, $serialiser, $type, $id) = @_;
		1237
		1238	$class->new (type => $type, id => $id)
		1239	}
		1240
		1241
		1242	=head1 ENCODING/CODESET FLAG NOTES
		1243
		1244	The interested reader might have seen a number of flags that signify
		1245	encodings or codesets - C<utf8>, C<latin1> and C<ascii>. There seems to be
		1246	some confusion on what these do, so here is a short comparison:
		1247
		1248	C<utf8> controls whether the JSON text created by C<encode> (and expected
		1249	by C<decode>) is UTF-8 encoded or not, while C<latin1> and C<ascii> only
		1250	control whether C<encode> escapes character values outside their respective
		1251	codeset range. Neither of these flags conflict with each other, although
		1252	some combinations make less sense than others.
		1253
		1254	Care has been taken to make all flags symmetrical with respect to
		1255	C<encode> and C<decode>, that is, texts encoded with any combination of
		1256	these flag values will be correctly decoded when the same flags are used
		1257	- in general, if you use different flag settings while encoding vs. when
		1258	decoding you likely have a bug somewhere.
		1259
		1260	Below comes a verbose discussion of these flags. Note that a "codeset" is
		1261	simply an abstract set of character-codepoint pairs, while an encoding
		1262	takes those codepoint numbers and I<encodes> them, in our case into
		1263	octets. Unicode is (among other things) a codeset, UTF-8 is an encoding,
		1264	and ISO-8859-1 (= latin 1) and ASCII are both codesets I<and> encodings at
		1265	the same time, which can be confusing.
		1266
		1267	=over 4
		1268
		1269	=item C<utf8> flag disabled
		1270
		1271	When C<utf8> is disabled (the default), then C<encode>/C<decode> generate
		1272	and expect Unicode strings, that is, characters with high ordinal Unicode
		1273	values (> 255) will be encoded as such characters, and likewise such
		1274	characters are decoded as-is, no changes to them will be done, except
		1275	"(re-)interpreting" them as Unicode codepoints or Unicode characters,
		1276	respectively (to Perl, these are the same thing in strings unless you do
		1277	funny/weird/dumb stuff).
		1278
		1279	This is useful when you want to do the encoding yourself (e.g. when you
		1280	want to have UTF-16 encoded JSON texts) or when some other layer does
		1281	the encoding for you (for example, when printing to a terminal using a
		1282	filehandle that transparently encodes to UTF-8 you certainly do NOT want
		1283	to UTF-8 encode your data first and have Perl encode it another time).
		1284
		1285	=item C<utf8> flag enabled
		1286
		1287	If the C<utf8>-flag is enabled, C<encode>/C<decode> will encode all
		1288	characters using the corresponding UTF-8 multi-byte sequence, and will
		1289	expect your input strings to be encoded as UTF-8, that is, no "character"
		1290	of the input string must have any value > 255, as UTF-8 does not allow
		1291	that.
		1292
		1293	The C<utf8> flag therefore switches between two modes: disabled means you
		1294	will get a Unicode string in Perl, enabled means you get a UTF-8 encoded
		1295	octet/binary string in Perl.
		1296
		1297	=item C<latin1> or C<ascii> flags enabled
		1298
		1299	With C<latin1> (or C<ascii>) enabled, C<encode> will escape characters
		1300	with ordinal values > 255 (> 127 with C<ascii>) and encode the remaining
		1301	characters as specified by the C<utf8> flag.
		1302
		1303	If C<utf8> is disabled, then the result is also correctly encoded in those
		1304	character sets (as both are proper subsets of Unicode, meaning that a
		1305	Unicode string with all character values < 256 is the same thing as a
		1306	ISO-8859-1 string, and a Unicode string with all character values < 128 is
		1307	the same thing as an ASCII string in Perl).
		1308
		1309	If C<utf8> is enabled, you still get a correct UTF-8-encoded string,
		1310	regardless of these flags, just some more characters will be escaped using
		1311	C<\uXXXX> then before.
		1312
		1313	Note that ISO-8859-1-I<encoded> strings are not compatible with UTF-8
		1314	encoding, while ASCII-encoded strings are. That is because the ISO-8859-1
		1315	encoding is NOT a subset of UTF-8 (despite the ISO-8859-1 I<codeset> being
		1316	a subset of Unicode), while ASCII is.
		1317
		1318	Surprisingly, C<decode> will ignore these flags and so treat all input
		1319	values as governed by the C<utf8> flag. If it is disabled, this allows you
		1320	to decode ISO-8859-1- and ASCII-encoded strings, as both strict subsets of
		1321	Unicode. If it is enabled, you can correctly decode UTF-8 encoded strings.
		1322
		1323	So neither C<latin1> nor C<ascii> are incompatible with the C<utf8> flag -
		1324	they only govern when the JSON output engine escapes a character or not.
		1325
		1326	The main use for C<latin1> is to relatively efficiently store binary data
		1327	as JSON, at the expense of breaking compatibility with most JSON decoders.
		1328
		1329	The main use for C<ascii> is to force the output to not contain characters
		1330	with values > 127, which means you can interpret the resulting string
		1331	as UTF-8, ISO-8859-1, ASCII, KOI8-R or most about any character set and
		1332	8-bit-encoding, and still get the same data structure back. This is useful
		1333	when your channel for JSON transfer is not 8-bit clean or the encoding
		1334	might be mangled in between (e.g. in mail), and works because ASCII is a
		1335	proper subset of most 8-bit and multibyte encodings in use in the world.
		1336
		1337	=back
		1338
		1339
		1340	=head2 JSON and ECMAscript
		1341
		1342	JSON syntax is based on how literals are represented in javascript (the
		1343	not-standardised predecessor of ECMAscript) which is presumably why it is
		1344	called "JavaScript Object Notation".
		1345
		1346	However, JSON is not a subset (and also not a superset of course) of
		1347	ECMAscript (the standard) or javascript (whatever browsers actually
		1348	implement).
		1349
		1350	If you want to use javascript's C<eval> function to "parse" JSON, you
		1351	might run into parse errors for valid JSON texts, or the resulting data
		1352	structure might not be queryable:
		1353
		1354	One of the problems is that U+2028 and U+2029 are valid characters inside
		1355	JSON strings, but are not allowed in ECMAscript string literals, so the
		1356	following Perl fragment will not output something that can be guaranteed
		1357	to be parsable by javascript's C<eval>:
		1358
		1359	use JSON::XS;
		1360
		1361	print encode_json [chr 0x2028];
		1362
		1363	The right fix for this is to use a proper JSON parser in your javascript
		1364	programs, and not rely on C<eval> (see for example Douglas Crockford's
		1365	F<json2.js> parser).
		1366
		1367	If this is not an option, you can, as a stop-gap measure, simply encode to
		1368	ASCII-only JSON:
		1369
		1370	use JSON::XS;
		1371
		1372	print JSON::XS->new->ascii->encode ([chr 0x2028]);
		1373
		1374	Note that this will enlarge the resulting JSON text quite a bit if you
		1375	have many non-ASCII characters. You might be tempted to run some regexes
		1376	to only escape U+2028 and U+2029, e.g.:
		1377
		1378	# DO NOT USE THIS!
		1379	my $json = JSON::XS->new->utf8->encode ([chr 0x2028]);
		1380	$json =~ s/\xe2\x80\xa8/\\u2028/g; # escape U+2028
		1381	$json =~ s/\xe2\x80\xa9/\\u2029/g; # escape U+2029
		1382	print $json;
		1383
		1384	Note that I<this is a bad idea>: the above only works for U+2028 and
		1385	U+2029 and thus only for fully ECMAscript-compliant parsers. Many existing
		1386	javascript implementations, however, have issues with other characters as
		1387	well - using C<eval> naively simply I<will> cause problems.
		1388
		1389	Another problem is that some javascript implementations reserve
		1390	some property names for their own purposes (which probably makes
		1391	them non-ECMAscript-compliant). For example, Iceweasel reserves the
		1392	C<__proto__> property name for its own purposes.
		1393
		1394	If that is a problem, you could parse try to filter the resulting JSON
		1395	output for these property strings, e.g.:
		1396
		1397	$json =~ s/"__proto__"\s*:/"__proto__renamed":/g;
		1398
		1399	This works because C<__proto__> is not valid outside of strings, so every
		1400	occurrence of C<"__proto__"\s*:> must be a string used as property name.
		1401
		1402	If you know of other incompatibilities, please let me know.
		1403
883		1404
884	=head2 JSON and YAML	1405	=head2 JSON and YAML
885		1406
886	You often hear that JSON is a subset (or a close subset) of YAML. This is,	1407	You often hear that JSON is a subset of YAML. This is, however, a mass
887	however, a mass hysteria and very far from the truth. In general, there is	1408	hysteria(*) and very far from the truth (as of the time of this writing),
888	no way to configure JSON::XS to output a data structure as valid YAML.	1409	so let me state it clearly: I<in general, there is no way to configure
		1410	JSON::XS to output a data structure as valid YAML> that works in all
		1411	cases.
889		1412
890	If you really must use JSON::XS to generate YAML, you should use this	1413	If you really must use JSON::XS to generate YAML, you should use this
891	algorithm (subject to change in future versions):	1414	algorithm (subject to change in future versions):
892		1415
893	my $to_yaml = JSON::XS->new->utf8->space_after (1);	1416	my $to_yaml = JSON::XS->new->utf8->space_after (1);
894	my $yaml = $to_yaml->encode ($ref) . "\n";	1417	my $yaml = $to_yaml->encode ($ref) . "\n";
895		1418
896	This will usually generate JSON texts that also parse as valid	1419	This will I<usually> generate JSON texts that also parse as valid
897	YAML. Please note that YAML has hardcoded limits on (simple) object key	1420	YAML. Please note that YAML has hardcoded limits on (simple) object key
898	lengths that JSON doesn't have, so you should make sure that your hash	1421	lengths that JSON doesn't have and also has different and incompatible
		1422	unicode character escape syntax, so you should make sure that your hash
899	keys are noticeably shorter than the 1024 characters YAML allows.	1423	keys are noticeably shorter than the 1024 "stream characters" YAML allows
		1424	and that you do not have characters with codepoint values outside the
		1425	Unicode BMP (basic multilingual page). YAML also does not allow C<\/>
		1426	sequences in strings (which JSON::XS does not I<currently> generate, but
		1427	other JSON generators might).
900		1428
901	There might be other incompatibilities that I am not aware of. In general	1429	There might be other incompatibilities that I am not aware of (or the YAML
		1430	specification has been changed yet again - it does so quite often). In
902	you should not try to generate YAML with a JSON generator or vice versa,	1431	general you should not try to generate YAML with a JSON generator or vice
903	or try to parse JSON with a YAML parser or vice versa: chances are high	1432	versa, or try to parse JSON with a YAML parser or vice versa: chances are
904	that you will run into severe interoperability problems.	1433	high that you will run into severe interoperability problems when you
		1434	least expect it.
		1435
		1436	=over 4
		1437
		1438	=item (*)
		1439
		1440	I have been pressured multiple times by Brian Ingerson (one of the
		1441	authors of the YAML specification) to remove this paragraph, despite him
		1442	acknowledging that the actual incompatibilities exist. As I was personally
		1443	bitten by this "JSON is YAML" lie, I refused and said I will continue to
		1444	educate people about these issues, so others do not run into the same
		1445	problem again and again. After this, Brian called me a (quote)I<complete
		1446	and worthless idiot>(unquote).
		1447
		1448	In my opinion, instead of pressuring and insulting people who actually
		1449	clarify issues with YAML and the wrong statements of some of its
		1450	proponents, I would kindly suggest reading the JSON spec (which is not
		1451	that difficult or long) and finally make YAML compatible to it, and
		1452	educating users about the changes, instead of spreading lies about the
		1453	real compatibility for many I<years> and trying to silence people who
		1454	point out that it isn't true.
		1455
		1456	Addendum/2009: the YAML 1.2 spec is still incompatible with JSON, even
		1457	though the incompatibilities have been documented (and are known to Brian)
		1458	for many years and the spec makes explicit claims that YAML is a superset
		1459	of JSON. It would be so easy to fix, but apparently, bullying people and
		1460	corrupting userdata is so much easier.
		1461
		1462	=back
905		1463
906		1464
907	=head2 SPEED	1465	=head2 SPEED
908		1466
909	It seems that JSON::XS is surprisingly fast, as shown in the following	1467	It seems that JSON::XS is surprisingly fast, as shown in the following
910	tables. They have been generated with the help of the C<eg/bench> program	1468	tables. They have been generated with the help of the C<eg/bench> program
911	in the JSON::XS distribution, to make it easy to compare on your own	1469	in the JSON::XS distribution, to make it easy to compare on your own
912	system.	1470	system.
913		1471
914	First comes a comparison between various modules using a very short	1472	First comes a comparison between various modules using
915	single-line JSON string:	1473	a very short single-line JSON string (also available at
		1474	L<http://dist.schmorp.de/misc/json/short.json>).
916		1475
917	{"method": "handleMessage", "params": ["user1", "we were just talking"], \	1476	{"method": "handleMessage", "params": ["user1",
918	"id": null, "array":[1,11,234,-5,1e5,1e7, true, false]}	1477	"we were just talking"], "id": null, "array":[1,11,234,-5,1e5,1e7,
		1478	1, 0]}
919		1479
920	It shows the number of encodes/decodes per second (JSON::XS uses	1480	It shows the number of encodes/decodes per second (JSON::XS uses
921	the functional interface, while JSON::XS/2 uses the OO interface	1481	the functional interface, while JSON::XS/2 uses the OO interface
922	with pretty-printing and hashkey sorting enabled, JSON::XS/3 enables	1482	with pretty-printing and hashkey sorting enabled, JSON::XS/3 enables
923	shrink). Higher is better:	1483	shrink. JSON::DWIW/DS uses the deserialise function, while JSON::DWIW::FJ
		1484	uses the from_json method). Higher is better:
924		1485
925	module | encode | decode |	1486	module | encode | decode |
926	-----------|------------|------------|	1487	--------------|------------|------------|
927	JSON 1.x | 4990.842 | 4088.813 |	1488	JSON::DWIW/DS | 86302.551 | 102300.098 |
928	JSON::DWIW | 51653.990 | 71575.154 |	1489	JSON::DWIW/FJ | 86302.551 | 75983.768 |
929	JSON::PC | 65948.176 | 74631.744 |	1490	JSON::PP | 15827.562 | 6638.658 |
930	JSON::PP | 8931.652 | 3817.168 |	1491	JSON::Syck | 63358.066 | 47662.545 |
931	JSON::Syck | 24877.248 | 27776.848 |	1492	JSON::XS | 511500.488 | 511500.488 |
932	JSON::XS | 388361.481 | 227951.304 |	1493	JSON::XS/2 | 291271.111 | 388361.481 |
933	JSON::XS/2 | 227951.304 | 218453.333 |	1494	JSON::XS/3 | 361577.931 | 361577.931 |
934	JSON::XS/3 | 338250.323 | 218453.333 |	1495	Storable | 66788.280 | 265462.278 |
935	Storable | 16500.016 | 135300.129 |
936	-----------+------------+------------+	1496	--------------+------------+------------+
937		1497
938	That is, JSON::XS is about five times faster than JSON::DWIW on encoding,	1498	That is, JSON::XS is almost six times faster than JSON::DWIW on encoding,
939	about three times faster on decoding, and over forty times faster	1499	about five times faster on decoding, and over thirty to seventy times
940	than JSON, even with pretty-printing and key sorting. It also compares	1500	faster than JSON's pure perl implementation. It also compares favourably
941	favourably to Storable for small amounts of data.	1501	to Storable for small amounts of data.
942		1502
943	Using a longer test string (roughly 18KB, generated from Yahoo! Locals	1503	Using a longer test string (roughly 18KB, generated from Yahoo! Locals
944	search API (http://nanoref.com/yahooapis/mgPdGg):	1504	search API (L<http://dist.schmorp.de/misc/json/long.json>).
945		1505
946	module | encode | decode |	1506	module | encode | decode |
947	-----------|------------|------------|	1507	--------------|------------|------------|
948	JSON 1.x | 55.260 | 34.971 |	1508	JSON::DWIW/DS | 1647.927 | 2673.916 |
949	JSON::DWIW | 825.228 | 1082.513 |	1509	JSON::DWIW/FJ | 1630.249 | 2596.128 |
950	JSON::PC | 3571.444 | 2394.829 |
951	JSON::PP | 210.987 | 32.574 |	1510	JSON::PP | 400.640 | 62.311 |
952	JSON::Syck | 552.551 | 787.544 |	1511	JSON::Syck | 1481.040 | 1524.869 |
953	JSON::XS | 5780.463 | 4854.519 |	1512	JSON::XS | 20661.596 | 9541.183 |
954	JSON::XS/2 | 3869.998 | 4798.975 |	1513	JSON::XS/2 | 10683.403 | 9416.938 |
955	JSON::XS/3 | 5862.880 | 4798.975 |	1514	JSON::XS/3 | 20661.596 | 9400.054 |
956	Storable | 4445.002 | 5235.027 |	1515	Storable | 19765.806 | 10000.725 |
957	-----------+------------+------------+	1516	--------------+------------+------------+
958		1517
959	Again, JSON::XS leads by far (except for Storable which non-surprisingly	1518	Again, JSON::XS leads by far (except for Storable which non-surprisingly
960	decodes faster).	1519	decodes a bit faster).
961		1520
962	On large strings containing lots of high Unicode characters, some modules	1521	On large strings containing lots of high Unicode characters, some modules
963	(such as JSON::PC) seem to decode faster than JSON::XS, but the result	1522	(such as JSON::PC) seem to decode faster than JSON::XS, but the result
964	will be broken due to missing (or wrong) Unicode handling. Others refuse	1523	will be broken due to missing (or wrong) Unicode handling. Others refuse
965	to decode or encode properly, so it was impossible to prepare a fair	1524	to decode or encode properly, so it was impossible to prepare a fair
…		…
986		1545
987	Third, JSON::XS recurses using the C stack when decoding objects and	1546	Third, JSON::XS recurses using the C stack when decoding objects and
988	arrays. The C stack is a limited resource: for instance, on my amd64	1547	arrays. The C stack is a limited resource: for instance, on my amd64
989	machine with 8MB of stack size I can decode around 180k nested arrays but	1548	machine with 8MB of stack size I can decode around 180k nested arrays but
990	only 14k nested JSON objects (due to perl itself recursing deeply on croak	1549	only 14k nested JSON objects (due to perl itself recursing deeply on croak
991	to free the temporary). If that is exceeded, the program crashes. to be	1550	to free the temporary). If that is exceeded, the program crashes. To be
992	conservative, the default nesting limit is set to 512. If your process	1551	conservative, the default nesting limit is set to 512. If your process
993	has a smaller stack, you should adjust this setting accordingly with the	1552	has a smaller stack, you should adjust this setting accordingly with the
994	C<max_depth> method.	1553	C<max_depth> method.
995		1554
996	And last but least, something else could bomb you that I forgot to think	1555	Something else could bomb you, too, that I forgot to think of. In that
997	of. In that case, you get to keep the pieces. I am always open for hints,	1556	case, you get to keep the pieces. I am always open for hints, though...
998	though...	1557
		1558	Also keep in mind that JSON::XS might leak contents of your Perl data
		1559	structures in its error messages, so when you serialise sensitive
		1560	information you might want to make sure that exceptions thrown by JSON::XS
		1561	will not end up in front of untrusted eyes.
999		1562
1000	If you are using JSON::XS to return packets to consumption	1563	If you are using JSON::XS to return packets to consumption
1001	by JavaScript scripts in a browser you should have a look at	1564	by JavaScript scripts in a browser you should have a look at
1002	L<http://jpsykes.com/47/practical-csrf-and-json-security> to see whether	1565	L<http://blog.archive.jpsykes.com/47/practical-csrf-and-json-security/> to
1003	you are vulnerable to some common attack vectors (which really are browser	1566	see whether you are vulnerable to some common attack vectors (which really
1004	design bugs, but it is still you who will have to deal with it, as major	1567	are browser design bugs, but it is still you who will have to deal with
1005	browser developers care only for features, not about doing security	1568	it, as major browser developers care only for features, not about getting
1006	right).	1569	security right).
1007		1570
1008		1571
		1572	=head1 "OLD" VS. "NEW" JSON (RFC 4627 VS. RFC 7159)
		1573
		1574	TL;DR: Due to security concerns, JSON::XS will not allow scalar data in
		1575	JSON texts by default - you need to create your own JSON::XS object and
		1576	enable C<allow_nonref>:
		1577
		1578
		1579	my $json = JSON::XS->new->allow_nonref;
		1580
		1581	$text = $json->encode ($data);
		1582	$data = $json->decode ($text);
		1583
		1584	The long version: JSON being an important and supposedly stable format,
		1585	the IETF standardised it as RFC 4627 in 2006. Unfortunately, the inventor
		1586	of JSON, Dougles Crockford, unilaterally changed the definition of JSON in
		1587	javascript. Rather than create a fork, the IETF decided to standardise the
		1588	new syntax (apparently, so Iw as told, without finding it very amusing).
		1589
		1590	The biggest difference between thed original JSON and the new JSON is that
		1591	the new JSON supports scalars (anything other than arrays and objects) at
		1592	the toplevel of a JSON text. While this is strictly backwards compatible
		1593	to older versions, it breaks a number of protocols that relied on sending
		1594	JSON back-to-back, and is a minor security concern.
		1595
		1596	For example, imagine you have two banks communicating, and on one side,
		1597	trhe JSON coder gets upgraded. Two messages, such as C<10> and C<1000>
		1598	might then be confused to mean C<101000>, something that couldn't happen
		1599	in the original JSON, because niether of these messages would be valid
		1600	JSON.
		1601
		1602	If one side accepts these messages, then an upgrade in the coder on either
		1603	side could result in this becoming exploitable.
		1604
		1605	This module has always allowed these messages as an optional extension, by
		1606	default disabled. The security concerns are the reason why the default is
		1607	still disabled, but future versions might/will likely upgrade to the newer
		1608	RFC as default format, so you are advised to check your implementation
		1609	and/or override the default with C<< ->allow_nonref (0) >> to ensure that
		1610	future versions are safe.
		1611
		1612
		1613	=head1 INTEROPERABILITY WITH OTHER MODULES
		1614
		1615	C<JSON::XS> uses the L<Types::Serialiser> module to provide boolean
		1616	constants. That means that the JSON true and false values will be
		1617	comaptible to true and false values of other modules that do the same,
		1618	such as L<JSON::PP> and L<CBOR::XS>.
		1619
		1620
		1621	=head1 INTEROPERABILITY WITH OTHER JSON DECODERS
		1622
		1623	As long as you only serialise data that can be directly expressed in JSON,
		1624	C<JSON::XS> is incapable of generating invalid JSON output (modulo bugs,
		1625	but C<JSON::XS> has found more bugs in the official JSON testsuite (1)
		1626	than the official JSON testsuite has found in C<JSON::XS> (0)).
		1627
		1628	When you have trouble decoding JSON generated by this module using other
		1629	decoders, then it is very likely that you have an encoding mismatch or the
		1630	other decoder is broken.
		1631
		1632	When decoding, C<JSON::XS> is strict by default and will likely catch all
		1633	errors. There are currently two settings that change this: C<relaxed>
		1634	makes C<JSON::XS> accept (but not generate) some non-standard extensions,
		1635	and C<allow_tags> will allow you to encode and decode Perl objects, at the
		1636	cost of not outputting valid JSON anymore.
		1637
		1638	=head2 TAGGED VALUE SYNTAX AND STANDARD JSON EN/DECODERS
		1639
		1640	When you use C<allow_tags> to use the extended (and also nonstandard and
		1641	invalid) JSON syntax for serialised objects, and you still want to decode
		1642	the generated When you want to serialise objects, you can run a regex
		1643	to replace the tagged syntax by standard JSON arrays (it only works for
		1644	"normal" package names without comma, newlines or single colons). First,
		1645	the readable Perl version:
		1646
		1647	# if your FREEZE methods return no values, you need this replace first:
		1648	$json =~ s/\( \s* (" (?: [^\\":,]+|\\.|::)* ") \s* \) \s* \[\s*\]/[$1]/gx;
		1649
		1650	# this works for non-empty constructor arg lists:
		1651	$json =~ s/\( \s* (" (?: [^\\":,]+|\\.|::)* ") \s* \) \s* \[/[$1,/gx;
		1652
		1653	And here is a less readable version that is easy to adapt to other
		1654	languages:
		1655
		1656	$json =~ s/\(\s*("([^\\":,]+|\\.|::)*")\s*\)\s*\[/[$1,/g;
		1657
		1658	Here is an ECMAScript version (same regex):
		1659
		1660	json = json.replace (/\(\s*("([^\\":,]+|\\.|::)*")\s*\)\s*\[/g, "[$1,");
		1661
		1662	Since this syntax converts to standard JSON arrays, it might be hard to
		1663	distinguish serialised objects from normal arrays. You can prepend a
		1664	"magic number" as first array element to reduce chances of a collision:
		1665
		1666	$json =~ s/\(\s*("([^\\":,]+|\\.|::)*")\s*\)\s*\[/["XU1peReLzT4ggEllLanBYq4G9VzliwKF",$1,/g;
		1667
		1668	And after decoding the JSON text, you could walk the data
		1669	structure looking for arrays with a first element of
		1670	C<XU1peReLzT4ggEllLanBYq4G9VzliwKF>.
		1671
		1672	The same approach can be used to create the tagged format with another
		1673	encoder. First, you create an array with the magic string as first member,
		1674	the classname as second, and constructor arguments last, encode it as part
		1675	of your JSON structure, and then:
		1676
		1677	$json =~ s/\[\s*"XU1peReLzT4ggEllLanBYq4G9VzliwKF"\s*,\s*("([^\\":,]+|\\.|::)*")\s*,/($1)[/g;
		1678
		1679	Again, this has some limitations - the magic string must not be encoded
		1680	with character escapes, and the constructor arguments must be non-empty.
		1681
		1682
		1683	=head1 RFC7159
		1684
		1685	Since this module was written, Google has written a new JSON RFC, RFC 7159
		1686	(and RFC7158). Unfortunately, this RFC breaks compatibility with both the
		1687	original JSON specification on www.json.org and RFC4627.
		1688
		1689	As far as I can see, you can get partial compatibility when parsing by
		1690	using C<< ->allow_nonref >>. However, consider the security implications
		1691	of doing so.
		1692
		1693	I haven't decided yet when to break compatibility with RFC4627 by default
		1694	(and potentially leave applications insecure) and change the default to
		1695	follow RFC7159, but application authors are well advised to call C<<
		1696	->allow_nonref(0) >> even if this is the current default, if they cannot
		1697	handle non-reference values, in preparation for the day when the default
		1698	will change.
		1699
		1700
1009	=head1 THREADS	1701	=head1 (I-)THREADS
1010		1702
1011	This module is I<not> guaranteed to be thread safe and there are no	1703	This module is I<not> guaranteed to be ithread (or MULTIPLICITY-) safe
1012	plans to change this until Perl gets thread support (as opposed to the	1704	and there are no plans to change this. Note that perl's builtin so-called
1013	horribly slow so-called "threads" which are simply slow and bloated	1705	threads/ithreads are officially deprecated and should not be used.
1014	process simulations - use fork, its I<much> faster, cheaper, better).
1015		1706
1016	(It might actually work, but you have been warned).	1707
		1708	=head1 THE PERILS OF SETLOCALE
		1709
		1710	Sometimes people avoid the Perl locale support and directly call the
		1711	system's setlocale function with C<LC_ALL>.
		1712
		1713	This breaks both perl and modules such as JSON::XS, as stringification of
		1714	numbers no longer works correctly (e.g. C<$x = 0.1; print "$x"+1> might
		1715	print C<1>, and JSON::XS might output illegal JSON as JSON::XS relies on
		1716	perl to stringify numbers).
		1717
		1718	The solution is simple: don't call C<setlocale>, or use it for only those
		1719	categories you need, such as C<LC_MESSAGES> or C<LC_CTYPE>.
		1720
		1721	If you need C<LC_NUMERIC>, you should enable it only around the code that
		1722	actually needs it (avoiding stringification of numbers), and restore it
		1723	afterwards.
1017		1724
1018		1725
1019	=head1 BUGS	1726	=head1 BUGS
1020		1727
1021	While the goal of this module is to be correct, that unfortunately does	1728	While the goal of this module is to be correct, that unfortunately does
1022	not mean its bug-free, only that I think its design is bug-free. It is	1729	not mean it's bug-free, only that I think its design is bug-free. If you
1023	still relatively early in its development. If you keep reporting bugs they	1730	keep reporting bugs they will be fixed swiftly, though.
1024	will be fixed swiftly, though.
1025		1731
1026	Please refrain from using rt.cpan.org or any other bug reporting	1732	Please refrain from using rt.cpan.org or any other bug reporting
1027	service. I put the contact address into my modules for a reason.	1733	service. I put the contact address into my modules for a reason.
1028		1734
1029	=cut	1735	=cut
1030		1736
1031	our $true = do { bless \(my $dummy = 1), "JSON::XS::Boolean" };	1737	BEGIN {
1032	our $false = do { bless \(my $dummy = 0), "JSON::XS::Boolean" };	1738	*true = \$Types::Serialiser::true;
		1739	*true = \&Types::Serialiser::true;
		1740	*false = \$Types::Serialiser::false;
		1741	*false = \&Types::Serialiser::false;
		1742	*is_bool = \&Types::Serialiser::is_bool;
1033		1743
1034	sub true() { $true }	1744	*JSON::XS::Boolean:: = *Types::Serialiser::Boolean::;
1035	sub false() { $false }
1036
1037	sub is_bool($) {
1038	UNIVERSAL::isa $_[0], "JSON::XS::Boolean"
1039	# or UNIVERSAL::isa $_[0], "JSON::Literal"
1040	}	1745	}
1041		1746
1042	XSLoader::load "JSON::XS", $VERSION;	1747	XSLoader::load "JSON::XS", $VERSION;
1043		1748
1044	package JSON::XS::Boolean;	1749	=head1 SEE ALSO
1045		1750
1046	use overload	1751	The F<json_xs> command line utility for quick experiments.
1047	"0+" => sub { ${$_[0]} },
1048	"++" => sub { $_[0] = ${$_[0]} + 1 },
1049	"--" => sub { $_[0] = ${$_[0]} - 1 },
1050	fallback => 1;
1051
1052	1;
1053		1752
1054	=head1 AUTHOR	1753	=head1 AUTHOR
1055		1754
1056	Marc Lehmann <schmorp@schmorp.de>	1755	Marc Lehmann <schmorp@schmorp.de>
1057	http://home.schmorp.de/	1756	http://home.schmorp.de/
1058		1757
1059	=cut	1758	=cut
1060		1759
		1760	1
		1761

Diff Legend

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