1 |
=head1 NAME |
2 |
|
3 |
JSON::XS - JSON serialising/deserialising, done correctly and fast |
4 |
|
5 |
=head1 SYNOPSIS |
6 |
|
7 |
use JSON::XS; |
8 |
|
9 |
# exported functions, they croak on error |
10 |
# and expect/generate UTF-8 |
11 |
|
12 |
$utf8_encoded_json_text = to_json $perl_hash_or_arrayref; |
13 |
$perl_hash_or_arrayref = from_json $utf8_encoded_json_text; |
14 |
|
15 |
# OO-interface |
16 |
|
17 |
$coder = JSON::XS->new->ascii->pretty->allow_nonref; |
18 |
$pretty_printed_unencoded = $coder->encode ($perl_scalar); |
19 |
$perl_scalar = $coder->decode ($unicode_json_text); |
20 |
|
21 |
=head1 DESCRIPTION |
22 |
|
23 |
This module converts Perl data structures to JSON and vice versa. Its |
24 |
primary goal is to be I<correct> and its secondary goal is to be |
25 |
I<fast>. To reach the latter goal it was written in C. |
26 |
|
27 |
As this is the n-th-something JSON module on CPAN, what was the reason |
28 |
to write yet another JSON module? While it seems there are many JSON |
29 |
modules, none of them correctly handle all corner cases, and in most cases |
30 |
their maintainers are unresponsive, gone missing, or not listening to bug |
31 |
reports for other reasons. |
32 |
|
33 |
See COMPARISON, below, for a comparison to some other JSON modules. |
34 |
|
35 |
See MAPPING, below, on how JSON::XS maps perl values to JSON values and |
36 |
vice versa. |
37 |
|
38 |
=head2 FEATURES |
39 |
|
40 |
=over 4 |
41 |
|
42 |
=item * correct unicode handling |
43 |
|
44 |
This module knows how to handle Unicode, and even documents how and when |
45 |
it does so. |
46 |
|
47 |
=item * round-trip integrity |
48 |
|
49 |
When you serialise a perl data structure using only datatypes supported |
50 |
by JSON, the deserialised data structure is identical on the Perl level. |
51 |
(e.g. the string "2.0" doesn't suddenly become "2" just because it looks |
52 |
like a number). |
53 |
|
54 |
=item * strict checking of JSON correctness |
55 |
|
56 |
There is no guessing, no generating of illegal JSON texts by default, |
57 |
and only JSON is accepted as input by default (the latter is a security |
58 |
feature). |
59 |
|
60 |
=item * fast |
61 |
|
62 |
Compared to other JSON modules, this module compares favourably in terms |
63 |
of speed, too. |
64 |
|
65 |
=item * simple to use |
66 |
|
67 |
This module has both a simple functional interface as well as an OO |
68 |
interface. |
69 |
|
70 |
=item * reasonably versatile output formats |
71 |
|
72 |
You can choose between the most compact guarenteed single-line format |
73 |
possible (nice for simple line-based protocols), a pure-ascii format |
74 |
(for when your transport is not 8-bit clean, still supports the whole |
75 |
unicode range), or a pretty-printed format (for when you want to read that |
76 |
stuff). Or you can combine those features in whatever way you like. |
77 |
|
78 |
=back |
79 |
|
80 |
=cut |
81 |
|
82 |
package JSON::XS; |
83 |
|
84 |
use strict; |
85 |
|
86 |
our $VERSION = '1.5'; |
87 |
our @ISA = qw(Exporter); |
88 |
|
89 |
our @EXPORT = qw(to_json from_json); |
90 |
|
91 |
use Exporter; |
92 |
use XSLoader; |
93 |
|
94 |
=head1 FUNCTIONAL INTERFACE |
95 |
|
96 |
The following convinience methods are provided by this module. They are |
97 |
exported by default: |
98 |
|
99 |
=over 4 |
100 |
|
101 |
=item $json_text = to_json $perl_scalar |
102 |
|
103 |
Converts the given Perl data structure (a simple scalar or a reference to |
104 |
a hash or array) to a UTF-8 encoded, binary string (that is, the string contains |
105 |
octets only). Croaks on error. |
106 |
|
107 |
This function call is functionally identical to: |
108 |
|
109 |
$json_text = JSON::XS->new->utf8->encode ($perl_scalar) |
110 |
|
111 |
except being faster. |
112 |
|
113 |
=item $perl_scalar = from_json $json_text |
114 |
|
115 |
The opposite of C<to_json>: expects an UTF-8 (binary) string and tries to |
116 |
parse that as an UTF-8 encoded JSON text, returning the resulting simple |
117 |
scalar or reference. Croaks on error. |
118 |
|
119 |
This function call is functionally identical to: |
120 |
|
121 |
$perl_scalar = JSON::XS->new->utf8->decode ($json_text) |
122 |
|
123 |
except being faster. |
124 |
|
125 |
=item $is_boolean = JSON::XS::is_bool $scalar |
126 |
|
127 |
Returns true if the passed scalar represents either JSON::XS::true or |
128 |
JSON::XS::false, two constants that act like C<1> and C<0>, respectively |
129 |
and are used to represent JSON C<true> and C<false> values in Perl. |
130 |
|
131 |
See MAPPING, below, for more information on how JSON values are mapped to |
132 |
Perl. |
133 |
|
134 |
=back |
135 |
|
136 |
|
137 |
=head1 OBJECT-ORIENTED INTERFACE |
138 |
|
139 |
The object oriented interface lets you configure your own encoding or |
140 |
decoding style, within the limits of supported formats. |
141 |
|
142 |
=over 4 |
143 |
|
144 |
=item $json = new JSON::XS |
145 |
|
146 |
Creates a new JSON::XS object that can be used to de/encode JSON |
147 |
strings. All boolean flags described below are by default I<disabled>. |
148 |
|
149 |
The mutators for flags all return the JSON object again and thus calls can |
150 |
be chained: |
151 |
|
152 |
my $json = JSON::XS->new->utf8->space_after->encode ({a => [1,2]}) |
153 |
=> {"a": [1, 2]} |
154 |
|
155 |
=item $json = $json->ascii ([$enable]) |
156 |
|
157 |
If C<$enable> is true (or missing), then the C<encode> method will not |
158 |
generate characters outside the code range C<0..127> (which is ASCII). Any |
159 |
unicode characters outside that range will be escaped using either a |
160 |
single \uXXXX (BMP characters) or a double \uHHHH\uLLLLL escape sequence, |
161 |
as per RFC4627. The resulting encoded JSON text can be treated as a native |
162 |
unicode string, an ascii-encoded, latin1-encoded or UTF-8 encoded string, |
163 |
or any other superset of ASCII. |
164 |
|
165 |
If C<$enable> is false, then the C<encode> method will not escape Unicode |
166 |
characters unless required by the JSON syntax or other flags. This results |
167 |
in a faster and more compact format. |
168 |
|
169 |
The main use for this flag is to produce JSON texts that can be |
170 |
transmitted over a 7-bit channel, as the encoded JSON texts will not |
171 |
contain any 8 bit characters. |
172 |
|
173 |
JSON::XS->new->ascii (1)->encode ([chr 0x10401]) |
174 |
=> ["\ud801\udc01"] |
175 |
|
176 |
=item $json = $json->latin1 ([$enable]) |
177 |
|
178 |
If C<$enable> is true (or missing), then the C<encode> method will encode |
179 |
the resulting JSON text as latin1 (or iso-8859-1), escaping any characters |
180 |
outside the code range C<0..255>. The resulting string can be treated as a |
181 |
latin1-encoded JSON text or a native unicode string. The C<decode> method |
182 |
will not be affected in any way by this flag, as C<decode> by default |
183 |
expects unicode, which is a strict superset of latin1. |
184 |
|
185 |
If C<$enable> is false, then the C<encode> method will not escape Unicode |
186 |
characters unless required by the JSON syntax or other flags. |
187 |
|
188 |
The main use for this flag is efficiently encoding binary data as JSON |
189 |
text, as most octets will not be escaped, resulting in a smaller encoded |
190 |
size. The disadvantage is that the resulting JSON text is encoded |
191 |
in latin1 (and must correctly be treated as such when storing and |
192 |
transfering), a rare encoding for JSON. It is therefore most useful when |
193 |
you want to store data structures known to contain binary data efficiently |
194 |
in files or databases, not when talking to other JSON encoders/decoders. |
195 |
|
196 |
JSON::XS->new->latin1->encode (["\x{89}\x{abc}"] |
197 |
=> ["\x{89}\\u0abc"] # (perl syntax, U+abc escaped, U+89 not) |
198 |
|
199 |
=item $json = $json->utf8 ([$enable]) |
200 |
|
201 |
If C<$enable> is true (or missing), then the C<encode> method will encode |
202 |
the JSON result into UTF-8, as required by many protocols, while the |
203 |
C<decode> method expects to be handled an UTF-8-encoded string. Please |
204 |
note that UTF-8-encoded strings do not contain any characters outside the |
205 |
range C<0..255>, they are thus useful for bytewise/binary I/O. In future |
206 |
versions, enabling this option might enable autodetection of the UTF-16 |
207 |
and UTF-32 encoding families, as described in RFC4627. |
208 |
|
209 |
If C<$enable> is false, then the C<encode> method will return the JSON |
210 |
string as a (non-encoded) unicode string, while C<decode> expects thus a |
211 |
unicode string. Any decoding or encoding (e.g. to UTF-8 or UTF-16) needs |
212 |
to be done yourself, e.g. using the Encode module. |
213 |
|
214 |
Example, output UTF-16BE-encoded JSON: |
215 |
|
216 |
use Encode; |
217 |
$jsontext = encode "UTF-16BE", JSON::XS->new->encode ($object); |
218 |
|
219 |
Example, decode UTF-32LE-encoded JSON: |
220 |
|
221 |
use Encode; |
222 |
$object = JSON::XS->new->decode (decode "UTF-32LE", $jsontext); |
223 |
|
224 |
=item $json = $json->pretty ([$enable]) |
225 |
|
226 |
This enables (or disables) all of the C<indent>, C<space_before> and |
227 |
C<space_after> (and in the future possibly more) flags in one call to |
228 |
generate the most readable (or most compact) form possible. |
229 |
|
230 |
Example, pretty-print some simple structure: |
231 |
|
232 |
my $json = JSON::XS->new->pretty(1)->encode ({a => [1,2]}) |
233 |
=> |
234 |
{ |
235 |
"a" : [ |
236 |
1, |
237 |
2 |
238 |
] |
239 |
} |
240 |
|
241 |
=item $json = $json->indent ([$enable]) |
242 |
|
243 |
If C<$enable> is true (or missing), then the C<encode> method will use a multiline |
244 |
format as output, putting every array member or object/hash key-value pair |
245 |
into its own line, identing them properly. |
246 |
|
247 |
If C<$enable> is false, no newlines or indenting will be produced, and the |
248 |
resulting JSON text is guarenteed not to contain any C<newlines>. |
249 |
|
250 |
This setting has no effect when decoding JSON texts. |
251 |
|
252 |
=item $json = $json->space_before ([$enable]) |
253 |
|
254 |
If C<$enable> is true (or missing), then the C<encode> method will add an extra |
255 |
optional space before the C<:> separating keys from values in JSON objects. |
256 |
|
257 |
If C<$enable> is false, then the C<encode> method will not add any extra |
258 |
space at those places. |
259 |
|
260 |
This setting has no effect when decoding JSON texts. You will also |
261 |
most likely combine this setting with C<space_after>. |
262 |
|
263 |
Example, space_before enabled, space_after and indent disabled: |
264 |
|
265 |
{"key" :"value"} |
266 |
|
267 |
=item $json = $json->space_after ([$enable]) |
268 |
|
269 |
If C<$enable> is true (or missing), then the C<encode> method will add an extra |
270 |
optional space after the C<:> separating keys from values in JSON objects |
271 |
and extra whitespace after the C<,> separating key-value pairs and array |
272 |
members. |
273 |
|
274 |
If C<$enable> is false, then the C<encode> method will not add any extra |
275 |
space at those places. |
276 |
|
277 |
This setting has no effect when decoding JSON texts. |
278 |
|
279 |
Example, space_before and indent disabled, space_after enabled: |
280 |
|
281 |
{"key": "value"} |
282 |
|
283 |
=item $json = $json->relaxed ([$enable]) |
284 |
|
285 |
If C<$enable> is true (or missing), then C<decode> will accept some |
286 |
extensions to normal JSON syntax (see below). C<encode> will not be |
287 |
affected in anyway. I<Be aware that this option makes you accept invalid |
288 |
JSON texts as if they were valid!>. I suggest only to use this option to |
289 |
parse application-specific files written by humans (configuration files, |
290 |
resource files etc.) |
291 |
|
292 |
If C<$enable> is false (the default), then C<decode> will only accept |
293 |
valid JSON texts. |
294 |
|
295 |
Currently accepted extensions are: |
296 |
|
297 |
=over 4 |
298 |
|
299 |
=item * list items can have an end-comma |
300 |
|
301 |
JSON I<separates> array elements and key-value pairs with commas. This |
302 |
can be annoying if you write JSON texts manually and want to be able to |
303 |
quickly append elements, so this extension accepts comma at the end of |
304 |
such items not just between them: |
305 |
|
306 |
[ |
307 |
1, |
308 |
2, <- this comma not normally allowed |
309 |
] |
310 |
{ |
311 |
"k1": "v1", |
312 |
"k2": "v2", <- this comma not normally allowed |
313 |
} |
314 |
|
315 |
=back |
316 |
|
317 |
=item $json = $json->canonical ([$enable]) |
318 |
|
319 |
If C<$enable> is true (or missing), then the C<encode> method will output JSON objects |
320 |
by sorting their keys. This is adding a comparatively high overhead. |
321 |
|
322 |
If C<$enable> is false, then the C<encode> method will output key-value |
323 |
pairs in the order Perl stores them (which will likely change between runs |
324 |
of the same script). |
325 |
|
326 |
This option is useful if you want the same data structure to be encoded as |
327 |
the same JSON text (given the same overall settings). If it is disabled, |
328 |
the same hash migh be encoded differently even if contains the same data, |
329 |
as key-value pairs have no inherent ordering in Perl. |
330 |
|
331 |
This setting has no effect when decoding JSON texts. |
332 |
|
333 |
=item $json = $json->allow_nonref ([$enable]) |
334 |
|
335 |
If C<$enable> is true (or missing), then the C<encode> method can convert a |
336 |
non-reference into its corresponding string, number or null JSON value, |
337 |
which is an extension to RFC4627. Likewise, C<decode> will accept those JSON |
338 |
values instead of croaking. |
339 |
|
340 |
If C<$enable> is false, then the C<encode> method will croak if it isn't |
341 |
passed an arrayref or hashref, as JSON texts must either be an object |
342 |
or array. Likewise, C<decode> will croak if given something that is not a |
343 |
JSON object or array. |
344 |
|
345 |
Example, encode a Perl scalar as JSON value with enabled C<allow_nonref>, |
346 |
resulting in an invalid JSON text: |
347 |
|
348 |
JSON::XS->new->allow_nonref->encode ("Hello, World!") |
349 |
=> "Hello, World!" |
350 |
|
351 |
=item $json = $json->allow_blessed ([$enable]) |
352 |
|
353 |
If C<$enable> is true (or missing), then the C<encode> method will not |
354 |
barf when it encounters a blessed reference. Instead, the value of the |
355 |
B<convert_blessed> option will decide wether C<null> (C<convert_blessed> |
356 |
disabled or no C<to_json> method found) or a representation of the |
357 |
object (C<convert_blessed> enabled and C<to_json> method found) is being |
358 |
encoded. Has no effect on C<decode>. |
359 |
|
360 |
If C<$enable> is false (the default), then C<encode> will throw an |
361 |
exception when it encounters a blessed object. |
362 |
|
363 |
=item $json = $json->convert_blessed ([$enable]) |
364 |
|
365 |
If C<$enable> is true (or missing), then C<encode>, upon encountering a |
366 |
blessed object, will check for the availability of the C<TO_JSON> method |
367 |
on the object's class. If found, it will be called in scalar context |
368 |
and the resulting scalar will be encoded instead of the object. If no |
369 |
C<TO_JSON> method is found, the value of C<allow_blessed> will decide what |
370 |
to do. |
371 |
|
372 |
The C<TO_JSON> method may safely call die if it wants. If C<TO_JSON> |
373 |
returns other blessed objects, those will be handled in the same |
374 |
way. C<TO_JSON> must take care of not causing an endless recursion cycle |
375 |
(== crash) in this case. The name of C<TO_JSON> was chosen because other |
376 |
methods called by the Perl core (== not by the user of the object) are |
377 |
usually in upper case letters and to avoid collisions with the C<to_json> |
378 |
function. |
379 |
|
380 |
This setting does not yet influence C<decode> in any way, but in the |
381 |
future, global hooks might get installed that influence C<decode> and are |
382 |
enabled by this setting. |
383 |
|
384 |
If C<$enable> is false, then the C<allow_blessed> setting will decide what |
385 |
to do when a blessed object is found. |
386 |
|
387 |
=item $json = $json->filter_json_object ([$coderef->($hashref)]) |
388 |
|
389 |
When C<$coderef> is specified, it will be called from C<decode> each |
390 |
time it decodes a JSON object. The only argument is a reference to the |
391 |
newly-created hash. If the code references returns a single scalar (which |
392 |
need not be a reference), this value (i.e. a copy of that scalar to avoid |
393 |
aliasing) is inserted into the deserialised data structure. If it returns |
394 |
an empty list (NOTE: I<not> C<undef>, which is a valid scalar), the |
395 |
original deserialised hash will be inserted. This setting can slow down |
396 |
decoding considerably. |
397 |
|
398 |
When C<$coderef> is omitted or undefined, any existing callback will |
399 |
be removed and C<decode> will not change the deserialised hash in any |
400 |
way. |
401 |
|
402 |
Example, convert all JSON objects into the integer 5: |
403 |
|
404 |
my $js = JSON::XS->new->filter_json_object (sub { 5 }); |
405 |
# returns [5] |
406 |
$js->decode ('[{}]') |
407 |
# throw an exception because allow_nonref is not enabled |
408 |
# so a lone 5 is not allowed. |
409 |
$js->decode ('{"a":1, "b":2}'); |
410 |
|
411 |
=item $json = $json->filter_json_single_key_object ($key [=> $coderef->($value)]) |
412 |
|
413 |
Works remotely similar to C<filter_json_object>, but is only called for |
414 |
JSON objects having a single key named C<$key>. |
415 |
|
416 |
This C<$coderef> is called before the one specified via |
417 |
C<filter_json_object>, if any. It gets passed the single value in the JSON |
418 |
object. If it returns a single value, it will be inserted into the data |
419 |
structure. If it returns nothing (not even C<undef> but the empty list), |
420 |
the callback from C<filter_json_object> will be called next, as if no |
421 |
single-key callback were specified. |
422 |
|
423 |
If C<$coderef> is omitted or undefined, the corresponding callback will be |
424 |
disabled. There can only ever be one callback for a given key. |
425 |
|
426 |
As this callback gets called less often then the C<filter_json_object> |
427 |
one, decoding speed will not usually suffer as much. Therefore, single-key |
428 |
objects make excellent targets to serialise Perl objects into, especially |
429 |
as single-key JSON objects are as close to the type-tagged value concept |
430 |
as JSON gets (its basically an ID/VALUE tuple). Of course, JSON does not |
431 |
support this in any way, so you need to make sure your data never looks |
432 |
like a serialised Perl hash. |
433 |
|
434 |
Typical names for the single object key are C<__class_whatever__>, or |
435 |
C<$__dollars_are_rarely_used__$> or C<}ugly_brace_placement>, or even |
436 |
things like C<__class_md5sum(classname)__>, to reduce the risk of clashing |
437 |
with real hashes. |
438 |
|
439 |
Example, decode JSON objects of the form C<< { "__widget__" => <id> } >> |
440 |
into the corresponding C<< $WIDGET{<id>} >> object: |
441 |
|
442 |
# return whatever is in $WIDGET{5}: |
443 |
JSON::XS |
444 |
->new |
445 |
->filter_json_single_key_object (__widget__ => sub { |
446 |
$WIDGET{ $_[0] } |
447 |
}) |
448 |
->decode ('{"__widget__": 5') |
449 |
|
450 |
# this can be used with a TO_JSON method in some "widget" class |
451 |
# for serialisation to json: |
452 |
sub WidgetBase::TO_JSON { |
453 |
my ($self) = @_; |
454 |
|
455 |
unless ($self->{id}) { |
456 |
$self->{id} = ..get..some..id..; |
457 |
$WIDGET{$self->{id}} = $self; |
458 |
} |
459 |
|
460 |
{ __widget__ => $self->{id} } |
461 |
} |
462 |
|
463 |
=item $json = $json->shrink ([$enable]) |
464 |
|
465 |
Perl usually over-allocates memory a bit when allocating space for |
466 |
strings. This flag optionally resizes strings generated by either |
467 |
C<encode> or C<decode> to their minimum size possible. This can save |
468 |
memory when your JSON texts are either very very long or you have many |
469 |
short strings. It will also try to downgrade any strings to octet-form |
470 |
if possible: perl stores strings internally either in an encoding called |
471 |
UTF-X or in octet-form. The latter cannot store everything but uses less |
472 |
space in general (and some buggy Perl or C code might even rely on that |
473 |
internal representation being used). |
474 |
|
475 |
The actual definition of what shrink does might change in future versions, |
476 |
but it will always try to save space at the expense of time. |
477 |
|
478 |
If C<$enable> is true (or missing), the string returned by C<encode> will |
479 |
be shrunk-to-fit, while all strings generated by C<decode> will also be |
480 |
shrunk-to-fit. |
481 |
|
482 |
If C<$enable> is false, then the normal perl allocation algorithms are used. |
483 |
If you work with your data, then this is likely to be faster. |
484 |
|
485 |
In the future, this setting might control other things, such as converting |
486 |
strings that look like integers or floats into integers or floats |
487 |
internally (there is no difference on the Perl level), saving space. |
488 |
|
489 |
=item $json = $json->max_depth ([$maximum_nesting_depth]) |
490 |
|
491 |
Sets the maximum nesting level (default C<512>) accepted while encoding |
492 |
or decoding. If the JSON text or Perl data structure has an equal or |
493 |
higher nesting level then this limit, then the encoder and decoder will |
494 |
stop and croak at that point. |
495 |
|
496 |
Nesting level is defined by number of hash- or arrayrefs that the encoder |
497 |
needs to traverse to reach a given point or the number of C<{> or C<[> |
498 |
characters without their matching closing parenthesis crossed to reach a |
499 |
given character in a string. |
500 |
|
501 |
Setting the maximum depth to one disallows any nesting, so that ensures |
502 |
that the object is only a single hash/object or array. |
503 |
|
504 |
The argument to C<max_depth> will be rounded up to the next highest power |
505 |
of two. If no argument is given, the highest possible setting will be |
506 |
used, which is rarely useful. |
507 |
|
508 |
See SECURITY CONSIDERATIONS, below, for more info on why this is useful. |
509 |
|
510 |
=item $json = $json->max_size ([$maximum_string_size]) |
511 |
|
512 |
Set the maximum length a JSON text may have (in bytes) where decoding is |
513 |
being attempted. The default is C<0>, meaning no limit. When C<decode> |
514 |
is called on a string longer then this number of characters it will not |
515 |
attempt to decode the string but throw an exception. This setting has no |
516 |
effect on C<encode> (yet). |
517 |
|
518 |
The argument to C<max_size> will be rounded up to the next B<highest> |
519 |
power of two (so may be more than requested). If no argument is given, the |
520 |
limit check will be deactivated (same as when C<0> is specified). |
521 |
|
522 |
See SECURITY CONSIDERATIONS, below, for more info on why this is useful. |
523 |
|
524 |
=item $json_text = $json->encode ($perl_scalar) |
525 |
|
526 |
Converts the given Perl data structure (a simple scalar or a reference |
527 |
to a hash or array) to its JSON representation. Simple scalars will be |
528 |
converted into JSON string or number sequences, while references to arrays |
529 |
become JSON arrays and references to hashes become JSON objects. Undefined |
530 |
Perl values (e.g. C<undef>) become JSON C<null> values. Neither C<true> |
531 |
nor C<false> values will be generated. |
532 |
|
533 |
=item $perl_scalar = $json->decode ($json_text) |
534 |
|
535 |
The opposite of C<encode>: expects a JSON text and tries to parse it, |
536 |
returning the resulting simple scalar or reference. Croaks on error. |
537 |
|
538 |
JSON numbers and strings become simple Perl scalars. JSON arrays become |
539 |
Perl arrayrefs and JSON objects become Perl hashrefs. C<true> becomes |
540 |
C<1>, C<false> becomes C<0> and C<null> becomes C<undef>. |
541 |
|
542 |
=item ($perl_scalar, $characters) = $json->decode_prefix ($json_text) |
543 |
|
544 |
This works like the C<decode> method, but instead of raising an exception |
545 |
when there is trailing garbage after the first JSON object, it will |
546 |
silently stop parsing there and return the number of characters consumed |
547 |
so far. |
548 |
|
549 |
This is useful if your JSON texts are not delimited by an outer protocol |
550 |
(which is not the brightest thing to do in the first place) and you need |
551 |
to know where the JSON text ends. |
552 |
|
553 |
JSON::XS->new->decode_prefix ("[1] the tail") |
554 |
=> ([], 3) |
555 |
|
556 |
=back |
557 |
|
558 |
|
559 |
=head1 MAPPING |
560 |
|
561 |
This section describes how JSON::XS maps Perl values to JSON values and |
562 |
vice versa. These mappings are designed to "do the right thing" in most |
563 |
circumstances automatically, preserving round-tripping characteristics |
564 |
(what you put in comes out as something equivalent). |
565 |
|
566 |
For the more enlightened: note that in the following descriptions, |
567 |
lowercase I<perl> refers to the Perl interpreter, while uppcercase I<Perl> |
568 |
refers to the abstract Perl language itself. |
569 |
|
570 |
|
571 |
=head2 JSON -> PERL |
572 |
|
573 |
=over 4 |
574 |
|
575 |
=item object |
576 |
|
577 |
A JSON object becomes a reference to a hash in Perl. No ordering of object |
578 |
keys is preserved (JSON does not preserver object key ordering itself). |
579 |
|
580 |
=item array |
581 |
|
582 |
A JSON array becomes a reference to an array in Perl. |
583 |
|
584 |
=item string |
585 |
|
586 |
A JSON string becomes a string scalar in Perl - Unicode codepoints in JSON |
587 |
are represented by the same codepoints in the Perl string, so no manual |
588 |
decoding is necessary. |
589 |
|
590 |
=item number |
591 |
|
592 |
A JSON number becomes either an integer, numeric (floating point) or |
593 |
string scalar in perl, depending on its range and any fractional parts. On |
594 |
the Perl level, there is no difference between those as Perl handles all |
595 |
the conversion details, but an integer may take slightly less memory and |
596 |
might represent more values exactly than (floating point) numbers. |
597 |
|
598 |
If the number consists of digits only, JSON::XS will try to represent |
599 |
it as an integer value. If that fails, it will try to represent it as |
600 |
a numeric (floating point) value if that is possible without loss of |
601 |
precision. Otherwise it will preserve the number as a string value. |
602 |
|
603 |
Numbers containing a fractional or exponential part will always be |
604 |
represented as numeric (floating point) values, possibly at a loss of |
605 |
precision. |
606 |
|
607 |
This might create round-tripping problems as numbers might become strings, |
608 |
but as Perl is typeless there is no other way to do it. |
609 |
|
610 |
=item true, false |
611 |
|
612 |
These JSON atoms become C<JSON::XS::true> and C<JSON::XS::false>, |
613 |
respectively. They are overloaded to act almost exactly like the numbers |
614 |
C<1> and C<0>. You can check wether a scalar is a JSON boolean by using |
615 |
the C<JSON::XS::is_bool> function. |
616 |
|
617 |
=item null |
618 |
|
619 |
A JSON null atom becomes C<undef> in Perl. |
620 |
|
621 |
=back |
622 |
|
623 |
|
624 |
=head2 PERL -> JSON |
625 |
|
626 |
The mapping from Perl to JSON is slightly more difficult, as Perl is a |
627 |
truly typeless language, so we can only guess which JSON type is meant by |
628 |
a Perl value. |
629 |
|
630 |
=over 4 |
631 |
|
632 |
=item hash references |
633 |
|
634 |
Perl hash references become JSON objects. As there is no inherent ordering |
635 |
in hash keys (or JSON objects), they will usually be encoded in a |
636 |
pseudo-random order that can change between runs of the same program but |
637 |
stays generally the same within a single run of a program. JSON::XS can |
638 |
optionally sort the hash keys (determined by the I<canonical> flag), so |
639 |
the same datastructure will serialise to the same JSON text (given same |
640 |
settings and version of JSON::XS), but this incurs a runtime overhead |
641 |
and is only rarely useful, e.g. when you want to compare some JSON text |
642 |
against another for equality. |
643 |
|
644 |
=item array references |
645 |
|
646 |
Perl array references become JSON arrays. |
647 |
|
648 |
=item other references |
649 |
|
650 |
Other unblessed references are generally not allowed and will cause an |
651 |
exception to be thrown, except for references to the integers C<0> and |
652 |
C<1>, which get turned into C<false> and C<true> atoms in JSON. You can |
653 |
also use C<JSON::XS::false> and C<JSON::XS::true> to improve readability. |
654 |
|
655 |
to_json [\0,JSON::XS::true] # yields [false,true] |
656 |
|
657 |
=item JSON::XS::true, JSON::XS::false |
658 |
|
659 |
These special values become JSON true and JSON false values, |
660 |
respectively. You cna alos use C<\1> and C<\0> directly if you want. |
661 |
|
662 |
=item blessed objects |
663 |
|
664 |
Blessed objects are not allowed. JSON::XS currently tries to encode their |
665 |
underlying representation (hash- or arrayref), but this behaviour might |
666 |
change in future versions. |
667 |
|
668 |
=item simple scalars |
669 |
|
670 |
Simple Perl scalars (any scalar that is not a reference) are the most |
671 |
difficult objects to encode: JSON::XS will encode undefined scalars as |
672 |
JSON null value, scalars that have last been used in a string context |
673 |
before encoding as JSON strings and anything else as number value: |
674 |
|
675 |
# dump as number |
676 |
to_json [2] # yields [2] |
677 |
to_json [-3.0e17] # yields [-3e+17] |
678 |
my $value = 5; to_json [$value] # yields [5] |
679 |
|
680 |
# used as string, so dump as string |
681 |
print $value; |
682 |
to_json [$value] # yields ["5"] |
683 |
|
684 |
# undef becomes null |
685 |
to_json [undef] # yields [null] |
686 |
|
687 |
You can force the type to be a string by stringifying it: |
688 |
|
689 |
my $x = 3.1; # some variable containing a number |
690 |
"$x"; # stringified |
691 |
$x .= ""; # another, more awkward way to stringify |
692 |
print $x; # perl does it for you, too, quite often |
693 |
|
694 |
You can force the type to be a number by numifying it: |
695 |
|
696 |
my $x = "3"; # some variable containing a string |
697 |
$x += 0; # numify it, ensuring it will be dumped as a number |
698 |
$x *= 1; # same thing, the choise is yours. |
699 |
|
700 |
You can not currently output JSON booleans or force the type in other, |
701 |
less obscure, ways. Tell me if you need this capability. |
702 |
|
703 |
=back |
704 |
|
705 |
|
706 |
=head1 COMPARISON |
707 |
|
708 |
As already mentioned, this module was created because none of the existing |
709 |
JSON modules could be made to work correctly. First I will describe the |
710 |
problems (or pleasures) I encountered with various existing JSON modules, |
711 |
followed by some benchmark values. JSON::XS was designed not to suffer |
712 |
from any of these problems or limitations. |
713 |
|
714 |
=over 4 |
715 |
|
716 |
=item JSON 1.07 |
717 |
|
718 |
Slow (but very portable, as it is written in pure Perl). |
719 |
|
720 |
Undocumented/buggy Unicode handling (how JSON handles unicode values is |
721 |
undocumented. One can get far by feeding it unicode strings and doing |
722 |
en-/decoding oneself, but unicode escapes are not working properly). |
723 |
|
724 |
No roundtripping (strings get clobbered if they look like numbers, e.g. |
725 |
the string C<2.0> will encode to C<2.0> instead of C<"2.0">, and that will |
726 |
decode into the number 2. |
727 |
|
728 |
=item JSON::PC 0.01 |
729 |
|
730 |
Very fast. |
731 |
|
732 |
Undocumented/buggy Unicode handling. |
733 |
|
734 |
No roundtripping. |
735 |
|
736 |
Has problems handling many Perl values (e.g. regex results and other magic |
737 |
values will make it croak). |
738 |
|
739 |
Does not even generate valid JSON (C<{1,2}> gets converted to C<{1:2}> |
740 |
which is not a valid JSON text. |
741 |
|
742 |
Unmaintained (maintainer unresponsive for many months, bugs are not |
743 |
getting fixed). |
744 |
|
745 |
=item JSON::Syck 0.21 |
746 |
|
747 |
Very buggy (often crashes). |
748 |
|
749 |
Very inflexible (no human-readable format supported, format pretty much |
750 |
undocumented. I need at least a format for easy reading by humans and a |
751 |
single-line compact format for use in a protocol, and preferably a way to |
752 |
generate ASCII-only JSON texts). |
753 |
|
754 |
Completely broken (and confusingly documented) Unicode handling (unicode |
755 |
escapes are not working properly, you need to set ImplicitUnicode to |
756 |
I<different> values on en- and decoding to get symmetric behaviour). |
757 |
|
758 |
No roundtripping (simple cases work, but this depends on wether the scalar |
759 |
value was used in a numeric context or not). |
760 |
|
761 |
Dumping hashes may skip hash values depending on iterator state. |
762 |
|
763 |
Unmaintained (maintainer unresponsive for many months, bugs are not |
764 |
getting fixed). |
765 |
|
766 |
Does not check input for validity (i.e. will accept non-JSON input and |
767 |
return "something" instead of raising an exception. This is a security |
768 |
issue: imagine two banks transfering money between each other using |
769 |
JSON. One bank might parse a given non-JSON request and deduct money, |
770 |
while the other might reject the transaction with a syntax error. While a |
771 |
good protocol will at least recover, that is extra unnecessary work and |
772 |
the transaction will still not succeed). |
773 |
|
774 |
=item JSON::DWIW 0.04 |
775 |
|
776 |
Very fast. Very natural. Very nice. |
777 |
|
778 |
Undocumented unicode handling (but the best of the pack. Unicode escapes |
779 |
still don't get parsed properly). |
780 |
|
781 |
Very inflexible. |
782 |
|
783 |
No roundtripping. |
784 |
|
785 |
Does not generate valid JSON texts (key strings are often unquoted, empty keys |
786 |
result in nothing being output) |
787 |
|
788 |
Does not check input for validity. |
789 |
|
790 |
=back |
791 |
|
792 |
|
793 |
=head2 JSON and YAML |
794 |
|
795 |
You often hear that JSON is a subset (or a close subset) of YAML. This is, |
796 |
however, a mass hysteria and very far from the truth. In general, there is |
797 |
no way to configure JSON::XS to output a data structure as valid YAML. |
798 |
|
799 |
If you really must use JSON::XS to generate YAML, you should use this |
800 |
algorithm (subject to change in future versions): |
801 |
|
802 |
my $to_yaml = JSON::XS->new->utf8->space_after (1); |
803 |
my $yaml = $to_yaml->encode ($ref) . "\n"; |
804 |
|
805 |
This will usually generate JSON texts that also parse as valid |
806 |
YAML. Please note that YAML has hardcoded limits on (simple) object key |
807 |
lengths that JSON doesn't have, so you should make sure that your hash |
808 |
keys are noticably shorter than the 1024 characters YAML allows. |
809 |
|
810 |
There might be other incompatibilities that I am not aware of. In general |
811 |
you should not try to generate YAML with a JSON generator or vice versa, |
812 |
or try to parse JSON with a YAML parser or vice versa: chances are high |
813 |
that you will run into severe interoperability problems. |
814 |
|
815 |
|
816 |
=head2 SPEED |
817 |
|
818 |
It seems that JSON::XS is surprisingly fast, as shown in the following |
819 |
tables. They have been generated with the help of the C<eg/bench> program |
820 |
in the JSON::XS distribution, to make it easy to compare on your own |
821 |
system. |
822 |
|
823 |
First comes a comparison between various modules using a very short |
824 |
single-line JSON string: |
825 |
|
826 |
{"method": "handleMessage", "params": ["user1", "we were just talking"], \ |
827 |
"id": null, "array":[1,11,234,-5,1e5,1e7, true, false]} |
828 |
|
829 |
It shows the number of encodes/decodes per second (JSON::XS uses |
830 |
the functional interface, while JSON::XS/2 uses the OO interface |
831 |
with pretty-printing and hashkey sorting enabled, JSON::XS/3 enables |
832 |
shrink). Higher is better: |
833 |
|
834 |
Storable | 15779.925 | 14169.946 | |
835 |
-----------+------------+------------+ |
836 |
module | encode | decode | |
837 |
-----------|------------|------------| |
838 |
JSON | 4990.842 | 4088.813 | |
839 |
JSON::DWIW | 51653.990 | 71575.154 | |
840 |
JSON::PC | 65948.176 | 74631.744 | |
841 |
JSON::PP | 8931.652 | 3817.168 | |
842 |
JSON::Syck | 24877.248 | 27776.848 | |
843 |
JSON::XS | 388361.481 | 227951.304 | |
844 |
JSON::XS/2 | 227951.304 | 218453.333 | |
845 |
JSON::XS/3 | 338250.323 | 218453.333 | |
846 |
Storable | 16500.016 | 135300.129 | |
847 |
-----------+------------+------------+ |
848 |
|
849 |
That is, JSON::XS is about five times faster than JSON::DWIW on encoding, |
850 |
about three times faster on decoding, and over fourty times faster |
851 |
than JSON, even with pretty-printing and key sorting. It also compares |
852 |
favourably to Storable for small amounts of data. |
853 |
|
854 |
Using a longer test string (roughly 18KB, generated from Yahoo! Locals |
855 |
search API (http://nanoref.com/yahooapis/mgPdGg): |
856 |
|
857 |
module | encode | decode | |
858 |
-----------|------------|------------| |
859 |
JSON | 55.260 | 34.971 | |
860 |
JSON::DWIW | 825.228 | 1082.513 | |
861 |
JSON::PC | 3571.444 | 2394.829 | |
862 |
JSON::PP | 210.987 | 32.574 | |
863 |
JSON::Syck | 552.551 | 787.544 | |
864 |
JSON::XS | 5780.463 | 4854.519 | |
865 |
JSON::XS/2 | 3869.998 | 4798.975 | |
866 |
JSON::XS/3 | 5862.880 | 4798.975 | |
867 |
Storable | 4445.002 | 5235.027 | |
868 |
-----------+------------+------------+ |
869 |
|
870 |
Again, JSON::XS leads by far (except for Storable which non-surprisingly |
871 |
decodes faster). |
872 |
|
873 |
On large strings containing lots of high unicode characters, some modules |
874 |
(such as JSON::PC) seem to decode faster than JSON::XS, but the result |
875 |
will be broken due to missing (or wrong) unicode handling. Others refuse |
876 |
to decode or encode properly, so it was impossible to prepare a fair |
877 |
comparison table for that case. |
878 |
|
879 |
|
880 |
=head1 SECURITY CONSIDERATIONS |
881 |
|
882 |
When you are using JSON in a protocol, talking to untrusted potentially |
883 |
hostile creatures requires relatively few measures. |
884 |
|
885 |
First of all, your JSON decoder should be secure, that is, should not have |
886 |
any buffer overflows. Obviously, this module should ensure that and I am |
887 |
trying hard on making that true, but you never know. |
888 |
|
889 |
Second, you need to avoid resource-starving attacks. That means you should |
890 |
limit the size of JSON texts you accept, or make sure then when your |
891 |
resources run out, thats just fine (e.g. by using a separate process that |
892 |
can crash safely). The size of a JSON text in octets or characters is |
893 |
usually a good indication of the size of the resources required to decode |
894 |
it into a Perl structure. While JSON::XS can check the size of the JSON |
895 |
text, it might be too late when you already have it in memory, so you |
896 |
might want to check the size before you accept the string. |
897 |
|
898 |
Third, JSON::XS recurses using the C stack when decoding objects and |
899 |
arrays. The C stack is a limited resource: for instance, on my amd64 |
900 |
machine with 8MB of stack size I can decode around 180k nested arrays but |
901 |
only 14k nested JSON objects (due to perl itself recursing deeply on croak |
902 |
to free the temporary). If that is exceeded, the program crashes. to be |
903 |
conservative, the default nesting limit is set to 512. If your process |
904 |
has a smaller stack, you should adjust this setting accordingly with the |
905 |
C<max_depth> method. |
906 |
|
907 |
And last but least, something else could bomb you that I forgot to think |
908 |
of. In that case, you get to keep the pieces. I am always open for hints, |
909 |
though... |
910 |
|
911 |
If you are using JSON::XS to return packets to consumption |
912 |
by javascript scripts in a browser you should have a look at |
913 |
L<http://jpsykes.com/47/practical-csrf-and-json-security> to see wether |
914 |
you are vulnerable to some common attack vectors (which really are browser |
915 |
design bugs, but it is still you who will have to deal with it, as major |
916 |
browser developers care only for features, not about doing security |
917 |
right). |
918 |
|
919 |
|
920 |
=head1 BUGS |
921 |
|
922 |
While the goal of this module is to be correct, that unfortunately does |
923 |
not mean its bug-free, only that I think its design is bug-free. It is |
924 |
still relatively early in its development. If you keep reporting bugs they |
925 |
will be fixed swiftly, though. |
926 |
|
927 |
=cut |
928 |
|
929 |
our $true = do { bless \(my $dummy = 1), "JSON::XS::Boolean" }; |
930 |
our $false = do { bless \(my $dummy = 0), "JSON::XS::Boolean" }; |
931 |
|
932 |
sub true() { $true } |
933 |
sub false() { $false } |
934 |
|
935 |
sub is_bool($) { |
936 |
UNIVERSAL::isa $_[0], "JSON::XS::Boolean" |
937 |
# or UNIVERSAL::isa $_[0], "JSON::Literal" |
938 |
} |
939 |
|
940 |
XSLoader::load "JSON::XS", $VERSION; |
941 |
|
942 |
package JSON::XS::Boolean; |
943 |
|
944 |
use overload |
945 |
"0+" => sub { ${$_[0]} }, |
946 |
"++" => sub { $_[0] = ${$_[0]} + 1 }, |
947 |
"--" => sub { $_[0] = ${$_[0]} - 1 }, |
948 |
fallback => 1; |
949 |
|
950 |
1; |
951 |
|
952 |
=head1 AUTHOR |
953 |
|
954 |
Marc Lehmann <schmorp@schmorp.de> |
955 |
http://home.schmorp.de/ |
956 |
|
957 |
=cut |
958 |
|