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 |
=head1 DESCRIPTION |
10 |
|
11 |
This module converts Perl data structures to JSON and vice versa. Its |
12 |
primary goal is to be I<correct> and its secondary goal is to be |
13 |
I<fast>. To reach the latter goal it was written in C. |
14 |
|
15 |
As this is the n-th-something JSON module on CPAN, what was the reason |
16 |
to write yet another JSON module? While it seems there are many JSON |
17 |
modules, none of them correctly handle all corner cases, and in most cases |
18 |
their maintainers are unresponsive, gone missing, or not listening to bug |
19 |
reports for other reasons. |
20 |
|
21 |
See COMPARISON, below, for a comparison to some other JSON modules. |
22 |
|
23 |
See MAPPING, below, on how JSON::XS maps perl values to JSON values and |
24 |
vice versa. |
25 |
|
26 |
=head2 FEATURES |
27 |
|
28 |
=over 4 |
29 |
|
30 |
=item * correct handling of unicode issues |
31 |
|
32 |
This module knows how to handle Unicode, and even documents how and when |
33 |
it does so. |
34 |
|
35 |
=item * round-trip integrity |
36 |
|
37 |
When you serialise a perl data structure using only datatypes supported |
38 |
by JSON, the deserialised data structure is identical on the Perl level. |
39 |
(e.g. the string "2.0" doesn't suddenly become "2"). |
40 |
|
41 |
=item * strict checking of JSON correctness |
42 |
|
43 |
There is no guessing, no generating of illegal JSON strings by default, |
44 |
and only JSON is accepted as input by default (the latter is a security |
45 |
feature). |
46 |
|
47 |
=item * fast |
48 |
|
49 |
Compared to other JSON modules, this module compares favourably in terms |
50 |
of speed, too. |
51 |
|
52 |
=item * simple to use |
53 |
|
54 |
This module has both a simple functional interface as well as an OO |
55 |
interface. |
56 |
|
57 |
=item * reasonably versatile output formats |
58 |
|
59 |
You can choose between the most compact guarenteed single-line format |
60 |
possible (nice for simple line-based protocols), a pure-ascii format (for |
61 |
when your transport is not 8-bit clean), or a pretty-printed format (for |
62 |
when you want to read that stuff). Or you can combine those features in |
63 |
whatever way you like. |
64 |
|
65 |
=back |
66 |
|
67 |
=cut |
68 |
|
69 |
package JSON::XS; |
70 |
|
71 |
BEGIN { |
72 |
$VERSION = '0.3'; |
73 |
@ISA = qw(Exporter); |
74 |
|
75 |
@EXPORT = qw(to_json from_json); |
76 |
require Exporter; |
77 |
|
78 |
require XSLoader; |
79 |
XSLoader::load JSON::XS::, $VERSION; |
80 |
} |
81 |
|
82 |
=head1 FUNCTIONAL INTERFACE |
83 |
|
84 |
The following convinience methods are provided by this module. They are |
85 |
exported by default: |
86 |
|
87 |
=over 4 |
88 |
|
89 |
=item $json_string = to_json $perl_scalar |
90 |
|
91 |
Converts the given Perl data structure (a simple scalar or a reference to |
92 |
a hash or array) to a UTF-8 encoded, binary string (that is, the string contains |
93 |
octets only). Croaks on error. |
94 |
|
95 |
This function call is functionally identical to C<< JSON::XS->new->utf8->encode ($perl_scalar) >>. |
96 |
|
97 |
=item $perl_scalar = from_json $json_string |
98 |
|
99 |
The opposite of C<to_json>: expects an UTF-8 (binary) string and tries to |
100 |
parse that as an UTF-8 encoded JSON string, returning the resulting simple |
101 |
scalar or reference. Croaks on error. |
102 |
|
103 |
This function call is functionally identical to C<< JSON::XS->new->utf8->decode ($json_string) >>. |
104 |
|
105 |
=back |
106 |
|
107 |
=head1 OBJECT-ORIENTED INTERFACE |
108 |
|
109 |
The object oriented interface lets you configure your own encoding or |
110 |
decoding style, within the limits of supported formats. |
111 |
|
112 |
=over 4 |
113 |
|
114 |
=item $json = new JSON::XS |
115 |
|
116 |
Creates a new JSON::XS object that can be used to de/encode JSON |
117 |
strings. All boolean flags described below are by default I<disabled>. |
118 |
|
119 |
The mutators for flags all return the JSON object again and thus calls can |
120 |
be chained: |
121 |
|
122 |
my $json = JSON::XS->new->utf8(1)->space_after(1)->encode ({a => [1,2]}) |
123 |
=> {"a": [1, 2]} |
124 |
|
125 |
=item $json = $json->ascii ([$enable]) |
126 |
|
127 |
If C<$enable> is true (or missing), then the C<encode> method will |
128 |
not generate characters outside the code range C<0..127>. Any unicode |
129 |
characters outside that range will be escaped using either a single |
130 |
\uXXXX (BMP characters) or a double \uHHHH\uLLLLL escape sequence, as per |
131 |
RFC4627. |
132 |
|
133 |
If C<$enable> is false, then the C<encode> method will not escape Unicode |
134 |
characters unless necessary. |
135 |
|
136 |
JSON::XS->new->ascii (1)->encode (chr 0x10401) |
137 |
=> \ud801\udc01 |
138 |
|
139 |
=item $json = $json->utf8 ([$enable]) |
140 |
|
141 |
If C<$enable> is true (or missing), then the C<encode> method will encode |
142 |
the JSON string into UTF-8, as required by many protocols, while the |
143 |
C<decode> method expects to be handled an UTF-8-encoded string. Please |
144 |
note that UTF-8-encoded strings do not contain any characters outside the |
145 |
range C<0..255>, they are thus useful for bytewise/binary I/O. |
146 |
|
147 |
If C<$enable> is false, then the C<encode> method will return the JSON |
148 |
string as a (non-encoded) unicode string, while C<decode> expects thus a |
149 |
unicode string. Any decoding or encoding (e.g. to UTF-8 or UTF-16) needs |
150 |
to be done yourself, e.g. using the Encode module. |
151 |
|
152 |
=item $json = $json->pretty ([$enable]) |
153 |
|
154 |
This enables (or disables) all of the C<indent>, C<space_before> and |
155 |
C<space_after> (and in the future possibly more) flags in one call to |
156 |
generate the most readable (or most compact) form possible. |
157 |
|
158 |
my $json = JSON::XS->new->pretty(1)->encode ({a => [1,2]}) |
159 |
=> |
160 |
{ |
161 |
"a" : [ |
162 |
1, |
163 |
2 |
164 |
] |
165 |
} |
166 |
|
167 |
=item $json = $json->indent ([$enable]) |
168 |
|
169 |
If C<$enable> is true (or missing), then the C<encode> method will use a multiline |
170 |
format as output, putting every array member or object/hash key-value pair |
171 |
into its own line, identing them properly. |
172 |
|
173 |
If C<$enable> is false, no newlines or indenting will be produced, and the |
174 |
resulting JSON strings is guarenteed not to contain any C<newlines>. |
175 |
|
176 |
This setting has no effect when decoding JSON strings. |
177 |
|
178 |
=item $json = $json->space_before ([$enable]) |
179 |
|
180 |
If C<$enable> is true (or missing), then the C<encode> method will add an extra |
181 |
optional space before the C<:> separating keys from values in JSON objects. |
182 |
|
183 |
If C<$enable> is false, then the C<encode> method will not add any extra |
184 |
space at those places. |
185 |
|
186 |
This setting has no effect when decoding JSON strings. You will also most |
187 |
likely combine this setting with C<space_after>. |
188 |
|
189 |
=item $json = $json->space_after ([$enable]) |
190 |
|
191 |
If C<$enable> is true (or missing), then the C<encode> method will add an extra |
192 |
optional space after the C<:> separating keys from values in JSON objects |
193 |
and extra whitespace after the C<,> separating key-value pairs and array |
194 |
members. |
195 |
|
196 |
If C<$enable> is false, then the C<encode> method will not add any extra |
197 |
space at those places. |
198 |
|
199 |
This setting has no effect when decoding JSON strings. |
200 |
|
201 |
=item $json = $json->canonical ([$enable]) |
202 |
|
203 |
If C<$enable> is true (or missing), then the C<encode> method will output JSON objects |
204 |
by sorting their keys. This is adding a comparatively high overhead. |
205 |
|
206 |
If C<$enable> is false, then the C<encode> method will output key-value |
207 |
pairs in the order Perl stores them (which will likely change between runs |
208 |
of the same script). |
209 |
|
210 |
This option is useful if you want the same data structure to be encoded as |
211 |
the same JSON string (given the same overall settings). If it is disabled, |
212 |
the same hash migh be encoded differently even if contains the same data, |
213 |
as key-value pairs have no inherent ordering in Perl. |
214 |
|
215 |
This setting has no effect when decoding JSON strings. |
216 |
|
217 |
=item $json = $json->allow_nonref ([$enable]) |
218 |
|
219 |
If C<$enable> is true (or missing), then the C<encode> method can convert a |
220 |
non-reference into its corresponding string, number or null JSON value, |
221 |
which is an extension to RFC4627. Likewise, C<decode> will accept those JSON |
222 |
values instead of croaking. |
223 |
|
224 |
If C<$enable> is false, then the C<encode> method will croak if it isn't |
225 |
passed an arrayref or hashref, as JSON strings must either be an object |
226 |
or array. Likewise, C<decode> will croak if given something that is not a |
227 |
JSON object or array. |
228 |
|
229 |
=item $json = $json->shrink ([$enable]) |
230 |
|
231 |
Perl usually over-allocates memory a bit when allocating space for |
232 |
strings. This flag optionally resizes strings generated by either |
233 |
C<encode> or C<decode> to their minimum size possible. This can save |
234 |
memory when your JSON strings are either very very long or you have many |
235 |
short strings. It will also try to downgrade any strings to octet-form |
236 |
if possible: perl stores strings internally either in an encoding called |
237 |
UTF-X or in octet-form. The latter cannot store everything but uses less |
238 |
space in general. |
239 |
|
240 |
If C<$enable> is true (or missing), the string returned by C<encode> will be shrunk-to-fit, |
241 |
while all strings generated by C<decode> will also be shrunk-to-fit. |
242 |
|
243 |
If C<$enable> is false, then the normal perl allocation algorithms are used. |
244 |
If you work with your data, then this is likely to be faster. |
245 |
|
246 |
In the future, this setting might control other things, such as converting |
247 |
strings that look like integers or floats into integers or floats |
248 |
internally (there is no difference on the Perl level), saving space. |
249 |
|
250 |
=item $json_string = $json->encode ($perl_scalar) |
251 |
|
252 |
Converts the given Perl data structure (a simple scalar or a reference |
253 |
to a hash or array) to its JSON representation. Simple scalars will be |
254 |
converted into JSON string or number sequences, while references to arrays |
255 |
become JSON arrays and references to hashes become JSON objects. Undefined |
256 |
Perl values (e.g. C<undef>) become JSON C<null> values. Neither C<true> |
257 |
nor C<false> values will be generated. |
258 |
|
259 |
=item $perl_scalar = $json->decode ($json_string) |
260 |
|
261 |
The opposite of C<encode>: expects a JSON string and tries to parse it, |
262 |
returning the resulting simple scalar or reference. Croaks on error. |
263 |
|
264 |
JSON numbers and strings become simple Perl scalars. JSON arrays become |
265 |
Perl arrayrefs and JSON objects become Perl hashrefs. C<true> becomes |
266 |
C<1>, C<false> becomes C<0> and C<null> becomes C<undef>. |
267 |
|
268 |
=back |
269 |
|
270 |
=head1 MAPPING |
271 |
|
272 |
This section describes how JSON::XS maps Perl values to JSON values and |
273 |
vice versa. These mappings are designed to "do the right thing" in most |
274 |
circumstances automatically, preserving round-tripping characteristics |
275 |
(what you put in comes out as something equivalent). |
276 |
|
277 |
For the more enlightened: note that in the following descriptions, |
278 |
lowercase I<perl> refers to the Perl interpreter, while uppcercase I<Perl> |
279 |
refers to the abstract Perl language itself. |
280 |
|
281 |
=head2 JSON -> PERL |
282 |
|
283 |
=over 4 |
284 |
|
285 |
=item object |
286 |
|
287 |
A JSON object becomes a reference to a hash in Perl. No ordering of object |
288 |
keys is preserved. |
289 |
|
290 |
=item array |
291 |
|
292 |
A JSON array becomes a reference to an array in Perl. |
293 |
|
294 |
=item string |
295 |
|
296 |
A JSON string becomes a string scalar in Perl - Unicode codepoints in JSON |
297 |
are represented by the same codepoints in the Perl string, so no manual |
298 |
decoding is necessary. |
299 |
|
300 |
=item number |
301 |
|
302 |
A JSON number becomes either an integer or numeric (floating point) |
303 |
scalar in perl, depending on its range and any fractional parts. On the |
304 |
Perl level, there is no difference between those as Perl handles all the |
305 |
conversion details, but an integer may take slightly less memory and might |
306 |
represent more values exactly than (floating point) numbers. |
307 |
|
308 |
=item true, false |
309 |
|
310 |
These JSON atoms become C<0>, C<1>, respectively. Information is lost in |
311 |
this process. Future versions might represent those values differently, |
312 |
but they will be guarenteed to act like these integers would normally in |
313 |
Perl. |
314 |
|
315 |
=item null |
316 |
|
317 |
A JSON null atom becomes C<undef> in Perl. |
318 |
|
319 |
=back |
320 |
|
321 |
=head2 PERL -> JSON |
322 |
|
323 |
The mapping from Perl to JSON is slightly more difficult, as Perl is a |
324 |
truly typeless language, so we can only guess which JSON type is meant by |
325 |
a Perl value. |
326 |
|
327 |
=over 4 |
328 |
|
329 |
=item hash references |
330 |
|
331 |
Perl hash references become JSON objects. As there is no inherent ordering |
332 |
in hash keys, they will usually be encoded in a pseudo-random order that |
333 |
can change between runs of the same program but stays generally the same |
334 |
within the single run of a program. JSON::XS can optionally sort the hash |
335 |
keys (determined by the I<canonical> flag), so the same datastructure |
336 |
will serialise to the same JSON text (given same settings and version of |
337 |
JSON::XS), but this incurs a runtime overhead. |
338 |
|
339 |
=item array references |
340 |
|
341 |
Perl array references become JSON arrays. |
342 |
|
343 |
=item blessed objects |
344 |
|
345 |
Blessed objects are not allowed. JSON::XS currently tries to encode their |
346 |
underlying representation (hash- or arrayref), but this behaviour might |
347 |
change in future versions. |
348 |
|
349 |
=item simple scalars |
350 |
|
351 |
Simple Perl scalars (any scalar that is not a reference) are the most |
352 |
difficult objects to encode: JSON::XS will encode undefined scalars as |
353 |
JSON null value, scalars that have last been used in a string context |
354 |
before encoding as JSON strings and anything else as number value: |
355 |
|
356 |
# dump as number |
357 |
to_json [2] # yields [2] |
358 |
to_json [-3.0e17] # yields [-3e+17] |
359 |
my $value = 5; to_json [$value] # yields [5] |
360 |
|
361 |
# used as string, so dump as string |
362 |
print $value; |
363 |
to_json [$value] # yields ["5"] |
364 |
|
365 |
# undef becomes null |
366 |
to_json [undef] # yields [null] |
367 |
|
368 |
You can force the type to be a string by stringifying it: |
369 |
|
370 |
my $x = 3.1; # some variable containing a number |
371 |
"$x"; # stringified |
372 |
$x .= ""; # another, more awkward way to stringify |
373 |
print $x; # perl does it for you, too, quite often |
374 |
|
375 |
You can force the type to be a number by numifying it: |
376 |
|
377 |
my $x = "3"; # some variable containing a string |
378 |
$x += 0; # numify it, ensuring it will be dumped as a number |
379 |
$x *= 1; # same thing, the choise is yours. |
380 |
|
381 |
You can not currently output JSON booleans or force the type in other, |
382 |
less obscure, ways. Tell me if you need this capability. |
383 |
|
384 |
=item circular data structures |
385 |
|
386 |
Those will be encoded until memory or stackspace runs out. |
387 |
|
388 |
=back |
389 |
|
390 |
=head1 COMPARISON |
391 |
|
392 |
As already mentioned, this module was created because none of the existing |
393 |
JSON modules could be made to work correctly. First I will describe the |
394 |
problems (or pleasures) I encountered with various existing JSON modules, |
395 |
followed by some benchmark values. JSON::XS was designed not to suffer |
396 |
from any of these problems or limitations. |
397 |
|
398 |
=over 4 |
399 |
|
400 |
=item JSON 1.07 |
401 |
|
402 |
Slow (but very portable, as it is written in pure Perl). |
403 |
|
404 |
Undocumented/buggy Unicode handling (how JSON handles unicode values is |
405 |
undocumented. One can get far by feeding it unicode strings and doing |
406 |
en-/decoding oneself, but unicode escapes are not working properly). |
407 |
|
408 |
No roundtripping (strings get clobbered if they look like numbers, e.g. |
409 |
the string C<2.0> will encode to C<2.0> instead of C<"2.0">, and that will |
410 |
decode into the number 2. |
411 |
|
412 |
=item JSON::PC 0.01 |
413 |
|
414 |
Very fast. |
415 |
|
416 |
Undocumented/buggy Unicode handling. |
417 |
|
418 |
No roundtripping. |
419 |
|
420 |
Has problems handling many Perl values (e.g. regex results and other magic |
421 |
values will make it croak). |
422 |
|
423 |
Does not even generate valid JSON (C<{1,2}> gets converted to C<{1:2}> |
424 |
which is not a valid JSON string. |
425 |
|
426 |
Unmaintained (maintainer unresponsive for many months, bugs are not |
427 |
getting fixed). |
428 |
|
429 |
=item JSON::Syck 0.21 |
430 |
|
431 |
Very buggy (often crashes). |
432 |
|
433 |
Very inflexible (no human-readable format supported, format pretty much |
434 |
undocumented. I need at least a format for easy reading by humans and a |
435 |
single-line compact format for use in a protocol, and preferably a way to |
436 |
generate ASCII-only JSON strings). |
437 |
|
438 |
Completely broken (and confusingly documented) Unicode handling (unicode |
439 |
escapes are not working properly, you need to set ImplicitUnicode to |
440 |
I<different> values on en- and decoding to get symmetric behaviour). |
441 |
|
442 |
No roundtripping (simple cases work, but this depends on wether the scalar |
443 |
value was used in a numeric context or not). |
444 |
|
445 |
Dumping hashes may skip hash values depending on iterator state. |
446 |
|
447 |
Unmaintained (maintainer unresponsive for many months, bugs are not |
448 |
getting fixed). |
449 |
|
450 |
Does not check input for validity (i.e. will accept non-JSON input and |
451 |
return "something" instead of raising an exception. This is a security |
452 |
issue: imagine two banks transfering money between each other using |
453 |
JSON. One bank might parse a given non-JSON request and deduct money, |
454 |
while the other might reject the transaction with a syntax error. While a |
455 |
good protocol will at least recover, that is extra unnecessary work and |
456 |
the transaction will still not succeed). |
457 |
|
458 |
=item JSON::DWIW 0.04 |
459 |
|
460 |
Very fast. Very natural. Very nice. |
461 |
|
462 |
Undocumented unicode handling (but the best of the pack. Unicode escapes |
463 |
still don't get parsed properly). |
464 |
|
465 |
Very inflexible. |
466 |
|
467 |
No roundtripping. |
468 |
|
469 |
Does not generate valid JSON (key strings are often unquoted, empty keys |
470 |
result in nothing being output) |
471 |
|
472 |
Does not check input for validity. |
473 |
|
474 |
=back |
475 |
|
476 |
=head2 SPEED |
477 |
|
478 |
It seems that JSON::XS is surprisingly fast, as shown in the following |
479 |
tables. They have been generated with the help of the C<eg/bench> program |
480 |
in the JSON::XS distribution, to make it easy to compare on your own |
481 |
system. |
482 |
|
483 |
First is a comparison between various modules using a very simple JSON |
484 |
string, showing the number of encodes/decodes per second (JSON::XS is |
485 |
the functional interface, while JSON::XS/2 is the OO interface with |
486 |
pretty-printing and hashkey sorting enabled). |
487 |
|
488 |
module | encode | decode | |
489 |
-----------|------------|------------| |
490 |
JSON | 14006 | 6820 | |
491 |
JSON::DWIW | 200937 | 120386 | |
492 |
JSON::PC | 85065 | 129366 | |
493 |
JSON::Syck | 59898 | 44232 | |
494 |
JSON::XS | 1171478 | 342435 | |
495 |
JSON::XS/2 | 730760 | 328714 | |
496 |
-----------+------------+------------+ |
497 |
|
498 |
That is, JSON::XS is 6 times faster than than JSON::DWIW and about 80 |
499 |
times faster than JSON, even with pretty-printing and key sorting. |
500 |
|
501 |
Using a longer test string (roughly 8KB, generated from Yahoo! Locals |
502 |
search API (http://nanoref.com/yahooapis/mgPdGg): |
503 |
|
504 |
module | encode | decode | |
505 |
-----------|------------|------------| |
506 |
JSON | 673 | 38 | |
507 |
JSON::DWIW | 5271 | 770 | |
508 |
JSON::PC | 9901 | 2491 | |
509 |
JSON::Syck | 2360 | 786 | |
510 |
JSON::XS | 37398 | 3202 | |
511 |
JSON::XS/2 | 13765 | 3153 | |
512 |
-----------+------------+------------+ |
513 |
|
514 |
Again, JSON::XS leads by far in the encoding case, while still beating |
515 |
every other module in the decoding case. |
516 |
|
517 |
Last example is an almost 8MB large hash with many large binary values |
518 |
(PNG files), resulting in a lot of escaping: |
519 |
|
520 |
=head1 RESOURCE LIMITS |
521 |
|
522 |
JSON::XS does not impose any limits on the size of JSON texts or Perl |
523 |
values they represent - if your machine cna handle it, JSON::XS will |
524 |
encode or decode it. Future versions might optionally impose structure |
525 |
depth and memory use resource limits. |
526 |
|
527 |
=head1 BUGS |
528 |
|
529 |
While the goal of this module is to be correct, that unfortunately does |
530 |
not mean its bug-free, only that I think its design is bug-free. It is |
531 |
still very young and not well-tested. If you keep reporting bugs they will |
532 |
be fixed swiftly, though. |
533 |
|
534 |
=cut |
535 |
|
536 |
1; |
537 |
|
538 |
=head1 AUTHOR |
539 |
|
540 |
Marc Lehmann <schmorp@schmorp.de> |
541 |
http://home.schmorp.de/ |
542 |
|
543 |
=cut |
544 |
|