1 |
NAME |
2 |
JSON::XS - JSON serialising/deserialising, done correctly and fast |
3 |
|
4 |
SYNOPSIS |
5 |
use JSON::XS; |
6 |
|
7 |
# exported functions, they croak on error |
8 |
# and expect/generate UTF-8 |
9 |
|
10 |
$utf8_encoded_json_text = to_json $perl_hash_or_arrayref; |
11 |
$perl_hash_or_arrayref = from_json $utf8_encoded_json_text; |
12 |
|
13 |
# objToJson and jsonToObj aliases to to_json and from_json |
14 |
# are exported for compatibility to the JSON module, |
15 |
# but should not be used in new code. |
16 |
|
17 |
# OO-interface |
18 |
|
19 |
$coder = JSON::XS->new->ascii->pretty->allow_nonref; |
20 |
$pretty_printed_unencoded = $coder->encode ($perl_scalar); |
21 |
$perl_scalar = $coder->decode ($unicode_json_text); |
22 |
|
23 |
DESCRIPTION |
24 |
This module converts Perl data structures to JSON and vice versa. Its |
25 |
primary goal is to be *correct* and its secondary goal is to be *fast*. |
26 |
To reach the latter goal it was written in C. |
27 |
|
28 |
As this is the n-th-something JSON module on CPAN, what was the reason |
29 |
to write yet another JSON module? While it seems there are many JSON |
30 |
modules, none of them correctly handle all corner cases, and in most |
31 |
cases their maintainers are unresponsive, gone missing, or not listening |
32 |
to bug reports for other reasons. |
33 |
|
34 |
See COMPARISON, below, for a comparison to some other JSON modules. |
35 |
|
36 |
See MAPPING, below, on how JSON::XS maps perl values to JSON values and |
37 |
vice versa. |
38 |
|
39 |
FEATURES |
40 |
* correct unicode handling |
41 |
This module knows how to handle Unicode, and even documents how and |
42 |
when it does so. |
43 |
|
44 |
* round-trip integrity |
45 |
When you serialise a perl data structure using only datatypes |
46 |
supported by JSON, the deserialised data structure is identical on |
47 |
the Perl level. (e.g. the string "2.0" doesn't suddenly become "2" |
48 |
just because it looks like a number). |
49 |
|
50 |
* strict checking of JSON correctness |
51 |
There is no guessing, no generating of illegal JSON texts by |
52 |
default, and only JSON is accepted as input by default (the latter |
53 |
is a security feature). |
54 |
|
55 |
* fast |
56 |
Compared to other JSON modules, this module compares favourably in |
57 |
terms of speed, too. |
58 |
|
59 |
* simple to use |
60 |
This module has both a simple functional interface as well as an OO |
61 |
interface. |
62 |
|
63 |
* reasonably versatile output formats |
64 |
You can choose between the most compact guarenteed single-line |
65 |
format possible (nice for simple line-based protocols), a pure-ascii |
66 |
format (for when your transport is not 8-bit clean, still supports |
67 |
the whole unicode range), or a pretty-printed format (for when you |
68 |
want to read that stuff). Or you can combine those features in |
69 |
whatever way you like. |
70 |
|
71 |
FUNCTIONAL INTERFACE |
72 |
The following convinience methods are provided by this module. They are |
73 |
exported by default: |
74 |
|
75 |
$json_text = to_json $perl_scalar |
76 |
Converts the given Perl data structure (a simple scalar or a |
77 |
reference to a hash or array) to a UTF-8 encoded, binary string |
78 |
(that is, the string contains octets only). Croaks on error. |
79 |
|
80 |
This function call is functionally identical to: |
81 |
|
82 |
$json_text = JSON::XS->new->utf8->encode ($perl_scalar) |
83 |
|
84 |
except being faster. |
85 |
|
86 |
$perl_scalar = from_json $json_text |
87 |
The opposite of "to_json": expects an UTF-8 (binary) string and |
88 |
tries to parse that as an UTF-8 encoded JSON text, returning the |
89 |
resulting simple scalar or reference. Croaks on error. |
90 |
|
91 |
This function call is functionally identical to: |
92 |
|
93 |
$perl_scalar = JSON::XS->new->utf8->decode ($json_text) |
94 |
|
95 |
except being faster. |
96 |
|
97 |
$is_boolean = JSON::XS::is_bool $scalar |
98 |
Returns true if the passed scalar represents either JSON::XS::true |
99 |
or JSON::XS::false, two constants that act like 1 and 0, |
100 |
respectively and are used to represent JSON "true" and "false" |
101 |
values in Perl. |
102 |
|
103 |
See MAPPING, below, for more information on how JSON values are |
104 |
mapped to Perl. |
105 |
|
106 |
OBJECT-ORIENTED INTERFACE |
107 |
The object oriented interface lets you configure your own encoding or |
108 |
decoding style, within the limits of supported formats. |
109 |
|
110 |
$json = new JSON::XS |
111 |
Creates a new JSON::XS object that can be used to de/encode JSON |
112 |
strings. All boolean flags described below are by default |
113 |
*disabled*. |
114 |
|
115 |
The mutators for flags all return the JSON object again and thus |
116 |
calls can be chained: |
117 |
|
118 |
my $json = JSON::XS->new->utf8->space_after->encode ({a => [1,2]}) |
119 |
=> {"a": [1, 2]} |
120 |
|
121 |
$json = $json->ascii ([$enable]) |
122 |
If $enable is true (or missing), then the "encode" method will not |
123 |
generate characters outside the code range 0..127 (which is ASCII). |
124 |
Any unicode characters outside that range will be escaped using |
125 |
either a single \uXXXX (BMP characters) or a double \uHHHH\uLLLLL |
126 |
escape sequence, as per RFC4627. The resulting encoded JSON text can |
127 |
be treated as a native unicode string, an ascii-encoded, |
128 |
latin1-encoded or UTF-8 encoded string, or any other superset of |
129 |
ASCII. |
130 |
|
131 |
If $enable is false, then the "encode" method will not escape |
132 |
Unicode characters unless required by the JSON syntax or other |
133 |
flags. This results in a faster and more compact format. |
134 |
|
135 |
The main use for this flag is to produce JSON texts that can be |
136 |
transmitted over a 7-bit channel, as the encoded JSON texts will not |
137 |
contain any 8 bit characters. |
138 |
|
139 |
JSON::XS->new->ascii (1)->encode ([chr 0x10401]) |
140 |
=> ["\ud801\udc01"] |
141 |
|
142 |
$json = $json->latin1 ([$enable]) |
143 |
If $enable is true (or missing), then the "encode" method will |
144 |
encode the resulting JSON text as latin1 (or iso-8859-1), escaping |
145 |
any characters outside the code range 0..255. The resulting string |
146 |
can be treated as a latin1-encoded JSON text or a native unicode |
147 |
string. The "decode" method will not be affected in any way by this |
148 |
flag, as "decode" by default expects unicode, which is a strict |
149 |
superset of latin1. |
150 |
|
151 |
If $enable is false, then the "encode" method will not escape |
152 |
Unicode characters unless required by the JSON syntax or other |
153 |
flags. |
154 |
|
155 |
The main use for this flag is efficiently encoding binary data as |
156 |
JSON text, as most octets will not be escaped, resulting in a |
157 |
smaller encoded size. The disadvantage is that the resulting JSON |
158 |
text is encoded in latin1 (and must correctly be treated as such |
159 |
when storing and transfering), a rare encoding for JSON. It is |
160 |
therefore most useful when you want to store data structures known |
161 |
to contain binary data efficiently in files or databases, not when |
162 |
talking to other JSON encoders/decoders. |
163 |
|
164 |
JSON::XS->new->latin1->encode (["\x{89}\x{abc}"] |
165 |
=> ["\x{89}\\u0abc"] # (perl syntax, U+abc escaped, U+89 not) |
166 |
|
167 |
$json = $json->utf8 ([$enable]) |
168 |
If $enable is true (or missing), then the "encode" method will |
169 |
encode the JSON result into UTF-8, as required by many protocols, |
170 |
while the "decode" method expects to be handled an UTF-8-encoded |
171 |
string. Please note that UTF-8-encoded strings do not contain any |
172 |
characters outside the range 0..255, they are thus useful for |
173 |
bytewise/binary I/O. In future versions, enabling this option might |
174 |
enable autodetection of the UTF-16 and UTF-32 encoding families, as |
175 |
described in RFC4627. |
176 |
|
177 |
If $enable is false, then the "encode" method will return the JSON |
178 |
string as a (non-encoded) unicode string, while "decode" expects |
179 |
thus a unicode string. Any decoding or encoding (e.g. to UTF-8 or |
180 |
UTF-16) needs to be done yourself, e.g. using the Encode module. |
181 |
|
182 |
Example, output UTF-16BE-encoded JSON: |
183 |
|
184 |
use Encode; |
185 |
$jsontext = encode "UTF-16BE", JSON::XS->new->encode ($object); |
186 |
|
187 |
Example, decode UTF-32LE-encoded JSON: |
188 |
|
189 |
use Encode; |
190 |
$object = JSON::XS->new->decode (decode "UTF-32LE", $jsontext); |
191 |
|
192 |
$json = $json->pretty ([$enable]) |
193 |
This enables (or disables) all of the "indent", "space_before" and |
194 |
"space_after" (and in the future possibly more) flags in one call to |
195 |
generate the most readable (or most compact) form possible. |
196 |
|
197 |
Example, pretty-print some simple structure: |
198 |
|
199 |
my $json = JSON::XS->new->pretty(1)->encode ({a => [1,2]}) |
200 |
=> |
201 |
{ |
202 |
"a" : [ |
203 |
1, |
204 |
2 |
205 |
] |
206 |
} |
207 |
|
208 |
$json = $json->indent ([$enable]) |
209 |
If $enable is true (or missing), then the "encode" method will use a |
210 |
multiline format as output, putting every array member or |
211 |
object/hash key-value pair into its own line, identing them |
212 |
properly. |
213 |
|
214 |
If $enable is false, no newlines or indenting will be produced, and |
215 |
the resulting JSON text is guarenteed not to contain any "newlines". |
216 |
|
217 |
This setting has no effect when decoding JSON texts. |
218 |
|
219 |
$json = $json->space_before ([$enable]) |
220 |
If $enable is true (or missing), then the "encode" method will add |
221 |
an extra optional space before the ":" separating keys from values |
222 |
in JSON objects. |
223 |
|
224 |
If $enable is false, then the "encode" method will not add any extra |
225 |
space at those places. |
226 |
|
227 |
This setting has no effect when decoding JSON texts. You will also |
228 |
most likely combine this setting with "space_after". |
229 |
|
230 |
Example, space_before enabled, space_after and indent disabled: |
231 |
|
232 |
{"key" :"value"} |
233 |
|
234 |
$json = $json->space_after ([$enable]) |
235 |
If $enable is true (or missing), then the "encode" method will add |
236 |
an extra optional space after the ":" separating keys from values in |
237 |
JSON objects and extra whitespace after the "," separating key-value |
238 |
pairs and array members. |
239 |
|
240 |
If $enable is false, then the "encode" method will not add any extra |
241 |
space at those places. |
242 |
|
243 |
This setting has no effect when decoding JSON texts. |
244 |
|
245 |
Example, space_before and indent disabled, space_after enabled: |
246 |
|
247 |
{"key": "value"} |
248 |
|
249 |
$json = $json->canonical ([$enable]) |
250 |
If $enable is true (or missing), then the "encode" method will |
251 |
output JSON objects by sorting their keys. This is adding a |
252 |
comparatively high overhead. |
253 |
|
254 |
If $enable is false, then the "encode" method will output key-value |
255 |
pairs in the order Perl stores them (which will likely change |
256 |
between runs of the same script). |
257 |
|
258 |
This option is useful if you want the same data structure to be |
259 |
encoded as the same JSON text (given the same overall settings). If |
260 |
it is disabled, the same hash migh be encoded differently even if |
261 |
contains the same data, as key-value pairs have no inherent ordering |
262 |
in Perl. |
263 |
|
264 |
This setting has no effect when decoding JSON texts. |
265 |
|
266 |
$json = $json->allow_nonref ([$enable]) |
267 |
If $enable is true (or missing), then the "encode" method can |
268 |
convert a non-reference into its corresponding string, number or |
269 |
null JSON value, which is an extension to RFC4627. Likewise, |
270 |
"decode" will accept those JSON values instead of croaking. |
271 |
|
272 |
If $enable is false, then the "encode" method will croak if it isn't |
273 |
passed an arrayref or hashref, as JSON texts must either be an |
274 |
object or array. Likewise, "decode" will croak if given something |
275 |
that is not a JSON object or array. |
276 |
|
277 |
Example, encode a Perl scalar as JSON value with enabled |
278 |
"allow_nonref", resulting in an invalid JSON text: |
279 |
|
280 |
JSON::XS->new->allow_nonref->encode ("Hello, World!") |
281 |
=> "Hello, World!" |
282 |
|
283 |
$json = $json->shrink ([$enable]) |
284 |
Perl usually over-allocates memory a bit when allocating space for |
285 |
strings. This flag optionally resizes strings generated by either |
286 |
"encode" or "decode" to their minimum size possible. This can save |
287 |
memory when your JSON texts are either very very long or you have |
288 |
many short strings. It will also try to downgrade any strings to |
289 |
octet-form if possible: perl stores strings internally either in an |
290 |
encoding called UTF-X or in octet-form. The latter cannot store |
291 |
everything but uses less space in general (and some buggy Perl or C |
292 |
code might even rely on that internal representation being used). |
293 |
|
294 |
The actual definition of what shrink does might change in future |
295 |
versions, but it will always try to save space at the expense of |
296 |
time. |
297 |
|
298 |
If $enable is true (or missing), the string returned by "encode" |
299 |
will be shrunk-to-fit, while all strings generated by "decode" will |
300 |
also be shrunk-to-fit. |
301 |
|
302 |
If $enable is false, then the normal perl allocation algorithms are |
303 |
used. If you work with your data, then this is likely to be faster. |
304 |
|
305 |
In the future, this setting might control other things, such as |
306 |
converting strings that look like integers or floats into integers |
307 |
or floats internally (there is no difference on the Perl level), |
308 |
saving space. |
309 |
|
310 |
$json = $json->max_depth ([$maximum_nesting_depth]) |
311 |
Sets the maximum nesting level (default 512) accepted while encoding |
312 |
or decoding. If the JSON text or Perl data structure has an equal or |
313 |
higher nesting level then this limit, then the encoder and decoder |
314 |
will stop and croak at that point. |
315 |
|
316 |
Nesting level is defined by number of hash- or arrayrefs that the |
317 |
encoder needs to traverse to reach a given point or the number of |
318 |
"{" or "[" characters without their matching closing parenthesis |
319 |
crossed to reach a given character in a string. |
320 |
|
321 |
Setting the maximum depth to one disallows any nesting, so that |
322 |
ensures that the object is only a single hash/object or array. |
323 |
|
324 |
The argument to "max_depth" will be rounded up to the next nearest |
325 |
power of two. |
326 |
|
327 |
See SECURITY CONSIDERATIONS, below, for more info on why this is |
328 |
useful. |
329 |
|
330 |
$json_text = $json->encode ($perl_scalar) |
331 |
Converts the given Perl data structure (a simple scalar or a |
332 |
reference to a hash or array) to its JSON representation. Simple |
333 |
scalars will be converted into JSON string or number sequences, |
334 |
while references to arrays become JSON arrays and references to |
335 |
hashes become JSON objects. Undefined Perl values (e.g. "undef") |
336 |
become JSON "null" values. Neither "true" nor "false" values will be |
337 |
generated. |
338 |
|
339 |
$perl_scalar = $json->decode ($json_text) |
340 |
The opposite of "encode": expects a JSON text and tries to parse it, |
341 |
returning the resulting simple scalar or reference. Croaks on error. |
342 |
|
343 |
JSON numbers and strings become simple Perl scalars. JSON arrays |
344 |
become Perl arrayrefs and JSON objects become Perl hashrefs. "true" |
345 |
becomes 1, "false" becomes 0 and "null" becomes "undef". |
346 |
|
347 |
($perl_scalar, $characters) = $json->decode_prefix ($json_text) |
348 |
This works like the "decode" method, but instead of raising an |
349 |
exception when there is trailing garbage after the first JSON |
350 |
object, it will silently stop parsing there and return the number of |
351 |
characters consumed so far. |
352 |
|
353 |
This is useful if your JSON texts are not delimited by an outer |
354 |
protocol (which is not the brightest thing to do in the first place) |
355 |
and you need to know where the JSON text ends. |
356 |
|
357 |
JSON::XS->new->decode_prefix ("[1] the tail") |
358 |
=> ([], 3) |
359 |
|
360 |
MAPPING |
361 |
This section describes how JSON::XS maps Perl values to JSON values and |
362 |
vice versa. These mappings are designed to "do the right thing" in most |
363 |
circumstances automatically, preserving round-tripping characteristics |
364 |
(what you put in comes out as something equivalent). |
365 |
|
366 |
For the more enlightened: note that in the following descriptions, |
367 |
lowercase *perl* refers to the Perl interpreter, while uppcercase *Perl* |
368 |
refers to the abstract Perl language itself. |
369 |
|
370 |
JSON -> PERL |
371 |
object |
372 |
A JSON object becomes a reference to a hash in Perl. No ordering of |
373 |
object keys is preserved (JSON does not preserver object key |
374 |
ordering itself). |
375 |
|
376 |
array |
377 |
A JSON array becomes a reference to an array in Perl. |
378 |
|
379 |
string |
380 |
A JSON string becomes a string scalar in Perl - Unicode codepoints |
381 |
in JSON are represented by the same codepoints in the Perl string, |
382 |
so no manual decoding is necessary. |
383 |
|
384 |
number |
385 |
A JSON number becomes either an integer or numeric (floating point) |
386 |
scalar in perl, depending on its range and any fractional parts. On |
387 |
the Perl level, there is no difference between those as Perl handles |
388 |
all the conversion details, but an integer may take slightly less |
389 |
memory and might represent more values exactly than (floating point) |
390 |
numbers. |
391 |
|
392 |
true, false |
393 |
These JSON atoms become "JSON::XS::true" and "JSON::XS::false", |
394 |
respectively. They are overloaded to act almost exactly like the |
395 |
numbers 1 and 0. You can check wether a scalar is a JSON boolean by |
396 |
using the "JSON::XS::is_bool" function. |
397 |
|
398 |
null |
399 |
A JSON null atom becomes "undef" in Perl. |
400 |
|
401 |
PERL -> JSON |
402 |
The mapping from Perl to JSON is slightly more difficult, as Perl is a |
403 |
truly typeless language, so we can only guess which JSON type is meant |
404 |
by a Perl value. |
405 |
|
406 |
hash references |
407 |
Perl hash references become JSON objects. As there is no inherent |
408 |
ordering in hash keys (or JSON objects), they will usually be |
409 |
encoded in a pseudo-random order that can change between runs of the |
410 |
same program but stays generally the same within a single run of a |
411 |
program. JSON::XS can optionally sort the hash keys (determined by |
412 |
the *canonical* flag), so the same datastructure will serialise to |
413 |
the same JSON text (given same settings and version of JSON::XS), |
414 |
but this incurs a runtime overhead and is only rarely useful, e.g. |
415 |
when you want to compare some JSON text against another for |
416 |
equality. |
417 |
|
418 |
array references |
419 |
Perl array references become JSON arrays. |
420 |
|
421 |
other references |
422 |
Other unblessed references are generally not allowed and will cause |
423 |
an exception to be thrown, except for references to the integers 0 |
424 |
and 1, which get turned into "false" and "true" atoms in JSON. You |
425 |
can also use "JSON::XS::false" and "JSON::XS::true" to improve |
426 |
readability. |
427 |
|
428 |
to_json [\0,JSON::XS::true] # yields [false,true] |
429 |
|
430 |
JSON::XS::true, JSON::XS::false |
431 |
These special values become JSON true and JSON false values, |
432 |
respectively. You cna alos use "\1" and "\0" directly if you want. |
433 |
|
434 |
blessed objects |
435 |
Blessed objects are not allowed. JSON::XS currently tries to encode |
436 |
their underlying representation (hash- or arrayref), but this |
437 |
behaviour might change in future versions. |
438 |
|
439 |
simple scalars |
440 |
Simple Perl scalars (any scalar that is not a reference) are the |
441 |
most difficult objects to encode: JSON::XS will encode undefined |
442 |
scalars as JSON null value, scalars that have last been used in a |
443 |
string context before encoding as JSON strings and anything else as |
444 |
number value: |
445 |
|
446 |
# dump as number |
447 |
to_json [2] # yields [2] |
448 |
to_json [-3.0e17] # yields [-3e+17] |
449 |
my $value = 5; to_json [$value] # yields [5] |
450 |
|
451 |
# used as string, so dump as string |
452 |
print $value; |
453 |
to_json [$value] # yields ["5"] |
454 |
|
455 |
# undef becomes null |
456 |
to_json [undef] # yields [null] |
457 |
|
458 |
You can force the type to be a string by stringifying it: |
459 |
|
460 |
my $x = 3.1; # some variable containing a number |
461 |
"$x"; # stringified |
462 |
$x .= ""; # another, more awkward way to stringify |
463 |
print $x; # perl does it for you, too, quite often |
464 |
|
465 |
You can force the type to be a number by numifying it: |
466 |
|
467 |
my $x = "3"; # some variable containing a string |
468 |
$x += 0; # numify it, ensuring it will be dumped as a number |
469 |
$x *= 1; # same thing, the choise is yours. |
470 |
|
471 |
You can not currently output JSON booleans or force the type in |
472 |
other, less obscure, ways. Tell me if you need this capability. |
473 |
|
474 |
COMPARISON |
475 |
As already mentioned, this module was created because none of the |
476 |
existing JSON modules could be made to work correctly. First I will |
477 |
describe the problems (or pleasures) I encountered with various existing |
478 |
JSON modules, followed by some benchmark values. JSON::XS was designed |
479 |
not to suffer from any of these problems or limitations. |
480 |
|
481 |
JSON 1.07 |
482 |
Slow (but very portable, as it is written in pure Perl). |
483 |
|
484 |
Undocumented/buggy Unicode handling (how JSON handles unicode values |
485 |
is undocumented. One can get far by feeding it unicode strings and |
486 |
doing en-/decoding oneself, but unicode escapes are not working |
487 |
properly). |
488 |
|
489 |
No roundtripping (strings get clobbered if they look like numbers, |
490 |
e.g. the string 2.0 will encode to 2.0 instead of "2.0", and that |
491 |
will decode into the number 2. |
492 |
|
493 |
JSON::PC 0.01 |
494 |
Very fast. |
495 |
|
496 |
Undocumented/buggy Unicode handling. |
497 |
|
498 |
No roundtripping. |
499 |
|
500 |
Has problems handling many Perl values (e.g. regex results and other |
501 |
magic values will make it croak). |
502 |
|
503 |
Does not even generate valid JSON ("{1,2}" gets converted to "{1:2}" |
504 |
which is not a valid JSON text. |
505 |
|
506 |
Unmaintained (maintainer unresponsive for many months, bugs are not |
507 |
getting fixed). |
508 |
|
509 |
JSON::Syck 0.21 |
510 |
Very buggy (often crashes). |
511 |
|
512 |
Very inflexible (no human-readable format supported, format pretty |
513 |
much undocumented. I need at least a format for easy reading by |
514 |
humans and a single-line compact format for use in a protocol, and |
515 |
preferably a way to generate ASCII-only JSON texts). |
516 |
|
517 |
Completely broken (and confusingly documented) Unicode handling |
518 |
(unicode escapes are not working properly, you need to set |
519 |
ImplicitUnicode to *different* values on en- and decoding to get |
520 |
symmetric behaviour). |
521 |
|
522 |
No roundtripping (simple cases work, but this depends on wether the |
523 |
scalar value was used in a numeric context or not). |
524 |
|
525 |
Dumping hashes may skip hash values depending on iterator state. |
526 |
|
527 |
Unmaintained (maintainer unresponsive for many months, bugs are not |
528 |
getting fixed). |
529 |
|
530 |
Does not check input for validity (i.e. will accept non-JSON input |
531 |
and return "something" instead of raising an exception. This is a |
532 |
security issue: imagine two banks transfering money between each |
533 |
other using JSON. One bank might parse a given non-JSON request and |
534 |
deduct money, while the other might reject the transaction with a |
535 |
syntax error. While a good protocol will at least recover, that is |
536 |
extra unnecessary work and the transaction will still not succeed). |
537 |
|
538 |
JSON::DWIW 0.04 |
539 |
Very fast. Very natural. Very nice. |
540 |
|
541 |
Undocumented unicode handling (but the best of the pack. Unicode |
542 |
escapes still don't get parsed properly). |
543 |
|
544 |
Very inflexible. |
545 |
|
546 |
No roundtripping. |
547 |
|
548 |
Does not generate valid JSON texts (key strings are often unquoted, |
549 |
empty keys result in nothing being output) |
550 |
|
551 |
Does not check input for validity. |
552 |
|
553 |
JSON and YAML |
554 |
You often hear that JSON is a subset (or a close subset) of YAML. This |
555 |
is, however, a mass hysteria and very far from the truth. In general, |
556 |
there is no way to configure JSON::XS to output a data structure as |
557 |
valid YAML. |
558 |
|
559 |
If you really must use JSON::XS to generate YAML, you should use this |
560 |
algorithm (subject to change in future versions): |
561 |
|
562 |
my $to_yaml = JSON::XS->new->utf8->space_after (1); |
563 |
my $yaml = $to_yaml->encode ($ref) . "\n"; |
564 |
|
565 |
This will usually generate JSON texts that also parse as valid YAML. |
566 |
Please note that YAML has hardcoded limits on (simple) object key |
567 |
lengths that JSON doesn't have, so you should make sure that your hash |
568 |
keys are noticably shorter than the 1024 characters YAML allows. |
569 |
|
570 |
There might be other incompatibilities that I am not aware of. In |
571 |
general you should not try to generate YAML with a JSON generator or |
572 |
vice versa, or try to parse JSON with a YAML parser or vice versa: |
573 |
chances are high that you will run into severe interoperability |
574 |
problems. |
575 |
|
576 |
SPEED |
577 |
It seems that JSON::XS is surprisingly fast, as shown in the following |
578 |
tables. They have been generated with the help of the "eg/bench" program |
579 |
in the JSON::XS distribution, to make it easy to compare on your own |
580 |
system. |
581 |
|
582 |
First comes a comparison between various modules using a very short |
583 |
single-line JSON string: |
584 |
|
585 |
{"method": "handleMessage", "params": ["user1", "we were just talking"], \ |
586 |
"id": null, "array":[1,11,234,-5,1e5,1e7, true, false]} |
587 |
|
588 |
It shows the number of encodes/decodes per second (JSON::XS uses the |
589 |
functional interface, while JSON::XS/2 uses the OO interface with |
590 |
pretty-printing and hashkey sorting enabled, JSON::XS/3 enables shrink). |
591 |
Higher is better: |
592 |
|
593 |
module | encode | decode | |
594 |
-----------|------------|------------| |
595 |
JSON | 7645.468 | 4208.613 | |
596 |
JSON::DWIW | 40721.398 | 77101.176 | |
597 |
JSON::PC | 65948.176 | 78251.940 | |
598 |
JSON::Syck | 22844.793 | 26479.192 | |
599 |
JSON::XS | 388361.481 | 199728.762 | |
600 |
JSON::XS/2 | 218453.333 | 192399.266 | |
601 |
JSON::XS/3 | 338250.323 | 192399.266 | |
602 |
Storable | 15779.925 | 14169.946 | |
603 |
-----------+------------+------------+ |
604 |
|
605 |
That is, JSON::XS is about five times faster than JSON::DWIW on |
606 |
encoding, about three times faster on decoding, and over fourty times |
607 |
faster than JSON, even with pretty-printing and key sorting. It also |
608 |
compares favourably to Storable for small amounts of data. |
609 |
|
610 |
Using a longer test string (roughly 18KB, generated from Yahoo! Locals |
611 |
search API (http://nanoref.com/yahooapis/mgPdGg): |
612 |
|
613 |
module | encode | decode | |
614 |
-----------|------------|------------| |
615 |
JSON | 254.685 | 37.665 | |
616 |
JSON::DWIW | 843.343 | 1049.731 | |
617 |
JSON::PC | 3602.116 | 2307.352 | |
618 |
JSON::Syck | 505.107 | 787.899 | |
619 |
JSON::XS | 5747.196 | 3690.220 | |
620 |
JSON::XS/2 | 3968.121 | 3676.634 | |
621 |
JSON::XS/3 | 6105.246 | 3662.508 | |
622 |
Storable | 4417.337 | 5285.161 | |
623 |
-----------+------------+------------+ |
624 |
|
625 |
Again, JSON::XS leads by far (except for Storable which non-surprisingly |
626 |
decodes faster). |
627 |
|
628 |
On large strings containing lots of high unicode characters, some |
629 |
modules (such as JSON::PC) seem to decode faster than JSON::XS, but the |
630 |
result will be broken due to missing (or wrong) unicode handling. Others |
631 |
refuse to decode or encode properly, so it was impossible to prepare a |
632 |
fair comparison table for that case. |
633 |
|
634 |
SECURITY CONSIDERATIONS |
635 |
When you are using JSON in a protocol, talking to untrusted potentially |
636 |
hostile creatures requires relatively few measures. |
637 |
|
638 |
First of all, your JSON decoder should be secure, that is, should not |
639 |
have any buffer overflows. Obviously, this module should ensure that and |
640 |
I am trying hard on making that true, but you never know. |
641 |
|
642 |
Second, you need to avoid resource-starving attacks. That means you |
643 |
should limit the size of JSON texts you accept, or make sure then when |
644 |
your resources run out, thats just fine (e.g. by using a separate |
645 |
process that can crash safely). The size of a JSON text in octets or |
646 |
characters is usually a good indication of the size of the resources |
647 |
required to decode it into a Perl structure. |
648 |
|
649 |
Third, JSON::XS recurses using the C stack when decoding objects and |
650 |
arrays. The C stack is a limited resource: for instance, on my amd64 |
651 |
machine with 8MB of stack size I can decode around 180k nested arrays |
652 |
but only 14k nested JSON objects (due to perl itself recursing deeply on |
653 |
croak to free the temporary). If that is exceeded, the program crashes. |
654 |
to be conservative, the default nesting limit is set to 512. If your |
655 |
process has a smaller stack, you should adjust this setting accordingly |
656 |
with the "max_depth" method. |
657 |
|
658 |
And last but least, something else could bomb you that I forgot to think |
659 |
of. In that case, you get to keep the pieces. I am always open for |
660 |
hints, though... |
661 |
|
662 |
If you are using JSON::XS to return packets to consumption by javascript |
663 |
scripts in a browser you should have a look at |
664 |
<http://jpsykes.com/47/practical-csrf-and-json-security> to see wether |
665 |
you are vulnerable to some common attack vectors (which really are |
666 |
browser design bugs, but it is still you who will have to deal with it, |
667 |
as major browser developers care only for features, not about doing |
668 |
security right). |
669 |
|
670 |
BUGS |
671 |
While the goal of this module is to be correct, that unfortunately does |
672 |
not mean its bug-free, only that I think its design is bug-free. It is |
673 |
still relatively early in its development. If you keep reporting bugs |
674 |
they will be fixed swiftly, though. |
675 |
|
676 |
AUTHOR |
677 |
Marc Lehmann <schmorp@schmorp.de> |
678 |
http://home.schmorp.de/ |
679 |
|