| 1 |
NAME |
| 2 |
JSON::XS - JSON serialising/deserialising, done correctly and fast |
| 3 |
|
| 4 |
JSON::XS - 正しくて高速な JSON シリアライザ/デシリアライザ |
| 5 |
(http://fleur.hio.jp/perldoc/mix/lib/JSON/XS.html) |
| 6 |
|
| 7 |
SYNOPSIS |
| 8 |
use JSON::XS; |
| 9 |
|
| 10 |
# exported functions, they croak on error |
| 11 |
# and expect/generate UTF-8 |
| 12 |
|
| 13 |
$utf8_encoded_json_text = encode_json $perl_hash_or_arrayref; |
| 14 |
$perl_hash_or_arrayref = decode_json $utf8_encoded_json_text; |
| 15 |
|
| 16 |
# OO-interface |
| 17 |
|
| 18 |
$coder = JSON::XS->new->ascii->pretty->allow_nonref; |
| 19 |
$pretty_printed_unencoded = $coder->encode ($perl_scalar); |
| 20 |
$perl_scalar = $coder->decode ($unicode_json_text); |
| 21 |
|
| 22 |
# Note that JSON version 2.0 and above will automatically use JSON::XS |
| 23 |
# if available, at virtually no speed overhead either, so you should |
| 24 |
# be able to just: |
| 25 |
|
| 26 |
use JSON; |
| 27 |
|
| 28 |
# and do the same things, except that you have a pure-perl fallback now. |
| 29 |
|
| 30 |
DESCRIPTION |
| 31 |
This module converts Perl data structures to JSON and vice versa. Its |
| 32 |
primary goal is to be *correct* and its secondary goal is to be *fast*. |
| 33 |
To reach the latter goal it was written in C. |
| 34 |
|
| 35 |
Beginning with version 2.0 of the JSON module, when both JSON and |
| 36 |
JSON::XS are installed, then JSON will fall back on JSON::XS (this can |
| 37 |
be overriden) with no overhead due to emulation (by inheritign |
| 38 |
constructor and methods). If JSON::XS is not available, it will fall |
| 39 |
back to the compatible JSON::PP module as backend, so using JSON instead |
| 40 |
of JSON::XS gives you a portable JSON API that can be fast when you need |
| 41 |
and doesn't require a C compiler when that is a problem. |
| 42 |
|
| 43 |
As this is the n-th-something JSON module on CPAN, what was the reason |
| 44 |
to write yet another JSON module? While it seems there are many JSON |
| 45 |
modules, none of them correctly handle all corner cases, and in most |
| 46 |
cases their maintainers are unresponsive, gone missing, or not listening |
| 47 |
to bug reports for other reasons. |
| 48 |
|
| 49 |
See COMPARISON, below, for a comparison to some other JSON modules. |
| 50 |
|
| 51 |
See MAPPING, below, on how JSON::XS maps perl values to JSON values and |
| 52 |
vice versa. |
| 53 |
|
| 54 |
FEATURES |
| 55 |
* correct Unicode handling |
| 56 |
|
| 57 |
This module knows how to handle Unicode, documents how and when it |
| 58 |
does so, and even documents what "correct" means. |
| 59 |
|
| 60 |
* round-trip integrity |
| 61 |
|
| 62 |
When you serialise a perl data structure using only datatypes |
| 63 |
supported by JSON, the deserialised data structure is identical on |
| 64 |
the Perl level. (e.g. the string "2.0" doesn't suddenly become "2" |
| 65 |
just because it looks like a number). There minor *are* exceptions |
| 66 |
to this, read the MAPPING section below to learn about those. |
| 67 |
|
| 68 |
* strict checking of JSON correctness |
| 69 |
|
| 70 |
There is no guessing, no generating of illegal JSON texts by |
| 71 |
default, and only JSON is accepted as input by default (the latter |
| 72 |
is a security feature). |
| 73 |
|
| 74 |
* fast |
| 75 |
|
| 76 |
Compared to other JSON modules and other serialisers such as |
| 77 |
Storable, this module usually compares favourably in terms of speed, |
| 78 |
too. |
| 79 |
|
| 80 |
* simple to use |
| 81 |
|
| 82 |
This module has both a simple functional interface as well as an |
| 83 |
objetc oriented interface interface. |
| 84 |
|
| 85 |
* reasonably versatile output formats |
| 86 |
|
| 87 |
You can choose between the most compact guaranteed-single-line |
| 88 |
format possible (nice for simple line-based protocols), a pure-ascii |
| 89 |
format (for when your transport is not 8-bit clean, still supports |
| 90 |
the whole Unicode range), or a pretty-printed format (for when you |
| 91 |
want to read that stuff). Or you can combine those features in |
| 92 |
whatever way you like. |
| 93 |
|
| 94 |
FUNCTIONAL INTERFACE |
| 95 |
The following convenience methods are provided by this module. They are |
| 96 |
exported by default: |
| 97 |
|
| 98 |
$json_text = encode_json $perl_scalar |
| 99 |
Converts the given Perl data structure to a UTF-8 encoded, binary |
| 100 |
string (that is, the string contains octets only). Croaks on error. |
| 101 |
|
| 102 |
This function call is functionally identical to: |
| 103 |
|
| 104 |
$json_text = JSON::XS->new->utf8->encode ($perl_scalar) |
| 105 |
|
| 106 |
except being faster. |
| 107 |
|
| 108 |
$perl_scalar = decode_json $json_text |
| 109 |
The opposite of "encode_json": expects an UTF-8 (binary) string and |
| 110 |
tries to parse that as an UTF-8 encoded JSON text, returning the |
| 111 |
resulting reference. Croaks on error. |
| 112 |
|
| 113 |
This function call is functionally identical to: |
| 114 |
|
| 115 |
$perl_scalar = JSON::XS->new->utf8->decode ($json_text) |
| 116 |
|
| 117 |
except being faster. |
| 118 |
|
| 119 |
$is_boolean = JSON::XS::is_bool $scalar |
| 120 |
Returns true if the passed scalar represents either JSON::XS::true |
| 121 |
or JSON::XS::false, two constants that act like 1 and 0, |
| 122 |
respectively and are used to represent JSON "true" and "false" |
| 123 |
values in Perl. |
| 124 |
|
| 125 |
See MAPPING, below, for more information on how JSON values are |
| 126 |
mapped to Perl. |
| 127 |
|
| 128 |
A FEW NOTES ON UNICODE AND PERL |
| 129 |
Since this often leads to confusion, here are a few very clear words on |
| 130 |
how Unicode works in Perl, modulo bugs. |
| 131 |
|
| 132 |
1. Perl strings can store characters with ordinal values > 255. |
| 133 |
This enables you to store Unicode characters as single characters in |
| 134 |
a Perl string - very natural. |
| 135 |
|
| 136 |
2. Perl does *not* associate an encoding with your strings. |
| 137 |
... until you force it to, e.g. when matching it against a regex, or |
| 138 |
printing the scalar to a file, in which case Perl either interprets |
| 139 |
your string as locale-encoded text, octets/binary, or as Unicode, |
| 140 |
depending on various settings. In no case is an encoding stored |
| 141 |
together with your data, it is *use* that decides encoding, not any |
| 142 |
magical meta data. |
| 143 |
|
| 144 |
3. The internal utf-8 flag has no meaning with regards to the encoding |
| 145 |
of your string. |
| 146 |
Just ignore that flag unless you debug a Perl bug, a module written |
| 147 |
in XS or want to dive into the internals of perl. Otherwise it will |
| 148 |
only confuse you, as, despite the name, it says nothing about how |
| 149 |
your string is encoded. You can have Unicode strings with that flag |
| 150 |
set, with that flag clear, and you can have binary data with that |
| 151 |
flag set and that flag clear. Other possibilities exist, too. |
| 152 |
|
| 153 |
If you didn't know about that flag, just the better, pretend it |
| 154 |
doesn't exist. |
| 155 |
|
| 156 |
4. A "Unicode String" is simply a string where each character can be |
| 157 |
validly interpreted as a Unicode codepoint. |
| 158 |
If you have UTF-8 encoded data, it is no longer a Unicode string, |
| 159 |
but a Unicode string encoded in UTF-8, giving you a binary string. |
| 160 |
|
| 161 |
5. A string containing "high" (> 255) character values is *not* a UTF-8 |
| 162 |
string. |
| 163 |
It's a fact. Learn to live with it. |
| 164 |
|
| 165 |
I hope this helps :) |
| 166 |
|
| 167 |
OBJECT-ORIENTED INTERFACE |
| 168 |
The object oriented interface lets you configure your own encoding or |
| 169 |
decoding style, within the limits of supported formats. |
| 170 |
|
| 171 |
$json = new JSON::XS |
| 172 |
Creates a new JSON::XS object that can be used to de/encode JSON |
| 173 |
strings. All boolean flags described below are by default |
| 174 |
*disabled*. |
| 175 |
|
| 176 |
The mutators for flags all return the JSON object again and thus |
| 177 |
calls can be chained: |
| 178 |
|
| 179 |
my $json = JSON::XS->new->utf8->space_after->encode ({a => [1,2]}) |
| 180 |
=> {"a": [1, 2]} |
| 181 |
|
| 182 |
$json = $json->ascii ([$enable]) |
| 183 |
$enabled = $json->get_ascii |
| 184 |
If $enable is true (or missing), then the "encode" method will not |
| 185 |
generate characters outside the code range 0..127 (which is ASCII). |
| 186 |
Any Unicode characters outside that range will be escaped using |
| 187 |
either a single \uXXXX (BMP characters) or a double \uHHHH\uLLLLL |
| 188 |
escape sequence, as per RFC4627. The resulting encoded JSON text can |
| 189 |
be treated as a native Unicode string, an ascii-encoded, |
| 190 |
latin1-encoded or UTF-8 encoded string, or any other superset of |
| 191 |
ASCII. |
| 192 |
|
| 193 |
If $enable is false, then the "encode" method will not escape |
| 194 |
Unicode characters unless required by the JSON syntax or other |
| 195 |
flags. This results in a faster and more compact format. |
| 196 |
|
| 197 |
See also the section *ENCODING/CODESET FLAG NOTES* later in this |
| 198 |
document. |
| 199 |
|
| 200 |
The main use for this flag is to produce JSON texts that can be |
| 201 |
transmitted over a 7-bit channel, as the encoded JSON texts will not |
| 202 |
contain any 8 bit characters. |
| 203 |
|
| 204 |
JSON::XS->new->ascii (1)->encode ([chr 0x10401]) |
| 205 |
=> ["\ud801\udc01"] |
| 206 |
|
| 207 |
$json = $json->latin1 ([$enable]) |
| 208 |
$enabled = $json->get_latin1 |
| 209 |
If $enable is true (or missing), then the "encode" method will |
| 210 |
encode the resulting JSON text as latin1 (or iso-8859-1), escaping |
| 211 |
any characters outside the code range 0..255. The resulting string |
| 212 |
can be treated as a latin1-encoded JSON text or a native Unicode |
| 213 |
string. The "decode" method will not be affected in any way by this |
| 214 |
flag, as "decode" by default expects Unicode, which is a strict |
| 215 |
superset of latin1. |
| 216 |
|
| 217 |
If $enable is false, then the "encode" method will not escape |
| 218 |
Unicode characters unless required by the JSON syntax or other |
| 219 |
flags. |
| 220 |
|
| 221 |
See also the section *ENCODING/CODESET FLAG NOTES* later in this |
| 222 |
document. |
| 223 |
|
| 224 |
The main use for this flag is efficiently encoding binary data as |
| 225 |
JSON text, as most octets will not be escaped, resulting in a |
| 226 |
smaller encoded size. The disadvantage is that the resulting JSON |
| 227 |
text is encoded in latin1 (and must correctly be treated as such |
| 228 |
when storing and transferring), a rare encoding for JSON. It is |
| 229 |
therefore most useful when you want to store data structures known |
| 230 |
to contain binary data efficiently in files or databases, not when |
| 231 |
talking to other JSON encoders/decoders. |
| 232 |
|
| 233 |
JSON::XS->new->latin1->encode (["\x{89}\x{abc}"] |
| 234 |
=> ["\x{89}\\u0abc"] # (perl syntax, U+abc escaped, U+89 not) |
| 235 |
|
| 236 |
$json = $json->utf8 ([$enable]) |
| 237 |
$enabled = $json->get_utf8 |
| 238 |
If $enable is true (or missing), then the "encode" method will |
| 239 |
encode the JSON result into UTF-8, as required by many protocols, |
| 240 |
while the "decode" method expects to be handled an UTF-8-encoded |
| 241 |
string. Please note that UTF-8-encoded strings do not contain any |
| 242 |
characters outside the range 0..255, they are thus useful for |
| 243 |
bytewise/binary I/O. In future versions, enabling this option might |
| 244 |
enable autodetection of the UTF-16 and UTF-32 encoding families, as |
| 245 |
described in RFC4627. |
| 246 |
|
| 247 |
If $enable is false, then the "encode" method will return the JSON |
| 248 |
string as a (non-encoded) Unicode string, while "decode" expects |
| 249 |
thus a Unicode string. Any decoding or encoding (e.g. to UTF-8 or |
| 250 |
UTF-16) needs to be done yourself, e.g. using the Encode module. |
| 251 |
|
| 252 |
See also the section *ENCODING/CODESET FLAG NOTES* later in this |
| 253 |
document. |
| 254 |
|
| 255 |
Example, output UTF-16BE-encoded JSON: |
| 256 |
|
| 257 |
use Encode; |
| 258 |
$jsontext = encode "UTF-16BE", JSON::XS->new->encode ($object); |
| 259 |
|
| 260 |
Example, decode UTF-32LE-encoded JSON: |
| 261 |
|
| 262 |
use Encode; |
| 263 |
$object = JSON::XS->new->decode (decode "UTF-32LE", $jsontext); |
| 264 |
|
| 265 |
$json = $json->pretty ([$enable]) |
| 266 |
This enables (or disables) all of the "indent", "space_before" and |
| 267 |
"space_after" (and in the future possibly more) flags in one call to |
| 268 |
generate the most readable (or most compact) form possible. |
| 269 |
|
| 270 |
Example, pretty-print some simple structure: |
| 271 |
|
| 272 |
my $json = JSON::XS->new->pretty(1)->encode ({a => [1,2]}) |
| 273 |
=> |
| 274 |
{ |
| 275 |
"a" : [ |
| 276 |
1, |
| 277 |
2 |
| 278 |
] |
| 279 |
} |
| 280 |
|
| 281 |
$json = $json->indent ([$enable]) |
| 282 |
$enabled = $json->get_indent |
| 283 |
If $enable is true (or missing), then the "encode" method will use a |
| 284 |
multiline format as output, putting every array member or |
| 285 |
object/hash key-value pair into its own line, indenting them |
| 286 |
properly. |
| 287 |
|
| 288 |
If $enable is false, no newlines or indenting will be produced, and |
| 289 |
the resulting JSON text is guaranteed not to contain any "newlines". |
| 290 |
|
| 291 |
This setting has no effect when decoding JSON texts. |
| 292 |
|
| 293 |
$json = $json->space_before ([$enable]) |
| 294 |
$enabled = $json->get_space_before |
| 295 |
If $enable is true (or missing), then the "encode" method will add |
| 296 |
an extra optional space before the ":" separating keys from values |
| 297 |
in JSON objects. |
| 298 |
|
| 299 |
If $enable is false, then the "encode" method will not add any extra |
| 300 |
space at those places. |
| 301 |
|
| 302 |
This setting has no effect when decoding JSON texts. You will also |
| 303 |
most likely combine this setting with "space_after". |
| 304 |
|
| 305 |
Example, space_before enabled, space_after and indent disabled: |
| 306 |
|
| 307 |
{"key" :"value"} |
| 308 |
|
| 309 |
$json = $json->space_after ([$enable]) |
| 310 |
$enabled = $json->get_space_after |
| 311 |
If $enable is true (or missing), then the "encode" method will add |
| 312 |
an extra optional space after the ":" separating keys from values in |
| 313 |
JSON objects and extra whitespace after the "," separating key-value |
| 314 |
pairs and array members. |
| 315 |
|
| 316 |
If $enable is false, then the "encode" method will not add any extra |
| 317 |
space at those places. |
| 318 |
|
| 319 |
This setting has no effect when decoding JSON texts. |
| 320 |
|
| 321 |
Example, space_before and indent disabled, space_after enabled: |
| 322 |
|
| 323 |
{"key": "value"} |
| 324 |
|
| 325 |
$json = $json->relaxed ([$enable]) |
| 326 |
$enabled = $json->get_relaxed |
| 327 |
If $enable is true (or missing), then "decode" will accept some |
| 328 |
extensions to normal JSON syntax (see below). "encode" will not be |
| 329 |
affected in anyway. *Be aware that this option makes you accept |
| 330 |
invalid JSON texts as if they were valid!*. I suggest only to use |
| 331 |
this option to parse application-specific files written by humans |
| 332 |
(configuration files, resource files etc.) |
| 333 |
|
| 334 |
If $enable is false (the default), then "decode" will only accept |
| 335 |
valid JSON texts. |
| 336 |
|
| 337 |
Currently accepted extensions are: |
| 338 |
|
| 339 |
* list items can have an end-comma |
| 340 |
|
| 341 |
JSON *separates* array elements and key-value pairs with commas. |
| 342 |
This can be annoying if you write JSON texts manually and want |
| 343 |
to be able to quickly append elements, so this extension accepts |
| 344 |
comma at the end of such items not just between them: |
| 345 |
|
| 346 |
[ |
| 347 |
1, |
| 348 |
2, <- this comma not normally allowed |
| 349 |
] |
| 350 |
{ |
| 351 |
"k1": "v1", |
| 352 |
"k2": "v2", <- this comma not normally allowed |
| 353 |
} |
| 354 |
|
| 355 |
* shell-style '#'-comments |
| 356 |
|
| 357 |
Whenever JSON allows whitespace, shell-style comments are |
| 358 |
additionally allowed. They are terminated by the first |
| 359 |
carriage-return or line-feed character, after which more |
| 360 |
white-space and comments are allowed. |
| 361 |
|
| 362 |
[ |
| 363 |
1, # this comment not allowed in JSON |
| 364 |
# neither this one... |
| 365 |
] |
| 366 |
|
| 367 |
$json = $json->canonical ([$enable]) |
| 368 |
$enabled = $json->get_canonical |
| 369 |
If $enable is true (or missing), then the "encode" method will |
| 370 |
output JSON objects by sorting their keys. This is adding a |
| 371 |
comparatively high overhead. |
| 372 |
|
| 373 |
If $enable is false, then the "encode" method will output key-value |
| 374 |
pairs in the order Perl stores them (which will likely change |
| 375 |
between runs of the same script). |
| 376 |
|
| 377 |
This option is useful if you want the same data structure to be |
| 378 |
encoded as the same JSON text (given the same overall settings). If |
| 379 |
it is disabled, the same hash might be encoded differently even if |
| 380 |
contains the same data, as key-value pairs have no inherent ordering |
| 381 |
in Perl. |
| 382 |
|
| 383 |
This setting has no effect when decoding JSON texts. |
| 384 |
|
| 385 |
$json = $json->allow_nonref ([$enable]) |
| 386 |
$enabled = $json->get_allow_nonref |
| 387 |
If $enable is true (or missing), then the "encode" method can |
| 388 |
convert a non-reference into its corresponding string, number or |
| 389 |
null JSON value, which is an extension to RFC4627. Likewise, |
| 390 |
"decode" will accept those JSON values instead of croaking. |
| 391 |
|
| 392 |
If $enable is false, then the "encode" method will croak if it isn't |
| 393 |
passed an arrayref or hashref, as JSON texts must either be an |
| 394 |
object or array. Likewise, "decode" will croak if given something |
| 395 |
that is not a JSON object or array. |
| 396 |
|
| 397 |
Example, encode a Perl scalar as JSON value with enabled |
| 398 |
"allow_nonref", resulting in an invalid JSON text: |
| 399 |
|
| 400 |
JSON::XS->new->allow_nonref->encode ("Hello, World!") |
| 401 |
=> "Hello, World!" |
| 402 |
|
| 403 |
$json = $json->allow_blessed ([$enable]) |
| 404 |
$enabled = $json->get_allow_blessed |
| 405 |
If $enable is true (or missing), then the "encode" method will not |
| 406 |
barf when it encounters a blessed reference. Instead, the value of |
| 407 |
the convert_blessed option will decide whether "null" |
| 408 |
("convert_blessed" disabled or no "TO_JSON" method found) or a |
| 409 |
representation of the object ("convert_blessed" enabled and |
| 410 |
"TO_JSON" method found) is being encoded. Has no effect on "decode". |
| 411 |
|
| 412 |
If $enable is false (the default), then "encode" will throw an |
| 413 |
exception when it encounters a blessed object. |
| 414 |
|
| 415 |
$json = $json->convert_blessed ([$enable]) |
| 416 |
$enabled = $json->get_convert_blessed |
| 417 |
If $enable is true (or missing), then "encode", upon encountering a |
| 418 |
blessed object, will check for the availability of the "TO_JSON" |
| 419 |
method on the object's class. If found, it will be called in scalar |
| 420 |
context and the resulting scalar will be encoded instead of the |
| 421 |
object. If no "TO_JSON" method is found, the value of |
| 422 |
"allow_blessed" will decide what to do. |
| 423 |
|
| 424 |
The "TO_JSON" method may safely call die if it wants. If "TO_JSON" |
| 425 |
returns other blessed objects, those will be handled in the same |
| 426 |
way. "TO_JSON" must take care of not causing an endless recursion |
| 427 |
cycle (== crash) in this case. The name of "TO_JSON" was chosen |
| 428 |
because other methods called by the Perl core (== not by the user of |
| 429 |
the object) are usually in upper case letters and to avoid |
| 430 |
collisions with any "to_json" function or method. |
| 431 |
|
| 432 |
This setting does not yet influence "decode" in any way, but in the |
| 433 |
future, global hooks might get installed that influence "decode" and |
| 434 |
are enabled by this setting. |
| 435 |
|
| 436 |
If $enable is false, then the "allow_blessed" setting will decide |
| 437 |
what to do when a blessed object is found. |
| 438 |
|
| 439 |
$json = $json->filter_json_object ([$coderef->($hashref)]) |
| 440 |
When $coderef is specified, it will be called from "decode" each |
| 441 |
time it decodes a JSON object. The only argument is a reference to |
| 442 |
the newly-created hash. If the code references returns a single |
| 443 |
scalar (which need not be a reference), this value (i.e. a copy of |
| 444 |
that scalar to avoid aliasing) is inserted into the deserialised |
| 445 |
data structure. If it returns an empty list (NOTE: *not* "undef", |
| 446 |
which is a valid scalar), the original deserialised hash will be |
| 447 |
inserted. This setting can slow down decoding considerably. |
| 448 |
|
| 449 |
When $coderef is omitted or undefined, any existing callback will be |
| 450 |
removed and "decode" will not change the deserialised hash in any |
| 451 |
way. |
| 452 |
|
| 453 |
Example, convert all JSON objects into the integer 5: |
| 454 |
|
| 455 |
my $js = JSON::XS->new->filter_json_object (sub { 5 }); |
| 456 |
# returns [5] |
| 457 |
$js->decode ('[{}]') |
| 458 |
# throw an exception because allow_nonref is not enabled |
| 459 |
# so a lone 5 is not allowed. |
| 460 |
$js->decode ('{"a":1, "b":2}'); |
| 461 |
|
| 462 |
$json = $json->filter_json_single_key_object ($key [=> |
| 463 |
$coderef->($value)]) |
| 464 |
Works remotely similar to "filter_json_object", but is only called |
| 465 |
for JSON objects having a single key named $key. |
| 466 |
|
| 467 |
This $coderef is called before the one specified via |
| 468 |
"filter_json_object", if any. It gets passed the single value in the |
| 469 |
JSON object. If it returns a single value, it will be inserted into |
| 470 |
the data structure. If it returns nothing (not even "undef" but the |
| 471 |
empty list), the callback from "filter_json_object" will be called |
| 472 |
next, as if no single-key callback were specified. |
| 473 |
|
| 474 |
If $coderef is omitted or undefined, the corresponding callback will |
| 475 |
be disabled. There can only ever be one callback for a given key. |
| 476 |
|
| 477 |
As this callback gets called less often then the |
| 478 |
"filter_json_object" one, decoding speed will not usually suffer as |
| 479 |
much. Therefore, single-key objects make excellent targets to |
| 480 |
serialise Perl objects into, especially as single-key JSON objects |
| 481 |
are as close to the type-tagged value concept as JSON gets (it's |
| 482 |
basically an ID/VALUE tuple). Of course, JSON does not support this |
| 483 |
in any way, so you need to make sure your data never looks like a |
| 484 |
serialised Perl hash. |
| 485 |
|
| 486 |
Typical names for the single object key are "__class_whatever__", or |
| 487 |
"$__dollars_are_rarely_used__$" or "}ugly_brace_placement", or even |
| 488 |
things like "__class_md5sum(classname)__", to reduce the risk of |
| 489 |
clashing with real hashes. |
| 490 |
|
| 491 |
Example, decode JSON objects of the form "{ "__widget__" => <id> }" |
| 492 |
into the corresponding $WIDGET{<id>} object: |
| 493 |
|
| 494 |
# return whatever is in $WIDGET{5}: |
| 495 |
JSON::XS |
| 496 |
->new |
| 497 |
->filter_json_single_key_object (__widget__ => sub { |
| 498 |
$WIDGET{ $_[0] } |
| 499 |
}) |
| 500 |
->decode ('{"__widget__": 5') |
| 501 |
|
| 502 |
# this can be used with a TO_JSON method in some "widget" class |
| 503 |
# for serialisation to json: |
| 504 |
sub WidgetBase::TO_JSON { |
| 505 |
my ($self) = @_; |
| 506 |
|
| 507 |
unless ($self->{id}) { |
| 508 |
$self->{id} = ..get..some..id..; |
| 509 |
$WIDGET{$self->{id}} = $self; |
| 510 |
} |
| 511 |
|
| 512 |
{ __widget__ => $self->{id} } |
| 513 |
} |
| 514 |
|
| 515 |
$json = $json->shrink ([$enable]) |
| 516 |
$enabled = $json->get_shrink |
| 517 |
Perl usually over-allocates memory a bit when allocating space for |
| 518 |
strings. This flag optionally resizes strings generated by either |
| 519 |
"encode" or "decode" to their minimum size possible. This can save |
| 520 |
memory when your JSON texts are either very very long or you have |
| 521 |
many short strings. It will also try to downgrade any strings to |
| 522 |
octet-form if possible: perl stores strings internally either in an |
| 523 |
encoding called UTF-X or in octet-form. The latter cannot store |
| 524 |
everything but uses less space in general (and some buggy Perl or C |
| 525 |
code might even rely on that internal representation being used). |
| 526 |
|
| 527 |
The actual definition of what shrink does might change in future |
| 528 |
versions, but it will always try to save space at the expense of |
| 529 |
time. |
| 530 |
|
| 531 |
If $enable is true (or missing), the string returned by "encode" |
| 532 |
will be shrunk-to-fit, while all strings generated by "decode" will |
| 533 |
also be shrunk-to-fit. |
| 534 |
|
| 535 |
If $enable is false, then the normal perl allocation algorithms are |
| 536 |
used. If you work with your data, then this is likely to be faster. |
| 537 |
|
| 538 |
In the future, this setting might control other things, such as |
| 539 |
converting strings that look like integers or floats into integers |
| 540 |
or floats internally (there is no difference on the Perl level), |
| 541 |
saving space. |
| 542 |
|
| 543 |
$json = $json->max_depth ([$maximum_nesting_depth]) |
| 544 |
$max_depth = $json->get_max_depth |
| 545 |
Sets the maximum nesting level (default 512) accepted while encoding |
| 546 |
or decoding. If the JSON text or Perl data structure has an equal or |
| 547 |
higher nesting level then this limit, then the encoder and decoder |
| 548 |
will stop and croak at that point. |
| 549 |
|
| 550 |
Nesting level is defined by number of hash- or arrayrefs that the |
| 551 |
encoder needs to traverse to reach a given point or the number of |
| 552 |
"{" or "[" characters without their matching closing parenthesis |
| 553 |
crossed to reach a given character in a string. |
| 554 |
|
| 555 |
Setting the maximum depth to one disallows any nesting, so that |
| 556 |
ensures that the object is only a single hash/object or array. |
| 557 |
|
| 558 |
The argument to "max_depth" will be rounded up to the next highest |
| 559 |
power of two. If no argument is given, the highest possible setting |
| 560 |
will be used, which is rarely useful. |
| 561 |
|
| 562 |
See SECURITY CONSIDERATIONS, below, for more info on why this is |
| 563 |
useful. |
| 564 |
|
| 565 |
$json = $json->max_size ([$maximum_string_size]) |
| 566 |
$max_size = $json->get_max_size |
| 567 |
Set the maximum length a JSON text may have (in bytes) where |
| 568 |
decoding is being attempted. The default is 0, meaning no limit. |
| 569 |
When "decode" is called on a string longer then this number of |
| 570 |
characters it will not attempt to decode the string but throw an |
| 571 |
exception. This setting has no effect on "encode" (yet). |
| 572 |
|
| 573 |
The argument to "max_size" will be rounded up to the next highest |
| 574 |
power of two (so may be more than requested). If no argument is |
| 575 |
given, the limit check will be deactivated (same as when 0 is |
| 576 |
specified). |
| 577 |
|
| 578 |
See SECURITY CONSIDERATIONS, below, for more info on why this is |
| 579 |
useful. |
| 580 |
|
| 581 |
$json_text = $json->encode ($perl_scalar) |
| 582 |
Converts the given Perl data structure (a simple scalar or a |
| 583 |
reference to a hash or array) to its JSON representation. Simple |
| 584 |
scalars will be converted into JSON string or number sequences, |
| 585 |
while references to arrays become JSON arrays and references to |
| 586 |
hashes become JSON objects. Undefined Perl values (e.g. "undef") |
| 587 |
become JSON "null" values. Neither "true" nor "false" values will be |
| 588 |
generated. |
| 589 |
|
| 590 |
$perl_scalar = $json->decode ($json_text) |
| 591 |
The opposite of "encode": expects a JSON text and tries to parse it, |
| 592 |
returning the resulting simple scalar or reference. Croaks on error. |
| 593 |
|
| 594 |
JSON numbers and strings become simple Perl scalars. JSON arrays |
| 595 |
become Perl arrayrefs and JSON objects become Perl hashrefs. "true" |
| 596 |
becomes 1, "false" becomes 0 and "null" becomes "undef". |
| 597 |
|
| 598 |
($perl_scalar, $characters) = $json->decode_prefix ($json_text) |
| 599 |
This works like the "decode" method, but instead of raising an |
| 600 |
exception when there is trailing garbage after the first JSON |
| 601 |
object, it will silently stop parsing there and return the number of |
| 602 |
characters consumed so far. |
| 603 |
|
| 604 |
This is useful if your JSON texts are not delimited by an outer |
| 605 |
protocol (which is not the brightest thing to do in the first place) |
| 606 |
and you need to know where the JSON text ends. |
| 607 |
|
| 608 |
JSON::XS->new->decode_prefix ("[1] the tail") |
| 609 |
=> ([], 3) |
| 610 |
|
| 611 |
INCREMENTAL PARSING |
| 612 |
[This section and the API it details is still EXPERIMENTAL] |
| 613 |
|
| 614 |
In some cases, there is the need for incremental parsing of JSON texts. |
| 615 |
While this module always has to keep both JSON text and resulting Perl |
| 616 |
data structure in memory at one time, it does allow you to parse a JSON |
| 617 |
stream incrementally. It does so by accumulating text until it has a |
| 618 |
full JSON object, which it then can decode. This process is similar to |
| 619 |
using "decode_prefix" to see if a full JSON object is available, but is |
| 620 |
much more efficient (JSON::XS will only attempt to parse the JSON text |
| 621 |
once it is sure it has enough text to get a decisive result, using a |
| 622 |
very simple but truly incremental parser). |
| 623 |
|
| 624 |
The following two methods deal with this. |
| 625 |
|
| 626 |
[void, scalar or list context] = $json->incr_parse ([$string]) |
| 627 |
This is the central parsing function. It can both append new text |
| 628 |
and extract objects from the stream accumulated so far (both of |
| 629 |
these functions are optional). |
| 630 |
|
| 631 |
If $string is given, then this string is appended to the already |
| 632 |
existing JSON fragment stored in the $json object. |
| 633 |
|
| 634 |
After that, if the function is called in void context, it will |
| 635 |
simply return without doing anything further. This can be used to |
| 636 |
add more text in as many chunks as you want. |
| 637 |
|
| 638 |
If the method is called in scalar context, then it will try to |
| 639 |
extract exactly *one* JSON object. If that is successful, it will |
| 640 |
return this object, otherwise it will return "undef". If there is a |
| 641 |
parse error, this method will croak just as "decode" would do (one |
| 642 |
can then use "incr_skip" to skip the errornous part). This is the |
| 643 |
most common way of using the method. |
| 644 |
|
| 645 |
And finally, in list context, it will try to extract as many objects |
| 646 |
from the stream as it can find and return them, or the empty list |
| 647 |
otherwise. For this to work, there must be no separators between the |
| 648 |
JSON objects or arrays, instead they must be concatenated |
| 649 |
back-to-back. If an error occurs, an exception will be raised as in |
| 650 |
the scalar context case. Note that in this case, any |
| 651 |
previously-parsed JSON texts will be lost. |
| 652 |
|
| 653 |
$lvalue_string = $json->incr_text |
| 654 |
This method returns the currently stored JSON fragment as an lvalue, |
| 655 |
that is, you can manipulate it. This *only* works when a preceding |
| 656 |
call to "incr_parse" in *scalar context* successfully returned an |
| 657 |
object. Under all other circumstances you must not call this |
| 658 |
function (I mean it. although in simple tests it might actually |
| 659 |
work, it *will* fail under real world conditions). As a special |
| 660 |
exception, you can also call this method before having parsed |
| 661 |
anything. |
| 662 |
|
| 663 |
This function is useful in two cases: a) finding the trailing text |
| 664 |
after a JSON object or b) parsing multiple JSON objects separated by |
| 665 |
non-JSON text (such as commas). |
| 666 |
|
| 667 |
$json->incr_skip |
| 668 |
This will reset the state of the incremental parser and will remove |
| 669 |
the parsed text from the input buffer. This is useful after |
| 670 |
"incr_parse" died, in which case the input buffer and incremental |
| 671 |
parser state is left unchanged, to skip the text parsed so far and |
| 672 |
to reset the parse state. |
| 673 |
|
| 674 |
LIMITATIONS |
| 675 |
All options that affect decoding are supported, except "allow_nonref". |
| 676 |
The reason for this is that it cannot be made to work sensibly: JSON |
| 677 |
objects and arrays are self-delimited, i.e. you can concatenate them |
| 678 |
back to back and still decode them perfectly. This does not hold true |
| 679 |
for JSON numbers, however. |
| 680 |
|
| 681 |
For example, is the string 1 a single JSON number, or is it simply the |
| 682 |
start of 12? Or is 12 a single JSON number, or the concatenation of 1 |
| 683 |
and 2? In neither case you can tell, and this is why JSON::XS takes the |
| 684 |
conservative route and disallows this case. |
| 685 |
|
| 686 |
EXAMPLES |
| 687 |
Some examples will make all this clearer. First, a simple example that |
| 688 |
works similarly to "decode_prefix": We want to decode the JSON object at |
| 689 |
the start of a string and identify the portion after the JSON object: |
| 690 |
|
| 691 |
my $text = "[1,2,3] hello"; |
| 692 |
|
| 693 |
my $json = new JSON::XS; |
| 694 |
|
| 695 |
my $obj = $json->incr_parse ($text) |
| 696 |
or die "expected JSON object or array at beginning of string"; |
| 697 |
|
| 698 |
my $tail = $json->incr_text; |
| 699 |
# $tail now contains " hello" |
| 700 |
|
| 701 |
Easy, isn't it? |
| 702 |
|
| 703 |
Now for a more complicated example: Imagine a hypothetical protocol |
| 704 |
where you read some requests from a TCP stream, and each request is a |
| 705 |
JSON array, without any separation between them (in fact, it is often |
| 706 |
useful to use newlines as "separators", as these get interpreted as |
| 707 |
whitespace at the start of the JSON text, which makes it possible to |
| 708 |
test said protocol with "telnet"...). |
| 709 |
|
| 710 |
Here is how you'd do it (it is trivial to write this in an event-based |
| 711 |
manner): |
| 712 |
|
| 713 |
my $json = new JSON::XS; |
| 714 |
|
| 715 |
# read some data from the socket |
| 716 |
while (sysread $socket, my $buf, 4096) { |
| 717 |
|
| 718 |
# split and decode as many requests as possible |
| 719 |
for my $request ($json->incr_parse ($buf)) { |
| 720 |
# act on the $request |
| 721 |
} |
| 722 |
} |
| 723 |
|
| 724 |
Another complicated example: Assume you have a string with JSON objects |
| 725 |
or arrays, all separated by (optional) comma characters (e.g. "[1],[2], |
| 726 |
[3]"). To parse them, we have to skip the commas between the JSON texts, |
| 727 |
and here is where the lvalue-ness of "incr_text" comes in useful: |
| 728 |
|
| 729 |
my $text = "[1],[2], [3]"; |
| 730 |
my $json = new JSON::XS; |
| 731 |
|
| 732 |
# void context, so no parsing done |
| 733 |
$json->incr_parse ($text); |
| 734 |
|
| 735 |
# now extract as many objects as possible. note the |
| 736 |
# use of scalar context so incr_text can be called. |
| 737 |
while (my $obj = $json->incr_parse) { |
| 738 |
# do something with $obj |
| 739 |
|
| 740 |
# now skip the optional comma |
| 741 |
$json->incr_text =~ s/^ \s* , //x; |
| 742 |
} |
| 743 |
|
| 744 |
Now lets go for a very complex example: Assume that you have a gigantic |
| 745 |
JSON array-of-objects, many gigabytes in size, and you want to parse it, |
| 746 |
but you cannot load it into memory fully (this has actually happened in |
| 747 |
the real world :). |
| 748 |
|
| 749 |
Well, you lost, you have to implement your own JSON parser. But JSON::XS |
| 750 |
can still help you: You implement a (very simple) array parser and let |
| 751 |
JSON decode the array elements, which are all full JSON objects on their |
| 752 |
own (this wouldn't work if the array elements could be JSON numbers, for |
| 753 |
example): |
| 754 |
|
| 755 |
my $json = new JSON::XS; |
| 756 |
|
| 757 |
# open the monster |
| 758 |
open my $fh, "<bigfile.json" |
| 759 |
or die "bigfile: $!"; |
| 760 |
|
| 761 |
# first parse the initial "[" |
| 762 |
for (;;) { |
| 763 |
sysread $fh, my $buf, 65536 |
| 764 |
or die "read error: $!"; |
| 765 |
$json->incr_parse ($buf); # void context, so no parsing |
| 766 |
|
| 767 |
# Exit the loop once we found and removed(!) the initial "[". |
| 768 |
# In essence, we are (ab-)using the $json object as a simple scalar |
| 769 |
# we append data to. |
| 770 |
last if $json->incr_text =~ s/^ \s* \[ //x; |
| 771 |
} |
| 772 |
|
| 773 |
# now we have the skipped the initial "[", so continue |
| 774 |
# parsing all the elements. |
| 775 |
for (;;) { |
| 776 |
# in this loop we read data until we got a single JSON object |
| 777 |
for (;;) { |
| 778 |
if (my $obj = $json->incr_parse) { |
| 779 |
# do something with $obj |
| 780 |
last; |
| 781 |
} |
| 782 |
|
| 783 |
# add more data |
| 784 |
sysread $fh, my $buf, 65536 |
| 785 |
or die "read error: $!"; |
| 786 |
$json->incr_parse ($buf); # void context, so no parsing |
| 787 |
} |
| 788 |
|
| 789 |
# in this loop we read data until we either found and parsed the |
| 790 |
# separating "," between elements, or the final "]" |
| 791 |
for (;;) { |
| 792 |
# first skip whitespace |
| 793 |
$json->incr_text =~ s/^\s*//; |
| 794 |
|
| 795 |
# if we find "]", we are done |
| 796 |
if ($json->incr_text =~ s/^\]//) { |
| 797 |
print "finished.\n"; |
| 798 |
exit; |
| 799 |
} |
| 800 |
|
| 801 |
# if we find ",", we can continue with the next element |
| 802 |
if ($json->incr_text =~ s/^,//) { |
| 803 |
last; |
| 804 |
} |
| 805 |
|
| 806 |
# if we find anything else, we have a parse error! |
| 807 |
if (length $json->incr_text) { |
| 808 |
die "parse error near ", $json->incr_text; |
| 809 |
} |
| 810 |
|
| 811 |
# else add more data |
| 812 |
sysread $fh, my $buf, 65536 |
| 813 |
or die "read error: $!"; |
| 814 |
$json->incr_parse ($buf); # void context, so no parsing |
| 815 |
} |
| 816 |
|
| 817 |
This is a complex example, but most of the complexity comes from the |
| 818 |
fact that we are trying to be correct (bear with me if I am wrong, I |
| 819 |
never ran the above example :). |
| 820 |
|
| 821 |
MAPPING |
| 822 |
This section describes how JSON::XS maps Perl values to JSON values and |
| 823 |
vice versa. These mappings are designed to "do the right thing" in most |
| 824 |
circumstances automatically, preserving round-tripping characteristics |
| 825 |
(what you put in comes out as something equivalent). |
| 826 |
|
| 827 |
For the more enlightened: note that in the following descriptions, |
| 828 |
lowercase *perl* refers to the Perl interpreter, while uppercase *Perl* |
| 829 |
refers to the abstract Perl language itself. |
| 830 |
|
| 831 |
JSON -> PERL |
| 832 |
object |
| 833 |
A JSON object becomes a reference to a hash in Perl. No ordering of |
| 834 |
object keys is preserved (JSON does not preserve object key ordering |
| 835 |
itself). |
| 836 |
|
| 837 |
array |
| 838 |
A JSON array becomes a reference to an array in Perl. |
| 839 |
|
| 840 |
string |
| 841 |
A JSON string becomes a string scalar in Perl - Unicode codepoints |
| 842 |
in JSON are represented by the same codepoints in the Perl string, |
| 843 |
so no manual decoding is necessary. |
| 844 |
|
| 845 |
number |
| 846 |
A JSON number becomes either an integer, numeric (floating point) or |
| 847 |
string scalar in perl, depending on its range and any fractional |
| 848 |
parts. On the Perl level, there is no difference between those as |
| 849 |
Perl handles all the conversion details, but an integer may take |
| 850 |
slightly less memory and might represent more values exactly than |
| 851 |
floating point numbers. |
| 852 |
|
| 853 |
If the number consists of digits only, JSON::XS will try to |
| 854 |
represent it as an integer value. If that fails, it will try to |
| 855 |
represent it as a numeric (floating point) value if that is possible |
| 856 |
without loss of precision. Otherwise it will preserve the number as |
| 857 |
a string value (in which case you lose roundtripping ability, as the |
| 858 |
JSON number will be re-encoded toa JSON string). |
| 859 |
|
| 860 |
Numbers containing a fractional or exponential part will always be |
| 861 |
represented as numeric (floating point) values, possibly at a loss |
| 862 |
of precision (in which case you might lose perfect roundtripping |
| 863 |
ability, but the JSON number will still be re-encoded as a JSON |
| 864 |
number). |
| 865 |
|
| 866 |
true, false |
| 867 |
These JSON atoms become "JSON::XS::true" and "JSON::XS::false", |
| 868 |
respectively. They are overloaded to act almost exactly like the |
| 869 |
numbers 1 and 0. You can check whether a scalar is a JSON boolean by |
| 870 |
using the "JSON::XS::is_bool" function. |
| 871 |
|
| 872 |
null |
| 873 |
A JSON null atom becomes "undef" in Perl. |
| 874 |
|
| 875 |
PERL -> JSON |
| 876 |
The mapping from Perl to JSON is slightly more difficult, as Perl is a |
| 877 |
truly typeless language, so we can only guess which JSON type is meant |
| 878 |
by a Perl value. |
| 879 |
|
| 880 |
hash references |
| 881 |
Perl hash references become JSON objects. As there is no inherent |
| 882 |
ordering in hash keys (or JSON objects), they will usually be |
| 883 |
encoded in a pseudo-random order that can change between runs of the |
| 884 |
same program but stays generally the same within a single run of a |
| 885 |
program. JSON::XS can optionally sort the hash keys (determined by |
| 886 |
the *canonical* flag), so the same datastructure will serialise to |
| 887 |
the same JSON text (given same settings and version of JSON::XS), |
| 888 |
but this incurs a runtime overhead and is only rarely useful, e.g. |
| 889 |
when you want to compare some JSON text against another for |
| 890 |
equality. |
| 891 |
|
| 892 |
array references |
| 893 |
Perl array references become JSON arrays. |
| 894 |
|
| 895 |
other references |
| 896 |
Other unblessed references are generally not allowed and will cause |
| 897 |
an exception to be thrown, except for references to the integers 0 |
| 898 |
and 1, which get turned into "false" and "true" atoms in JSON. You |
| 899 |
can also use "JSON::XS::false" and "JSON::XS::true" to improve |
| 900 |
readability. |
| 901 |
|
| 902 |
encode_json [\0,JSON::XS::true] # yields [false,true] |
| 903 |
|
| 904 |
JSON::XS::true, JSON::XS::false |
| 905 |
These special values become JSON true and JSON false values, |
| 906 |
respectively. You can also use "\1" and "\0" directly if you want. |
| 907 |
|
| 908 |
blessed objects |
| 909 |
Blessed objects are not directly representable in JSON. See the |
| 910 |
"allow_blessed" and "convert_blessed" methods on various options on |
| 911 |
how to deal with this: basically, you can choose between throwing an |
| 912 |
exception, encoding the reference as if it weren't blessed, or |
| 913 |
provide your own serialiser method. |
| 914 |
|
| 915 |
simple scalars |
| 916 |
Simple Perl scalars (any scalar that is not a reference) are the |
| 917 |
most difficult objects to encode: JSON::XS will encode undefined |
| 918 |
scalars as JSON "null" values, scalars that have last been used in a |
| 919 |
string context before encoding as JSON strings, and anything else as |
| 920 |
number value: |
| 921 |
|
| 922 |
# dump as number |
| 923 |
encode_json [2] # yields [2] |
| 924 |
encode_json [-3.0e17] # yields [-3e+17] |
| 925 |
my $value = 5; encode_json [$value] # yields [5] |
| 926 |
|
| 927 |
# used as string, so dump as string |
| 928 |
print $value; |
| 929 |
encode_json [$value] # yields ["5"] |
| 930 |
|
| 931 |
# undef becomes null |
| 932 |
encode_json [undef] # yields [null] |
| 933 |
|
| 934 |
You can force the type to be a JSON string by stringifying it: |
| 935 |
|
| 936 |
my $x = 3.1; # some variable containing a number |
| 937 |
"$x"; # stringified |
| 938 |
$x .= ""; # another, more awkward way to stringify |
| 939 |
print $x; # perl does it for you, too, quite often |
| 940 |
|
| 941 |
You can force the type to be a JSON number by numifying it: |
| 942 |
|
| 943 |
my $x = "3"; # some variable containing a string |
| 944 |
$x += 0; # numify it, ensuring it will be dumped as a number |
| 945 |
$x *= 1; # same thing, the choice is yours. |
| 946 |
|
| 947 |
You can not currently force the type in other, less obscure, ways. |
| 948 |
Tell me if you need this capability (but don't forget to explain why |
| 949 |
it's needed :). |
| 950 |
|
| 951 |
ENCODING/CODESET FLAG NOTES |
| 952 |
The interested reader might have seen a number of flags that signify |
| 953 |
encodings or codesets - "utf8", "latin1" and "ascii". There seems to be |
| 954 |
some confusion on what these do, so here is a short comparison: |
| 955 |
|
| 956 |
"utf8" controls whether the JSON text created by "encode" (and expected |
| 957 |
by "decode") is UTF-8 encoded or not, while "latin1" and "ascii" only |
| 958 |
control whether "encode" escapes character values outside their |
| 959 |
respective codeset range. Neither of these flags conflict with each |
| 960 |
other, although some combinations make less sense than others. |
| 961 |
|
| 962 |
Care has been taken to make all flags symmetrical with respect to |
| 963 |
"encode" and "decode", that is, texts encoded with any combination of |
| 964 |
these flag values will be correctly decoded when the same flags are used |
| 965 |
- in general, if you use different flag settings while encoding vs. when |
| 966 |
decoding you likely have a bug somewhere. |
| 967 |
|
| 968 |
Below comes a verbose discussion of these flags. Note that a "codeset" |
| 969 |
is simply an abstract set of character-codepoint pairs, while an |
| 970 |
encoding takes those codepoint numbers and *encodes* them, in our case |
| 971 |
into octets. Unicode is (among other things) a codeset, UTF-8 is an |
| 972 |
encoding, and ISO-8859-1 (= latin 1) and ASCII are both codesets *and* |
| 973 |
encodings at the same time, which can be confusing. |
| 974 |
|
| 975 |
"utf8" flag disabled |
| 976 |
When "utf8" is disabled (the default), then "encode"/"decode" |
| 977 |
generate and expect Unicode strings, that is, characters with high |
| 978 |
ordinal Unicode values (> 255) will be encoded as such characters, |
| 979 |
and likewise such characters are decoded as-is, no canges to them |
| 980 |
will be done, except "(re-)interpreting" them as Unicode codepoints |
| 981 |
or Unicode characters, respectively (to Perl, these are the same |
| 982 |
thing in strings unless you do funny/weird/dumb stuff). |
| 983 |
|
| 984 |
This is useful when you want to do the encoding yourself (e.g. when |
| 985 |
you want to have UTF-16 encoded JSON texts) or when some other layer |
| 986 |
does the encoding for you (for example, when printing to a terminal |
| 987 |
using a filehandle that transparently encodes to UTF-8 you certainly |
| 988 |
do NOT want to UTF-8 encode your data first and have Perl encode it |
| 989 |
another time). |
| 990 |
|
| 991 |
"utf8" flag enabled |
| 992 |
If the "utf8"-flag is enabled, "encode"/"decode" will encode all |
| 993 |
characters using the corresponding UTF-8 multi-byte sequence, and |
| 994 |
will expect your input strings to be encoded as UTF-8, that is, no |
| 995 |
"character" of the input string must have any value > 255, as UTF-8 |
| 996 |
does not allow that. |
| 997 |
|
| 998 |
The "utf8" flag therefore switches between two modes: disabled means |
| 999 |
you will get a Unicode string in Perl, enabled means you get an |
| 1000 |
UTF-8 encoded octet/binary string in Perl. |
| 1001 |
|
| 1002 |
"latin1" or "ascii" flags enabled |
| 1003 |
With "latin1" (or "ascii") enabled, "encode" will escape characters |
| 1004 |
with ordinal values > 255 (> 127 with "ascii") and encode the |
| 1005 |
remaining characters as specified by the "utf8" flag. |
| 1006 |
|
| 1007 |
If "utf8" is disabled, then the result is also correctly encoded in |
| 1008 |
those character sets (as both are proper subsets of Unicode, meaning |
| 1009 |
that a Unicode string with all character values < 256 is the same |
| 1010 |
thing as a ISO-8859-1 string, and a Unicode string with all |
| 1011 |
character values < 128 is the same thing as an ASCII string in |
| 1012 |
Perl). |
| 1013 |
|
| 1014 |
If "utf8" is enabled, you still get a correct UTF-8-encoded string, |
| 1015 |
regardless of these flags, just some more characters will be escaped |
| 1016 |
using "\uXXXX" then before. |
| 1017 |
|
| 1018 |
Note that ISO-8859-1-*encoded* strings are not compatible with UTF-8 |
| 1019 |
encoding, while ASCII-encoded strings are. That is because the |
| 1020 |
ISO-8859-1 encoding is NOT a subset of UTF-8 (despite the ISO-8859-1 |
| 1021 |
*codeset* being a subset of Unicode), while ASCII is. |
| 1022 |
|
| 1023 |
Surprisingly, "decode" will ignore these flags and so treat all |
| 1024 |
input values as governed by the "utf8" flag. If it is disabled, this |
| 1025 |
allows you to decode ISO-8859-1- and ASCII-encoded strings, as both |
| 1026 |
strict subsets of Unicode. If it is enabled, you can correctly |
| 1027 |
decode UTF-8 encoded strings. |
| 1028 |
|
| 1029 |
So neither "latin1" nor "ascii" are incompatible with the "utf8" |
| 1030 |
flag - they only govern when the JSON output engine escapes a |
| 1031 |
character or not. |
| 1032 |
|
| 1033 |
The main use for "latin1" is to relatively efficiently store binary |
| 1034 |
data as JSON, at the expense of breaking compatibility with most |
| 1035 |
JSON decoders. |
| 1036 |
|
| 1037 |
The main use for "ascii" is to force the output to not contain |
| 1038 |
characters with values > 127, which means you can interpret the |
| 1039 |
resulting string as UTF-8, ISO-8859-1, ASCII, KOI8-R or most about |
| 1040 |
any character set and 8-bit-encoding, and still get the same data |
| 1041 |
structure back. This is useful when your channel for JSON transfer |
| 1042 |
is not 8-bit clean or the encoding might be mangled in between (e.g. |
| 1043 |
in mail), and works because ASCII is a proper subset of most 8-bit |
| 1044 |
and multibyte encodings in use in the world. |
| 1045 |
|
| 1046 |
JSON and YAML |
| 1047 |
You often hear that JSON is a subset of YAML. This is, however, a mass |
| 1048 |
hysteria(*) and very far from the truth (as of the time of this |
| 1049 |
writing), so let me state it clearly: *in general, there is no way to |
| 1050 |
configure JSON::XS to output a data structure as valid YAML* that works |
| 1051 |
in all cases. |
| 1052 |
|
| 1053 |
If you really must use JSON::XS to generate YAML, you should use this |
| 1054 |
algorithm (subject to change in future versions): |
| 1055 |
|
| 1056 |
my $to_yaml = JSON::XS->new->utf8->space_after (1); |
| 1057 |
my $yaml = $to_yaml->encode ($ref) . "\n"; |
| 1058 |
|
| 1059 |
This will *usually* generate JSON texts that also parse as valid YAML. |
| 1060 |
Please note that YAML has hardcoded limits on (simple) object key |
| 1061 |
lengths that JSON doesn't have and also has different and incompatible |
| 1062 |
unicode handling, so you should make sure that your hash keys are |
| 1063 |
noticeably shorter than the 1024 "stream characters" YAML allows and |
| 1064 |
that you do not have characters with codepoint values outside the |
| 1065 |
Unicode BMP (basic multilingual page). YAML also does not allow "\/" |
| 1066 |
sequences in strings (which JSON::XS does not *currently* generate, but |
| 1067 |
other JSON generators might). |
| 1068 |
|
| 1069 |
There might be other incompatibilities that I am not aware of (or the |
| 1070 |
YAML specification has been changed yet again - it does so quite often). |
| 1071 |
In general you should not try to generate YAML with a JSON generator or |
| 1072 |
vice versa, or try to parse JSON with a YAML parser or vice versa: |
| 1073 |
chances are high that you will run into severe interoperability problems |
| 1074 |
when you least expect it. |
| 1075 |
|
| 1076 |
(*) I have been pressured multiple times by Brian Ingerson (one of the |
| 1077 |
authors of the YAML specification) to remove this paragraph, despite |
| 1078 |
him acknowledging that the actual incompatibilities exist. As I was |
| 1079 |
personally bitten by this "JSON is YAML" lie, I refused and said I |
| 1080 |
will continue to educate people about these issues, so others do not |
| 1081 |
run into the same problem again and again. After this, Brian called |
| 1082 |
me a (quote)*complete and worthless idiot*(unquote). |
| 1083 |
|
| 1084 |
In my opinion, instead of pressuring and insulting people who |
| 1085 |
actually clarify issues with YAML and the wrong statements of some |
| 1086 |
of its proponents, I would kindly suggest reading the JSON spec |
| 1087 |
(which is not that difficult or long) and finally make YAML |
| 1088 |
compatible to it, and educating users about the changes, instead of |
| 1089 |
spreading lies about the real compatibility for many *years* and |
| 1090 |
trying to silence people who point out that it isn't true. |
| 1091 |
|
| 1092 |
SPEED |
| 1093 |
It seems that JSON::XS is surprisingly fast, as shown in the following |
| 1094 |
tables. They have been generated with the help of the "eg/bench" program |
| 1095 |
in the JSON::XS distribution, to make it easy to compare on your own |
| 1096 |
system. |
| 1097 |
|
| 1098 |
First comes a comparison between various modules using a very short |
| 1099 |
single-line JSON string (also available at |
| 1100 |
<http://dist.schmorp.de/misc/json/short.json>). |
| 1101 |
|
| 1102 |
{"method": "handleMessage", "params": ["user1", "we were just talking"], \ |
| 1103 |
"id": null, "array":[1,11,234,-5,1e5,1e7, true, false]} |
| 1104 |
|
| 1105 |
It shows the number of encodes/decodes per second (JSON::XS uses the |
| 1106 |
functional interface, while JSON::XS/2 uses the OO interface with |
| 1107 |
pretty-printing and hashkey sorting enabled, JSON::XS/3 enables shrink). |
| 1108 |
Higher is better: |
| 1109 |
|
| 1110 |
module | encode | decode | |
| 1111 |
-----------|------------|------------| |
| 1112 |
JSON 1.x | 4990.842 | 4088.813 | |
| 1113 |
JSON::DWIW | 51653.990 | 71575.154 | |
| 1114 |
JSON::PC | 65948.176 | 74631.744 | |
| 1115 |
JSON::PP | 8931.652 | 3817.168 | |
| 1116 |
JSON::Syck | 24877.248 | 27776.848 | |
| 1117 |
JSON::XS | 388361.481 | 227951.304 | |
| 1118 |
JSON::XS/2 | 227951.304 | 218453.333 | |
| 1119 |
JSON::XS/3 | 338250.323 | 218453.333 | |
| 1120 |
Storable | 16500.016 | 135300.129 | |
| 1121 |
-----------+------------+------------+ |
| 1122 |
|
| 1123 |
That is, JSON::XS is about five times faster than JSON::DWIW on |
| 1124 |
encoding, about three times faster on decoding, and over forty times |
| 1125 |
faster than JSON, even with pretty-printing and key sorting. It also |
| 1126 |
compares favourably to Storable for small amounts of data. |
| 1127 |
|
| 1128 |
Using a longer test string (roughly 18KB, generated from Yahoo! Locals |
| 1129 |
search API (<http://dist.schmorp.de/misc/json/long.json>). |
| 1130 |
|
| 1131 |
module | encode | decode | |
| 1132 |
-----------|------------|------------| |
| 1133 |
JSON 1.x | 55.260 | 34.971 | |
| 1134 |
JSON::DWIW | 825.228 | 1082.513 | |
| 1135 |
JSON::PC | 3571.444 | 2394.829 | |
| 1136 |
JSON::PP | 210.987 | 32.574 | |
| 1137 |
JSON::Syck | 552.551 | 787.544 | |
| 1138 |
JSON::XS | 5780.463 | 4854.519 | |
| 1139 |
JSON::XS/2 | 3869.998 | 4798.975 | |
| 1140 |
JSON::XS/3 | 5862.880 | 4798.975 | |
| 1141 |
Storable | 4445.002 | 5235.027 | |
| 1142 |
-----------+------------+------------+ |
| 1143 |
|
| 1144 |
Again, JSON::XS leads by far (except for Storable which non-surprisingly |
| 1145 |
decodes faster). |
| 1146 |
|
| 1147 |
On large strings containing lots of high Unicode characters, some |
| 1148 |
modules (such as JSON::PC) seem to decode faster than JSON::XS, but the |
| 1149 |
result will be broken due to missing (or wrong) Unicode handling. Others |
| 1150 |
refuse to decode or encode properly, so it was impossible to prepare a |
| 1151 |
fair comparison table for that case. |
| 1152 |
|
| 1153 |
SECURITY CONSIDERATIONS |
| 1154 |
When you are using JSON in a protocol, talking to untrusted potentially |
| 1155 |
hostile creatures requires relatively few measures. |
| 1156 |
|
| 1157 |
First of all, your JSON decoder should be secure, that is, should not |
| 1158 |
have any buffer overflows. Obviously, this module should ensure that and |
| 1159 |
I am trying hard on making that true, but you never know. |
| 1160 |
|
| 1161 |
Second, you need to avoid resource-starving attacks. That means you |
| 1162 |
should limit the size of JSON texts you accept, or make sure then when |
| 1163 |
your resources run out, that's just fine (e.g. by using a separate |
| 1164 |
process that can crash safely). The size of a JSON text in octets or |
| 1165 |
characters is usually a good indication of the size of the resources |
| 1166 |
required to decode it into a Perl structure. While JSON::XS can check |
| 1167 |
the size of the JSON text, it might be too late when you already have it |
| 1168 |
in memory, so you might want to check the size before you accept the |
| 1169 |
string. |
| 1170 |
|
| 1171 |
Third, JSON::XS recurses using the C stack when decoding objects and |
| 1172 |
arrays. The C stack is a limited resource: for instance, on my amd64 |
| 1173 |
machine with 8MB of stack size I can decode around 180k nested arrays |
| 1174 |
but only 14k nested JSON objects (due to perl itself recursing deeply on |
| 1175 |
croak to free the temporary). If that is exceeded, the program crashes. |
| 1176 |
To be conservative, the default nesting limit is set to 512. If your |
| 1177 |
process has a smaller stack, you should adjust this setting accordingly |
| 1178 |
with the "max_depth" method. |
| 1179 |
|
| 1180 |
Something else could bomb you, too, that I forgot to think of. In that |
| 1181 |
case, you get to keep the pieces. I am always open for hints, though... |
| 1182 |
|
| 1183 |
Also keep in mind that JSON::XS might leak contents of your Perl data |
| 1184 |
structures in its error messages, so when you serialise sensitive |
| 1185 |
information you might want to make sure that exceptions thrown by |
| 1186 |
JSON::XS will not end up in front of untrusted eyes. |
| 1187 |
|
| 1188 |
If you are using JSON::XS to return packets to consumption by JavaScript |
| 1189 |
scripts in a browser you should have a look at |
| 1190 |
<http://jpsykes.com/47/practical-csrf-and-json-security> to see whether |
| 1191 |
you are vulnerable to some common attack vectors (which really are |
| 1192 |
browser design bugs, but it is still you who will have to deal with it, |
| 1193 |
as major browser developers care only for features, not about getting |
| 1194 |
security right). |
| 1195 |
|
| 1196 |
THREADS |
| 1197 |
This module is *not* guaranteed to be thread safe and there are no plans |
| 1198 |
to change this until Perl gets thread support (as opposed to the |
| 1199 |
horribly slow so-called "threads" which are simply slow and bloated |
| 1200 |
process simulations - use fork, it's *much* faster, cheaper, better). |
| 1201 |
|
| 1202 |
(It might actually work, but you have been warned). |
| 1203 |
|
| 1204 |
BUGS |
| 1205 |
While the goal of this module is to be correct, that unfortunately does |
| 1206 |
not mean it's bug-free, only that I think its design is bug-free. It is |
| 1207 |
still relatively early in its development. If you keep reporting bugs |
| 1208 |
they will be fixed swiftly, though. |
| 1209 |
|
| 1210 |
Please refrain from using rt.cpan.org or any other bug reporting |
| 1211 |
service. I put the contact address into my modules for a reason. |
| 1212 |
|
| 1213 |
SEE ALSO |
| 1214 |
The json_xs command line utility for quick experiments. |
| 1215 |
|
| 1216 |
AUTHOR |
| 1217 |
Marc Lehmann <schmorp@schmorp.de> |
| 1218 |
http://home.schmorp.de/ |
| 1219 |
|
