1XS(3) User Contributed Perl Documentation XS(3)
2
6 Cpanel::JSON::XS - cPanel fork of JSON::XS, fast and correct
7 serializing
8
10 use Cpanel::JSON::XS;
11 # exported functions, they croak on error
13 # and expect/generate UTF-8
14 $utf8_encoded_json_text = encode_json $perl_hash_or_arrayref;
16 $perl_hash_or_arrayref = decode_json $utf8_encoded_json_text;
17 # OO-interface
19 $coder = Cpanel::JSON::XS->new->ascii->pretty->allow_nonref;
21 $pretty_printed_unencoded = $coder->encode ($perl_scalar);
22 $perl_scalar = $coder->decode ($unicode_json_text);
23 # Note that 5.6 misses most smart utf8 and encoding functionalities
25 # of newer releases.
26 # Note that L<JSON::MaybeXS> will automatically use Cpanel::JSON::XS
28 # if available, at virtually no speed overhead either, so you should
29 # be able to just:
30 use JSON::MaybeXS;
32 # and do the same things, except that you have a pure-perl fallback now.
34 Note that this module will be replaced by a new JSON::Safe module soon,
36 with the same API just guaranteed safe defaults.
37
39 This module converts Perl data structures to JSON and vice versa. Its
40 primary goal is to be correct and its secondary goal is to be fast. To
41 reach the latter goal it was written in C.
42 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
47 listening to bug reports for other reasons.
48 See below for the cPanel fork.
50 See MAPPING, below, on how Cpanel::JSON::XS maps perl values to JSON
52 values and vice versa.
53 FEATURES
55 • correct Unicode handling
56 This module knows how to handle Unicode with Perl version higher
58 than 5.8.5, documents how and when it does so, and even documents
59 what "correct" means.
60 • round-trip integrity
62 When you serialize a perl data structure using only data types
64 supported by JSON and Perl, the deserialized data structure is
65 identical on the Perl level. (e.g. the string "2.0" doesn't
66 suddenly become "2" just because it looks like a number). There are
67 minor exceptions to this, read the MAPPING section below to learn
68 about those.
69 • strict checking of JSON correctness
71 There is no guessing, no generating of illegal JSON texts by
73 default, and only JSON is accepted as input by default. the latter
74 is a security feature.
75 • fast
77 Compared to other JSON modules and other serializers such as
79 Storable, this module usually compares favourably in terms of
80 speed, too.
81 • simple to use
83 This module has both a simple functional interface as well as an
85 object oriented interface.
86 • reasonably versatile output formats
88 You can choose between the most compact guaranteed-single-line
90 format possible (nice for simple line-based protocols), a pure-
91 ASCII format (for when your transport is not 8-bit clean, still
92 supports the whole Unicode range), or a pretty-printed format (for
93 when you want to read that stuff). Or you can combine those
94 features in whatever way you like.
95 cPanel fork
97 Since the original author MLEHMANN has no public bugtracker, this
98 cPanel fork sits now on github.
99 src repo: <https://github.com/rurban/Cpanel-JSON-XS> original:
101 <http://cvs.schmorp.de/JSON-XS/>
102 RT: <https://github.com/rurban/Cpanel-JSON-XS/issues> or
104 <https://rt.cpan.org/Public/Dist/Display.html?Queue=Cpanel-JSON-XS>
105 Changes to JSON::XS
107 - stricter decode_json() as documented. non-refs are disallowed.
109 added a 2nd optional argument. decode() honors now allow_nonref.
110 - fixed encode of numbers for dual-vars. Different string
112 representations are preserved, but numbers with temporary strings
113 which represent the same number are here treated as numbers, not
114 strings. Cpanel::JSON::XS is a bit slower, but preserves numeric
115 types better.
116 - numbers ending with .0 stray numbers, are not converted to
118 integers. [#63] dual-vars which are represented as number not
119 integer (42+"bar" != 5.8.9) are now encoded as number (=> 42.0)
120 because internally it's now a NOK type. However !!1 which is
121 wrongly encoded in 5.8 as "1"/1.0 is still represented as integer.
122 - different handling of inf/nan. Default now to null, optionally with
124 stringify_infnan() to "inf"/"nan". [#28, #32]
125 - added "binary" extension, non-JSON and non JSON parsable, allows
127 "\xNN" and "\NNN" sequences.
128 - 5.6.2 support; sacrificing some utf8 features (assuming bytes
130 all-over), no multi-byte unicode characters with 5.6.
131 - interop for true/false overloading. JSON::XS, JSON::PP and Mojo::JSON
133 representations for booleans are accepted and JSON::XS accepts
134 Cpanel::JSON::XS booleans [#13, #37]
135 Fixed overloading of booleans. Cpanel::JSON::XS::true stringifies
136 again
137 to "1", not "true", analog to all other JSON modules.
138 - native boolean mapping of yes and no to true and false, as in
140 YAML::XS.
141 In perl "!0" is yes, "!1" is no.
142 The JSON value true maps to 1, false maps to 0. [#39]
143 - support arbitrary stringification with encode, with convert_blessed
145 and allow_blessed.
146 - ithread support. Cpanel::JSON::XS is thread-safe, JSON::XS not
148 - is_bool can be called as method, JSON::XS::is_bool not.
150 - performance optimizations for threaded Perls
152 - relaxed mode, allowing many popular extensions
154 - additional fixes for:
156 - [cpan #88061] AIX atof without USE_LONG_DOUBLE
158 - #10 unshare_hek crash
160 - #7, #29 avoid re-blessing where possible. It fails in JSON::XS for
162 READONLY values, i.e. restricted hashes.
163 - #41 overloading of booleans, use the object not the reference.
165 - #62 -Dusequadmath conversion and no SEGV.
167 - #72 parsing of values followed \0, like 1\0 does fail.
169 - #72 parsing of illegal unicode or non-unicode characters.
171 - #96 locale-insensitive numeric conversion.
173 - #154 numeric conversion fixed since 5.22, using the same strtold as perl5.
175 - #167 sort tied hashes with canonical.
177 - public maintenance and bugtracker
179 - use ppport.h, sanify XS.xs comment styles, harness C coding style
181 - common::sense is optional. When available it is not used in the
183 published production module, just during development and testing.
184 - extended testsuite, passes all http://seriot.ch/parsing_json.html
186 tests. In fact it is the only know JSON decoder which does so,
187 while also being the fastest.
188 - support many more options and methods from JSON::PP:
190 stringify_infnan, allow_unknown, allow_stringify, allow_barekey,
191 encode_stringify, allow_bignum, allow_singlequote, sort_by
192 (partially), escape_slash, convert_blessed, ... optional
193 decode_json(, allow_nonref) arg.
194 relaxed implements allow_dupkeys.
195 - support all 5 unicode BOM's: UTF-8, UTF-16LE, UTF-16BE, UTF-32LE,
197 UTF-32BE, encoding internally to UTF-8.
198
200 The following convenience methods are provided by this module. They are
201 exported by default:
202 $json_text = encode_json $perl_scalar, [json_type]
204 Converts the given Perl data structure to a UTF-8 encoded, binary
205 string (that is, the string contains octets only). Croaks on error.
206 This function call is functionally identical to:
208 $json_text = Cpanel::JSON::XS->new->utf8->encode ($perl_scalar, $json_type)
210 Except being faster.
212 For the type argument see Cpanel::JSON::XS::Type.
214 $perl_scalar = decode_json $json_text [, $allow_nonref [, my $json_type
216 ] ]
217 The opposite of "encode_json": expects an UTF-8 (binary) string of
218 an json reference and tries to parse that as an UTF-8 encoded JSON
219 text, returning the resulting reference. Croaks on error.
220 This function call is functionally identical to:
222 $perl_scalar = Cpanel::JSON::XS->new->utf8->decode ($json_text, $json_type)
224 except being faster.
226 Note that older decode_json versions in Cpanel::JSON::XS older than
228 3.0116 and JSON::XS did not set allow_nonref but allowed them due
229 to a bug in the decoder.
230 If the new 2nd optional $allow_nonref argument is set and not
232 false, the "allow_nonref" option will be set and the function will
233 act is described as in the relaxed RFC 7159 allowing all values
234 such as objects, arrays, strings, numbers, "null", "true", and
235 "false". See ""OLD" VS. "NEW" JSON (RFC 4627 VS. RFC 7159)" below,
236 why you don't want to do that.
237 For the 3rd optional type argument see Cpanel::JSON::XS::Type.
239 $is_boolean = Cpanel::JSON::XS::is_bool $scalar
241 Returns true if the passed scalar represents either
242 "JSON::PP::true" or "JSON::PP::false", two constants that act like
243 1 and 0, respectively and are used to represent JSON "true" and
244 "false" values in Perl. (Also recognizes the booleans produced by
245 JSON::XS.)
246 See MAPPING, below, for more information on how JSON values are
248 mapped to Perl.
249
251 from_json
252 from_json has been renamed to decode_json
253 to_json
255 to_json has been renamed to encode_json
256
258 Since this often leads to confusion, here are a few very clear words on
259 how Unicode works in Perl, modulo bugs.
260 1. Perl strings can store characters with ordinal values > 255.
262 This enables you to store Unicode characters as single characters
263 in a Perl string - very natural.
264 2. Perl does not associate an encoding with your strings.
266 ... until you force it to, e.g. when matching it against a regex,
267 or printing the scalar to a file, in which case Perl either
268 interprets your string as locale-encoded text, octets/binary, or as
269 Unicode, depending on various settings. In no case is an encoding
270 stored together with your data, it is use that decides encoding,
271 not any magical meta data.
272 3. The internal utf-8 flag has no meaning with regards to the encoding
274 of your string.
275 4. A "Unicode String" is simply a string where each character can be
276 validly interpreted as a Unicode code point.
277 If you have UTF-8 encoded data, it is no longer a Unicode string,
278 but a Unicode string encoded in UTF-8, giving you a binary string.
279 5. A string containing "high" (> 255) character values is not a UTF-8
281 string.
282 6. Unicode noncharacters only warn, as in core.
283 The 66 Unicode noncharacters U+FDD0..U+FDEF, and U+*FFFE, U+*FFFF
284 just warn, see <http://www.unicode.org/versions/corrigendum9.html>.
285 But illegal surrogate pairs fail to parse.
286 7. Raw non-Unicode characters above U+10FFFF are disallowed.
288 Raw non-Unicode characters outside the valid unicode range fail to
289 parse, because "A string is a sequence of zero or more Unicode
290 characters" RFC 7159 section 1 and "JSON text SHALL be encoded in
291 Unicode RFC 7159 section 8.1. We use now the UTF8_DISALLOW_SUPER
292 flag when parsing unicode.
293 I hope this helps :)
295
297 The object oriented interface lets you configure your own encoding or
298 decoding style, within the limits of supported formats.
299 $json = new Cpanel::JSON::XS
301 Creates a new JSON object that can be used to de/encode JSON
302 strings. All boolean flags described below are by default disabled.
303 The mutators for flags all return the JSON object again and thus
305 calls can be chained:
306 my $json = Cpanel::JSON::XS->new->utf8->space_after->encode ({a => [1,2]})
308 => {"a": [1, 2]}
309 $json = $json->ascii ([$enable])
311 $enabled = $json->get_ascii
312 If $enable is true (or missing), then the "encode" method will not
313 generate characters outside the code range 0..127 (which is ASCII).
314 Any Unicode characters outside that range will be escaped using
315 either a single "\uXXXX" (BMP characters) or a double
316 "\uHHHH\uLLLLL" escape sequence, as per RFC4627. The resulting
317 encoded JSON text can be treated as a native Unicode string, an
318 ascii-encoded, latin1-encoded or UTF-8 encoded string, or any other
319 superset of ASCII.
320 If $enable is false, then the "encode" method will not escape
322 Unicode characters unless required by the JSON syntax or other
323 flags. This results in a faster and more compact format.
324 See also the section ENCODING/CODESET FLAG NOTES later in this
326 document.
327 The main use for this flag is to produce JSON texts that can be
329 transmitted over a 7-bit channel, as the encoded JSON texts will
330 not contain any 8 bit characters.
331 Cpanel::JSON::XS->new->ascii (1)->encode ([chr 0x10401])
333 => ["\ud801\udc01"]
334 $json = $json->latin1 ([$enable])
336 $enabled = $json->get_latin1
337 If $enable is true (or missing), then the "encode" method will
338 encode the resulting JSON text as latin1 (or ISO-8859-1), escaping
339 any characters outside the code range 0..255. The resulting string
340 can be treated as a latin1-encoded JSON text or a native Unicode
341 string. The "decode" method will not be affected in any way by this
342 flag, as "decode" by default expects Unicode, which is a strict
343 superset of latin1.
344 If $enable is false, then the "encode" method will not escape
346 Unicode characters unless required by the JSON syntax or other
347 flags.
348 See also the section ENCODING/CODESET FLAG NOTES later in this
350 document.
351 The main use for this flag is efficiently encoding binary data as
353 JSON text, as most octets will not be escaped, resulting in a
354 smaller encoded size. The disadvantage is that the resulting JSON
355 text is encoded in latin1 (and must correctly be treated as such
356 when storing and transferring), a rare encoding for JSON. It is
357 therefore most useful when you want to store data structures known
358 to contain binary data efficiently in files or databases, not when
359 talking to other JSON encoders/decoders.
360 Cpanel::JSON::XS->new->latin1->encode (["\x{89}\x{abc}"]
362 => ["\x{89}\\u0abc"] # (perl syntax, U+abc escaped, U+89 not)
363 $json = $json->binary ([$enable])
365 $enabled = $json = $json->get_binary
366 If the $enable argument is true (or missing), then the "encode"
367 method will not try to detect an UTF-8 encoding in any JSON string,
368 it will strictly interpret it as byte sequence. The result might
369 contain new "\xNN" sequences, which is unparsable JSON. The
370 "decode" method forbids "\uNNNN" sequences and accepts "\xNN" and
371 octal "\NNN" sequences.
372 There is also a special logic for perl 5.6 and utf8. 5.6 encodes
374 any string to utf-8 automatically when seeing a codepoint >= 0x80
375 and < 0x100. With the binary flag enabled decode the perl utf8
376 encoded string to the original byte encoding and encode this with
377 "\xNN" escapes. This will result to the same encodings as with
378 newer perls. But note that binary multi-byte codepoints with 5.6
379 will result in "illegal unicode character in binary string" errors,
380 unlike with newer perls.
381 If $enable is false, then the "encode" method will smartly try to
383 detect Unicode characters unless required by the JSON syntax or
384 other flags and hex and octal sequences are forbidden.
385 See also the section ENCODING/CODESET FLAG NOTES later in this
387 document.
388 The main use for this flag is to avoid the smart unicode detection
390 and possible double encoding. The disadvantage is that the
391 resulting JSON text is encoded in new "\xNN" and in latin1
392 characters and must correctly be treated as such when storing and
393 transferring, a rare encoding for JSON. It will produce non-
394 readable JSON strings in the browser. It is therefore most useful
395 when you want to store data structures known to contain binary data
396 efficiently in files or databases, not when talking to other JSON
397 encoders/decoders. The binary decoding method can also be used
398 when an encoder produced a non-JSON conformant hex or octal
399 encoding "\xNN" or "\NNN".
400 Cpanel::JSON::XS->new->binary->encode (["\x{89}\x{abc}"])
402 5.6: Error: malformed or illegal unicode character in binary string
403 >=5.8: ['\x89\xe0\xaa\xbc']
404 Cpanel::JSON::XS->new->binary->encode (["\x{89}\x{bc}"])
406 => ["\x89\xbc"]
407 Cpanel::JSON::XS->new->binary->decode (["\x89\ua001"])
409 Error: malformed or illegal unicode character in binary string
410 Cpanel::JSON::XS->new->decode (["\x89"])
412 Error: illegal hex character in non-binary string
413 $json = $json->utf8 ([$enable])
415 $enabled = $json->get_utf8
416 If $enable is true (or missing), then the "encode" method will
417 encode the JSON result into UTF-8, as required by many protocols,
418 while the "decode" method expects to be handled an UTF-8-encoded
419 string. Please note that UTF-8-encoded strings do not contain any
420 characters outside the range 0..255, they are thus useful for
421 bytewise/binary I/O. In future versions, enabling this option might
422 enable autodetection of the UTF-16 and UTF-32 encoding families, as
423 described in RFC4627.
424 If $enable is false, then the "encode" method will return the JSON
426 string as a (non-encoded) Unicode string, while "decode" expects
427 thus a Unicode string. Any decoding or encoding (e.g. to UTF-8 or
428 UTF-16) needs to be done yourself, e.g. using the Encode module.
429 See also the section ENCODING/CODESET FLAG NOTES later in this
431 document.
432 Example, output UTF-16BE-encoded JSON:
434 use Encode;
436 $jsontext = encode "UTF-16BE", Cpanel::JSON::XS->new->encode ($object);
437 Example, decode UTF-32LE-encoded JSON:
439 use Encode;
441 $object = Cpanel::JSON::XS->new->decode (decode "UTF-32LE", $jsontext);
442 $json = $json->pretty ([$enable])
444 This enables (or disables) all of the "indent", "space_before" and
445 "space_after" (and in the future possibly more) flags in one call
446 to generate the most readable (or most compact) form possible.
447 Example, pretty-print some simple structure:
449 my $json = Cpanel::JSON::XS->new->pretty(1)->encode ({a => [1,2]})
451 =>
452 {
453 "a" : [
454 1,
455 2
456 ]
457 }
458 $json = $json->indent ([$enable])
460 $enabled = $json->get_indent
461 If $enable is true (or missing), then the "encode" method will use
462 a multiline format as output, putting every array member or
463 object/hash key-value pair into its own line, indenting them
464 properly.
465 If $enable is false, no newlines or indenting will be produced, and
467 the resulting JSON text is guaranteed not to contain any
468 "newlines".
469 This setting has no effect when decoding JSON texts.
471 $json = $json->indent_length([$number_of_spaces])
473 $length = $json->get_indent_length()
474 Set the indent length (default 3). This option is only useful when
475 you also enable indent or pretty. The acceptable range is from 0
476 (no indentation) to 15
477 $json = $json->space_before ([$enable])
479 $enabled = $json->get_space_before
480 If $enable is true (or missing), then the "encode" method will add
481 an extra optional space before the ":" separating keys from values
482 in JSON objects.
483 If $enable is false, then the "encode" method will not add any
485 extra space at those places.
486 This setting has no effect when decoding JSON texts. You will also
488 most likely combine this setting with "space_after".
489 Example, space_before enabled, space_after and indent disabled:
491 {"key" :"value"}
493 $json = $json->space_after ([$enable])
495 $enabled = $json->get_space_after
496 If $enable is true (or missing), then the "encode" method will add
497 an extra optional space after the ":" separating keys from values
498 in JSON objects and extra whitespace after the "," separating key-
499 value pairs and array members.
500 If $enable is false, then the "encode" method will not add any
502 extra space at those places.
503 This setting has no effect when decoding JSON texts.
505 Example, space_before and indent disabled, space_after enabled:
507 {"key": "value"}
509 $json = $json->relaxed ([$enable])
511 $enabled = $json->get_relaxed
512 If $enable is true (or missing), then "decode" will accept some
513 extensions to normal JSON syntax (see below). "encode" will not be
514 affected in anyway. Be aware that this option makes you accept
515 invalid JSON texts as if they were valid!. I suggest only to use
516 this option to parse application-specific files written by humans
517 (configuration files, resource files etc.)
518 If $enable is false (the default), then "decode" will only accept
520 valid JSON texts.
521 Currently accepted extensions are:
523 • list items can have an end-comma
525 JSON separates array elements and key-value pairs with commas.
527 This can be annoying if you write JSON texts manually and want
528 to be able to quickly append elements, so this extension
529 accepts comma at the end of such items not just between them:
530 [
532 1,
533 2, <- this comma not normally allowed
534 ]
535 {
536 "k1": "v1",
537 "k2": "v2", <- this comma not normally allowed
538 }
539 • shell-style '#'-comments
541 Whenever JSON allows whitespace, shell-style comments are
543 additionally allowed. They are terminated by the first
544 carriage-return or line-feed character, after which more white-
545 space and comments are allowed.
546 [
548 1, # this comment not allowed in JSON
549 # neither this one...
550 ]
551 • literal ASCII TAB characters in strings
553 Literal ASCII TAB characters are now allowed in strings (and
555 treated as "\t") in relaxed mode. Despite JSON mandates, that
556 TAB character is substituted for "\t" sequence.
557 [
559 "Hello\tWorld",
560 "Hello<TAB>World", # literal <TAB> would not normally be allowed
561 ]
562 • allow_singlequote
564 Single quotes are accepted instead of double quotes. See the
566 "allow_singlequote" option.
567 { "foo":'bar' }
569 { 'foo':"bar" }
570 { 'foo':'bar' }
571 • allow_barekey
573 Accept unquoted object keys instead of with mandatory double
575 quotes. See the "allow_barekey" option.
576 { foo:"bar" }
578 • allow_dupkeys
580 Allow decoding of duplicate keys in hashes. By default
582 duplicate keys are forbidden. See
583 <http://seriot.ch/parsing_json.php#24>: RFC 7159 section 4:
584 "The names within an object should be unique." See the
585 "allow_dupkeys" option.
586 $json = $json->canonical ([$enable])
588 $enabled = $json->get_canonical
589 If $enable is true (or missing), then the "encode" method will
590 output JSON objects by sorting their keys. This is adding a
591 comparatively high overhead.
592 If $enable is false, then the "encode" method will output key-value
594 pairs in the order Perl stores them (which will likely change
595 between runs of the same script, and can change even within the
596 same run from 5.18 onwards).
597 This option is useful if you want the same data structure to be
599 encoded as the same JSON text (given the same overall settings). If
600 it is disabled, the same hash might be encoded differently even if
601 contains the same data, as key-value pairs have no inherent
602 ordering in Perl.
603 This setting has no effect when decoding JSON texts.
605 This is now also done with tied hashes, contrary to JSON::XS. But
607 note that with most large tied hashes stored as tree it is advised
608 to sort the iterator already and don't sort the hash output here.
609 Most such iterators are already sorted, as such e.g. DB_File with
610 "DB_BTREE".
611 $json = $json->sort_by (undef, 0, 1 or a block)
613 This currently only (un)sets the "canonical" option, and ignores
614 custom sort blocks.
615 This setting has no effect when decoding JSON texts.
617 This setting has currently no effect on tied hashes.
619 $json = $json->escape_slash ([$enable])
621 $enabled = $json->get_escape_slash
622 According to the JSON Grammar, the forward slash character (U+002F)
623 "/" need to be escaped. But by default strings are encoded without
624 escaping slashes in all perl JSON encoders.
625 If $enable is true (or missing), then "encode" will escape slashes,
627 "\/".
628 This setting has no effect when decoding JSON texts.
630 $json = $json->unblessed_bool ([$enable])
632 $enabled = $json->get_unblessed_bool
633 $json = $json->unblessed_bool([$enable])
634 If $enable is true (or missing), then "decode" will return Perl
636 non-object boolean variables (1 and 0) for JSON booleans ("true"
637 and "false"). If $enable is false, then "decode" will return
638 "JSON::PP::Boolean" objects for JSON booleans.
639 $json = $json->allow_singlequote ([$enable])
641 $enabled = $json->get_allow_singlequote
642 $json = $json->allow_singlequote([$enable])
643 If $enable is true (or missing), then "decode" will accept JSON
645 strings quoted by single quotations that are invalid JSON format.
646 $json->allow_singlequote->decode({"foo":'bar'});
648 $json->allow_singlequote->decode({'foo':"bar"});
649 $json->allow_singlequote->decode({'foo':'bar'});
650 This is also enabled with "relaxed". As same as the "relaxed"
652 option, this option may be used to parse application-specific files
653 written by humans.
654 $json = $json->allow_barekey ([$enable])
656 $enabled = $json->get_allow_barekey
657 $json = $json->allow_barekey([$enable])
658 If $enable is true (or missing), then "decode" will accept bare
660 keys of JSON object that are invalid JSON format.
661 Same as with the "relaxed" option, this option may be used to parse
663 application-specific files written by humans.
664 $json->allow_barekey->decode('{foo:"bar"}');
666 $json = $json->allow_bignum ([$enable])
668 $enabled = $json->get_allow_bignum
669 $json = $json->allow_bignum([$enable])
670 If $enable is true (or missing), then "decode" will convert the big
672 integer Perl cannot handle as integer into a Math::BigInt object
673 and convert a floating number (any) into a Math::BigFloat.
674 On the contrary, "encode" converts "Math::BigInt" objects and
676 "Math::BigFloat" objects into JSON numbers with "allow_blessed"
677 enable.
678 $json->allow_nonref->allow_blessed->allow_bignum;
680 $bigfloat = $json->decode('2.000000000000000000000000001');
681 print $json->encode($bigfloat);
682 # => 2.000000000000000000000000001
683 See "MAPPING" about the normal conversion of JSON number.
685 $json = $json->allow_bigint ([$enable])
687 This option is obsolete and replaced by allow_bignum.
688 $json = $json->allow_nonref ([$enable])
690 $enabled = $json->get_allow_nonref
691 If $enable is true (or missing), then the "encode" method can
692 convert a non-reference into its corresponding string, number or
693 null JSON value, which is an extension to RFC4627. Likewise,
694 "decode" will accept those JSON values instead of croaking.
695 If $enable is false, then the "encode" method will croak if it
697 isn't passed an arrayref or hashref, as JSON texts must either be
698 an object or array. Likewise, "decode" will croak if given
699 something that is not a JSON object or array.
700 Example, encode a Perl scalar as JSON value with enabled
702 "allow_nonref", resulting in an invalid JSON text:
703 Cpanel::JSON::XS->new->allow_nonref->encode ("Hello, World!")
705 => "Hello, World!"
706 $json = $json->allow_unknown ([$enable])
708 $enabled = $json->get_allow_unknown
709 If $enable is true (or missing), then "encode" will not throw an
710 exception when it encounters values it cannot represent in JSON
711 (for example, filehandles) but instead will encode a JSON "null"
712 value. Note that blessed objects are not included here and are
713 handled separately by c<allow_nonref>.
714 If $enable is false (the default), then "encode" will throw an
716 exception when it encounters anything it cannot encode as JSON.
717 This option does not affect "decode" in any way, and it is
719 recommended to leave it off unless you know your communications
720 partner.
721 $json = $json->allow_stringify ([$enable])
723 $enabled = $json->get_allow_stringify
724 If $enable is true (or missing), then "encode" will stringify the
725 non-object perl value or reference. Note that blessed objects are
726 not included here and are handled separately by "allow_blessed" and
727 "convert_blessed". String references are stringified to the string
728 value, other references as in perl.
729 This option does not affect "decode" in any way.
731 This option is special to this module, it is not supported by other
733 encoders. So it is not recommended to use it.
734 $json = $json->require_types ([$enable])
736 $enable = $json->get_require_types
737 $json = $json->require_types([$enable])
738 If $enable is true (or missing), then "encode" will require either
740 enabled "type_all_string" or second argument with supplied JSON
741 types. See Cpanel::JSON::XS::Type. When "type_all_string" is not
742 enabled or second argument is not provided (or is undef), then
743 "encode" croaks. It also croaks when the type for provided
744 structure in "encode" is incomplete.
745 $json = $json->type_all_string ([$enable])
747 $enable = $json->get_type_all_string
748 $json = $json->type_all_string([$enable])
749 If $enable is true (or missing), then "encode" will always produce
751 stable deterministic JSON string types in resulted output.
752 When $enable is false, then result of encoded JSON output may be
754 different for different Perl versions and may depends on loaded
755 modules.
756 This is useful it you need deterministic JSON types, independently
758 of used Perl version and other modules, but do not want to write
759 complicated type definitions for Cpanel::JSON::XS::Type.
760 $json = $json->allow_dupkeys ([$enable])
762 $enabled = $json->get_allow_dupkeys
763 If $enable is true (or missing), then the "decode" method will not
764 die when it encounters duplicate keys in a hash. "allow_dupkeys"
765 is also enabled in the "relaxed" mode.
766 The JSON spec allows duplicate name in objects but recommends to
768 disable it, however with Perl hashes they are impossible, parsing
769 JSON in Perl silently ignores duplicate names, using the last value
770 found.
771 See <http://seriot.ch/parsing_json.php#24>: RFC 7159 section 4:
773 "The names within an object should be unique."
774 $json = $json->allow_blessed ([$enable])
776 $enabled = $json->get_allow_blessed
777 If $enable is true (or missing), then the "encode" method will not
778 barf when it encounters a blessed reference. Instead, the value of
779 the convert_blessed option will decide whether "null"
780 ("convert_blessed" disabled or no "TO_JSON" method found) or a
781 representation of the object ("convert_blessed" enabled and
782 "TO_JSON" method found) is being encoded. Has no effect on
783 "decode".
784 If $enable is false (the default), then "encode" will throw an
786 exception when it encounters a blessed object without
787 "convert_blessed" and a "TO_JSON" method.
788 This setting has no effect on "decode".
790 $json = $json->convert_blessed ([$enable])
792 $enabled = $json->get_convert_blessed
793 If $enable is true (or missing), then "encode", upon encountering a
794 blessed object, will check for the availability of the "TO_JSON"
795 method on the object's class. If found, it will be called in scalar
796 context and the resulting scalar will be encoded instead of the
797 object. If no "TO_JSON" method is found, a stringification overload
798 method is tried next. If both are not found, the value of
799 "allow_blessed" will decide what to do.
800 The "TO_JSON" method may safely call die if it wants. If "TO_JSON"
802 returns other blessed objects, those will be handled in the same
803 way. "TO_JSON" must take care of not causing an endless recursion
804 cycle (== crash) in this case. The same care must be taken with
805 calling encode in stringify overloads (even if this works by luck
806 in older perls) or other callbacks. The name of "TO_JSON" was
807 chosen because other methods called by the Perl core (== not by the
808 user of the object) are usually in upper case letters and to avoid
809 collisions with any "to_json" function or method.
810 If $enable is false (the default), then "encode" will not consider
812 this type of conversion.
813 This setting has no effect on "decode".
815 $json = $json->allow_tags ([$enable])
817 $enabled = $json->get_allow_tags
818 See "OBJECT SERIALIZATION" for details.
819 If $enable is true (or missing), then "encode", upon encountering a
821 blessed object, will check for the availability of the "FREEZE"
822 method on the object's class. If found, it will be used to
823 serialize the object into a nonstandard tagged JSON value (that
824 JSON decoders cannot decode).
825 It also causes "decode" to parse such tagged JSON values and
827 deserialize them via a call to the "THAW" method.
828 If $enable is false (the default), then "encode" will not consider
830 this type of conversion, and tagged JSON values will cause a parse
831 error in "decode", as if tags were not part of the grammar.
832 $json = $json->filter_json_object ([$coderef->($hashref)])
834 When $coderef is specified, it will be called from "decode" each
835 time it decodes a JSON object. The only argument is a reference to
836 the newly-created hash. If the code references returns a single
837 scalar (which need not be a reference), this value (i.e. a copy of
838 that scalar to avoid aliasing) is inserted into the deserialized
839 data structure. If it returns an empty list (NOTE: not "undef",
840 which is a valid scalar), the original deserialized hash will be
841 inserted. This setting can slow down decoding considerably.
842 When $coderef is omitted or undefined, any existing callback will
844 be removed and "decode" will not change the deserialized hash in
845 any way.
846 Example, convert all JSON objects into the integer 5:
848 my $js = Cpanel::JSON::XS->new->filter_json_object (sub { 5 });
850 # returns [5]
851 $js->decode ('[{}]')
852 # throw an exception because allow_nonref is not enabled
853 # so a lone 5 is not allowed.
854 $js->decode ('{"a":1, "b":2}');
855 $json = $json->filter_json_single_key_object ($key [=>
857 $coderef->($value)])
858 Works remotely similar to "filter_json_object", but is only called
859 for JSON objects having a single key named $key.
860 This $coderef is called before the one specified via
862 "filter_json_object", if any. It gets passed the single value in
863 the JSON object. If it returns a single value, it will be inserted
864 into the data structure. If it returns nothing (not even "undef"
865 but the empty list), the callback from "filter_json_object" will be
866 called next, as if no single-key callback were specified.
867 If $coderef is omitted or undefined, the corresponding callback
869 will be disabled. There can only ever be one callback for a given
870 key.
871 As this callback gets called less often then the
873 "filter_json_object" one, decoding speed will not usually suffer as
874 much. Therefore, single-key objects make excellent targets to
875 serialize Perl objects into, especially as single-key JSON objects
876 are as close to the type-tagged value concept as JSON gets (it's
877 basically an ID/VALUE tuple). Of course, JSON does not support this
878 in any way, so you need to make sure your data never looks like a
879 serialized Perl hash.
880 Typical names for the single object key are "__class_whatever__",
882 or "$__dollars_are_rarely_used__$" or "}ugly_brace_placement", or
883 even things like "__class_md5sum(classname)__", to reduce the risk
884 of clashing with real hashes.
885 Example, decode JSON objects of the form "{ "__widget__" => <id> }"
887 into the corresponding $WIDGET{<id>} object:
888 # return whatever is in $WIDGET{5}:
890 Cpanel::JSON::XS
891 ->new
892 ->filter_json_single_key_object (__widget__ => sub {
893 $WIDGET{ $_[0] }
894 })
895 ->decode ('{"__widget__": 5')
896 # this can be used with a TO_JSON method in some "widget" class
898 # for serialization to json:
899 sub WidgetBase::TO_JSON {
900 my ($self) = @_;
901 unless ($self->{id}) {
903 $self->{id} = ..get..some..id..;
904 $WIDGET{$self->{id}} = $self;
905 }
906 { __widget__ => $self->{id} }
908 }
909 $json = $json->shrink ([$enable])
911 $enabled = $json->get_shrink
912 Perl usually over-allocates memory a bit when allocating space for
913 strings. This flag optionally resizes strings generated by either
914 "encode" or "decode" to their minimum size possible. This can save
915 memory when your JSON texts are either very very long or you have
916 many short strings. It will also try to downgrade any strings to
917 octet-form if possible: perl stores strings internally either in an
918 encoding called UTF-X or in octet-form. The latter cannot store
919 everything but uses less space in general (and some buggy Perl or C
920 code might even rely on that internal representation being used).
921 The actual definition of what shrink does might change in future
923 versions, but it will always try to save space at the expense of
924 time.
925 If $enable is true (or missing), the string returned by "encode"
927 will be shrunk-to-fit, while all strings generated by "decode" will
928 also be shrunk-to-fit.
929 If $enable is false, then the normal perl allocation algorithms are
931 used. If you work with your data, then this is likely to be
932 faster.
933 In the future, this setting might control other things, such as
935 converting strings that look like integers or floats into integers
936 or floats internally (there is no difference on the Perl level),
937 saving space.
938 $json = $json->max_depth ([$maximum_nesting_depth])
940 $max_depth = $json->get_max_depth
941 Sets the maximum nesting level (default 512) accepted while
942 encoding or decoding. If a higher nesting level is detected in JSON
943 text or a Perl data structure, then the encoder and decoder will
944 stop and croak at that point.
945 Nesting level is defined by number of hash- or arrayrefs that the
947 encoder needs to traverse to reach a given point or the number of
948 "{" or "[" characters without their matching closing parenthesis
949 crossed to reach a given character in a string.
950 Setting the maximum depth to one disallows any nesting, so that
952 ensures that the object is only a single hash/object or array.
953 If no argument is given, the highest possible setting will be used,
955 which is rarely useful.
956 Note that nesting is implemented by recursion in C. The default
958 value has been chosen to be as large as typical operating systems
959 allow without crashing.
960 See "SECURITY CONSIDERATIONS", below, for more info on why this is
962 useful.
963 $json = $json->max_size ([$maximum_string_size])
965 $max_size = $json->get_max_size
966 Set the maximum length a JSON text may have (in bytes) where
967 decoding is being attempted. The default is 0, meaning no limit.
968 When "decode" is called on a string that is longer then this many
969 bytes, it will not attempt to decode the string but throw an
970 exception. This setting has no effect on "encode" (yet).
971 If no argument is given, the limit check will be deactivated (same
973 as when 0 is specified).
974 See "SECURITY CONSIDERATIONS", below, for more info on why this is
976 useful.
977 $json->stringify_infnan ([$infnan_mode = 1])
979 $infnan_mode = $json->get_stringify_infnan
980 Get or set how Cpanel::JSON::XS encodes "inf", "-inf" or "nan" for
981 numeric values. Also qnan, snan or negative nan on some platforms.
982 "null": infnan_mode = 0. Similar to most JSON modules in other
984 languages. Always null.
985 stringified: infnan_mode = 1. As in Mojo::JSON. Platform specific
987 strings. Stringified via sprintf(%g), with double quotes.
988 inf/nan: infnan_mode = 2. As in JSON::XS, and older releases.
990 Passes through platform dependent values, invalid JSON. Stringified
991 via sprintf(%g), but without double quotes.
992 "inf/-inf/nan": infnan_mode = 3. Platform independent inf/nan/-inf
994 strings. No QNAN/SNAN/negative NAN support, unified to "nan". Much
995 easier to detect, but may conflict with valid strings.
996 $json_text = $json->encode ($perl_scalar, $json_type)
998 Converts the given Perl data structure (a simple scalar or a
999 reference to a hash or array) to its JSON representation. Simple
1000 scalars will be converted into JSON string or number sequences,
1001 while references to arrays become JSON arrays and references to
1002 hashes become JSON objects. Undefined Perl values (e.g. "undef")
1003 become JSON "null" values. Neither "true" nor "false" values will
1004 be generated.
1005 For the type argument see Cpanel::JSON::XS::Type.
1007 $perl_scalar = $json->decode ($json_text, my $json_type)
1009 The opposite of "encode": expects a JSON text and tries to parse
1010 it, returning the resulting simple scalar or reference. Croaks on
1011 error.
1012 JSON numbers and strings become simple Perl scalars. JSON arrays
1014 become Perl arrayrefs and JSON objects become Perl hashrefs. "true"
1015 becomes 1, "false" becomes 0 and "null" becomes "undef".
1016 For the type argument see Cpanel::JSON::XS::Type.
1018 ($perl_scalar, $characters) = $json->decode_prefix ($json_text)
1020 This works like the "decode" method, but instead of raising an
1021 exception when there is trailing garbage after the first JSON
1022 object, it will silently stop parsing there and return the number
1023 of characters consumed so far.
1024 This is useful if your JSON texts are not delimited by an outer
1026 protocol and you need to know where the JSON text ends.
1027 Cpanel::JSON::XS->new->decode_prefix ("[1] the tail")
1029 => ([1], 3)
1030 $json->to_json ($perl_hash_or_arrayref)
1032 Deprecated method for perl 5.8 and newer. Use encode_json instead.
1033 $json->from_json ($utf8_encoded_json_text)
1035 Deprecated method for perl 5.8 and newer. Use decode_json instead.
1036
1038 In some cases, there is the need for incremental parsing of JSON texts.
1039 While this module always has to keep both JSON text and resulting Perl
1040 data structure in memory at one time, it does allow you to parse a JSON
1041 stream incrementally. It does so by accumulating text until it has a
1042 full JSON object, which it then can decode. This process is similar to
1043 using "decode_prefix" to see if a full JSON object is available, but is
1044 much more efficient (and can be implemented with a minimum of method
1045 calls).
1046 Cpanel::JSON::XS will only attempt to parse the JSON text once it is
1048 sure it has enough text to get a decisive result, using a very simple
1049 but truly incremental parser. This means that it sometimes won't stop
1050 as early as the full parser, for example, it doesn't detect mismatched
1051 parentheses. The only thing it guarantees is that it starts decoding as
1052 soon as a syntactically valid JSON text has been seen. This means you
1053 need to set resource limits (e.g. "max_size") to ensure the parser will
1054 stop parsing in the presence if syntax errors.
1055 The following methods implement this incremental parser.
1057 [void, scalar or list context] = $json->incr_parse ([$string])
1059 This is the central parsing function. It can both append new text
1060 and extract objects from the stream accumulated so far (both of
1061 these functions are optional).
1062 If $string is given, then this string is appended to the already
1064 existing JSON fragment stored in the $json object.
1065 After that, if the function is called in void context, it will
1067 simply return without doing anything further. This can be used to
1068 add more text in as many chunks as you want.
1069 If the method is called in scalar context, then it will try to
1071 extract exactly one JSON object. If that is successful, it will
1072 return this object, otherwise it will return "undef". If there is a
1073 parse error, this method will croak just as "decode" would do (one
1074 can then use "incr_skip" to skip the erroneous part). This is the
1075 most common way of using the method.
1076 And finally, in list context, it will try to extract as many
1078 objects from the stream as it can find and return them, or the
1079 empty list otherwise. For this to work, there must be no separators
1080 between the JSON objects or arrays, instead they must be
1081 concatenated back-to-back. If an error occurs, an exception will be
1082 raised as in the scalar context case. Note that in this case, any
1083 previously-parsed JSON texts will be lost.
1084 Example: Parse some JSON arrays/objects in a given string and
1086 return them.
1087 my @objs = Cpanel::JSON::XS->new->incr_parse ("[5][7][1,2]");
1089 $lvalue_string = $json->incr_text (>5.8 only)
1091 This method returns the currently stored JSON fragment as an
1092 lvalue, that is, you can manipulate it. This only works when a
1093 preceding call to "incr_parse" in scalar context successfully
1094 returned an object, and 2. only with Perl >= 5.8
1095 Under all other circumstances you must not call this function (I
1097 mean it. although in simple tests it might actually work, it will
1098 fail under real world conditions). As a special exception, you can
1099 also call this method before having parsed anything.
1100 This function is useful in two cases: a) finding the trailing text
1102 after a JSON object or b) parsing multiple JSON objects separated
1103 by non-JSON text (such as commas).
1104 $json->incr_skip
1106 This will reset the state of the incremental parser and will remove
1107 the parsed text from the input buffer so far. This is useful after
1108 "incr_parse" died, in which case the input buffer and incremental
1109 parser state is left unchanged, to skip the text parsed so far and
1110 to reset the parse state.
1111 The difference to "incr_reset" is that only text until the parse
1113 error occurred is removed.
1114 $json->incr_reset
1116 This completely resets the incremental parser, that is, after this
1117 call, it will be as if the parser had never parsed anything.
1118 This is useful if you want to repeatedly parse JSON objects and
1120 want to ignore any trailing data, which means you have to reset the
1121 parser after each successful decode.
1122 LIMITATIONS
1124 All options that affect decoding are supported, except "allow_nonref".
1125 The reason for this is that it cannot be made to work sensibly: JSON
1126 objects and arrays are self-delimited, i.e. you can concatenate them
1127 back to back and still decode them perfectly. This does not hold true
1128 for JSON numbers, however.
1129 For example, is the string 1 a single JSON number, or is it simply the
1131 start of 12? Or is 12 a single JSON number, or the concatenation of 1
1132 and 2? In neither case you can tell, and this is why Cpanel::JSON::XS
1133 takes the conservative route and disallows this case.
1134 EXAMPLES
1136 Some examples will make all this clearer. First, a simple example that
1137 works similarly to "decode_prefix": We want to decode the JSON object
1138 at the start of a string and identify the portion after the JSON
1139 object:
1140 my $text = "[1,2,3] hello";
1142 my $json = new Cpanel::JSON::XS;
1144 my $obj = $json->incr_parse ($text)
1146 or die "expected JSON object or array at beginning of string";
1147 my $tail = $json->incr_text;
1149 # $tail now contains " hello"
1150 Easy, isn't it?
1152 Now for a more complicated example: Imagine a hypothetical protocol
1154 where you read some requests from a TCP stream, and each request is a
1155 JSON array, without any separation between them (in fact, it is often
1156 useful to use newlines as "separators", as these get interpreted as
1157 whitespace at the start of the JSON text, which makes it possible to
1158 test said protocol with "telnet"...).
1159 Here is how you'd do it (it is trivial to write this in an event-based
1161 manner):
1162 my $json = new Cpanel::JSON::XS;
1164 # read some data from the socket
1166 while (sysread $socket, my $buf, 4096) {
1167 # split and decode as many requests as possible
1169 for my $request ($json->incr_parse ($buf)) {
1170 # act on the $request
1171 }
1172 }
1173 Another complicated example: Assume you have a string with JSON objects
1175 or arrays, all separated by (optional) comma characters (e.g. "[1],[2],
1176 [3]"). To parse them, we have to skip the commas between the JSON
1177 texts, and here is where the lvalue-ness of "incr_text" comes in
1178 useful:
1179 my $text = "[1],[2], [3]";
1181 my $json = new Cpanel::JSON::XS;
1182 # void context, so no parsing done
1184 $json->incr_parse ($text);
1185 # now extract as many objects as possible. note the
1187 # use of scalar context so incr_text can be called.
1188 while (my $obj = $json->incr_parse) {
1189 # do something with $obj
1190 # now skip the optional comma
1192 $json->incr_text =~ s/^ \s* , //x;
1193 }
1194 Now lets go for a very complex example: Assume that you have a gigantic
1196 JSON array-of-objects, many gigabytes in size, and you want to parse
1197 it, but you cannot load it into memory fully (this has actually
1198 happened in the real world :).
1199 Well, you lost, you have to implement your own JSON parser. But
1201 Cpanel::JSON::XS can still help you: You implement a (very simple)
1202 array parser and let JSON decode the array elements, which are all full
1203 JSON objects on their own (this wouldn't work if the array elements
1204 could be JSON numbers, for example):
1205 my $json = new Cpanel::JSON::XS;
1207 # open the monster
1209 open my $fh, "<bigfile.json"
1210 or die "bigfile: $!";
1211 # first parse the initial "["
1213 for (;;) {
1214 sysread $fh, my $buf, 65536
1215 or die "read error: $!";
1216 $json->incr_parse ($buf); # void context, so no parsing
1217 # Exit the loop once we found and removed(!) the initial "[".
1219 # In essence, we are (ab-)using the $json object as a simple scalar
1220 # we append data to.
1221 last if $json->incr_text =~ s/^ \s* \[ //x;
1222 }
1223 # now we have the skipped the initial "[", so continue
1225 # parsing all the elements.
1226 for (;;) {
1227 # in this loop we read data until we got a single JSON object
1228 for (;;) {
1229 if (my $obj = $json->incr_parse) {
1230 # do something with $obj
1231 last;
1232 }
1233 # add more data
1235 sysread $fh, my $buf, 65536
1236 or die "read error: $!";
1237 $json->incr_parse ($buf); # void context, so no parsing
1238 }
1239 # in this loop we read data until we either found and parsed the
1241 # separating "," between elements, or the final "]"
1242 for (;;) {
1243 # first skip whitespace
1244 $json->incr_text =~ s/^\s*//;
1245 # if we find "]", we are done
1247 if ($json->incr_text =~ s/^\]//) {
1248 print "finished.\n";
1249 exit;
1250 }
1251 # if we find ",", we can continue with the next element
1253 if ($json->incr_text =~ s/^,//) {
1254 last;
1255 }
1256 # if we find anything else, we have a parse error!
1258 if (length $json->incr_text) {
1259 die "parse error near ", $json->incr_text;
1260 }
1261 # else add more data
1263 sysread $fh, my $buf, 65536
1264 or die "read error: $!";
1265 $json->incr_parse ($buf); # void context, so no parsing
1266 }
1267 This is a complex example, but most of the complexity comes from the
1269 fact that we are trying to be correct (bear with me if I am wrong, I
1270 never ran the above example :).
1271
1273 Detect all unicode Byte Order Marks on decode. Which are UTF-8,
1274 UTF-16LE, UTF-16BE, UTF-32LE and UTF-32BE.
1275 The BOM encoding is set only for one specific decode call, it does not
1277 change the state of the JSON object.
1278 Warning: With perls older than 5.20 you need load the Encode module
1280 before loading a multibyte BOM, i.e. >= UTF-16. Otherwise an error is
1281 thrown. This is an implementation limitation and might get fixed later.
1282 See <https://tools.ietf.org/html/rfc7159#section-8.1> "JSON text SHALL
1284 be encoded in UTF-8, UTF-16, or UTF-32."
1285 "Implementations MUST NOT add a byte order mark to the beginning of a
1287 JSON text", "implementations (...) MAY ignore the presence of a byte
1288 order mark rather than treating it as an error".
1289 See also <http://www.unicode.org/faq/utf_bom.html#BOM>.
1291 Beware that Cpanel::JSON::XS is currently the only JSON module which
1293 does accept and decode a BOM.
1294 The latest JSON spec
1296 <https://www.greenbytes.de/tech/webdav/rfc8259.html#character.encoding>
1297 forbid the usage of UTF-16 or UTF-32, the character encoding is UTF-8.
1298 Thus in subsequent updates BOM's of UTF-16 or UTF-32 will throw an
1299 error.
1300
1302 This section describes how Cpanel::JSON::XS maps Perl values to JSON
1303 values and vice versa. These mappings are designed to "do the right
1304 thing" in most circumstances automatically, preserving round-tripping
1305 characteristics (what you put in comes out as something equivalent).
1306 For the more enlightened: note that in the following descriptions,
1308 lowercase perl refers to the Perl interpreter, while uppercase Perl
1309 refers to the abstract Perl language itself.
1310 JSON -> PERL
1312 object
1313 A JSON object becomes a reference to a hash in Perl. No ordering of
1314 object keys is preserved (JSON does not preserve object key
1315 ordering itself).
1316 array
1318 A JSON array becomes a reference to an array in Perl.
1319 string
1321 A JSON string becomes a string scalar in Perl - Unicode codepoints
1322 in JSON are represented by the same codepoints in the Perl string,
1323 so no manual decoding is necessary.
1324 number
1326 A JSON number becomes either an integer, numeric (floating point)
1327 or string scalar in perl, depending on its range and any fractional
1328 parts. On the Perl level, there is no difference between those as
1329 Perl handles all the conversion details, but an integer may take
1330 slightly less memory and might represent more values exactly than
1331 floating point numbers.
1332 If the number consists of digits only, Cpanel::JSON::XS will try to
1334 represent it as an integer value. If that fails, it will try to
1335 represent it as a numeric (floating point) value if that is
1336 possible without loss of precision. Otherwise it will preserve the
1337 number as a string value (in which case you lose roundtripping
1338 ability, as the JSON number will be re-encoded to a JSON string).
1339 Numbers containing a fractional or exponential part will always be
1341 represented as numeric (floating point) values, possibly at a loss
1342 of precision (in which case you might lose perfect roundtripping
1343 ability, but the JSON number will still be re-encoded as a JSON
1344 number).
1345 Note that precision is not accuracy - binary floating point values
1347 cannot represent most decimal fractions exactly, and when
1348 converting from and to floating point, "Cpanel::JSON::XS" only
1349 guarantees precision up to but not including the least significant
1350 bit.
1351 true, false
1353 When "unblessed_bool" is set to true, then JSON "true" becomes 1
1354 and JSON "false" becomes 0.
1355 Otherwise these JSON atoms become "JSON::PP::true" and
1357 "JSON::PP::false", respectively. They are "JSON::PP::Boolean"
1358 objects and are overloaded to act almost exactly like the numbers 1
1359 and 0. You can check whether a scalar is a JSON boolean by using
1360 the "Cpanel::JSON::XS::is_bool" function.
1361 The other round, from perl to JSON, "!0" which is represented as
1363 "yes" becomes "true", and "!1" which is represented as "no" becomes
1364 "false".
1365 Via Cpanel::JSON::XS::Type you can now even force negation in
1367 "encode", without overloading of "!":
1368 my $false = Cpanel::JSON::XS::false;
1370 print($json->encode([!$false], [JSON_TYPE_BOOL]));
1371 => [true]
1372 null
1374 A JSON null atom becomes "undef" in Perl.
1375 shell-style comments ("# text")
1377 As a nonstandard extension to the JSON syntax that is enabled by
1378 the "relaxed" setting, shell-style comments are allowed. They can
1379 start anywhere outside strings and go till the end of the line.
1380 tagged values ("(tag)value").
1382 Another nonstandard extension to the JSON syntax, enabled with the
1383 "allow_tags" setting, are tagged values. In this implementation,
1384 the tag must be a perl package/class name encoded as a JSON string,
1385 and the value must be a JSON array encoding optional constructor
1386 arguments.
1387 See "OBJECT SERIALIZATION", below, for details.
1389 PERL -> JSON
1391 The mapping from Perl to JSON is slightly more difficult, as Perl is a
1392 truly typeless language, so we can only guess which JSON type is meant
1393 by a Perl value.
1394 hash references
1396 Perl hash references become JSON objects. As there is no inherent
1397 ordering in hash keys (or JSON objects), they will usually be
1398 encoded in a pseudo-random order that can change between runs of
1399 the same program but stays generally the same within a single run
1400 of a program. Cpanel::JSON::XS can optionally sort the hash keys
1401 (determined by the canonical flag), so the same datastructure will
1402 serialize to the same JSON text (given same settings and version of
1403 Cpanel::JSON::XS), but this incurs a runtime overhead and is only
1404 rarely useful, e.g. when you want to compare some JSON text against
1405 another for equality.
1406 array references
1408 Perl array references become JSON arrays.
1409 other references
1411 Other unblessed references are generally not allowed and will cause
1412 an exception to be thrown, except for references to the integers 0
1413 and 1, which get turned into "false" and "true" atoms in JSON.
1414 With the option "allow_stringify", you can ignore the exception and
1416 return the stringification of the perl value.
1417 With the option "allow_unknown", you can ignore the exception and
1419 return "null" instead.
1420 encode_json [\"x"] # => cannot encode reference to scalar 'SCALAR(0x..)'
1422 # unless the scalar is 0 or 1
1423 encode_json [\0, \1] # yields [false,true]
1424 allow_stringify->encode_json [\"x"] # yields "x" unlike JSON::PP
1426 allow_unknown->encode_json [\"x"] # yields null as in JSON::PP
1427 Cpanel::JSON::XS::true, Cpanel::JSON::XS::false
1429 These special values become JSON true and JSON false values,
1430 respectively. You can also use "\1" and "\0" or "!0" and "!1"
1431 directly if you want.
1432 encode_json [Cpanel::JSON::XS::false, Cpanel::JSON::XS::true] # yields [false,true]
1434 encode_json [!1, !0], [JSON_TYPE_BOOL, JSON_TYPE_BOOL] # yields [false,true]
1435 eq/ne comparisons with true, false:
1437 false is eq to the empty string or the string 'false' or the
1439 special empty string "!!0" or "!1", i.e. "SV_NO", or the numbers 0
1440 or 0.0.
1441 true is eq to the string 'true' or to the special string "!0" (i.e.
1443 "SV_YES") or to the numbers 1 or 1.0.
1444 blessed objects
1446 Blessed objects are not directly representable in JSON, but
1447 "Cpanel::JSON::XS" allows various optional ways of handling
1448 objects. See "OBJECT SERIALIZATION", below, for details.
1449 See the "allow_blessed" and "convert_blessed" methods on various
1451 options on how to deal with this: basically, you can choose between
1452 throwing an exception, encoding the reference as if it weren't
1453 blessed, use the objects overloaded stringification method or
1454 provide your own serializer method.
1455 simple scalars
1457 Simple Perl scalars (any scalar that is not a reference) are the
1458 most difficult objects to encode: Cpanel::JSON::XS will encode
1459 undefined scalars or inf/nan as JSON "null" values and other
1460 scalars to either number or string in non-deterministic way which
1461 may be affected or changed by Perl version or any other loaded Perl
1462 module.
1463 If you want to have stable and deterministic types in JSON encoder
1465 then use Cpanel::JSON::XS::Type.
1466 Alternative way for deterministic types is to use "type_all_string"
1468 method when all perl scalars are encoded to JSON strings.
1469 Non-deterministic behavior is following: scalars that have last
1471 been used in a string context before encoding as JSON strings, and
1472 anything else as number value:
1473 # dump as number
1475 encode_json [2] # yields [2]
1476 encode_json [-3.0e17] # yields [-3e+17]
1477 my $value = 5; encode_json [$value] # yields [5]
1478 # used as string, but the two representations are for the same number
1480 print $value;
1481 encode_json [$value] # yields [5]
1482 # used as different string (non-matching dual-var)
1484 my $str = '0 but true';
1485 my $num = 1 + $str;
1486 encode_json [$num, $str] # yields [1,"0 but true"]
1487 # undef becomes null
1489 encode_json [undef] # yields [null]
1490 # inf or nan becomes null, unless you answered
1492 # "Do you want to handle inf/nan as strings" with yes
1493 encode_json [9**9**9] # yields [null]
1494 You can force the type to be a JSON string by stringifying it:
1496 my $x = 3.1; # some variable containing a number
1498 "$x"; # stringified
1499 $x .= ""; # another, more awkward way to stringify
1500 print $x; # perl does it for you, too, quite often
1501 You can force the type to be a JSON number by numifying it:
1503 my $x = "3"; # some variable containing a string
1505 $x += 0; # numify it, ensuring it will be dumped as a number
1506 $x *= 1; # same thing, the choice is yours.
1507 Note that numerical precision has the same meaning as under Perl
1509 (so binary to decimal conversion follows the same rules as in Perl,
1510 which can differ to other languages). Also, your perl interpreter
1511 might expose extensions to the floating point numbers of your
1512 platform, such as infinities or NaN's - these cannot be represented
1513 in JSON, and thus null is returned instead. Optionally you can
1514 configure it to stringify inf and nan values.
1515 OBJECT SERIALIZATION
1517 As JSON cannot directly represent Perl objects, you have to choose
1518 between a pure JSON representation (without the ability to deserialize
1519 the object automatically again), and a nonstandard extension to the
1520 JSON syntax, tagged values.
1521 SERIALIZATION
1523 What happens when "Cpanel::JSON::XS" encounters a Perl object depends
1525 on the "allow_blessed", "convert_blessed" and "allow_tags" settings,
1526 which are used in this order:
1527 1. "allow_tags" is enabled and the object has a "FREEZE" method.
1529 In this case, "Cpanel::JSON::XS" uses the Types::Serialiser object
1530 serialization protocol to create a tagged JSON value, using a
1531 nonstandard extension to the JSON syntax.
1532 This works by invoking the "FREEZE" method on the object, with the
1534 first argument being the object to serialize, and the second
1535 argument being the constant string "JSON" to distinguish it from
1536 other serializers.
1537 The "FREEZE" method can return any number of values (i.e. zero or
1539 more). These values and the paclkage/classname of the object will
1540 then be encoded as a tagged JSON value in the following format:
1541 ("classname")[FREEZE return values...]
1543 e.g.:
1545 ("URI")["http://www.google.com/"]
1547 ("MyDate")[2013,10,29]
1548 ("ImageData::JPEG")["Z3...VlCg=="]
1549 For example, the hypothetical "My::Object" "FREEZE" method might
1551 use the objects "type" and "id" members to encode the object:
1552 sub My::Object::FREEZE {
1554 my ($self, $serializer) = @_;
1555 ($self->{type}, $self->{id})
1557 }
1558 2. "convert_blessed" is enabled and the object has a "TO_JSON" method.
1560 In this case, the "TO_JSON" method of the object is invoked in
1561 scalar context. It must return a single scalar that can be directly
1562 encoded into JSON. This scalar replaces the object in the JSON
1563 text.
1564 For example, the following "TO_JSON" method will convert all URI
1566 objects to JSON strings when serialized. The fact that these values
1567 originally were URI objects is lost.
1568 sub URI::TO_JSON {
1570 my ($uri) = @_;
1571 $uri->as_string
1572 }
1573 3. "convert_blessed" is enabled and the object has a stringification
1575 overload.
1576 In this case, the overloaded "" method of the object is invoked in
1577 scalar context. It must return a single scalar that can be directly
1578 encoded into JSON. This scalar replaces the object in the JSON
1579 text.
1580 For example, the following "" method will convert all URI objects
1582 to JSON strings when serialized. The fact that these values
1583 originally were URI objects is lost.
1584 package URI;
1586 use overload '""' => sub { shift->as_string };
1587 4. "allow_blessed" is enabled.
1589 The object will be serialized as a JSON null value.
1590 5. none of the above
1592 If none of the settings are enabled or the respective methods are
1593 missing, "Cpanel::JSON::XS" throws an exception.
1594 DESERIALIZATION
1596 For deserialization there are only two cases to consider: either
1598 nonstandard tagging was used, in which case "allow_tags" decides, or
1599 objects cannot be automatically be deserialized, in which case you can
1600 use postprocessing or the "filter_json_object" or
1601 "filter_json_single_key_object" callbacks to get some real objects our
1602 of your JSON.
1603 This section only considers the tagged value case: I a tagged JSON
1605 object is encountered during decoding and "allow_tags" is disabled, a
1606 parse error will result (as if tagged values were not part of the
1607 grammar).
1608 If "allow_tags" is enabled, "Cpanel::JSON::XS" will look up the "THAW"
1610 method of the package/classname used during serialization (it will not
1611 attempt to load the package as a Perl module). If there is no such
1612 method, the decoding will fail with an error.
1613 Otherwise, the "THAW" method is invoked with the classname as first
1615 argument, the constant string "JSON" as second argument, and all the
1616 values from the JSON array (the values originally returned by the
1617 "FREEZE" method) as remaining arguments.
1618 The method must then return the object. While technically you can
1620 return any Perl scalar, you might have to enable the "enable_nonref"
1621 setting to make that work in all cases, so better return an actual
1622 blessed reference.
1623 As an example, let's implement a "THAW" function that regenerates the
1625 "My::Object" from the "FREEZE" example earlier:
1626 sub My::Object::THAW {
1628 my ($class, $serializer, $type, $id) = @_;
1629 $class->new (type => $type, id => $id)
1631 }
1632 See the "SECURITY CONSIDERATIONS" section below. Allowing external json
1634 objects being deserialized to perl objects is usually a very bad idea.
1635
1637 The interested reader might have seen a number of flags that signify
1638 encodings or codesets - "utf8", "latin1", "binary" and "ascii". There
1639 seems to be some confusion on what these do, so here is a short
1640 comparison:
1641 "utf8" controls whether the JSON text created by "encode" (and expected
1643 by "decode") is UTF-8 encoded or not, while "latin1" and "ascii" only
1644 control whether "encode" escapes character values outside their
1645 respective codeset range. Neither of these flags conflict with each
1646 other, although some combinations make less sense than others.
1647 Care has been taken to make all flags symmetrical with respect to
1649 "encode" and "decode", that is, texts encoded with any combination of
1650 these flag values will be correctly decoded when the same flags are
1651 used - in general, if you use different flag settings while encoding
1652 vs. when decoding you likely have a bug somewhere.
1653 Below comes a verbose discussion of these flags. Note that a "codeset"
1655 is simply an abstract set of character-codepoint pairs, while an
1656 encoding takes those codepoint numbers and encodes them, in our case
1657 into octets. Unicode is (among other things) a codeset, UTF-8 is an
1658 encoding, and ISO-8859-1 (= latin 1) and ASCII are both codesets and
1659 encodings at the same time, which can be confusing.
1660 "utf8" flag disabled
1662 When "utf8" is disabled (the default), then "encode"/"decode"
1663 generate and expect Unicode strings, that is, characters with high
1664 ordinal Unicode values (> 255) will be encoded as such characters,
1665 and likewise such characters are decoded as-is, no changes to them
1666 will be done, except "(re-)interpreting" them as Unicode codepoints
1667 or Unicode characters, respectively (to Perl, these are the same
1668 thing in strings unless you do funny/weird/dumb stuff).
1669 This is useful when you want to do the encoding yourself (e.g. when
1671 you want to have UTF-16 encoded JSON texts) or when some other
1672 layer does the encoding for you (for example, when printing to a
1673 terminal using a filehandle that transparently encodes to UTF-8 you
1674 certainly do NOT want to UTF-8 encode your data first and have Perl
1675 encode it another time).
1676 "utf8" flag enabled
1678 If the "utf8"-flag is enabled, "encode"/"decode" will encode all
1679 characters using the corresponding UTF-8 multi-byte sequence, and
1680 will expect your input strings to be encoded as UTF-8, that is, no
1681 "character" of the input string must have any value > 255, as UTF-8
1682 does not allow that.
1683 The "utf8" flag therefore switches between two modes: disabled
1685 means you will get a Unicode string in Perl, enabled means you get
1686 an UTF-8 encoded octet/binary string in Perl.
1687 "latin1", "binary" or "ascii" flags enabled
1689 With "latin1" (or "ascii") enabled, "encode" will escape characters
1690 with ordinal values > 255 (> 127 with "ascii") and encode the
1691 remaining characters as specified by the "utf8" flag. With
1692 "binary" enabled, ordinal values > 255 are illegal.
1693 If "utf8" is disabled, then the result is also correctly encoded in
1695 those character sets (as both are proper subsets of Unicode,
1696 meaning that a Unicode string with all character values < 256 is
1697 the same thing as a ISO-8859-1 string, and a Unicode string with
1698 all character values < 128 is the same thing as an ASCII string in
1699 Perl).
1700 If "utf8" is enabled, you still get a correct UTF-8-encoded string,
1702 regardless of these flags, just some more characters will be
1703 escaped using "\uXXXX" then before.
1704 Note that ISO-8859-1-encoded strings are not compatible with UTF-8
1706 encoding, while ASCII-encoded strings are. That is because the
1707 ISO-8859-1 encoding is NOT a subset of UTF-8 (despite the
1708 ISO-8859-1 codeset being a subset of Unicode), while ASCII is.
1709 Surprisingly, "decode" will ignore these flags and so treat all
1711 input values as governed by the "utf8" flag. If it is disabled,
1712 this allows you to decode ISO-8859-1- and ASCII-encoded strings, as
1713 both strict subsets of Unicode. If it is enabled, you can correctly
1714 decode UTF-8 encoded strings.
1715 So neither "latin1", "binary" nor "ascii" are incompatible with the
1717 "utf8" flag - they only govern when the JSON output engine escapes
1718 a character or not.
1719 The main use for "latin1" or "binary" is to relatively efficiently
1721 store binary data as JSON, at the expense of breaking compatibility
1722 with most JSON decoders.
1723 The main use for "ascii" is to force the output to not contain
1725 characters with values > 127, which means you can interpret the
1726 resulting string as UTF-8, ISO-8859-1, ASCII, KOI8-R or most about
1727 any character set and 8-bit-encoding, and still get the same data
1728 structure back. This is useful when your channel for JSON transfer
1729 is not 8-bit clean or the encoding might be mangled in between
1730 (e.g. in mail), and works because ASCII is a proper subset of most
1731 8-bit and multibyte encodings in use in the world.
1732 JSON and ECMAscript
1734 JSON syntax is based on how literals are represented in javascript (the
1735 not-standardized predecessor of ECMAscript) which is presumably why it
1736 is called "JavaScript Object Notation".
1737 However, JSON is not a subset (and also not a superset of course) of
1739 ECMAscript (the standard) or javascript (whatever browsers actually
1740 implement).
1741 If you want to use javascript's "eval" function to "parse" JSON, you
1743 might run into parse errors for valid JSON texts, or the resulting data
1744 structure might not be queryable:
1745 One of the problems is that U+2028 and U+2029 are valid characters
1747 inside JSON strings, but are not allowed in ECMAscript string literals,
1748 so the following Perl fragment will not output something that can be
1749 guaranteed to be parsable by javascript's "eval":
1750 use Cpanel::JSON::XS;
1752 print encode_json [chr 0x2028];
1754 The right fix for this is to use a proper JSON parser in your
1756 javascript programs, and not rely on "eval" (see for example Douglas
1757 Crockford's json2.js parser).
1758 If this is not an option, you can, as a stop-gap measure, simply encode
1760 to ASCII-only JSON:
1761 use Cpanel::JSON::XS;
1763 print Cpanel::JSON::XS->new->ascii->encode ([chr 0x2028]);
1765 Note that this will enlarge the resulting JSON text quite a bit if you
1767 have many non-ASCII characters. You might be tempted to run some
1768 regexes to only escape U+2028 and U+2029, e.g.:
1769 # DO NOT USE THIS!
1771 my $json = Cpanel::JSON::XS->new->utf8->encode ([chr 0x2028]);
1772 $json =~ s/\xe2\x80\xa8/\\u2028/g; # escape U+2028
1773 $json =~ s/\xe2\x80\xa9/\\u2029/g; # escape U+2029
1774 print $json;
1775 Note that this is a bad idea: the above only works for U+2028 and
1777 U+2029 and thus only for fully ECMAscript-compliant parsers. Many
1778 existing javascript implementations, however, have issues with other
1779 characters as well - using "eval" naively simply will cause problems.
1780 Another problem is that some javascript implementations reserve some
1782 property names for their own purposes (which probably makes them non-
1783 ECMAscript-compliant). For example, Iceweasel reserves the "__proto__"
1784 property name for its own purposes.
1785 If that is a problem, you could parse try to filter the resulting JSON
1787 output for these property strings, e.g.:
1788 $json =~ s/"__proto__"\s*:/"__proto__renamed":/g;
1790 This works because "__proto__" is not valid outside of strings, so
1792 every occurrence of ""__proto__"\s*:" must be a string used as property
1793 name.
1794 Unicode non-characters between U+FFFD and U+10FFFF are decoded either
1796 to the recommended U+FFFD REPLACEMENT CHARACTER (see Unicode PR #121:
1797 Recommended Practice for Replacement Characters), or in the binary or
1798 relaxed mode left as is, keeping the illegal non-characters as before.
1799 Raw non-Unicode characters outside the valid unicode range fail now to
1801 parse, because "A string is a sequence of zero or more Unicode
1802 characters" RFC 7159 section 1 and "JSON text SHALL be encoded in
1803 Unicode RFC 7159 section 8.1. We use now the UTF8_DISALLOW_SUPER flag
1804 when parsing unicode.
1805 If you know of other incompatibilities, please let me know.
1807 JSON and YAML
1809 You often hear that JSON is a subset of YAML. in general, there is no
1810 way to configure JSON::XS to output a data structure as valid YAML that
1811 works in all cases. If you really must use Cpanel::JSON::XS to
1812 generate YAML, you should use this algorithm (subject to change in
1813 future versions):
1814 my $to_yaml = Cpanel::JSON::XS->new->utf8->space_after (1);
1816 my $yaml = $to_yaml->encode ($ref) . "\n";
1817 This will usually generate JSON texts that also parse as valid YAML.
1819 SPEED
1821 It seems that JSON::XS is surprisingly fast, as shown in the following
1822 tables. They have been generated with the help of the "eg/bench"
1823 program in the JSON::XS distribution, to make it easy to compare on
1824 your own system.
1825 JSON::XS is with Data::MessagePack and Sereal one of the fastest
1827 serializers, because JSON and JSON::XS do not support backrefs (no
1828 graph structures), only trees. Storable supports backrefs, i.e. graphs.
1829 Data::MessagePack encodes its data binary (as Storable) and supports
1830 only very simple subset of JSON.
1831 First comes a comparison between various modules using a very short
1833 single-line JSON string (also available at
1834 <http://dist.schmorp.de/misc/json/short.json>).
1835 {"method": "handleMessage", "params": ["user1",
1837 "we were just talking"], "id": null, "array":[1,11,234,-5,1e5,1e7,
1838 1, 0]}
1839 It shows the number of encodes/decodes per second (JSON::XS uses the
1841 functional interface, while Cpanel::JSON::XS/2 uses the OO interface
1842 with pretty-printing and hash key sorting enabled, Cpanel::JSON::XS/3
1843 enables shrink. JSON::DWIW/DS uses the deserialize function, while
1844 JSON::DWIW::FJ uses the from_json method). Higher is better:
1845 module | encode | decode |
1847 --------------|------------|------------|
1848 JSON::DWIW/DS | 86302.551 | 102300.098 |
1849 JSON::DWIW/FJ | 86302.551 | 75983.768 |
1850 JSON::PP | 15827.562 | 6638.658 |
1851 JSON::Syck | 63358.066 | 47662.545 |
1852 JSON::XS | 511500.488 | 511500.488 |
1853 JSON::XS/2 | 291271.111 | 388361.481 |
1854 JSON::XS/3 | 361577.931 | 361577.931 |
1855 Storable | 66788.280 | 265462.278 |
1856 --------------+------------+------------+
1857 That is, JSON::XS is almost six times faster than JSON::DWIW on
1859 encoding, about five times faster on decoding, and over thirty to
1860 seventy times faster than JSON's pure perl implementation. It also
1861 compares favourably to Storable for small amounts of data.
1862 Using a longer test string (roughly 18KB, generated from Yahoo! Locals
1864 search API (<http://dist.schmorp.de/misc/json/long.json>).
1865 module | encode | decode |
1867 --------------|------------|------------|
1868 JSON::DWIW/DS | 1647.927 | 2673.916 |
1869 JSON::DWIW/FJ | 1630.249 | 2596.128 |
1870 JSON::PP | 400.640 | 62.311 |
1871 JSON::Syck | 1481.040 | 1524.869 |
1872 JSON::XS | 20661.596 | 9541.183 |
1873 JSON::XS/2 | 10683.403 | 9416.938 |
1874 JSON::XS/3 | 20661.596 | 9400.054 |
1875 Storable | 19765.806 | 10000.725 |
1876 --------------+------------+------------+
1877 Again, JSON::XS leads by far (except for Storable which non-
1879 surprisingly decodes a bit faster).
1880 On large strings containing lots of high Unicode characters, some
1882 modules (such as JSON::PC) seem to decode faster than JSON::XS, but the
1883 result will be broken due to missing (or wrong) Unicode handling.
1884 Others refuse to decode or encode properly, so it was impossible to
1885 prepare a fair comparison table for that case.
1886 For updated graphs see
1888 <https://github.com/Sereal/Sereal/wiki/Sereal-Comparison-Graphs>
1889
1891 As long as you only serialize data that can be directly expressed in
1892 JSON, "Cpanel::JSON::XS" is incapable of generating invalid JSON output
1893 (modulo bugs, but "JSON::XS" has found more bugs in the official JSON
1894 testsuite (1) than the official JSON testsuite has found in "JSON::XS"
1895 (0)). "Cpanel::JSON::XS" is currently the only known JSON decoder
1896 which passes all <http://seriot.ch/parsing_json.html> tests, while
1897 being the fastest also.
1898 When you have trouble decoding JSON generated by this module using
1900 other decoders, then it is very likely that you have an encoding
1901 mismatch or the other decoder is broken.
1902 When decoding, "JSON::XS" is strict by default and will likely catch
1904 all errors. There are currently two settings that change this:
1905 "relaxed" makes "JSON::XS" accept (but not generate) some non-standard
1906 extensions, and "allow_tags" or "allow_blessed" will allow you to
1907 encode and decode Perl objects, at the cost of being totally insecure
1908 and not outputting valid JSON anymore.
1909 JSON-XS-3.01 broke interoperability with JSON-2.90 with booleans. See
1911 JSON.
1912 Cpanel::JSON::XS needs to know the JSON and JSON::XS versions to be
1914 able work with those objects, especially when encoding a booleans like
1915 "{"is_true":true}". So you need to load these modules before.
1916 true/false overloading and boolean representations are supported.
1918 JSON::XS and JSON::PP representations are accepted and older JSON::XS
1920 accepts Cpanel::JSON::XS booleans. All JSON modules JSON, JSON, PP,
1921 JSON::XS, Cpanel::JSON::XS produce JSON::PP::Boolean objects, just Mojo
1922 and JSON::YAJL not. Mojo produces Mojo::JSON::_Bool and
1923 JSON::YAJL::Parser just an unblessed IV.
1924 Cpanel::JSON::XS accepts JSON::PP::Boolean and Mojo::JSON::_Bool
1926 objects as booleans.
1927 I cannot think of any reason to still use JSON::XS anymore.
1929 TAGGED VALUE SYNTAX AND STANDARD JSON EN/DECODERS
1931 When you use "allow_tags" to use the extended (and also nonstandard and
1932 invalid) JSON syntax for serialized objects, and you still want to
1933 decode the generated serialize objects, you can run a regex to replace
1934 the tagged syntax by standard JSON arrays (it only works for "normal"
1935 package names without comma, newlines or single colons). First, the
1936 readable Perl version:
1937 # if your FREEZE methods return no values, you need this replace first:
1939 $json =~ s/\( \s* (" (?: [^\\":,]+|\\.|::)* ") \s* \) \s* \[\s*\]/[$1]/gx;
1940 # this works for non-empty constructor arg lists:
1942 $json =~ s/\( \s* (" (?: [^\\":,]+|\\.|::)* ") \s* \) \s* \[/[$1,/gx;
1943 And here is a less readable version that is easy to adapt to other
1945 languages:
1946 $json =~ s/\(\s*("([^\\":,]+|\\.|::)*")\s*\)\s*\[/[$1,/g;
1948 Here is an ECMAScript version (same regex):
1950 json = json.replace (/\(\s*("([^\\":,]+|\\.|::)*")\s*\)\s*\[/g, "[$1,");
1952 Since this syntax converts to standard JSON arrays, it might be hard to
1954 distinguish serialized objects from normal arrays. You can prepend a
1955 "magic number" as first array element to reduce chances of a collision:
1956 $json =~ s/\(\s*("([^\\":,]+|\\.|::)*")\s*\)\s*\[/["XU1peReLzT4ggEllLanBYq4G9VzliwKF",$1,/g;
1958 And after decoding the JSON text, you could walk the data structure
1960 looking for arrays with a first element of
1961 "XU1peReLzT4ggEllLanBYq4G9VzliwKF".
1962 The same approach can be used to create the tagged format with another
1964 encoder. First, you create an array with the magic string as first
1965 member, the classname as second, and constructor arguments last, encode
1966 it as part of your JSON structure, and then:
1967 $json =~ s/\[\s*"XU1peReLzT4ggEllLanBYq4G9VzliwKF"\s*,\s*("([^\\":,]+|\\.|::)*")\s*,/($1)[/g;
1969 Again, this has some limitations - the magic string must not be encoded
1971 with character escapes, and the constructor arguments must be non-
1972 empty.
1973
1975 Since this module was written, Google has written a new JSON RFC, RFC
1976 7159 (and RFC7158). Unfortunately, this RFC breaks compatibility with
1977 both the original JSON specification on www.json.org and RFC4627.
1978 As far as I can see, you can get partial compatibility when parsing by
1980 using "->allow_nonref". However, consider the security implications of
1981 doing so.
1982 I haven't decided yet when to break compatibility with RFC4627 by
1984 default (and potentially leave applications insecure) and change the
1985 default to follow RFC7159, but application authors are well advised to
1986 call "->allow_nonref(0)" even if this is the current default, if they
1987 cannot handle non-reference values, in preparation for the day when the
1988 default will change.
1989
1991 JSON::XS and Cpanel::JSON::XS are not only fast. JSON is generally the
1992 most secure serializing format, because it is the only one besides
1993 Data::MessagePack, which does not deserialize objects per default. For
1994 all languages, not just perl. The binary variant BSON (MongoDB) does
1995 more but is unsafe.
1996 It is trivial for any attacker to create such serialized objects in
1998 JSON and trick perl into expanding them, thereby triggering certain
1999 methods. Watch <https://www.youtube.com/watch?v=Gzx6KlqiIZE> for an
2000 exploit demo for "CVE-2015-1592 SixApart MovableType Storable Perl Code
2001 Execution" for a deserializer which expands objects. Deserializing
2002 even coderefs (methods, functions) or external data would be considered
2003 the most dangerous.
2004 Security relevant overview of serializers regarding deserializing
2006 objects by default:
2007 Objects Coderefs External Data
2009 Data::Dumper YES YES YES
2011 Storable YES NO (def) NO
2012 Sereal YES NO NO
2013 YAML YES NO NO
2014 B::C YES YES YES
2015 B::Bytecode YES YES YES
2016 BSON YES YES NO
2017 JSON::SL YES NO YES
2018 JSON NO (def) NO NO
2019 Data::MessagePack NO NO NO
2020 XML NO NO YES
2021 Pickle YES YES YES
2023 PHP Deserialize YES NO NO
2024 When you are using JSON in a protocol, talking to untrusted potentially
2026 hostile creatures requires relatively few measures.
2027 First of all, your JSON decoder should be secure, that is, should not
2029 have any buffer overflows. Obviously, this module should ensure that.
2030 Second, you need to avoid resource-starving attacks. That means you
2032 should limit the size of JSON texts you accept, or make sure then when
2033 your resources run out, that's just fine (e.g. by using a separate
2034 process that can crash safely). The size of a JSON text in octets or
2035 characters is usually a good indication of the size of the resources
2036 required to decode it into a Perl structure. While JSON::XS can check
2037 the size of the JSON text, it might be too late when you already have
2038 it in memory, so you might want to check the size before you accept the
2039 string.
2040 Third, Cpanel::JSON::XS recurses using the C stack when decoding
2042 objects and arrays. The C stack is a limited resource: for instance, on
2043 my amd64 machine with 8MB of stack size I can decode around 180k nested
2044 arrays but only 14k nested JSON objects (due to perl itself recursing
2045 deeply on croak to free the temporary). If that is exceeded, the
2046 program crashes. To be conservative, the default nesting limit is set
2047 to 512. If your process has a smaller stack, you should adjust this
2048 setting accordingly with the "max_depth" method.
2049 Also keep in mind that Cpanel::JSON::XS might leak contents of your
2051 Perl data structures in its error messages, so when you serialize
2052 sensitive information you might want to make sure that exceptions
2053 thrown by JSON::XS will not end up in front of untrusted eyes.
2054 If you are using Cpanel::JSON::XS to return packets to consumption by
2056 JavaScript scripts in a browser you should have a look at
2057 <http://blog.archive.jpsykes.com/47/practical-csrf-and-json-security/>
2058 to see whether you are vulnerable to some common attack vectors (which
2059 really are browser design bugs, but it is still you who will have to
2060 deal with it, as major browser developers care only for features, not
2061 about getting security right). You might also want to also look at
2062 Mojo::JSON special escape rules to prevent from XSS attacks.
2063
2065 TL;DR: Due to security concerns, Cpanel::JSON::XS will not allow scalar
2066 data in JSON texts by default - you need to create your own
2067 Cpanel::JSON::XS object and enable "allow_nonref":
2068 my $json = JSON::XS->new->allow_nonref;
2070 $text = $json->encode ($data);
2072 $data = $json->decode ($text);
2073 The long version: JSON being an important and supposedly stable format,
2075 the IETF standardized it as RFC 4627 in 2006. Unfortunately the
2076 inventor of JSON Douglas Crockford unilaterally changed the definition
2077 of JSON in javascript. Rather than create a fork, the IETF decided to
2078 standardize the new syntax (apparently, so I as told, without finding
2079 it very amusing).
2080 The biggest difference between the original JSON and the new JSON is
2082 that the new JSON supports scalars (anything other than arrays and
2083 objects) at the top-level of a JSON text. While this is strictly
2084 backwards compatible to older versions, it breaks a number of protocols
2085 that relied on sending JSON back-to-back, and is a minor security
2086 concern.
2087 For example, imagine you have two banks communicating, and on one side,
2089 the JSON coder gets upgraded. Two messages, such as 10 and 1000 might
2090 then be confused to mean 101000, something that couldn't happen in the
2091 original JSON, because neither of these messages would be valid JSON.
2092 If one side accepts these messages, then an upgrade in the coder on
2094 either side could result in this becoming exploitable.
2095 This module has always allowed these messages as an optional extension,
2097 by default disabled. The security concerns are the reason why the
2098 default is still disabled, but future versions might/will likely
2099 upgrade to the newer RFC as default format, so you are advised to check
2100 your implementation and/or override the default with "->allow_nonref
2101 (0)" to ensure that future versions are safe.
2102
2104 Cpanel::JSON::XS has proper ithreads support, unlike JSON::XS. If you
2105 encounter any bugs with thread support please report them.
2106 From Version 4.00 - 4.19 you couldn't encode true with threads::shared
2108 magic.
2109
2111 While the goal of the Cpanel::JSON::XS module is to be correct, that
2112 unfortunately does not mean it's bug-free, only that the author thinks
2113 its design is bug-free. If you keep reporting bugs and tests they will
2114 be fixed swiftly, though.
2115 Since the JSON::XS author refuses to use a public bugtracker and
2117 prefers private emails, we use the tracker at github, so you might want
2118 to report any issues twice. Once in private to MLEHMANN to be fixed in
2119 JSON::XS and one to our the public tracker. Issues fixed by JSON::XS
2120 with a new release will also be backported to Cpanel::JSON::XS and
2121 5.6.2, as long as cPanel relies on 5.6.2 and Cpanel::JSON::XS as our
2122 serializer of choice.
2123 <https://github.com/rurban/Cpanel-JSON-XS/issues>
2125
2127 This module is available under the same licences as perl, the Artistic
2128 license and the GPL.
2129
2131 The cpanel_json_xs command line utility for quick experiments.
2132 JSON, JSON::XS, JSON::MaybeXS, Mojo::JSON, Mojo::JSON::MaybeXS,
2134 JSON::SL, JSON::DWIW, JSON::YAJL, JSON::Any, Test::JSON,
2135 Locale::Wolowitz, <https://metacpan.org/search?q=JSON>
2136 <https://tools.ietf.org/html/rfc7159>
2138 <https://tools.ietf.org/html/rfc4627>
2140
2142 Reini Urban <rurban@cpan.org>
2143 Marc Lehmann <schmorp@schmorp.de>, http://home.schmorp.de/
2145
2147 Reini Urban <rurban@cpan.org>
2148perl v5.32.1 2021-01-27 XS(3)