ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/JSON-XS/XS.pm
(Generate patch)

Comparing JSON-XS/XS.pm (file contents):
Revision 1.22 by root, Sun Mar 25 02:37:00 2007 UTC vs.
Revision 1.44 by root, Mon Jun 25 04:08:17 2007 UTC

85 85
86package JSON::XS; 86package JSON::XS;
87 87
88use strict; 88use strict;
89 89
90BEGIN {
91 our $VERSION = '0.8'; 90our $VERSION = '1.4';
92 our @ISA = qw(Exporter); 91our @ISA = qw(Exporter);
93 92
94 our @EXPORT = qw(to_json from_json objToJson jsonToObj); 93our @EXPORT = qw(to_json from_json objToJson jsonToObj);
95 require Exporter;
96 94
97 require XSLoader; 95use Exporter;
98 XSLoader::load JSON::XS::, $VERSION; 96use XSLoader;
99}
100 97
101=head1 FUNCTIONAL INTERFACE 98=head1 FUNCTIONAL INTERFACE
102 99
103The following convinience methods are provided by this module. They are 100The following convinience methods are provided by this module. They are
104exported by default: 101exported by default:
127 124
128 $perl_scalar = JSON::XS->new->utf8->decode ($json_text) 125 $perl_scalar = JSON::XS->new->utf8->decode ($json_text)
129 126
130except being faster. 127except being faster.
131 128
129=item $is_boolean = JSON::XS::is_bool $scalar
130
131Returns true if the passed scalar represents either JSON::XS::true or
132JSON::XS::false, two constants that act like C<1> and C<0>, respectively
133and are used to represent JSON C<true> and C<false> values in Perl.
134
135See MAPPING, below, for more information on how JSON values are mapped to
136Perl.
137
132=back 138=back
139
133 140
134=head1 OBJECT-ORIENTED INTERFACE 141=head1 OBJECT-ORIENTED INTERFACE
135 142
136The object oriented interface lets you configure your own encoding or 143The object oriented interface lets you configure your own encoding or
137decoding style, within the limits of supported formats. 144decoding style, within the limits of supported formats.
153 160
154If C<$enable> is true (or missing), then the C<encode> method will not 161If C<$enable> is true (or missing), then the C<encode> method will not
155generate characters outside the code range C<0..127> (which is ASCII). Any 162generate characters outside the code range C<0..127> (which is ASCII). Any
156unicode characters outside that range will be escaped using either a 163unicode characters outside that range will be escaped using either a
157single \uXXXX (BMP characters) or a double \uHHHH\uLLLLL escape sequence, 164single \uXXXX (BMP characters) or a double \uHHHH\uLLLLL escape sequence,
158as per RFC4627. 165as per RFC4627. The resulting encoded JSON text can be treated as a native
166unicode string, an ascii-encoded, latin1-encoded or UTF-8 encoded string,
167or any other superset of ASCII.
159 168
160If C<$enable> is false, then the C<encode> method will not escape Unicode 169If C<$enable> is false, then the C<encode> method will not escape Unicode
161characters unless required by the JSON syntax. This results in a faster 170characters unless required by the JSON syntax or other flags. This results
162and more compact format. 171in a faster and more compact format.
172
173The main use for this flag is to produce JSON texts that can be
174transmitted over a 7-bit channel, as the encoded JSON texts will not
175contain any 8 bit characters.
163 176
164 JSON::XS->new->ascii (1)->encode ([chr 0x10401]) 177 JSON::XS->new->ascii (1)->encode ([chr 0x10401])
165 => ["\ud801\udc01"] 178 => ["\ud801\udc01"]
179
180=item $json = $json->latin1 ([$enable])
181
182If C<$enable> is true (or missing), then the C<encode> method will encode
183the resulting JSON text as latin1 (or iso-8859-1), escaping any characters
184outside the code range C<0..255>. The resulting string can be treated as a
185latin1-encoded JSON text or a native unicode string. The C<decode> method
186will not be affected in any way by this flag, as C<decode> by default
187expects unicode, which is a strict superset of latin1.
188
189If C<$enable> is false, then the C<encode> method will not escape Unicode
190characters unless required by the JSON syntax or other flags.
191
192The main use for this flag is efficiently encoding binary data as JSON
193text, as most octets will not be escaped, resulting in a smaller encoded
194size. The disadvantage is that the resulting JSON text is encoded
195in latin1 (and must correctly be treated as such when storing and
196transfering), a rare encoding for JSON. It is therefore most useful when
197you want to store data structures known to contain binary data efficiently
198in files or databases, not when talking to other JSON encoders/decoders.
199
200 JSON::XS->new->latin1->encode (["\x{89}\x{abc}"]
201 => ["\x{89}\\u0abc"] # (perl syntax, U+abc escaped, U+89 not)
166 202
167=item $json = $json->utf8 ([$enable]) 203=item $json = $json->utf8 ([$enable])
168 204
169If C<$enable> is true (or missing), then the C<encode> method will encode 205If C<$enable> is true (or missing), then the C<encode> method will encode
170the JSON result into UTF-8, as required by many protocols, while the 206the JSON result into UTF-8, as required by many protocols, while the
280resulting in an invalid JSON text: 316resulting in an invalid JSON text:
281 317
282 JSON::XS->new->allow_nonref->encode ("Hello, World!") 318 JSON::XS->new->allow_nonref->encode ("Hello, World!")
283 => "Hello, World!" 319 => "Hello, World!"
284 320
321=item $json = $json->allow_blessed ([$enable])
322
323If C<$enable> is true (or missing), then the C<encode> method will not
324barf when it encounters a blessed reference. Instead, the value of the
325B<convert_blessed> option will decide wether C<null> (C<convert_blessed>
326disabled or no C<to_json> method found) or a representation of the
327object (C<convert_blessed> enabled and C<to_json> method found) is being
328encoded. Has no effect on C<decode>.
329
330If C<$enable> is false (the default), then C<encode> will throw an
331exception when it encounters a blessed object.
332
333=item $json = $json->convert_blessed ([$enable])
334
335If C<$enable> is true (or missing), then C<encode>, upon encountering a
336blessed object, will check for the availability of the C<TO_JSON> method
337on the object's class. If found, it will be called in scalar context
338and the resulting scalar will be encoded instead of the object. If no
339C<TO_JSON> method is found, the value of C<allow_blessed> will decide what
340to do.
341
342The C<TO_JSON> method may safely call die if it wants. If C<TO_JSON>
343returns other blessed objects, those will be handled in the same
344way. C<TO_JSON> must take care of not causing an endless recursion cycle
345(== crash) in this case. The name of C<TO_JSON> was chosen because other
346methods called by the Perl core (== not the user of the object) are
347usually in upper case letters and to avoid collisions with the C<to_json>
348function.
349
350If C<$enable> is false, then the C<allow_blessed> setting will decide what
351to do when a blessed object is found.
352
285=item $json = $json->shrink ([$enable]) 353=item $json = $json->shrink ([$enable])
286 354
287Perl usually over-allocates memory a bit when allocating space for 355Perl usually over-allocates memory a bit when allocating space for
288strings. This flag optionally resizes strings generated by either 356strings. This flag optionally resizes strings generated by either
289C<encode> or C<decode> to their minimum size possible. This can save 357C<encode> or C<decode> to their minimum size possible. This can save
290memory when your JSON texts are either very very long or you have many 358memory when your JSON texts are either very very long or you have many
291short strings. It will also try to downgrade any strings to octet-form 359short strings. It will also try to downgrade any strings to octet-form
292if possible: perl stores strings internally either in an encoding called 360if possible: perl stores strings internally either in an encoding called
293UTF-X or in octet-form. The latter cannot store everything but uses less 361UTF-X or in octet-form. The latter cannot store everything but uses less
294space in general. 362space in general (and some buggy Perl or C code might even rely on that
363internal representation being used).
295 364
365The actual definition of what shrink does might change in future versions,
366but it will always try to save space at the expense of time.
367
296If C<$enable> is true (or missing), the string returned by C<encode> will be shrunk-to-fit, 368If C<$enable> is true (or missing), the string returned by C<encode> will
297while all strings generated by C<decode> will also be shrunk-to-fit. 369be shrunk-to-fit, while all strings generated by C<decode> will also be
370shrunk-to-fit.
298 371
299If C<$enable> is false, then the normal perl allocation algorithms are used. 372If C<$enable> is false, then the normal perl allocation algorithms are used.
300If you work with your data, then this is likely to be faster. 373If you work with your data, then this is likely to be faster.
301 374
302In the future, this setting might control other things, such as converting 375In the future, this setting might control other things, such as converting
303strings that look like integers or floats into integers or floats 376strings that look like integers or floats into integers or floats
304internally (there is no difference on the Perl level), saving space. 377internally (there is no difference on the Perl level), saving space.
378
379=item $json = $json->max_depth ([$maximum_nesting_depth])
380
381Sets the maximum nesting level (default C<512>) accepted while encoding
382or decoding. If the JSON text or Perl data structure has an equal or
383higher nesting level then this limit, then the encoder and decoder will
384stop and croak at that point.
385
386Nesting level is defined by number of hash- or arrayrefs that the encoder
387needs to traverse to reach a given point or the number of C<{> or C<[>
388characters without their matching closing parenthesis crossed to reach a
389given character in a string.
390
391Setting the maximum depth to one disallows any nesting, so that ensures
392that the object is only a single hash/object or array.
393
394The argument to C<max_depth> will be rounded up to the next nearest power
395of two.
396
397See SECURITY CONSIDERATIONS, below, for more info on why this is useful.
305 398
306=item $json_text = $json->encode ($perl_scalar) 399=item $json_text = $json->encode ($perl_scalar)
307 400
308Converts the given Perl data structure (a simple scalar or a reference 401Converts the given Perl data structure (a simple scalar or a reference
309to a hash or array) to its JSON representation. Simple scalars will be 402to a hash or array) to its JSON representation. Simple scalars will be
319 412
320JSON numbers and strings become simple Perl scalars. JSON arrays become 413JSON numbers and strings become simple Perl scalars. JSON arrays become
321Perl arrayrefs and JSON objects become Perl hashrefs. C<true> becomes 414Perl arrayrefs and JSON objects become Perl hashrefs. C<true> becomes
322C<1>, C<false> becomes C<0> and C<null> becomes C<undef>. 415C<1>, C<false> becomes C<0> and C<null> becomes C<undef>.
323 416
417=item ($perl_scalar, $characters) = $json->decode_prefix ($json_text)
418
419This works like the C<decode> method, but instead of raising an exception
420when there is trailing garbage after the first JSON object, it will
421silently stop parsing there and return the number of characters consumed
422so far.
423
424This is useful if your JSON texts are not delimited by an outer protocol
425(which is not the brightest thing to do in the first place) and you need
426to know where the JSON text ends.
427
428 JSON::XS->new->decode_prefix ("[1] the tail")
429 => ([], 3)
430
324=back 431=back
432
325 433
326=head1 MAPPING 434=head1 MAPPING
327 435
328This section describes how JSON::XS maps Perl values to JSON values and 436This section describes how JSON::XS maps Perl values to JSON values and
329vice versa. These mappings are designed to "do the right thing" in most 437vice versa. These mappings are designed to "do the right thing" in most
331(what you put in comes out as something equivalent). 439(what you put in comes out as something equivalent).
332 440
333For the more enlightened: note that in the following descriptions, 441For the more enlightened: note that in the following descriptions,
334lowercase I<perl> refers to the Perl interpreter, while uppcercase I<Perl> 442lowercase I<perl> refers to the Perl interpreter, while uppcercase I<Perl>
335refers to the abstract Perl language itself. 443refers to the abstract Perl language itself.
444
336 445
337=head2 JSON -> PERL 446=head2 JSON -> PERL
338 447
339=over 4 448=over 4
340 449
361conversion details, but an integer may take slightly less memory and might 470conversion details, but an integer may take slightly less memory and might
362represent more values exactly than (floating point) numbers. 471represent more values exactly than (floating point) numbers.
363 472
364=item true, false 473=item true, false
365 474
366These JSON atoms become C<0>, C<1>, respectively. Information is lost in 475These JSON atoms become C<JSON::XS::true> and C<JSON::XS::false>,
367this process. Future versions might represent those values differently, 476respectively. They are overloaded to act almost exactly like the numbers
368but they will be guarenteed to act like these integers would normally in 477C<1> and C<0>. You can check wether a scalar is a JSON boolean by using
369Perl. 478the C<JSON::XS::is_bool> function.
370 479
371=item null 480=item null
372 481
373A JSON null atom becomes C<undef> in Perl. 482A JSON null atom becomes C<undef> in Perl.
374 483
375=back 484=back
485
376 486
377=head2 PERL -> JSON 487=head2 PERL -> JSON
378 488
379The mapping from Perl to JSON is slightly more difficult, as Perl is a 489The mapping from Perl to JSON is slightly more difficult, as Perl is a
380truly typeless language, so we can only guess which JSON type is meant by 490truly typeless language, so we can only guess which JSON type is meant by
383=over 4 493=over 4
384 494
385=item hash references 495=item hash references
386 496
387Perl hash references become JSON objects. As there is no inherent ordering 497Perl hash references become JSON objects. As there is no inherent ordering
388in hash keys, they will usually be encoded in a pseudo-random order that 498in hash keys (or JSON objects), they will usually be encoded in a
389can change between runs of the same program but stays generally the same 499pseudo-random order that can change between runs of the same program but
390within a single run of a program. JSON::XS can optionally sort the hash 500stays generally the same within a single run of a program. JSON::XS can
391keys (determined by the I<canonical> flag), so the same datastructure 501optionally sort the hash keys (determined by the I<canonical> flag), so
392will serialise to the same JSON text (given same settings and version of 502the same datastructure will serialise to the same JSON text (given same
393JSON::XS), but this incurs a runtime overhead. 503settings and version of JSON::XS), but this incurs a runtime overhead
504and is only rarely useful, e.g. when you want to compare some JSON text
505against another for equality.
394 506
395=item array references 507=item array references
396 508
397Perl array references become JSON arrays. 509Perl array references become JSON arrays.
510
511=item other references
512
513Other unblessed references are generally not allowed and will cause an
514exception to be thrown, except for references to the integers C<0> and
515C<1>, which get turned into C<false> and C<true> atoms in JSON. You can
516also use C<JSON::XS::false> and C<JSON::XS::true> to improve readability.
517
518 to_json [\0,JSON::XS::true] # yields [false,true]
519
520=item JSON::XS::true, JSON::XS::false
521
522These special values become JSON true and JSON false values,
523respectively. You cna alos use C<\1> and C<\0> directly if you want.
398 524
399=item blessed objects 525=item blessed objects
400 526
401Blessed objects are not allowed. JSON::XS currently tries to encode their 527Blessed objects are not allowed. JSON::XS currently tries to encode their
402underlying representation (hash- or arrayref), but this behaviour might 528underlying representation (hash- or arrayref), but this behaviour might
435 $x *= 1; # same thing, the choise is yours. 561 $x *= 1; # same thing, the choise is yours.
436 562
437You can not currently output JSON booleans or force the type in other, 563You can not currently output JSON booleans or force the type in other,
438less obscure, ways. Tell me if you need this capability. 564less obscure, ways. Tell me if you need this capability.
439 565
440=item circular data structures
441
442Those will be encoded until memory or stackspace runs out.
443
444=back 566=back
567
445 568
446=head1 COMPARISON 569=head1 COMPARISON
447 570
448As already mentioned, this module was created because none of the existing 571As already mentioned, this module was created because none of the existing
449JSON modules could be made to work correctly. First I will describe the 572JSON modules could be made to work correctly. First I will describe the
527 650
528Does not check input for validity. 651Does not check input for validity.
529 652
530=back 653=back
531 654
655
656=head2 JSON and YAML
657
658You often hear that JSON is a subset (or a close subset) of YAML. This is,
659however, a mass hysteria and very far from the truth. In general, there is
660no way to configure JSON::XS to output a data structure as valid YAML.
661
662If you really must use JSON::XS to generate YAML, you should use this
663algorithm (subject to change in future versions):
664
665 my $to_yaml = JSON::XS->new->utf8->space_after (1);
666 my $yaml = $to_yaml->encode ($ref) . "\n";
667
668This will usually generate JSON texts that also parse as valid
669YAML. Please note that YAML has hardcoded limits on (simple) object key
670lengths that JSON doesn't have, so you should make sure that your hash
671keys are noticably shorter than the 1024 characters YAML allows.
672
673There might be other incompatibilities that I am not aware of. In general
674you should not try to generate YAML with a JSON generator or vice versa,
675or try to parse JSON with a YAML parser or vice versa: chances are high
676that you will run into severe interoperability problems.
677
678
532=head2 SPEED 679=head2 SPEED
533 680
534It seems that JSON::XS is surprisingly fast, as shown in the following 681It seems that JSON::XS is surprisingly fast, as shown in the following
535tables. They have been generated with the help of the C<eg/bench> program 682tables. They have been generated with the help of the C<eg/bench> program
536in the JSON::XS distribution, to make it easy to compare on your own 683in the JSON::XS distribution, to make it easy to compare on your own
537system. 684system.
538 685
539First comes a comparison between various modules using a very short JSON 686First comes a comparison between various modules using a very short
540string: 687single-line JSON string:
541 688
542 {"method": "handleMessage", "params": ["user1", "we were just talking"], "id": null} 689 {"method": "handleMessage", "params": ["user1", "we were just talking"], \
690 "id": null, "array":[1,11,234,-5,1e5,1e7, true, false]}
543 691
544It shows the number of encodes/decodes per second (JSON::XS uses the 692It shows the number of encodes/decodes per second (JSON::XS uses
545functional interface, while JSON::XS/2 uses the OO interface with 693the functional interface, while JSON::XS/2 uses the OO interface
546pretty-printing and hashkey sorting enabled). Higher is better: 694with pretty-printing and hashkey sorting enabled, JSON::XS/3 enables
695shrink). Higher is better:
547 696
548 module | encode | decode | 697 module | encode | decode |
549 -----------|------------|------------| 698 -----------|------------|------------|
550 JSON | 11488.516 | 7823.035 | 699 JSON | 7645.468 | 4208.613 |
551 JSON::DWIW | 94708.054 | 129094.260 | 700 JSON::DWIW | 40721.398 | 77101.176 |
552 JSON::PC | 63884.157 | 128528.212 | 701 JSON::PC | 65948.176 | 78251.940 |
553 JSON::Syck | 34898.677 | 42096.911 | 702 JSON::Syck | 22844.793 | 26479.192 |
554 JSON::XS | 654027.064 | 396423.669 | 703 JSON::XS | 388361.481 | 199728.762 |
555 JSON::XS/2 | 371564.190 | 371725.613 | 704 JSON::XS/2 | 218453.333 | 192399.266 |
705 JSON::XS/3 | 338250.323 | 192399.266 |
706 Storable | 15779.925 | 14169.946 |
556 -----------+------------+------------+ 707 -----------+------------+------------+
557 708
558That is, JSON::XS is more than six times faster than JSON::DWIW on 709That is, JSON::XS is about five times faster than JSON::DWIW on encoding,
559encoding, more than three times faster on decoding, and about thirty times 710about three times faster on decoding, and over fourty times faster
560faster than JSON, even with pretty-printing and key sorting. 711than JSON, even with pretty-printing and key sorting. It also compares
712favourably to Storable for small amounts of data.
561 713
562Using a longer test string (roughly 18KB, generated from Yahoo! Locals 714Using a longer test string (roughly 18KB, generated from Yahoo! Locals
563search API (http://nanoref.com/yahooapis/mgPdGg): 715search API (http://nanoref.com/yahooapis/mgPdGg):
564 716
565 module | encode | decode | 717 module | encode | decode |
566 -----------|------------|------------| 718 -----------|------------|------------|
567 JSON | 273.023 | 44.674 | 719 JSON | 254.685 | 37.665 |
568 JSON::DWIW | 1089.383 | 1145.704 | 720 JSON::DWIW | 843.343 | 1049.731 |
569 JSON::PC | 3097.419 | 2393.921 | 721 JSON::PC | 3602.116 | 2307.352 |
570 JSON::Syck | 514.060 | 843.053 | 722 JSON::Syck | 505.107 | 787.899 |
571 JSON::XS | 6479.668 | 3636.364 | 723 JSON::XS | 5747.196 | 3690.220 |
572 JSON::XS/2 | 3774.221 | 3599.124 | 724 JSON::XS/2 | 3968.121 | 3676.634 |
725 JSON::XS/3 | 6105.246 | 3662.508 |
726 Storable | 4417.337 | 5285.161 |
573 -----------+------------+------------+ 727 -----------+------------+------------+
574 728
575Again, JSON::XS leads by far. 729Again, JSON::XS leads by far (except for Storable which non-surprisingly
730decodes faster).
576 731
577On large strings containing lots of high unicode characters, some modules 732On large strings containing lots of high unicode characters, some modules
578(such as JSON::PC) seem to decode faster than JSON::XS, but the result 733(such as JSON::PC) seem to decode faster than JSON::XS, but the result
579will be broken due to missing (or wrong) unicode handling. Others refuse 734will be broken due to missing (or wrong) unicode handling. Others refuse
580to decode or encode properly, so it was impossible to prepare a fair 735to decode or encode properly, so it was impossible to prepare a fair
581comparison table for that case. 736comparison table for that case.
582 737
583=head1 RESOURCE LIMITS
584 738
585JSON::XS does not impose any limits on the size of JSON texts or Perl 739=head1 SECURITY CONSIDERATIONS
586values they represent - if your machine can handle it, JSON::XS will 740
587encode or decode it. Future versions might optionally impose structure 741When you are using JSON in a protocol, talking to untrusted potentially
588depth and memory use resource limits. 742hostile creatures requires relatively few measures.
743
744First of all, your JSON decoder should be secure, that is, should not have
745any buffer overflows. Obviously, this module should ensure that and I am
746trying hard on making that true, but you never know.
747
748Second, you need to avoid resource-starving attacks. That means you should
749limit the size of JSON texts you accept, or make sure then when your
750resources run out, thats just fine (e.g. by using a separate process that
751can crash safely). The size of a JSON text in octets or characters is
752usually a good indication of the size of the resources required to decode
753it into a Perl structure.
754
755Third, JSON::XS recurses using the C stack when decoding objects and
756arrays. The C stack is a limited resource: for instance, on my amd64
757machine with 8MB of stack size I can decode around 180k nested arrays but
758only 14k nested JSON objects (due to perl itself recursing deeply on croak
759to free the temporary). If that is exceeded, the program crashes. to be
760conservative, the default nesting limit is set to 512. If your process
761has a smaller stack, you should adjust this setting accordingly with the
762C<max_depth> method.
763
764And last but least, something else could bomb you that I forgot to think
765of. In that case, you get to keep the pieces. I am always open for hints,
766though...
767
768If you are using JSON::XS to return packets to consumption
769by javascript scripts in a browser you should have a look at
770L<http://jpsykes.com/47/practical-csrf-and-json-security> to see wether
771you are vulnerable to some common attack vectors (which really are browser
772design bugs, but it is still you who will have to deal with it, as major
773browser developers care only for features, not about doing security
774right).
775
589 776
590=head1 BUGS 777=head1 BUGS
591 778
592While the goal of this module is to be correct, that unfortunately does 779While the goal of this module is to be correct, that unfortunately does
593not mean its bug-free, only that I think its design is bug-free. It is 780not mean its bug-free, only that I think its design is bug-free. It is
594still very young and not well-tested. If you keep reporting bugs they will 781still relatively early in its development. If you keep reporting bugs they
595be fixed swiftly, though. 782will be fixed swiftly, though.
596 783
597=cut 784=cut
785
786our $true = do { bless \(my $dummy = 1), "JSON::XS::Boolean" };
787our $false = do { bless \(my $dummy = 0), "JSON::XS::Boolean" };
788
789sub true() { $true }
790sub false() { $false }
791
792sub is_bool($) {
793 UNIVERSAL::isa $_[0], "JSON::XS::Boolean"
794# or UNIVERSAL::isa $_[0], "JSON::Literal"
795}
796
797XSLoader::load "JSON::XS", $VERSION;
798
799package JSON::XS::Boolean;
800
801use overload
802 "0+" => sub { ${$_[0]} },
803 "++" => sub { $_[0] = ${$_[0]} + 1 },
804 "--" => sub { $_[0] = ${$_[0]} - 1 },
805 fallback => 1;
598 806
5991; 8071;
600 808
601=head1 AUTHOR 809=head1 AUTHOR
602 810

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines