… | |
… | |
49 | to write yet another JSON module? While it seems there are many JSON |
49 | to write yet another JSON module? While it seems there are many JSON |
50 | modules, none of them correctly handle all corner cases, and in most cases |
50 | modules, none of them correctly handle all corner cases, and in most cases |
51 | their maintainers are unresponsive, gone missing, or not listening to bug |
51 | their maintainers are unresponsive, gone missing, or not listening to bug |
52 | reports for other reasons. |
52 | reports for other reasons. |
53 | |
53 | |
54 | See COMPARISON, below, for a comparison to some other JSON modules. |
|
|
55 | |
|
|
56 | See MAPPING, below, on how JSON::XS maps perl values to JSON values and |
54 | See MAPPING, below, on how JSON::XS maps perl values to JSON values and |
57 | vice versa. |
55 | vice versa. |
58 | |
56 | |
59 | =head2 FEATURES |
57 | =head2 FEATURES |
60 | |
58 | |
… | |
… | |
66 | so, and even documents what "correct" means. |
64 | so, and even documents what "correct" means. |
67 | |
65 | |
68 | =item * round-trip integrity |
66 | =item * round-trip integrity |
69 | |
67 | |
70 | When you serialise a perl data structure using only data types supported |
68 | When you serialise a perl data structure using only data types supported |
71 | by JSON, the deserialised data structure is identical on the Perl level. |
69 | by JSON and Perl, the deserialised data structure is identical on the Perl |
72 | (e.g. the string "2.0" doesn't suddenly become "2" just because it looks |
70 | level. (e.g. the string "2.0" doesn't suddenly become "2" just because |
73 | like a number). There minor I<are> exceptions to this, read the MAPPING |
71 | it looks like a number). There I<are> minor exceptions to this, read the |
74 | section below to learn about those. |
72 | MAPPING section below to learn about those. |
75 | |
73 | |
76 | =item * strict checking of JSON correctness |
74 | =item * strict checking of JSON correctness |
77 | |
75 | |
78 | There is no guessing, no generating of illegal JSON texts by default, |
76 | There is no guessing, no generating of illegal JSON texts by default, |
79 | and only JSON is accepted as input by default (the latter is a security |
77 | and only JSON is accepted as input by default (the latter is a security |
… | |
… | |
85 | this module usually compares favourably in terms of speed, too. |
83 | this module usually compares favourably in terms of speed, too. |
86 | |
84 | |
87 | =item * simple to use |
85 | =item * simple to use |
88 | |
86 | |
89 | This module has both a simple functional interface as well as an object |
87 | This module has both a simple functional interface as well as an object |
90 | oriented interface interface. |
88 | oriented interface. |
91 | |
89 | |
92 | =item * reasonably versatile output formats |
90 | =item * reasonably versatile output formats |
93 | |
91 | |
94 | You can choose between the most compact guaranteed-single-line format |
92 | You can choose between the most compact guaranteed-single-line format |
95 | possible (nice for simple line-based protocols), a pure-ASCII format |
93 | possible (nice for simple line-based protocols), a pure-ASCII format |
… | |
… | |
101 | |
99 | |
102 | =cut |
100 | =cut |
103 | |
101 | |
104 | package JSON::XS; |
102 | package JSON::XS; |
105 | |
103 | |
106 | use strict; |
104 | use common::sense; |
107 | |
105 | |
108 | our $VERSION = '2.222'; |
106 | our $VERSION = 2.34; |
109 | our @ISA = qw(Exporter); |
107 | our @ISA = qw(Exporter); |
110 | |
108 | |
111 | our @EXPORT = qw(encode_json decode_json to_json from_json); |
109 | our @EXPORT = qw(encode_json decode_json); |
112 | |
|
|
113 | sub to_json($) { |
|
|
114 | require Carp; |
|
|
115 | Carp::croak ("JSON::XS::to_json has been renamed to encode_json, either downgrade to pre-2.0 versions of JSON::XS or rename the call"); |
|
|
116 | } |
|
|
117 | |
|
|
118 | sub from_json($) { |
|
|
119 | require Carp; |
|
|
120 | Carp::croak ("JSON::XS::from_json has been renamed to decode_json, either downgrade to pre-2.0 versions of JSON::XS or rename the call"); |
|
|
121 | } |
|
|
122 | |
110 | |
123 | use Exporter; |
111 | use Exporter; |
124 | use XSLoader; |
112 | use XSLoader; |
125 | |
113 | |
126 | =head1 FUNCTIONAL INTERFACE |
114 | =head1 FUNCTIONAL INTERFACE |
… | |
… | |
434 | If C<$enable> is true (or missing), then the C<encode> method will output JSON objects |
422 | If C<$enable> is true (or missing), then the C<encode> method will output JSON objects |
435 | by sorting their keys. This is adding a comparatively high overhead. |
423 | by sorting their keys. This is adding a comparatively high overhead. |
436 | |
424 | |
437 | If C<$enable> is false, then the C<encode> method will output key-value |
425 | If C<$enable> is false, then the C<encode> method will output key-value |
438 | pairs in the order Perl stores them (which will likely change between runs |
426 | pairs in the order Perl stores them (which will likely change between runs |
439 | of the same script). |
427 | of the same script, and can change even within the same run from 5.18 |
|
|
428 | onwards). |
440 | |
429 | |
441 | This option is useful if you want the same data structure to be encoded as |
430 | This option is useful if you want the same data structure to be encoded as |
442 | the same JSON text (given the same overall settings). If it is disabled, |
431 | the same JSON text (given the same overall settings). If it is disabled, |
443 | the same hash might be encoded differently even if contains the same data, |
432 | the same hash might be encoded differently even if contains the same data, |
444 | as key-value pairs have no inherent ordering in Perl. |
433 | as key-value pairs have no inherent ordering in Perl. |
445 | |
434 | |
446 | This setting has no effect when decoding JSON texts. |
435 | This setting has no effect when decoding JSON texts. |
|
|
436 | |
|
|
437 | This setting has currently no effect on tied hashes. |
447 | |
438 | |
448 | =item $json = $json->allow_nonref ([$enable]) |
439 | =item $json = $json->allow_nonref ([$enable]) |
449 | |
440 | |
450 | =item $enabled = $json->get_allow_nonref |
441 | =item $enabled = $json->get_allow_nonref |
451 | |
442 | |
… | |
… | |
666 | |
657 | |
667 | See SECURITY CONSIDERATIONS, below, for more info on why this is useful. |
658 | See SECURITY CONSIDERATIONS, below, for more info on why this is useful. |
668 | |
659 | |
669 | =item $json_text = $json->encode ($perl_scalar) |
660 | =item $json_text = $json->encode ($perl_scalar) |
670 | |
661 | |
671 | Converts the given Perl data structure (a simple scalar or a reference |
662 | Converts the given Perl value or data structure to its JSON |
672 | to a hash or array) to its JSON representation. Simple scalars will be |
663 | representation. Croaks on error. |
673 | converted into JSON string or number sequences, while references to arrays |
|
|
674 | become JSON arrays and references to hashes become JSON objects. Undefined |
|
|
675 | Perl values (e.g. C<undef>) become JSON C<null> values. Neither C<true> |
|
|
676 | nor C<false> values will be generated. |
|
|
677 | |
664 | |
678 | =item $perl_scalar = $json->decode ($json_text) |
665 | =item $perl_scalar = $json->decode ($json_text) |
679 | |
666 | |
680 | The opposite of C<encode>: expects a JSON text and tries to parse it, |
667 | The opposite of C<encode>: expects a JSON text and tries to parse it, |
681 | returning the resulting simple scalar or reference. Croaks on error. |
668 | returning the resulting simple scalar or reference. Croaks on error. |
682 | |
|
|
683 | JSON numbers and strings become simple Perl scalars. JSON arrays become |
|
|
684 | Perl arrayrefs and JSON objects become Perl hashrefs. C<true> becomes |
|
|
685 | C<1>, C<false> becomes C<0> and C<null> becomes C<undef>. |
|
|
686 | |
669 | |
687 | =item ($perl_scalar, $characters) = $json->decode_prefix ($json_text) |
670 | =item ($perl_scalar, $characters) = $json->decode_prefix ($json_text) |
688 | |
671 | |
689 | This works like the C<decode> method, but instead of raising an exception |
672 | This works like the C<decode> method, but instead of raising an exception |
690 | when there is trailing garbage after the first JSON object, it will |
673 | when there is trailing garbage after the first JSON object, it will |
… | |
… | |
713 | calls). |
696 | calls). |
714 | |
697 | |
715 | JSON::XS will only attempt to parse the JSON text once it is sure it |
698 | JSON::XS will only attempt to parse the JSON text once it is sure it |
716 | has enough text to get a decisive result, using a very simple but |
699 | has enough text to get a decisive result, using a very simple but |
717 | truly incremental parser. This means that it sometimes won't stop as |
700 | truly incremental parser. This means that it sometimes won't stop as |
718 | early as the full parser, for example, it doesn't detect parenthese |
701 | early as the full parser, for example, it doesn't detect mismatched |
719 | mismatches. The only thing it guarantees is that it starts decoding as |
702 | parentheses. The only thing it guarantees is that it starts decoding as |
720 | soon as a syntactically valid JSON text has been seen. This means you need |
703 | soon as a syntactically valid JSON text has been seen. This means you need |
721 | to set resource limits (e.g. C<max_size>) to ensure the parser will stop |
704 | to set resource limits (e.g. C<max_size>) to ensure the parser will stop |
722 | parsing in the presence if syntax errors. |
705 | parsing in the presence if syntax errors. |
723 | |
706 | |
724 | The following methods implement this incremental parser. |
707 | The following methods implement this incremental parser. |
… | |
… | |
740 | |
723 | |
741 | If the method is called in scalar context, then it will try to extract |
724 | If the method is called in scalar context, then it will try to extract |
742 | exactly I<one> JSON object. If that is successful, it will return this |
725 | exactly I<one> JSON object. If that is successful, it will return this |
743 | object, otherwise it will return C<undef>. If there is a parse error, |
726 | object, otherwise it will return C<undef>. If there is a parse error, |
744 | this method will croak just as C<decode> would do (one can then use |
727 | this method will croak just as C<decode> would do (one can then use |
745 | C<incr_skip> to skip the errornous part). This is the most common way of |
728 | C<incr_skip> to skip the erroneous part). This is the most common way of |
746 | using the method. |
729 | using the method. |
747 | |
730 | |
748 | And finally, in list context, it will try to extract as many objects |
731 | And finally, in list context, it will try to extract as many objects |
749 | from the stream as it can find and return them, or the empty list |
732 | from the stream as it can find and return them, or the empty list |
750 | otherwise. For this to work, there must be no separators between the JSON |
733 | otherwise. For this to work, there must be no separators between the JSON |
751 | objects or arrays, instead they must be concatenated back-to-back. If |
734 | objects or arrays, instead they must be concatenated back-to-back. If |
752 | an error occurs, an exception will be raised as in the scalar context |
735 | an error occurs, an exception will be raised as in the scalar context |
753 | case. Note that in this case, any previously-parsed JSON texts will be |
736 | case. Note that in this case, any previously-parsed JSON texts will be |
754 | lost. |
737 | lost. |
755 | |
738 | |
|
|
739 | Example: Parse some JSON arrays/objects in a given string and return |
|
|
740 | them. |
|
|
741 | |
|
|
742 | my @objs = JSON::XS->new->incr_parse ("[5][7][1,2]"); |
|
|
743 | |
756 | =item $lvalue_string = $json->incr_text |
744 | =item $lvalue_string = $json->incr_text |
757 | |
745 | |
758 | This method returns the currently stored JSON fragment as an lvalue, that |
746 | This method returns the currently stored JSON fragment as an lvalue, that |
759 | is, you can manipulate it. This I<only> works when a preceding call to |
747 | is, you can manipulate it. This I<only> works when a preceding call to |
760 | C<incr_parse> in I<scalar context> successfully returned an object. Under |
748 | C<incr_parse> in I<scalar context> successfully returned an object. Under |
… | |
… | |
767 | JSON object or b) parsing multiple JSON objects separated by non-JSON text |
755 | JSON object or b) parsing multiple JSON objects separated by non-JSON text |
768 | (such as commas). |
756 | (such as commas). |
769 | |
757 | |
770 | =item $json->incr_skip |
758 | =item $json->incr_skip |
771 | |
759 | |
772 | This will reset the state of the incremental parser and will remove the |
760 | This will reset the state of the incremental parser and will remove |
773 | parsed text from the input buffer. This is useful after C<incr_parse> |
761 | the parsed text from the input buffer so far. This is useful after |
774 | died, in which case the input buffer and incremental parser state is left |
762 | C<incr_parse> died, in which case the input buffer and incremental parser |
775 | unchanged, to skip the text parsed so far and to reset the parse state. |
763 | state is left unchanged, to skip the text parsed so far and to reset the |
|
|
764 | parse state. |
|
|
765 | |
|
|
766 | The difference to C<incr_reset> is that only text until the parse error |
|
|
767 | occurred is removed. |
776 | |
768 | |
777 | =item $json->incr_reset |
769 | =item $json->incr_reset |
778 | |
770 | |
779 | This completely resets the incremental parser, that is, after this call, |
771 | This completely resets the incremental parser, that is, after this call, |
780 | it will be as if the parser had never parsed anything. |
772 | it will be as if the parser had never parsed anything. |
781 | |
773 | |
782 | This is useful if you want ot repeatedly parse JSON objects and want to |
774 | This is useful if you want to repeatedly parse JSON objects and want to |
783 | ignore any trailing data, which means you have to reset the parser after |
775 | ignore any trailing data, which means you have to reset the parser after |
784 | each successful decode. |
776 | each successful decode. |
785 | |
777 | |
786 | =back |
778 | =back |
787 | |
779 | |
… | |
… | |
978 | If the number consists of digits only, JSON::XS will try to represent |
970 | If the number consists of digits only, JSON::XS will try to represent |
979 | it as an integer value. If that fails, it will try to represent it as |
971 | it as an integer value. If that fails, it will try to represent it as |
980 | a numeric (floating point) value if that is possible without loss of |
972 | a numeric (floating point) value if that is possible without loss of |
981 | precision. Otherwise it will preserve the number as a string value (in |
973 | precision. Otherwise it will preserve the number as a string value (in |
982 | which case you lose roundtripping ability, as the JSON number will be |
974 | which case you lose roundtripping ability, as the JSON number will be |
983 | re-encoded toa JSON string). |
975 | re-encoded to a JSON string). |
984 | |
976 | |
985 | Numbers containing a fractional or exponential part will always be |
977 | Numbers containing a fractional or exponential part will always be |
986 | represented as numeric (floating point) values, possibly at a loss of |
978 | represented as numeric (floating point) values, possibly at a loss of |
987 | precision (in which case you might lose perfect roundtripping ability, but |
979 | precision (in which case you might lose perfect roundtripping ability, but |
988 | the JSON number will still be re-encoded as a JSON number). |
980 | the JSON number will still be re-encoded as a JSON number). |
989 | |
981 | |
|
|
982 | Note that precision is not accuracy - binary floating point values cannot |
|
|
983 | represent most decimal fractions exactly, and when converting from and to |
|
|
984 | floating point, JSON::XS only guarantees precision up to but not including |
|
|
985 | the least significant bit. |
|
|
986 | |
990 | =item true, false |
987 | =item true, false |
991 | |
988 | |
992 | These JSON atoms become C<JSON::XS::true> and C<JSON::XS::false>, |
989 | These JSON atoms become C<JSON::XS::true> and C<JSON::XS::false>, |
993 | respectively. They are overloaded to act almost exactly like the numbers |
990 | respectively. They are overloaded to act almost exactly like the numbers |
994 | C<1> and C<0>. You can check whether a scalar is a JSON boolean by using |
991 | C<1> and C<0>. You can check whether a scalar is a JSON boolean by using |
… | |
… | |
1009 | |
1006 | |
1010 | =over 4 |
1007 | =over 4 |
1011 | |
1008 | |
1012 | =item hash references |
1009 | =item hash references |
1013 | |
1010 | |
1014 | Perl hash references become JSON objects. As there is no inherent ordering |
1011 | Perl hash references become JSON objects. As there is no inherent |
1015 | in hash keys (or JSON objects), they will usually be encoded in a |
1012 | ordering in hash keys (or JSON objects), they will usually be encoded |
1016 | pseudo-random order that can change between runs of the same program but |
1013 | in a pseudo-random order. JSON::XS can optionally sort the hash keys |
1017 | stays generally the same within a single run of a program. JSON::XS can |
1014 | (determined by the I<canonical> flag), so the same datastructure will |
1018 | optionally sort the hash keys (determined by the I<canonical> flag), so |
1015 | serialise to the same JSON text (given same settings and version of |
1019 | the same datastructure will serialise to the same JSON text (given same |
1016 | JSON::XS), but this incurs a runtime overhead and is only rarely useful, |
1020 | settings and version of JSON::XS), but this incurs a runtime overhead |
1017 | e.g. when you want to compare some JSON text against another for equality. |
1021 | and is only rarely useful, e.g. when you want to compare some JSON text |
|
|
1022 | against another for equality. |
|
|
1023 | |
1018 | |
1024 | =item array references |
1019 | =item array references |
1025 | |
1020 | |
1026 | Perl array references become JSON arrays. |
1021 | Perl array references become JSON arrays. |
1027 | |
1022 | |
… | |
… | |
1080 | $x *= 1; # same thing, the choice is yours. |
1075 | $x *= 1; # same thing, the choice is yours. |
1081 | |
1076 | |
1082 | You can not currently force the type in other, less obscure, ways. Tell me |
1077 | You can not currently force the type in other, less obscure, ways. Tell me |
1083 | if you need this capability (but don't forget to explain why it's needed |
1078 | if you need this capability (but don't forget to explain why it's needed |
1084 | :). |
1079 | :). |
|
|
1080 | |
|
|
1081 | Note that numerical precision has the same meaning as under Perl (so |
|
|
1082 | binary to decimal conversion follows the same rules as in Perl, which |
|
|
1083 | can differ to other languages). Also, your perl interpreter might expose |
|
|
1084 | extensions to the floating point numbers of your platform, such as |
|
|
1085 | infinities or NaN's - these cannot be represented in JSON, and it is an |
|
|
1086 | error to pass those in. |
1085 | |
1087 | |
1086 | =back |
1088 | =back |
1087 | |
1089 | |
1088 | |
1090 | |
1089 | =head1 ENCODING/CODESET FLAG NOTES |
1091 | =head1 ENCODING/CODESET FLAG NOTES |
… | |
… | |
1116 | =item C<utf8> flag disabled |
1118 | =item C<utf8> flag disabled |
1117 | |
1119 | |
1118 | When C<utf8> is disabled (the default), then C<encode>/C<decode> generate |
1120 | When C<utf8> is disabled (the default), then C<encode>/C<decode> generate |
1119 | and expect Unicode strings, that is, characters with high ordinal Unicode |
1121 | and expect Unicode strings, that is, characters with high ordinal Unicode |
1120 | values (> 255) will be encoded as such characters, and likewise such |
1122 | values (> 255) will be encoded as such characters, and likewise such |
1121 | characters are decoded as-is, no canges to them will be done, except |
1123 | characters are decoded as-is, no changes to them will be done, except |
1122 | "(re-)interpreting" them as Unicode codepoints or Unicode characters, |
1124 | "(re-)interpreting" them as Unicode codepoints or Unicode characters, |
1123 | respectively (to Perl, these are the same thing in strings unless you do |
1125 | respectively (to Perl, these are the same thing in strings unless you do |
1124 | funny/weird/dumb stuff). |
1126 | funny/weird/dumb stuff). |
1125 | |
1127 | |
1126 | This is useful when you want to do the encoding yourself (e.g. when you |
1128 | This is useful when you want to do the encoding yourself (e.g. when you |
… | |
… | |
1182 | proper subset of most 8-bit and multibyte encodings in use in the world. |
1184 | proper subset of most 8-bit and multibyte encodings in use in the world. |
1183 | |
1185 | |
1184 | =back |
1186 | =back |
1185 | |
1187 | |
1186 | |
1188 | |
|
|
1189 | =head2 JSON and ECMAscript |
|
|
1190 | |
|
|
1191 | JSON syntax is based on how literals are represented in javascript (the |
|
|
1192 | not-standardised predecessor of ECMAscript) which is presumably why it is |
|
|
1193 | called "JavaScript Object Notation". |
|
|
1194 | |
|
|
1195 | However, JSON is not a subset (and also not a superset of course) of |
|
|
1196 | ECMAscript (the standard) or javascript (whatever browsers actually |
|
|
1197 | implement). |
|
|
1198 | |
|
|
1199 | If you want to use javascript's C<eval> function to "parse" JSON, you |
|
|
1200 | might run into parse errors for valid JSON texts, or the resulting data |
|
|
1201 | structure might not be queryable: |
|
|
1202 | |
|
|
1203 | One of the problems is that U+2028 and U+2029 are valid characters inside |
|
|
1204 | JSON strings, but are not allowed in ECMAscript string literals, so the |
|
|
1205 | following Perl fragment will not output something that can be guaranteed |
|
|
1206 | to be parsable by javascript's C<eval>: |
|
|
1207 | |
|
|
1208 | use JSON::XS; |
|
|
1209 | |
|
|
1210 | print encode_json [chr 0x2028]; |
|
|
1211 | |
|
|
1212 | The right fix for this is to use a proper JSON parser in your javascript |
|
|
1213 | programs, and not rely on C<eval> (see for example Douglas Crockford's |
|
|
1214 | F<json2.js> parser). |
|
|
1215 | |
|
|
1216 | If this is not an option, you can, as a stop-gap measure, simply encode to |
|
|
1217 | ASCII-only JSON: |
|
|
1218 | |
|
|
1219 | use JSON::XS; |
|
|
1220 | |
|
|
1221 | print JSON::XS->new->ascii->encode ([chr 0x2028]); |
|
|
1222 | |
|
|
1223 | Note that this will enlarge the resulting JSON text quite a bit if you |
|
|
1224 | have many non-ASCII characters. You might be tempted to run some regexes |
|
|
1225 | to only escape U+2028 and U+2029, e.g.: |
|
|
1226 | |
|
|
1227 | # DO NOT USE THIS! |
|
|
1228 | my $json = JSON::XS->new->utf8->encode ([chr 0x2028]); |
|
|
1229 | $json =~ s/\xe2\x80\xa8/\\u2028/g; # escape U+2028 |
|
|
1230 | $json =~ s/\xe2\x80\xa9/\\u2029/g; # escape U+2029 |
|
|
1231 | print $json; |
|
|
1232 | |
|
|
1233 | Note that I<this is a bad idea>: the above only works for U+2028 and |
|
|
1234 | U+2029 and thus only for fully ECMAscript-compliant parsers. Many existing |
|
|
1235 | javascript implementations, however, have issues with other characters as |
|
|
1236 | well - using C<eval> naively simply I<will> cause problems. |
|
|
1237 | |
|
|
1238 | Another problem is that some javascript implementations reserve |
|
|
1239 | some property names for their own purposes (which probably makes |
|
|
1240 | them non-ECMAscript-compliant). For example, Iceweasel reserves the |
|
|
1241 | C<__proto__> property name for its own purposes. |
|
|
1242 | |
|
|
1243 | If that is a problem, you could parse try to filter the resulting JSON |
|
|
1244 | output for these property strings, e.g.: |
|
|
1245 | |
|
|
1246 | $json =~ s/"__proto__"\s*:/"__proto__renamed":/g; |
|
|
1247 | |
|
|
1248 | This works because C<__proto__> is not valid outside of strings, so every |
|
|
1249 | occurrence of C<"__proto__"\s*:> must be a string used as property name. |
|
|
1250 | |
|
|
1251 | If you know of other incompatibilities, please let me know. |
|
|
1252 | |
|
|
1253 | |
1187 | =head2 JSON and YAML |
1254 | =head2 JSON and YAML |
1188 | |
1255 | |
1189 | You often hear that JSON is a subset of YAML. This is, however, a mass |
1256 | You often hear that JSON is a subset of YAML. This is, however, a mass |
1190 | hysteria(*) and very far from the truth (as of the time of this writing), |
1257 | hysteria(*) and very far from the truth (as of the time of this writing), |
1191 | so let me state it clearly: I<in general, there is no way to configure |
1258 | so let me state it clearly: I<in general, there is no way to configure |
… | |
… | |
1199 | my $yaml = $to_yaml->encode ($ref) . "\n"; |
1266 | my $yaml = $to_yaml->encode ($ref) . "\n"; |
1200 | |
1267 | |
1201 | This will I<usually> generate JSON texts that also parse as valid |
1268 | This will I<usually> generate JSON texts that also parse as valid |
1202 | YAML. Please note that YAML has hardcoded limits on (simple) object key |
1269 | YAML. Please note that YAML has hardcoded limits on (simple) object key |
1203 | lengths that JSON doesn't have and also has different and incompatible |
1270 | lengths that JSON doesn't have and also has different and incompatible |
1204 | unicode handling, so you should make sure that your hash keys are |
1271 | unicode character escape syntax, so you should make sure that your hash |
1205 | noticeably shorter than the 1024 "stream characters" YAML allows and that |
1272 | keys are noticeably shorter than the 1024 "stream characters" YAML allows |
1206 | you do not have characters with codepoint values outside the Unicode BMP |
1273 | and that you do not have characters with codepoint values outside the |
1207 | (basic multilingual page). YAML also does not allow C<\/> sequences in |
1274 | Unicode BMP (basic multilingual page). YAML also does not allow C<\/> |
1208 | strings (which JSON::XS does not I<currently> generate, but other JSON |
1275 | sequences in strings (which JSON::XS does not I<currently> generate, but |
1209 | generators might). |
1276 | other JSON generators might). |
1210 | |
1277 | |
1211 | There might be other incompatibilities that I am not aware of (or the YAML |
1278 | There might be other incompatibilities that I am not aware of (or the YAML |
1212 | specification has been changed yet again - it does so quite often). In |
1279 | specification has been changed yet again - it does so quite often). In |
1213 | general you should not try to generate YAML with a JSON generator or vice |
1280 | general you should not try to generate YAML with a JSON generator or vice |
1214 | versa, or try to parse JSON with a YAML parser or vice versa: chances are |
1281 | versa, or try to parse JSON with a YAML parser or vice versa: chances are |
… | |
… | |
1233 | that difficult or long) and finally make YAML compatible to it, and |
1300 | that difficult or long) and finally make YAML compatible to it, and |
1234 | educating users about the changes, instead of spreading lies about the |
1301 | educating users about the changes, instead of spreading lies about the |
1235 | real compatibility for many I<years> and trying to silence people who |
1302 | real compatibility for many I<years> and trying to silence people who |
1236 | point out that it isn't true. |
1303 | point out that it isn't true. |
1237 | |
1304 | |
|
|
1305 | Addendum/2009: the YAML 1.2 spec is still incompatible with JSON, even |
|
|
1306 | though the incompatibilities have been documented (and are known to Brian) |
|
|
1307 | for many years and the spec makes explicit claims that YAML is a superset |
|
|
1308 | of JSON. It would be so easy to fix, but apparently, bullying people and |
|
|
1309 | corrupting userdata is so much easier. |
|
|
1310 | |
1238 | =back |
1311 | =back |
1239 | |
1312 | |
1240 | |
1313 | |
1241 | =head2 SPEED |
1314 | =head2 SPEED |
1242 | |
1315 | |
… | |
… | |
1249 | a very short single-line JSON string (also available at |
1322 | a very short single-line JSON string (also available at |
1250 | L<http://dist.schmorp.de/misc/json/short.json>). |
1323 | L<http://dist.schmorp.de/misc/json/short.json>). |
1251 | |
1324 | |
1252 | {"method": "handleMessage", "params": ["user1", |
1325 | {"method": "handleMessage", "params": ["user1", |
1253 | "we were just talking"], "id": null, "array":[1,11,234,-5,1e5,1e7, |
1326 | "we were just talking"], "id": null, "array":[1,11,234,-5,1e5,1e7, |
1254 | true, false]} |
1327 | 1, 0]} |
1255 | |
1328 | |
1256 | It shows the number of encodes/decodes per second (JSON::XS uses |
1329 | It shows the number of encodes/decodes per second (JSON::XS uses |
1257 | the functional interface, while JSON::XS/2 uses the OO interface |
1330 | the functional interface, while JSON::XS/2 uses the OO interface |
1258 | with pretty-printing and hashkey sorting enabled, JSON::XS/3 enables |
1331 | with pretty-printing and hashkey sorting enabled, JSON::XS/3 enables |
1259 | shrink). Higher is better: |
1332 | shrink. JSON::DWIW/DS uses the deserialise function, while JSON::DWIW::FJ |
|
|
1333 | uses the from_json method). Higher is better: |
1260 | |
1334 | |
1261 | module | encode | decode | |
1335 | module | encode | decode | |
1262 | -----------|------------|------------| |
1336 | --------------|------------|------------| |
1263 | JSON 1.x | 4990.842 | 4088.813 | |
1337 | JSON::DWIW/DS | 86302.551 | 102300.098 | |
1264 | JSON::DWIW | 51653.990 | 71575.154 | |
1338 | JSON::DWIW/FJ | 86302.551 | 75983.768 | |
1265 | JSON::PC | 65948.176 | 74631.744 | |
1339 | JSON::PP | 15827.562 | 6638.658 | |
1266 | JSON::PP | 8931.652 | 3817.168 | |
1340 | JSON::Syck | 63358.066 | 47662.545 | |
1267 | JSON::Syck | 24877.248 | 27776.848 | |
1341 | JSON::XS | 511500.488 | 511500.488 | |
1268 | JSON::XS | 388361.481 | 227951.304 | |
1342 | JSON::XS/2 | 291271.111 | 388361.481 | |
1269 | JSON::XS/2 | 227951.304 | 218453.333 | |
1343 | JSON::XS/3 | 361577.931 | 361577.931 | |
1270 | JSON::XS/3 | 338250.323 | 218453.333 | |
1344 | Storable | 66788.280 | 265462.278 | |
1271 | Storable | 16500.016 | 135300.129 | |
|
|
1272 | -----------+------------+------------+ |
1345 | --------------+------------+------------+ |
1273 | |
1346 | |
1274 | That is, JSON::XS is about five times faster than JSON::DWIW on encoding, |
1347 | That is, JSON::XS is almost six times faster than JSON::DWIW on encoding, |
1275 | about three times faster on decoding, and over forty times faster |
1348 | about five times faster on decoding, and over thirty to seventy times |
1276 | than JSON, even with pretty-printing and key sorting. It also compares |
1349 | faster than JSON's pure perl implementation. It also compares favourably |
1277 | favourably to Storable for small amounts of data. |
1350 | to Storable for small amounts of data. |
1278 | |
1351 | |
1279 | Using a longer test string (roughly 18KB, generated from Yahoo! Locals |
1352 | Using a longer test string (roughly 18KB, generated from Yahoo! Locals |
1280 | search API (L<http://dist.schmorp.de/misc/json/long.json>). |
1353 | search API (L<http://dist.schmorp.de/misc/json/long.json>). |
1281 | |
1354 | |
1282 | module | encode | decode | |
1355 | module | encode | decode | |
1283 | -----------|------------|------------| |
1356 | --------------|------------|------------| |
1284 | JSON 1.x | 55.260 | 34.971 | |
1357 | JSON::DWIW/DS | 1647.927 | 2673.916 | |
1285 | JSON::DWIW | 825.228 | 1082.513 | |
1358 | JSON::DWIW/FJ | 1630.249 | 2596.128 | |
1286 | JSON::PC | 3571.444 | 2394.829 | |
|
|
1287 | JSON::PP | 210.987 | 32.574 | |
1359 | JSON::PP | 400.640 | 62.311 | |
1288 | JSON::Syck | 552.551 | 787.544 | |
1360 | JSON::Syck | 1481.040 | 1524.869 | |
1289 | JSON::XS | 5780.463 | 4854.519 | |
1361 | JSON::XS | 20661.596 | 9541.183 | |
1290 | JSON::XS/2 | 3869.998 | 4798.975 | |
1362 | JSON::XS/2 | 10683.403 | 9416.938 | |
1291 | JSON::XS/3 | 5862.880 | 4798.975 | |
1363 | JSON::XS/3 | 20661.596 | 9400.054 | |
1292 | Storable | 4445.002 | 5235.027 | |
1364 | Storable | 19765.806 | 10000.725 | |
1293 | -----------+------------+------------+ |
1365 | --------------+------------+------------+ |
1294 | |
1366 | |
1295 | Again, JSON::XS leads by far (except for Storable which non-surprisingly |
1367 | Again, JSON::XS leads by far (except for Storable which non-surprisingly |
1296 | decodes faster). |
1368 | decodes a bit faster). |
1297 | |
1369 | |
1298 | On large strings containing lots of high Unicode characters, some modules |
1370 | On large strings containing lots of high Unicode characters, some modules |
1299 | (such as JSON::PC) seem to decode faster than JSON::XS, but the result |
1371 | (such as JSON::PC) seem to decode faster than JSON::XS, but the result |
1300 | will be broken due to missing (or wrong) Unicode handling. Others refuse |
1372 | will be broken due to missing (or wrong) Unicode handling. Others refuse |
1301 | to decode or encode properly, so it was impossible to prepare a fair |
1373 | to decode or encode properly, so it was impossible to prepare a fair |
… | |
… | |
1337 | information you might want to make sure that exceptions thrown by JSON::XS |
1409 | information you might want to make sure that exceptions thrown by JSON::XS |
1338 | will not end up in front of untrusted eyes. |
1410 | will not end up in front of untrusted eyes. |
1339 | |
1411 | |
1340 | If you are using JSON::XS to return packets to consumption |
1412 | If you are using JSON::XS to return packets to consumption |
1341 | by JavaScript scripts in a browser you should have a look at |
1413 | by JavaScript scripts in a browser you should have a look at |
1342 | L<http://jpsykes.com/47/practical-csrf-and-json-security> to see whether |
1414 | L<http://blog.archive.jpsykes.com/47/practical-csrf-and-json-security/> to |
1343 | you are vulnerable to some common attack vectors (which really are browser |
1415 | see whether you are vulnerable to some common attack vectors (which really |
1344 | design bugs, but it is still you who will have to deal with it, as major |
1416 | are browser design bugs, but it is still you who will have to deal with |
1345 | browser developers care only for features, not about getting security |
1417 | it, as major browser developers care only for features, not about getting |
1346 | right). |
1418 | security right). |
1347 | |
1419 | |
1348 | |
1420 | |
1349 | =head1 THREADS |
1421 | =head1 THREADS |
1350 | |
1422 | |
1351 | This module is I<not> guaranteed to be thread safe and there are no |
1423 | This module is I<not> guaranteed to be thread safe and there are no |
1352 | plans to change this until Perl gets thread support (as opposed to the |
1424 | plans to change this until Perl gets thread support (as opposed to the |
1353 | horribly slow so-called "threads" which are simply slow and bloated |
1425 | horribly slow so-called "threads" which are simply slow and bloated |
1354 | process simulations - use fork, it's I<much> faster, cheaper, better). |
1426 | process simulations - use fork, it's I<much> faster, cheaper, better). |
1355 | |
1427 | |
1356 | (It might actually work, but you have been warned). |
1428 | (It might actually work, but you have been warned). |
|
|
1429 | |
|
|
1430 | |
|
|
1431 | =head1 THE PERILS OF SETLOCALE |
|
|
1432 | |
|
|
1433 | Sometimes people avoid the Perl locale support and directly call the |
|
|
1434 | system's setlocale function with C<LC_ALL>. |
|
|
1435 | |
|
|
1436 | This breaks both perl and modules such as JSON::XS, as stringification of |
|
|
1437 | numbers no longer works correctly (e.g. C<$x = 0.1; print "$x"+1> might |
|
|
1438 | print C<1>, and JSON::XS might output illegal JSON as JSON::XS relies on |
|
|
1439 | perl to stringify numbers). |
|
|
1440 | |
|
|
1441 | The solution is simple: don't call C<setlocale>, or use it for only those |
|
|
1442 | categories you need, such as C<LC_MESSAGES> or C<LC_CTYPE>. |
|
|
1443 | |
|
|
1444 | If you need C<LC_NUMERIC>, you should enable it only around the code that |
|
|
1445 | actually needs it (avoiding stringification of numbers), and restore it |
|
|
1446 | afterwards. |
1357 | |
1447 | |
1358 | |
1448 | |
1359 | =head1 BUGS |
1449 | =head1 BUGS |
1360 | |
1450 | |
1361 | While the goal of this module is to be correct, that unfortunately does |
1451 | While the goal of this module is to be correct, that unfortunately does |