--- JSON-XS/XS.pm 2007/12/19 11:42:52 1.79 +++ JSON-XS/XS.pm 2014/03/02 22:09:38 1.154 @@ -2,6 +2,8 @@ JSON::XS - JSON serialising/deserialising, done correctly and fast +=encoding utf-8 + JSON::XS - 正しくて高速な JSON シリアライザ/デシリアライザ (http://fleur.hio.jp/perldoc/mix/lib/JSON/XS.html) @@ -37,7 +39,7 @@ Beginning with version 2.0 of the JSON module, when both JSON and JSON::XS are installed, then JSON will fall back on JSON::XS (this can be -overriden) with no overhead due to emulation (by inheritign constructor +overridden) with no overhead due to emulation (by inheriting constructor and methods). If JSON::XS is not available, it will fall back to the compatible JSON::PP module as backend, so using JSON instead of JSON::XS gives you a portable JSON API that can be fast when you need and doesn't @@ -49,8 +51,6 @@ their maintainers are unresponsive, gone missing, or not listening to bug reports for other reasons. -See COMPARISON, below, for a comparison to some other JSON modules. - See MAPPING, below, on how JSON::XS maps perl values to JSON values and vice versa. @@ -60,15 +60,16 @@ =item * correct Unicode handling -This module knows how to handle Unicode, and even documents how and when -it does so. +This module knows how to handle Unicode, documents how and when it does +so, and even documents what "correct" means. =item * round-trip integrity -When you serialise a perl data structure using only datatypes supported -by JSON, the deserialised data structure is identical on the Perl level. -(e.g. the string "2.0" doesn't suddenly become "2" just because it looks -like a number). +When you serialise a perl data structure using only data types supported +by JSON and Perl, the deserialised data structure is identical on the Perl +level. (e.g. the string "2.0" doesn't suddenly become "2" just because +it looks like a number). There I minor exceptions to this, read the +MAPPING section below to learn about those. =item * strict checking of JSON correctness @@ -78,18 +79,18 @@ =item * fast -Compared to other JSON modules, this module compares favourably in terms -of speed, too. +Compared to other JSON modules and other serialisers such as Storable, +this module usually compares favourably in terms of speed, too. =item * simple to use -This module has both a simple functional interface as well as an OO -interface. +This module has both a simple functional interface as well as an object +oriented interface. =item * reasonably versatile output formats -You can choose between the most compact guaranteed single-line format -possible (nice for simple line-based protocols), a pure-ascii format +You can choose between the most compact guaranteed-single-line format +possible (nice for simple line-based protocols), a pure-ASCII format (for when your transport is not 8-bit clean, still supports the whole Unicode range), or a pretty-printed format (for when you want to read that stuff). Or you can combine those features in whatever way you like. @@ -100,26 +101,18 @@ package JSON::XS; -use strict; +use common::sense; -our $VERSION = '2.01'; +our $VERSION = 3.01; our @ISA = qw(Exporter); -our @EXPORT = qw(encode_json decode_json to_json from_json); - -sub to_json($) { - require Carp; - 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"); -} - -sub from_json($) { - require Carp; - 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"); -} +our @EXPORT = qw(encode_json decode_json); use Exporter; use XSLoader; +use Types::Serialiser (); + =head1 FUNCTIONAL INTERFACE The following convenience methods are provided by this module. They are @@ -136,7 +129,7 @@ $json_text = JSON::XS->new->utf8->encode ($perl_scalar) -except being faster. +Except being faster. =item $perl_scalar = decode_json $json_text @@ -148,16 +141,7 @@ $perl_scalar = JSON::XS->new->utf8->decode ($json_text) -except being faster. - -=item $is_boolean = JSON::XS::is_bool $scalar - -Returns true if the passed scalar represents either JSON::XS::true or -JSON::XS::false, two constants that act like C<1> and C<0>, respectively -and are used to represent JSON C and C values in Perl. - -See MAPPING, below, for more information on how JSON values are mapped to -Perl. +Except being faster. =back @@ -176,11 +160,11 @@ =item 2. Perl does I associate an encoding with your strings. -Unless you force it to, e.g. when matching it against a regex, or printing -the scalar to a file, in which case Perl either interprets your string as -locale-encoded text, octets/binary, or as Unicode, depending on various -settings. In no case is an encoding stored together with your data, it is -I that decides encoding, not any magical metadata. +... until you force it to, e.g. when matching it against a regex, or +printing the scalar to a file, in which case Perl either interprets your +string as locale-encoded text, octets/binary, or as Unicode, depending +on various settings. In no case is an encoding stored together with your +data, it is I that decides encoding, not any magical meta data. =item 3. The internal utf-8 flag has no meaning with regards to the encoding of your string. @@ -196,7 +180,7 @@ exist. =item 4. A "Unicode String" is simply a string where each character can be -validly interpreted as a Unicode codepoint. +validly interpreted as a Unicode code point. If you have UTF-8 encoded data, it is no longer a Unicode string, but a Unicode string encoded in UTF-8, giving you a binary string. @@ -244,6 +228,9 @@ characters unless required by the JSON syntax or other flags. This results in a faster and more compact format. +See also the section I later in this +document. + The main use for this flag is to produce JSON texts that can be transmitted over a 7-bit channel, as the encoded JSON texts will not contain any 8 bit characters. @@ -265,6 +252,9 @@ If C<$enable> is false, then the C method will not escape Unicode characters unless required by the JSON syntax or other flags. +See also the section I later in this +document. + The main use for this flag is efficiently encoding binary data as JSON text, as most octets will not be escaped, resulting in a smaller encoded size. The disadvantage is that the resulting JSON text is encoded @@ -293,6 +283,9 @@ Unicode string. Any decoding or encoding (e.g. to UTF-8 or UTF-16) needs to be done yourself, e.g. using the Encode module. +See also the section I later in this +document. + Example, output UTF-16BE-encoded JSON: use Encode; @@ -424,7 +417,8 @@ If C<$enable> is false, then the C method will output key-value pairs in the order Perl stores them (which will likely change between runs -of the same script). +of the same script, and can change even within the same run from 5.18 +onwards). This option is useful if you want the same data structure to be encoded as the same JSON text (given the same overall settings). If it is disabled, @@ -433,6 +427,8 @@ This setting has no effect when decoding JSON texts. +This setting has currently no effect on tied hashes. + =item $json = $json->allow_nonref ([$enable]) =item $enabled = $json->get_allow_nonref @@ -453,30 +449,48 @@ JSON::XS->new->allow_nonref->encode ("Hello, World!") => "Hello, World!" +=item $json = $json->allow_unknown ([$enable]) + +=item $enabled = $json->get_allow_unknown + +If C<$enable> is true (or missing), then C will I throw an +exception when it encounters values it cannot represent in JSON (for +example, filehandles) but instead will encode a JSON C value. Note +that blessed objects are not included here and are handled separately by +c. + +If C<$enable> is false (the default), then C will throw an +exception when it encounters anything it cannot encode as JSON. + +This option does not affect C in any way, and it is recommended to +leave it off unless you know your communications partner. + =item $json = $json->allow_blessed ([$enable]) =item $enabled = $json->get_allow_blessed +See L for details. + If C<$enable> is true (or missing), then the C method will not -barf when it encounters a blessed reference. Instead, the value of the -B option will decide whether C (C -disabled or no C method found) or a representation of the -object (C enabled and C method found) is being -encoded. Has no effect on C. +barf when it encounters a blessed reference that it cannot convert +otherwise. Instead, a JSON C value is encoded instead of the object. If C<$enable> is false (the default), then C will throw an -exception when it encounters a blessed object. +exception when it encounters a blessed object that it cannot convert +otherwise. + +This setting has no effect on C. =item $json = $json->convert_blessed ([$enable]) =item $enabled = $json->get_convert_blessed +See L for details. + If C<$enable> is true (or missing), then C, upon encountering a blessed object, will check for the availability of the C method -on the object's class. If found, it will be called in scalar context -and the resulting scalar will be encoded instead of the object. If no -C method is found, the value of C will decide what -to do. +on the object's class. If found, it will be called in scalar context and +the resulting scalar will be encoded instead of the object. The C method may safely call die if it wants. If C returns other blessed objects, those will be handled in the same @@ -486,12 +500,28 @@ usually in upper case letters and to avoid collisions with any C function or method. -This setting does not yet influence C in any way, but in the -future, global hooks might get installed that influence C and are -enabled by this setting. +If C<$enable> is false (the default), then C will not consider +this type of conversion. + +This setting has no effect on C. + +=item $json = $json->allow_tags ([$enable]) + +=item $enabled = $json->allow_tags + +See L for details. -If C<$enable> is false, then the C setting will decide what -to do when a blessed object is found. +If C<$enable> is true (or missing), then C, upon encountering a +blessed object, will check for the availability of the C method on +the object's class. If found, it will be used to serialise the object into +a nonstandard tagged JSON value (that JSON decoders cannot decode). + +It also causes C to parse such tagged JSON values and deserialise +them via a call to the C method. + +If C<$enable> is false (the default), then C will not consider +this type of conversion, and tagged JSON values will cause a parse error +in C, as if tags were not part of the grammar. =item $json = $json->filter_json_object ([$coderef->($hashref)]) @@ -602,9 +632,9 @@ =item $max_depth = $json->get_max_depth Sets the maximum nesting level (default C<512>) accepted while encoding -or decoding. If the JSON text or Perl data structure has an equal or -higher nesting level then this limit, then the encoder and decoder will -stop and croak at that point. +or decoding. If a higher nesting level is detected in JSON text or a Perl +data structure, then the encoder and decoder will stop and croak at that +point. Nesting level is defined by number of hash- or arrayrefs that the encoder needs to traverse to reach a given point or the number of C<{> or C<[> @@ -614,9 +644,12 @@ Setting the maximum depth to one disallows any nesting, so that ensures that the object is only a single hash/object or array. -The argument to C will be rounded up to the next highest power -of two. If no argument is given, the highest possible setting will be -used, which is rarely useful. +If no argument is given, the highest possible setting will be used, which +is rarely useful. + +Note that nesting is implemented by recursion in C. The default value has +been chosen to be as large as typical operating systems allow without +crashing. See SECURITY CONSIDERATIONS, below, for more info on why this is useful. @@ -626,34 +659,25 @@ Set the maximum length a JSON text may have (in bytes) where decoding is being attempted. The default is C<0>, meaning no limit. When C -is called on a string longer then this number of characters it will not +is called on a string that is longer then this many bytes, it will not attempt to decode the string but throw an exception. This setting has no effect on C (yet). -The argument to C will be rounded up to the next B -power of two (so may be more than requested). If no argument is given, the -limit check will be deactivated (same as when C<0> is specified). +If no argument is given, the limit check will be deactivated (same as when +C<0> is specified). See SECURITY CONSIDERATIONS, below, for more info on why this is useful. =item $json_text = $json->encode ($perl_scalar) -Converts the given Perl data structure (a simple scalar or a reference -to a hash or array) to its JSON representation. Simple scalars will be -converted into JSON string or number sequences, while references to arrays -become JSON arrays and references to hashes become JSON objects. Undefined -Perl values (e.g. C) become JSON C values. Neither C -nor C values will be generated. +Converts the given Perl value or data structure to its JSON +representation. Croaks on error. =item $perl_scalar = $json->decode ($json_text) The opposite of C: expects a JSON text and tries to parse it, returning the resulting simple scalar or reference. Croaks on error. -JSON numbers and strings become simple Perl scalars. JSON arrays become -Perl arrayrefs and JSON objects become Perl hashrefs. C becomes -C<1>, C becomes C<0> and C becomes C. - =item ($perl_scalar, $characters) = $json->decode_prefix ($json_text) This works like the C method, but instead of raising an exception @@ -662,8 +686,7 @@ so far. This is useful if your JSON texts are not delimited by an outer protocol -(which is not the brightest thing to do in the first place) and you need -to know where the JSON text ends. +and you need to know where the JSON text ends. JSON::XS->new->decode_prefix ("[1] the tail") => ([], 3) @@ -671,6 +694,250 @@ =back +=head1 INCREMENTAL PARSING + +In some cases, there is the need for incremental parsing of JSON +texts. While this module always has to keep both JSON text and resulting +Perl data structure in memory at one time, it does allow you to parse a +JSON stream incrementally. It does so by accumulating text until it has +a full JSON object, which it then can decode. This process is similar to +using C to see if a full JSON object is available, but +is much more efficient (and can be implemented with a minimum of method +calls). + +JSON::XS will only attempt to parse the JSON text once it is sure it +has enough text to get a decisive result, using a very simple but +truly incremental parser. This means that it sometimes won't stop as +early as the full parser, for example, it doesn't detect mismatched +parentheses. The only thing it guarantees is that it starts decoding as +soon as a syntactically valid JSON text has been seen. This means you need +to set resource limits (e.g. C) to ensure the parser will stop +parsing in the presence if syntax errors. + +The following methods implement this incremental parser. + +=over 4 + +=item [void, scalar or list context] = $json->incr_parse ([$string]) + +This is the central parsing function. It can both append new text and +extract objects from the stream accumulated so far (both of these +functions are optional). + +If C<$string> is given, then this string is appended to the already +existing JSON fragment stored in the C<$json> object. + +After that, if the function is called in void context, it will simply +return without doing anything further. This can be used to add more text +in as many chunks as you want. + +If the method is called in scalar context, then it will try to extract +exactly I JSON object. If that is successful, it will return this +object, otherwise it will return C. If there is a parse error, +this method will croak just as C would do (one can then use +C to skip the erroneous part). This is the most common way of +using the method. + +And finally, in list context, it will try to extract as many objects +from the stream as it can find and return them, or the empty list +otherwise. For this to work, there must be no separators between the JSON +objects or arrays, instead they must be concatenated back-to-back. If +an error occurs, an exception will be raised as in the scalar context +case. Note that in this case, any previously-parsed JSON texts will be +lost. + +Example: Parse some JSON arrays/objects in a given string and return +them. + + my @objs = JSON::XS->new->incr_parse ("[5][7][1,2]"); + +=item $lvalue_string = $json->incr_text + +This method returns the currently stored JSON fragment as an lvalue, that +is, you can manipulate it. This I works when a preceding call to +C in I successfully returned an object. Under +all other circumstances you must not call this function (I mean it. +although in simple tests it might actually work, it I fail under +real world conditions). As a special exception, you can also call this +method before having parsed anything. + +This function is useful in two cases: a) finding the trailing text after a +JSON object or b) parsing multiple JSON objects separated by non-JSON text +(such as commas). + +=item $json->incr_skip + +This will reset the state of the incremental parser and will remove +the parsed text from the input buffer so far. This is useful after +C died, in which case the input buffer and incremental parser +state is left unchanged, to skip the text parsed so far and to reset the +parse state. + +The difference to C is that only text until the parse error +occurred is removed. + +=item $json->incr_reset + +This completely resets the incremental parser, that is, after this call, +it will be as if the parser had never parsed anything. + +This is useful if you want to repeatedly parse JSON objects and want to +ignore any trailing data, which means you have to reset the parser after +each successful decode. + +=back + +=head2 LIMITATIONS + +All options that affect decoding are supported, except +C. The reason for this is that it cannot be made to work +sensibly: JSON objects and arrays are self-delimited, i.e. you can +concatenate them back to back and still decode them perfectly. This does +not hold true for JSON numbers, however. + +For example, is the string C<1> a single JSON number, or is it simply the +start of C<12>? Or is C<12> a single JSON number, or the concatenation +of C<1> and C<2>? In neither case you can tell, and this is why JSON::XS +takes the conservative route and disallows this case. + +=head2 EXAMPLES + +Some examples will make all this clearer. First, a simple example that +works similarly to C: We want to decode the JSON object at +the start of a string and identify the portion after the JSON object: + + my $text = "[1,2,3] hello"; + + my $json = new JSON::XS; + + my $obj = $json->incr_parse ($text) + or die "expected JSON object or array at beginning of string"; + + my $tail = $json->incr_text; + # $tail now contains " hello" + +Easy, isn't it? + +Now for a more complicated example: Imagine a hypothetical protocol where +you read some requests from a TCP stream, and each request is a JSON +array, without any separation between them (in fact, it is often useful to +use newlines as "separators", as these get interpreted as whitespace at +the start of the JSON text, which makes it possible to test said protocol +with C...). + +Here is how you'd do it (it is trivial to write this in an event-based +manner): + + my $json = new JSON::XS; + + # read some data from the socket + while (sysread $socket, my $buf, 4096) { + + # split and decode as many requests as possible + for my $request ($json->incr_parse ($buf)) { + # act on the $request + } + } + +Another complicated example: Assume you have a string with JSON objects +or arrays, all separated by (optional) comma characters (e.g. C<[1],[2], +[3]>). To parse them, we have to skip the commas between the JSON texts, +and here is where the lvalue-ness of C comes in useful: + + my $text = "[1],[2], [3]"; + my $json = new JSON::XS; + + # void context, so no parsing done + $json->incr_parse ($text); + + # now extract as many objects as possible. note the + # use of scalar context so incr_text can be called. + while (my $obj = $json->incr_parse) { + # do something with $obj + + # now skip the optional comma + $json->incr_text =~ s/^ \s* , //x; + } + +Now lets go for a very complex example: Assume that you have a gigantic +JSON array-of-objects, many gigabytes in size, and you want to parse it, +but you cannot load it into memory fully (this has actually happened in +the real world :). + +Well, you lost, you have to implement your own JSON parser. But JSON::XS +can still help you: You implement a (very simple) array parser and let +JSON decode the array elements, which are all full JSON objects on their +own (this wouldn't work if the array elements could be JSON numbers, for +example): + + my $json = new JSON::XS; + + # open the monster + open my $fh, "incr_parse ($buf); # void context, so no parsing + + # Exit the loop once we found and removed(!) the initial "[". + # In essence, we are (ab-)using the $json object as a simple scalar + # we append data to. + last if $json->incr_text =~ s/^ \s* \[ //x; + } + + # now we have the skipped the initial "[", so continue + # parsing all the elements. + for (;;) { + # in this loop we read data until we got a single JSON object + for (;;) { + if (my $obj = $json->incr_parse) { + # do something with $obj + last; + } + + # add more data + sysread $fh, my $buf, 65536 + or die "read error: $!"; + $json->incr_parse ($buf); # void context, so no parsing + } + + # in this loop we read data until we either found and parsed the + # separating "," between elements, or the final "]" + for (;;) { + # first skip whitespace + $json->incr_text =~ s/^\s*//; + + # if we find "]", we are done + if ($json->incr_text =~ s/^\]//) { + print "finished.\n"; + exit; + } + + # if we find ",", we can continue with the next element + if ($json->incr_text =~ s/^,//) { + last; + } + + # if we find anything else, we have a parse error! + if (length $json->incr_text) { + die "parse error near ", $json->incr_text; + } + + # else add more data + sysread $fh, my $buf, 65536 + or die "read error: $!"; + $json->incr_parse ($buf); # void context, so no parsing + } + +This is a complex example, but most of the complexity comes from the fact +that we are trying to be correct (bear with me if I am wrong, I never ran +the above example :). + + + =head1 MAPPING This section describes how JSON::XS maps Perl values to JSON values and @@ -708,31 +975,52 @@ string scalar in perl, depending on its range and any fractional parts. On the Perl level, there is no difference between those as Perl handles all the conversion details, but an integer may take slightly less memory and -might represent more values exactly than (floating point) numbers. +might represent more values exactly than floating point numbers. If the number consists of digits only, JSON::XS will try to represent it as an integer value. If that fails, it will try to represent it as a numeric (floating point) value if that is possible without loss of -precision. Otherwise it will preserve the number as a string value. +precision. Otherwise it will preserve the number as a string value (in +which case you lose roundtripping ability, as the JSON number will be +re-encoded to a JSON string). Numbers containing a fractional or exponential part will always be represented as numeric (floating point) values, possibly at a loss of -precision. +precision (in which case you might lose perfect roundtripping ability, but +the JSON number will still be re-encoded as a JSON number). -This might create round-tripping problems as numbers might become strings, -but as Perl is typeless there is no other way to do it. +Note that precision is not accuracy - binary floating point values cannot +represent most decimal fractions exactly, and when converting from and to +floating point, JSON::XS only guarantees precision up to but not including +the least significant bit. =item true, false -These JSON atoms become C and C, -respectively. They are overloaded to act almost exactly like the numbers -C<1> and C<0>. You can check whether a scalar is a JSON boolean by using -the C function. +These JSON atoms become C and +C, respectively. They are overloaded to act +almost exactly like the numbers C<1> and C<0>. You can check whether +a scalar is a JSON boolean by using the C +function (after C, of course). =item null A JSON null atom becomes C in Perl. +=item shell-style comments (C<< # I >>) + +As a nonstandard extension to the JSON syntax that is enabled by the +C setting, shell-style comments are allowed. They can start +anywhere outside strings and go till the end of the line. + +=item tagged values (C<< (I)I >>). + +Another nonstandard extension to the JSON syntax, enabled with the +C setting, are tagged values. In this implementation, the +I must be a perl package/class name encoded as a JSON string, and the +I must be a JSON array encoding optional constructor arguments. + +See L, below, for details. + =back @@ -746,15 +1034,13 @@ =item hash references -Perl hash references become JSON objects. As there is no inherent ordering -in hash keys (or JSON objects), they will usually be encoded in a -pseudo-random order that can change between runs of the same program but -stays generally the same within a single run of a program. JSON::XS can -optionally sort the hash keys (determined by the I flag), so -the same datastructure will serialise to the same JSON text (given same -settings and version of JSON::XS), but this incurs a runtime overhead -and is only rarely useful, e.g. when you want to compare some JSON text -against another for equality. +Perl hash references become JSON objects. As there is no inherent +ordering in hash keys (or JSON objects), they will usually be encoded +in a pseudo-random order. JSON::XS can optionally sort the hash keys +(determined by the I flag), so the same datastructure will +serialise to the same JSON text (given same settings and version of +JSON::XS), but this incurs a runtime overhead and is only rarely useful, +e.g. when you want to compare some JSON text against another for equality. =item array references @@ -764,28 +1050,33 @@ Other unblessed references are generally not allowed and will cause an exception to be thrown, except for references to the integers C<0> and -C<1>, which get turned into C and C atoms in JSON. You can -also use C and C to improve readability. +C<1>, which get turned into C and C atoms in JSON. + +Since C uses the boolean model from L, you +can also C and then use C +and C to improve readability. - encode_json [\0,JSON::XS::true] # yields [false,true] + use Types::Serialiser; + encode_json [\0, Types::Serialiser::true] # yields [false,true] -=item JSON::XS::true, JSON::XS::false +=item Types::Serialiser::true, Types::Serialiser::false -These special values become JSON true and JSON false values, -respectively. You can also use C<\1> and C<\0> directly if you want. +These special values from the L module become JSON true +and JSON false values, respectively. You can also use C<\1> and C<\0> +directly if you want. =item blessed objects -Blessed objects are not allowed. JSON::XS currently tries to encode their -underlying representation (hash- or arrayref), but this behaviour might -change in future versions. +Blessed objects are not directly representable in JSON, but C +allows various ways of handling objects. See L, +below, for details. =item simple scalars Simple Perl scalars (any scalar that is not a reference) are the most difficult objects to encode: JSON::XS will encode undefined scalars as -JSON null value, scalars that have last been used in a string context -before encoding as JSON strings and anything else as number value: +JSON C values, scalars that have last been used in a string context +before encoding as JSON strings, and anything else as number value: # dump as number encode_json [2] # yields [2] @@ -813,103 +1104,297 @@ $x *= 1; # same thing, the choice is yours. You can not currently force the type in other, less obscure, ways. Tell me -if you need this capability. +if you need this capability (but don't forget to explain why it's needed +:). -=back +Note that numerical precision has the same meaning as under Perl (so +binary to decimal conversion follows the same rules as in Perl, which +can differ to other languages). Also, your perl interpreter might expose +extensions to the floating point numbers of your platform, such as +infinities or NaN's - these cannot be represented in JSON, and it is an +error to pass those in. +=back -=head1 COMPARISON +=head2 OBJECT SERIALISATION -As already mentioned, this module was created because none of the existing -JSON modules could be made to work correctly. First I will describe the -problems (or pleasures) I encountered with various existing JSON modules, -followed by some benchmark values. JSON::XS was designed not to suffer -from any of these problems or limitations. +As JSON cannot directly represent Perl objects, you have to choose between +a pure JSON representation (without the ability to deserialise the object +automatically again), and a nonstandard extension to the JSON syntax, +tagged values. + +=head3 SERIALISATION + +What happens when C encounters a Perl object depends on the +C, C and C settings, which are +used in this order: =over 4 -=item JSON 1.07 +=item 1. C is enabled and the object has a C method. -Slow (but very portable, as it is written in pure Perl). +In this case, C uses the L object +serialisation protocol to create a tagged JSON value, using a nonstandard +extension to the JSON syntax. -Undocumented/buggy Unicode handling (how JSON handles Unicode values is -undocumented. One can get far by feeding it Unicode strings and doing -en-/decoding oneself, but Unicode escapes are not working properly). +This works by invoking the C method on the object, with the first +argument being the object to serialise, and the second argument being the +constant string C to distinguish it from other serialisers. -No round-tripping (strings get clobbered if they look like numbers, e.g. -the string C<2.0> will encode to C<2.0> instead of C<"2.0">, and that will -decode into the number 2. +The C method can return any number of values (i.e. zero or +more). These values and the paclkage/classname of the object will then be +encoded as a tagged JSON value in the following format: -=item JSON::PC 0.01 + ("classname")[FREEZE return values...] -Very fast. +e.g.: -Undocumented/buggy Unicode handling. + ("URI")["http://www.google.com/"] + ("MyDate")[2013,10,29] + ("ImageData::JPEG")["Z3...VlCg=="] -No round-tripping. +For example, the hypothetical C C method might use the +objects C and C members to encode the object: -Has problems handling many Perl values (e.g. regex results and other magic -values will make it croak). + sub My::Object::FREEZE { + my ($self, $serialiser) = @_; -Does not even generate valid JSON (C<{1,2}> gets converted to C<{1:2}> -which is not a valid JSON text. + ($self->{type}, $self->{id}) + } -Unmaintained (maintainer unresponsive for many months, bugs are not -getting fixed). +=item 2. C is enabled and the object has a C method. -=item JSON::Syck 0.21 +In this case, the C method of the object is invoked in scalar +context. It must return a single scalar that can be directly encoded into +JSON. This scalar replaces the object in the JSON text. + +For example, the following C method will convert all L +objects to JSON strings when serialised. The fatc that these values +originally were L objects is lost. + + sub URI::TO_JSON { + my ($uri) = @_; + $uri->as_string + } + +=item 3. C is enabled. -Very buggy (often crashes). +The object will be serialised as a JSON null value. -Very inflexible (no human-readable format supported, format pretty much -undocumented. I need at least a format for easy reading by humans and a -single-line compact format for use in a protocol, and preferably a way to -generate ASCII-only JSON texts). +=item 4. none of the above -Completely broken (and confusingly documented) Unicode handling (Unicode -escapes are not working properly, you need to set ImplicitUnicode to -I values on en- and decoding to get symmetric behaviour). +If none of the settings are enabled or the respective methods are missing, +C throws an exception. + +=back -No round-tripping (simple cases work, but this depends on whether the scalar -value was used in a numeric context or not). +=head3 DESERIALISATION -Dumping hashes may skip hash values depending on iterator state. +For deserialisation there are only two cases to consider: either +nonstandard tagging was used, in which case C decides, +or objects cannot be automatically be deserialised, in which +case you can use postprocessing or the C or +C callbacks to get some real objects our of +your JSON. + +This section only considers the tagged value case: I a tagged JSON object +is encountered during decoding and C is disabled, a parse +error will result (as if tagged values were not part of the grammar). + +If C is enabled, C will look up the C method +of the package/classname used during serialisation (it will not attempt +to load the package as a Perl module). If there is no such method, the +decoding will fail with an error. + +Otherwise, the C method is invoked with the classname as first +argument, the constant string C as second argument, and all the +values from the JSON array (the values originally returned by the +C method) as remaining arguments. + +The method must then return the object. While technically you can return +any Perl scalar, you might have to enable the C setting to +make that work in all cases, so better return an actual blessed reference. -Unmaintained (maintainer unresponsive for many months, bugs are not -getting fixed). +As an example, let's implement a C function that regenerates the +C from the C example earlier: -Does not check input for validity (i.e. will accept non-JSON input and -return "something" instead of raising an exception. This is a security -issue: imagine two banks transferring money between each other using -JSON. One bank might parse a given non-JSON request and deduct money, -while the other might reject the transaction with a syntax error. While a -good protocol will at least recover, that is extra unnecessary work and -the transaction will still not succeed). + sub My::Object::THAW { + my ($class, $serialiser, $type, $id) = @_; -=item JSON::DWIW 0.04 + $class->new (type => $type, id => $id) + } -Very fast. Very natural. Very nice. -Undocumented Unicode handling (but the best of the pack. Unicode escapes -still don't get parsed properly). +=head1 ENCODING/CODESET FLAG NOTES -Very inflexible. +The interested reader might have seen a number of flags that signify +encodings or codesets - C, C and C. There seems to be +some confusion on what these do, so here is a short comparison: + +C controls whether the JSON text created by C (and expected +by C) is UTF-8 encoded or not, while C and C only +control whether C escapes character values outside their respective +codeset range. Neither of these flags conflict with each other, although +some combinations make less sense than others. + +Care has been taken to make all flags symmetrical with respect to +C and C, that is, texts encoded with any combination of +these flag values will be correctly decoded when the same flags are used +- in general, if you use different flag settings while encoding vs. when +decoding you likely have a bug somewhere. + +Below comes a verbose discussion of these flags. Note that a "codeset" is +simply an abstract set of character-codepoint pairs, while an encoding +takes those codepoint numbers and I them, in our case into +octets. Unicode is (among other things) a codeset, UTF-8 is an encoding, +and ISO-8859-1 (= latin 1) and ASCII are both codesets I encodings at +the same time, which can be confusing. -No round-tripping. +=over 4 -Does not generate valid JSON texts (key strings are often unquoted, empty keys -result in nothing being output) +=item C flag disabled -Does not check input for validity. +When C is disabled (the default), then C/C generate +and expect Unicode strings, that is, characters with high ordinal Unicode +values (> 255) will be encoded as such characters, and likewise such +characters are decoded as-is, no changes to them will be done, except +"(re-)interpreting" them as Unicode codepoints or Unicode characters, +respectively (to Perl, these are the same thing in strings unless you do +funny/weird/dumb stuff). + +This is useful when you want to do the encoding yourself (e.g. when you +want to have UTF-16 encoded JSON texts) or when some other layer does +the encoding for you (for example, when printing to a terminal using a +filehandle that transparently encodes to UTF-8 you certainly do NOT want +to UTF-8 encode your data first and have Perl encode it another time). + +=item C flag enabled + +If the C-flag is enabled, C/C will encode all +characters using the corresponding UTF-8 multi-byte sequence, and will +expect your input strings to be encoded as UTF-8, that is, no "character" +of the input string must have any value > 255, as UTF-8 does not allow +that. + +The C flag therefore switches between two modes: disabled means you +will get a Unicode string in Perl, enabled means you get an UTF-8 encoded +octet/binary string in Perl. + +=item C or C flags enabled + +With C (or C) enabled, C will escape characters +with ordinal values > 255 (> 127 with C) and encode the remaining +characters as specified by the C flag. + +If C is disabled, then the result is also correctly encoded in those +character sets (as both are proper subsets of Unicode, meaning that a +Unicode string with all character values < 256 is the same thing as a +ISO-8859-1 string, and a Unicode string with all character values < 128 is +the same thing as an ASCII string in Perl). + +If C is enabled, you still get a correct UTF-8-encoded string, +regardless of these flags, just some more characters will be escaped using +C<\uXXXX> then before. + +Note that ISO-8859-1-I strings are not compatible with UTF-8 +encoding, while ASCII-encoded strings are. That is because the ISO-8859-1 +encoding is NOT a subset of UTF-8 (despite the ISO-8859-1 I being +a subset of Unicode), while ASCII is. + +Surprisingly, C will ignore these flags and so treat all input +values as governed by the C flag. If it is disabled, this allows you +to decode ISO-8859-1- and ASCII-encoded strings, as both strict subsets of +Unicode. If it is enabled, you can correctly decode UTF-8 encoded strings. + +So neither C nor C are incompatible with the C flag - +they only govern when the JSON output engine escapes a character or not. + +The main use for C is to relatively efficiently store binary data +as JSON, at the expense of breaking compatibility with most JSON decoders. + +The main use for C is to force the output to not contain characters +with values > 127, which means you can interpret the resulting string +as UTF-8, ISO-8859-1, ASCII, KOI8-R or most about any character set and +8-bit-encoding, and still get the same data structure back. This is useful +when your channel for JSON transfer is not 8-bit clean or the encoding +might be mangled in between (e.g. in mail), and works because ASCII is a +proper subset of most 8-bit and multibyte encodings in use in the world. =back +=head2 JSON and ECMAscript + +JSON syntax is based on how literals are represented in javascript (the +not-standardised predecessor of ECMAscript) which is presumably why it is +called "JavaScript Object Notation". + +However, JSON is not a subset (and also not a superset of course) of +ECMAscript (the standard) or javascript (whatever browsers actually +implement). + +If you want to use javascript's C function to "parse" JSON, you +might run into parse errors for valid JSON texts, or the resulting data +structure might not be queryable: + +One of the problems is that U+2028 and U+2029 are valid characters inside +JSON strings, but are not allowed in ECMAscript string literals, so the +following Perl fragment will not output something that can be guaranteed +to be parsable by javascript's C: + + use JSON::XS; + + print encode_json [chr 0x2028]; + +The right fix for this is to use a proper JSON parser in your javascript +programs, and not rely on C (see for example Douglas Crockford's +F parser). + +If this is not an option, you can, as a stop-gap measure, simply encode to +ASCII-only JSON: + + use JSON::XS; + + print JSON::XS->new->ascii->encode ([chr 0x2028]); + +Note that this will enlarge the resulting JSON text quite a bit if you +have many non-ASCII characters. You might be tempted to run some regexes +to only escape U+2028 and U+2029, e.g.: + + # DO NOT USE THIS! + my $json = JSON::XS->new->utf8->encode ([chr 0x2028]); + $json =~ s/\xe2\x80\xa8/\\u2028/g; # escape U+2028 + $json =~ s/\xe2\x80\xa9/\\u2029/g; # escape U+2029 + print $json; + +Note that I: the above only works for U+2028 and +U+2029 and thus only for fully ECMAscript-compliant parsers. Many existing +javascript implementations, however, have issues with other characters as +well - using C naively simply I cause problems. + +Another problem is that some javascript implementations reserve +some property names for their own purposes (which probably makes +them non-ECMAscript-compliant). For example, Iceweasel reserves the +C<__proto__> property name for its own purposes. + +If that is a problem, you could parse try to filter the resulting JSON +output for these property strings, e.g.: + + $json =~ s/"__proto__"\s*:/"__proto__renamed":/g; + +This works because C<__proto__> is not valid outside of strings, so every +occurrence of C<"__proto__"\s*:> must be a string used as property name. + +If you know of other incompatibilities, please let me know. + + =head2 JSON and YAML -You often hear that JSON is a subset (or a close subset) of YAML. This is, -however, a mass hysteria and very far from the truth. In general, there is -no way to configure JSON::XS to output a data structure as valid YAML. +You often hear that JSON is a subset of YAML. This is, however, a mass +hysteria(*) and very far from the truth (as of the time of this writing), +so let me state it clearly: I that works in all +cases. If you really must use JSON::XS to generate YAML, you should use this algorithm (subject to change in future versions): @@ -917,15 +1402,50 @@ my $to_yaml = JSON::XS->new->utf8->space_after (1); my $yaml = $to_yaml->encode ($ref) . "\n"; -This will usually generate JSON texts that also parse as valid +This will I generate JSON texts that also parse as valid YAML. Please note that YAML has hardcoded limits on (simple) object key -lengths that JSON doesn't have, so you should make sure that your hash -keys are noticeably shorter than the 1024 characters YAML allows. +lengths that JSON doesn't have and also has different and incompatible +unicode character escape syntax, so you should make sure that your hash +keys are noticeably shorter than the 1024 "stream characters" YAML allows +and that you do not have characters with codepoint values outside the +Unicode BMP (basic multilingual page). YAML also does not allow C<\/> +sequences in strings (which JSON::XS does not I generate, but +other JSON generators might). + +There might be other incompatibilities that I am not aware of (or the YAML +specification has been changed yet again - it does so quite often). In +general you should not try to generate YAML with a JSON generator or vice +versa, or try to parse JSON with a YAML parser or vice versa: chances are +high that you will run into severe interoperability problems when you +least expect it. -There might be other incompatibilities that I am not aware of. In general -you should not try to generate YAML with a JSON generator or vice versa, -or try to parse JSON with a YAML parser or vice versa: chances are high -that you will run into severe interoperability problems. +=over 4 + +=item (*) + +I have been pressured multiple times by Brian Ingerson (one of the +authors of the YAML specification) to remove this paragraph, despite him +acknowledging that the actual incompatibilities exist. As I was personally +bitten by this "JSON is YAML" lie, I refused and said I will continue to +educate people about these issues, so others do not run into the same +problem again and again. After this, Brian called me a (quote)I(unquote). + +In my opinion, instead of pressuring and insulting people who actually +clarify issues with YAML and the wrong statements of some of its +proponents, I would kindly suggest reading the JSON spec (which is not +that difficult or long) and finally make YAML compatible to it, and +educating users about the changes, instead of spreading lies about the +real compatibility for many I and trying to silence people who +point out that it isn't true. + +Addendum/2009: the YAML 1.2 spec is still incompatible with JSON, even +though the incompatibilities have been documented (and are known to Brian) +for many years and the spec makes explicit claims that YAML is a superset +of JSON. It would be so easy to fix, but apparently, bullying people and +corrupting userdata is so much easier. + +=back =head2 SPEED @@ -935,53 +1455,54 @@ in the JSON::XS distribution, to make it easy to compare on your own system. -First comes a comparison between various modules using a very short -single-line JSON string: - - {"method": "handleMessage", "params": ["user1", "we were just talking"], \ - "id": null, "array":[1,11,234,-5,1e5,1e7, true, false]} +First comes a comparison between various modules using +a very short single-line JSON string (also available at +L). + + {"method": "handleMessage", "params": ["user1", + "we were just talking"], "id": null, "array":[1,11,234,-5,1e5,1e7, + 1, 0]} It shows the number of encodes/decodes per second (JSON::XS uses the functional interface, while JSON::XS/2 uses the OO interface with pretty-printing and hashkey sorting enabled, JSON::XS/3 enables -shrink). Higher is better: +shrink. JSON::DWIW/DS uses the deserialise function, while JSON::DWIW::FJ +uses the from_json method). Higher is better: - module | encode | decode | - -----------|------------|------------| - JSON 1.x | 4990.842 | 4088.813 | - JSON::DWIW | 51653.990 | 71575.154 | - JSON::PC | 65948.176 | 74631.744 | - JSON::PP | 8931.652 | 3817.168 | - JSON::Syck | 24877.248 | 27776.848 | - JSON::XS | 388361.481 | 227951.304 | - JSON::XS/2 | 227951.304 | 218453.333 | - JSON::XS/3 | 338250.323 | 218453.333 | - Storable | 16500.016 | 135300.129 | - -----------+------------+------------+ - -That is, JSON::XS is about five times faster than JSON::DWIW on encoding, -about three times faster on decoding, and over forty times faster -than JSON, even with pretty-printing and key sorting. It also compares -favourably to Storable for small amounts of data. + module | encode | decode | + --------------|------------|------------| + JSON::DWIW/DS | 86302.551 | 102300.098 | + JSON::DWIW/FJ | 86302.551 | 75983.768 | + JSON::PP | 15827.562 | 6638.658 | + JSON::Syck | 63358.066 | 47662.545 | + JSON::XS | 511500.488 | 511500.488 | + JSON::XS/2 | 291271.111 | 388361.481 | + JSON::XS/3 | 361577.931 | 361577.931 | + Storable | 66788.280 | 265462.278 | + --------------+------------+------------+ + +That is, JSON::XS is almost six times faster than JSON::DWIW on encoding, +about five times faster on decoding, and over thirty to seventy times +faster than JSON's pure perl implementation. It also compares favourably +to Storable for small amounts of data. Using a longer test string (roughly 18KB, generated from Yahoo! Locals -search API (http://nanoref.com/yahooapis/mgPdGg): +search API (L). - module | encode | decode | - -----------|------------|------------| - JSON 1.x | 55.260 | 34.971 | - JSON::DWIW | 825.228 | 1082.513 | - JSON::PC | 3571.444 | 2394.829 | - JSON::PP | 210.987 | 32.574 | - JSON::Syck | 552.551 | 787.544 | - JSON::XS | 5780.463 | 4854.519 | - JSON::XS/2 | 3869.998 | 4798.975 | - JSON::XS/3 | 5862.880 | 4798.975 | - Storable | 4445.002 | 5235.027 | - -----------+------------+------------+ + module | encode | decode | + --------------|------------|------------| + JSON::DWIW/DS | 1647.927 | 2673.916 | + JSON::DWIW/FJ | 1630.249 | 2596.128 | + JSON::PP | 400.640 | 62.311 | + JSON::Syck | 1481.040 | 1524.869 | + JSON::XS | 20661.596 | 9541.183 | + JSON::XS/2 | 10683.403 | 9416.938 | + JSON::XS/3 | 20661.596 | 9400.054 | + Storable | 19765.806 | 10000.725 | + --------------+------------+------------+ Again, JSON::XS leads by far (except for Storable which non-surprisingly -decodes faster). +decodes a bit faster). On large strings containing lots of high Unicode characters, some modules (such as JSON::PC) seem to decode faster than JSON::XS, but the result @@ -1017,17 +1538,106 @@ has a smaller stack, you should adjust this setting accordingly with the C method. -And last but least, something else could bomb you that I forgot to think -of. In that case, you get to keep the pieces. I am always open for hints, -though... +Something else could bomb you, too, that I forgot to think of. In that +case, you get to keep the pieces. I am always open for hints, though... + +Also keep in mind that JSON::XS might leak contents of your Perl data +structures in its error messages, so when you serialise sensitive +information you might want to make sure that exceptions thrown by JSON::XS +will not end up in front of untrusted eyes. If you are using JSON::XS to return packets to consumption by JavaScript scripts in a browser you should have a look at -L to see whether -you are vulnerable to some common attack vectors (which really are browser -design bugs, but it is still you who will have to deal with it, as major -browser developers care only for features, not about getting security -right). +L to +see whether you are vulnerable to some common attack vectors (which really +are browser design bugs, but it is still you who will have to deal with +it, as major browser developers care only for features, not about getting +security right). + + +=head1 INTEROPERABILITY WITH OTHER MODULES + +C uses the L module to provide boolean +constants. That means that the JSON true and false values will be +comaptible to true and false values of iother modules that do the same, +such as L and L. + + +=head1 INTEROPERABILITY WITH OTHER JSON DECODERS + +As long as you only serialise data that can be directly expressed in JSON, +C is incapable of generating invalid JSON output (modulo bugs, +but C has found more bugs in the official JSON testsuite (1) +than the official JSON testsuite has found in C (0)). + +When you have trouble decoding JSON generated by this module using other +decoders, then it is very likely that you have an encoding mismatch or the +other decoder is broken. + +When decoding, C is strict by default and will likely catch all +errors. There are currently two settings that change this: C +makes C accept (but not generate) some non-standard extensions, +and C will allow you to encode and decode Perl objects, at the +cost of not outputting valid JSON anymore. + +=head2 TAGGED VALUE SYNTAX AND STANDARD JSON EN/DECODERS + +When you use C to use the extended (and also nonstandard and +invalid) JSON syntax for serialised objects, and you still want to decode +the generated When you want to serialise objects, you can run a regex +to replace the tagged syntax by standard JSON arrays (it only works for +"normal" packagesnames without comma, newlines or single colons). First, +the readable Perl version: + + # if your FREEZE methods return no values, you need this replace first: + $json =~ s/\( \s* (" (?: [^\\":,]+|\\.|::)* ") \s* \) \s* \[\s*\]/[$1]/gx; + + # this works for non-empty constructor arg lists: + $json =~ s/\( \s* (" (?: [^\\":,]+|\\.|::)* ") \s* \) \s* \[/[$1,/gx; + +And here is a less readable version that is easy to adapt to other +languages: + + $json =~ s/\(\s*("([^\\":,]+|\\.|::)*")\s*\)\s*\[/[$1,/g; + +Here is an ECMAScript version (same regex): + + json = json.replace (/\(\s*("([^\\":,]+|\\.|::)*")\s*\)\s*\[/g, "[$1,"); + +Since this syntax converts to standard JSON arrays, it might be hard to +distinguish serialised objects from normal arrays. You can prepend a +"magic number" as first array element to reduce chances of a collision: + + $json =~ s/\(\s*("([^\\":,]+|\\.|::)*")\s*\)\s*\[/["XU1peReLzT4ggEllLanBYq4G9VzliwKF",$1,/g; + +And after decoding the JSON text, you could walk the data +structure looking for arrays with a first element of +C. + +The same approach can be used to create the tagged format with another +encoder. First, you create an array with the magic string as first member, +the classname as second, and constructor arguments last, encode it as part +of your JSON structure, and then: + + $json =~ s/\[\s*"XU1peReLzT4ggEllLanBYq4G9VzliwKF"\s*,\s*("([^\\":,]+|\\.|::)*")\s*,/($1)[/g; + +Again, this has some limitations - the magic string must not be encoded +with character escapes, and the constructor arguments must be non-empty. + + +=head1 RFC7158 + +Since this module was written, Google has written a new JSON RFC, RFC +7158. Unfortunately, this RFC breaks compatibility with both the original +JSON specification on www.json.org and RFC4627. + +As far as I can see, you can get partial compatibility when parsing by +using C<< ->allow_nonref >>. However, consider thew security implications +of doing so. + +I haven't decided yet whether to break compatibility with RFC4627 by +default (and potentially leave applications insecure), or change the +default to follow RFC7158. =head1 THREADS @@ -1035,45 +1645,55 @@ This module is I guaranteed to be thread safe and there are no plans to change this until Perl gets thread support (as opposed to the horribly slow so-called "threads" which are simply slow and bloated -process simulations - use fork, its I faster, cheaper, better). +process simulations - use fork, it's I faster, cheaper, better). (It might actually work, but you have been warned). +=head1 THE PERILS OF SETLOCALE + +Sometimes people avoid the Perl locale support and directly call the +system's setlocale function with C. + +This breaks both perl and modules such as JSON::XS, as stringification of +numbers no longer works correctly (e.g. C<$x = 0.1; print "$x"+1> might +print C<1>, and JSON::XS might output illegal JSON as JSON::XS relies on +perl to stringify numbers). + +The solution is simple: don't call C, or use it for only those +categories you need, such as C or C. + +If you need C, you should enable it only around the code that +actually needs it (avoiding stringification of numbers), and restore it +afterwards. + + =head1 BUGS While the goal of this module is to be correct, that unfortunately does -not mean its bug-free, only that I think its design is bug-free. It is -still relatively early in its development. If you keep reporting bugs they -will be fixed swiftly, though. +not mean it's bug-free, only that I think its design is bug-free. If you +keep reporting bugs they will be fixed swiftly, though. Please refrain from using rt.cpan.org or any other bug reporting service. I put the contact address into my modules for a reason. =cut -our $true = do { bless \(my $dummy = 1), "JSON::XS::Boolean" }; -our $false = do { bless \(my $dummy = 0), "JSON::XS::Boolean" }; - -sub true() { $true } -sub false() { $false } +BEGIN { + *true = \$Types::Serialiser::true; + *true = \&Types::Serialiser::true; + *false = \$Types::Serialiser::false; + *false = \&Types::Serialiser::false; + *is_bool = \&Types::Serialiser::is_bool; -sub is_bool($) { - UNIVERSAL::isa $_[0], "JSON::XS::Boolean" -# or UNIVERSAL::isa $_[0], "JSON::Literal" + *JSON::XS::Boolean:: = *Types::Serialiser::Boolean::; } XSLoader::load "JSON::XS", $VERSION; -package JSON::XS::Boolean; +=head1 SEE ALSO -use overload - "0+" => sub { ${$_[0]} }, - "++" => sub { $_[0] = ${$_[0]} + 1 }, - "--" => sub { $_[0] = ${$_[0]} - 1 }, - fallback => 1; - -1; +The F command line utility for quick experiments. =head1 AUTHOR @@ -1082,3 +1702,5 @@ =cut +1 +