File: //usr/local/share/man/man3/Cpanel::JSON::XS.3pm
.\" Automatically generated by Pod::Man 4.11 (Pod::Simple 3.35)
.\"
.\" Standard preamble:
.\" ========================================================================
.de Sp \" Vertical space (when we can't use .PP)
.if t .sp .5v
.if n .sp
..
.de Vb \" Begin verbatim text
.ft CW
.nf
.ne \\$1
..
.de Ve \" End verbatim text
.ft R
.fi
..
.\" Set up some character translations and predefined strings. \*(-- will
.\" give an unbreakable dash, \*(PI will give pi, \*(L" will give a left
.\" double quote, and \*(R" will give a right double quote. \*(C+ will
.\" give a nicer C++. Capital omega is used to do unbreakable dashes and
.\" therefore won't be available. \*(C` and \*(C' expand to `' in nroff,
.\" nothing in troff, for use with C<>.
.tr \(*W-
.ds C+ C\v'-.1v'\h'-1p'\s-2+\h'-1p'+\s0\v'.1v'\h'-1p'
.ie n \{\
. ds -- \(*W-
. ds PI pi
. if (\n(.H=4u)&(1m=24u) .ds -- \(*W\h'-12u'\(*W\h'-12u'-\" diablo 10 pitch
. if (\n(.H=4u)&(1m=20u) .ds -- \(*W\h'-12u'\(*W\h'-8u'-\" diablo 12 pitch
. ds L" ""
. ds R" ""
. ds C` ""
. ds C' ""
'br\}
.el\{\
. ds -- \|\(em\|
. ds PI \(*p
. ds L" ``
. ds R" ''
. ds C`
. ds C'
'br\}
.\"
.\" Escape single quotes in literal strings from groff's Unicode transform.
.ie \n(.g .ds Aq \(aq
.el .ds Aq '
.\"
.\" If the F register is >0, we'll generate index entries on stderr for
.\" titles (.TH), headers (.SH), subsections (.SS), items (.Ip), and index
.\" entries marked with X<> in POD. Of course, you'll have to process the
.\" output yourself in some meaningful fashion.
.\"
.\" Avoid warning from groff about undefined register 'F'.
.de IX
..
.nr rF 0
.if \n(.g .if rF .nr rF 1
.if (\n(rF:(\n(.g==0)) \{\
. if \nF \{\
. de IX
. tm Index:\\$1\t\\n%\t"\\$2"
..
. if !\nF==2 \{\
. nr % 0
. nr F 2
. \}
. \}
.\}
.rr rF
.\" ========================================================================
.\"
.IX Title "XS 3"
.TH XS 3 "2021-10-14" "perl v5.26.3" "User Contributed Perl Documentation"
.\" For nroff, turn off justification. Always turn off hyphenation; it makes
.\" way too many mistakes in technical documents.
.if n .ad l
.nh
.SH "NAME"
Cpanel::JSON::XS \- cPanel fork of JSON::XS, fast and correct serializing
.SH "SYNOPSIS"
.IX Header "SYNOPSIS"
.Vb 1
\& use Cpanel::JSON::XS;
\&
\& # exported functions, they croak on error
\& # and expect/generate UTF\-8
\&
\& $utf8_encoded_json_text = encode_json $perl_hash_or_arrayref;
\& $perl_hash_or_arrayref = decode_json $utf8_encoded_json_text;
\&
\& # OO\-interface
\&
\& $coder = Cpanel::JSON::XS\->new\->ascii\->pretty\->allow_nonref;
\& $pretty_printed_unencoded = $coder\->encode ($perl_scalar);
\& $perl_scalar = $coder\->decode ($unicode_json_text);
\&
\& # Note that 5.6 misses most smart utf8 and encoding functionalities
\& # of newer releases.
\&
\& # Note that L<JSON::MaybeXS> will automatically use Cpanel::JSON::XS
\& # if available, at virtually no speed overhead either, so you should
\& # be able to just:
\&
\& use JSON::MaybeXS;
\&
\& # and do the same things, except that you have a pure\-perl fallback now.
\&
\& Note that this module will be replaced by a new JSON::Safe module soon,
\& with the same API just guaranteed safe defaults.
.Ve
.SH "DESCRIPTION"
.IX Header "DESCRIPTION"
This module converts Perl data structures to \s-1JSON\s0 and vice versa. Its
primary goal is to be \fIcorrect\fR and its secondary goal is to be
\&\fIfast\fR. To reach the latter goal it was written in C.
.PP
As this is the n\-th-something \s-1JSON\s0 module on \s-1CPAN,\s0 what was the reason
to write yet another \s-1JSON\s0 module? While it seems there are many \s-1JSON\s0
modules, none of them correctly handle all corner cases, and in most cases
their maintainers are unresponsive, gone missing, or not listening to bug
reports for other reasons.
.PP
See below for the cPanel fork.
.PP
See \s-1MAPPING,\s0 below, on how Cpanel::JSON::XS maps perl values to \s-1JSON\s0
values and vice versa.
.SS "\s-1FEATURES\s0"
.IX Subsection "FEATURES"
.IP "\(bu" 4
correct Unicode handling
.Sp
This module knows how to handle Unicode with Perl version higher than 5.8.5,
documents how and when it does so, and even documents what \*(L"correct\*(R" means.
.IP "\(bu" 4
round-trip integrity
.Sp
When you serialize a perl data structure using only data types supported
by \s-1JSON\s0 and Perl, the deserialized data structure is identical on the Perl
level. (e.g. the string \*(L"2.0\*(R" doesn't suddenly become \*(L"2\*(R" just because
it looks like a number). There \fIare\fR minor exceptions to this, read the
\&\s-1MAPPING\s0 section below to learn about those.
.IP "\(bu" 4
strict checking of \s-1JSON\s0 correctness
.Sp
There is no guessing, no generating of illegal \s-1JSON\s0 texts by default,
and only \s-1JSON\s0 is accepted as input by default. the latter is a security
feature.
.IP "\(bu" 4
fast
.Sp
Compared to other \s-1JSON\s0 modules and other serializers such as Storable,
this module usually compares favourably in terms of speed, too.
.IP "\(bu" 4
simple to use
.Sp
This module has both a simple functional interface as well as an object
oriented interface.
.IP "\(bu" 4
reasonably versatile output formats
.Sp
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.
.SS "cPanel fork"
.IX Subsection "cPanel fork"
Since the original author \s-1MLEHMANN\s0 has no public
bugtracker, this cPanel fork sits now on github.
.PP
src repo: <https://github.com/rurban/Cpanel\-JSON\-XS>
original: <http://cvs.schmorp.de/JSON\-XS/>
.PP
\&\s-1RT:\s0 <https://github.com/rurban/Cpanel\-JSON\-XS/issues>
or <https://rt.cpan.org/Public/Dist/Display.html?Queue=Cpanel\-JSON\-XS>
.PP
\&\fBChanges to \s-1JSON::XS\s0\fR
.PP
\&\- stricter \fBdecode_json()\fR as documented. non-refs are disallowed.
added a 2nd optional argument. \fBdecode()\fR honors now allow_nonref.
.PP
\&\- fixed encode of numbers for dual-vars. Different string
representations are preserved, but numbers with temporary strings
which represent the same number are here treated as numbers, not
strings. Cpanel::JSON::XS is a bit slower, but preserves numeric
types better.
.PP
\&\- numbers ending with .0 stray numbers, are not converted to
integers. [#63] dual-vars which are represented as number not
integer (42+\*(L"bar\*(R" != 5.8.9) are now encoded as number (=> 42.0)
because internally it's now a \s-1NOK\s0 type. However !!1 which is
wrongly encoded in 5.8 as \*(L"1\*(R"/1.0 is still represented as integer.
.PP
\&\- different handling of inf/nan. Default now to null, optionally with
\fBstringify_infnan()\fR to \*(L"inf\*(R"/\*(L"nan\*(R". [#28, #32]
.PP
\&\- added \f(CW\*(C`binary\*(C'\fR extension, non-JSON and non \s-1JSON\s0 parsable, allows
\f(CW\*(C`\exNN\*(C'\fR and \f(CW\*(C`\eNNN\*(C'\fR sequences.
.PP
\&\- 5.6.2 support; sacrificing some utf8 features (assuming bytes
all-over), no multi-byte unicode characters with 5.6.
.PP
\&\- interop for true/false overloading. \s-1JSON::XS, JSON::PP\s0 and Mojo::JSON
representations for booleans are accepted and \s-1JSON::XS\s0 accepts
Cpanel::JSON::XS booleans [#13, #37]
Fixed overloading of booleans. Cpanel::JSON::XS::true stringifies again
to \*(L"1\*(R", not \*(L"true\*(R", analog to all other \s-1JSON\s0 modules.
.PP
\&\- native boolean mapping of yes and no to true and false, as in \s-1YAML::XS.\s0
In perl \f(CW\*(C`!0\*(C'\fR is yes, \f(CW\*(C`!1\*(C'\fR is no.
The \s-1JSON\s0 value true maps to 1, false maps to 0. [#39]
.PP
\&\- support arbitrary stringification with encode, with convert_blessed
and allow_blessed.
.PP
\&\- ithread support. Cpanel::JSON::XS is thread-safe, \s-1JSON::XS\s0 not
.PP
\&\- is_bool can be called as method, JSON::XS::is_bool not.
.PP
\&\- performance optimizations for threaded Perls
.PP
\&\- relaxed mode, allowing many popular extensions
.PP
\&\- additional fixes for:
.PP
.Vb 1
\& \- [cpan #88061] AIX atof without USE_LONG_DOUBLE
\&
\& \- #10 unshare_hek crash
\&
\& \- #7, #29 avoid re\-blessing where possible. It fails in JSON::XS for
\& READONLY values, i.e. restricted hashes.
\&
\& \- #41 overloading of booleans, use the object not the reference.
\&
\& \- #62 \-Dusequadmath conversion and no SEGV.
\&
\& \- #72 parsing of values followed \e0, like 1\e0 does fail.
\&
\& \- #72 parsing of illegal unicode or non\-unicode characters.
\&
\& \- #96 locale\-insensitive numeric conversion.
\&
\& \- #154 numeric conversion fixed since 5.22, using the same strtold as perl5.
\&
\& \- #167 sort tied hashes with canonical.
.Ve
.PP
\&\- public maintenance and bugtracker
.PP
\&\- use ppport.h, sanify \s-1XS\s0.xs comment styles, harness C coding style
.PP
\&\- common::sense is optional. When available it is not used in the
published production module, just during development and testing.
.PP
\&\- extended testsuite, passes all http://seriot.ch/parsing_json.html
tests. In fact it is the only know \s-1JSON\s0 decoder which does so,
while also being the fastest.
.PP
\&\- support many more options and methods from \s-1JSON::PP:\s0
stringify_infnan, allow_unknown, allow_stringify, allow_barekey,
encode_stringify, allow_bignum, allow_singlequote, sort_by
(partially), escape_slash, convert_blessed, ... optional
decode_json(, allow_nonref) arg.
relaxed implements allow_dupkeys.
.PP
\&\- support all 5 unicode \s-1BOM\s0's: \s-1UTF\-8, UTF\-16LE, UTF\-16BE, UTF\-32LE,\s0
\s-1UTF\-32BE,\s0 encoding internally to \s-1UTF\-8.\s0
.SH "FUNCTIONAL INTERFACE"
.IX Header "FUNCTIONAL INTERFACE"
The following convenience methods are provided by this module. They are
exported by default:
.ie n .IP "$json_text = encode_json $perl_scalar, [json_type]" 4
.el .IP "\f(CW$json_text\fR = encode_json \f(CW$perl_scalar\fR, [json_type]" 4
.IX Item "$json_text = encode_json $perl_scalar, [json_type]"
Converts the given Perl data structure to a \s-1UTF\-8\s0 encoded, binary string
(that is, the string contains octets only). Croaks on error.
.Sp
This function call is functionally identical to:
.Sp
.Vb 1
\& $json_text = Cpanel::JSON::XS\->new\->utf8\->encode ($perl_scalar, $json_type)
.Ve
.Sp
Except being faster.
.Sp
For the type argument see Cpanel::JSON::XS::Type.
.ie n .IP "$perl_scalar = decode_json $json_text [, $allow_nonref [, my $json_type ] ]" 4
.el .IP "\f(CW$perl_scalar\fR = decode_json \f(CW$json_text\fR [, \f(CW$allow_nonref\fR [, my \f(CW$json_type\fR ] ]" 4
.IX Item "$perl_scalar = decode_json $json_text [, $allow_nonref [, my $json_type ] ]"
The opposite of \f(CW\*(C`encode_json\*(C'\fR: expects an \s-1UTF\-8\s0 (binary) string of an
json reference and tries to parse that as an \s-1UTF\-8\s0 encoded \s-1JSON\s0 text,
returning the resulting reference. Croaks on error.
.Sp
This function call is functionally identical to:
.Sp
.Vb 1
\& $perl_scalar = Cpanel::JSON::XS\->new\->utf8\->decode ($json_text, $json_type)
.Ve
.Sp
except being faster.
.Sp
Note that older decode_json versions in Cpanel::JSON::XS older than
3.0116 and \s-1JSON::XS\s0 did not set allow_nonref but allowed them due to a
bug in the decoder.
.Sp
If the new 2nd optional \f(CW$allow_nonref\fR argument is set and not false, the
\&\f(CW\*(C`allow_nonref\*(C'\fR option will be set and the function will act is described
as in the relaxed \s-1RFC 7159\s0 allowing all values such as objects,
arrays, strings, numbers, \*(L"null\*(R", \*(L"true\*(R", and \*(L"false\*(R".
See "\*(L"\s-1OLD\*(R" VS. \*(L"NEW\*(R" JSON\s0 (\s-1RFC 4627 VS. RFC 7159\s0)" below, why you don't
want to do that.
.Sp
For the 3rd optional type argument see Cpanel::JSON::XS::Type.
.ie n .IP "$is_boolean = Cpanel::JSON::XS::is_bool $scalar" 4
.el .IP "\f(CW$is_boolean\fR = Cpanel::JSON::XS::is_bool \f(CW$scalar\fR" 4
.IX Item "$is_boolean = Cpanel::JSON::XS::is_bool $scalar"
Returns true if the passed scalar represents either \f(CW\*(C`JSON::PP::true\*(C'\fR
or \f(CW\*(C`JSON::PP::false\*(C'\fR, two constants that act like \f(CW1\fR and \f(CW0\fR,
respectively and are used to represent \s-1JSON\s0 \f(CW\*(C`true\*(C'\fR and \f(CW\*(C`false\*(C'\fR
values in Perl. (Also recognizes the booleans produced by \s-1JSON::XS\s0.)
.Sp
See \s-1MAPPING,\s0 below, for more information on how \s-1JSON\s0 values are mapped
to Perl.
.SH "DEPRECATED FUNCTIONS"
.IX Header "DEPRECATED FUNCTIONS"
.IP "from_json" 4
.IX Item "from_json"
from_json has been renamed to decode_json
.IP "to_json" 4
.IX Item "to_json"
to_json has been renamed to encode_json
.SH "A FEW NOTES ON UNICODE AND PERL"
.IX Header "A FEW NOTES ON UNICODE AND PERL"
Since this often leads to confusion, here are a few very clear words on
how Unicode works in Perl, modulo bugs.
.IP "1. Perl strings can store characters with ordinal values > 255." 4
.IX Item "1. Perl strings can store characters with ordinal values > 255."
This enables you to store Unicode characters as single characters in a
Perl string \- very natural.
.IP "2. Perl does \fInot\fR associate an encoding with your strings." 4
.IX Item "2. Perl does not associate an encoding with your strings."
\&... 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 \fIuse\fR that decides encoding, not any
magical meta data.
.IP "3. The internal utf\-8 flag has no meaning with regards to the encoding of your string." 4
.IX Item "3. The internal utf-8 flag has no meaning with regards to the encoding of your string."
.PD 0
.ie n .IP "4. A ""Unicode String"" is simply a string where each character can be validly interpreted as a Unicode code point." 4
.el .IP "4. A ``Unicode String'' is simply a string where each character can be validly interpreted as a Unicode code point." 4
.IX Item "4. A Unicode String is simply a string where each character can be validly interpreted as a Unicode code point."
.PD
If you have \s-1UTF\-8\s0 encoded data, it is no longer a Unicode string, but
a Unicode string encoded in \s-1UTF\-8,\s0 giving you a binary string.
.ie n .IP "5. A string containing ""high"" (> 255) character values is \fInot\fR a \s-1UTF\-8\s0 string." 4
.el .IP "5. A string containing ``high'' (> 255) character values is \fInot\fR a \s-1UTF\-8\s0 string." 4
.IX Item "5. A string containing high (> 255) character values is not a UTF-8 string."
.PD 0
.IP "6. Unicode noncharacters only warn, as in core." 4
.IX Item "6. Unicode noncharacters only warn, as in core."
.PD
The 66 Unicode noncharacters U+FDD0..U+FDEF, and U+*FFFE, U+*FFFF just
warn, see <http://www.unicode.org/versions/corrigendum9.html>. But
illegal surrogate pairs fail to parse.
.IP "7. Raw non-Unicode characters above U+10FFFF are disallowed." 4
.IX Item "7. Raw non-Unicode characters above U+10FFFF are disallowed."
Raw non-Unicode characters outside the valid unicode range fail to
parse, because \*(L"A string is a sequence of zero or more Unicode
characters\*(R" \s-1RFC 7159\s0 section 1 and "\s-1JSON\s0 text \s-1SHALL\s0 be encoded in
Unicode \s-1RFC 7159\s0 section 8.1. We use now the \s-1UTF8_DISALLOW_SUPER\s0
flag when parsing unicode.
.PP
I hope this helps :)
.SH "OBJECT-ORIENTED INTERFACE"
.IX Header "OBJECT-ORIENTED INTERFACE"
The object oriented interface lets you configure your own encoding or
decoding style, within the limits of supported formats.
.ie n .IP "$json = new Cpanel::JSON::XS" 4
.el .IP "\f(CW$json\fR = new Cpanel::JSON::XS" 4
.IX Item "$json = new Cpanel::JSON::XS"
Creates a new \s-1JSON\s0 object that can be used to de/encode \s-1JSON\s0
strings. All boolean flags described below are by default \fIdisabled\fR.
.Sp
The mutators for flags all return the \s-1JSON\s0 object again and thus calls can
be chained:
.Sp
.Vb 2
\& my $json = Cpanel::JSON::XS\->new\->utf8\->space_after\->encode ({a => [1,2]})
\& => {"a": [1, 2]}
.Ve
.ie n .IP "$json = $json\->ascii ([$enable])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->ascii ([$enable])" 4
.IX Item "$json = $json->ascii ([$enable])"
.PD 0
.ie n .IP "$enabled = $json\->get_ascii" 4
.el .IP "\f(CW$enabled\fR = \f(CW$json\fR\->get_ascii" 4
.IX Item "$enabled = $json->get_ascii"
.PD
If \f(CW$enable\fR is true (or missing), then the \f(CW\*(C`encode\*(C'\fR method will not
generate characters outside the code range \f(CW0..127\fR (which is \s-1ASCII\s0). Any
Unicode characters outside that range will be escaped using either a
single \f(CW\*(C`\euXXXX\*(C'\fR (\s-1BMP\s0 characters) or a double \f(CW\*(C`\euHHHH\euLLLLL\*(C'\fR escape sequence,
as per \s-1RFC4627.\s0 The resulting encoded \s-1JSON\s0 text can be treated as a native
Unicode string, an ascii-encoded, latin1\-encoded or \s-1UTF\-8\s0 encoded string,
or any other superset of \s-1ASCII.\s0
.Sp
If \f(CW$enable\fR is false, then the \f(CW\*(C`encode\*(C'\fR method will not escape Unicode
characters unless required by the \s-1JSON\s0 syntax or other flags. This results
in a faster and more compact format.
.Sp
See also the section \fI\s-1ENCODING/CODESET FLAG NOTES\s0\fR later in this
document.
.Sp
The main use for this flag is to produce \s-1JSON\s0 texts that can be
transmitted over a 7\-bit channel, as the encoded \s-1JSON\s0 texts will not
contain any 8 bit characters.
.Sp
.Vb 2
\& Cpanel::JSON::XS\->new\->ascii (1)\->encode ([chr 0x10401])
\& => ["\eud801\eudc01"]
.Ve
.ie n .IP "$json = $json\->latin1 ([$enable])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->latin1 ([$enable])" 4
.IX Item "$json = $json->latin1 ([$enable])"
.PD 0
.ie n .IP "$enabled = $json\->get_latin1" 4
.el .IP "\f(CW$enabled\fR = \f(CW$json\fR\->get_latin1" 4
.IX Item "$enabled = $json->get_latin1"
.PD
If \f(CW$enable\fR is true (or missing), then the \f(CW\*(C`encode\*(C'\fR method will encode
the resulting \s-1JSON\s0 text as latin1 (or \s-1ISO\-8859\-1\s0), escaping any characters
outside the code range \f(CW0..255\fR. The resulting string can be treated as a
latin1\-encoded \s-1JSON\s0 text or a native Unicode string. The \f(CW\*(C`decode\*(C'\fR method
will not be affected in any way by this flag, as \f(CW\*(C`decode\*(C'\fR by default
expects Unicode, which is a strict superset of latin1.
.Sp
If \f(CW$enable\fR is false, then the \f(CW\*(C`encode\*(C'\fR method will not escape Unicode
characters unless required by the \s-1JSON\s0 syntax or other flags.
.Sp
See also the section \fI\s-1ENCODING/CODESET FLAG NOTES\s0\fR later in this
document.
.Sp
The main use for this flag is efficiently encoding binary data as \s-1JSON\s0
text, as most octets will not be escaped, resulting in a smaller encoded
size. The disadvantage is that the resulting \s-1JSON\s0 text is encoded
in latin1 (and must correctly be treated as such when storing and
transferring), a rare encoding for \s-1JSON.\s0 It is therefore most useful when
you want to store data structures known to contain binary data efficiently
in files or databases, not when talking to other \s-1JSON\s0 encoders/decoders.
.Sp
.Vb 2
\& Cpanel::JSON::XS\->new\->latin1\->encode (["\ex{89}\ex{abc}"]
\& => ["\ex{89}\e\eu0abc"] # (perl syntax, U+abc escaped, U+89 not)
.Ve
.ie n .IP "$json = $json\->binary ([$enable])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->binary ([$enable])" 4
.IX Item "$json = $json->binary ([$enable])"
.PD 0
.ie n .IP "$enabled = $json = $json\->get_binary" 4
.el .IP "\f(CW$enabled\fR = \f(CW$json\fR = \f(CW$json\fR\->get_binary" 4
.IX Item "$enabled = $json = $json->get_binary"
.PD
If the \f(CW$enable\fR argument is true (or missing), then the \f(CW\*(C`encode\*(C'\fR
method will not try to detect an \s-1UTF\-8\s0 encoding in any \s-1JSON\s0 string, it
will strictly interpret it as byte sequence. The result might contain
new \f(CW\*(C`\exNN\*(C'\fR sequences, which is \fBunparsable \s-1JSON\s0\fR. The \f(CW\*(C`decode\*(C'\fR
method forbids \f(CW\*(C`\euNNNN\*(C'\fR sequences and accepts \f(CW\*(C`\exNN\*(C'\fR and octal
\&\f(CW\*(C`\eNNN\*(C'\fR sequences.
.Sp
There is also a special logic for perl 5.6 and utf8. 5.6 encodes any
string to utf\-8 automatically when seeing a codepoint >= \f(CW0x80\fR and
< \f(CW0x100\fR. With the binary flag enabled decode the perl utf8 encoded
string to the original byte encoding and encode this with \f(CW\*(C`\exNN\*(C'\fR
escapes. This will result to the same encodings as with newer
perls. But note that binary multi-byte codepoints with 5.6 will
result in \f(CW\*(C`illegal unicode character in binary string\*(C'\fR errors,
unlike with newer perls.
.Sp
If \f(CW$enable\fR is false, then the \f(CW\*(C`encode\*(C'\fR method will smartly try to
detect Unicode characters unless required by the \s-1JSON\s0 syntax or other
flags and hex and octal sequences are forbidden.
.Sp
See also the section \fI\s-1ENCODING/CODESET FLAG NOTES\s0\fR later in this
document.
.Sp
The main use for this flag is to avoid the smart unicode detection and
possible double encoding. The disadvantage is that the resulting \s-1JSON\s0
text is encoded in new \f(CW\*(C`\exNN\*(C'\fR and in latin1 characters and must
correctly be treated as such when storing and transferring, a rare
encoding for \s-1JSON.\s0 It will produce non-readable \s-1JSON\s0 strings in the
browser. It is therefore most useful when you want to store data
structures known to contain binary data efficiently in files or
databases, not when talking to other \s-1JSON\s0 encoders/decoders. The
binary decoding method can also be used when an encoder produced a
non-JSON conformant hex or octal encoding \f(CW\*(C`\exNN\*(C'\fR or \f(CW\*(C`\eNNN\*(C'\fR.
.Sp
.Vb 3
\& Cpanel::JSON::XS\->new\->binary\->encode (["\ex{89}\ex{abc}"])
\& 5.6: Error: malformed or illegal unicode character in binary string
\& >=5.8: [\*(Aq\ex89\exe0\exaa\exbc\*(Aq]
\&
\& Cpanel::JSON::XS\->new\->binary\->encode (["\ex{89}\ex{bc}"])
\& => ["\ex89\exbc"]
\&
\& Cpanel::JSON::XS\->new\->binary\->decode (["\ex89\eua001"])
\& Error: malformed or illegal unicode character in binary string
\&
\& Cpanel::JSON::XS\->new\->decode (["\ex89"])
\& Error: illegal hex character in non\-binary string
.Ve
.ie n .IP "$json = $json\->utf8 ([$enable])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->utf8 ([$enable])" 4
.IX Item "$json = $json->utf8 ([$enable])"
.PD 0
.ie n .IP "$enabled = $json\->get_utf8" 4
.el .IP "\f(CW$enabled\fR = \f(CW$json\fR\->get_utf8" 4
.IX Item "$enabled = $json->get_utf8"
.PD
If \f(CW$enable\fR is true (or missing), then the \f(CW\*(C`encode\*(C'\fR method will encode
the \s-1JSON\s0 result into \s-1UTF\-8,\s0 as required by many protocols, while the
\&\f(CW\*(C`decode\*(C'\fR method expects to be handled an UTF\-8\-encoded string. Please
note that UTF\-8\-encoded strings do not contain any characters outside the
range \f(CW0..255\fR, they are thus useful for bytewise/binary I/O. In future
versions, enabling this option might enable autodetection of the \s-1UTF\-16\s0
and \s-1UTF\-32\s0 encoding families, as described in \s-1RFC4627.\s0
.Sp
If \f(CW$enable\fR is false, then the \f(CW\*(C`encode\*(C'\fR method will return the \s-1JSON\s0
string as a (non-encoded) Unicode string, while \f(CW\*(C`decode\*(C'\fR expects thus a
Unicode string. Any decoding or encoding (e.g. to \s-1UTF\-8\s0 or \s-1UTF\-16\s0) needs
to be done yourself, e.g. using the Encode module.
.Sp
See also the section \fI\s-1ENCODING/CODESET FLAG NOTES\s0\fR later in this
document.
.Sp
Example, output UTF\-16BE\-encoded \s-1JSON:\s0
.Sp
.Vb 2
\& use Encode;
\& $jsontext = encode "UTF\-16BE", Cpanel::JSON::XS\->new\->encode ($object);
.Ve
.Sp
Example, decode UTF\-32LE\-encoded \s-1JSON:\s0
.Sp
.Vb 2
\& use Encode;
\& $object = Cpanel::JSON::XS\->new\->decode (decode "UTF\-32LE", $jsontext);
.Ve
.ie n .IP "$json = $json\->pretty ([$enable])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->pretty ([$enable])" 4
.IX Item "$json = $json->pretty ([$enable])"
This enables (or disables) all of the \f(CW\*(C`indent\*(C'\fR, \f(CW\*(C`space_before\*(C'\fR and
\&\f(CW\*(C`space_after\*(C'\fR (and in the future possibly more) flags in one call to
generate the most readable (or most compact) form possible.
.Sp
Example, pretty-print some simple structure:
.Sp
.Vb 8
\& my $json = Cpanel::JSON::XS\->new\->pretty(1)\->encode ({a => [1,2]})
\& =>
\& {
\& "a" : [
\& 1,
\& 2
\& ]
\& }
.Ve
.ie n .IP "$json = $json\->indent ([$enable])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->indent ([$enable])" 4
.IX Item "$json = $json->indent ([$enable])"
.PD 0
.ie n .IP "$enabled = $json\->get_indent" 4
.el .IP "\f(CW$enabled\fR = \f(CW$json\fR\->get_indent" 4
.IX Item "$enabled = $json->get_indent"
.PD
If \f(CW$enable\fR is true (or missing), then the \f(CW\*(C`encode\*(C'\fR method will use
a multiline format as output, putting every array member or
object/hash key-value pair into its own line, indenting them properly.
.Sp
If \f(CW$enable\fR is false, no newlines or indenting will be produced, and the
resulting \s-1JSON\s0 text is guaranteed not to contain any \f(CW\*(C`newlines\*(C'\fR.
.Sp
This setting has no effect when decoding \s-1JSON\s0 texts.
.ie n .IP "$json = $json\->indent_length([$number_of_spaces])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->indent_length([$number_of_spaces])" 4
.IX Item "$json = $json->indent_length([$number_of_spaces])"
.PD 0
.ie n .IP "$length = $json\->\fBget_indent_length()\fR" 4
.el .IP "\f(CW$length\fR = \f(CW$json\fR\->\fBget_indent_length()\fR" 4
.IX Item "$length = $json->get_indent_length()"
.PD
Set the indent length (default \f(CW3\fR).
This option is only useful when you also enable indent or pretty.
The acceptable range is from 0 (no indentation) to 15
.ie n .IP "$json = $json\->space_before ([$enable])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->space_before ([$enable])" 4
.IX Item "$json = $json->space_before ([$enable])"
.PD 0
.ie n .IP "$enabled = $json\->get_space_before" 4
.el .IP "\f(CW$enabled\fR = \f(CW$json\fR\->get_space_before" 4
.IX Item "$enabled = $json->get_space_before"
.PD
If \f(CW$enable\fR is true (or missing), then the \f(CW\*(C`encode\*(C'\fR method will add an extra
optional space before the \f(CW\*(C`:\*(C'\fR separating keys from values in \s-1JSON\s0 objects.
.Sp
If \f(CW$enable\fR is false, then the \f(CW\*(C`encode\*(C'\fR method will not add any extra
space at those places.
.Sp
This setting has no effect when decoding \s-1JSON\s0 texts. You will also
most likely combine this setting with \f(CW\*(C`space_after\*(C'\fR.
.Sp
Example, space_before enabled, space_after and indent disabled:
.Sp
.Vb 1
\& {"key" :"value"}
.Ve
.ie n .IP "$json = $json\->space_after ([$enable])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->space_after ([$enable])" 4
.IX Item "$json = $json->space_after ([$enable])"
.PD 0
.ie n .IP "$enabled = $json\->get_space_after" 4
.el .IP "\f(CW$enabled\fR = \f(CW$json\fR\->get_space_after" 4
.IX Item "$enabled = $json->get_space_after"
.PD
If \f(CW$enable\fR is true (or missing), then the \f(CW\*(C`encode\*(C'\fR method will add
an extra optional space after the \f(CW\*(C`:\*(C'\fR separating keys from values in
\&\s-1JSON\s0 objects and extra whitespace after the \f(CW\*(C`,\*(C'\fR separating key-value
pairs and array members.
.Sp
If \f(CW$enable\fR is false, then the \f(CW\*(C`encode\*(C'\fR method will not add any extra
space at those places.
.Sp
This setting has no effect when decoding \s-1JSON\s0 texts.
.Sp
Example, space_before and indent disabled, space_after enabled:
.Sp
.Vb 1
\& {"key": "value"}
.Ve
.ie n .IP "$json = $json\->relaxed ([$enable])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->relaxed ([$enable])" 4
.IX Item "$json = $json->relaxed ([$enable])"
.PD 0
.ie n .IP "$enabled = $json\->get_relaxed" 4
.el .IP "\f(CW$enabled\fR = \f(CW$json\fR\->get_relaxed" 4
.IX Item "$enabled = $json->get_relaxed"
.PD
If \f(CW$enable\fR is true (or missing), then \f(CW\*(C`decode\*(C'\fR will accept some
extensions to normal \s-1JSON\s0 syntax (see below). \f(CW\*(C`encode\*(C'\fR will not be
affected in anyway. \fIBe aware that this option makes you accept invalid
\&\s-1JSON\s0 texts as if they were valid!\fR. I suggest only to use this option to
parse application-specific files written by humans (configuration files,
resource files etc.)
.Sp
If \f(CW$enable\fR is false (the default), then \f(CW\*(C`decode\*(C'\fR will only accept
valid \s-1JSON\s0 texts.
.Sp
Currently accepted extensions are:
.RS 4
.IP "\(bu" 4
list items can have an end-comma
.Sp
\&\s-1JSON\s0 \fIseparates\fR array elements and key-value pairs with commas. This
can be annoying if you write \s-1JSON\s0 texts manually and want to be able to
quickly append elements, so this extension accepts comma at the end of
such items not just between them:
.Sp
.Vb 8
\& [
\& 1,
\& 2, <\- this comma not normally allowed
\& ]
\& {
\& "k1": "v1",
\& "k2": "v2", <\- this comma not normally allowed
\& }
.Ve
.IP "\(bu" 4
shell-style '#'\-comments
.Sp
Whenever \s-1JSON\s0 allows whitespace, shell-style comments are additionally
allowed. They are terminated by the first carriage-return or line-feed
character, after which more white-space and comments are allowed.
.Sp
.Vb 4
\& [
\& 1, # this comment not allowed in JSON
\& # neither this one...
\& ]
.Ve
.IP "\(bu" 4
literal \s-1ASCII TAB\s0 characters in strings
.Sp
Literal \s-1ASCII TAB\s0 characters are now allowed in strings (and treated as
\&\f(CW\*(C`\et\*(C'\fR) in relaxed mode. Despite \s-1JSON\s0 mandates, that \s-1TAB\s0 character is
substituted for \*(L"\et\*(R" sequence.
.Sp
.Vb 4
\& [
\& "Hello\etWorld",
\& "Hello<TAB>World", # literal <TAB> would not normally be allowed
\& ]
.Ve
.IP "\(bu" 4
allow_singlequote
.Sp
Single quotes are accepted instead of double quotes. See the
\&\*(L"allow_singlequote\*(R" option.
.Sp
.Vb 3
\& { "foo":\*(Aqbar\*(Aq }
\& { \*(Aqfoo\*(Aq:"bar" }
\& { \*(Aqfoo\*(Aq:\*(Aqbar\*(Aq }
.Ve
.IP "\(bu" 4
allow_barekey
.Sp
Accept unquoted object keys instead of with mandatory double quotes. See the
\&\*(L"allow_barekey\*(R" option.
.Sp
.Vb 1
\& { foo:"bar" }
.Ve
.IP "\(bu" 4
allow_dupkeys
.Sp
Allow decoding of duplicate keys in hashes. By default duplicate keys are forbidden.
See <http://seriot.ch/parsing_json.php#24>:
\&\s-1RFC 7159\s0 section 4: \*(L"The names within an object should be unique.\*(R"
See the \*(L"allow_dupkeys\*(R" option.
.RE
.RS 4
.RE
.ie n .IP "$json = $json\->canonical ([$enable])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->canonical ([$enable])" 4
.IX Item "$json = $json->canonical ([$enable])"
.PD 0
.ie n .IP "$enabled = $json\->get_canonical" 4
.el .IP "\f(CW$enabled\fR = \f(CW$json\fR\->get_canonical" 4
.IX Item "$enabled = $json->get_canonical"
.PD
If \f(CW$enable\fR is true (or missing), then the \f(CW\*(C`encode\*(C'\fR method will
output \s-1JSON\s0 objects by sorting their keys. This is adding a
comparatively high overhead.
.Sp
If \f(CW$enable\fR is false, then the \f(CW\*(C`encode\*(C'\fR method will output key-value
pairs in the order Perl stores them (which will likely change between runs
of the same script, and can change even within the same run from 5.18
onwards).
.Sp
This option is useful if you want the same data structure to be encoded as
the same \s-1JSON\s0 text (given the same overall settings). If it is disabled,
the same hash might be encoded differently even if contains the same data,
as key-value pairs have no inherent ordering in Perl.
.Sp
This setting has no effect when decoding \s-1JSON\s0 texts.
.Sp
This is now also done with tied hashes, contrary to \s-1JSON::XS\s0.
But note that with most large tied hashes stored as tree it is advised to
sort the iterator already and don't sort the hash output here. Most such
iterators are already sorted, as such e.g. DB_File with \f(CW\*(C`DB_BTREE\*(C'\fR.
.ie n .IP "$json = $json\->sort_by (undef, 0, 1 or a block)" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->sort_by (undef, 0, 1 or a block)" 4
.IX Item "$json = $json->sort_by (undef, 0, 1 or a block)"
This currently only (un)sets the \f(CW\*(C`canonical\*(C'\fR option, and ignores
custom sort blocks.
.Sp
This setting has no effect when decoding \s-1JSON\s0 texts.
.Sp
This setting has currently no effect on tied hashes.
.ie n .IP "$json = $json\->escape_slash ([$enable])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->escape_slash ([$enable])" 4
.IX Item "$json = $json->escape_slash ([$enable])"
.PD 0
.ie n .IP "$enabled = $json\->get_escape_slash" 4
.el .IP "\f(CW$enabled\fR = \f(CW$json\fR\->get_escape_slash" 4
.IX Item "$enabled = $json->get_escape_slash"
.PD
According to the \s-1JSON\s0 Grammar, the \fIforward slash\fR character (U+002F)
\&\f(CW"/"\fR need to be escaped. But by default strings are encoded without
escaping slashes in all perl \s-1JSON\s0 encoders.
.Sp
If \f(CW$enable\fR is true (or missing), then \f(CW\*(C`encode\*(C'\fR will escape slashes,
\&\f(CW"\e/"\fR.
.Sp
This setting has no effect when decoding \s-1JSON\s0 texts.
.ie n .IP "$json = $json\->unblessed_bool ([$enable])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->unblessed_bool ([$enable])" 4
.IX Item "$json = $json->unblessed_bool ([$enable])"
.PD 0
.ie n .IP "$enabled = $json\->get_unblessed_bool" 4
.el .IP "\f(CW$enabled\fR = \f(CW$json\fR\->get_unblessed_bool" 4
.IX Item "$enabled = $json->get_unblessed_bool"
.PD
.Vb 1
\& $json = $json\->unblessed_bool([$enable])
.Ve
.Sp
If \f(CW$enable\fR is true (or missing), then \f(CW\*(C`decode\*(C'\fR will return
Perl non-object boolean variables (1 and 0) for \s-1JSON\s0 booleans
(\f(CW\*(C`true\*(C'\fR and \f(CW\*(C`false\*(C'\fR). If \f(CW$enable\fR is false, then \f(CW\*(C`decode\*(C'\fR
will return \f(CW\*(C`JSON::PP::Boolean\*(C'\fR objects for \s-1JSON\s0 booleans.
.ie n .IP "$json = $json\->allow_singlequote ([$enable])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->allow_singlequote ([$enable])" 4
.IX Item "$json = $json->allow_singlequote ([$enable])"
.PD 0
.ie n .IP "$enabled = $json\->get_allow_singlequote" 4
.el .IP "\f(CW$enabled\fR = \f(CW$json\fR\->get_allow_singlequote" 4
.IX Item "$enabled = $json->get_allow_singlequote"
.PD
.Vb 1
\& $json = $json\->allow_singlequote([$enable])
.Ve
.Sp
If \f(CW$enable\fR is true (or missing), then \f(CW\*(C`decode\*(C'\fR will accept
\&\s-1JSON\s0 strings quoted by single quotations that are invalid \s-1JSON\s0
format.
.Sp
.Vb 3
\& $json\->allow_singlequote\->decode({"foo":\*(Aqbar\*(Aq});
\& $json\->allow_singlequote\->decode({\*(Aqfoo\*(Aq:"bar"});
\& $json\->allow_singlequote\->decode({\*(Aqfoo\*(Aq:\*(Aqbar\*(Aq});
.Ve
.Sp
This is also enabled with \f(CW\*(C`relaxed\*(C'\fR.
As same as the \f(CW\*(C`relaxed\*(C'\fR option, this option may be used to parse
application-specific files written by humans.
.ie n .IP "$json = $json\->allow_barekey ([$enable])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->allow_barekey ([$enable])" 4
.IX Item "$json = $json->allow_barekey ([$enable])"
.PD 0
.ie n .IP "$enabled = $json\->get_allow_barekey" 4
.el .IP "\f(CW$enabled\fR = \f(CW$json\fR\->get_allow_barekey" 4
.IX Item "$enabled = $json->get_allow_barekey"
.PD
.Vb 1
\& $json = $json\->allow_barekey([$enable])
.Ve
.Sp
If \f(CW$enable\fR is true (or missing), then \f(CW\*(C`decode\*(C'\fR will accept
bare keys of \s-1JSON\s0 object that are invalid \s-1JSON\s0 format.
.Sp
Same as with the \f(CW\*(C`relaxed\*(C'\fR option, this option may be used to parse
application-specific files written by humans.
.Sp
.Vb 1
\& $json\->allow_barekey\->decode(\*(Aq{foo:"bar"}\*(Aq);
.Ve
.ie n .IP "$json = $json\->allow_bignum ([$enable])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->allow_bignum ([$enable])" 4
.IX Item "$json = $json->allow_bignum ([$enable])"
.PD 0
.ie n .IP "$enabled = $json\->get_allow_bignum" 4
.el .IP "\f(CW$enabled\fR = \f(CW$json\fR\->get_allow_bignum" 4
.IX Item "$enabled = $json->get_allow_bignum"
.PD
.Vb 1
\& $json = $json\->allow_bignum([$enable])
.Ve
.Sp
If \f(CW$enable\fR is true (or missing), then \f(CW\*(C`decode\*(C'\fR will convert
the big integer Perl cannot handle as integer into a Math::BigInt
object and convert a floating number (any) into a Math::BigFloat.
.Sp
On the contrary, \f(CW\*(C`encode\*(C'\fR converts \f(CW\*(C`Math::BigInt\*(C'\fR objects and
\&\f(CW\*(C`Math::BigFloat\*(C'\fR objects into \s-1JSON\s0 numbers with \f(CW\*(C`allow_blessed\*(C'\fR
enable.
.Sp
.Vb 4
\& $json\->allow_nonref\->allow_blessed\->allow_bignum;
\& $bigfloat = $json\->decode(\*(Aq2.000000000000000000000000001\*(Aq);
\& print $json\->encode($bigfloat);
\& # => 2.000000000000000000000000001
.Ve
.Sp
See \*(L"\s-1MAPPING\*(R"\s0 about the normal conversion of \s-1JSON\s0 number.
.ie n .IP "$json = $json\->allow_bigint ([$enable])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->allow_bigint ([$enable])" 4
.IX Item "$json = $json->allow_bigint ([$enable])"
This option is obsolete and replaced by allow_bignum.
.ie n .IP "$json = $json\->allow_nonref ([$enable])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->allow_nonref ([$enable])" 4
.IX Item "$json = $json->allow_nonref ([$enable])"
.PD 0
.ie n .IP "$enabled = $json\->get_allow_nonref" 4
.el .IP "\f(CW$enabled\fR = \f(CW$json\fR\->get_allow_nonref" 4
.IX Item "$enabled = $json->get_allow_nonref"
.PD
If \f(CW$enable\fR is true (or missing), then the \f(CW\*(C`encode\*(C'\fR method can
convert a non-reference into its corresponding string, number or null
\&\s-1JSON\s0 value, which is an extension to \s-1RFC4627.\s0 Likewise, \f(CW\*(C`decode\*(C'\fR will
accept those \s-1JSON\s0 values instead of croaking.
.Sp
If \f(CW$enable\fR is false, then the \f(CW\*(C`encode\*(C'\fR method will croak if it isn't
passed an arrayref or hashref, as \s-1JSON\s0 texts must either be an object
or array. Likewise, \f(CW\*(C`decode\*(C'\fR will croak if given something that is not a
\&\s-1JSON\s0 object or array.
.Sp
Example, encode a Perl scalar as \s-1JSON\s0 value with enabled \f(CW\*(C`allow_nonref\*(C'\fR,
resulting in an invalid \s-1JSON\s0 text:
.Sp
.Vb 2
\& Cpanel::JSON::XS\->new\->allow_nonref\->encode ("Hello, World!")
\& => "Hello, World!"
.Ve
.ie n .IP "$json = $json\->allow_unknown ([$enable])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->allow_unknown ([$enable])" 4
.IX Item "$json = $json->allow_unknown ([$enable])"
.PD 0
.ie n .IP "$enabled = $json\->get_allow_unknown" 4
.el .IP "\f(CW$enabled\fR = \f(CW$json\fR\->get_allow_unknown" 4
.IX Item "$enabled = $json->get_allow_unknown"
.PD
If \f(CW$enable\fR is true (or missing), then \f(CW\*(C`encode\*(C'\fR will \fInot\fR throw an
exception when it encounters values it cannot represent in \s-1JSON\s0 (for
example, filehandles) but instead will encode a \s-1JSON\s0 \f(CW\*(C`null\*(C'\fR value. Note
that blessed objects are not included here and are handled separately by
c<allow_nonref>.
.Sp
If \f(CW$enable\fR is false (the default), then \f(CW\*(C`encode\*(C'\fR will throw an
exception when it encounters anything it cannot encode as \s-1JSON.\s0
.Sp
This option does not affect \f(CW\*(C`decode\*(C'\fR in any way, and it is recommended to
leave it off unless you know your communications partner.
.ie n .IP "$json = $json\->allow_stringify ([$enable])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->allow_stringify ([$enable])" 4
.IX Item "$json = $json->allow_stringify ([$enable])"
.PD 0
.ie n .IP "$enabled = $json\->get_allow_stringify" 4
.el .IP "\f(CW$enabled\fR = \f(CW$json\fR\->get_allow_stringify" 4
.IX Item "$enabled = $json->get_allow_stringify"
.PD
If \f(CW$enable\fR is true (or missing), then \f(CW\*(C`encode\*(C'\fR will stringify the
non-object perl value or reference. Note that blessed objects are not
included here and are handled separately by \f(CW\*(C`allow_blessed\*(C'\fR and
\&\f(CW\*(C`convert_blessed\*(C'\fR. String references are stringified to the string
value, other references as in perl.
.Sp
This option does not affect \f(CW\*(C`decode\*(C'\fR in any way.
.Sp
This option is special to this module, it is not supported by other
encoders. So it is not recommended to use it.
.ie n .IP "$json = $json\->require_types ([$enable])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->require_types ([$enable])" 4
.IX Item "$json = $json->require_types ([$enable])"
.PD 0
.ie n .IP "$enable = $json\->get_require_types" 4
.el .IP "\f(CW$enable\fR = \f(CW$json\fR\->get_require_types" 4
.IX Item "$enable = $json->get_require_types"
.PD
.Vb 1
\& $json = $json\->require_types([$enable])
.Ve
.Sp
If \f(CW$enable\fR is true (or missing), then \f(CW\*(C`encode\*(C'\fR will require
either enabled \f(CW\*(C`type_all_string\*(C'\fR or second argument with supplied \s-1JSON\s0 types.
See Cpanel::JSON::XS::Type. When \f(CW\*(C`type_all_string\*(C'\fR is not enabled or
second argument is not provided (or is undef), then \f(CW\*(C`encode\*(C'\fR
croaks. It also croaks when the type for provided structure in
\&\f(CW\*(C`encode\*(C'\fR is incomplete.
.ie n .IP "$json = $json\->type_all_string ([$enable])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->type_all_string ([$enable])" 4
.IX Item "$json = $json->type_all_string ([$enable])"
.PD 0
.ie n .IP "$enable = $json\->get_type_all_string" 4
.el .IP "\f(CW$enable\fR = \f(CW$json\fR\->get_type_all_string" 4
.IX Item "$enable = $json->get_type_all_string"
.PD
.Vb 1
\& $json = $json\->type_all_string([$enable])
.Ve
.Sp
If \f(CW$enable\fR is true (or missing), then \f(CW\*(C`encode\*(C'\fR will always
produce stable deterministic \s-1JSON\s0 string types in resulted output.
.Sp
When \f(CW$enable\fR is false, then result of encoded \s-1JSON\s0 output may be
different for different Perl versions and may depends on loaded modules.
.Sp
This is useful it you need deterministic \s-1JSON\s0 types, independently of used
Perl version and other modules, but do not want to write complicated type
definitions for Cpanel::JSON::XS::Type.
.ie n .IP "$json = $json\->allow_dupkeys ([$enable])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->allow_dupkeys ([$enable])" 4
.IX Item "$json = $json->allow_dupkeys ([$enable])"
.PD 0
.ie n .IP "$enabled = $json\->get_allow_dupkeys" 4
.el .IP "\f(CW$enabled\fR = \f(CW$json\fR\->get_allow_dupkeys" 4
.IX Item "$enabled = $json->get_allow_dupkeys"
.PD
If \f(CW$enable\fR is true (or missing), then the \f(CW\*(C`decode\*(C'\fR method will not
die when it encounters duplicate keys in a hash.
\&\f(CW\*(C`allow_dupkeys\*(C'\fR is also enabled in the \f(CW\*(C`relaxed\*(C'\fR mode.
.Sp
The \s-1JSON\s0 spec allows duplicate name in objects but recommends to
disable it, however with Perl hashes they are impossible, parsing
\&\s-1JSON\s0 in Perl silently ignores duplicate names, using the last value
found.
.Sp
See <http://seriot.ch/parsing_json.php#24>:
\&\s-1RFC 7159\s0 section 4: \*(L"The names within an object should be unique.\*(R"
.ie n .IP "$json = $json\->allow_blessed ([$enable])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->allow_blessed ([$enable])" 4
.IX Item "$json = $json->allow_blessed ([$enable])"
.PD 0
.ie n .IP "$enabled = $json\->get_allow_blessed" 4
.el .IP "\f(CW$enabled\fR = \f(CW$json\fR\->get_allow_blessed" 4
.IX Item "$enabled = $json->get_allow_blessed"
.PD
If \f(CW$enable\fR is true (or missing), then the \f(CW\*(C`encode\*(C'\fR method will not
barf when it encounters a blessed reference. Instead, the value of the
\&\fBconvert_blessed\fR option will decide whether \f(CW\*(C`null\*(C'\fR (\f(CW\*(C`convert_blessed\*(C'\fR
disabled or no \f(CW\*(C`TO_JSON\*(C'\fR method found) or a representation of the
object (\f(CW\*(C`convert_blessed\*(C'\fR enabled and \f(CW\*(C`TO_JSON\*(C'\fR method found) is being
encoded. Has no effect on \f(CW\*(C`decode\*(C'\fR.
.Sp
If \f(CW$enable\fR is false (the default), then \f(CW\*(C`encode\*(C'\fR will throw an
exception when it encounters a blessed object without \f(CW\*(C`convert_blessed\*(C'\fR
and a \f(CW\*(C`TO_JSON\*(C'\fR method.
.Sp
This setting has no effect on \f(CW\*(C`decode\*(C'\fR.
.ie n .IP "$json = $json\->convert_blessed ([$enable])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->convert_blessed ([$enable])" 4
.IX Item "$json = $json->convert_blessed ([$enable])"
.PD 0
.ie n .IP "$enabled = $json\->get_convert_blessed" 4
.el .IP "\f(CW$enabled\fR = \f(CW$json\fR\->get_convert_blessed" 4
.IX Item "$enabled = $json->get_convert_blessed"
.PD
If \f(CW$enable\fR is true (or missing), then \f(CW\*(C`encode\*(C'\fR, upon encountering a
blessed object, will check for the availability of the \f(CW\*(C`TO_JSON\*(C'\fR 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
\&\f(CW\*(C`TO_JSON\*(C'\fR method is found, a stringification overload method is tried next.
If both are not found, the value of \f(CW\*(C`allow_blessed\*(C'\fR will decide what
to do.
.Sp
The \f(CW\*(C`TO_JSON\*(C'\fR method may safely call die if it wants. If \f(CW\*(C`TO_JSON\*(C'\fR
returns other blessed objects, those will be handled in the same
way. \f(CW\*(C`TO_JSON\*(C'\fR must take care of not causing an endless recursion
cycle (== crash) in this case. The same care must be taken with
calling encode in stringify overloads (even if this works by luck in
older perls) or other callbacks. The name of \f(CW\*(C`TO_JSON\*(C'\fR was chosen
because other methods called by the Perl core (== not by the user of
the object) are usually in upper case letters and to avoid collisions
with any \f(CW\*(C`to_json\*(C'\fR function or method.
.Sp
If \f(CW$enable\fR is false (the default), then \f(CW\*(C`encode\*(C'\fR will not consider
this type of conversion.
.Sp
This setting has no effect on \f(CW\*(C`decode\*(C'\fR.
.ie n .IP "$json = $json\->allow_tags ([$enable])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->allow_tags ([$enable])" 4
.IX Item "$json = $json->allow_tags ([$enable])"
.PD 0
.ie n .IP "$enabled = $json\->get_allow_tags" 4
.el .IP "\f(CW$enabled\fR = \f(CW$json\fR\->get_allow_tags" 4
.IX Item "$enabled = $json->get_allow_tags"
.PD
See \*(L"\s-1OBJECT SERIALIZATION\*(R"\s0 for details.
.Sp
If \f(CW$enable\fR is true (or missing), then \f(CW\*(C`encode\*(C'\fR, upon encountering a
blessed object, will check for the availability of the \f(CW\*(C`FREEZE\*(C'\fR method on
the object's class. If found, it will be used to serialize the object into
a nonstandard tagged \s-1JSON\s0 value (that \s-1JSON\s0 decoders cannot decode).
.Sp
It also causes \f(CW\*(C`decode\*(C'\fR to parse such tagged \s-1JSON\s0 values and deserialize
them via a call to the \f(CW\*(C`THAW\*(C'\fR method.
.Sp
If \f(CW$enable\fR is false (the default), then \f(CW\*(C`encode\*(C'\fR will not consider
this type of conversion, and tagged \s-1JSON\s0 values will cause a parse error
in \f(CW\*(C`decode\*(C'\fR, as if tags were not part of the grammar.
.ie n .IP "$json = $json\->filter_json_object ([$coderef\->($hashref)])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->filter_json_object ([$coderef\->($hashref)])" 4
.IX Item "$json = $json->filter_json_object ([$coderef->($hashref)])"
When \f(CW$coderef\fR is specified, it will be called from \f(CW\*(C`decode\*(C'\fR each
time it decodes a \s-1JSON\s0 object. The only argument is a reference to the
newly-created hash. If the code references returns a single scalar (which
need not be a reference), this value (i.e. a copy of that scalar to avoid
aliasing) is inserted into the deserialized data structure. If it returns
an empty list (\s-1NOTE:\s0 \fInot\fR \f(CW\*(C`undef\*(C'\fR, which is a valid scalar), the
original deserialized hash will be inserted. This setting can slow down
decoding considerably.
.Sp
When \f(CW$coderef\fR is omitted or undefined, any existing callback will
be removed and \f(CW\*(C`decode\*(C'\fR will not change the deserialized hash in any
way.
.Sp
Example, convert all \s-1JSON\s0 objects into the integer 5:
.Sp
.Vb 6
\& my $js = Cpanel::JSON::XS\->new\->filter_json_object (sub { 5 });
\& # returns [5]
\& $js\->decode (\*(Aq[{}]\*(Aq)
\& # throw an exception because allow_nonref is not enabled
\& # so a lone 5 is not allowed.
\& $js\->decode (\*(Aq{"a":1, "b":2}\*(Aq);
.Ve
.ie n .IP "$json = $json\->filter_json_single_key_object ($key [=> $coderef\->($value)])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->filter_json_single_key_object ($key [=> \f(CW$coderef\fR\->($value)])" 4
.IX Item "$json = $json->filter_json_single_key_object ($key [=> $coderef->($value)])"
Works remotely similar to \f(CW\*(C`filter_json_object\*(C'\fR, but is only called for
\&\s-1JSON\s0 objects having a single key named \f(CW$key\fR.
.Sp
This \f(CW$coderef\fR is called before the one specified via
\&\f(CW\*(C`filter_json_object\*(C'\fR, if any. It gets passed the single value in the \s-1JSON\s0
object. If it returns a single value, it will be inserted into the data
structure. If it returns nothing (not even \f(CW\*(C`undef\*(C'\fR but the empty list),
the callback from \f(CW\*(C`filter_json_object\*(C'\fR will be called next, as if no
single-key callback were specified.
.Sp
If \f(CW$coderef\fR is omitted or undefined, the corresponding callback will be
disabled. There can only ever be one callback for a given key.
.Sp
As this callback gets called less often then the \f(CW\*(C`filter_json_object\*(C'\fR
one, decoding speed will not usually suffer as much. Therefore, single-key
objects make excellent targets to serialize Perl objects into, especially
as single-key \s-1JSON\s0 objects are as close to the type-tagged value concept
as \s-1JSON\s0 gets (it's basically an \s-1ID/VALUE\s0 tuple). Of course, \s-1JSON\s0 does not
support this in any way, so you need to make sure your data never looks
like a serialized Perl hash.
.Sp
Typical names for the single object key are \f(CW\*(C`_\|_class_whatever_\|_\*(C'\fR, or
\&\f(CW\*(C`$_\|_dollars_are_rarely_used_\|_$\*(C'\fR or \f(CW\*(C`}ugly_brace_placement\*(C'\fR, or even
things like \f(CW\*(C`_\|_class_md5sum(classname)_\|_\*(C'\fR, to reduce the risk of clashing
with real hashes.
.Sp
Example, decode \s-1JSON\s0 objects of the form \f(CW\*(C`{ "_\|_widget_\|_" => <id> }\*(C'\fR
into the corresponding \f(CW$WIDGET{<id>}\fR object:
.Sp
.Vb 7
\& # return whatever is in $WIDGET{5}:
\& Cpanel::JSON::XS
\& \->new
\& \->filter_json_single_key_object (_\|_widget_\|_ => sub {
\& $WIDGET{ $_[0] }
\& })
\& \->decode (\*(Aq{"_\|_widget_\|_": 5\*(Aq)
\&
\& # this can be used with a TO_JSON method in some "widget" class
\& # for serialization to json:
\& sub WidgetBase::TO_JSON {
\& my ($self) = @_;
\&
\& unless ($self\->{id}) {
\& $self\->{id} = ..get..some..id..;
\& $WIDGET{$self\->{id}} = $self;
\& }
\&
\& { _\|_widget_\|_ => $self\->{id} }
\& }
.Ve
.ie n .IP "$json = $json\->shrink ([$enable])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->shrink ([$enable])" 4
.IX Item "$json = $json->shrink ([$enable])"
.PD 0
.ie n .IP "$enabled = $json\->get_shrink" 4
.el .IP "\f(CW$enabled\fR = \f(CW$json\fR\->get_shrink" 4
.IX Item "$enabled = $json->get_shrink"
.PD
Perl usually over-allocates memory a bit when allocating space for
strings. This flag optionally resizes strings generated by either
\&\f(CW\*(C`encode\*(C'\fR or \f(CW\*(C`decode\*(C'\fR to their minimum size possible. This can save
memory when your \s-1JSON\s0 texts are either very very long or you have many
short strings. It will also try to downgrade any strings to octet-form
if possible: perl stores strings internally either in an encoding called
UTF-X or in octet-form. The latter cannot store everything but uses less
space in general (and some buggy Perl or C code might even rely on that
internal representation being used).
.Sp
The actual definition of what shrink does might change in future versions,
but it will always try to save space at the expense of time.
.Sp
If \f(CW$enable\fR is true (or missing), the string returned by \f(CW\*(C`encode\*(C'\fR will
be shrunk-to-fit, while all strings generated by \f(CW\*(C`decode\*(C'\fR will also be
shrunk-to-fit.
.Sp
If \f(CW$enable\fR is false, then the normal perl allocation algorithms are used.
If you work with your data, then this is likely to be faster.
.Sp
In the future, this setting might control other things, such as converting
strings that look like integers or floats into integers or floats
internally (there is no difference on the Perl level), saving space.
.ie n .IP "$json = $json\->max_depth ([$maximum_nesting_depth])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->max_depth ([$maximum_nesting_depth])" 4
.IX Item "$json = $json->max_depth ([$maximum_nesting_depth])"
.PD 0
.ie n .IP "$max_depth = $json\->get_max_depth" 4
.el .IP "\f(CW$max_depth\fR = \f(CW$json\fR\->get_max_depth" 4
.IX Item "$max_depth = $json->get_max_depth"
.PD
Sets the maximum nesting level (default \f(CW512\fR) accepted while encoding
or decoding. If a higher nesting level is detected in \s-1JSON\s0 text or a Perl
data structure, then the encoder and decoder will stop and croak at that
point.
.Sp
Nesting level is defined by number of hash\- or arrayrefs that the encoder
needs to traverse to reach a given point or the number of \f(CW\*(C`{\*(C'\fR or \f(CW\*(C`[\*(C'\fR
characters without their matching closing parenthesis crossed to reach a
given character in a string.
.Sp
Setting the maximum depth to one disallows any nesting, so that ensures
that the object is only a single hash/object or array.
.Sp
If no argument is given, the highest possible setting will be used, which
is rarely useful.
.Sp
Note that nesting is implemented by recursion in C. The default value has
been chosen to be as large as typical operating systems allow without
crashing.
.Sp
See \*(L"\s-1SECURITY CONSIDERATIONS\*(R"\s0, below, for more info on why this is useful.
.ie n .IP "$json = $json\->max_size ([$maximum_string_size])" 4
.el .IP "\f(CW$json\fR = \f(CW$json\fR\->max_size ([$maximum_string_size])" 4
.IX Item "$json = $json->max_size ([$maximum_string_size])"
.PD 0
.ie n .IP "$max_size = $json\->get_max_size" 4
.el .IP "\f(CW$max_size\fR = \f(CW$json\fR\->get_max_size" 4
.IX Item "$max_size = $json->get_max_size"
.PD
Set the maximum length a \s-1JSON\s0 text may have (in bytes) where decoding is
being attempted. The default is \f(CW0\fR, meaning no limit. When \f(CW\*(C`decode\*(C'\fR
is called on a string that is longer then this many bytes, it will not
attempt to decode the string but throw an exception. This setting has no
effect on \f(CW\*(C`encode\*(C'\fR (yet).
.Sp
If no argument is given, the limit check will be deactivated (same as when
\&\f(CW0\fR is specified).
.Sp
See \*(L"\s-1SECURITY CONSIDERATIONS\*(R"\s0, below, for more info on why this is useful.
.ie n .IP "$json\->stringify_infnan ([$infnan_mode = 1])" 4
.el .IP "\f(CW$json\fR\->stringify_infnan ([$infnan_mode = 1])" 4
.IX Item "$json->stringify_infnan ([$infnan_mode = 1])"
.PD 0
.ie n .IP "$infnan_mode = $json\->get_stringify_infnan" 4
.el .IP "\f(CW$infnan_mode\fR = \f(CW$json\fR\->get_stringify_infnan" 4
.IX Item "$infnan_mode = $json->get_stringify_infnan"
.PD
Get or set how Cpanel::JSON::XS encodes \f(CW\*(C`inf\*(C'\fR, \f(CW\*(C`\-inf\*(C'\fR or \f(CW\*(C`nan\*(C'\fR for numeric
values. Also qnan, snan or negative nan on some platforms.
.Sp
\&\f(CW\*(C`null\*(C'\fR: infnan_mode = 0. Similar to most \s-1JSON\s0 modules in other languages.
Always null.
.Sp
stringified: infnan_mode = 1. As in Mojo::JSON. Platform specific strings.
Stringified via sprintf(%g), with double quotes.
.Sp
inf/nan: infnan_mode = 2. As in \s-1JSON::XS,\s0 and older releases.
Passes through platform dependent values, invalid \s-1JSON.\s0 Stringified via
sprintf(%g), but without double quotes.
.Sp
\&\*(L"inf/\-inf/nan\*(R": infnan_mode = 3. Platform independent inf/nan/\-inf
strings. No QNAN/SNAN/negative \s-1NAN\s0 support, unified to \*(L"nan\*(R". Much
easier to detect, but may conflict with valid strings.
.ie n .IP "$json_text = $json\->encode ($perl_scalar, $json_type)" 4
.el .IP "\f(CW$json_text\fR = \f(CW$json\fR\->encode ($perl_scalar, \f(CW$json_type\fR)" 4
.IX Item "$json_text = $json->encode ($perl_scalar, $json_type)"
Converts the given Perl data structure (a simple scalar or a reference
to a hash or array) to its \s-1JSON\s0 representation. Simple scalars will be
converted into \s-1JSON\s0 string or number sequences, while references to
arrays become \s-1JSON\s0 arrays and references to hashes become \s-1JSON\s0
objects. Undefined Perl values (e.g. \f(CW\*(C`undef\*(C'\fR) become \s-1JSON\s0 \f(CW\*(C`null\*(C'\fR
values. Neither \f(CW\*(C`true\*(C'\fR nor \f(CW\*(C`false\*(C'\fR values will be generated.
.Sp
For the type argument see Cpanel::JSON::XS::Type.
.ie n .IP "$perl_scalar = $json\->decode ($json_text, my $json_type)" 4
.el .IP "\f(CW$perl_scalar\fR = \f(CW$json\fR\->decode ($json_text, my \f(CW$json_type\fR)" 4
.IX Item "$perl_scalar = $json->decode ($json_text, my $json_type)"
The opposite of \f(CW\*(C`encode\*(C'\fR: expects a \s-1JSON\s0 text and tries to parse it,
returning the resulting simple scalar or reference. Croaks on error.
.Sp
\&\s-1JSON\s0 numbers and strings become simple Perl scalars. \s-1JSON\s0 arrays become
Perl arrayrefs and \s-1JSON\s0 objects become Perl hashrefs. \f(CW\*(C`true\*(C'\fR becomes
\&\f(CW1\fR, \f(CW\*(C`false\*(C'\fR becomes \f(CW0\fR and \f(CW\*(C`null\*(C'\fR becomes \f(CW\*(C`undef\*(C'\fR.
.Sp
For the type argument see Cpanel::JSON::XS::Type.
.ie n .IP "($perl_scalar, $characters) = $json\->decode_prefix ($json_text)" 4
.el .IP "($perl_scalar, \f(CW$characters\fR) = \f(CW$json\fR\->decode_prefix ($json_text)" 4
.IX Item "($perl_scalar, $characters) = $json->decode_prefix ($json_text)"
This works like the \f(CW\*(C`decode\*(C'\fR method, but instead of raising an exception
when there is trailing garbage after the first \s-1JSON\s0 object, it will
silently stop parsing there and return the number of characters consumed
so far.
.Sp
This is useful if your \s-1JSON\s0 texts are not delimited by an outer protocol
and you need to know where the \s-1JSON\s0 text ends.
.Sp
.Vb 2
\& Cpanel::JSON::XS\->new\->decode_prefix ("[1] the tail")
\& => ([1], 3)
.Ve
.ie n .IP "$json\->to_json ($perl_hash_or_arrayref)" 4
.el .IP "\f(CW$json\fR\->to_json ($perl_hash_or_arrayref)" 4
.IX Item "$json->to_json ($perl_hash_or_arrayref)"
Deprecated method for perl 5.8 and newer. Use encode_json instead.
.ie n .IP "$json\->from_json ($utf8_encoded_json_text)" 4
.el .IP "\f(CW$json\fR\->from_json ($utf8_encoded_json_text)" 4
.IX Item "$json->from_json ($utf8_encoded_json_text)"
Deprecated method for perl 5.8 and newer. Use decode_json instead.
.SH "INCREMENTAL PARSING"
.IX Header "INCREMENTAL PARSING"
In some cases, there is the need for incremental parsing of \s-1JSON\s0
texts. While this module always has to keep both \s-1JSON\s0 text and resulting
Perl data structure in memory at one time, it does allow you to parse a
\&\s-1JSON\s0 stream incrementally. It does so by accumulating text until it has
a full \s-1JSON\s0 object, which it then can decode. This process is similar to
using \f(CW\*(C`decode_prefix\*(C'\fR to see if a full \s-1JSON\s0 object is available, but
is much more efficient (and can be implemented with a minimum of method
calls).
.PP
Cpanel::JSON::XS will only attempt to parse the \s-1JSON\s0 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 \s-1JSON\s0 text has been seen. This means
you need to set resource limits (e.g. \f(CW\*(C`max_size\*(C'\fR) to ensure the
parser will stop parsing in the presence if syntax errors.
.PP
The following methods implement this incremental parser.
.ie n .IP "[void, scalar or list context] = $json\->incr_parse ([$string])" 4
.el .IP "[void, scalar or list context] = \f(CW$json\fR\->incr_parse ([$string])" 4
.IX 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).
.Sp
If \f(CW$string\fR is given, then this string is appended to the already
existing \s-1JSON\s0 fragment stored in the \f(CW$json\fR object.
.Sp
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.
.Sp
If the method is called in scalar context, then it will try to extract
exactly \fIone\fR \s-1JSON\s0 object. If that is successful, it will return this
object, otherwise it will return \f(CW\*(C`undef\*(C'\fR. If there is a parse error,
this method will croak just as \f(CW\*(C`decode\*(C'\fR would do (one can then use
\&\f(CW\*(C`incr_skip\*(C'\fR to skip the erroneous part). This is the most common way of
using the method.
.Sp
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 \s-1JSON\s0
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 \s-1JSON\s0 texts will be
lost.
.Sp
Example: Parse some \s-1JSON\s0 arrays/objects in a given string and return
them.
.Sp
.Vb 1
\& my @objs = Cpanel::JSON::XS\->new\->incr_parse ("[5][7][1,2]");
.Ve
.ie n .IP "$lvalue_string = $json\->incr_text (>5.8 only)" 4
.el .IP "\f(CW$lvalue_string\fR = \f(CW$json\fR\->incr_text (>5.8 only)" 4
.IX Item "$lvalue_string = $json->incr_text (>5.8 only)"
This method returns the currently stored \s-1JSON\s0 fragment as an lvalue, that
is, you can manipulate it. This \fIonly\fR works when a preceding call to
\&\f(CW\*(C`incr_parse\*(C'\fR in \fIscalar context\fR successfully returned an object, and
2. only with Perl >= 5.8
.Sp
Under all other circumstances you must not call this function (I mean
it. although in simple tests it might actually work, it \fIwill\fR fail
under real world conditions). As a special exception, you can also
call this method before having parsed anything.
.Sp
This function is useful in two cases: a) finding the trailing text after a
\&\s-1JSON\s0 object or b) parsing multiple \s-1JSON\s0 objects separated by non-JSON text
(such as commas).
.ie n .IP "$json\->incr_skip" 4
.el .IP "\f(CW$json\fR\->incr_skip" 4
.IX Item "$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
\&\f(CW\*(C`incr_parse\*(C'\fR 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.
.Sp
The difference to \f(CW\*(C`incr_reset\*(C'\fR is that only text until the parse error
occurred is removed.
.ie n .IP "$json\->incr_reset" 4
.el .IP "\f(CW$json\fR\->incr_reset" 4
.IX 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.
.Sp
This is useful if you want to repeatedly parse \s-1JSON\s0 objects and want to
ignore any trailing data, which means you have to reset the parser after
each successful decode.
.SS "\s-1LIMITATIONS\s0"
.IX Subsection "LIMITATIONS"
All options that affect decoding are supported, except
\&\f(CW\*(C`allow_nonref\*(C'\fR. The reason for this is that it cannot be made to work
sensibly: \s-1JSON\s0 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 \s-1JSON\s0 numbers, however.
.PP
For example, is the string \f(CW1\fR a single \s-1JSON\s0 number, or is it simply
the start of \f(CW12\fR? Or is \f(CW12\fR a single \s-1JSON\s0 number, or the
concatenation of \f(CW1\fR and \f(CW2\fR? In neither case you can tell, and this
is why Cpanel::JSON::XS takes the conservative route and disallows
this case.
.SS "\s-1EXAMPLES\s0"
.IX Subsection "EXAMPLES"
Some examples will make all this clearer. First, a simple example that
works similarly to \f(CW\*(C`decode_prefix\*(C'\fR: We want to decode the \s-1JSON\s0 object at
the start of a string and identify the portion after the \s-1JSON\s0 object:
.PP
.Vb 1
\& my $text = "[1,2,3] hello";
\&
\& my $json = new Cpanel::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"
.Ve
.PP
Easy, isn't it?
.PP
Now for a more complicated example: Imagine a hypothetical protocol where
you read some requests from a \s-1TCP\s0 stream, and each request is a \s-1JSON\s0
array, without any separation between them (in fact, it is often useful to
use newlines as \*(L"separators\*(R", as these get interpreted as whitespace at
the start of the \s-1JSON\s0 text, which makes it possible to test said protocol
with \f(CW\*(C`telnet\*(C'\fR...).
.PP
Here is how you'd do it (it is trivial to write this in an event-based
manner):
.PP
.Vb 1
\& my $json = new Cpanel::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
\& }
\& }
.Ve
.PP
Another complicated example: Assume you have a string with \s-1JSON\s0 objects
or arrays, all separated by (optional) comma characters (e.g. \f(CW\*(C`[1],[2],
[3]\*(C'\fR). To parse them, we have to skip the commas between the \s-1JSON\s0 texts,
and here is where the lvalue-ness of \f(CW\*(C`incr_text\*(C'\fR comes in useful:
.PP
.Vb 2
\& my $text = "[1],[2], [3]";
\& my $json = new Cpanel::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/^ \es* , //x;
\& }
.Ve
.PP
Now lets go for a very complex example: Assume that you have a gigantic
\&\s-1JSON\s0 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 :).
.PP
Well, you lost, you have to implement your own \s-1JSON\s0 parser. But
Cpanel::JSON::XS can still help you: You implement a (very simple)
array parser and let \s-1JSON\s0 decode the array elements, which are all
full \s-1JSON\s0 objects on their own (this wouldn't work if the array
elements could be \s-1JSON\s0 numbers, for example):
.PP
.Vb 1
\& my $json = new Cpanel::JSON::XS;
\&
\& # open the monster
\& open my $fh, "<bigfile.json"
\& or die "bigfile: $!";
\&
\& # first parse the initial "["
\& for (;;) {
\& sysread $fh, my $buf, 65536
\& or die "read error: $!";
\& $json\->incr_parse ($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/^ \es* \e[ //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
\& $json\->incr_text =~ s/^\es*//;
\&
\& # if we find "]", we are done
\& if ($json\->incr_text =~ s/^\e]//) {
\& print "finished.\en";
\& 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
\& }
.Ve
.PP
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 :).
.SH "BOM"
.IX Header "BOM"
Detect all unicode \fBByte Order Marks\fR on decode.
Which are \s-1UTF\-8, UTF\-16LE, UTF\-16BE, UTF\-32LE\s0 and \s-1UTF\-32BE.\s0
.PP
The \s-1BOM\s0 encoding is set only for one specific decode call, it does not
change the state of the \s-1JSON\s0 object.
.PP
\&\fBWarning\fR: With perls older than 5.20 you need load the Encode module
before loading a multibyte \s-1BOM,\s0 i.e. >= \s-1UTF\-16.\s0 Otherwise an error is
thrown. This is an implementation limitation and might get fixed later.
.PP
See <https://tools.ietf.org/html/rfc7159#section\-8.1>
\&\fI\*(L"\s-1JSON\s0 text \s-1SHALL\s0 be encoded in \s-1UTF\-8, UTF\-16,\s0 or \s-1UTF\-32.\*(R"\s0\fR
.PP
\&\fI\*(L"Implementations \s-1MUST NOT\s0 add a byte order mark to the beginning of a
\&\s-1JSON\s0 text\*(R", \*(L"implementations (...) \s-1MAY\s0 ignore the presence of a byte
order mark rather than treating it as an error\*(R".\fR
.PP
See also <http://www.unicode.org/faq/utf_bom.html#BOM>.
.PP
Beware that Cpanel::JSON::XS is currently the only \s-1JSON\s0 module which
does accept and decode a \s-1BOM.\s0
.PP
The latest \s-1JSON\s0 spec
<https://www.greenbytes.de/tech/webdav/rfc8259.html#character.encoding>
forbid the usage of \s-1UTF\-16\s0 or \s-1UTF\-32,\s0 the character encoding is \s-1UTF\-8.\s0
Thus in subsequent updates \s-1BOM\s0's of \s-1UTF\-16\s0 or \s-1UTF\-32\s0 will throw an error.
.SH "MAPPING"
.IX Header "MAPPING"
This section describes how Cpanel::JSON::XS maps Perl values to \s-1JSON\s0
values and vice versa. These mappings are designed to \*(L"do the right
thing\*(R" in most circumstances automatically, preserving round-tripping
characteristics (what you put in comes out as something equivalent).
.PP
For the more enlightened: note that in the following descriptions,
lowercase \fIperl\fR refers to the Perl interpreter, while uppercase \fIPerl\fR
refers to the abstract Perl language itself.
.SS "\s-1JSON\s0 \-> \s-1PERL\s0"
.IX Subsection "JSON -> PERL"
.IP "object" 4
.IX Item "object"
A \s-1JSON\s0 object becomes a reference to a hash in Perl. No ordering of object
keys is preserved (\s-1JSON\s0 does not preserve object key ordering itself).
.IP "array" 4
.IX Item "array"
A \s-1JSON\s0 array becomes a reference to an array in Perl.
.IP "string" 4
.IX Item "string"
A \s-1JSON\s0 string becomes a string scalar in Perl \- Unicode codepoints in \s-1JSON\s0
are represented by the same codepoints in the Perl string, so no manual
decoding is necessary.
.IP "number" 4
.IX Item "number"
A \s-1JSON\s0 number becomes either an integer, numeric (floating point) or
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.
.Sp
If the number consists of digits only, Cpanel::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 (in which case you lose roundtripping ability, as the
\&\s-1JSON\s0 number will be re-encoded to a \s-1JSON\s0 string).
.Sp
Numbers containing a fractional or exponential part will always be
represented as numeric (floating point) values, possibly at a loss of
precision (in which case you might lose perfect roundtripping ability, but
the \s-1JSON\s0 number will still be re-encoded as a \s-1JSON\s0 number).
.Sp
Note that precision is not accuracy \- binary floating point values
cannot represent most decimal fractions exactly, and when converting
from and to floating point, \f(CW\*(C`Cpanel::JSON::XS\*(C'\fR only guarantees precision
up to but not including the least significant bit.
.IP "true, false" 4
.IX Item "true, false"
When \f(CW\*(C`unblessed_bool\*(C'\fR is set to true, then \s-1JSON\s0 \f(CW\*(C`true\*(C'\fR becomes \f(CW1\fR and
\&\s-1JSON\s0 \f(CW\*(C`false\*(C'\fR becomes \f(CW0\fR.
.Sp
Otherwise these \s-1JSON\s0 atoms become \f(CW\*(C`JSON::PP::true\*(C'\fR and
\&\f(CW\*(C`JSON::PP::false\*(C'\fR, respectively. They are \f(CW\*(C`JSON::PP::Boolean\*(C'\fR
objects and are overloaded to act almost exactly like the numbers \f(CW1\fR
and \f(CW0\fR. You can check whether a scalar is a \s-1JSON\s0 boolean by using
the \f(CW\*(C`Cpanel::JSON::XS::is_bool\*(C'\fR function.
.Sp
The other round, from perl to \s-1JSON,\s0 \f(CW\*(C`!0\*(C'\fR which is represented as
\&\f(CW\*(C`yes\*(C'\fR becomes \f(CW\*(C`true\*(C'\fR, and \f(CW\*(C`!1\*(C'\fR which is represented as
\&\f(CW\*(C`no\*(C'\fR becomes \f(CW\*(C`false\*(C'\fR.
.Sp
Via Cpanel::JSON::XS::Type you can now even force negation in \f(CW\*(C`encode\*(C'\fR,
without overloading of \f(CW\*(C`!\*(C'\fR:
.Sp
.Vb 3
\& my $false = Cpanel::JSON::XS::false;
\& print($json\->encode([!$false], [JSON_TYPE_BOOL]));
\& => [true]
.Ve
.IP "null" 4
.IX Item "null"
A \s-1JSON\s0 null atom becomes \f(CW\*(C`undef\*(C'\fR in Perl.
.ie n .IP "shell-style comments (""# \fItext\fP"")" 4
.el .IP "shell-style comments (\f(CW# \f(CItext\f(CW\fR)" 4
.IX Item "shell-style comments (# text)"
As a nonstandard extension to the \s-1JSON\s0 syntax that is enabled by the
\&\f(CW\*(C`relaxed\*(C'\fR setting, shell-style comments are allowed. They can start
anywhere outside strings and go till the end of the line.
.ie n .IP "tagged values (""(\fItag\fP)\fIvalue\fP"")." 4
.el .IP "tagged values (\f(CW(\f(CItag\f(CW)\f(CIvalue\f(CW\fR)." 4
.IX Item "tagged values ((tag)value)."
Another nonstandard extension to the \s-1JSON\s0 syntax, enabled with the
\&\f(CW\*(C`allow_tags\*(C'\fR setting, are tagged values. In this implementation, the
\&\fItag\fR must be a perl package/class name encoded as a \s-1JSON\s0 string, and the
\&\fIvalue\fR must be a \s-1JSON\s0 array encoding optional constructor arguments.
.Sp
See \*(L"\s-1OBJECT SERIALIZATION\*(R"\s0, below, for details.
.SS "\s-1PERL\s0 \-> \s-1JSON\s0"
.IX Subsection "PERL -> JSON"
The mapping from Perl to \s-1JSON\s0 is slightly more difficult, as Perl is a
truly typeless language, so we can only guess which \s-1JSON\s0 type is meant by
a Perl value.
.IP "hash references" 4
.IX Item "hash references"
Perl hash references become \s-1JSON\s0 objects. As there is no inherent ordering
in hash keys (or \s-1JSON\s0 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. Cpanel::JSON::XS can
optionally sort the hash keys (determined by the \fIcanonical\fR flag), so
the same datastructure will serialize to the same \s-1JSON\s0 text (given same
settings and version of Cpanel::JSON::XS), but this incurs a runtime overhead
and is only rarely useful, e.g. when you want to compare some \s-1JSON\s0 text
against another for equality.
.IP "array references" 4
.IX Item "array references"
Perl array references become \s-1JSON\s0 arrays.
.IP "other references" 4
.IX Item "other references"
Other unblessed references are generally not allowed and will cause an
exception to be thrown, except for references to the integers \f(CW0\fR and
\&\f(CW1\fR, which get turned into \f(CW\*(C`false\*(C'\fR and \f(CW\*(C`true\*(C'\fR atoms in \s-1JSON.\s0
.Sp
With the option \f(CW\*(C`allow_stringify\*(C'\fR, you can ignore the exception and return
the stringification of the perl value.
.Sp
With the option \f(CW\*(C`allow_unknown\*(C'\fR, you can ignore the exception and
return \f(CW\*(C`null\*(C'\fR instead.
.Sp
.Vb 3
\& encode_json [\e"x"] # => cannot encode reference to scalar \*(AqSCALAR(0x..)\*(Aq
\& # unless the scalar is 0 or 1
\& encode_json [\e0, \e1] # yields [false,true]
\&
\& allow_stringify\->encode_json [\e"x"] # yields "x" unlike JSON::PP
\& allow_unknown\->encode_json [\e"x"] # yields null as in JSON::PP
.Ve
.IP "Cpanel::JSON::XS::true, Cpanel::JSON::XS::false" 4
.IX Item "Cpanel::JSON::XS::true, Cpanel::JSON::XS::false"
These special values become \s-1JSON\s0 true and \s-1JSON\s0 false values,
respectively. You can also use \f(CW\*(C`\e1\*(C'\fR and \f(CW\*(C`\e0\*(C'\fR or \f(CW\*(C`!0\*(C'\fR and \f(CW\*(C`!1\*(C'\fR
directly if you want.
.Sp
.Vb 2
\& encode_json [Cpanel::JSON::XS::false, Cpanel::JSON::XS::true] # yields [false,true]
\& encode_json [!1, !0], [JSON_TYPE_BOOL, JSON_TYPE_BOOL] # yields [false,true]
.Ve
.Sp
eq/ne comparisons with true, false:
.Sp
false is eq to the empty string or the string 'false' or the special
empty string \f(CW\*(C`!!0\*(C'\fR or \f(CW\*(C`!1\*(C'\fR, i.e. \f(CW\*(C`SV_NO\*(C'\fR, or the numbers 0 or 0.0.
.Sp
true is eq to the string 'true' or to the special string \f(CW\*(C`!0\*(C'\fR
(i.e. \f(CW\*(C`SV_YES\*(C'\fR) or to the numbers 1 or 1.0.
.IP "blessed objects" 4
.IX Item "blessed objects"
Blessed objects are not directly representable in \s-1JSON,\s0 but
\&\f(CW\*(C`Cpanel::JSON::XS\*(C'\fR allows various optional ways of handling
objects. See \*(L"\s-1OBJECT SERIALIZATION\*(R"\s0, below, for details.
.Sp
See the \f(CW\*(C`allow_blessed\*(C'\fR and \f(CW\*(C`convert_blessed\*(C'\fR methods on various
options on how to deal with this: basically, you can choose between
throwing an exception, encoding the reference as if it weren't
blessed, use the objects overloaded stringification method or provide
your own serializer method.
.IP "simple scalars" 4
.IX Item "simple scalars"
Simple Perl scalars (any scalar that is not a reference) are the most
difficult objects to encode: Cpanel::JSON::XS will encode undefined
scalars or inf/nan as \s-1JSON\s0 \f(CW\*(C`null\*(C'\fR values and other scalars to either
number or string in non-deterministic way which may be affected or
changed by Perl version or any other loaded Perl module.
.Sp
If you want to have stable and deterministic types in \s-1JSON\s0 encoder then
use Cpanel::JSON::XS::Type.
.Sp
Alternative way for deterministic types is to use \f(CW\*(C`type_all_string\*(C'\fR
method when all perl scalars are encoded to \s-1JSON\s0 strings.
.Sp
Non-deterministic behavior is following: scalars that have last been
used in a string context before encoding as \s-1JSON\s0 strings, and anything
else as number value:
.Sp
.Vb 4
\& # dump as number
\& encode_json [2] # yields [2]
\& encode_json [\-3.0e17] # yields [\-3e+17]
\& my $value = 5; encode_json [$value] # yields [5]
\&
\& # used as string, but the two representations are for the same number
\& print $value;
\& encode_json [$value] # yields [5]
\&
\& # used as different string (non\-matching dual\-var)
\& my $str = \*(Aq0 but true\*(Aq;
\& my $num = 1 + $str;
\& encode_json [$num, $str] # yields [1,"0 but true"]
\&
\& # undef becomes null
\& encode_json [undef] # yields [null]
\&
\& # inf or nan becomes null, unless you answered
\& # "Do you want to handle inf/nan as strings" with yes
\& encode_json [9**9**9] # yields [null]
.Ve
.Sp
You can force the type to be a \s-1JSON\s0 string by stringifying it:
.Sp
.Vb 4
\& my $x = 3.1; # some variable containing a number
\& "$x"; # stringified
\& $x .= ""; # another, more awkward way to stringify
\& print $x; # perl does it for you, too, quite often
.Ve
.Sp
You can force the type to be a \s-1JSON\s0 number by numifying it:
.Sp
.Vb 3
\& my $x = "3"; # some variable containing a string
\& $x += 0; # numify it, ensuring it will be dumped as a number
\& $x *= 1; # same thing, the choice is yours.
.Ve
.Sp
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 \s-1JSON,\s0 and thus
null is returned instead. Optionally you can configure it to stringify
inf and nan values.
.SS "\s-1OBJECT SERIALIZATION\s0"
.IX Subsection "OBJECT SERIALIZATION"
As \s-1JSON\s0 cannot directly represent Perl objects, you have to choose between
a pure \s-1JSON\s0 representation (without the ability to deserialize the object
automatically again), and a nonstandard extension to the \s-1JSON\s0 syntax,
tagged values.
.PP
\fI\s-1SERIALIZATION\s0\fR
.IX Subsection "SERIALIZATION"
.PP
What happens when \f(CW\*(C`Cpanel::JSON::XS\*(C'\fR encounters a Perl object depends
on the \f(CW\*(C`allow_blessed\*(C'\fR, \f(CW\*(C`convert_blessed\*(C'\fR and \f(CW\*(C`allow_tags\*(C'\fR
settings, which are used in this order:
.ie n .IP "1. ""allow_tags"" is enabled and the object has a ""FREEZE"" method." 4
.el .IP "1. \f(CWallow_tags\fR is enabled and the object has a \f(CWFREEZE\fR method." 4
.IX Item "1. allow_tags is enabled and the object has a FREEZE method."
In this case, \f(CW\*(C`Cpanel::JSON::XS\*(C'\fR uses the Types::Serialiser object
serialization protocol to create a tagged \s-1JSON\s0 value, using a nonstandard
extension to the \s-1JSON\s0 syntax.
.Sp
This works by invoking the \f(CW\*(C`FREEZE\*(C'\fR method on the object, with the first
argument being the object to serialize, and the second argument being the
constant string \f(CW\*(C`JSON\*(C'\fR to distinguish it from other serializers.
.Sp
The \f(CW\*(C`FREEZE\*(C'\fR 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 \s-1JSON\s0 value in the following format:
.Sp
.Vb 1
\& ("classname")[FREEZE return values...]
.Ve
.Sp
e.g.:
.Sp
.Vb 3
\& ("URI")["http://www.google.com/"]
\& ("MyDate")[2013,10,29]
\& ("ImageData::JPEG")["Z3...VlCg=="]
.Ve
.Sp
For example, the hypothetical \f(CW\*(C`My::Object\*(C'\fR \f(CW\*(C`FREEZE\*(C'\fR method might use the
objects \f(CW\*(C`type\*(C'\fR and \f(CW\*(C`id\*(C'\fR members to encode the object:
.Sp
.Vb 2
\& sub My::Object::FREEZE {
\& my ($self, $serializer) = @_;
\&
\& ($self\->{type}, $self\->{id})
\& }
.Ve
.ie n .IP "2. ""convert_blessed"" is enabled and the object has a ""TO_JSON"" method." 4
.el .IP "2. \f(CWconvert_blessed\fR is enabled and the object has a \f(CWTO_JSON\fR method." 4
.IX Item "2. convert_blessed is enabled and the object has a TO_JSON method."
In this case, the \f(CW\*(C`TO_JSON\*(C'\fR method of the object is invoked in scalar
context. It must return a single scalar that can be directly encoded into
\&\s-1JSON.\s0 This scalar replaces the object in the \s-1JSON\s0 text.
.Sp
For example, the following \f(CW\*(C`TO_JSON\*(C'\fR method will convert all \s-1URI\s0
objects to \s-1JSON\s0 strings when serialized. The fact that these values
originally were \s-1URI\s0 objects is lost.
.Sp
.Vb 4
\& sub URI::TO_JSON {
\& my ($uri) = @_;
\& $uri\->as_string
\& }
.Ve
.ie n .IP "3. ""convert_blessed"" is enabled and the object has a stringification overload." 4
.el .IP "3. \f(CWconvert_blessed\fR is enabled and the object has a stringification overload." 4
.IX Item "3. convert_blessed is enabled and the object has a stringification overload."
In this case, the overloaded \f(CW""\fR method of the object is invoked in scalar
context. It must return a single scalar that can be directly encoded into
\&\s-1JSON.\s0 This scalar replaces the object in the \s-1JSON\s0 text.
.Sp
For example, the following \f(CW""\fR method will convert all \s-1URI\s0
objects to \s-1JSON\s0 strings when serialized. The fact that these values
originally were \s-1URI\s0 objects is lost.
.Sp
.Vb 2
\& package URI;
\& use overload \*(Aq""\*(Aq => sub { shift\->as_string };
.Ve
.ie n .IP "4. ""allow_blessed"" is enabled." 4
.el .IP "4. \f(CWallow_blessed\fR is enabled." 4
.IX Item "4. allow_blessed is enabled."
The object will be serialized as a \s-1JSON\s0 null value.
.IP "5. none of the above" 4
.IX Item "5. none of the above"
If none of the settings are enabled or the respective methods are missing,
\&\f(CW\*(C`Cpanel::JSON::XS\*(C'\fR throws an exception.
.PP
\fI\s-1DESERIALIZATION\s0\fR
.IX Subsection "DESERIALIZATION"
.PP
For deserialization there are only two cases to consider: either
nonstandard tagging was used, in which case \f(CW\*(C`allow_tags\*(C'\fR decides,
or objects cannot be automatically be deserialized, in which
case you can use postprocessing or the \f(CW\*(C`filter_json_object\*(C'\fR or
\&\f(CW\*(C`filter_json_single_key_object\*(C'\fR callbacks to get some real objects our of
your \s-1JSON.\s0
.PP
This section only considers the tagged value case: I a tagged \s-1JSON\s0 object
is encountered during decoding and \f(CW\*(C`allow_tags\*(C'\fR is disabled, a parse
error will result (as if tagged values were not part of the grammar).
.PP
If \f(CW\*(C`allow_tags\*(C'\fR is enabled, \f(CW\*(C`Cpanel::JSON::XS\*(C'\fR will look up the \f(CW\*(C`THAW\*(C'\fR method
of the package/classname used during serialization (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.
.PP
Otherwise, the \f(CW\*(C`THAW\*(C'\fR method is invoked with the classname as first
argument, the constant string \f(CW\*(C`JSON\*(C'\fR as second argument, and all the
values from the \s-1JSON\s0 array (the values originally returned by the
\&\f(CW\*(C`FREEZE\*(C'\fR method) as remaining arguments.
.PP
The method must then return the object. While technically you can return
any Perl scalar, you might have to enable the \f(CW\*(C`enable_nonref\*(C'\fR setting to
make that work in all cases, so better return an actual blessed reference.
.PP
As an example, let's implement a \f(CW\*(C`THAW\*(C'\fR function that regenerates the
\&\f(CW\*(C`My::Object\*(C'\fR from the \f(CW\*(C`FREEZE\*(C'\fR example earlier:
.PP
.Vb 2
\& sub My::Object::THAW {
\& my ($class, $serializer, $type, $id) = @_;
\&
\& $class\->new (type => $type, id => $id)
\& }
.Ve
.PP
See the \*(L"\s-1SECURITY CONSIDERATIONS\*(R"\s0 section below. Allowing external
json objects being deserialized to perl objects is usually a very bad
idea.
.SH "ENCODING/CODESET FLAG NOTES"
.IX Header "ENCODING/CODESET FLAG NOTES"
The interested reader might have seen a number of flags that signify
encodings or codesets \- \f(CW\*(C`utf8\*(C'\fR, \f(CW\*(C`latin1\*(C'\fR, \f(CW\*(C`binary\*(C'\fR and
\&\f(CW\*(C`ascii\*(C'\fR. There seems to be some confusion on what these do, so here
is a short comparison:
.PP
\&\f(CW\*(C`utf8\*(C'\fR controls whether the \s-1JSON\s0 text created by \f(CW\*(C`encode\*(C'\fR (and expected
by \f(CW\*(C`decode\*(C'\fR) is \s-1UTF\-8\s0 encoded or not, while \f(CW\*(C`latin1\*(C'\fR and \f(CW\*(C`ascii\*(C'\fR only
control whether \f(CW\*(C`encode\*(C'\fR escapes character values outside their respective
codeset range. Neither of these flags conflict with each other, although
some combinations make less sense than others.
.PP
Care has been taken to make all flags symmetrical with respect to
\&\f(CW\*(C`encode\*(C'\fR and \f(CW\*(C`decode\*(C'\fR, 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.
.PP
Below comes a verbose discussion of these flags. Note that a \*(L"codeset\*(R" is
simply an abstract set of character-codepoint pairs, while an encoding
takes those codepoint numbers and \fIencodes\fR them, in our case into
octets. Unicode is (among other things) a codeset, \s-1UTF\-8\s0 is an encoding,
and \s-1ISO\-8859\-1\s0 (= latin 1) and \s-1ASCII\s0 are both codesets \fIand\fR encodings at
the same time, which can be confusing.
.ie n .IP """utf8"" flag disabled" 4
.el .IP "\f(CWutf8\fR flag disabled" 4
.IX Item "utf8 flag disabled"
When \f(CW\*(C`utf8\*(C'\fR is disabled (the default), then \f(CW\*(C`encode\*(C'\fR/\f(CW\*(C`decode\*(C'\fR 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
\&\*(L"(re\-)interpreting\*(R" them as Unicode codepoints or Unicode characters,
respectively (to Perl, these are the same thing in strings unless you do
funny/weird/dumb stuff).
.Sp
This is useful when you want to do the encoding yourself (e.g. when you
want to have \s-1UTF\-16\s0 encoded \s-1JSON\s0 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 \s-1UTF\-8\s0 you certainly do \s-1NOT\s0 want
to \s-1UTF\-8\s0 encode your data first and have Perl encode it another time).
.ie n .IP """utf8"" flag enabled" 4
.el .IP "\f(CWutf8\fR flag enabled" 4
.IX Item "utf8 flag enabled"
If the \f(CW\*(C`utf8\*(C'\fR\-flag is enabled, \f(CW\*(C`encode\*(C'\fR/\f(CW\*(C`decode\*(C'\fR will encode all
characters using the corresponding \s-1UTF\-8\s0 multi-byte sequence, and will
expect your input strings to be encoded as \s-1UTF\-8,\s0 that is, no \*(L"character\*(R"
of the input string must have any value > 255, as \s-1UTF\-8\s0 does not allow
that.
.Sp
The \f(CW\*(C`utf8\*(C'\fR flag therefore switches between two modes: disabled means you
will get a Unicode string in Perl, enabled means you get an \s-1UTF\-8\s0 encoded
octet/binary string in Perl.
.ie n .IP """latin1"", ""binary"" or ""ascii"" flags enabled" 4
.el .IP "\f(CWlatin1\fR, \f(CWbinary\fR or \f(CWascii\fR flags enabled" 4
.IX Item "latin1, binary or ascii flags enabled"
With \f(CW\*(C`latin1\*(C'\fR (or \f(CW\*(C`ascii\*(C'\fR) enabled, \f(CW\*(C`encode\*(C'\fR will escape
characters with ordinal values > 255 (> 127 with \f(CW\*(C`ascii\*(C'\fR) and encode
the remaining characters as specified by the \f(CW\*(C`utf8\*(C'\fR flag.
With \f(CW\*(C`binary\*(C'\fR enabled, ordinal values > 255 are illegal.
.Sp
If \f(CW\*(C`utf8\*(C'\fR 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
\&\s-1ISO\-8859\-1\s0 string, and a Unicode string with all character values < 128 is
the same thing as an \s-1ASCII\s0 string in Perl).
.Sp
If \f(CW\*(C`utf8\*(C'\fR is enabled, you still get a correct UTF\-8\-encoded string,
regardless of these flags, just some more characters will be escaped using
\&\f(CW\*(C`\euXXXX\*(C'\fR then before.
.Sp
Note that \s-1ISO\-8859\-1\-\s0\fIencoded\fR strings are not compatible with \s-1UTF\-8\s0
encoding, while ASCII-encoded strings are. That is because the \s-1ISO\-8859\-1\s0
encoding is \s-1NOT\s0 a subset of \s-1UTF\-8\s0 (despite the \s-1ISO\-8859\-1\s0 \fIcodeset\fR being
a subset of Unicode), while \s-1ASCII\s0 is.
.Sp
Surprisingly, \f(CW\*(C`decode\*(C'\fR will ignore these flags and so treat all input
values as governed by the \f(CW\*(C`utf8\*(C'\fR flag. If it is disabled, this allows you
to decode \s-1ISO\-8859\-1\-\s0 and ASCII-encoded strings, as both strict subsets of
Unicode. If it is enabled, you can correctly decode \s-1UTF\-8\s0 encoded strings.
.Sp
So neither \f(CW\*(C`latin1\*(C'\fR, \f(CW\*(C`binary\*(C'\fR nor \f(CW\*(C`ascii\*(C'\fR are incompatible with the
\&\f(CW\*(C`utf8\*(C'\fR flag \- they only govern when the \s-1JSON\s0 output engine escapes a
character or not.
.Sp
The main use for \f(CW\*(C`latin1\*(C'\fR or \f(CW\*(C`binary\*(C'\fR is to relatively efficiently
store binary data as \s-1JSON,\s0 at the expense of breaking compatibility
with most \s-1JSON\s0 decoders.
.Sp
The main use for \f(CW\*(C`ascii\*(C'\fR is to force the output to not contain characters
with values > 127, which means you can interpret the resulting string
as \s-1UTF\-8, ISO\-8859\-1, ASCII, KOI8\-R\s0 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 \s-1JSON\s0 transfer is not 8\-bit clean or the encoding
might be mangled in between (e.g. in mail), and works because \s-1ASCII\s0 is a
proper subset of most 8\-bit and multibyte encodings in use in the world.
.SS "\s-1JSON\s0 and ECMAscript"
.IX Subsection "JSON and ECMAscript"
\&\s-1JSON\s0 syntax is based on how literals are represented in javascript (the
not-standardized predecessor of ECMAscript) which is presumably why it is
called \*(L"JavaScript Object Notation\*(R".
.PP
However, \s-1JSON\s0 is not a subset (and also not a superset of course) of
ECMAscript (the standard) or javascript (whatever browsers actually
implement).
.PP
If you want to use javascript's \f(CW\*(C`eval\*(C'\fR function to \*(L"parse\*(R" \s-1JSON,\s0 you
might run into parse errors for valid \s-1JSON\s0 texts, or the resulting data
structure might not be queryable:
.PP
One of the problems is that U+2028 and U+2029 are valid characters inside
\&\s-1JSON\s0 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 \f(CW\*(C`eval\*(C'\fR:
.PP
.Vb 1
\& use Cpanel::JSON::XS;
\&
\& print encode_json [chr 0x2028];
.Ve
.PP
The right fix for this is to use a proper \s-1JSON\s0 parser in your javascript
programs, and not rely on \f(CW\*(C`eval\*(C'\fR (see for example Douglas Crockford's
\&\fIjson2.js\fR parser).
.PP
If this is not an option, you can, as a stop-gap measure, simply encode to
ASCII-only \s-1JSON:\s0
.PP
.Vb 1
\& use Cpanel::JSON::XS;
\&
\& print Cpanel::JSON::XS\->new\->ascii\->encode ([chr 0x2028]);
.Ve
.PP
Note that this will enlarge the resulting \s-1JSON\s0 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.:
.PP
.Vb 5
\& # DO NOT USE THIS!
\& my $json = Cpanel::JSON::XS\->new\->utf8\->encode ([chr 0x2028]);
\& $json =~ s/\exe2\ex80\exa8/\e\eu2028/g; # escape U+2028
\& $json =~ s/\exe2\ex80\exa9/\e\eu2029/g; # escape U+2029
\& print $json;
.Ve
.PP
Note that \fIthis is a bad idea\fR: 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 \f(CW\*(C`eval\*(C'\fR naively simply \fIwill\fR cause problems.
.PP
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
\&\f(CW\*(C`_\|_proto_\|_\*(C'\fR property name for its own purposes.
.PP
If that is a problem, you could parse try to filter the resulting \s-1JSON\s0
output for these property strings, e.g.:
.PP
.Vb 1
\& $json =~ s/"_\|_proto_\|_"\es*:/"_\|_proto_\|_renamed":/g;
.Ve
.PP
This works because \f(CW\*(C`_\|_proto_\|_\*(C'\fR is not valid outside of strings, so every
occurrence of \f(CW\*(C`"_\|_proto_\|_"\es*:\*(C'\fR must be a string used as property name.
.PP
Unicode non-characters between U+FFFD and U+10FFFF are decoded either
to the recommended U+FFFD \s-1REPLACEMENT CHARACTER\s0 (see Unicode \s-1PR\s0 #121:
Recommended Practice for Replacement Characters), or in the binary or
relaxed mode left as is, keeping the illegal non-characters as before.
.PP
Raw non-Unicode characters outside the valid unicode range fail now to
parse, because \*(L"A string is a sequence of zero or more Unicode
characters\*(R" \s-1RFC 7159\s0 section 1 and "\s-1JSON\s0 text \s-1SHALL\s0 be encoded in
Unicode \s-1RFC 7159\s0 section 8.1. We use now the \s-1UTF8_DISALLOW_SUPER\s0
flag when parsing unicode.
.PP
If you know of other incompatibilities, please let me know.
.SS "\s-1JSON\s0 and \s-1YAML\s0"
.IX Subsection "JSON and YAML"
You often hear that \s-1JSON\s0 is a subset of \s-1YAML.\s0 \fIin general, there is
no way to configure \s-1JSON::XS\s0 to output a data structure as valid \s-1YAML\s0\fR
that works in all cases. If you really must use Cpanel::JSON::XS to
generate \s-1YAML,\s0 you should use this algorithm (subject to change in
future versions):
.PP
.Vb 2
\& my $to_yaml = Cpanel::JSON::XS\->new\->utf8\->space_after (1);
\& my $yaml = $to_yaml\->encode ($ref) . "\en";
.Ve
.PP
This will \fIusually\fR generate \s-1JSON\s0 texts that also parse as valid
\&\s-1YAML.\s0
.SS "\s-1SPEED\s0"
.IX Subsection "SPEED"
It seems that \s-1JSON::XS\s0 is surprisingly fast, as shown in the following
tables. They have been generated with the help of the \f(CW\*(C`eg/bench\*(C'\fR program
in the \s-1JSON::XS\s0 distribution, to make it easy to compare on your own
system.
.PP
\&\s-1JSON::XS\s0 is with Data::MessagePack and Sereal one of the fastest
serializers, because \s-1JSON\s0 and \s-1JSON::XS\s0 do not support backrefs (no
graph structures), only trees. Storable supports backrefs,
i.e. graphs. Data::MessagePack encodes its data binary (as Storable)
and supports only very simple subset of \s-1JSON.\s0
.PP
First comes a comparison between various modules using
a very short single-line \s-1JSON\s0 string (also available at
<http://dist.schmorp.de/misc/json/short.json>).
.PP
.Vb 3
\& {"method": "handleMessage", "params": ["user1",
\& "we were just talking"], "id": null, "array":[1,11,234,\-5,1e5,1e7,
\& 1, 0]}
.Ve
.PP
It shows the number of encodes/decodes per second (\s-1JSON::XS\s0 uses
the functional interface, while Cpanel::JSON::XS/2 uses the \s-1OO\s0 interface
with pretty-printing and hash key sorting enabled, Cpanel::JSON::XS/3 enables
shrink. \s-1JSON::DWIW/DS\s0 uses the deserialize function, while \s-1JSON::DWIW::FJ\s0
uses the from_json method). Higher is better:
.PP
.Vb 11
\& 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 |
\& \-\-\-\-\-\-\-\-\-\-\-\-\-\-+\-\-\-\-\-\-\-\-\-\-\-\-+\-\-\-\-\-\-\-\-\-\-\-\-+
.Ve
.PP
That is, \s-1JSON::XS\s0 is almost six times faster than \s-1JSON::DWIW\s0 on encoding,
about five times faster on decoding, and over thirty to seventy times
faster than \s-1JSON\s0's pure perl implementation. It also compares favourably
to Storable for small amounts of data.
.PP
Using a longer test string (roughly 18KB, generated from Yahoo! Locals
search \s-1API\s0 (<http://dist.schmorp.de/misc/json/long.json>).
.PP
.Vb 11
\& 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 |
\& \-\-\-\-\-\-\-\-\-\-\-\-\-\-+\-\-\-\-\-\-\-\-\-\-\-\-+\-\-\-\-\-\-\-\-\-\-\-\-+
.Ve
.PP
Again, \s-1JSON::XS\s0 leads by far (except for Storable which non-surprisingly
decodes a bit faster).
.PP
On large strings containing lots of high Unicode characters, some modules
(such as \s-1JSON::PC\s0) seem to decode faster than \s-1JSON::XS,\s0 but the result
will be broken due to missing (or wrong) Unicode handling. Others refuse
to decode or encode properly, so it was impossible to prepare a fair
comparison table for that case.
.PP
For updated graphs see <https://github.com/Sereal/Sereal/wiki/Sereal\-Comparison\-Graphs>
.SH "INTEROP with JSON and JSON::XS and other JSON modules"
.IX Header "INTEROP with JSON and JSON::XS and other JSON modules"
As long as you only serialize data that can be directly expressed in
\&\s-1JSON,\s0 \f(CW\*(C`Cpanel::JSON::XS\*(C'\fR is incapable of generating invalid \s-1JSON\s0
output (modulo bugs, but \f(CW\*(C`JSON::XS\*(C'\fR has found more bugs in the
official \s-1JSON\s0 testsuite (1) than the official \s-1JSON\s0 testsuite has found
in \f(CW\*(C`JSON::XS\*(C'\fR (0)).
\&\f(CW\*(C`Cpanel::JSON::XS\*(C'\fR is currently the only known \s-1JSON\s0 decoder which passes all
<http://seriot.ch/parsing_json.html> tests, while being the fastest also.
.PP
When you have trouble decoding \s-1JSON\s0 generated by this module using other
decoders, then it is very likely that you have an encoding mismatch or the
other decoder is broken.
.PP
When decoding, \f(CW\*(C`JSON::XS\*(C'\fR is strict by default and will likely catch
all errors. There are currently two settings that change this:
\&\f(CW\*(C`relaxed\*(C'\fR makes \f(CW\*(C`JSON::XS\*(C'\fR accept (but not generate) some
non-standard extensions, and \f(CW\*(C`allow_tags\*(C'\fR or \f(CW\*(C`allow_blessed\*(C'\fR will
allow you to encode and decode Perl objects, at the cost of being
totally insecure and not outputting valid \s-1JSON\s0 anymore.
.PP
\&\s-1JSON\-XS\-3.01\s0 broke interoperability with \s-1JSON\-2.90\s0 with booleans. See \s-1JSON\s0.
.PP
Cpanel::JSON::XS needs to know the \s-1JSON\s0 and \s-1JSON::XS\s0 versions to be able work
with those objects, especially when encoding a booleans like \f(CW\*(C`{"is_true":true}\*(C'\fR.
So you need to load these modules before.
.PP
true/false overloading and boolean representations are supported.
.PP
\&\s-1JSON::XS\s0 and \s-1JSON::PP\s0 representations are accepted and older \s-1JSON::XS\s0
accepts Cpanel::JSON::XS booleans. All \s-1JSON\s0 modules \s-1JSON, JSON, PP,
JSON::XS,\s0 Cpanel::JSON::XS produce JSON::PP::Boolean objects, just
Mojo and \s-1JSON::YAJL\s0 not. Mojo produces Mojo::JSON::_Bool and
JSON::YAJL::Parser just an unblessed \s-1IV.\s0
.PP
Cpanel::JSON::XS accepts JSON::PP::Boolean and Mojo::JSON::_Bool
objects as booleans.
.PP
I cannot think of any reason to still use \s-1JSON::XS\s0 anymore.
.SS "\s-1TAGGED VALUE SYNTAX AND STANDARD JSON EN/DECODERS\s0"
.IX Subsection "TAGGED VALUE SYNTAX AND STANDARD JSON EN/DECODERS"
When you use \f(CW\*(C`allow_tags\*(C'\fR to use the extended (and also nonstandard
and invalid) \s-1JSON\s0 syntax for serialized objects, and you still want to
decode the generated serialize objects, you can run a regex to replace
the tagged syntax by standard \s-1JSON\s0 arrays (it only works for \*(L"normal\*(R"
package names without comma, newlines or single colons). First, the
readable Perl version:
.PP
.Vb 2
\& # if your FREEZE methods return no values, you need this replace first:
\& $json =~ s/\e( \es* (" (?: [^\e\e":,]+|\e\e.|::)* ") \es* \e) \es* \e[\es*\e]/[$1]/gx;
\&
\& # this works for non\-empty constructor arg lists:
\& $json =~ s/\e( \es* (" (?: [^\e\e":,]+|\e\e.|::)* ") \es* \e) \es* \e[/[$1,/gx;
.Ve
.PP
And here is a less readable version that is easy to adapt to other
languages:
.PP
.Vb 1
\& $json =~ s/\e(\es*("([^\e\e":,]+|\e\e.|::)*")\es*\e)\es*\e[/[$1,/g;
.Ve
.PP
Here is an ECMAScript version (same regex):
.PP
.Vb 1
\& json = json.replace (/\e(\es*("([^\e\e":,]+|\e\e.|::)*")\es*\e)\es*\e[/g, "[$1,");
.Ve
.PP
Since this syntax converts to standard \s-1JSON\s0 arrays, it might be hard to
distinguish serialized objects from normal arrays. You can prepend a
\&\*(L"magic number\*(R" as first array element to reduce chances of a collision:
.PP
.Vb 1
\& $json =~ s/\e(\es*("([^\e\e":,]+|\e\e.|::)*")\es*\e)\es*\e[/["XU1peReLzT4ggEllLanBYq4G9VzliwKF",$1,/g;
.Ve
.PP
And after decoding the \s-1JSON\s0 text, you could walk the data
structure looking for arrays with a first element of
\&\f(CW\*(C`XU1peReLzT4ggEllLanBYq4G9VzliwKF\*(C'\fR.
.PP
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 \s-1JSON\s0 structure, and then:
.PP
.Vb 1
\& $json =~ s/\e[\es*"XU1peReLzT4ggEllLanBYq4G9VzliwKF"\es*,\es*("([^\e\e":,]+|\e\e.|::)*")\es*,/($1)[/g;
.Ve
.PP
Again, this has some limitations \- the magic string must not be encoded
with character escapes, and the constructor arguments must be non-empty.
.SH "RFC7159"
.IX Header "RFC7159"
Since this module was written, Google has written a new \s-1JSON RFC, RFC 7159\s0
(and \s-1RFC7158\s0). Unfortunately, this \s-1RFC\s0 breaks compatibility with both the
original \s-1JSON\s0 specification on www.json.org and \s-1RFC4627.\s0
.PP
As far as I can see, you can get partial compatibility when parsing by
using \f(CW\*(C`\->allow_nonref\*(C'\fR. However, consider the security implications
of doing so.
.PP
I haven't decided yet when to break compatibility with \s-1RFC4627\s0 by default
(and potentially leave applications insecure) and change the default to
follow \s-1RFC7159,\s0 but application authors are well advised to call \f(CW\*(C`\->allow_nonref(0)\*(C'\fR even if this is the current default, if they cannot
handle non-reference values, in preparation for the day when the default
will change.
.SH "SECURITY CONSIDERATIONS"
.IX Header "SECURITY CONSIDERATIONS"
\&\s-1JSON::XS\s0 and Cpanel::JSON::XS are not only fast. \s-1JSON\s0 is generally the
most secure serializing format, because it is the only one besides
Data::MessagePack, which does not deserialize objects per default. For
all languages, not just perl. The binary variant \s-1BSON\s0 (MongoDB) does
more but is unsafe.
.PP
It is trivial for any attacker to create such serialized objects in
\&\s-1JSON\s0 and trick perl into expanding them, thereby triggering certain
methods. Watch <https://www.youtube.com/watch?v=Gzx6KlqiIZE> for an
exploit demo for \*(L"\s-1CVE\-2015\-1592\s0 SixApart MovableType Storable Perl
Code Execution\*(R" for a deserializer which expands objects.
Deserializing even coderefs (methods, functions) or external
data would be considered the most dangerous.
.PP
Security relevant overview of serializers regarding deserializing
objects by default:
.PP
.Vb 1
\& Objects Coderefs External Data
\&
\& Data::Dumper YES YES YES
\& Storable YES NO (def) NO
\& Sereal YES NO NO
\& YAML YES NO NO
\& B::C YES YES YES
\& B::Bytecode YES YES YES
\& BSON YES YES NO
\& JSON::SL YES NO YES
\& JSON NO (def) NO NO
\& Data::MessagePack NO NO NO
\& XML NO NO YES
\&
\& Pickle YES YES YES
\& PHP Deserialize YES NO NO
.Ve
.PP
When you are using \s-1JSON\s0 in a protocol, talking to untrusted potentially
hostile creatures requires relatively few measures.
.PP
First of all, your \s-1JSON\s0 decoder should be secure, that is, should not have
any buffer overflows. Obviously, this module should ensure that.
.PP
Second, you need to avoid resource-starving attacks. That means you should
limit the size of \s-1JSON\s0 texts you accept, or make sure then when your
resources run out, that's just fine (e.g. by using a separate process that
can crash safely). The size of a \s-1JSON\s0 text in octets or characters is
usually a good indication of the size of the resources required to decode
it into a Perl structure. While \s-1JSON::XS\s0 can check the size of the \s-1JSON\s0
text, it might be too late when you already have it in memory, so you
might want to check the size before you accept the string.
.PP
Third, Cpanel::JSON::XS recurses using the C stack when decoding objects and
arrays. The C stack is a limited resource: for instance, on my amd64
machine with 8MB of stack size I can decode around 180k nested arrays but
only 14k nested \s-1JSON\s0 objects (due to perl itself recursing deeply on croak
to free the temporary). If that is exceeded, the program crashes. To be
conservative, the default nesting limit is set to 512. If your process
has a smaller stack, you should adjust this setting accordingly with the
\&\f(CW\*(C`max_depth\*(C'\fR method.
.PP
Also keep in mind that Cpanel::JSON::XS might leak contents of your Perl data
structures in its error messages, so when you serialize sensitive
information you might want to make sure that exceptions thrown by \s-1JSON::XS\s0
will not end up in front of untrusted eyes.
.PP
If you are using Cpanel::JSON::XS to return packets to consumption
by JavaScript scripts in a browser you should have a look at
<http://blog.archive.jpsykes.com/47/practical\-csrf\-and\-json\-security/> 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). You might also want to also look at Mojo::JSON
special escape rules to prevent from \s-1XSS\s0 attacks.
.ie n .SH """OLD"" VS. ""NEW"" JSON (RFC 4627 VS. RFC 7159)"
.el .SH "``OLD'' VS. ``NEW'' JSON (RFC 4627 VS. RFC 7159)"
.IX Header "OLD VS. NEW JSON (RFC 4627 VS. RFC 7159)"
\&\s-1TL\s0;DR: Due to security concerns, Cpanel::JSON::XS will not allow
scalar data in \s-1JSON\s0 texts by default \- you need to create your own
Cpanel::JSON::XS object and enable \f(CW\*(C`allow_nonref\*(C'\fR:
.PP
.Vb 1
\& my $json = JSON::XS\->new\->allow_nonref;
\&
\& $text = $json\->encode ($data);
\& $data = $json\->decode ($text);
.Ve
.PP
The long version: \s-1JSON\s0 being an important and supposedly stable format,
the \s-1IETF\s0 standardized it as \s-1RFC 4627\s0 in 2006. Unfortunately the inventor
of \s-1JSON\s0 Douglas Crockford unilaterally changed the definition of \s-1JSON\s0 in
javascript. Rather than create a fork, the \s-1IETF\s0 decided to standardize the
new syntax (apparently, so I as told, without finding it very amusing).
.PP
The biggest difference between the original \s-1JSON\s0 and the new \s-1JSON\s0 is that
the new \s-1JSON\s0 supports scalars (anything other than arrays and objects) at
the top-level of a \s-1JSON\s0 text. While this is strictly backwards compatible
to older versions, it breaks a number of protocols that relied on sending
\&\s-1JSON\s0 back-to-back, and is a minor security concern.
.PP
For example, imagine you have two banks communicating, and on one side,
the \s-1JSON\s0 coder gets upgraded. Two messages, such as \f(CW10\fR and \f(CW1000\fR
might then be confused to mean \f(CW101000\fR, something that couldn't happen
in the original \s-1JSON,\s0 because neither of these messages would be valid
\&\s-1JSON.\s0
.PP
If one side accepts these messages, then an upgrade in the coder on either
side could result in this becoming exploitable.
.PP
This module has always allowed these messages as an optional extension, by
default disabled. The security concerns are the reason why the default is
still disabled, but future versions might/will likely upgrade to the newer
\&\s-1RFC\s0 as default format, so you are advised to check your implementation
and/or override the default with \f(CW\*(C`\->allow_nonref (0)\*(C'\fR to ensure that
future versions are safe.
.SH "THREADS"
.IX Header "THREADS"
Cpanel::JSON::XS has proper ithreads support, unlike \s-1JSON::XS.\s0 If you
encounter any bugs with thread support please report them.
.PP
From Version 4.00 \- 4.19 you couldn't encode true with threads::shared
magic.
.SH "BUGS"
.IX Header "BUGS"
While the goal of the Cpanel::JSON::XS module is to be correct, that
unfortunately does not mean it's bug-free, only that the author thinks
its design is bug-free. If you keep reporting bugs and tests they will
be fixed swiftly, though.
.PP
Since the \s-1JSON::XS\s0 author refuses to use a public bugtracker and
prefers private emails, we use the tracker at \fBgithub\fR, so you might want
to report any issues twice. Once in private to \s-1MLEHMANN\s0 to be fixed in
\&\s-1JSON::XS\s0 and one to our the public tracker. Issues fixed by \s-1JSON::XS\s0
with a new release will also be backported to Cpanel::JSON::XS and
5.6.2, as long as cPanel relies on 5.6.2 and Cpanel::JSON::XS as our
serializer of choice.
.PP
<https://github.com/rurban/Cpanel\-JSON\-XS/issues>
.SH "LICENSE"
.IX Header "LICENSE"
This module is available under the same licences as perl, the Artistic
license and the \s-1GPL.\s0
.SH "SEE ALSO"
.IX Header "SEE ALSO"
The \fIcpanel_json_xs\fR command line utility for quick experiments.
.PP
\&\s-1JSON\s0, \s-1JSON::XS\s0, JSON::MaybeXS, Mojo::JSON, Mojo::JSON::MaybeXS,
\&\s-1JSON::SL\s0, \s-1JSON::DWIW\s0, \s-1JSON::YAJL\s0, JSON::Any, Test::JSON,
Locale::Wolowitz,
<https://metacpan.org/search?q=JSON>
.PP
<https://tools.ietf.org/html/rfc7159>
.PP
<https://tools.ietf.org/html/rfc4627>
.SH "AUTHOR"
.IX Header "AUTHOR"
Reini Urban <rurban@cpan.org>
.PP
Marc Lehmann <schmorp@schmorp.de>, http://home.schmorp.de/
.SH "MAINTAINER"
.IX Header "MAINTAINER"
Reini Urban <rurban@cpan.org>