NAME

VERSION

SYNOPSIS

    use CGI::Ex::JSONDump;

    my $js = JSONDump(\%complex_data, {pretty => 1});

    ### OR

    my $js = CGI::Ex::JSONDump->new({pretty => 1})->dump(\%complex_data);

DESCRIPTION

CGI::Ex::JSONDump has roughly the same output as JSON::objToJson, but with the following differences:

    - CGI::Ex::JSONDump is much much lighter and smaller (a whopping 134 lines).
    - It dumps Javascript in more browser friendly format (handling of </script> tags).
    - It removes unknown key types by default instead of dying.
    - It allows for a general handler to handle unknown key types.
    - It allows for fine grain control of all whitespace.
    - It allows for skipping keys by name or by regex.
    - It dumps both data structures and scalar types.

METHODS

new

Create a CGI::Ex::JSONDump object. Takes arguments hashref as single argument.

    my $obj = CGI::Ex::JSONDump->new(\%args);
    

See the arguments section for a list of the possible arguments.

dump

Takes a perl data structure or scalar string or number and returns a string containing the javascript representation of that string (in Javascript object notation - JSON).

js_escape

Takes a scalar string or number and returns a javascript escaped string that will embed properly in javascript. All numbers and strings of nested data structures are passed through this method.

FUNCTIONS

JSONDump

A wrapper around the new and dump methods. Takes a structure to dump and optional args to pass to the new routine.

    JSONDump($data, $args);
    

Is the same as:

    CGI::Ex::JSONDump->new($args)->dump($data);
    

ARGUMENTS

pretty

0 or 1. Default 0 (false). If true then dumped structures will include whitespace to make them more readable.

     JSONDump({a => [1, 2]}, {pretty => 0});
     JSONDump({a => [1, 2]}, {pretty => 1});

     Would print

     {"a":[1,2]}
     {
       "a" : [
         1,
         2
       ]
     }
    

single_quote

0 or 1. Default 0 (false). If true then escaped values will be quoted with single quotes. Otherwise values are quoted with double quotes.

     JSONDump("a", {single_quote => 0});
     JSONDump("a", {single_quote => 1});

     Would print

     "a"
     'a'
    

sort_keys

0 or 1. Default 1 (true)

If true, then key/value pairs of hashrefs will be output in sorted order.

play_coderefs

0 or 1. Default 0 (false). If true, then any code refs will be executed and the returned string will be dumped.

If false, then keys of hashrefs that contain coderefs will be skipped (unless the handle_unknown_types property is set). Coderefs that are in arrayrefs will show up as "CODE(0x814c648)" unless the handle_unknown_types property is set.

handle_unknown_types

Default undef. If true it should contain a coderef that will be called if any unknown types are encountered. The only default known types are scalar string or number values, unblessed HASH refs and ARRAY refs (and CODE refs if the play_coderefs property is set). All other types will be passed to the handle_unknown_types method call.

    JSONDump({a => bless({}, 'A'), b => 1}, {
        handle_unknown_types => sub {
            my $self = shift; # a JSON object
            my $data = shift; # the object to dump

            return $self->js_escape("Ref=" . ref $data);
        },
        pretty => 0,
    });

    Would print

    {"a":"Ref=A","b":1}
    

If the handle_unknown_types method is not set then keys hashrefs that have values with unknown types will not be included in the javascript output.

    JSONDump({a => bless({}, 'A'), b => 1}, {pretty => 0});

    Would print

    {"b":1}
    

skip_keys

Should contain an arrayref of keys or a hashref whose keys are the keys to skip. Default is unset. Any keys of hashrefs (including nested hashrefs) that are listed in the skip_keys item will not be included in the javascript output.

    JSONDump({a => 1, b => 1}, {skip_keys => ['a'], pretty => 0});

    Would print

    {"b":1}
    

skip_keys_qr

Similar to skip_keys but should contain a regex. Any keys of hashrefs (including nested hashrefs) that match the skip_keys_qr regex will not be included in the javascript output.

    JSONDump({a => 1, _b => 1}, {skip_keys_qr => qr/^_/, pretty => 0});

    Would print

    {"a":1}
    

indent

The level to indent each nested data structure level if pretty is true. Default is " " (two spaces).

hash_nl

The whitespace to add after each hashref key/value pair if pretty is true. Default is "\n".

hash_sep

The separator and whitespace to put between each hashref key/value pair if pretty is true. Default is " : ".

array_nl

The whitespace to add after each arrayref entry if pretty is true. Default is "\n".

str_nl

The whitespace to add in between newline separated strings if pretty is true or the output line is greater than 80 characters. Default is "\n" (if pretty is true).

    JSONDump("This is a long string\n"
             ."with plenty of embedded newlines\n"
             ."and is greater than 80 characters.\n", {pretty => 1});

    Would print

    "This is a long string\n"
      +"with plenty of embedded newlines\n"
      +"and is greater than 80 characters.\n"

    JSONDump("This is a long string\n"
             ."with plenty of embedded newlines\n"
             ."and is greater than 80 characters.\n", {pretty => 1, str_nl => ""});

    Would print

    "This is a long string\nwith plenty of embedded newlines\nand is greater than 80 characters.\n"
    

If the string is less than 80 characters, or if str_nl is set to "", then the escaped string will be contained on a single line. Setting pretty to 0 effectively sets str_nl equal to "".

no_tag_splitting

Default off. If JSON is embedded in an HTML document and the JSON contains "<html>", "</html>", "<script>", "</script>", "<!--", or , "-->" tags, they are split apart with a quote, a +, and a quote. This allows the embedded tags to not affect the currently playing JavaScript.

However, if the JSON that is output is intended for deserialization by another non-javascript-engine JSON parser, this splitting behavior may cause errors when the JSON is imported. To avoid the splitting behavior in these cases you can use the no_tag_splitting flag to turn off the behavior.

    JSONDump("<html><!-- comment --><script></script></html>");

    Would print

    "<htm"+"l><!-"+"- comment --"+"><scrip"+"t></scrip"+"t></htm"+"l>"

    With the flag

    JSONDump("<html><!-- comment --><script></script></html>", {no_tag_splitting => 1});

    Would print

    "<html><!-- comment --><script></script></html>"
    

LICENSE

AUTHORS