# put data back to how it was
$data = flat_to_complex($flat);
$data = query_to_complex($query);
################################################
### some html form somewhere
<form>
<input type="text" name="foo.bar.baz" value="brum">
<input type="text" name="bing:2" value="blang">
<input type="text" name="'key with :, ., and \''.red" value="blue">
</form>
### when the form is submitted to the following code
use CGI;
use Data::URIEncode qw(query_to_complex);
my $q = CGI->new;
my $data = query_to_complex($q);
### data will look like
$data = {
foo => {
bar => {
baz = "brum",
},
},
bing => [
undef,
undef,
"blang",
],
"key with :, ., and '" => {
red = "blue",
},
};
DESCRIPTION
The world of the web works off of URI's. The Query string portion of
URIs already support encoding of key/value paired data - they just don't
natively allow for for complex data structures.
There are modules or encodings that do support arbitrarily complex data
structures. JSON, YAML and Data::Dumper all have their own way of
encoding complex structures. But then to pass them across the web, you
usually still have to URL encode them and pass them via a form
parameter.
Data::URIEncode allows for encoding and decoding complex (multi level
datastructures) using native Query String manipulators (such as CGI.pm).
It takes complex data and turns it into a flat hashref which can then be
turned into a URI query string using URL encoding. It also takes a flat
hashref of data passed in and translates it back to a complex structure.
One benefit of using Data::URIEncode is that a standard submission from
a standard html form can automatically be translated into complex data
even though it arrived in a "flat" form. This somewhat mimics the
abilities of XForms without introducing the complexity of XForms.
Another benefit is that sparse data can be represented in a more compact
form than JSON or YAML are able to provide. However, complex data with
long key names will be more verbose as the full data hierarchy must be
repeated for each value.
RULES
For each of the following rules, the $data can be translated to $flat
and $query by calling complex_to_flat and complex_to_query respectively.
The $flat and $query can be translated back into $data using
flat_to_complex and query_to_complex respectively.
Single quotes may be used to enclose complex strings.
Any key containing a colon ":", a dot ".", or a single quote "'"
must be quoted with single quotes and have enclosed single quotes
escaped.
$data = {"foo.bar" => "baz"}
$flat === {"'foo.bar'" => "baz"}
$query eq "'foo.bar'=baz" # the ' will be swapped with %27
########
$data = {"foo:bar" => "baz"}
$flat === {"'foo:bar'" => "baz"}
$query eq "'foo:bar'=baz" # the ' will be swapped with %27
########
$data = {"" => "baz"}
$flat === {"''" => "baz"}
$query eq "''=baz" # the ' will be swapped with %27
########
$data = {"'" => "baz"}
$flat === {"'\\''" => "baz"}
$query eq "'\\''=baz" # the ' will be swapped with %27 and the \ will be replaced with %5C
Single quotes were chosen as double quotes are most commonly used in
HTML forms, thus allowing escaped single quotes more easily inside
the double quoted name.
Undefined values are not included in the flattened data
$data = {foo => undef, bar => 1}
$flat === {bar => 1}
$query eq "bar=1"
Arrays created by flat_to_complex and query_to_complex must obey the
value of the $MAX_ARRAY_EXPAND variable.
FUNCTIONS
flat_to_complex
Takes a hashref of simple key value pairs. Returns a data structure
based on the the parsed key value pairs. The parsing proceeds
according to the rules listed in RULES.
complex_to_flat
Takes a complex data structure and turns it into a flat hashref
(single level key/value pairs only). The parsing proceeds according
to the rules listed in RULES.
complex_to_query
Similar to complex_to_flat, except that the flattened hashref is
then translated into query string suitable for use in a URI.
my $str = complex_to_query({foo => ['a','b']});
# $str eq "foo:0=a&foo:1=b"
query_to_complex
Takes one of a string, a reference to a string, a hash, or a CGI.pm
compatible object and translates it into a complex data structure.
Similar to flat_to_complex, exempt that a first step is taken to
access the query parameters from the CGI compatible object or
string. If a string or string ref is given, the CGI module is used
to parse the string into an initial flat hash of key value pairs
(using the param method). If another module is desired over, CGI.pm
you must initialize it with the data to be parsed prior to passing
the object to the query_to_complex function.
my $data = query_to_complex("foo.bar:0=baz");
my $data = query_to_complex(\ "foo.bar:0=baz");
my $data = query_to_complex({"foo.bar:0" => "baz"}); # same as flat_to_complex
my $cgi = CGI->new(\ "foo.bar:0=baz");
my $data = query_to_complex($cgi);
my $cgi = CGI->new; # use the values passed in from STDIN
my $data = query_to_complex($cgi);
VARIABLES
$MAX_ARRAY_EXPAND
Default value is 100. This variable is used to determine how large
flat_to_complex will allow an array to be expanded beyond its
current size. An array can grow as large as you have memory, but
intermediate values must exist.
Without this value, somebody could specify foo:1000000000000=bar and
your server would attempt to set the 1000000000000th index of the
foo value to bar.
The string "foo:101=bar" would die, but the string
"foo:50=bar&foo:101=baz" would not die because the intermediate
foo->[50] increments the foo arrayref by 51 and the subsequent
foo->[101] call increments the foo arrayref by only 51.
$DUMP_BLESSED_DATA
Default is true. If true, blessed hashrefs and arrayrefs will also
be added to the flat data returned by complex_to_flat. If false,
bless hashrefs and arrayrefs will be skipped.
BUGS
Circular refs are not detected. Any attempt to dump a struture with
cirular refs will result in an infinite loop. There is no immediate plan
to add circular ref tracking.
AUTHOR
Paul Seamons perlspam at seamons dot com
LICENSE
This library may be distributed under the same terms as Perl itself.
