view release on metacpan or  search on metacpan

lib/CGI/Thin.pm  view on Meta::CPAN

            So $cgi_data{'color'} will be the scalar value selected.
        3)  The user selects exactly more than one color.
            So $cgi_data{'color'} will be a reference to an array of the values selected.

        To fix this you could call the parser by giving it a list of keys that you want
        to force to be arrays.  In this case like...

          use CGI::Thin;
          my %cgi_data = &Parse_CGI ('color');

        Now it they pick exactly one color, $cgi_data{'color'} will be a reference to
        an array of the one value selected.  And thus there will be no need for
        special cases later in the code.

=head1 BUGS

=head2 Fixed

=over 4

=item *

Added %([0-9a-fA-F]{2} to the regular expression to avoid illegal escapes

=item *

Now split the key/value pairs by [;&] not just the ampersand

=back

=head2 Pending

=over 4

=item *

Long headers lines that have been broken over multiple lines in
multipart/form-data don't seem to be handled.

=item *

Large file uploads (like 150MB) will clobber main memory.  One possible addition is
to change how multipart/form-data is read and to spit files directly to the temp directory
and return to the script a filename so it can be retreived from there.

=item *

Any thoughts on adapting it for use withing a mod_perl environment?

Under Apache::Registry, which emulates a CGI environmnet, it should be.
Under plain ol' mod_perl, we need to interact with the
Apache::Request class a bit instead of %ENV and STDIN.

This feature may be added in the next incarnation of the module, or possibly a companion
CGI::Thin::Mod_Perlish may be created to do it if the code will be too different.

=back

=head1 SEE ALSO

CGI::Thin::Cookies

=head1 SUPPORT

    Visit CGI::Thin's web site at
        http://www.PlatypiVentures.com/perl/modules/cgi_thin.shtml
    Send email to
        mailto:cgi_thin@PlatypiVentures.com

=head1 AUTHOR

    R. Geoffrey Avery
    CPAN ID: RGEOFFREY
    modules@PlatypiVentures.com
    http://www.PlatypiVentures.com/perl

=head1 COPYRIGHT

This module is free software, you may redistribute it or modify in under the same terms as Perl itself.

=cut

############################################# main pod documentation end ##

################################################ subroutine header begin ##
################################################## subroutine header end ##

sub Parse_CGI
{
	my %hash = ();

	foreach my $entry (split(/[&;]/, $ENV{'QUERY_STRING'})) {
		&Insert_Item (\%hash, &Divide_Item ($entry));
	}

	if ((defined $ENV{'CONTENT_TYPE'}) && ($ENV{'CONTENT_TYPE'} =~ m|multipart/form-data|si)) {
		my $in;
		read(STDIN, $in, $ENV{'CONTENT_LENGTH'});

		### Find the field "boundary" string.
		my $boundary = substr($in, 0, index($in, "\r\n") - 1);
		### Tokenize the input.
		my @args = split(/\s*$boundary\s*/s, $in);
		### remove extra pieces before first and after last boundary
		shift @args;
		pop @args;

		foreach my $entry (@args) {
			# Split the token into header and content
			my ($head, $item) = split(/\r\n\r\n/ios, $entry, 2);

			# ... name="CGI_FILE" filename="myfile.txt" ....
			# so this is a bit of a trick, based on the double
			# occurence of 'name'.
			my ($name, $file) = ($head =~ /name="(.*?)"/gios);

			my $mimetype;
			if ($head =~ /Content-type:\s*(\S+)/gios) {
				$mimetype = $1;
			}