in reply to Re^5: Building data structures from CGI params
in thread Building data structures from CGI params

That line is generated by the workaround

[...]

This is the bug, it forgot to split on \0, because Vars joins on \0

Oh, I see what you mean. Hm. I hate that NUL stuff...

I guess the best thing would be to add a config'able (probably on by default) splitting on them to the processing. I like that better than expecting a ->param, since that breaks it away from the contract of purely hash->hash.

I'll see if I can't get that into a new release in the next few weeks. Annoyingly busy... :|

Replies are listed 'Best First'.
Re^7: Building data structures from CGI params
by Anonymous Monk on Aug 09, 2012 at 08:30 UTC

    I like that better than expecting a ->param, since that breaks it away from the contract of purely hash->hash.

    Adding a branch breaks no contracts

    sub build_cgi_struct { my ($iv, $errs, $conf) = @_; use Scalar::Util qw(blessed); my $blessed = blessed $iv; my $splitnull = 1; if( $blessed and UNIVERSAL::can( $iv, 'param') ){ $splitnull = 0; if( $iv->isa('CGI') ){ $iv = $iv->{"param"} ; # cheat :) } else { $iv = { map { $_ => [ $iv->param($_) ] } $iv->param }; } } ... ... if $splitnull;
Re^7: Building data structures from CGI params
by fullermd (Priest) on Oct 01, 2012 at 16:27 UTC

    I'll see if I can't get that into a new release in the next few weeks.

    Well, OK, it's still just a few weeks. FSVO "few" :)

    I've uploaded a new version that does nullsplits by default (as well as one or two other minor tweaks).