/ 
CGI-Path-1.04
River stage zero No dependents
/ CGI::Path

NAME

CGI::Path - module to aid in traversing one or more paths

SYNOPSIS

CGI::Path allows for easy navigation through a set of steps, a path. It uses a session extensively (managed by default via Apache::Session) to hopefully simplify path based cgis.

A PATH

A path is a package, like CGI::Path::Skel. The path needs to be @ISA CGI::Path. The package can contain the step methods as described below. You can also make a directory for the path, like CGI/Path/Skel, where the direectory will contain a package for each step. This could be done from your $ENV{PERL5LIB}.

path_hash

The path_hash is what helps generate the path_array, which is just an array of steps. It is a hash to allow for easy overrides, since it is sort of hard to override the third element of an array through a series of news.

The path_hash needs a key named 'initial_step', and then steps that point down the line, like so

path_hash => {
  initial_step => 'page_one',
  page_one     => 'page_two',
  page_two     => 'page_three',
},

since page_three doesn't point anywhere, the path_array ends. You can just override $self->path_hash, and have it return a hash ref as above.

It is quite easy to look at $ENV{PATH_INFO} and control multiple paths through a single cgi. I offer the following as a simple example

sub path_hash { my $self = shift; my $sub_path = ''; if($ENV{PATH_INFO} && $ENV{PATH_INFO} =~ m@/(\w+)@) { $sub_path = $1; } my $sub_path_hash = { '' => { initial_step => 'main', main => '', }, };

### this is the generic path for adding something
if($sub_path =~ /^add_(\w+)$/ && !exists $sub_path_hash->{$sub_path}) {
  $sub_path_hash->{$sub_path} = {
    initial_step          => $sub_path,
    $sub_path             => "${sub_path}_confirm",
    "${sub_path}_confirm" => "${sub_path}_receipt",
  };
}
$sub_path = '' unless(exists $sub_path_hash->{$sub_path});
return $sub_path_hash->{$sub_path};
}

The above path_hash method was used to manage a series of distinct add paths. Distinct paths added users, categories, blogs and entries. Each path was to handled differently, but they each had a path similar to the add_user path, which looked like this

add_user => add_user_confirm => add_user_receipt

my_module

my_module by default is something like CGI::Path::Skel. You can override $self->my_module and have it return a scalar containing your my_module. Module overrides are done based on my_module.

my_content

my_module by default is something like path/skel. It defaults to a variant of my_module. You can override $self->my_content and have it return a scalar your my_content. html content gets printed based on my_content.

path_array

The path_array is formed from path_hash. It is an array ref of the steps in the path.

navigate

$self->navigate walks through a path of steps, where each step corresponds to a .htm content file and a .val validation hash.

A step corresponds to a .htm content file. The .htm and .val need to share the base same name.

$self->{this_step} is hash ref containing the following previous_step => the last step this_step => the current step validate_ref => the validation ref for the current step

Generally, navigate generates the form (see below), and for each step does the following

-- Get the validate ref (val_ref) for the given page -- Comparing the val_ref to the form see if info exists for the step -- Validate according to the val_ref -- If validation fails, or if info doesn't exist, process the page and stop

More specifically, the following methods can be called for a step, in the given order.

step details/possible uses --------------------------------------------- ${step}_hook_pre initializations, must return 0 or step gets skipped info_exists checks to see if you have info for this step ${step}_info_complete can be used to make sure you have all the info you need

validate                contains the following
${step}_pre_validate    stuff to check before validate proper
validate_proper         runs the .val file validation
${step}_post_validate   stuff to run after validate proper

${step}_hash_fill       return a hash ref of things to add to $self->fill
                        fill is a hash ref of what fills the forms
${step}_hash_form       perhaps set stuff for $self->{my_form}
                        my_form is a hash ref that gets passed to the process method
${step}_hash_errors     set errors
${step}_step            do actual stuff for the step
${step}_hook_post       last chance

generate_form

The goal is that the programmer just look at $self->form for form or session information. To help facilitate this goal, I use the following

$self->this_form           - form from the current hit
$self->{session_only} = [] - things that get deleted from this_form and get inserted from the session
$self->{session_wins} = [] - this_form wins by default, set this if you want something just from the session

The code then sets the form with the following line

$self->{form} = {%{$self->session}, %{$this_form}, %{$form}};

magic_fill

magic_fill is written to help aid in rapid development. It is a simple, space-delimited file of key/value pairs, like so

address                       123 Fake Street
email,email_address,from      cpan@spack.net

I split on the first white space, then split on commas for the key names. In the above example, I would end up with a ref like this

{
  address       => '123 Fake Street',
  email         => 'cpan@spack.net',
  email_address => 'cpan@spack.net',
  from          => 'cpan@spack.net',
}

Once I have a ref, those values will get filled into forms as pages are displayed. Makes it nice to fill forms with dummy data and test the flow of your script.

magic_fill is turned off by default. The method allow_magic_fill determines if magic_fill is on. By default allow_magic_fill just looks at $self->{allow_magic_fill} and returns true or false accordingly. magic_fill_filename points to the location of your file.

When you new up your CGI::Path object you just need to do something like the following

my $self = CGI::Path->new({ allow_magic_fill => 1, magic_fill_filename => "/path/to/magic_fill_file", });

You can use variable values using the magic_fill_interpolation_hash. By default you can use Template::Toolkit tags, like so

currenttime [% localtime %]

Currently, the following are included by default in the magic_fill_interpolation_hash

script    - a good guess at the name of your script
_script   - the stuff after the last _ in the above script
localtime - scalar (localtime),
time      - time,

I also include %ENV

Two other keys are not available by default, based on micro seconds namely

micro      - join(".", &Time::HiRes::gettimeofday()), which really tries to get you a unique value
micro_part - (&Time::HiRes::gettimeofday())[1];, which is just the micro seconds

To make these swaps available you need to set $self->{allow_magic_micro} to a true value.

Session management

CGI::Path uses Apache::Session::File by default for session management. If you use this default you will need to write the following methods

session_dir      - returns the directory where the session files will go
session_lock_dir - returns the directory where the session lock files will go

AUTHOR

Copyright 2003-2004, Earl J. Cahill. All rights reserved.

This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself.

Address bug reports and comments to: cpan@spack.net.

When sending bug reports, please provide the version of CGI::Path, the version of Perl, and the name and version of the operating system you are using.