File Coverage

blib/lib/Catalyst.pm

Criterion	Covered	Total	%
statement	1129	1279	88.2
branch	427	572	74.6
condition	171	230	74.3
subroutine	191	208	91.8
pod	109	110	99.0
total	2027	2399	84.4

package Catalyst;

use Moose;
use Moose::Meta::Class ();
extends 'Catalyst::Component';
use Moose::Util qw/find_meta/;
use namespace::clean -except => 'meta';
use Catalyst::Exception;
use Catalyst::Exception::Detach;
use Catalyst::Exception::Go;
use Catalyst::Log;
use Catalyst::Request;
use Catalyst::Request::Upload;
use Catalyst::Response;
use Catalyst::Utils;
use Catalyst::Controller;
use Data::OptList;
use Devel::InnerPackage ();
use Module::Pluggable::Object ();
use Text::SimpleTable ();
use Path::Class::Dir ();
use Path::Class::File ();
use URI ();
use URI::http;
use URI::https;
use HTML::Entities;
use Tree::Simple qw/use_weak_refs/;
use Tree::Simple::Visitor::FindByUID;
use Class::C3::Adopt::NEXT;
use List::Util qw/uniq/;
use attributes;
use String::RewritePrefix;
use Catalyst::EngineLoader;
use utf8;
use Carp qw/croak carp shortmess/;
use Try::Tiny;
use Safe::Isa;
use Moose::Util 'find_meta';
use Plack::Middleware::Conditional;
use Plack::Middleware::ReverseProxy;
use Plack::Middleware::IIS6ScriptNameFix;
use Plack::Middleware::IIS7KeepAliveFix;
use Plack::Middleware::LighttpdScriptNameFix;
use Plack::Middleware::ContentLength;
use Plack::Middleware::Head;
use Plack::Middleware::HTTPExceptions;
use Plack::Middleware::FixMissingBodyInRedirect;
use Plack::Middleware::MethodOverride;
use Plack::Middleware::RemoveRedundantBody;
use Catalyst::Middleware::Stash;
use Plack::Util;
use Class::Load 'load_class';
use Encode 2.21 'decode_utf8', 'encode_utf8';
use Scalar::Util;

our $VERSION = '5.90131';
$VERSION =~ tr/_//d;

BEGIN { require 5.008003; }

has stack => (is => 'ro', default => sub { [] });
has state => (is => 'rw', default => 0);
has stats => (is => 'rw');
has action => (is => 'rw');
has counter => (is => 'rw', default => sub { {} });
has request => (
    is => 'rw',
    default => sub {
        my $self = shift;
        my $class = ref $self;
        my $composed_request_class = $class->composed_request_class;
        return $composed_request_class->new( $self->_build_request_constructor_args);
    },
    predicate => 'has_request',
    lazy => 1,
);
sub _build_request_constructor_args {
    my $self = shift;
    my %p = ( _log => $self->log );
    $p{_uploadtmp} = $self->_uploadtmp if $self->_has_uploadtmp;
    $p{data_handlers} = {$self->registered_data_handlers};
    $p{_use_hash_multivalue} = $self->config->{use_hash_multivalue_in_request}
      if $self->config->{use_hash_multivalue_in_request};
    \%p;
}

sub composed_request_class {
  my $class = shift;
  return $class->_composed_request_class if $class->_composed_request_class;

  my @traits = (@{$class->request_class_traits||[]}, @{$class->config->{request_class_traits}||[]});

  # For each trait listed, figure out what the namespace is.  First we try the $trait
  # as it is in the config.  Then try $MyApp::TraitFor::Request:$trait. Last we try
  # Catalyst::TraitFor::Request::$trait.  If none load, throw error.

  my $trait_ns = 'TraitFor::Request';
  my @normalized_traits = map {
    Class::Load::load_first_existing_class($_, $class.'::'.$trait_ns.'::'. $_, 'Catalyst::'.$trait_ns.'::'.$_)
  } @traits;

  if ($class->debug && scalar(@normalized_traits)) {
      my $column_width = Catalyst::Utils::term_width() - 6;
      my $t = Text::SimpleTable->new($column_width);
      $t->row($_) for @normalized_traits;
      $class->log->debug( "Composed Request Class Traits:\n" . $t->draw . "\n" );
  }

  return $class->_composed_request_class(Moose::Util::with_traits($class->request_class, @normalized_traits));
}

has response => (
    is => 'rw',
    default => sub {
        my $self = shift;
        my $class = ref $self;
        my $composed_response_class = $class->composed_response_class;
        return $composed_response_class->new( $self->_build_response_constructor_args);
    },
    predicate=>'has_response',
    lazy => 1,
);
sub _build_response_constructor_args {
    return +{
      _log => $_[0]->log,
      encoding => $_[0]->encoding,
    };
}

sub composed_response_class {
  my $class = shift;
  return $class->_composed_response_class if $class->_composed_response_class;

  my @traits = (@{$class->response_class_traits||[]}, @{$class->config->{response_class_traits}||[]});

  my $trait_ns = 'TraitFor::Response';
  my @normalized_traits = map {
    Class::Load::load_first_existing_class($_, $class.'::'.$trait_ns.'::'. $_, 'Catalyst::'.$trait_ns.'::'.$_)
  } @traits;

  if ($class->debug && scalar(@normalized_traits)) {
      my $column_width = Catalyst::Utils::term_width() - 6;
      my $t = Text::SimpleTable->new($column_width);
      $t->row($_) for @normalized_traits;
      $class->log->debug( "Composed Response Class Traits:\n" . $t->draw . "\n" );
  }

  return $class->_composed_response_class(Moose::Util::with_traits($class->response_class, @normalized_traits));
}

has namespace => (is => 'rw');

sub depth { scalar @{ shift->stack || [] }; }
sub comp { shift->component(@_) }

sub req {
    my $self = shift; return $self->request(@_);
}
sub res {
    my $self = shift; return $self->response(@_);
}

# For backwards compatibility
sub finalize_output { shift->finalize_body(@_) };

# For statistics
our $COUNT     = 1;
our $START     = time;
our $RECURSION = 1000;
our $DETACH    = Catalyst::Exception::Detach->new;
our $GO        = Catalyst::Exception::Go->new;

#I imagine that very few of these really
#need to be class variables. if any.
#maybe we should just make them attributes with a default?
__PACKAGE__->mk_classdata($_)
  for qw/components arguments dispatcher engine log dispatcher_class
  engine_loader context_class request_class response_class stats_class
  setup_finished _psgi_app loading_psgi_file run_options _psgi_middleware
  _data_handlers _encoding _encode_check finalized_default_middleware
  request_class_traits response_class_traits stats_class_traits
  _composed_request_class _composed_response_class _composed_stats_class/;

__PACKAGE__->dispatcher_class('Catalyst::Dispatcher');
__PACKAGE__->request_class('Catalyst::Request');
__PACKAGE__->response_class('Catalyst::Response');
__PACKAGE__->stats_class('Catalyst::Stats');

sub composed_stats_class {
  my $class = shift;
  return $class->_composed_stats_class if $class->_composed_stats_class;

  my @traits = (@{$class->stats_class_traits||[]}, @{$class->config->{stats_class_traits}||[]});

  my $trait_ns = 'TraitFor::Stats';
  my @normalized_traits = map {
    Class::Load::load_first_existing_class($_, $class.'::'.$trait_ns.'::'. $_, 'Catalyst::'.$trait_ns.'::'.$_)
  } @traits;

  if ($class->debug && scalar(@normalized_traits)) {
      my $column_width = Catalyst::Utils::term_width() - 6;
      my $t = Text::SimpleTable->new($column_width);
      $t->row($_) for @normalized_traits;
      $class->log->debug( "Composed Stats Class Traits:\n" . $t->draw . "\n" );
  }

  return $class->_composed_stats_class(Moose::Util::with_traits($class->stats_class, @normalized_traits));
}

__PACKAGE__->_encode_check(Encode::FB_CROAK | Encode::LEAVE_SRC);

sub import {
    my ( $class, @arguments ) = @_;

    # We have to limit $class to Catalyst to avoid pushing Catalyst upon every
    # callers @ISA.
    return unless $class eq 'Catalyst';

    my $caller = caller();
    return if $caller eq 'main';

    my $meta = Moose::Meta::Class->initialize($caller);
    unless ( $caller->isa('Catalyst') ) {
        my @superclasses = ($meta->superclasses, $class, 'Catalyst::Controller');
        $meta->superclasses(@superclasses);
    }
    # Avoid possible C3 issues if 'Moose::Object' is already on RHS of MyApp
    $meta->superclasses(grep { $_ ne 'Moose::Object' } $meta->superclasses);

    unless( $meta->has_method('meta') ){
        if ($Moose::VERSION >= 1.15) {
            $meta->_add_meta_method('meta');
        }
        else {
            $meta->add_method(meta => sub { Moose::Meta::Class->initialize("${caller}") } );
        }
    }

    $caller->arguments( [@arguments] );
    $caller->setup_home;
}

sub _application { $_[0] }

=encoding UTF-8

=head1 NAME

Catalyst - The Elegant MVC Web Application Framework

=head1 SYNOPSIS

See the L<Catalyst::Manual> distribution for comprehensive
documentation and tutorials.

    # Install Catalyst::Devel for helpers and other development tools
    # use the helper to create a new application
    catalyst.pl MyApp

    # add models, views, controllers
    script/myapp_create.pl model MyDatabase DBIC::Schema create=static dbi:SQLite:/path/to/db
    script/myapp_create.pl view MyTemplate TT
    script/myapp_create.pl controller Search

    # built in testserver -- use -r to restart automatically on changes
    # --help to see all available options
    script/myapp_server.pl

    # command line testing interface
    script/myapp_test.pl /yada

    ### in lib/MyApp.pm
    use Catalyst qw/-Debug/; # include plugins here as well

    ### In lib/MyApp/Controller/Root.pm (autocreated)
    sub foo : Chained('/') Args() { # called for /foo, /foo/1, /foo/1/2, etc.
        my ( $self, $c, @args ) = @_; # args are qw/1 2/ for /foo/1/2
        $c->stash->{template} = 'foo.tt'; # set the template
        # lookup something from db -- stash vars are passed to TT
        $c->stash->{data} =
          $c->model('Database::Foo')->search( { country => $args[0] } );
        if ( $c->req->params->{bar} ) { # access GET or POST parameters
            $c->forward( 'bar' ); # process another action
            # do something else after forward returns
        }
    }

    # The foo.tt TT template can use the stash data from the database
    [% WHILE (item = data.next) %]
        [% item.foo %]
    [% END %]

    # called for /bar/of/soap, /bar/of/soap/10, etc.
    sub bar : Chained('/') PathPart('/bar/of/soap') Args() { ... }

    # called after all actions are finished
    sub end : Action {
        my ( $self, $c ) = @_;
        if ( scalar @{ $c->error } ) { ... } # handle errors
        return if $c->res->body; # already have a response
        $c->forward( 'MyApp::View::TT' ); # render template
    }

See L<Catalyst::Manual::Intro> for additional information.

=head1 DESCRIPTION

Catalyst is a modern framework for making web applications without the
pain usually associated with this process. This document is a reference
to the main Catalyst application. If you are a new user, we suggest you
start with L<Catalyst::Manual::Tutorial> or L<Catalyst::Manual::Intro>.

See L<Catalyst::Manual> for more documentation.

Catalyst plugins can be loaded by naming them as arguments to the "use
Catalyst" statement. Omit the C<Catalyst::Plugin::> prefix from the
plugin name, i.e., C<Catalyst::Plugin::My::Module> becomes
C<My::Module>.

    use Catalyst qw/My::Module/;

If your plugin starts with a name other than C<Catalyst::Plugin::>, you can
fully qualify the name by using a unary plus:

    use Catalyst qw/
        My::Module
        +Fully::Qualified::Plugin::Name
    /;

Special flags like C<-Debug> can also be specified as
arguments when Catalyst is loaded:

    use Catalyst qw/-Debug My::Module/;

The position of plugins and flags in the chain is important, because
they are loaded in the order in which they appear.

The following flags are supported:

=head2 -Debug

Enables debug output. You can also force this setting from the system
environment with CATALYST_DEBUG or <MYAPP>_DEBUG. The environment
settings override the application, with <MYAPP>_DEBUG having the highest
priority.

This sets the log level to 'debug' and enables full debug output on the
error screen. If you only want the latter, see L<< $c->debug >>.

=head2 -Home

Forces Catalyst to use a specific home directory, e.g.:

    use Catalyst qw[-Home=/usr/mst];

This can also be done in the shell environment by setting either the
C<CATALYST_HOME> environment variable or C<MYAPP_HOME>; where C<MYAPP>
is replaced with the uppercased name of your application, any "::" in
the name will be replaced with underscores, e.g. MyApp::Web should use
MYAPP_WEB_HOME. If both variables are set, the MYAPP_HOME one will be used.

If none of these are set, Catalyst will attempt to automatically detect the
home directory. If you are working in a development environment, Catalyst
will try and find the directory containing either Makefile.PL, Build.PL,
dist.ini, or cpanfile. If the application has been installed into the system
(i.e. you have done C<make install>), then Catalyst will use the path to your
application module, without the .pm extension (e.g., /foo/MyApp if your
application was installed at /foo/MyApp.pm)

=head2 -Log

    use Catalyst '-Log=warn,fatal,error';

Specifies a comma-delimited list of log levels.

=head2 -Stats

Enables statistics collection and reporting.

   use Catalyst qw/-Stats=1/;

You can also force this setting from the system environment with CATALYST_STATS
or <MYAPP>_STATS. The environment settings override the application, with
<MYAPP>_STATS having the highest priority.

Stats are also enabled if L<< debugging |/"-Debug" >> is enabled.

=head1 METHODS

=head2 INFORMATION ABOUT THE CURRENT REQUEST

=head2 $c->action

Returns a L<Catalyst::Action> object for the current action, which
stringifies to the action name. See L<Catalyst::Action>.

=head2 $c->namespace

Returns the namespace of the current action, i.e., the URI prefix
corresponding to the controller of the current action. For example:

    # in Controller::Foo::Bar
    $c->namespace; # returns 'foo/bar';

=head2 $c->request

=head2 $c->req

Returns the current L<Catalyst::Request> object, giving access to
information about the current client request (including parameters,
cookies, HTTP headers, etc.). See L<Catalyst::Request>.

There is a predicate method C<has_request> that returns true if the
request object has been created.  This is something you might need to
check if you are writing plugins that run before a request is finalized.

=head2 REQUEST FLOW HANDLING

=head2 $c->forward( $action [, \@arguments ] )

=head2 $c->forward( $class, $method, [, \@arguments ] )

=head2 $c->forward( $component_instance, $method, [, \@arguments ] )


This is one way of calling another action (method) in the same or
a different controller. You can also use C<< $self->my_method($c, @args) >>
in the same controller or C<< $c->controller('MyController')->my_method($c, @args) >>
in a different controller.
The main difference is that 'forward' uses some of the Catalyst request
cycle overhead, including debugging, which may be useful to you. On the
other hand, there are some complications to using 'forward', restrictions
on values returned from 'forward', and it may not handle errors as you prefer.
Whether you use 'forward' or not is up to you; it is not considered superior to
the other ways to call a method.

'forward' calls  another action, by its private name. If you give a
class name but no method, C<process()> is called. You may also optionally
pass arguments in an arrayref. The action will receive the arguments in
C<@_> and C<< $c->req->args >>. Upon returning from the function,
C<< $c->req->args >> will be restored to the previous values.

Any data C<return>ed from the action forwarded to, will be returned by the
call to forward.

    my $foodata = $c->forward('/foo');
    $c->forward('index');
    $c->forward(qw/Model::DBIC::Foo do_stuff/);
    $c->forward('View::TT');

Note that L<< forward|/"$c->forward( $action [, \@arguments ] )" >> implies
an C<< eval { } >> around the call (actually
L<< execute|/"$c->execute( $class, $coderef )" >> does), thus rendering all
exceptions thrown by the called action non-fatal and pushing them onto
$c->error instead. If you want C<die> to propagate you need to do something
like:

    $c->forward('foo');
    die join "\n", @{ $c->error } if @{ $c->error };

Or make sure to always return true values from your actions and write
your code like this:

    $c->forward('foo') || return;

Another note is that C<< $c->forward >> always returns a scalar because it
actually returns $c->state which operates in a scalar context.
Thus, something like:

    return @array;

in an action that is forwarded to is going to return a scalar,
i.e. how many items are in that array, which is probably not what you want.
If you need to return an array then return a reference to it,
or stash it like so:

    $c->stash->{array} = \@array;

and access it from the stash.

Keep in mind that the C<end> method used is that of the caller action. So a C<< $c->detach >> inside a forwarded action would run the C<end> method from the original action requested.

If you call c<forward> with the name of a component class or instance, rather than an action name
or instance, we invoke the C<process> action on that class or instance, or whatever action you
specific via the second argument $method.

=cut

sub forward { my $c = shift; no warnings 'recursion'; $c->dispatcher->forward( $c, @_ ) }

=head2 $c->detach( $action [, \@arguments ] )

=head2 $c->detach( $class, $method, [, \@arguments ] )

=head2 $c->detach()

The same as L<< forward|/"$c->forward( $action [, \@arguments ] )" >>, but
doesn't return to the previous action when processing is finished.

When called with no arguments it escapes the processing chain entirely.

=cut

sub detach { my $c = shift; $c->dispatcher->detach( $c, @_ ) }

=head2 $c->visit( $action [, \@arguments ] )

=head2 $c->visit( $action [, \@captures, \@arguments ] )

=head2 $c->visit( $class, $method, [, \@arguments ] )

=head2 $c->visit( $class, $method, [, \@captures, \@arguments ] )

Almost the same as L<< forward|/"$c->forward( $action [, \@arguments ] )" >>,
but does a full dispatch, instead of just calling the new C<$action> /
C<< $class->$method >>. This means that C<begin>, C<auto> and the method
you go to are called, just like a new request.

In addition both C<< $c->action >> and C<< $c->namespace >> are localized.
This means, for example, that C<< $c->action >> methods such as
L<name|Catalyst::Action/name>, L<class|Catalyst::Action/class> and
L<reverse|Catalyst::Action/reverse> return information for the visited action
when they are invoked within the visited action.  This is different from the
behavior of L<< forward|/"$c->forward( $action [, \@arguments ] )" >>, which
continues to use the $c->action object from the caller action even when
invoked from the called action.

C<< $c->stash >> is kept unchanged.

In effect, L<< visit|/"$c->visit( $action [, \@captures, \@arguments ] )" >>
allows you to "wrap" another action, just as it would have been called by
dispatching from a URL, while the analogous
L<< go|/"$c->go( $action [, \@captures, \@arguments ] )" >> allows you to
transfer control to another action as if it had been reached directly from a URL.

=cut

sub visit { my $c = shift; $c->dispatcher->visit( $c, @_ ) }

=head2 $c->go( $action [, \@arguments ] )

=head2 $c->go( $action [, \@captures, \@arguments ] )

=head2 $c->go( $class, $method, [, \@arguments ] )

=head2 $c->go( $class, $method, [, \@captures, \@arguments ] )

The relationship between C<go> and
L<< visit|/"$c->visit( $action [, \@captures, \@arguments ] )" >> is the same as
the relationship between
L<< forward|/"$c->forward( $class, $method, [, \@arguments ] )" >> and
L<< detach|/"$c->detach( $action [, \@arguments ] )" >>. Like C<< $c->visit >>,
C<< $c->go >> will perform a full dispatch on the specified action or method,
with localized C<< $c->action >> and C<< $c->namespace >>. Like C<detach>,
C<go> escapes the processing of the current request chain on completion, and
does not return to its caller.

@arguments are arguments to the final destination of $action. @captures are
arguments to the intermediate steps, if any, on the way to the final sub of
$action.

=cut

sub go { my $c = shift; $c->dispatcher->go( $c, @_ ) }

=head2 $c->response

=head2 $c->res

Returns the current L<Catalyst::Response> object, see there for details.

There is a predicate method C<has_response> that returns true if the
request object has been created.  This is something you might need to
check if you are writing plugins that run before a request is finalized.

=head2 $c->stash

Returns a hashref to the stash, which may be used to store data and pass
it between components during a request. You can also set hash keys by
passing arguments. The stash is automatically sent to the view. The
stash is cleared at the end of a request; it cannot be used for
persistent storage (for this you must use a session; see
L<Catalyst::Plugin::Session> for a complete system integrated with
Catalyst).

    $c->stash->{foo} = $bar;
    $c->stash( { moose => 'majestic', qux => 0 } );
    $c->stash( bar => 1, gorch => 2 ); # equivalent to passing a hashref

    # stash is automatically passed to the view for use in a template
    $c->forward( 'MyApp::View::TT' );

The stash hash is currently stored in the PSGI C<$env> and is managed by
L<Catalyst::Middleware::Stash>.  Since it's part of the C<$env> items in
the stash can be accessed in sub applications mounted under your main
L<Catalyst> application.  For example if you delegate the response of an
action to another L<Catalyst> application, that sub application will have
access to all the stash keys of the main one, and if can of course add
more keys of its own.  However those new keys will not 'bubble' back up
to the main application.

For more information the best thing to do is to review the test case:
t/middleware-stash.t in the distribution /t directory.

=cut

sub stash {
  my $c = shift;
  $c->log->error("You are requesting the stash but you don't have a context") unless blessed $c;
  return Catalyst::Middleware::Stash::get_stash($c->req->env)->(@_);
}

=head2 $c->error

=head2 $c->error($error, ...)

=head2 $c->error($arrayref)

Returns an arrayref containing error messages.  If Catalyst encounters an
error while processing a request, it stores the error in $c->error.  This
method should only be used to store fatal error messages.

    my @error = @{ $c->error };

Add a new error.

    $c->error('Something bad happened');

Calling this will always return an arrayref (if there are no errors it
will be an empty arrayref.

=cut

sub error {
    my $c = shift;
    if ( $_[0] ) {
        my $error = ref $_[0] eq 'ARRAY' ? $_[0] : [@_];
        croak @$error unless ref $c;
        push @{ $c->{error} }, @$error;
    }
    elsif ( defined $_[0] ) { $c->{error} = undef }
    return $c->{error} || [];
}

=head2 $c->state

Contains the return value of the last executed action.
Note that << $c->state >> operates in a scalar context which means that all
values it returns are scalar.

Please note that if an action throws an exception, the value of state
should no longer be considered the return if the last action.  It is generally
going to be 0, which indicates an error state.  Examine $c->error for error
details.

=head2 $c->clear_errors

Clear errors.  You probably don't want to clear the errors unless you are
implementing a custom error screen.

This is equivalent to running

    $c->error(0);

=cut

sub clear_errors {
    my $c = shift;
    $c->error(0);
}

=head2 $c->has_errors

Returns true if you have errors

=cut

sub has_errors { scalar(@{shift->error}) ? 1:0 }

=head2 $c->last_error

Returns the most recent error in the stack (the one most recently added...)
or nothing if there are no errors.  This does not modify the contents of the
error stack.

=cut

sub last_error {
  my (@errs) = @{shift->error};
  return scalar(@errs) ? $errs[-1]: undef;
}

=head2 shift_errors

shifts the most recently added error off the error stack and returns it.  Returns
nothing if there are no more errors.

=cut

sub shift_errors {
    my ($self) = @_;
    my @errors = @{$self->error};
    my $err = shift(@errors);
    $self->{error} = \@errors;
    return $err;
}

=head2 pop_errors

pops the most recently added error off the error stack and returns it.  Returns
nothing if there are no more errors.

=cut

sub pop_errors {
    my ($self) = @_;
    my @errors = @{$self->error};
    my $err = pop(@errors);
    $self->{error} = \@errors;
    return $err;
}

sub _comp_search_prefixes {
    my $c = shift;
    return map $c->components->{ $_ }, $c->_comp_names_search_prefixes(@_);
}

# search components given a name and some prefixes
sub _comp_names_search_prefixes {
    my ( $c, $name, @prefixes ) = @_;
    my $appclass = ref $c || $c;
    my $filter   = "^${appclass}::(" . join( '|', @prefixes ) . ')::';
    $filter = qr/$filter/; # Compile regex now rather than once per loop

    # map the original component name to the sub part that we will search against
    my %eligible = map { my $n = $_; $n =~ s{^$appclass\::[^:]+::}{}; $_ => $n; }
        grep { /$filter/ } keys %{ $c->components };

    # undef for a name will return all
    return keys %eligible if !defined $name;

    my $query  = $name->$_isa('Regexp') ? $name : qr/^$name$/i;
    my @result = grep { $eligible{$_} =~ m{$query} } keys %eligible;

    return @result if @result;

    # if we were given a regexp to search against, we're done.
    return if $name->$_isa('Regexp');

    # skip regexp fallback if configured
    return
        if $appclass->config->{disable_component_resolution_regex_fallback};

    # regexp fallback
    $query  = qr/$name/i;
    @result = grep { $eligible{ $_ } =~ m{$query} } keys %eligible;

    # no results? try against full names
    if( !@result ) {
        @result = grep { m{$query} } keys %eligible;
    }

    # don't warn if we didn't find any results, it just might not exist
    if( @result ) {
        # Disgusting hack to work out correct method name
        my $warn_for = lc $prefixes[0];
        my $msg = "Used regexp fallback for \$c->${warn_for}('${name}'), which found '" .
           (join '", "', @result) . "'. Relying on regexp fallback behavior for " .
           "component resolution is unreliable and unsafe.";
        my $short = $result[0];
        # remove the component namespace prefix
        $short =~ s/.*?(Model|Controller|View):://;
        my $shortmess = Carp::shortmess('');
        if ($shortmess =~ m#Catalyst/Plugin#) {
           $msg .= " You probably need to set '$short' instead of '${name}' in this " .
              "plugin's config";
        } elsif ($shortmess =~ m#Catalyst/lib/(View|Controller)#) {
           $msg .= " You probably need to set '$short' instead of '${name}' in this " .
              "component's config";
        } else {
           $msg .= " You probably meant \$c->${warn_for}('$short') instead of \$c->${warn_for}('${name}'), " .
              "but if you really wanted to search, pass in a regexp as the argument " .
              "like so: \$c->${warn_for}(qr/${name}/)";
        }
        $c->log->warn( "${msg}$shortmess" );
    }

    return @result;
}

# Find possible names for a prefix
sub _comp_names {
    my ( $c, @prefixes ) = @_;
    my $appclass = ref $c || $c;

    my $filter = "^${appclass}::(" . join( '|', @prefixes ) . ')::';

    my @names = map { s{$filter}{}; $_; }
        $c->_comp_names_search_prefixes( undef, @prefixes );

    return @names;
}

# Filter a component before returning by calling ACCEPT_CONTEXT if available

sub _filter_component {
    my ( $c, $comp, @args ) = @_;

    if(ref $comp eq 'CODE') {
      $comp = $comp->();
    }

    if ( eval { $comp->can('ACCEPT_CONTEXT'); } ) {
      return $comp->ACCEPT_CONTEXT( $c, @args );
    }

    $c->log->warn("You called component '${\$comp->catalyst_component_name}' with arguments [@args], but this component does not ACCEPT_CONTEXT, so args are ignored.") if scalar(@args) && $c->debug;

    return $comp;
}

=head2 COMPONENT ACCESSORS

=head2 $c->controller($name)

Gets a L<Catalyst::Controller> instance by name.

    $c->controller('Foo')->do_stuff;

If the name is omitted, will return the controller for the dispatched
action.

If you want to search for controllers, pass in a regexp as the argument.

    # find all controllers that start with Foo
    my @foo_controllers = $c->controller(qr{^Foo});


=cut

sub controller {
    my ( $c, $name, @args ) = @_;

    my $appclass = ref($c) || $c;
    if( $name ) {
        unless ( $name->$_isa('Regexp') ) { # Direct component hash lookup to avoid costly regexps
            my $comps = $c->components;
            my $check = $appclass."::Controller::".$name;
            return $c->_filter_component( $comps->{$check}, @args ) if exists $comps->{$check};
            foreach my $path (@{$appclass->config->{ setup_components }->{ search_extra }}) {
                next unless $path =~ /.*::Controller/;
                $check = $path."::".$name;
                return $c->_filter_component( $comps->{$check}, @args ) if exists $comps->{$check};
            }
        }
        my @result = $c->_comp_search_prefixes( $name, qw/Controller C/ );
        return map { $c->_filter_component( $_, @args ) } @result if ref $name;
        return $c->_filter_component( $result[ 0 ], @args );
    }

    return $c->component( $c->action->class );
}

=head2 $c->model($name)

Gets a L<Catalyst::Model> instance by name.

    $c->model('Foo')->do_stuff;

Any extra arguments are directly passed to ACCEPT_CONTEXT, if the model
defines ACCEPT_CONTEXT.  If it does not, the args are discarded.

If the name is omitted, it will look for
 - a model object in $c->stash->{current_model_instance}, then
 - a model name in $c->stash->{current_model}, then
 - a config setting 'default_model', or
 - check if there is only one model, and return it if that's the case.

If you want to search for models, pass in a regexp as the argument.

    # find all models that start with Foo
    my @foo_models = $c->model(qr{^Foo});

=cut

sub model {
    my ( $c, $name, @args ) = @_;
    my $appclass = ref($c) || $c;
    if( $name ) {
        unless ( $name->$_isa('Regexp') ) { # Direct component hash lookup to avoid costly regexps
            my $comps = $c->components;
            my $check = $appclass."::Model::".$name;
            return $c->_filter_component( $comps->{$check}, @args ) if exists $comps->{$check};
            foreach my $path (@{$appclass->config->{ setup_components }->{ search_extra }}) {
                next unless $path =~ /.*::Model/;
                $check = $path."::".$name;
                return $c->_filter_component( $comps->{$check}, @args ) if exists $comps->{$check};
            }
        }
        my @result = $c->_comp_search_prefixes( $name, qw/Model M/ );
        return map { $c->_filter_component( $_, @args ) } @result if ref $name;
        return $c->_filter_component( $result[ 0 ], @args );
    }

    if (ref $c) {
        return $c->stash->{current_model_instance}
          if $c->stash->{current_model_instance};
        return $c->model( $c->stash->{current_model} )
          if $c->stash->{current_model};
    }
    return $c->model( $appclass->config->{default_model} )
      if $appclass->config->{default_model};

    my( $comp, $rest ) = $c->_comp_search_prefixes( undef, qw/Model M/);

    if( $rest ) {
        $c->log->warn( Carp::shortmess('Calling $c->model() will return a random model unless you specify one of:') );
        $c->log->warn( '* $c->config(default_model => "the name of the default model to use")' );
        $c->log->warn( '* $c->stash->{current_model} # the name of the model to use for this request' );
        $c->log->warn( '* $c->stash->{current_model_instance} # the instance of the model to use for this request' );
        $c->log->warn( 'NB: in version 5.81, the "random" behavior will not work at all.' );
    }

    return $c->_filter_component( $comp );
}


=head2 $c->view($name)

Gets a L<Catalyst::View> instance by name.

    $c->view('Foo')->do_stuff;

Any extra arguments are directly passed to ACCEPT_CONTEXT.

If the name is omitted, it will look for
 - a view object in $c->stash->{current_view_instance}, then
 - a view name in $c->stash->{current_view}, then
 - a config setting 'default_view', or
 - check if there is only one view, and return it if that's the case.

If you want to search for views, pass in a regexp as the argument.

    # find all views that start with Foo
    my @foo_views = $c->view(qr{^Foo});

=cut

sub view {
    my ( $c, $name, @args ) = @_;

    my $appclass = ref($c) || $c;
    if( $name ) {
        unless ( $name->$_isa('Regexp') ) { # Direct component hash lookup to avoid costly regexps
            my $comps = $c->components;
            my $check = $appclass."::View::".$name;
            if( exists $comps->{$check} ) {
                return $c->_filter_component( $comps->{$check}, @args );
            }
            else {
                $c->log->warn( "Attempted to use view '$check', but does not exist" );
            }
            foreach my $path (@{$appclass->config->{ setup_components }->{ search_extra }}) {
                next unless $path =~ /.*::View/;
                $check = $path."::".$name;
                return $c->_filter_component( $comps->{$check}, @args ) if exists $comps->{$check};
            }
        }
        my @result = $c->_comp_search_prefixes( $name, qw/View V/ );
        return map { $c->_filter_component( $_, @args ) } @result if ref $name;
        return $c->_filter_component( $result[ 0 ], @args );
    }

    if (ref $c) {
        return $c->stash->{current_view_instance}
          if $c->stash->{current_view_instance};
        return $c->view( $c->stash->{current_view} )
          if $c->stash->{current_view};
    }
    return $c->view( $appclass->config->{default_view} )
      if $appclass->config->{default_view};

    my( $comp, $rest ) = $c->_comp_search_prefixes( undef, qw/View V/);

    if( $rest ) {
        $c->log->warn( 'Calling $c->view() will return a random view unless you specify one of:' );
        $c->log->warn( '* $c->config(default_view => "the name of the default view to use")' );
        $c->log->warn( '* $c->stash->{current_view} # the name of the view to use for this request' );
        $c->log->warn( '* $c->stash->{current_view_instance} # the instance of the view to use for this request' );
        $c->log->warn( 'NB: in version 5.81, the "random" behavior will not work at all.' );
    }

    return $c->_filter_component( $comp );
}

=head2 $c->controllers

Returns the available names which can be passed to $c->controller

=cut

sub controllers {
    my ( $c ) = @_;
    return $c->_comp_names(qw/Controller C/);
}

=head2 $c->models

Returns the available names which can be passed to $c->model

=cut

sub models {
    my ( $c ) = @_;
    return $c->_comp_names(qw/Model M/);
}


=head2 $c->views

Returns the available names which can be passed to $c->view

=cut

sub views {
    my ( $c ) = @_;
    return $c->_comp_names(qw/View V/);
}

=head2 $c->comp($name)

=head2 $c->component($name)

Gets a component object by name. This method is not recommended,
unless you want to get a specific component by full
class. C<< $c->controller >>, C<< $c->model >>, and C<< $c->view >>
should be used instead.

If C<$name> is a regexp, a list of components matched against the full
component name will be returned.

If Catalyst can't find a component by name, it will fallback to regex
matching by default. To disable this behaviour set
disable_component_resolution_regex_fallback to a true value.

    __PACKAGE__->config( disable_component_resolution_regex_fallback => 1 );

=cut

sub component {
    my ( $c, $name, @args ) = @_;

    if( $name ) {
        my $comps = $c->components;

        if( !ref $name ) {
            # is it the exact name?
            return $c->_filter_component( $comps->{ $name }, @args )
                       if exists $comps->{ $name };

            # perhaps we just omitted "MyApp"?
            my $composed = ( ref $c || $c ) . "::${name}";
            return $c->_filter_component( $comps->{ $composed }, @args )
                       if exists $comps->{ $composed };

            # search all of the models, views and controllers
            my( $comp ) = $c->_comp_search_prefixes( $name, qw/Model M Controller C View V/ );
            return $c->_filter_component( $comp, @args ) if $comp;
        }

        return
            if $c->config->{disable_component_resolution_regex_fallback};

        # This is here so $c->comp( '::M::' ) works
        my $query = ref $name ? $name : qr{$name}i;

        my @result = grep { m{$query} } keys %{ $c->components };
        return map { $c->_filter_component( $_, @args ) } @result if ref $name;

        if( $result[ 0 ] ) {
            $c->log->warn( Carp::shortmess(qq(Found results for "${name}" using regexp fallback)) );
            $c->log->warn( 'Relying on the regexp fallback behavior for component resolution' );
            $c->log->warn( 'is unreliable and unsafe. You have been warned' );
            return $c->_filter_component( $result[ 0 ], @args );
        }

        # I would expect to return an empty list here, but that breaks back-compat
    }

    # fallback
    return sort keys %{ $c->components };
}

=head2 CLASS DATA AND HELPER CLASSES

=head2 $c->config

Returns or takes a hashref containing the application's configuration.

    __PACKAGE__->config( { db => 'dsn:SQLite:foo.db' } );

You can also use a C<YAML>, C<XML> or L<Config::General> config file
like C<myapp.conf> in your applications home directory. See
L<Catalyst::Plugin::ConfigLoader>.

=head3 Cascading configuration

The config method is present on all Catalyst components, and configuration
will be merged when an application is started. Configuration loaded with
L<Catalyst::Plugin::ConfigLoader> takes precedence over other configuration,
followed by configuration in your top level C<MyApp> class. These two
configurations are merged, and then configuration data whose hash key matches a
component name is merged with configuration for that component.

The configuration for a component is then passed to the C<new> method when a
component is constructed.

For example:

    MyApp->config({ 'Model::Foo' => { bar => 'baz', overrides => 'me' } });
    MyApp::Model::Foo->config({ quux => 'frob', overrides => 'this' });

will mean that C<MyApp::Model::Foo> receives the following data when
constructed:

    MyApp::Model::Foo->new({
        bar => 'baz',
        quux => 'frob',
        overrides => 'me',
    });

It's common practice to use a Moose attribute
on the receiving component to access the config value.

    package MyApp::Model::Foo;

    use Moose;

    # this attr will receive 'baz' at construction time
    has 'bar' => (
        is  => 'rw',
        isa => 'Str',
    );

You can then get the value 'baz' by calling $c->model('Foo')->bar
(or $self->bar inside code in the model).

B<NOTE:> you MUST NOT call C<< $self->config >> or C<< __PACKAGE__->config >>
as a way of reading config within your code, as this B<will not> give you the
correctly merged config back. You B<MUST> take the config values supplied to
the constructor and use those instead.

=cut

around config => sub {
    my $orig = shift;
    my $c = shift;

    croak('Setting config after setup has been run is not allowed.')
        if ( @_ and $c->setup_finished );

    $c->$orig(@_);
};

=head2 $c->log

Returns the logging object instance. Unless it is already set, Catalyst
sets this up with a L<Catalyst::Log> object. To use your own log class,
set the logger with the C<< __PACKAGE__->log >> method prior to calling
C<< __PACKAGE__->setup >>.

 __PACKAGE__->log( MyLogger->new );
 __PACKAGE__->setup;

And later:

    $c->log->info( 'Now logging with my own logger!' );

Your log class should implement the methods described in
L<Catalyst::Log>.

=head2 has_encoding

Returned True if there's a valid encoding

=head2 clear_encoding

Clears the encoding for the current context

=head2 encoding

Sets or gets the application encoding.  Setting encoding takes either an
Encoding object or a string that we try to resolve via L<Encode::find_encoding>.

You would expect to get the encoding object back if you attempt to set it.  If
there is a failure you will get undef returned and an error message in the log.

=cut

sub has_encoding { shift->encoding ? 1:0 }

sub clear_encoding {
    my $c = shift;
    if(blessed $c) {
        $c->encoding(undef);
    } else {
        $c->log->error("You can't clear encoding on the application");
    }
}

sub encoding {
    my $c = shift;
    my $encoding;

    if ( scalar @_ ) {

        # Don't let one change this once we are too far into the response
        if(blessed $c && $c->res->finalized_headers) {
          Carp::croak("You may not change the encoding once the headers are finalized");
          return;
        }

        # Let it be set to undef
        if (my $wanted = shift)  {
            $encoding = Encode::find_encoding($wanted)
              or Carp::croak( qq/Unknown encoding '$wanted'/ );
            binmode(STDERR, ':encoding(' . $encoding->name . ')');
        }
        else {
            binmode(STDERR);
        }

        $encoding = ref $c
                  ? $c->{encoding} = $encoding
                  : $c->_encoding($encoding);
    } else {
      $encoding = ref $c && exists $c->{encoding}
                ? $c->{encoding}
                : $c->_encoding;
    }

    return $encoding;
}

=head2 $c->debug

Returns 1 if debug mode is enabled, 0 otherwise.

You can enable debug mode in several ways:

=over

=item By calling myapp_server.pl with the -d flag

=item With the environment variables MYAPP_DEBUG, or CATALYST_DEBUG

=item The -Debug option in your MyApp.pm

=item By declaring C<sub debug { 1 }> in your MyApp.pm.

=back

The first three also set the log level to 'debug'.

Calling C<< $c->debug(1) >> has no effect.

=cut

sub debug { 0 }

=head2 $c->dispatcher

Returns the dispatcher instance. See L<Catalyst::Dispatcher>.

=head2 $c->engine

Returns the engine instance. See L<Catalyst::Engine>.


=head2 UTILITY METHODS

=head2 $c->path_to(@path)

Merges C<@path> with C<< $c->config->{home} >> and returns a
L<Path::Class::Dir> object. Note you can usually use this object as
a filename, but sometimes you will have to explicitly stringify it
yourself by calling the C<< ->stringify >> method.

For example:

    $c->path_to( 'db', 'sqlite.db' );

=cut

sub path_to {
    my ( $c, @path ) = @_;
    my $path = Path::Class::Dir->new( $c->config->{home}, @path );
    if ( -d $path ) { return $path }
    else { return Path::Class::File->new( $c->config->{home}, @path ) }
}

sub plugin {
    my ( $class, $name, $plugin, @args ) = @_;

    # See block comment in t/unit_core_plugin.t
    $class->log->warn(qq/Adding plugin using the ->plugin method is deprecated, and will be removed in a future release/);

    $class->_register_plugin( $plugin, 1 );

    eval { $plugin->import };
    $class->mk_classdata($name);
    my $obj;
    eval { $obj = $plugin->new(@args) };

    if ($@) {
        Catalyst::Exception->throw( message =>
              qq/Couldn't instantiate instant plugin "$plugin", "$@"/ );
    }

    $class->$name($obj);
    $class->log->debug(qq/Initialized instant plugin "$plugin" as "$name"/)
      if $class->debug;
}

=head2 MyApp->setup

Initializes the dispatcher and engine, loads any plugins, and loads the
model, view, and controller components. You may also specify an array
of plugins to load here, if you choose to not load them in the C<use
Catalyst> line.

    MyApp->setup;
    MyApp->setup( qw/-Debug/ );

B<Note:> You B<should not> wrap this method with method modifiers
or bad things will happen - wrap the C<setup_finalize> method instead.

B<Note:> You can create a custom setup stage that will execute when the
application is starting.  Use this to customize setup.

    MyApp->setup(-Custom=value);

    sub setup_custom {
      my ($class, $value) = @_;
    }

Can be handy if you want to hook into the setup phase.

=cut

sub setup {
    my ( $class, @arguments ) = @_;
    croak('Running setup more than once')
        if ( $class->setup_finished );

    unless ( $class->isa('Catalyst') ) {

        Catalyst::Exception->throw(
            message => qq/'$class' does not inherit from Catalyst/ );
    }

    if ( $class->arguments ) {
        @arguments = ( @arguments, @{ $class->arguments } );
    }

    # Process options
    my $flags = {};

    foreach (@arguments) {

        if (/^-Debug$/) {
            $flags->{log} =
              ( $flags->{log} ) ? 'debug,' . $flags->{log} : 'debug';
        }
        elsif (/^-(\w+)=?(.*)$/) {
            $flags->{ lc $1 } = $2;
        }
        else {
            push @{ $flags->{plugins} }, $_;
        }
    }

    $class->setup_home( delete $flags->{home} );

    $class->setup_log( delete $flags->{log} );
    $class->setup_plugins( delete $flags->{plugins} );

    $class->setup_data_handlers();
    $class->setup_dispatcher( delete $flags->{dispatcher} );
    if (my $engine = delete $flags->{engine}) {
        $class->log->warn("Specifying the engine in ->setup is no longer supported, see Catalyst::Upgrading");
    }
    $class->setup_engine();
    $class->setup_stats( delete $flags->{stats} );

    for my $flag ( sort keys %{$flags} ) {

        if ( my $code = $class->can( 'setup_' . $flag ) ) {
            &$code( $class, delete $flags->{$flag} );
        }
        else {
            $class->log->warn(qq/Unknown flag "$flag"/);
        }
    }

    eval { require Catalyst::Devel; };
    if( !$@ && $ENV{CATALYST_SCRIPT_GEN} && ( $ENV{CATALYST_SCRIPT_GEN} < $Catalyst::Devel::CATALYST_SCRIPT_GEN ) ) {
        $class->log->warn(<<"EOF");
You are running an old script!

  Please update by running (this will overwrite existing files):
    catalyst.pl -force -scripts $class

  or (this will not overwrite existing files):
    catalyst.pl -scripts $class

EOF
    }

    # Call plugins setup, this is stupid and evil.
    # Also screws C3 badly on 5.10, hack to avoid.
    {
        no warnings qw/redefine/;
        local *setup = sub { };
        $class->setup unless $Catalyst::__AM_RESTARTING;
    }

    # If you are expecting configuration info as part of your setup, it needs
    # to get called here and below, since we need the above line to support
    # ConfigLoader based configs.

    $class->setup_encoding();
    $class->setup_middleware();

    # Initialize our data structure
    $class->components( {} );

    $class->setup_components;

    if ( $class->debug ) {
        my @plugins = map { "$_  " . ( $_->VERSION || '' ) } $class->registered_plugins;

        if (@plugins) {
            my $column_width = Catalyst::Utils::term_width() - 6;
            my $t = Text::SimpleTable->new($column_width);
            $t->row($_) for @plugins;
            $class->log->debug( "Loaded plugins:\n" . $t->draw . "\n" );
        }

        my @middleware = map {
          ref $_ eq 'CODE' ?
            "Inline Coderef" :
              (ref($_) .'  '. ($_->can('VERSION') ? $_->VERSION || '' : '')
                || '')  } $class->registered_middlewares;

        if (@middleware) {
            my $column_width = Catalyst::Utils::term_width() - 6;
            my $t = Text::SimpleTable->new($column_width);
            $t->row($_) for @middleware;
            $class->log->debug( "Loaded PSGI Middleware:\n" . $t->draw . "\n" );
        }

        my %dh = $class->registered_data_handlers;
        if (my @data_handlers = keys %dh) {
            my $column_width = Catalyst::Utils::term_width() - 6;
            my $t = Text::SimpleTable->new($column_width);
            $t->row($_) for @data_handlers;
            $class->log->debug( "Loaded Request Data Handlers:\n" . $t->draw . "\n" );
        }

        my $dispatcher = $class->dispatcher;
        my $engine     = $class->engine;
        my $home       = $class->config->{home};

        $class->log->debug(sprintf(q/Loaded dispatcher "%s"/, blessed($dispatcher)));
        $class->log->debug(sprintf(q/Loaded engine "%s"/, blessed($engine)));

        $home
          ? ( -d $home )
          ? $class->log->debug(qq/Found home "$home"/)
          : $class->log->debug(qq/Home "$home" doesn't exist/)
          : $class->log->debug(q/Couldn't find home/);

        my $column_width = Catalyst::Utils::term_width() - 8 - 9;

        my $t = Text::SimpleTable->new( [ $column_width, 'Class' ], [ 8, 'Type' ] );
        for my $comp ( sort keys %{ $class->components } ) {
            my $type = ref $class->components->{$comp} ? 'instance' : 'class';
            $t->row( $comp, $type );
        }
        $class->log->debug( "Loaded components:\n" . $t->draw . "\n" )
          if ( keys %{ $class->components } );
    }

    # Add our self to components, since we are also a component
    if( $class->isa('Catalyst::Controller') ){
      $class->components->{$class} = $class;
    }

    $class->setup_actions;

    if ( $class->debug ) {
        my $name = $class->config->{name} || 'Application';
        $class->log->info("$name powered by Catalyst $Catalyst::VERSION");
    }

    if ($class->config->{case_sensitive}) {
        $class->log->warn($class . "->config->{case_sensitive} is set.");
        $class->log->warn("This setting is deprecated and planned to be removed in Catalyst 5.81.");
    }

    # call these so we pre setup the composed classes
    $class->composed_request_class;
    $class->composed_response_class;
    $class->composed_stats_class;

    $class->setup_finalize;

    # Flush the log for good measure (in case something turned off 'autoflush' early)
    $class->log->_flush() if $class->log->can('_flush');

    return $class || 1; # Just in case someone named their Application 0...
}

=head2 $app->setup_finalize

A hook to attach modifiers to. This method does not do anything except set the
C<setup_finished> accessor.

Applying method modifiers to the C<setup> method doesn't work, because of quirky things done for plugin setup.

Example:

    after setup_finalize => sub {
        my $app = shift;

        ## do stuff here..
    };

=cut

sub setup_finalize {
    my ($class) = @_;
    $class->setup_finished(1);
}

=head2 $c->uri_for( $path?, @args?, \%query_values?, \$fragment? )

=head2 $c->uri_for( $action, \@captures?, @args?, \%query_values?, \$fragment? )

=head2 $c->uri_for( $action, [@captures, @args], \%query_values?, \$fragment? )

Constructs an absolute L<URI> object based on the application root, the
provided path, and the additional arguments and query parameters provided.
When used as a string, provides a textual URI.  If you need more flexibility
than this (i.e. the option to provide relative URIs etc.) see
L<Catalyst::Plugin::SmartURI>.

If no arguments are provided, the URI for the current action is returned.
To return the current action and also provide @args, use
C<< $c->uri_for( $c->action, @args ) >>.

If the first argument is a string, it is taken as a public URI path relative
to C<< $c->namespace >> (if it doesn't begin with a forward slash) or
relative to the application root (if it does). It is then merged with
C<< $c->request->base >>; any C<@args> are appended as additional path
components; and any C<%query_values> are appended as C<?foo=bar> parameters.

B<NOTE> If you are using this 'stringy' first argument, we skip encoding and
allow you to declare something like:

    $c->uri_for('/foo/bar#baz')

Where 'baz' is a URI fragment.  We consider this first argument string to be
'expert' mode where you are expected to create a valid URL and we for the most
part just pass it through without a lot of internal effort to escape and encode.

If the first argument is a L<Catalyst::Action> it represents an action which
will have its path resolved using C<< $c->dispatcher->uri_for_action >>. The
optional C<\@captures> argument (an arrayref) allows passing the captured
variables that are needed to fill in the paths of Chained and Regex actions;
once the path is resolved, C<uri_for> continues as though a path was
provided, appending any arguments or parameters and creating an absolute
URI.

The captures for the current request can be found in
C<< $c->request->captures >>, and actions can be resolved using
C<< Catalyst::Controller->action_for($name) >>. If you have a private action
path, use C<< $c->uri_for_action >> instead.

  # Equivalent to $c->req->uri
  $c->uri_for($c->action, $c->req->captures,
      @{ $c->req->args }, $c->req->params);

  # For the Foo action in the Bar controller
  $c->uri_for($c->controller('Bar')->action_for('Foo'));

  # Path to a static resource
  $c->uri_for('/static/images/logo.png');

In general the scheme of the generated URI object will follow the incoming request
however if your targeted action or action chain has the Scheme attribute it will
use that instead.

Also, if the targeted Action or Action chain declares Args/CaptureArgs that have
type constraints, we will require that your proposed URL verify on those declared
constraints.

=cut

sub uri_for {
    my ( $c, $path, @args ) = @_;

    if ( $path->$_isa('Catalyst::Controller') ) {
        $path = $path->path_prefix;
        $path =~ s{/+\z}{};
        $path .= '/';
    }

    my $fragment =  ((scalar(@args) && ref($args[-1]) eq 'SCALAR') ? ${pop @args} : undef );

    unless(blessed $path) {
      if (defined($path) and $path =~ s/#(.+)$//)  {
        if(defined($1) and defined $fragment) {
          carp "Abiguious fragment declaration: You cannot define a fragment in '$path' and as an argument '$fragment'";
        }
        if(defined($1)) {
          $fragment = $1;
        }
      }
    }

    my $params =
      ( scalar @args && ref $args[$#args] eq 'HASH' ? pop @args : {} );

    undef($path) if (defined $path && $path eq '');

    carp "uri_for called with undef argument" if grep { ! defined $_ } @args;

    my $target_action = $path->$_isa('Catalyst::Action') ? $path : undef;
    if ( $path->$_isa('Catalyst::Action') ) { # action object
        s|/|%2F|g for @args;
        my $captures = [ map { s|/|%2F|g; $_; }
                        ( scalar @args && ref $args[0] eq 'ARRAY'
                         ? @{ shift(@args) }
                         : ()) ];

        my $action = $path;
        my $expanded_action = $c->dispatcher->expand_action( $action );
        my $num_captures = $expanded_action->number_of_captures;

        # ->uri_for( $action, \@captures_and_args, \%query_values? )
        if( !@args && $action->number_of_args && @$captures > $num_captures ) {
          unshift @args, splice @$captures, $num_captures;
        }

        if($num_captures) {
          unless($expanded_action->match_captures_constraints($c, $captures)) {
            $c->log->debug("captures [@{$captures}] do not match the type constraints in actionchain ending with '$expanded_action'")
                if $c->debug;
            return undef;
          }
        }

        $path = $c->dispatcher->uri_for_action($action, $captures);
        if (not defined $path) {
            $c->log->debug(qq/Can't find uri_for action '$action' @$captures/)
                if $c->debug;
            return undef;
        }
        $path = '/' if $path eq '';

        # At this point @encoded_args is the remaining Args (all captures removed).
        if($expanded_action->has_args_constraints) {
          unless($expanded_action->match_args($c,\@args)) {
             $c->log->debug("args [@args] do not match the type constraints in action '$expanded_action'")
                if $c->debug;
             return undef;
          }
        }
    }

    unshift(@args, $path);

    unless (defined $path && $path =~ s!^/!!) { # in-place strip
        my $namespace = $c->namespace;
        if (defined $path) { # cheesy hack to handle path '../foo'
           $namespace =~ s{(?:^|/)[^/]+$}{} while $args[0] =~ s{^\.\./}{};
        }
        unshift(@args, $namespace || '');
    }

    # join args with '/', or a blank string
    my $args = join('/', grep { defined($_) } @args);
    $args =~ s/\?/%3F/g; # STUPID STUPID SPECIAL CASE
    $args =~ s!^/+!!;

    my ($base, $class) = ('/', 'URI::_generic');
    if(blessed($c)) {
      $base = $c->req->base;
      if($target_action) {
        $target_action = $c->dispatcher->expand_action($target_action);
        if(my $s = $target_action->scheme) {
          $s = lc($s);
          $class = "URI::$s";
          $base->scheme($s);
        } else {
          $class = ref($base);
        }
      } else {
        $class = ref($base);
      }

      $base =~ s{(?<!/)$}{/};
    }

    my $query = '';
    if (my @keys = keys %$params) {
      # somewhat lifted from URI::_query's query_form
      $query = '?'.join('&', map {
          my $val = $params->{$_};
          my $key = encode_utf8($_);
          # using the URI::Escape pattern here so utf8 chars survive
          $key =~ s/([^A-Za-z0-9\-_.!~*'() ])/$URI::Escape::escapes{$1}/go;
          $key =~ s/ /+/g;

          $val = '' unless defined $val;
          (map {
              my $param = encode_utf8($_);
              # using the URI::Escape pattern here so utf8 chars survive
              $param =~ s/([^A-Za-z0-9\-_.!~*'() ])/$URI::Escape::escapes{$1}/go;
              $param =~ s/ /+/g;

              "${key}=$param";
          } ( ref $val eq 'ARRAY' ? @$val : $val ));
      } @keys);
    }

    $base = encode_utf8 $base;
    $base =~ s/([^$URI::uric])/$URI::Escape::escapes{$1}/go;
    $args = encode_utf8 $args;
    $args =~ s/([^$URI::uric])/$URI::Escape::escapes{$1}/go;

    if(defined $fragment) {
      if(blessed $path) {
        $fragment = encode_utf8($fragment);
        $fragment =~ s/([^A-Za-z0-9\-_.!~*'() ])/$URI::Escape::escapes{$1}/go;
        $fragment =~ s/ /+/g;
      }
      $query .= "#$fragment";
    }

    my $res = bless(\"${base}${args}${query}", $class);
    $res;
}

=head2 $c->uri_for_action( $path, \@captures_and_args?, @args?, \%query_values? )

=head2 $c->uri_for_action( $action, \@captures_and_args?, @args?, \%query_values? )

=over

=item $path

A private path to the Catalyst action you want to create a URI for.

This is a shortcut for calling C<< $c->dispatcher->get_action_by_path($path)
>> and passing the resulting C<$action> and the remaining arguments to C<<
$c->uri_for >>.

You can also pass in a Catalyst::Action object, in which case it is passed to
C<< $c->uri_for >>.

Note that although the path looks like a URI that dispatches to the wanted action, it is not a URI, but an internal path to that action.

For example, if the action looks like:

 package MyApp::Controller::Users;

 sub lst : Path('the-list') {}

You can use:

 $c->uri_for_action('/users/lst')

and it will create the URI /users/the-list.

=item \@captures_and_args?

Optional array reference of Captures (i.e. C<CaptureArgs> or C<< $c->req->captures >>)
and arguments to the request. Usually used with L<Catalyst::DispatchType::Chained>
to interpolate all the parameters in the URI.

=item @args?

Optional list of extra arguments - can be supplied in the
C<< \@captures_and_args? >> array ref, or here - whichever is easier for your
code.

Your action can have zero, a fixed or a variable number of args (e.g.
C<< Args(1) >> for a fixed number or C<< Args() >> for a variable number)..

=item \%query_values?

Optional array reference of query parameters to append. E.g.

  { foo => 'bar' }

will generate

  /rest/of/your/uri?foo=bar

=back

=cut

sub uri_for_action {
    my ( $c, $path, @args ) = @_;
    my $action = blessed($path)
      ? $path
      : $c->dispatcher->get_action_by_path($path);
    unless (defined $action) {
      croak "Can't find action for path '$path'";
    }
    return $c->uri_for( $action, @args );
}

=head2 $c->welcome_message

Returns the Catalyst welcome HTML page.

=cut

sub welcome_message {
    my $c      = shift;
    my $name   = $c->config->{name};
    my $logo   = $c->uri_for('/static/images/catalyst_logo.png');
    my $prefix = Catalyst::Utils::appprefix( ref $c );
    $c->response->content_type('text/html; charset=utf-8');
    return <<"EOF";
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
    "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
    <head>
    <meta http-equiv="Content-Language" content="en" />
    <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
        <title>$name on Catalyst $VERSION</title>
        <style type="text/css">
            body {
                color: #000;
                background-color: #eee;
            }
            div#content {
                width: 640px;
                margin-left: auto;
                margin-right: auto;
                margin-top: 10px;
                margin-bottom: 10px;
                text-align: left;
                background-color: #ccc;
                border: 1px solid #aaa;
            }
            p, h1, h2 {
                margin-left: 20px;
                margin-right: 20px;
                font-family: verdana, tahoma, sans-serif;
            }
            a {
                font-family: verdana, tahoma, sans-serif;
            }
            :link, :visited {
                    text-decoration: none;
                    color: #b00;
                    border-bottom: 1px dotted #bbb;
            }
            :link:hover, :visited:hover {
                    color: #555;
            }
            div#topbar {
                margin: 0px;
            }
            pre {
                margin: 10px;
                padding: 8px;
            }
            div#answers {
                padding: 8px;
                margin: 10px;
                background-color: #fff;
                border: 1px solid #aaa;
            }
            h1 {
                font-size: 0.9em;
                font-weight: normal;
                text-align: center;
            }
            h2 {
                font-size: 1.0em;
            }
            p {
                font-size: 0.9em;
            }
            p img {
                float: right;
                margin-left: 10px;
            }
            span#appname {
                font-weight: bold;
                font-size: 1.6em;
            }
        </style>
    </head>
    <body>
        <div id="content">
            <div id="topbar">
                <h1><span id="appname">$name</span> on <a href="http://catalyst.perl.org">Catalyst</a>
                    $VERSION</h1>
             </div>
             <div id="answers">
                 <p>
                 <img src="$logo" alt="Catalyst Logo" />
                 </p>
                 <p>Welcome to the  world of Catalyst.
                    This <a href="http://en.wikipedia.org/wiki/MVC">MVC</a>
                    framework will make web development something you had
                    never expected it to be: Fun, rewarding, and quick.</p>
                 <h2>What to do now?</h2>
                 <p>That really depends  on what <b>you</b> want to do.
                    We do, however, provide you with a few starting points.</p>
                 <p>If you want to jump right into web development with Catalyst
                    you might want to start with a tutorial.</p>
<pre>perldoc <a href="https://metacpan.org/module/Catalyst::Manual::Tutorial">Catalyst::Manual::Tutorial</a></code>
</pre>
<p>Afterwards you can go on to check out a more complete look at our features.</p>
<pre>
<code>perldoc <a href="https://metacpan.org/module/Catalyst::Manual::Intro">Catalyst::Manual::Intro</a>
<!-- Something else should go here, but the Catalyst::Manual link seems unhelpful -->
</code></pre>
                 <h2>What to do next?</h2>
                 <p>Next it's time to write an actual application. Use the
                    helper scripts to generate <a href="https://metacpan.org/search?q=Catalyst%3A%3AController">controllers</a>,
                    <a href="https://metacpan.org/search?q=Catalyst%3A%3AModel">models</a>, and
                    <a href="https://metacpan.org/search?q=Catalyst%3A%3AView">views</a>;
                    they can save you a lot of work.</p>
                    <pre><code>script/${prefix}_create.pl --help</code></pre>
                    <p>Also, be sure to check out the vast and growing
                    collection of <a href="http://search.cpan.org/search?query=Catalyst">plugins for Catalyst on CPAN</a>;
                    you are likely to find what you need there.
                    </p>

                 <h2>Need help?</h2>
                 <p>Catalyst has a very active community. Here are the main places to
                    get in touch with us.</p>
                 <ul>
                     <li>
                         <a href="http://dev.catalyst.perl.org">Wiki</a>
                     </li>
                     <li>
                         <a href="http://lists.scsys.co.uk/cgi-bin/mailman/listinfo/catalyst">Mailing-List</a>
                     </li>
                     <li>
                         <a href="irc://irc.perl.org/catalyst">IRC channel #catalyst on irc.perl.org</a>
                     </li>
                 </ul>
                 <h2>In conclusion</h2>
                 <p>The Catalyst team hopes you will enjoy using Catalyst as much
                    as we enjoyed making it. Please contact us if you have ideas
                    for improvement or other feedback.</p>
             </div>
         </div>
    </body>
</html>
EOF
}

=head2 run_options

Contains a hash of options passed from the application script, including
the original ARGV the script received, the processed values from that
ARGV and any extra arguments to the script which were not processed.

This can be used to add custom options to your application's scripts
and setup your application differently depending on the values of these
options.

=head1 INTERNAL METHODS

These methods are not meant to be used by end users.

=head2 $c->components

Returns a hash of components.

=head2 $c->context_class

Returns or sets the context class.

=head2 $c->counter

Returns a hashref containing coderefs and execution counts (needed for
deep recursion detection).

=head2 $c->depth

Returns the number of actions on the current internal execution stack.

=head2 $c->dispatch

Dispatches a request to actions.

=cut

sub dispatch { my $c = shift; $c->dispatcher->dispatch( $c, @_ ) }

=head2 $c->dispatcher_class

Returns or sets the dispatcher class.

=head2 $c->dump_these

Returns a list of 2-element array references (name, structure) pairs
that will be dumped on the error page in debug mode.

=cut

sub dump_these {
    my $c = shift;
    [ Request => $c->req ],
    [ Response => $c->res ],
    [ Stash => $c->stash ],
    [ Config => $c->config ];
}

=head2 $c->engine_class

Returns or sets the engine class.

=head2 $c->execute( $class, $coderef )

Execute a coderef in given class and catch exceptions. Errors are available
via $c->error.

=cut

sub execute {
    my ( $c, $class, $code ) = @_;
    $class = $c->component($class) || $class;
    #$c->state(0);

    if ( $c->depth >= $RECURSION ) {
        my $action = $code->reverse();
        $action = "/$action" unless $action =~ /->/;
        my $error = qq/Deep recursion detected calling "${action}"/;
        $c->log->error($error);
        $c->error($error);
        $c->state(0);
        return $c->state;
    }

    my $stats_info = $c->_stats_start_execute( $code ) if $c->use_stats;

    push( @{ $c->stack }, $code );

    no warnings 'recursion';
    # N.B. This used to be combined, but I have seen $c get clobbered if so, and
    #      I have no idea how, ergo $ret (which appears to fix the issue)
    eval { my $ret = $code->execute( $class, $c, @{ $c->req->args } ) || 0; $c->state( $ret ) };

    $c->_stats_finish_execute( $stats_info ) if $c->use_stats and $stats_info;

    my $last = pop( @{ $c->stack } );

    if ( my $error = $@ ) {
        #rethow if this can be handled by middleware
        if ( $c->_handle_http_exception($error) ) {
            foreach my $err (@{$c->error}) {
                $c->log->error($err);
            }
            $c->clear_errors;
            $c->log->_flush if $c->log->can('_flush');

            $error->can('rethrow') ? $error->rethrow : croak $error;
        }
        if ( blessed($error) and $error->isa('Catalyst::Exception::Detach') ) {
            $error->rethrow if $c->depth > 1;
        }
        elsif ( blessed($error) and $error->isa('Catalyst::Exception::Go') ) {
            $error->rethrow if $c->depth > 0;
        }
        else {
            unless ( ref $error ) {
                no warnings 'uninitialized';
                chomp $error;
                my $class = $last->class;
                my $name  = $last->name;
                $error = qq/Caught exception in $class->$name "$error"/;
            }
            $c->error($error);
        }
        #$c->state(0);
    }
    return $c->state;
}

sub _stats_start_execute {
    my ( $c, $code ) = @_;
    my $appclass = ref($c) || $c;
    return if ( ( $code->name =~ /^_.*/ )
        && ( !$appclass->config->{show_internal_actions} ) );

    my $action_name = $code->reverse();
    $c->counter->{$action_name}++;

    my $action = $action_name;
    $action = "/$action" unless $action =~ /->/;

    # determine if the call was the result of a forward
    # this is done by walking up the call stack and looking for a calling
    # sub of Catalyst::forward before the eval
    my $callsub = q{};
    for my $index ( 2 .. 11 ) {
        last
        if ( ( caller($index) )[0] eq 'Catalyst'
            && ( caller($index) )[3] eq '(eval)' );

        if ( ( caller($index) )[3] =~ /forward$/ ) {
            $callsub = ( caller($index) )[3];
            $action  = "-> $action";
            last;
        }
    }

    my $uid = $action_name . $c->counter->{$action_name};

    # is this a root-level call or a forwarded call?
    if ( $callsub =~ /forward$/ ) {
        my $parent = $c->stack->[-1];

        # forward, locate the caller
        if ( defined $parent && exists $c->counter->{"$parent"} ) {
            $c->stats->profile(
                begin  => $action,
                parent => "$parent" . $c->counter->{"$parent"},
                uid    => $uid,
            );
        }
        else {

            # forward with no caller may come from a plugin
            $c->stats->profile(
                begin => $action,
                uid   => $uid,
            );
        }
    }
    else {

        # root-level call
        $c->stats->profile(
            begin => $action,
            uid   => $uid,
        );
    }
    return $action;

}

sub _stats_finish_execute {
    my ( $c, $info ) = @_;
    $c->stats->profile( end => $info );
}

=head2 $c->finalize

Finalizes the request.

=cut

sub finalize {
    my $c = shift;

    for my $error ( @{ $c->error } ) {
        $c->log->error($error);
    }

    # Support skipping finalize for psgix.io style 'jailbreak'.  Used to support
    # stuff like cometd and websockets

    if($c->request->_has_io_fh) {
      $c->log_response;
      return;
    }

    # Allow engine to handle finalize flow (for POE)
    my $engine = $c->engine;
    if ( my $code = $engine->can('finalize') ) {
        $engine->$code($c);
    }
    else {

        $c->finalize_uploads;

        # Error
        if ( $#{ $c->error } >= 0 ) {
            $c->finalize_error;
        }

        $c->finalize_encoding;
        $c->finalize_headers unless $c->response->finalized_headers;
        $c->finalize_body;
    }

    $c->log_response;

    $c->log_stats if $c->use_stats;

    return $c->response->status;
}

=head2 $c->log_stats

Logs statistics.

=cut

sub log_stats {
    my $c = shift;

    my $elapsed = $c->stats->elapsed;
    my $av = $elapsed == 0 ? '??' : sprintf '%.3f', 1 / $elapsed;
    $c->log->info(
        "Request took ${elapsed}s ($av/s)\n" . $c->stats->report . "\n" );
}


=head2 $c->finalize_body

Finalizes body.

=cut

sub finalize_body { my $c = shift; $c->engine->finalize_body( $c, @_ ) }

=head2 $c->finalize_cookies

Finalizes cookies.

=cut

sub finalize_cookies { my $c = shift; $c->engine->finalize_cookies( $c, @_ ) }

=head2 $c->finalize_error

Finalizes error.  If there is only one error in L</error> and it is an object that
does C<as_psgi> or C<code> we rethrow the error and presume it caught by middleware
up the ladder.  Otherwise we return the debugging error page (in debug mode) or we
return the default error page (production mode).

=cut

sub finalize_error {
    my $c = shift;
    if($#{$c->error} > 0) {
        $c->engine->finalize_error( $c, @_ );
    } else {
        my ($error) = @{$c->error};
        if ( $c->_handle_http_exception($error) ) {
            # In the case where the error 'knows what it wants', becauses its PSGI
            # aware, just rethow and let middleware catch it
            $error->can('rethrow') ? $error->rethrow : croak $error;
        } else {
            $c->engine->finalize_error( $c, @_ )
        }
    }
}

=head2 $c->finalize_headers

Finalizes headers.

=cut

sub finalize_headers {
    my $c = shift;

    my $response = $c->response; #accessor calls can add up?

    # Check if we already finalized headers
    return if $response->finalized_headers;

    # Handle redirects
    if ( my $location = $response->redirect ) {
        $c->log->debug(qq/Redirecting to "$location"/) if $c->debug;
        $response->header( Location => $location );
    }

    # Remove incorrectly added body and content related meta data when returning
    # an information response, or a response the is required to not include a body

    $c->finalize_cookies;

    # This currently is a NOOP but I don't want to remove it since I guess people
    # might have Response subclasses that use it for something... (JNAP)
    $c->response->finalize_headers();

    # Done
    $response->finalized_headers(1);
}

=head2 $c->finalize_encoding

Make sure your body is encoded properly IF you set an encoding.  By
default the encoding is UTF-8 but you can disable it by explicitly setting the
encoding configuration value to undef.

We can only encode when the body is a scalar.  Methods for encoding via the
streaming interfaces (such as C<write> and C<write_fh> on L<Catalyst::Response>
are available).

See L</ENCODING>.

=cut

sub finalize_encoding {
    my $c = shift;
    my $res = $c->res || return;

    # Warn if the set charset is different from the one you put into encoding.  We need
    # to do this early since encodable_response is false for this condition and we need
    # to match the debug output for backcompat (there's a test for this...) -JNAP
    if(
      $res->content_type_charset and $c->encoding and
      (uc($c->encoding->mime_name) ne uc($res->content_type_charset))
    ) {
        my $ct = lc($res->content_type_charset);
        $c->log->debug("Catalyst encoding config is set to encode in '" .
            $c->encoding->mime_name .
            "', content type is '$ct', not encoding ");
    }

    if(
      ($res->encodable_response) and
      (defined($res->body)) and
      (ref(\$res->body) eq 'SCALAR')
    ) {
        # if you are finding yourself here and your body is already encoded correctly
        # and you want to turn this off, use $c->clear_encoding to prevent encoding
        # at this step, or set encoding to undef in the config to do so for the whole
        # application.  See the ENCODING documentaiton for better notes.
        $c->res->body( $c->encoding->encode( $c->res->body, $c->_encode_check ) );

        # Set the charset if necessary.  This might be a bit bonkers since encodable response
        # is false when the set charset is not the same as the encoding mimetype (maybe
        # confusing action at a distance here..
        # Don't try to set the charset if one already exists or if headers are already finalized
        $c->res->content_type($c->res->content_type . "; charset=" . $c->encoding->mime_name)
          unless($c->res->content_type_charset ||
                ($c->res->_context && $c->res->finalized_headers && !$c->res->_has_response_cb));
    }
}

=head2 $c->finalize_output

An alias for finalize_body.

=head2 $c->finalize_read

Finalizes the input after reading is complete.

=cut

sub finalize_read { my $c = shift; $c->engine->finalize_read( $c, @_ ) }

=head2 $c->finalize_uploads

Finalizes uploads. Cleans up any temporary files.

=cut

sub finalize_uploads { my $c = shift; $c->engine->finalize_uploads( $c, @_ ) }

=head2 $c->get_action( $action, $namespace )

Gets an action in a given namespace.

=cut

sub get_action { my $c = shift; $c->dispatcher->get_action(@_) }

=head2 $c->get_actions( $action, $namespace )

Gets all actions of a given name in a namespace and all parent
namespaces.

=cut

sub get_actions { my $c = shift; $c->dispatcher->get_actions( $c, @_ ) }

=head2 $app->handle_request( @arguments )

Called to handle each HTTP request.

=cut

sub handle_request {
    my ( $class, @arguments ) = @_;

    # Always expect worst case!
    my $status = -1;
    try {
        if ($class->debug) {
            my $secs = time - $START || 1;
            my $av = sprintf '%.3f', $COUNT / $secs;
            my $time = localtime time;
            $class->log->info("*** Request $COUNT ($av/s) [$$] [$time] ***");
        }

        my $c = $class->prepare(@arguments);
        $c->dispatch;
        $status = $c->finalize;
    } catch {
        #rethow if this can be handled by middleware
        if ( $class->_handle_http_exception($_) ) {
            $_->can('rethrow') ? $_->rethrow : croak $_;
        }
        chomp(my $error = $_);
        $class->log->error(qq/Caught exception in engine "$error"/);
    };

    $COUNT++;

    if(my $coderef = $class->log->can('_flush')){
        $class->log->$coderef();
    }
    return $status;
}

=head2 $class->prepare( @arguments )

Creates a Catalyst context from an engine-specific request (Apache, CGI,
etc.).

=cut

has _uploadtmp => (
    is => 'ro',
    predicate => '_has_uploadtmp',
);

sub prepare {
    my ( $class, @arguments ) = @_;

    # XXX
    # After the app/ctxt split, this should become an attribute based on something passed
    # into the application.
    $class->context_class( ref $class || $class ) unless $class->context_class;

    my $uploadtmp = $class->config->{uploadtmp};
    my $c = $class->context_class->new({ $uploadtmp ? (_uploadtmp => $uploadtmp) : ()});

    $c->response->_context($c);
    $c->stats($class->stats_class->new)->enable($c->use_stats);

    if ( $c->debug || $c->config->{enable_catalyst_header} ) {
        $c->res->headers->header( 'X-Catalyst' => $Catalyst::VERSION );
    }

    try {
        # Allow engine to direct the prepare flow (for POE)
        if ( my $prepare = $c->engine->can('prepare') ) {
            $c->engine->$prepare( $c, @arguments );
        }
        else {
            $c->prepare_request(@arguments);
            $c->prepare_connection;
            $c->prepare_query_parameters;
            $c->prepare_headers; # Just hooks, no longer needed - they just
            $c->prepare_cookies; # cause the lazy attribute on req to build
            $c->prepare_path;

            # Prepare the body for reading, either by prepare_body
            # or the user, if they are using $c->read
            $c->prepare_read;

            # Parse the body unless the user wants it on-demand
            unless ( ref($c)->config->{parse_on_demand} ) {
                $c->prepare_body;
            }
        }
        $c->prepare_action;
    }
    # VERY ugly and probably shouldn't rely on ->finalize actually working
    catch {
        # failed prepare is always due to an invalid request, right?
        # Note we call finalize and then die here, which escapes
        # finalize being called in the enclosing block..
        # It in fact couldn't be called, as we don't return $c..
        # This is a mess - but I'm unsure you can fix this without
        # breaking compat for people doing crazy things (we should set
        # the 400 and just return the ctx here IMO, letting finalize get called
        # above...
        if ( $c->_handle_http_exception($_) ) {
            foreach my $err (@{$c->error}) {
                $c->log->error($err);
            }
            $c->clear_errors;
            $c->log->_flush if $c->log->can('_flush');
            $_->can('rethrow') ? $_->rethrow : croak $_;
        } else {
            $c->response->status(400);
            $c->response->content_type('text/plain');
            $c->response->body('Bad Request');
            $c->finalize;
            die $_;
        }
    };

    $c->log_request;
    $c->{stash} = $c->stash;
    Scalar::Util::weaken($c->{stash});

    return $c;
}

=head2 $c->prepare_action

Prepares action. See L<Catalyst::Dispatcher>.

=cut

sub prepare_action {
    my $c = shift;
    my $ret = $c->dispatcher->prepare_action( $c, @_);

    if($c->encoding) {
        foreach (@{$c->req->arguments}, @{$c->req->captures}) {
          $_ = $c->_handle_param_unicode_decoding($_);
        }
    }

    return $ret;
}


=head2 $c->prepare_body

Prepares message body.

=cut

sub prepare_body {
    my $c = shift;

    return if $c->request->_has_body;

    # Initialize on-demand data
    $c->engine->prepare_body( $c, @_ );
    $c->prepare_parameters;
    $c->prepare_uploads;
}

=head2 $c->prepare_body_chunk( $chunk )

Prepares a chunk of data before sending it to L<HTTP::Body>.

See L<Catalyst::Engine>.

=cut

sub prepare_body_chunk {
    my $c = shift;
    $c->engine->prepare_body_chunk( $c, @_ );
}

=head2 $c->prepare_body_parameters

Prepares body parameters.

=cut

sub prepare_body_parameters {
    my $c = shift;
    $c->request->prepare_body_parameters( $c, @_ );
}

=head2 $c->prepare_connection

Prepares connection.

=cut

sub prepare_connection {
    my $c = shift;
    $c->request->prepare_connection($c);
}

=head2 $c->prepare_cookies

Prepares cookies by ensuring that the attribute on the request
object has been built.

=cut

sub prepare_cookies { my $c = shift; $c->request->cookies }

=head2 $c->prepare_headers

Prepares request headers by ensuring that the attribute on the request
object has been built.

=cut

sub prepare_headers { my $c = shift; $c->request->headers }

=head2 $c->prepare_parameters

Prepares parameters.

=cut

sub prepare_parameters {
    my $c = shift;
    $c->prepare_body_parameters;
    $c->engine->prepare_parameters( $c, @_ );
}

=head2 $c->prepare_path

Prepares path and base.

=cut

sub prepare_path { my $c = shift; $c->engine->prepare_path( $c, @_ ) }

=head2 $c->prepare_query_parameters

Prepares query parameters.

=cut

sub prepare_query_parameters {
    my $c = shift;

    $c->engine->prepare_query_parameters( $c, @_ );
}

=head2 $c->log_request

Writes information about the request to the debug logs.  This includes:

=over 4

=item * Request method, path, and remote IP address

=item * Query keywords (see L<Catalyst::Request/query_keywords>)

=item * Request parameters

=item * File uploads

=back

=cut

sub log_request {
    my $c = shift;

    return unless $c->debug;

    my($dump) = grep {$_->[0] eq 'Request' } $c->dump_these;
    my $request = $dump->[1];

    my ( $method, $path, $address ) = ( $request->method, $request->path, $request->address );
    $method ||= '';
    $path = '/' unless length $path;
    $address ||= '';

    $path =~ s/%([0-9A-Fa-f]{2})/chr(hex($1))/eg;
    $path = decode_utf8($path);

    $c->log->debug(qq/"$method" request for "$path" from "$address"/);

    $c->log_request_headers($request->headers);

    if ( my $keywords = $request->query_keywords ) {
        $c->log->debug("Query keywords are: $keywords");
    }

    $c->log_request_parameters( query => $request->query_parameters, $request->_has_body ? (body => $request->body_parameters) : () );

    $c->log_request_uploads($request);
}

=head2 $c->log_response

Writes information about the response to the debug logs by calling
C<< $c->log_response_status_line >> and C<< $c->log_response_headers >>.

=cut

sub log_response {
    my $c = shift;

    return unless $c->debug;

    my($dump) = grep {$_->[0] eq 'Response' } $c->dump_these;
    my $response = $dump->[1];

    $c->log_response_status_line($response);
    $c->log_response_headers($response->headers);
}

=head2 $c->log_response_status_line($response)

Writes one line of information about the response to the debug logs.  This includes:

=over 4

=item * Response status code

=item * Content-Type header (if present)

=item * Content-Length header (if present)

=back

=cut

sub log_response_status_line {
    my ($c, $response) = @_;

    $c->log->debug(
        sprintf(
            'Response Code: %s; Content-Type: %s; Content-Length: %s',
            $response->status                            || 'unknown',
            $response->headers->header('Content-Type')   || 'unknown',
            $response->headers->header('Content-Length') || 'unknown'
        )
    );
}

=head2 $c->log_response_headers($headers);

Hook method which can be wrapped by plugins to log the response headers.
No-op in the default implementation.

=cut

sub log_response_headers {}

=head2 $c->log_request_parameters( query => {}, body => {} )

Logs request parameters to debug logs

=cut

sub log_request_parameters {
    my $c          = shift;
    my %all_params = @_;

    return unless $c->debug;

    my $column_width = Catalyst::Utils::term_width() - 44;
    foreach my $type (qw(query body)) {
        my $params = $all_params{$type};
        next if ! keys %$params;
        my $t = Text::SimpleTable->new( [ 35, 'Parameter' ], [ $column_width, 'Value' ] );
        for my $key ( sort keys %$params ) {
            my @values = ();
            if(ref $params eq 'Hash::MultiValue') {
                @values = $params->get_all($key);
            } else {
                my $param = $params->{$key};
                if( defined($param) ) {
                    @values = ref $param eq 'ARRAY' ? @$param : $param;
                }
            }
            $t->row( $key.( scalar @values > 1 ? ' [multiple]' : ''), join(', ', @values) );
        }
        $c->log->debug( ucfirst($type) . " Parameters are:\n" . $t->draw );
    }
}

=head2 $c->log_request_uploads

Logs file uploads included in the request to the debug logs.
The parameter name, filename, file type, and file size are all included in
the debug logs.

=cut

sub log_request_uploads {
    my $c = shift;
    my $request = shift;
    return unless $c->debug;
    my $uploads = $request->uploads;
    if ( keys %$uploads ) {
        my $t = Text::SimpleTable->new(
            [ 12, 'Parameter' ],
            [ 26, 'Filename' ],
            [ 18, 'Type' ],
            [ 9,  'Size' ]
        );
        for my $key ( sort keys %$uploads ) {
            my $upload = $uploads->{$key};
            for my $u ( ref $upload eq 'ARRAY' ? @{$upload} : ($upload) ) {
                $t->row( $key, $u->filename, $u->type, $u->size );
            }
        }
        $c->log->debug( "File Uploads are:\n" . $t->draw );
    }
}

=head2 $c->log_request_headers($headers);

Hook method which can be wrapped by plugins to log the request headers.
No-op in the default implementation.

=cut

sub log_request_headers {}

=head2 $c->log_headers($type => $headers)

Logs L<HTTP::Headers> (either request or response) to the debug logs.

=cut

sub log_headers {
    my $c       = shift;
    my $type    = shift;
    my $headers = shift;    # an HTTP::Headers instance

    return unless $c->debug;

    my $column_width = Catalyst::Utils::term_width() - 28;
    my $t = Text::SimpleTable->new( [ 15, 'Header Name' ], [ $column_width, 'Value' ] );
    $headers->scan(
        sub {
            my ( $name, $value ) = @_;
            $t->row( $name, $value );
        }
    );
    $c->log->debug( ucfirst($type) . " Headers:\n" . $t->draw );
}


=head2 $c->prepare_read

Prepares the input for reading.

=cut

sub prepare_read { my $c = shift; $c->engine->prepare_read( $c, @_ ) }

=head2 $c->prepare_request

Prepares the engine request.

=cut

sub prepare_request { my $c = shift; $c->engine->prepare_request( $c, @_ ) }

=head2 $c->prepare_uploads

Prepares uploads.

=cut

sub prepare_uploads {
    my $c = shift;
    $c->engine->prepare_uploads( $c, @_ );
}

=head2 $c->prepare_write

Prepares the output for writing.

=cut

sub prepare_write { my $c = shift; $c->engine->prepare_write( $c, @_ ) }

=head2 $c->request_class

Returns or sets the request class. Defaults to L<Catalyst::Request>.

=head2 $app->request_class_traits

An arrayref of L<Moose::Role>s which are applied to the request class.  You can
name the full namespace of the role, or a namespace suffix, which will then
be tried against the following standard namespace prefixes.

    $MyApp::TraitFor::Request::$trait_suffix
    Catalyst::TraitFor::Request::$trait_suffix

So for example if you set:

    MyApp->request_class_traits(['Foo']);

We try each possible role in turn (and throw an error if none load)

    Foo
    MyApp::TraitFor::Request::Foo
    Catalyst::TraitFor::Request::Foo

The namespace part 'TraitFor::Request' was chosen to assist in backwards
compatibility with L<CatalystX::RoleApplicator> which previously provided
these features in a stand alone package.

=head2 $app->composed_request_class

This is the request class which has been composed with any request_class_traits.

=head2 $c->response_class

Returns or sets the response class. Defaults to L<Catalyst::Response>.

=head2 $app->response_class_traits

An arrayref of L<Moose::Role>s which are applied to the response class.  You can
name the full namespace of the role, or a namespace suffix, which will then
be tried against the following standard namespace prefixes.

    $MyApp::TraitFor::Response::$trait_suffix
    Catalyst::TraitFor::Response::$trait_suffix

So for example if you set:

    MyApp->response_class_traits(['Foo']);

We try each possible role in turn (and throw an error if none load)

    Foo
    MyApp::TraitFor::Response::Foo
    Catalyst::TraitFor::Responset::Foo

The namespace part 'TraitFor::Response' was chosen to assist in backwards
compatibility with L<CatalystX::RoleApplicator> which previously provided
these features in a stand alone package.


=head2 $app->composed_response_class

This is the request class which has been composed with any response_class_traits.

=head2 $c->read( [$maxlength] )

Reads a chunk of data from the request body. This method is designed to
be used in a while loop, reading C<$maxlength> bytes on every call.
C<$maxlength> defaults to the size of the request if not specified.

You have to set C<< MyApp->config(parse_on_demand => 1) >> to use this
directly.

Warning: If you use read(), Catalyst will not process the body,
so you will not be able to access POST parameters or file uploads via
$c->request.  You must handle all body parsing yourself.

=cut

sub read { my $c = shift; return $c->request->read( @_ ) }

=head2 $c->run

Starts the engine.

=cut

sub run {
  my $app = shift;
  $app->_make_immutable_if_needed;
  $app->engine_loader->needs_psgi_engine_compat_hack ?
    $app->engine->run($app, @_) :
      $app->engine->run( $app, $app->_finalized_psgi_app, @_ );
}

sub _make_immutable_if_needed {
    my $class = shift;
    my $meta = find_meta($class);
    my $isa_ca = $class->isa('Class::Accessor::Fast') || $class->isa('Class::Accessor');
    if (
        $meta->is_immutable
        && ! { $meta->immutable_options }->{replace_constructor}
        && $isa_ca
    ) {
        warn("You made your application class ($class) immutable, "
            . "but did not inline the\nconstructor. "
            . "This will break catalyst, as your app \@ISA "
            . "Class::Accessor(::Fast)?\nPlease pass "
            . "(replace_constructor => 1)\nwhen making your class immutable.\n");
    }
    unless ($meta->is_immutable) {
        # XXX - FIXME warning here as you should make your app immutable yourself.
        $meta->make_immutable(
            replace_constructor => 1,
        );
    }
}

=head2 $c->set_action( $action, $code, $namespace, $attrs )

Sets an action in a given namespace.

=cut

sub set_action { my $c = shift; $c->dispatcher->set_action( $c, @_ ) }

=head2 $c->setup_actions($component)

Sets up actions for a component.

=cut

sub setup_actions { my $c = shift; $c->dispatcher->setup_actions( $c, @_ ) }

=head2 $c->setup_components

This method is called internally to set up the application's components.

It finds modules by calling the L<locate_components> method, expands them to
package names with the L<expand_component_module> method, and then installs
each component into the application.

The C<setup_components> config option is passed to both of the above methods.

Installation of each component is performed by the L<setup_component> method,
below.

=cut

sub setup_components {
    my $class = shift;

    my $config  = $class->config->{ setup_components };

    my @comps = $class->locate_components($config);

    my $deprecatedcatalyst_component_names = grep { /::[CMV]::/ } @comps;
    $class->log->warn(qq{Your application is using the deprecated ::[MVC]:: type naming scheme.\n}.
        qq{Please switch your class names to ::Model::, ::View:: and ::Controller: as appropriate.\n}
    ) if $deprecatedcatalyst_component_names;

    for my $component ( @comps ) {

        # We pass ignore_loaded here so that overlay files for (e.g.)
        # Model::DBI::Schema sub-classes are loaded - if it's in @comps
        # we know M::P::O found a file on disk so this is safe

        Catalyst::Utils::ensure_class_loaded( $component, { ignore_loaded => 1 } );
    }

    for my $component (@comps) {
        my $instance = $class->components->{ $component } = $class->delayed_setup_component($component);
    }

    # Inject a component or wrap a stand alone class in an adaptor. This makes a list
    # of named components in the configuration that are not actually existing (not a
    # real file).

    my @injected = $class->setup_injected_components;

    # All components are registered, now we need to 'init' them.
    foreach my $component_name (@comps, @injected) {
      $class->components->{$component_name} = $class->components->{$component_name}->() if
        (ref($class->components->{$component_name}) || '') eq 'CODE';
    }
}

=head2 $app->setup_injected_components

Called by setup_compoents to setup components that are injected.

=cut

sub setup_injected_components {
    my ($class) = @_;
    my @injected_components = keys %{$class->config->{inject_components} ||+{}};

    foreach my $injected_comp_name(@injected_components) {
        $class->setup_injected_component(
          $injected_comp_name,
          $class->config->{inject_components}->{$injected_comp_name});
    }

    return map { $class ."::" . $_ }
      @injected_components;
}

=head2 $app->setup_injected_component( $injected_component_name, $config )

Setup a given injected component.

=cut

sub setup_injected_component {
    my ($class, $injected_comp_name, $config) = @_;
    if(my $component_class = $config->{from_component}) {
        my @roles = @{$config->{roles} ||[]};
        Catalyst::Utils::inject_component(
          into => $class,
          component => $component_class,
          (scalar(@roles) ? (traits => \@roles) : ()),
          as => $injected_comp_name);
    }
}

=head2 $app->inject_component($MyApp_Component_name => \%args);

Add a component that is injected at setup:

    MyApp->inject_component( 'Model::Foo' => { from_component => 'Common::Foo' } );

Must be called before ->setup.  Expects a component name for your
current application and \%args where

=over 4

=item from_component

The target component being injected into your application

=item roles

An arrayref of L<Moose::Role>s that are applied to your component.

=back

Example

    MyApp->inject_component(
      'Model::Foo' => {
        from_component => 'Common::Model::Foo',
        roles => ['Role1', 'Role2'],
      });

=head2 $app->inject_components

Inject a list of components:

    MyApp->inject_components(
      'Model::FooOne' => {
        from_component => 'Common::Model::Foo',
        roles => ['Role1', 'Role2'],
      },
      'Model::FooTwo' => {
        from_component => 'Common::Model::Foo',
        roles => ['Role1', 'Role2'],
      });

=cut

sub inject_component {
  my ($app, $name, $args) = @_;
  die "Component $name exists" if
    $app->config->{inject_components}->{$name};
  $app->config->{inject_components}->{$name} = $args;
}

sub inject_components {
  my $app = shift;
  while(@_) {
    $app->inject_component(shift, shift);
  }
}

=head2 $c->locate_components( $setup_component_config )

This method is meant to provide a list of component modules that should be
setup for the application.  By default, it will use L<Module::Pluggable>.

Specify a C<setup_components> config option to pass additional options directly
to L<Module::Pluggable>. To add additional search paths, specify a key named
C<search_extra> as an array reference. Items in the array beginning with C<::>
will have the application class name prepended to them.

=cut

sub locate_components {
    my $class  = shift;
    my $config = shift;

    my @paths   = qw( ::M ::Model ::V ::View ::C ::Controller );
    my $extra   = $config->{ search_extra } || [];

    unshift @paths, @$extra;

    my @comps = map { sort { length($a) <=> length($b) } Module::Pluggable::Object->new(
      search_path => [ map { s/^(?=::)/$class/; $_; } ($_) ],
      %$config
    )->plugins } @paths;

    return @comps;
}

=head2 $c->expand_component_module( $component, $setup_component_config )

Components found by C<locate_components> will be passed to this method, which
is expected to return a list of component (package) names to be set up.

=cut

sub expand_component_module {
    my ($class, $module) = @_;
    return Devel::InnerPackage::list_packages( $module );
}

=head2 $app->delayed_setup_component

Returns a coderef that points to a setup_component instance.  Used
internally for when you want to delay setup until the first time
the component is called.

=cut

sub delayed_setup_component {
  my($class, $component, @more) = @_;
  return sub {
    return my $instance = $class->setup_component($component, @more);
  };
}

=head2 $c->setup_component

=cut

sub setup_component {
    my( $class, $component ) = @_;

    unless ( $component->can( 'COMPONENT' ) ) {
        return $component;
    }

    my $config = $class->config_for($component);
    # Stash catalyst_component_name in the config here, so that custom COMPONENT
    # methods also pass it. local to avoid pointlessly shitting in config
    # for the debug screen, as $component is already the key name.
    local $config->{catalyst_component_name} = $component;

    my $instance = eval {
      $component->COMPONENT( $class, $config );
    } || do {
      my $error = $@;
      chomp $error;
      Catalyst::Exception->throw(
        message => qq/Couldn't instantiate component "$component", "$error"/
      );
    };

    unless (blessed $instance) {
        my $metaclass = Moose::Util::find_meta($component);
        my $method_meta = $metaclass->find_method_by_name('COMPONENT');
        my $component_method_from = $method_meta->associated_metaclass->name;
        my $value = defined($instance) ? $instance : 'undef';
        Catalyst::Exception->throw(
            message =>
            qq/Couldn't instantiate component "$component", COMPONENT() method (from $component_method_from) didn't return an object-like value (value was $value)./
        );
    }

    my @expanded_components = $instance->can('expand_modules')
      ? $instance->expand_modules( $component, $config )
      : $class->expand_component_module( $component, $config );
    for my $component (@expanded_components) {
      next if $class->components->{ $component };
      $class->components->{ $component } = $class->setup_component($component);
    }

    return $instance;
}

=head2 $app->config_for( $component_name )

Return the application level configuration (which is not yet merged with any
local component configuration, via $component_class->config) for the named
component or component object. Example:

    MyApp->config(
      'Model::Foo' => { a => 1, b => 2},
    );

    my $config = MyApp->config_for('MyApp::Model::Foo');

In this case $config is the hashref C<< {a=>1, b=>2} >>.

This is also handy for looking up configuration for a plugin, to make sure you follow
existing L<Catalyst> standards for where a plugin should put its configuration.

=cut

sub config_for {
    my ($class, $component_name) = @_;
    my $component_suffix = Catalyst::Utils::class2classsuffix($component_name);
    my $config = $class->config->{ $component_suffix } || {};

    return $config;
}

=head2 $c->setup_dispatcher

Sets up dispatcher.

=cut

sub setup_dispatcher {
    my ( $class, $dispatcher ) = @_;

    if ($dispatcher) {
        $dispatcher = 'Catalyst::Dispatcher::' . $dispatcher;
    }

    if ( my $env = Catalyst::Utils::env_value( $class, 'DISPATCHER' ) ) {
        $dispatcher = 'Catalyst::Dispatcher::' . $env;
    }

    unless ($dispatcher) {
        $dispatcher = $class->dispatcher_class;
    }

    load_class($dispatcher);

    # dispatcher instance
    $class->dispatcher( $dispatcher->new );
}

=head2 $c->setup_engine

Sets up engine.

=cut

sub engine_class {
    my ($class, $requested_engine) = @_;

    if (!$class->engine_loader || $requested_engine) {
        $class->engine_loader(
            Catalyst::EngineLoader->new({
                application_name => $class,
                (defined $requested_engine
                     ? (catalyst_engine_class => $requested_engine) : ()),
            }),
        );
    }

    $class->engine_loader->catalyst_engine_class;
}

sub setup_engine {
    my ($class, $requested_engine) = @_;

    my $engine = do {
        my $loader = $class->engine_loader;

        if (!$loader || $requested_engine) {
            $loader = Catalyst::EngineLoader->new({
                application_name => $class,
                (defined $requested_engine
                     ? (requested_engine => $requested_engine) : ()),
            }),

            $class->engine_loader($loader);
        }

        $loader->catalyst_engine_class;
    };

    # Don't really setup_engine -- see _setup_psgi_app for explanation.
    return if $class->loading_psgi_file;

    load_class($engine);

    if ($ENV{MOD_PERL}) {
        my $apache = $class->engine_loader->auto;

        my $meta = find_meta($class);
        my $was_immutable = $meta->is_immutable;
        my %immutable_options = $meta->immutable_options;
        $meta->make_mutable if $was_immutable;

        $meta->add_method(handler => sub {
            my $r = shift;
            my $psgi_app = $class->_finalized_psgi_app;
            $apache->call_app($r, $psgi_app);
        });

        $meta->make_immutable(%immutable_options) if $was_immutable;
    }

    $class->engine( $engine->new );

    return;
}

## This exists just to supply a prebuild psgi app for mod_perl and for the
## build in server support (back compat support for pre psgi port behavior).
## This is so that we don't build a new psgi app for each request when using
## the mod_perl handler or the built in servers (http and fcgi, etc).

sub _finalized_psgi_app {
    my ($app) = @_;

    unless ($app->_psgi_app) {
        my $psgi_app = $app->_setup_psgi_app;
        $app->_psgi_app($psgi_app);
    }

    return $app->_psgi_app;
}

## Look for a psgi file like 'myapp_web.psgi' (if the app is MyApp::Web) in the
## home directory and load that and return it (just assume it is doing the
## right thing :) ).  If that does not exist, call $app->psgi_app, wrap that
## in default_middleware and return it ( this is for backward compatibility
## with pre psgi port behavior ).

sub _setup_psgi_app {
    my ($app) = @_;

    for my $home (Path::Class::Dir->new($app->config->{home})) {
        my $psgi_file = $home->file(
            Catalyst::Utils::appprefix($app) . '.psgi',
        );

        next unless -e $psgi_file;

        # If $psgi_file calls ->setup_engine, it's doing so to load
        # Catalyst::Engine::PSGI. But if it does that, we're only going to
        # throw away the loaded PSGI-app and load the 5.9 Catalyst::Engine
        # anyway. So set a flag (ick) that tells setup_engine not to populate
        # $c->engine or do any other things we might regret.

        $app->loading_psgi_file(1);
        my $psgi_app = Plack::Util::load_psgi($psgi_file);
        $app->loading_psgi_file(0);

        return $psgi_app
            unless $app->engine_loader->needs_psgi_engine_compat_hack;

        warn <<"EOW";
Found a legacy Catalyst::Engine::PSGI .psgi file at ${psgi_file}.

Its content has been ignored. Please consult the Catalyst::Upgrading
documentation on how to upgrade from Catalyst::Engine::PSGI.
EOW
    }

    return $app->apply_default_middlewares($app->psgi_app);
}

=head2 $c->apply_default_middlewares

Adds the following L<Plack> middlewares to your application, since they are
useful and commonly needed:

L<Plack::Middleware::LighttpdScriptNameFix> (if you are using Lighttpd),
L<Plack::Middleware::IIS6ScriptNameFix> (always applied since this middleware
is smart enough to conditionally apply itself).

We will also automatically add L<Plack::Middleware::ReverseProxy> if we notice
that your HTTP $env variable C<REMOTE_ADDR> is '127.0.0.1'.  This is usually
an indication that your server is running behind a proxy frontend.  However in
2014 this is often not the case.  We preserve this code for backwards compatibility
however I B<highly> recommend that if you are running the server behind a front
end proxy that you clearly indicate so with the C<using_frontend_proxy> configuration
setting to true for your environment configurations that run behind a proxy.  This
way if you change your front end proxy address someday your code would inexplicably
stop working as expected.

Additionally if we detect we are using Nginx, we add a bit of custom middleware
to solve some problems with the way that server handles $ENV{PATH_INFO} and
$ENV{SCRIPT_NAME}.

Please B<NOTE> that if you do use C<using_frontend_proxy> the middleware is now
adding via C<registered_middleware> rather than this method.

If you are using Lighttpd or IIS6 you may wish to apply these middlewares.  In
general this is no longer a common case but we have this here for backward
compatibility.

=cut


sub apply_default_middlewares {
    my ($app, $psgi_app) = @_;

    # Don't add this conditional IF we are explicitly saying we want the
    # frontend proxy support.  We don't need it here since if that is the
    # case it will be always loaded in the default_middleware.

    unless($app->config->{using_frontend_proxy}) {
      $psgi_app = Plack::Middleware::Conditional->wrap(
          $psgi_app,
          builder   => sub { Plack::Middleware::ReverseProxy->wrap($_[0]) },
          condition => sub {
              my ($env) = @_;
              return if $app->config->{ignore_frontend_proxy};
              return $env->{REMOTE_ADDR} && $env->{REMOTE_ADDR} eq '127.0.0.1';
          },
      );
    }

    # If we're running under Lighttpd, swap PATH_INFO and SCRIPT_NAME
    # http://lists.scsys.co.uk/pipermail/catalyst/2006-June/008361.html
    $psgi_app = Plack::Middleware::Conditional->wrap(
        $psgi_app,
        builder   => sub { Plack::Middleware::LighttpdScriptNameFix->wrap($_[0]) },
        condition => sub {
            my ($env) = @_;
            return unless $env->{SERVER_SOFTWARE} && $env->{SERVER_SOFTWARE} =~ m!lighttpd[-/]1\.(\d+\.\d+)!;
            return unless $1 < 4.23;
            1;
        },
    );

    # we're applying this unconditionally as the middleware itself already makes
    # sure it doesn't fuck things up if it's not running under one of the right
    # IIS versions
    $psgi_app = Plack::Middleware::IIS6ScriptNameFix->wrap($psgi_app);

    # And another IIS issue, this time with IIS7.
    $psgi_app = Plack::Middleware::Conditional->wrap(
        $psgi_app,
        builder => sub { Plack::Middleware::IIS7KeepAliveFix->wrap($_[0]) },
        condition => sub {
            my ($env) = @_;
            return $env->{SERVER_SOFTWARE} && $env->{SERVER_SOFTWARE} =~ m!IIS/7\.[0-9]!;
        },
    );

    return $psgi_app;
}

=head2 App->psgi_app

=head2 App->to_app

Returns a PSGI application code reference for the catalyst application
C<$c>. This is the bare application created without the C<apply_default_middlewares>
method called.  We do however apply C<registered_middleware> since those are
integral to how L<Catalyst> functions.  Also, unlike starting your application
with a generated server script (via L<Catalyst::Devel> and C<catalyst.pl>) we do
not attempt to return a valid L<PSGI> application using any existing C<${myapp}.psgi>
scripts in your $HOME directory.

B<NOTE> C<apply_default_middlewares> was originally created when the first PSGI
port was done for v5.90000.  These are middlewares that are added to achieve
backward compatibility with older applications.  If you start your application
using one of the supplied server scripts (generated with L<Catalyst::Devel> and
the project skeleton script C<catalyst.pl>) we apply C<apply_default_middlewares>
automatically.  This was done so that pre and post PSGI port applications would
work the same way.

This is what you want to be using to retrieve the PSGI application code
reference of your Catalyst application for use in a custom F<.psgi> or in your
own created server modules.

=cut

*to_app = \&psgi_app;

sub psgi_app {
    my ($app) = @_;
    my $psgi = $app->engine->build_psgi_app($app);
    return $app->Catalyst::Utils::apply_registered_middleware($psgi);
}

=head2 $c->setup_home

Sets up the home directory.

=cut

sub setup_home {
    my ( $class, $home ) = @_;

    if ( my $env = Catalyst::Utils::env_value( $class, 'HOME' ) ) {
        $home = $env;
    }

    $home ||= Catalyst::Utils::home($class);

    if ($home) {
        #I remember recently being scolded for assigning config values like this
        $class->config->{home} ||= $home;
        $class->config->{root} ||= Path::Class::Dir->new($home)->subdir('root');
    }
}

=head2 $c->setup_encoding

Sets up the input/output encoding. See L<ENCODING>

=cut

sub setup_encoding {
    my $c = shift;
    if( exists($c->config->{encoding}) && !defined($c->config->{encoding}) ) {
        # Ok, so the user has explicitly said "I don't want encoding..."
        return;
    } else {
      my $enc = defined($c->config->{encoding}) ?
        delete $c->config->{encoding} : 'UTF-8'; # not sure why we delete it... (JNAP)
      $c->encoding($enc);
    }
}

=head2 handle_unicode_encoding_exception

Hook to let you customize how encoding errors are handled. By default
we just throw an exception and the default error page will pick it up.
Receives a hashref of debug information. Example of call (from the
Catalyst internals):

  my $decoded_after_fail = $c->handle_unicode_encoding_exception({
        param_value => $value,
        error_msg => $_,
        encoding_step => 'params',
   });

The calling code expects to receive a decoded string or an exception.

You can override this for custom handling of unicode errors. By
default we just die. If you want a custom response here, one approach
is to throw an HTTP style exception, instead of returning a decoded
string or throwing a generic exception.

    sub handle_unicode_encoding_exception {
      my ($c, $params) = @_;
      HTTP::Exception::BAD_REQUEST->throw(status_message=>$params->{error_msg});
    }

Alternatively you can 'catch' the error, stash it and write handling code later
in your application:

    sub handle_unicode_encoding_exception {
      my ($c, $params) = @_;
      $c->stash(BAD_UNICODE_DATA=>$params);
      # return a dummy string.
      return 1;
    }

<B>NOTE:</b> Please keep in mind that once an error like this occurs,
the request setup is still ongoing, which means the state of C<$c> and
related context parts like the request and response may not be setup
up correctly (since we haven't finished the setup yet). If you throw
an exception the setup is aborted.

=cut

sub handle_unicode_encoding_exception {
    my ( $self, $exception_ctx ) = @_;
    die $exception_ctx->{error_msg};
}

# Some unicode helpers cargo culted from the old plugin.  These could likely
# be neater.

sub _handle_unicode_decoding {
    my ( $self, $value ) = @_;

    return unless defined $value;

    ## I think this mess is to support the old nested
    if ( ref $value eq 'ARRAY' ) {
        foreach ( @$value ) {
            $_ = $self->_handle_unicode_decoding($_);
        }
        return $value;
    }
    elsif ( ref $value eq 'HASH' ) {
        foreach (keys %$value) {
            my $encoded_key = $self->_handle_param_unicode_decoding($_);
            $value->{$encoded_key} = $self->_handle_unicode_decoding($value->{$_});

            # If the key was encoded we now have two (the original and current so
            # delete the original.
            delete $value->{$_} if $_ ne $encoded_key;
        }
        return $value;
    }
    else {
        return $self->_handle_param_unicode_decoding($value);
    }
}

sub _handle_param_unicode_decoding {
    my ( $self, $value, $check ) = @_;
    return unless defined $value; # not in love with just ignoring undefs - jnap
    return $value if blessed($value); #don't decode when the value is an object.

    my $enc = $self->encoding;

    return $value unless $enc; # don't decode if no encoding is specified

    $check ||= $self->_encode_check;
    return try {
      $enc->decode( $value, $check);
    }
    catch {
        return $self->handle_unicode_encoding_exception({
            param_value => $value,
            error_msg => $_,
            encoding_step => 'params',
        });
    };
}

=head2 $c->setup_log

Sets up log by instantiating a L<Catalyst::Log|Catalyst::Log> object and
passing it to C<log()>. Pass in a comma-delimited list of levels to set the
log to.

This method also installs a C<debug> method that returns a true value into the
catalyst subclass if the "debug" level is passed in the comma-delimited list,
or if the C<$CATALYST_DEBUG> environment variable is set to a true value.

Note that if the log has already been setup, by either a previous call to
C<setup_log> or by a call such as C<< __PACKAGE__->log( MyLogger->new ) >>,
that this method won't actually set up the log object.

=cut

sub setup_log {
    my ( $class, $levels ) = @_;

    $levels ||= '';
    $levels =~ s/^\s+//;
    $levels =~ s/\s+$//;
    my %levels = map { $_ => 1 } split /\s*,\s*/, $levels;

    my $env_debug = Catalyst::Utils::env_value( $class, 'DEBUG' );
    if ( defined $env_debug ) {
        $levels{debug} = 1 if $env_debug; # Ugly!
        delete($levels{debug}) unless $env_debug;
    }

    unless ( $class->log ) {
        $class->log( Catalyst::Log->new(keys %levels) );
    }

    if ( $levels{debug} ) {
        Class::MOP::get_metaclass_by_name($class)->add_method('debug' => sub { 1 });
        $class->log->debug('Debug messages enabled');
    }
}

=head2 $c->setup_plugins

Sets up plugins.

=cut

=head2 $c->setup_stats

Sets up timing statistics class.

=cut

sub setup_stats {
    my ( $class, $stats ) = @_;

    Catalyst::Utils::ensure_class_loaded($class->stats_class);

    my $env = Catalyst::Utils::env_value( $class, 'STATS' );
    if ( defined($env) ? $env : ($stats || $class->debug ) ) {
        Class::MOP::get_metaclass_by_name($class)->add_method('use_stats' => sub { 1 });
        $class->log->debug('Statistics enabled');
    }
}


=head2 $c->registered_plugins

Returns a sorted list of the plugins which have either been stated in the
import list.

If passed a given plugin name, it will report a boolean value indicating
whether or not that plugin is loaded.  A fully qualified name is required if
the plugin name does not begin with C<Catalyst::Plugin::>.

 if ($c->registered_plugins('Some::Plugin')) {
     ...
 }

=cut

{

    sub registered_plugins {
        my $proto = shift;
        return sort keys %{ $proto->_plugins } unless @_;
        my $plugin = shift;
        return 1 if exists $proto->_plugins->{$plugin};
        return exists $proto->_plugins->{"Catalyst::Plugin::$plugin"};
    }

    sub _register_plugin {
        my ( $proto, $plugin, $instant ) = @_;
        my $class = ref $proto || $proto;

        load_class( $plugin );
        $class->log->warn( "$plugin inherits from 'Catalyst::Component' - this is deprecated and will not work in 5.81" )
            if $plugin->isa( 'Catalyst::Component' );
        my $plugin_meta = Moose::Meta::Class->create($plugin);
        if (!$plugin_meta->has_method('new')
            && ( $plugin->isa('Class::Accessor::Fast') || $plugin->isa('Class::Accessor') ) ) {
            $plugin_meta->add_method('new', Moose::Object->meta->get_method('new'))
        }
        if (!$instant && !$proto->_plugins->{$plugin}) {
            my $meta = Class::MOP::get_metaclass_by_name($class);
            $meta->superclasses($plugin, $meta->superclasses);
        }
        $proto->_plugins->{$plugin} = 1;
        return $class;
    }

    sub _default_plugins { return qw() }

    sub setup_plugins {
        my ( $class, $plugins ) = @_;

        $class->_plugins( {} ) unless $class->_plugins;
        $plugins = [ grep {
            m/Unicode::Encoding/ ? do {
                $class->log->warn(
                    'Unicode::Encoding plugin is auto-applied,'
                    . ' please remove this from your appclass'
                    . ' and make sure to define "encoding" config'
                );
                unless (exists $class->config->{'encoding'}) {
                  $class->config->{'encoding'} = 'UTF-8';
                }
                () }
                : $_
        } @$plugins ];
        push @$plugins, $class->_default_plugins;
        $plugins = Data::OptList::mkopt($plugins || []);

        my @plugins = map {
            [ Catalyst::Utils::resolve_namespace(
                  $class . '::Plugin',
                  'Catalyst::Plugin', $_->[0]
              ),
              $_->[1],
            ]
         } @{ $plugins };

        for my $plugin ( reverse @plugins ) {
            load_class($plugin->[0], $plugin->[1]);
            my $meta = find_meta($plugin->[0]);
            next if $meta && $meta->isa('Moose::Meta::Role');

            $class->_register_plugin($plugin->[0]);
        }

        my @roles =
            map  { $_->[0]->name, $_->[1] }
            grep { blessed($_->[0]) && $_->[0]->isa('Moose::Meta::Role') }
            map  { [find_meta($_->[0]), $_->[1]] }
            @plugins;

        Moose::Util::apply_all_roles(
            $class => @roles
        ) if @roles;
    }
}

=head2 default_middleware

Returns a list of instantiated PSGI middleware objects which is the default
middleware that is active for this application (taking any configuration
options into account, excluding your custom added middleware via the C<psgi_middleware>
configuration option).  You can override this method if you wish to change
the default middleware (although do so at risk since some middleware is vital
to application function.)

The current default middleware list is:

      Catalyst::Middleware::Stash
      Plack::Middleware::HTTPExceptions
      Plack::Middleware::RemoveRedundantBody
      Plack::Middleware::FixMissingBodyInRedirect
      Plack::Middleware::ContentLength
      Plack::Middleware::MethodOverride
      Plack::Middleware::Head

If the configuration setting C<using_frontend_proxy> is true we add:

      Plack::Middleware::ReverseProxy

If the configuration setting C<using_frontend_proxy_path> is true we add:

      Plack::Middleware::ReverseProxyPath

But B<NOTE> that L<Plack::Middleware::ReverseProxyPath> is not a dependency of the
L<Catalyst> distribution so if you want to use this option you should add it to
your project distribution file.

These middlewares will be added at L</setup_middleware> during the
L</setup> phase of application startup.

=cut

sub default_middleware {
    my $class = shift;
    my @mw = (
      Catalyst::Middleware::Stash->new,
      Plack::Middleware::HTTPExceptions->new,
      Plack::Middleware::RemoveRedundantBody->new,
      Plack::Middleware::FixMissingBodyInRedirect->new,
      Plack::Middleware::ContentLength->new,
      Plack::Middleware::MethodOverride->new,
      Plack::Middleware::Head->new);

    if($class->config->{using_frontend_proxy}) {
        push @mw, Plack::Middleware::ReverseProxy->new;
    }

    if($class->config->{using_frontend_proxy_path}) {
        if(Class::Load::try_load_class('Plack::Middleware::ReverseProxyPath')) {
            push @mw, Plack::Middleware::ReverseProxyPath->new;
        } else {
          $class->log->error("Cannot use configuration 'using_frontend_proxy_path' because 'Plack::Middleware::ReverseProxyPath' is not installed");
        }
    }

    return @mw;
}

=head2 registered_middlewares

Read only accessor that returns an array of all the middleware in the order
that they were added (which is the REVERSE of the order they will be applied).

The values returned will be either instances of L<Plack::Middleware> or of a
compatible interface, or a coderef, which is assumed to be inlined middleware

=head2 setup_middleware (?@middleware)

Read configuration information stored in configuration key C<psgi_middleware> or
from passed @args.

See under L</CONFIGURATION> information regarding C<psgi_middleware> and how
to use it to enable L<Plack::Middleware>

This method is automatically called during 'setup' of your application, so
you really don't need to invoke it.  However you may do so if you find the idea
of loading middleware via configuration weird :).  For example:

    package MyApp;

    use Catalyst;

    __PACKAGE__->setup_middleware('Head');
    __PACKAGE__->setup;

When we read middleware definitions from configuration, we reverse the list
which sounds odd but is likely how you expect it to work if you have prior
experience with L<Plack::Builder> or if you previously used the plugin
L<Catalyst::Plugin::EnableMiddleware> (which is now considered deprecated)

So basically your middleware handles an incoming request from the first
registered middleware, down and handles the response from the last middleware
up.

=cut

sub registered_middlewares {
    my $class = shift;
    if(my $middleware = $class->_psgi_middleware) {
        my @mw = ($class->default_middleware, @$middleware);

        if($class->config->{using_frontend_proxy}) {
          push @mw, Plack::Middleware::ReverseProxy->new;
        }

        return @mw;
    } else {
        die "You cannot call ->registered_middlewares until middleware has been setup";
    }
}

sub setup_middleware {
    my $class = shift;
    my @middleware_definitions;

    # If someone calls this method you can add middleware with args.  However if its
    # called without an arg we need to setup the configuration middleware.
    if(@_) {
      @middleware_definitions = reverse(@_);
    } else {
      @middleware_definitions = reverse(@{$class->config->{'psgi_middleware'}||[]})
        unless $class->finalized_default_middleware;
      $class->finalized_default_middleware(1); # Only do this once, just in case some people call setup over and over...
    }

    my @middleware = ();
    while(my $next = shift(@middleware_definitions)) {
        if(ref $next) {
            if(Scalar::Util::blessed $next && $next->can('wrap')) {
                push @middleware, $next;
            } elsif(ref $next eq 'CODE') {
                push @middleware, $next;
            } elsif(ref $next eq 'HASH') {
                my $namespace = shift @middleware_definitions;
                my $mw = $class->Catalyst::Utils::build_middleware($namespace, %$next);
                push @middleware, $mw;
            } else {
              die "I can't handle middleware definition ${\ref $next}";
            }
        } else {
          my $mw = $class->Catalyst::Utils::build_middleware($next);
          push @middleware, $mw;
        }
    }

    my @existing = @{$class->_psgi_middleware || []};
    $class->_psgi_middleware([@middleware,@existing,]);
}

=head2 registered_data_handlers

A read only copy of registered Data Handlers returned as a Hash, where each key
is a content type and each value is a subref that attempts to decode that content
type.

=head2 setup_data_handlers (?@data_handler)

Read configuration information stored in configuration key C<data_handlers> or
from passed @args.

See under L</CONFIGURATION> information regarding C<data_handlers>.

This method is automatically called during 'setup' of your application, so
you really don't need to invoke it.

=head2 default_data_handlers

Default Data Handlers that come bundled with L<Catalyst>.  Currently there are
only two default data handlers, for 'application/json' and an alternative to
'application/x-www-form-urlencoded' which supposed nested form parameters via
L<CGI::Struct> or via L<CGI::Struct::XS> IF you've installed it.

The 'application/json' data handler is used to parse incoming JSON into a Perl
data structure.  It uses L<JSON::MaybeXS>.  This allows you to fail back to
L<JSON::PP>, which is a Pure Perl JSON decoder, and has the smallest dependency
impact.

Because we don't wish to add more dependencies to L<Catalyst>, if you wish to
use this new feature we recommend installing L<Cpanel::JSON::XS> in order to get
the best performance.  You should add either to your dependency list
(Makefile.PL, dist.ini, cpanfile, etc.)

=cut

sub registered_data_handlers {
    my $class = shift;
    if(my $data_handlers = $class->_data_handlers) {
        return %$data_handlers;
    } else {
        $class->setup_data_handlers;
        return $class->registered_data_handlers;
    }
}

sub setup_data_handlers {
    my ($class, %data_handler_callbacks) = @_;
    %data_handler_callbacks = (
      %{$class->default_data_handlers},
      %{$class->config->{'data_handlers'}||+{}},
      %data_handler_callbacks);

    $class->_data_handlers(\%data_handler_callbacks);
}

sub default_data_handlers {
    my ($class) = @_;
    return +{
      'application/x-www-form-urlencoded' => sub {
          my ($fh, $req) = @_;
          my $params = $req->_use_hash_multivalue ? $req->body_parameters->mixed : $req->body_parameters;
          Class::Load::load_first_existing_class('CGI::Struct::XS', 'CGI::Struct')
            ->can('build_cgi_struct')->($params);
      },
      'application/json' => sub {
          my ($fh, $req) = @_;
          require JSON::MaybeXS;
          my $slurped;
          return eval {
            local $/;
            $slurped = $fh->getline;
            JSON::MaybeXS::decode_json($slurped); # decode_json does utf8 decoding for us
          } || Catalyst::Exception->throw(sprintf "Error Parsing POST '%s', Error: %s", (defined($slurped) ? $slurped : 'undef') ,$@);
        },
    };
}

sub _handle_http_exception {
    my ( $self, $error ) = @_;
    if (
           !$self->config->{always_catch_http_exceptions}
        && blessed $error
        && (
            $error->can('as_psgi')
            || (   $error->can('code')
                && $error->code =~ m/^[1-5][0-9][0-9]$/ )
        )
      )
    {
        return 1;
    }
}

=head2 $c->stack

Returns an arrayref of the internal execution stack (actions that are
currently executing).

=head2 $c->stats

Returns the current timing statistics object. By default Catalyst uses
L<Catalyst::Stats|Catalyst::Stats>, but can be set otherwise with
L<< stats_class|/"$c->stats_class" >>.

Even if L<< -Stats|/"-Stats" >> is not enabled, the stats object is still
available. By enabling it with C<< $c->stats->enabled(1) >>, it can be used to
profile explicitly, although MyApp.pm still won't profile nor output anything
by itself.

=head2 $c->stats_class

Returns or sets the stats (timing statistics) class. L<Catalyst::Stats|Catalyst::Stats> is used by default.

=head2 $app->stats_class_traits

A arrayref of L<Moose::Role>s that are applied to the stats_class before creating it.

=head2 $app->composed_stats_class

this is the stats_class composed with any 'stats_class_traits'.  You can
name the full namespace of the role, or a namespace suffix, which will then
be tried against the following standard namespace prefixes.

    $MyApp::TraitFor::Stats::$trait_suffix
    Catalyst::TraitFor::Stats::$trait_suffix

So for example if you set:

    MyApp->stats_class_traits(['Foo']);

We try each possible role in turn (and throw an error if none load)

    Foo
    MyApp::TraitFor::Stats::Foo
    Catalyst::TraitFor::Stats::Foo

The namespace part 'TraitFor::Stats' was chosen to assist in backwards
compatibility with L<CatalystX::RoleApplicator> which previously provided
these features in a stand alone package.

=head2 $c->use_stats

Returns 1 when L<< stats collection|/"-Stats" >> is enabled.

Note that this is a static method, not an accessor and should be overridden
by declaring C<sub use_stats { 1 }> in your MyApp.pm, not by calling C<< $c->use_stats(1) >>.

=cut

sub use_stats { 0 }


=head2 $c->write( $data )

Writes $data to the output stream. When using this method directly, you
will need to manually set the C<Content-Length> header to the length of
your output data, if known.

=cut

sub write {
    my $c = shift;

    # Finalize headers if someone manually writes output (for compat)
    $c->finalize_headers;

    return $c->response->write( @_ );
}

=head2 version

Returns the Catalyst version number. Mostly useful for "powered by"
messages in template systems.

=cut

sub version { return $Catalyst::VERSION }

=head1 CONFIGURATION

There are a number of 'base' config variables which can be set:

=over

=item *

C<always_catch_http_exceptions> - As of version 5.90060 Catalyst
rethrows errors conforming to the interface described by
L<Plack::Middleware::HTTPExceptions> and lets the middleware deal with it.
Set true to get the deprecated behaviour and have Catalyst catch HTTP exceptions.

=item *

C<default_model> - The default model picked if you say C<< $c->model >>. See L<< /$c->model($name) >>.

=item *

C<default_view> - The default view to be rendered or returned when C<< $c->view >> is called. See L<< /$c->view($name) >>.

=item *

C<disable_component_resolution_regex_fallback> - Turns
off the deprecated component resolution functionality so
that if any of the component methods (e.g. C<< $c->controller('Foo') >>)
are called then regex search will not be attempted on string values and
instead C<undef> will be returned.

=item *

C<home> - The application home directory. In an uninstalled application,
this is the top level application directory. In an installed application,
this will be the directory containing C<< MyApp.pm >>.

=item *

C<ignore_frontend_proxy> - See L</PROXY SUPPORT>

=item *

C<name> - The name of the application in debug messages and the debug and
welcome screens

=item *

C<parse_on_demand> - The request body (for example file uploads) will not be parsed
until it is accessed. This allows you to (for example) check authentication (and reject
the upload) before actually receiving all the data. See L</ON-DEMAND PARSER>

=item *

C<root> - The root directory for templates. Usually this is just a
subdirectory of the home directory, but you can set it to change the
templates to a different directory.

=item *

C<search_extra> - Array reference passed to Module::Pluggable to for additional
namespaces from which components will be loaded (and constructed and stored in
C<< $c->components >>).

=item *

C<show_internal_actions> - If true, causes internal actions such as C<< _DISPATCH >>
to be shown in hit debug tables in the test server.

=item *

C<use_request_uri_for_path> - Controls if the C<REQUEST_URI> or C<PATH_INFO> environment
variable should be used for determining the request path.

Most web server environments pass the requested path to the application using environment variables,
from which Catalyst has to reconstruct the request base (i.e. the top level path to / in the application,
exposed as C<< $c->request->base >>) and the request path below that base.

There are two methods of doing this, both of which have advantages and disadvantages. Which method is used
is determined by the C<< $c->config(use_request_uri_for_path) >> setting (which can either be true or false).

=over

=item use_request_uri_for_path => 0

This is the default (and the) traditional method that Catalyst has used for determining the path information.
The path is generated from a combination of the C<PATH_INFO> and C<SCRIPT_NAME> environment variables.
The allows the application to behave correctly when C<mod_rewrite> is being used to redirect requests
into the application, as these variables are adjusted by mod_rewrite to take account for the redirect.

However this method has the major disadvantage that it is impossible to correctly decode some elements
of the path, as RFC 3875 says: "C<< Unlike a URI path, the PATH_INFO is not URL-encoded, and cannot
contain path-segment parameters. >>" This means PATH_INFO is B<always> decoded, and therefore Catalyst
can't distinguish / vs %2F in paths (in addition to other encoded values).

=item use_request_uri_for_path => 1

This method uses the C<REQUEST_URI> and C<SCRIPT_NAME> environment variables. As C<REQUEST_URI> is never
decoded, this means that applications using this mode can correctly handle URIs including the %2F character
(i.e. with C<AllowEncodedSlashes> set to C<On> in Apache).

Given that this method of path resolution is provably more correct, it is recommended that you use
this unless you have a specific need to deploy your application in a non-standard environment, and you are
aware of the implications of not being able to handle encoded URI paths correctly.

However it also means that in a number of cases when the app isn't installed directly at a path, but instead
is having paths rewritten into it (e.g. as a .cgi/fcgi in a public_html directory, with mod_rewrite in a
.htaccess file, or when SSI is used to rewrite pages into the app, or when sub-paths of the app are exposed
at other URIs than that which the app is 'normally' based at with C<mod_rewrite>), the resolution of
C<< $c->request->base >> will be incorrect.

=back

=item *

C<using_frontend_proxy> - See L</PROXY SUPPORT>.

=item *

C<using_frontend_proxy_path> - Enabled L<Plack::Middleware::ReverseProxyPath> on your application (if
installed, otherwise log an error).  This is useful if your application is not running on the
'root' (or /) of your host server.  B<NOTE> if you use this feature you should add the required
middleware to your project dependency list since its not automatically a dependency of L<Catalyst>.
This has been done since not all people need this feature and we wish to restrict the growth of
L<Catalyst> dependencies.

=item *

C<encoding> - See L</ENCODING>

This now defaults to 'UTF-8'.  You my turn it off by setting this configuration
value to undef.

=item *

C<abort_chain_on_error_fix>

Defaults to true.

When there is an error in an action chain, the default behavior is to
abort the processing of the remaining actions to avoid running them
when the application is in an unexpected state.

Before version 5.90070, the default used to be false. To keep the old
behaviour, you can explicitly set the value to false. E.g.

    __PACKAGE__->config(abort_chain_on_error_fix => 0);

If this setting is set to false, then the remaining actions are
performed and the error is caught at the end of the chain.


=item *

C<use_hash_multivalue_in_request>

In L<Catalyst::Request> the methods C<query_parameters>, C<body_parametes>
and C<parameters> return a hashref where values might be scalar or an arrayref
depending on the incoming data.  In many cases this can be undesirable as it
leads one to writing defensive code like the following:

    my ($val) = ref($c->req->parameters->{a}) ?
      @{$c->req->parameters->{a}} :
        $c->req->parameters->{a};

Setting this configuration item to true will make L<Catalyst> populate the
attributes underlying these methods with an instance of L<Hash::MultiValue>
which is used by L<Plack::Request> and others to solve this very issue.  You
may prefer this behavior to the default, if so enable this option (be warned
if you enable it in a legacy application we are not sure if it is completely
backwardly compatible).

=item *

C<skip_complex_post_part_handling>

When creating body parameters from a POST, if we run into a multipart POST
that does not contain uploads, but instead contains inlined complex data
(very uncommon) we cannot reliably convert that into field => value pairs.  So
instead we create an instance of L<Catalyst::Request::PartData>.  If this causes
issue for you, you can disable this by setting C<skip_complex_post_part_handling>
to true (default is false).

=item *

C<skip_body_param_unicode_decoding>

Generally we decode incoming POST params based on your declared encoding (the
default for this is to decode UTF-8).  If this is causing you trouble and you
do not wish to turn all encoding support off (with the C<encoding> configuration
parameter) you may disable this step atomically by setting this configuration
parameter to true.

=item *

C<do_not_decode_query>

If true, then do not try to character decode any wide characters in your
request URL query or keywords.  Most readings of the relevant specifications
suggest these should be UTF-* encoded, which is the default that L<Catalyst>
will use, however if you are creating a lot of URLs manually or have external
evil clients, this might cause you trouble.  If you find the changes introduced
in Catalyst version 5.90080+ break some of your query code, you may disable
the UTF-8 decoding globally using this configuration.

This setting takes precedence over C<default_query_encoding>

=item *

C<do_not_check_query_encoding>

Catalyst versions 5.90080 - 5.90106 would decode query parts of an incoming
request but would not raise an exception when the decoding failed due to
incorrect unicode.  It now does, but if this change is giving you trouble
you may disable it by setting this configuration to true.

=item *

C<default_query_encoding>

By default we decode query and keywords in your request URL using UTF-8, which
is our reading of the relevant specifications.  This setting allows one to
specify a fixed value for how to decode your query.  You might need this if
you are doing a lot of custom encoding of your URLs and not using UTF-8.

=item *

C<use_chained_args_0_special_case>

In older versions of Catalyst, when more than one action matched the same path
AND all those matching actions declared Args(0), we'd break the tie by choosing
the first action defined.  We now normalized how Args(0) works so that it
follows the same rule as Args(N), which is to say when we need to break a tie
we choose the LAST action defined.  If this breaks your code and you don't
have time to update to follow the new normalized approach, you may set this
value to true and it will globally revert to the original chaining behavior.

=item *

C<psgi_middleware> - See L<PSGI MIDDLEWARE>.

=item *

C<data_handlers> - See L<DATA HANDLERS>.

=item *

C<stats_class_traits>

An arrayref of L<Moose::Role>s that get composed into your stats class.

=item *

C<request_class_traits>

An arrayref of L<Moose::Role>s that get composed into your request class.

=item *

C<response_class_traits>

An arrayref of L<Moose::Role>s that get composed into your response class.

=item *

C<inject_components>

A Hashref of L<Catalyst::Component> subclasses that are 'injected' into configuration.
For example:

    MyApp->config({
      inject_components => {
        'Controller::Err' => { from_component => 'Local::Controller::Errors' },
        'Model::Zoo' => { from_component => 'Local::Model::Foo' },
        'Model::Foo' => { from_component => 'Local::Model::Foo', roles => ['TestRole'] },
      },
      'Controller::Err' => { a => 100, b=>200, namespace=>'error' },
      'Model::Zoo' => { a => 2 },
      'Model::Foo' => { a => 100 },
    });

Generally L<Catalyst> looks for components in your Model/View or Controller directories.
However for cases when you which to use an existing component and you don't need any
customization (where for when you can apply a role to customize it) you may inject those
components into your application.  Please note any configuration should be done 'in the
normal way', with a key under configuration named after the component affix, as in the
above example.

Using this type of injection allows you to construct significant amounts of your application
with only configuration!.  This may or may not lead to increased code understanding.

Please not you may also call the ->inject_components application method as well, although
you must do so BEFORE setup.

=back

=head1 EXCEPTIONS

Generally when you throw an exception inside an Action (or somewhere in
your stack, such as in a model that an Action is calling) that exception
is caught by Catalyst and unless you either catch it yourself (via eval
or something like L<Try::Tiny> or by reviewing the L</error> stack, it
will eventually reach L</finalize_errors> and return either the debugging
error stack page, or the default error page.  However, if your exception
can be caught by L<Plack::Middleware::HTTPExceptions>, L<Catalyst> will
instead rethrow it so that it can be handled by that middleware (which
is part of the default middleware).  For example this would allow

    use HTTP::Throwable::Factory 'http_throw';

    sub throws_exception :Local {
      my ($self, $c) = @_;

      http_throw(SeeOther => { location =>
        $c->uri_for($self->action_for('redirect')) });

    }

=head1 INTERNAL ACTIONS

Catalyst uses internal actions like C<_DISPATCH>, C<_BEGIN>, C<_AUTO>,
C<_ACTION>, and C<_END>. These are by default not shown in the private
action table, but you can make them visible with a config parameter.

    MyApp->config(show_internal_actions => 1);

=head1 ON-DEMAND PARSER

The request body is usually parsed at the beginning of a request,
but if you want to handle input yourself, you can enable on-demand
parsing with a config parameter.

    MyApp->config(parse_on_demand => 1);

=head1 PROXY SUPPORT

Many production servers operate using the common double-server approach,
with a lightweight frontend web server passing requests to a larger
backend server. An application running on the backend server must deal
with two problems: the remote user always appears to be C<127.0.0.1> and
the server's hostname will appear to be C<localhost> regardless of the
virtual host that the user connected through.

Catalyst will automatically detect this situation when you are running
the frontend and backend servers on the same machine. The following
changes are made to the request.

    $c->req->address is set to the user's real IP address, as read from
    the HTTP X-Forwarded-For header.

    The host value for $c->req->base and $c->req->uri is set to the real
    host, as read from the HTTP X-Forwarded-Host header.

Additionally, you may be running your backend application on an insecure
connection (port 80) while your frontend proxy is running under SSL.  If there
is a discrepancy in the ports, use the HTTP header C<X-Forwarded-Port> to
tell Catalyst what port the frontend listens on.  This will allow all URIs to
be created properly.

In the case of passing in:

    X-Forwarded-Port: 443

All calls to C<uri_for> will result in an https link, as is expected.

Obviously, your web server must support these headers for this to work.

In a more complex server farm environment where you may have your
frontend proxy server(s) on different machines, you will need to set a
configuration option to tell Catalyst to read the proxied data from the
headers.

    MyApp->config(using_frontend_proxy => 1);

If you do not wish to use the proxy support at all, you may set:

    MyApp->config(ignore_frontend_proxy => 0);

=head2 Note about psgi files

Note that if you supply your own .psgi file, calling
C<< MyApp->psgi_app(@_); >>, then B<this will not happen automatically>.

You either need to apply L<Plack::Middleware::ReverseProxy> yourself
in your psgi, for example:

    builder {
        enable "Plack::Middleware::ReverseProxy";
        MyApp->psgi_app
    };

This will unconditionally add the ReverseProxy support, or you need to call
C<< $app = MyApp->apply_default_middlewares($app) >> (to conditionally
apply the support depending upon your config).

See L<Catalyst::PSGI> for more information.

=head1 THREAD SAFETY

Catalyst has been tested under Apache 2's threading C<mpm_worker>,
C<mpm_winnt>, and the standalone forking HTTP server on Windows. We
believe the Catalyst core to be thread-safe.

If you plan to operate in a threaded environment, remember that all other
modules you are using must also be thread-safe. Some modules, most notably
L<DBD::SQLite>, are not thread-safe.

=head1 DATA HANDLERS

The L<Catalyst::Request> object uses L<HTTP::Body> to populate 'classic' HTML
form parameters and URL search query fields.  However it has become common
for various alternative content types to be PUT or POSTed to your controllers
and actions.  People working on RESTful APIs, or using AJAX often use JSON,
XML and other content types when communicating with an application server.  In
order to better support this use case, L<Catalyst> defines a global configuration
option, C<data_handlers>, which lets you associate a content type with a coderef
that parses that content type into something Perl can readily access.

    package MyApp::Web;

    use Catalyst;
    use JSON::MaybeXS;

    __PACKAGE__->config(
      data_handlers => {
        'application/json' => sub { local $/; decode_json $_->getline },
      },
      ## Any other configuration.
    );

    __PACKAGE__->setup;

By default L<Catalyst> comes with a generic JSON data handler similar to the
example given above, which uses L<JSON::MaybeXS> to provide either L<JSON::PP>
(a pure Perl, dependency free JSON parser) or L<Cpanel::JSON::XS> if you have
it installed (if you want the faster XS parser, add it to you project Makefile.PL
or dist.ini, cpanfile, etc.)

The C<data_handlers> configuration is a hashref whose keys are HTTP Content-Types
(matched against the incoming request type using a regexp such as to be case
insensitive) and whose values are coderefs that receive a localized version of
C<$_> which is a filehandle object pointing to received body.

This feature is considered an early access release and we reserve the right
to alter the interface in order to provide a performant and secure solution to
alternative request body content.  Your reports welcomed!

=head1 PSGI MIDDLEWARE

You can define middleware, defined as L<Plack::Middleware> or a compatible
interface in configuration.  Your middleware definitions are in the form of an
arrayref under the configuration key C<psgi_middleware>.  Here's an example
with details to follow:

    package MyApp::Web;

    use Catalyst;
    use Plack::Middleware::StackTrace;

    my $stacktrace_middleware = Plack::Middleware::StackTrace->new;

    __PACKAGE__->config(
      'psgi_middleware', [
        'Debug',
        '+MyApp::Custom',
        $stacktrace_middleware,
        'Session' => {store => 'File'},
        sub {
          my $app = shift;
          return sub {
            my $env = shift;
            $env->{myapp.customkey} = 'helloworld';
            $app->($env);
          },
        },
      ],
    );

    __PACKAGE__->setup;

So the general form is:

    __PACKAGE__->config(psgi_middleware => \@middleware_definitions);

Where C<@middleware> is one or more of the following, applied in the REVERSE of
the order listed (to make it function similarly to L<Plack::Builder>:

Alternatively, you may also define middleware by calling the L</setup_middleware>
package method:

    package MyApp::Web;

    use Catalyst;

    __PACKAGE__->setup_middleware( \@middleware_definitions);
    __PACKAGE__->setup;

In the case where you do both (use 'setup_middleware' and configuration) the
package call to setup_middleware will be applied earlier (in other words its
middleware will wrap closer to the application).  Keep this in mind since in
some cases the order of middleware is important.

The two approaches are not exclusive.

=over 4

=item Middleware Object

An already initialized object that conforms to the L<Plack::Middleware>
specification:

    my $stacktrace_middleware = Plack::Middleware::StackTrace->new;

    __PACKAGE__->config(
      'psgi_middleware', [
        $stacktrace_middleware,
      ]);


=item coderef

A coderef that is an inlined middleware:

    __PACKAGE__->config(
      'psgi_middleware', [
        sub {
          my $app = shift;
          return sub {
            my $env = shift;
            if($env->{PATH_INFO} =~m/forced/) {
              Plack::App::File
                ->new(file=>TestApp->path_to(qw/share static forced.txt/))
                ->call($env);
            } else {
              return $app->($env);
            }
         },
      },
    ]);



=item a scalar

We assume the scalar refers to a namespace after normalizing it using the
following rules:

(1) If the scalar is prefixed with a "+" (as in C<+MyApp::Foo>) then the full string
is assumed to be 'as is', and we just install and use the middleware.

(2) If the scalar begins with "Plack::Middleware" or your application namespace
(the package name of your Catalyst application subclass), we also assume then
that it is a full namespace, and use it.

(3) Lastly, we then assume that the scalar is a partial namespace, and attempt to
resolve it first by looking for it under your application namespace (for example
if you application is "MyApp::Web" and the scalar is "MyMiddleware", we'd look
under "MyApp::Web::Middleware::MyMiddleware") and if we don't find it there, we
will then look under the regular L<Plack::Middleware> namespace (i.e. for the
previous we'd try "Plack::Middleware::MyMiddleware").  We look under your application
namespace first to let you 'override' common L<Plack::Middleware> locally, should
you find that a good idea.

Examples:

    package MyApp::Web;

    __PACKAGE__->config(
      'psgi_middleware', [
        'Debug',  ## MyAppWeb::Middleware::Debug->wrap or Plack::Middleware::Debug->wrap
        'Plack::Middleware::Stacktrace', ## Plack::Middleware::Stacktrace->wrap
        '+MyApp::Custom',  ## MyApp::Custom->wrap
      ],
    );

=item a scalar followed by a hashref

Just like the previous, except the following C<HashRef> is used as arguments
to initialize the middleware object.

    __PACKAGE__->config(
      'psgi_middleware', [
         'Session' => {store => 'File'},
    ]);

=back

Please see L<PSGI> for more on middleware.

=head1 ENCODING

Starting in L<Catalyst> version 5.90080 encoding is automatically enabled
and set to encode all body responses to UTF8 when possible and applicable.
Following is documentation on this process.  If you are using an older
version of L<Catalyst> you should review documentation for that version since
a lot has changed.

By default encoding is now 'UTF-8'.  You may turn it off by setting
the encoding configuration to undef.

    MyApp->config(encoding => undef);

This is recommended for temporary backwards compatibility only.

To turn it off for a single request use the L<clear_encoding>
method to turn off encoding for this request.  This can be useful
when you are setting the body to be an arbitrary block of bytes,
especially if that block happens to be a block of UTF8 text.

Encoding is automatically applied when the content-type is set to
a type that can be encoded.  Currently we encode when the content type
matches the following regular expression:

    $content_type =~ /^text|xml$|javascript$/

Encoding is set on the application, but it is copied to the context object
so that you can override it on a request basis.

Be default we don't automatically encode 'application/json' since the most
common approaches to generating this type of response (Either via L<Catalyst::View::JSON>
or L<Catalyst::Action::REST>) will do so already and we want to avoid double
encoding issues.

If you are producing JSON response in an unconventional manner (such
as via a template or manual strings) you should perform the UTF8 encoding
manually as well such as to conform to the JSON specification.

NOTE: We also examine the value of $c->response->content_encoding.  If
you set this (like for example 'gzip', and manually gzipping the body)
we assume that you have done all the necessary encoding yourself, since
we cannot encode the gzipped contents.  If you use a plugin like
L<Catalyst::Plugin::Compress> you need to update to a modern version in order
to have this function correctly  with the new UTF8 encoding code, or you
can use L<Plack::Middleware::Deflater> or (probably best) do your compression on
a front end proxy.

=head2 Methods

=over 4

=item encoding

Returns an instance of an C<Encode> encoding

    print $c->encoding->name

=item handle_unicode_encoding_exception ($exception_context)

Method called when decoding process for a request fails.

An C<$exception_context> hashref is provided to allow you to override the
behaviour of your application when given data with incorrect encodings.

The default method throws exceptions in the case of invalid request parameters
(resulting in a 500 error), but ignores errors in upload filenames.

The keys passed in the C<$exception_context> hash are:

=over

=item param_value

The value which was not able to be decoded.

=item error_msg

The exception received from L<Encode>.

=item encoding_step

What type of data was being decoded. Valid values are (currently)
C<params> - for request parameters / arguments / captures
and C<uploads> - for request upload filenames.

=back

=back

=head1 SUPPORT

IRC:

    Join #catalyst on irc.perl.org.

Mailing Lists:

    http://lists.scsys.co.uk/cgi-bin/mailman/listinfo/catalyst
    http://lists.scsys.co.uk/cgi-bin/mailman/listinfo/catalyst-dev

Web:

    http://catalyst.perl.org

Wiki:

    http://dev.catalyst.perl.org

=head1 SEE ALSO

=head2 L<Task::Catalyst> - All you need to start with Catalyst

=head2 L<Catalyst::Manual> - The Catalyst Manual

=head2 L<Catalyst::Component>, L<Catalyst::Controller> - Base classes for components

=head2 L<Catalyst::Engine> - Core engine

=head2 L<Catalyst::Log> - Log class.

=head2 L<Catalyst::Request> - Request object

=head2 L<Catalyst::Response> - Response object

=head2 L<Catalyst::Test> - The test suite.

=head1 PROJECT FOUNDER

sri: Sebastian Riedel <[e-mail]>

=head1 CONTRIBUTORS

abw: Andy Wardley

acme: Leon Brocard <[e-mail]>

abraxxa: Alexander Hartmaier <[e-mail]>

andrewalker: André Walker <[e-mail]>

Andrew Bramble

Andrew Ford <[e-mail]>

Andrew Ruthven

andyg: Andy Grundman <[e-mail]>

audreyt: Audrey Tang

bricas: Brian Cassidy <[e-mail]>

Caelum: Rafael Kitover <[e-mail]>

chansen: Christian Hansen

Chase Venters <[e-mail]>

chicks: Christopher Hicks

Chisel Wright <[e-mail]>

Danijel Milicevic <[e-mail]>

davewood: David Schmidt <[e-mail]>

David Kamholz <[e-mail]>

David Naughton <[e-mail]>

David E. Wheeler

dhoss: Devin Austin <[e-mail]>

dkubb: Dan Kubb <[e-mail]>

Drew Taylor

dwc: Daniel Westermann-Clark <[e-mail]>

esskar: Sascha Kiefer

fireartist: Carl Franks <[e-mail]>

frew: Arthur Axel "fREW" Schmidt <[e-mail]>

gabb: Danijel Milicevic

Gary Ashton Jones

Gavin Henry <[e-mail]>

Geoff Richards

groditi: Guillermo Roditi <[e-mail]>

hobbs: Andrew Rodland <[e-mail]>

ilmari: Dagfinn Ilmari Mannsåker <[e-mail]>

jcamacho: Juan Camacho

jester: Jesse Sheidlower <[e-mail]>

jhannah: Jay Hannah <[e-mail]>

Jody Belka

Johan Lindstrom

jon: Jon Schutz <[e-mail]>

Jonathan Rockway <[e-mail]>

Kieren Diment <[e-mail]>

konobi: Scott McWhirter <[e-mail]>

marcus: Marcus Ramberg <[e-mail]>

miyagawa: Tatsuhiko Miyagawa <[e-mail]>

mgrimes: Mark Grimes <[e-mail]>

mst: Matt S. Trout <[e-mail]>

mugwump: Sam Vilain

naughton: David Naughton

ningu: David Kamholz <[e-mail]>

nothingmuch: Yuval Kogman <[e-mail]>

numa: Dan Sully <[e-mail]>

obra: Jesse Vincent

Octavian Rasnita

omega: Andreas Marienborg

Oleg Kostyuk <[e-mail]>

phaylon: Robert Sedlacek <[e-mail]>

rafl: Florian Ragwitz <[e-mail]>

random: Roland Lammel <[e-mail]>

revmischa: Mischa Spiegelmock <[e-mail]>

Robert Sedlacek <[e-mail]>

rrwo: Robert Rothenberg <[e-mail]>

SpiceMan: Marcel Montes

sky: Arthur Bergman

szbalint: Balint Szilakszi <[e-mail]>

t0m: Tomas Doran <[e-mail]>

Ulf Edvinsson

vanstyn: Henry Van Styn <[e-mail]>

Viljo Marrandi <[e-mail]>

Will Hawes <[e-mail]>

willert: Sebastian Willert <[e-mail]>

wreis: Wallace Reis <[e-mail]>

Yuval Kogman <[e-mail]>

rainboxx: Matthias Dietrich <[e-mail]>

dd070: Dhaval Dhanani <[e-mail]>

Upasana <[e-mail]>

John Napiorkowski (jnap) <[e-mail]>

=head1 COPYRIGHT

Copyright (c) 2005-2015, the above named PROJECT FOUNDER and CONTRIBUTORS.

=head1 LICENSE

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

=cut

no Moose;

__PACKAGE__->meta->make_immutable;

1;