Dancer::Request - interface for accessing incoming requests


version 1.3521


This class implements a common interface for accessing incoming requests in a Dancer application.

In a route handler, the current request object can be accessed by the request method, like in the following example:

    get '/foo' => sub {
        request->params; # request, params parsed as a hash ref
        request->body; # returns the request body, unparsed
        request->path; # the path requested by the client
        # ...

A route handler should not read the environment by itself, but should instead use the current request object.



The constructor of the class, used internally by Dancer's core to create request objects.

It uses the environment hash table given to build the request object:

    Dancer::Request->new(env => \%ENV);

It also accepts the is_forward boolean flag, if the new request object is the result of a forward.


Used internally to define some default values and parse parameters.

new_for_request($method, $path, $params, $body, $headers)

An alternate constructor convenient for test scripts which creates a request object with the arguments given.

forward($request, $new_location)

Create a new request which is a clone of the current one, apart from the path location, which points instead to the new location. This is used internally to chain requests using the forward keyword.

Note that the new location should be a hash reference. Only one key is required, the to_url, that should point to the URL that forward will use. Optional values are the key params to a hash of parameters to be added to the current request parameters, and the key options that points to a hash of options about the redirect (for instance, method pointing to a new request method).


Flag that will be set to true if the request has been forwarded.


Return a string representing the request object (eg: "GET /some/path")


Return the HTTP method used by the client to access the application.

While this method returns the method string as provided by the environment, it's better to use one of the following boolean accessors if you want to inspect the requested method.


Return the IP address of the client.


Return the remote host of the client. This only works with web servers configured to do a reverse DNS lookup on the client's IP address.


Return the protocol (HTTP/1.0 or HTTP/1.1) used for the request.


Return the port of the server.


An alias to request_uri()


Return the raw, undecoded request URI path.


Return remote user if defined.


Return script_name from the environment.


Return the scheme of the request


Return true of false, indicating whether the connection is secure


Return true if the method requested by the client is 'GET'


Return true if the method requested by the client is 'HEAD'


Return true if the method requested by the client is 'PATCH'


Return true if the method requested by the client is 'POST'


Return true if the method requested by the client is 'PUT'


Return true if the method requested by the client is 'DELETE'


Return the path requested by the client.


Returns an absolute URI for the base of the application. Returns a URI object (which stringifies to the URL, as you'd expect).


Same thing as base above, except it removes the last trailing slash in the path if it is the only path.

This means that if your base is http://myserver/, uri_base will return http://myserver (notice no trailing slash). This is considered very useful when using templates to do the following thing:

    <link rel="stylesheet" href="<% request.uri_base %>/css/style.css" />

uri_for(path, params)

Constructs a URI from the base and the passed path. If params (hashref) is supplied, these are added to the query string of the uri. If the base is http://localhost:5000/foo, request->uri_for('/bar', { baz => 'baz' }) would return http://localhost:5000/foo/bar?baz=baz. Returns a URI object (which stringifies to the URL, as you'd expect).


Called in scalar context, returns a hashref of params, either from the specified source (see below for more info on that) or merging all sources.

So, you can use, for instance:

    my $foo = params->{foo}

If called in list context, returns a list of key => value pairs, so you could use:

    my %allparams = params;

If the incoming form data contains multiple values for the same key, they will be returned as an arrayref.

Fetching only params from a given source

If a required source isn't specified, a mixed hashref (or list of key value pairs, in list context) will be returned; this will contain params from all sources (route, query, body).

In practical terms, this means that if the param foo is passed both on the querystring and in a POST body, you can only access one of them.

If you want to see only params from a given source, you can say so by passing the $source param to params():

    my %querystring_params = params('query');
    my %route_params       = params('route');
    my %post_params        = params('body');

If source equals route, then only params parsed from the route pattern are returned.

If source equals query, then only params parsed from the query string are returned.

If source equals body, then only params sent in the request body will be returned.

If another value is given for $source, then an exception is triggered.


Alias to the params accessor, for backward-compatibility with CGI interface.


Alias to the method accessor, for backward-compatibility with CGI interface.


Alias to the PSGI input handle (<request->env->{psgi.input}>)


Return the content type of the request.


Return the content length of the request.


Return the value of the given header, if present. If the header has multiple values, return a list with its values if called in list context, or the first one if called in scalar context.


Returns the HTTP::Header object used to store all the headers.


Return the raw body of the request, unparsed.

NOTE: the behaviour of this keyword has changed. Originally, the entire raw request body was kept in RAM for this accessor, but that's not ideal if you handle large requests (file uploads, etc), so in 1.3143 that was ditched, and the body accessor replaced by a convenience method which would get the temp file handle that HTTP::Body uses, read it for you and return the content, so that if you did want the raw body, it was there. However, HTTP::Body only creates a temp file for certain types of request, leading to unpredictable behaviour and confusion - see issue #1140.

So, handling of the raw request body is now controlled by a configuration setting, raw_request_body_in_ram, which controls whether or not the raw request body will be kept in RAM when it's parsed; if this is set to a false value, then the body accessor will not return anything, giving you lower memory usage, at the cost of not having access to the raw (unparsed) request body.


Return true if the value of the header X-Requested-With is XMLHttpRequest.


Return the current environment as a hashref.

Note that a request's environment is not always reflected by the global variable %ENV (e.g., when running via Plack::Handler::FCGI). In consequence, it is recommended to always rely on the values returned by env(), and not to access %ENV directly.


Returns a reference to a hash containing uploads. Values can be either a Dancer::Request::Upload object, or an arrayref of Dancer::Request::Upload objects.

You should probably use the upload($name) accessor instead of manually accessing the uploads hash table.


Context-aware accessor for uploads. It's a wrapper around an access to the hash table provided by uploads(). It looks at the calling context and returns a corresponding value.

If you have many file uploads under the same name, and call upload('name') in an array context, the accessor will unroll the ARRAY ref for you:

    my @uploads = request->upload('many_uploads'); # OK

Whereas with a manual access to the hash table, you'll end up with one element in @uploads, being the ARRAY ref:

    my @uploads = request->uploads->{'many_uploads'}; # $uploads[0]: ARRAY(0xXXXXX)

That is why this accessor should be used instead of a manual access to uploads.


Given a request to these are the values returned by the various request-> method calls:

  uri          /request-methods?a=1
  request_uri  /request-methods?a=1
  path         /request-methods
  path_info    /request-methods
  method       GET
  port         5000
  protocol     HTTP/1.1
  scheme       http

HTTP environment variables

All HTTP environment variables that are in %ENV will be provided in the Dancer::Request object through specific accessors, here are those supported:

agent (alias for user_agent)

Looks for HTTP_X_FORWARDED_FOR if X_FORWARDED_FOR is not there.


If you app is on a non-standard port, you can expect this to return the hostname and port, e.g.



This module has been written by Alexis Sukrieh and was mostly inspired by Plack::Request, written by Tatsuiko Miyagawa.

Tatsuiko Miyagawa also gave a hand for the PSGI interface.


This module is released under the same terms as Perl itself.




Dancer Core Developers


This software is copyright (c) 2010 by Alexis Sukrieh.

This is free software; you can redistribute it and/or modify it under the same terms as the Perl 5 programming language system itself.