shell bypass 403
package Mojo::Message::Request;
use Mojo::Base 'Mojo::Message';
use Digest::SHA qw(sha1_base64);
use Mojo::Cookie::Request;
use Mojo::Util qw(b64_encode b64_decode sha1_sum);
use Mojo::URL;
my ($SEED, $COUNTER) = ($$ . time . rand, int rand 0xffffff);
has env => sub { {} };
has method => 'GET';
has [qw(proxy reverse_proxy)];
has request_id => sub {
my $b64
= substr(sha1_base64($SEED . ($COUNTER = ($COUNTER + 1) % 0xffffff)), 0, 8);
$b64 =~ tr!+/!-_!;
return $b64;
};
has url => sub { Mojo::URL->new };
has via_proxy => 1;
sub clone {
my $self = shift;
# Dynamic requests cannot be cloned
return undef unless my $content = $self->content->clone;
my $clone = $self->new(
content => $content,
method => $self->method,
url => $self->url->clone,
version => $self->version
);
$clone->{proxy} = $self->{proxy}->clone if $self->{proxy};
return $clone;
}
sub cookies {
my $self = shift;
# Parse cookies
my $headers = $self->headers;
return [map { @{Mojo::Cookie::Request->parse($_)} } $headers->cookie]
unless @_;
# Add cookies
my @cookies = map { ref $_ eq 'HASH' ? Mojo::Cookie::Request->new($_) : $_ }
$headers->cookie || (), @_;
$headers->cookie(join '; ', @cookies);
return $self;
}
sub every_param { shift->params->every_param(@_) }
sub extract_start_line {
my ($self, $bufref) = @_;
# Ignore any leading empty lines
return undef unless $$bufref =~ s/^\s*(.*?)\x0d?\x0a//;
# We have a (hopefully) full request-line
return !$self->error({message => 'Bad request start-line'})
unless $1 =~ /^(\S+)\s+(\S+)\s+HTTP\/(\d\.\d)$/;
my $url = $self->method($1)->version($3)->url;
my $target = $2;
return !!$url->host_port($target) if $1 eq 'CONNECT';
return !!$url->parse($target)->fragment(undef) if $target =~ /^[^:\/?#]+:/;
return !!$url->path_query($target);
}
sub fix_headers {
my $self = shift;
$self->{fix} ? return $self : $self->SUPER::fix_headers(@_);
# Host
my $url = $self->url;
my $headers = $self->headers;
$headers->host($url->host_port) unless $headers->host;
# Basic authentication
if ((my $info = $url->userinfo) && !$headers->authorization) {
$headers->authorization('Basic ' . b64_encode($info, ''));
}
# Basic proxy authentication
return $self unless (my $proxy = $self->proxy) && $self->via_proxy;
return $self unless my $info = $proxy->userinfo;
$headers->proxy_authorization('Basic ' . b64_encode($info, ''))
unless $headers->proxy_authorization;
return $self;
}
sub get_start_line_chunk {
my ($self, $offset) = @_;
$self->_start_line->emit(progress => 'start_line', $offset);
return substr $self->{start_buffer}, $offset, 131072;
}
sub is_handshake { lc($_[0]->headers->upgrade // '') eq 'websocket' }
sub is_secure {
my $url = shift->url;
return ($url->protocol || $url->base->protocol) eq 'https';
}
sub is_xhr {
(shift->headers->header('X-Requested-With') // '') =~ /XMLHttpRequest/i;
}
sub param { shift->params->param(@_) }
sub params {
my $self = shift;
return $self->{params}
||= $self->body_params->clone->append($self->query_params);
}
sub parse {
my $self = shift;
my ($env, $chunk) = ref $_[0] ? (shift, '') : (undef, shift);
# Parse CGI environment
$self->env($env)->_parse_env($env) if $env;
# Parse normal message
if (($self->{state} // '') ne 'cgi') { $self->SUPER::parse($chunk) }
# Parse CGI content
else { $self->content($self->content->parse_body($chunk))->SUPER::parse('') }
# Check if we can fix things that require all headers
return $self unless $self->is_finished;
# Base URL
my $base = $self->url->base;
$base->scheme('http') unless $base->scheme;
my $headers = $self->headers;
if (!$base->host && (my $host = $headers->host)) { $base->host_port($host) }
# Basic authentication
if (my $basic = _basic($headers->authorization)) { $base->userinfo($basic) }
# Basic proxy authentication
my $basic = _basic($headers->proxy_authorization);
$self->proxy(Mojo::URL->new->userinfo($basic)) if $basic;
# "X-Forwarded-Proto"
$base->scheme('https')
if $self->reverse_proxy
&& ($headers->header('X-Forwarded-Proto') // '') eq 'https';
return $self;
}
sub query_params { shift->url->query }
sub start_line_size { length shift->_start_line->{start_buffer} }
sub _basic { $_[0] && $_[0] =~ /Basic (.+)$/ ? b64_decode $1 : undef }
sub _parse_env {
my ($self, $env) = @_;
# Bypass normal message parser
$self->{state} = 'cgi';
# Extract headers
my $headers = $self->headers;
my $url = $self->url;
my $base = $url->base;
for my $name (keys %$env) {
my $value = $env->{$name};
next unless $name =~ s/^HTTP_//i;
$name =~ y/_/-/;
$headers->header($name => $value);
# Host/Port
$value =~ s/:(\d+)$// ? $base->host($value)->port($1) : $base->host($value)
if $name eq 'HOST';
}
# Content-Type is a special case on some servers
$headers->content_type($env->{CONTENT_TYPE}) if $env->{CONTENT_TYPE};
# Content-Length is a special case on some servers
$headers->content_length($env->{CONTENT_LENGTH}) if $env->{CONTENT_LENGTH};
# Query
$url->query->parse($env->{QUERY_STRING}) if $env->{QUERY_STRING};
# Method
$self->method($env->{REQUEST_METHOD}) if $env->{REQUEST_METHOD};
# Scheme/Version
$base->scheme($1) and $self->version($2)
if ($env->{SERVER_PROTOCOL} // '') =~ m!^([^/]+)/([^/]+)$!;
# HTTPS
$base->scheme('https') if uc($env->{HTTPS} // '') eq 'ON';
# Path
my $path = $url->path->parse($env->{PATH_INFO} ? $env->{PATH_INFO} : '');
# Base path
if (my $value = $env->{SCRIPT_NAME}) {
# Make sure there is a trailing slash (important for merging)
$base->path->parse($value =~ m!/$! ? $value : "$value/");
# Remove SCRIPT_NAME prefix if necessary
my $buffer = $path->to_string;
$value =~ s!^/|/$!!g;
$buffer =~ s!^/?\Q$value\E/?!!;
$buffer =~ s!^/!!;
$path->parse($buffer);
}
}
sub _start_line {
my $self = shift;
return $self if defined $self->{start_buffer};
# Path
my $url = $self->url;
my $path = $url->path_query;
$path = "/$path" unless $path =~ m!^/!;
# CONNECT
my $method = uc $self->method;
if ($method eq 'CONNECT') {
my $port = $url->port // ($url->protocol eq 'https' ? '443' : '80');
$path = $url->ihost . ":$port";
}
# Proxy
elsif ($self->proxy && $self->via_proxy && $url->protocol ne 'https') {
$path = $url->clone->userinfo(undef) unless $self->is_handshake;
}
$self->{start_buffer} = "$method $path HTTP/@{[$self->version]}\x0d\x0a";
return $self;
}
1;
=encoding utf8
=head1 NAME
Mojo::Message::Request - HTTP request
=head1 SYNOPSIS
use Mojo::Message::Request;
# Parse
my $req = Mojo::Message::Request->new;
$req->parse("GET /foo HTTP/1.0\x0d\x0a");
$req->parse("Content-Length: 12\x0d\x0a");
$req->parse("Content-Type: text/plain\x0d\x0a\x0d\x0a");
$req->parse('Hello World!');
say $req->method;
say $req->headers->content_type;
say $req->body;
# Build
my $req = Mojo::Message::Request->new;
$req->url->parse('http://127.0.0.1/foo/bar');
$req->method('GET');
say $req->to_string;
=head1 DESCRIPTION
L<Mojo::Message::Request> is a container for HTTP requests, based on
L<RFC 7230|http://tools.ietf.org/html/rfc7230>,
L<RFC 7231|http://tools.ietf.org/html/rfc7231>,
L<RFC 7235|http://tools.ietf.org/html/rfc7235> and
L<RFC 2817|http://tools.ietf.org/html/rfc2817>.
=head1 EVENTS
L<Mojo::Message::Request> inherits all events from L<Mojo::Message>.
=head1 ATTRIBUTES
L<Mojo::Message::Request> inherits all attributes from L<Mojo::Message> and
implements the following new ones.
=head2 env
my $env = $req->env;
$req = $req->env({PATH_INFO => '/'});
Direct access to the C<CGI> or C<PSGI> environment hash if available.
# Check CGI version
my $version = $req->env->{GATEWAY_INTERFACE};
# Check PSGI version
my $version = $req->env->{'psgi.version'};
=head2 method
my $method = $req->method;
$req = $req->method('POST');
HTTP request method, defaults to C<GET>.
=head2 proxy
my $url = $req->proxy;
$req = $req->proxy(Mojo::URL->new('http://127.0.0.1:3000'));
Proxy URL for request.
=head2 reverse_proxy
my $bool = $req->reverse_proxy;
$req = $req->reverse_proxy($bool);
Request has been performed through a reverse proxy.
=head2 request_id
my $id = $req->request_id;
$req = $req->request_id('aee7d5d8');
Request ID, defaults to a reasonably unique value.
=head2 url
my $url = $req->url;
$req = $req->url(Mojo::URL->new);
HTTP request URL, defaults to a L<Mojo::URL> object.
# Get request information
my $info = $req->url->to_abs->userinfo;
my $host = $req->url->to_abs->host;
my $path = $req->url->to_abs->path;
=head2 via_proxy
my $bool = $req->via_proxy;
$req = $req->via_proxy($bool);
Request can be performed through a proxy server.
=head1 METHODS
L<Mojo::Message::Request> inherits all methods from L<Mojo::Message> and
implements the following new ones.
=head2 clone
my $clone = $req->clone;
Return a new L<Mojo::Message::Request> object cloned from this request if
possible, otherwise return C<undef>.
=head2 cookies
my $cookies = $req->cookies;
$req = $req->cookies(Mojo::Cookie::Request->new);
$req = $req->cookies({name => 'foo', value => 'bar'});
Access request cookies, usually L<Mojo::Cookie::Request> objects.
# Names of all cookies
say $_->name for @{$req->cookies};
=head2 every_param
my $values = $req->every_param('foo');
Similar to L</"param">, but returns all values sharing the same name as an
array reference.
# Get first value
say $req->every_param('foo')->[0];
=head2 extract_start_line
my $bool = $req->extract_start_line(\$str);
Extract request-line from string.
=head2 fix_headers
$req = $req->fix_headers;
Make sure request has all required headers.
=head2 get_start_line_chunk
my $bytes = $req->get_start_line_chunk($offset);
Get a chunk of request-line data starting from a specific position. Note that
this method finalizes the request.
=head2 is_handshake
my $bool = $req->is_handshake;
Check C<Upgrade> header for C<websocket> value.
=head2 is_secure
my $bool = $req->is_secure;
Check if connection is secure.
=head2 is_xhr
my $bool = $req->is_xhr;
Check C<X-Requested-With> header for C<XMLHttpRequest> value.
=head2 param
my $value = $req->param('foo');
Access C<GET> and C<POST> parameters extracted from the query string and
C<application/x-www-form-urlencoded> or C<multipart/form-data> message body. If
there are multiple values sharing the same name, and you want to access more
than just the last one, you can use L</"every_param">. Note that this method
caches all data, so it should not be called before the entire request body has
been received. Parts of the request body need to be loaded into memory to parse
C<POST> parameters, so you have to make sure it is not excessively large.
There's a 16MiB limit for requests by default.
=head2 params
my $params = $req->params;
All C<GET> and C<POST> parameters extracted from the query string and
C<application/x-www-form-urlencoded> or C<multipart/form-data> message body,
usually a L<Mojo::Parameters> object. Note that this method caches all data, so
it should not be called before the entire request body has been received. Parts
of the request body need to be loaded into memory to parse C<POST> parameters,
so you have to make sure it is not excessively large. There's a 16MiB limit for
requests by default.
# Get parameter names and values
my $hash = $req->params->to_hash;
=head2 parse
$req = $req->parse('GET /foo/bar HTTP/1.1');
$req = $req->parse({PATH_INFO => '/'});
Parse HTTP request chunks or environment hash.
=head2 query_params
my $params = $req->query_params;
All C<GET> parameters, usually a L<Mojo::Parameters> object.
# Turn GET parameters to hash and extract value
say $req->query_params->to_hash->{foo};
=head2 start_line_size
my $size = $req->start_line_size;
Size of the request-line in bytes. Note that this method finalizes the request.
=head1 SEE ALSO
L<Mojolicious>, L<Mojolicious::Guides>, L<https://mojolicious.org>.
=cut