Retrieve REQUEST_URI before calling ->canonical since that decodes URI

[catagits/HTTP-Request-AsCGI.git] / lib / HTTP / Request / AsCGI.pm

diff --git a/lib/HTTP/Request/AsCGI.pm b/lib/HTTP/Request/AsCGI.pm
index 4356c93..274dda9 100644 (file)
--- a/lib/HTTP/Request/AsCGI.pm
+++ b/lib/HTTP/Request/AsCGI.pm
@@ -5,10 +5,14 @@ use warnings;
 use bytes;
 use base 'Class::Accessor::Fast';
 
+our $VERSION = '1.1';
+
 use Carp;
 use HTTP::Response;
 use IO::Handle;
 use IO::File;
+use URI ();
+use URI::Escape ();
 
 __PACKAGE__->mk_accessors(qw[ environment request stdin stdout stderr ]);
 
@@ -23,6 +27,13 @@ __PACKAGE__->mk_accessors(qw[ environment request stdin stdout stderr ]);
 
 *enviroment = \&environment;
 
+my %reserved = map { sprintf('%02x', ord($_)) => 1 } split //, $URI::reserved;
+sub _uri_safe_unescape {
+    my ($s) = @_;
+    $s =~ s/%([a-fA-F0-9]{2})/$reserved{lc($1)} ? "%$1" : pack('C', hex($1))/ge;
+    $s
+}
+
 sub new {
     my $class   = shift;
     my $request = shift;
@@ -43,6 +54,9 @@ sub new {
     $uri->port(80)          unless $uri->port;
     $uri->host_port($host)  unless !$host || ( $host eq $uri->host_port );
 
+    # Get it before canonicalized so REQUEST_URI can be as raw as possible
+    my $request_uri = $uri->path_query;
+
     $uri = $uri->canonical;
 
     my $environment = {
@@ -59,11 +73,18 @@ sub new {
         REMOTE_ADDR       => '127.0.0.1',
         REMOTE_HOST       => 'localhost',
         REMOTE_PORT       => int( rand(64000) + 1000 ),                   # not in RFC 3875
-        REQUEST_URI       => $uri->path_query,                            # not in RFC 3875
+        REQUEST_URI       => $request_uri,                                # not in RFC 3875
         REQUEST_METHOD    => $request->method,
         @_
     };
 
+    # RFC 3875 says PATH_INFO is not URI-encoded. That's really
+    # annoying for applications that you can't tell "%2F" vs "/", but
+    # doing the partial decoding then makes it impossible to tell
+    # "%252F" vs "%2F". Encoding everything is more compatible to what
+    # web servers like Apache or lighttpd do, anyways.
+    $environment->{PATH_INFO} = URI::Escape::uri_unescape($environment->{PATH_INFO});
+
     foreach my $field ( $request->headers->header_field_names ) {
 
         my $key = uc("HTTP_$field");
@@ -138,7 +159,7 @@ sub setup {
 
     {
         no warnings 'uninitialized';
-        %ENV = %{ $self->environment };
+        %ENV = (%ENV, %{ $self->environment });
     }
 
     if ( $INC{'CGI.pm'} ) {
@@ -163,7 +184,7 @@ sub response {
         $headers .= $line;
         last if $headers =~ /\x0d?\x0a\x0d?\x0a$/;
     }
-    
+
     unless ( defined $headers ) {
         $headers = "HTTP/1.1 500 Internal Server Error\x0d\x0a";
     }
@@ -190,7 +211,7 @@ sub response {
         $response->code($code);
         $response->message($message);
     }
-    
+
     my $length = ( stat( $self->stdout ) )[7] - tell( $self->stdout );
 
     if ( $response->code == 500 && !$length ) {
@@ -288,36 +309,36 @@ __END__
     use CGI;
     use HTTP::Request;
     use HTTP::Request::AsCGI;
-    
+
     my $request = HTTP::Request->new( GET => 'http://www.host.com/' );
     my $stdout;
-    
+
     {
         my $c = HTTP::Request::AsCGI->new($request)->setup;
         my $q = CGI->new;
-        
+
         print $q->header,
               $q->start_html('Hello World'),
               $q->h1('Hello World'),
               $q->end_html;
-        
+
         $stdout = $c->stdout;
-        
+
         # environment and descriptors will automatically be restored
         # when $c is destructed.
     }
-    
+
     while ( my $line = $stdout->getline ) {
         print $line;
     }
-    
+
 =head1 DESCRIPTION
 
 Provides a convenient way of setting up an CGI environment from an HTTP::Request.
 
 =head1 METHODS
 
-=over 4 
+=over 4
 
 =item new ( $request [, key => value ] )
 
@@ -326,7 +347,7 @@ by optional pairs of environment key and value.
 
 =item environment
 
-Returns a hashref containing the environment that will be used in setup. 
+Returns a hashref containing the environment that will be used in setup.
 Changing the hashref after setup has been called will have no effect.
 
 =item setup