changes so that we skip encoding under programmatic situations

[catagits/Catalyst-Runtime.git] / lib / Catalyst / Request / Upload.pm

diff --git a/lib/Catalyst/Request/Upload.pm b/lib/Catalyst/Request/Upload.pm
index d8e58be..6df2dff 100644 (file)
--- a/lib/Catalyst/Request/Upload.pm
+++ b/lib/Catalyst/Request/Upload.pm
@@ -15,6 +15,8 @@ has size => (is => 'rw');
 has tempname => (is => 'rw');
 has type => (is => 'rw');
 has basename => (is => 'ro', lazy_build => 1);
+has raw_basename => (is => 'ro', lazy_build => 1);
+has charset => (is=>'ro', predicate=>'has_charset');
 
 has fh => (
   is => 'rw',
@@ -29,17 +31,21 @@ has fh => (
       Catalyst::Exception->throw(
           message => qq/Can't open '$filename': '$!'/ );
     }
-
     return $fh;
   },
 );
 
 sub _build_basename {
+    my $basename = shift->raw_basename;
+    $basename =~ s|[^\w\.-]+|_|g;
+    return $basename;
+}
+
+sub _build_raw_basename {
     my $self = shift;
     my $basename = $self->filename;
     $basename =~ s|\\|/|g;
     $basename = ( File::Spec::Unix->splitpath($basename) )[2];
-    $basename =~ s|[^\w\.-]+|_|g;
     return $basename;
 }
 
@@ -58,13 +64,16 @@ Catalyst::Request::Upload - handles file upload requests
     $upload->basename;
     $upload->copy_to;
     $upload->fh;
+    $upload->decoded_fh
     $upload->filename;
     $upload->headers;
     $upload->link_to;
     $upload->size;
     $upload->slurp;
+    $upload->decoded_slurp;
     $upload->tempname;
     $upload->type;
+    $upload->charset;
 
 To specify where Catalyst should put the temporary files, set the 'uploadtmp'
 option in the Catalyst config. If unset, Catalyst will use the system temp dir.
@@ -97,10 +106,56 @@ sub copy_to {
     return File::Copy::copy( $self->tempname, @_ );
 }
 
+=head2 $upload->is_utf8_encoded
+
+Returns true of the upload defines a character set at that value is 'UTF-8'.
+This does not try to inspect your upload and make any guesses if the Content
+Type charset is undefined.
+
+=cut
+
+sub is_utf8_encoded {
+    my $self = shift;
+    if(my $charset = $self->charset) {
+      return $charset eq 'UTF-8' ? 1 : 0;
+    }
+    return 0;
+}
+
 =head2 $upload->fh
 
 Opens a temporary file (see tempname below) and returns an L<IO::File> handle.
 
+This is a filehandle that is opened with no additional IO Layers.
+
+=head2 $upload->decoded_fh(?$encoding)
+
+Returns a filehandle that has binmode set to UTF-8 if a UTF-8 character set
+is found. This also accepts an override encoding value that you can use to
+force a particular L<PerlIO> layer.  If neither are found the filehandle is
+set to :raw.
+
+This is useful if you are pulling the file into code and inspecting bit and
+maybe then sending those bits back as the response.  (Please not this is not
+a suitable filehandle to set in the body; use C<fh> if you are doing that).
+
+Please note that using this method sets the underlying filehandle IO layer
+so once you use this method if you go back and use the C<fh> method you
+still get the IO layer applied.
+
+=cut
+
+sub decoded_fh {
+    my ($self, $layer) = @_;
+    my $fh  = $self->fh;
+
+    $layer = ":encoding(UTF-8)" if !$layer && $self->is_utf8_encoded;
+    $layer = ':raw' unless $layer;
+
+    binmode($fh, $layer);
+    return $fh;
+}
+
 =head2 $upload->filename
 
 Returns the client-supplied filename.
@@ -127,13 +182,17 @@ sub link_to {
 
 Returns the size of the uploaded file in bytes.
 
-=head2 $upload->slurp
+=head2 $upload->slurp(?$encoding)
+
+Optionally accepts an argument to define an IO Layer (which is applied to
+the filehandle via binmode; if no layer is defined the default is set to
+":raw".
 
 Returns a scalar containing the contents of the temporary file.
 
 Note that this will cause the filehandle pointed to by C<< $upload->fh >> to
 be reset to the start of the file using seek and the file handle to be put
-into binary mode.
+into whatever encoding mode is applied.
 
 =cut
 
@@ -158,9 +217,39 @@ sub slurp {
     return $content;
 }
 
+=head2 $upload->decoded_slurp(?$encoding)
+
+Works just like C<slurp> except we use C<decoded_fh> instead of C<fh> to
+open a filehandle to slurp.  This means if your upload charset is UTF8
+we binmode the filehandle to that encoding.
+
+=cut
+
+sub decoded_slurp {
+    my ( $self, $layer ) = @_;
+    my $handle = $self->decoded_fh($layer);
+
+    my $content = undef;
+    $handle->seek(0, IO::File::SEEK_SET);
+    while ( $handle->sysread( my $buffer, 8192 ) ) {
+        $content .= $buffer;
+    }
+
+    $handle->seek(0, IO::File::SEEK_SET);
+    return $content;
+}
+
 =head2 $upload->basename
 
-Returns basename for C<filename>.
+Returns basename for C<filename>.  This filters the name through a regexp
+C<basename =~ s|[^\w\.-]+|_|g> to make it safe for filesystems that don't
+like advanced characters.  This will of course filter UTF8 characters.
+If you need the exact basename unfiltered use C<raw_basename>.
+
+=head2 $upload->raw_basename
+
+Just like C<basename> but without filtering the filename for characters that
+don't always write to a filesystem.
 
 =head2 $upload->tempname
 
@@ -170,6 +259,11 @@ Returns the path to the temporary file.
 
 Returns the client-supplied Content-Type.
 
+=head2 $upload->charset
+
+The character set information part of the content type, if any.  Useful if you
+need to figure out any encodings on the file upload.
+
 =head2 meta
 
 Provided by Moose