Re: [Encode::Guess] ambiguous result !?

Christophe,

Thanks for your mail.  So far Encode::Guess does not guess the  
encoding for filehandle not because it is impossible but because  
Encode::Guess takes a very conservative -- even paranoia --  
approach.  For example, Ambiguity raises exception, not preferred  
encoding.

One way to enable guessing filehandle goes like this.

sub open_and_guess{
   my $filename = shift;
   open my $fh, "<:raw", $filename or return; # or die if you like
   my $head = <$fh>; # may not work for UTF-(16|32);
   my $enc  = guess_encoding($head);
   ref $enc or die $enc; # or return $enc
   my $encname = $enc->name;
   seek $fh, 0, 0;
   binmode $fh, ":encoding($encname)";
   return $fh;
}

Here we open, guess, and reopen but this does not work for general  
case; Not all files are seekable and reopenable (i.e. pipes and  
sockets).

To know more about Encode::Guess, try

http://search.cpan.org/~dankogai/Encode-2.12/lib/Encode/Guess.pm

Yours,

Dan the Encode Maintainer


On Oct 24, 2005, at 19:12 , HERMIER Christophe wrote:

> Hello,
>
> I am using the Encode::Guess module to detect the encoding of a  
> file before opening it.
> Basically I believe that I can have various sorts of unicode  
> encodings or latin-1.
>
> What I want to get is an encoding string to give back to "open".
> my code goes like this :
>
> my $codage      = guess_encoding ( $debut );
>
> if ( UNIVERSAL::isa ( $codage, "Encode::utf8" ) )
> {
>         $codage = ":utf8";
> }
> elsif ( UNIVERSAL::isa ( $codage, "Encode::Unicode" ) )
> {
>         $codage = ":encoding(utf-16)";
> }
> else
> {
>         $codage = ":encoding(iso-8859-1)";
> }
>
>
> The problem is with the "Encode::Unicode" case : I don't know if it  
> is UTF16-LE ou UTF16-BE
> and it could even be UTF32....
>
> Is there a way to know that ???
>
>
> BTW, I checked your homepage (<http://www.dan.co.jp/>http:// 
> www.dan.co.jp/) first but it does not seem to work ?
>
> Regards,
> Christophe.
>
>
>