Encoding problem(reposting in more detail)

[Nik]

Same question that remaines unanswered although i ahve asked it in many irc channels and forum places is why the greek file name string arent UTF8 as i saved them when creating the file. It it were then it would have benn appeared correctly in the popup menu and they wouldnt require 3 re-encodingsa back and forth from utf8 => iso-8859-7 and so on. here is the code again:

my @files = <../data/text/*.txt>;
my @display_files = map /([^\/]+)\.txt/, @files;
Encode::from_to($_, "ISO-8859-7", "utf8") for @display_files; #1st con
+version in order to appear correctly in the popup menu

print br;
print start_form( action=>'index.pl' );
    print h1( {class=>'lime'}, "&#917;&#960;&#941;&#955;&#949;&#958;&#
+949; &#964;&#959; &#954;&#949;&#943;&#956;&#949;&#957;&#959; &#960;&#
+959;&#965; &#963;&#949; &#949;&#957;&#948;&#953;&#945;&#966;&#941;&#9
+61;&#949;&#953; => ",
                                popup_menu( -name=>'select', -values=>
+\@display_files ),
                                submit('&#917;&#956;&#966;&#940;&#957;
+&#953;&#963;&#951;'));
print end_form;

my $passage = param('select') || "&#913;&#961;&#967;&#953;&#954;&#942;
+ &#931;&#949;&#955;&#943;&#948;&#945;!";
Encode::from_to($passage, "utf8", "ISO-8859-7") if param();
#2nd conversion in order for the user selected file from the popup men
+u to be able to be opened

if ( param('select') )
{
    open(FILE, "<../data/text/$passage.txt") or die $!;
         local $/;
         $data = <FILE>;
    close(FILE);
    
    Encode::from_to($passage, "ISO-8859-7", "utf8");
#3nd conversion in order for the user selected file from the popup men
+u to be able to be inserted as "UTF8" in the database.

    $select = $dbh->prepare( "UPDATE guestlog SET passage=?, date=?, c
+ounter=counter+1 WHERE host=?" );
    $select->execute( $passage, $date, $host );
}
else more code

As you can see this re-encoding stuff its becaming very tredious and not only to this script(index.pl) but in other as well.

[john_oshea]

It appears that you've got filenames in ISO-8859-7, so I'm wondering if you couldn't just serve the html page(s) out with ISO-8859-7 encoding as well? This would reduce the number of from_to conversions to the final one that inputs the contents of $passage into the database.

I've not seen your previous postings, so apologies in advance if this is repeating previous comments -hope that helps.

[Nik]

Well that can work and i though of it but i want to keep using UTF8 and if i find a reason of wwhy nto greek file strings arent UTF8 be default as i save them and correct ti then i wont have to use encoding at all.

[john_oshea]

OK, I've done some digging in your previous posts and have found this snippet from you:

Because windows are unable of saving greek filenames also in utf8 format(they save only the file contents iam afraid)

which makes me wonder if Windows is saving the filenames correctly and that you actually have a display issue in cmd.exe. Googling for 'windows greek filenames cmd.exe' gets me this as one of the results, which has, about halfway down, the following:

While the reason for their existance might be odd, whitespaces, dots, percentage signs are perfectly legal characters for file names in Windows 2000 (Though they weren't in DOS). The only illegal characters for file names in Windows are: \ / : * ? " < > | Anything else is OK, including international characters which may seem odd to you if your FS is FAT and your codepage is not set correctly (or simply because you don't speak russian/greek/hebrew/arabic/etc).

Unfortunately I don't 'do' Windows well enough to confirm if this is the issue or not, but, at the very least, something like this should get you started.

Hope that helps