The basic problem is that while your program is busy importing the data, it cannot tell Tk to do updates to its windows.

The easiest solution if you can manage that is to spawn off a separate program to do the actual data import. That way, you can even create a way to do fully unattended data imports without a GUI.

Another way would be to spawn a thread which does the importing, but using Tk with threads is likely to cause you no end of fancy problems because the resources allocated to Tk in the main thread will (or will not) get shared or recreated in the spawned thread. The same likely holds true when forking a child process.

Another thing you could do is to modify your data import loop to look like this:

my $start_time = time(); # remember when we started
my $progress;

while (<$input_file>) {
    ...
    if (time - $start_time > 2) { # 2 seconds have passed
        $progress = create_progress_dialog();
        Tk->update(); # I don't know if there is such a thing or wheth
+er Tk needs it
    };
};

undef $progress; # hide progress dialog if we have one
[download]

Alternatively, you can try to use the Tk facilities like Tk->after():

use Time::HiRes qw(time);

my $progress;
my $popup = Tk->after( 2, sub { $progress = create_progress_dialog } )
+;

my $last_poll = time();
while (<$input_file>) {
    # let Tk process its events every 0.1 seconds:
    if (time - $last_poll > 0.1) {
        Tk->process_one_event();
    };
    ...
};

if ($progress) {
    undef $progress;
};
undef $popup; # cancel our Tk->after() timeout
[download]

That way, Tk can also respond when the user clicks on a window or tries to minimize/move it.

Thank you. Your method "Modifying your data import loop" works great. I noticed now, that much of the importing time in my script is due to a "sort" I must due during importing. Any way to keep track also of this time? What I have so far is following (after implementig your idea):

sub import {
$dbh = DBI->connect( "dbi:SQLite:files/database/data3.db" ) || die "Ca
+nnot connect: $DBI::errstr";

my $start_time = time(); # remember when we started

foreach my $row_db ( sort { deaccent($a->[($selected_order+2)]) cmp de
+accent($b->[($selected_order+2)]) or $a->[($selected_order+2)] cmp $b
+->[($selected_order+2)] } @$all_select_glossary_orderd ) {

my ($ID, $a, $b,$c, $d, $e, $f) = @$row_db;

### DOING HERE SOMETIHN WITH MY DATA (IMPORT/PRINTING)

    if (time - $start_time > 2) { # 2 seconds have passed
      display_note();
      $mw->update();
    }    
}
$dbh->disconnect;
}


#subrutine for better alphab. sorting
sub deaccent {
    my $in = $_[0];
    return lc($in) unless ( $in =~ y/\xC0-\xFF// ); #short circuit if 
+no upper chars
    # translterate what we can (for speed)
    $in =~ tr/ÀÁÂÃÄÅàáâãäåÇçÈÉÊËèéêëÌÍÎÏìíîïÒÓÔÕÖØòóôõöøÑñÙÚÛÜùúûüÝÿý/
+AAAAAAaaaaaaCcEEEEeeeeIIIIiiiiOOOOOOooooooNnUUUUuuuuYyy/;
    # and substitute the rest
    #my %trans = qw(Æ AE æ ae Þ TH þ th Ð TH ð th ß ss);
    #$in =~ s/([ÆæÞþÐðß])/$trans{$1}/g;
    $in =~ tr/'//d; # d for delete
    return lc($in);
}
[download]

You can also move your updating right into the block passed to sort:

    my $last_time = time;
    sort {
            if (time - $last_time > 1) {
                $mw->update();
            };
            $a <=> $b
         } @list;
[download]

Alternatively, think about caching the search comparisons or creating a string key to sort on. That might be much faster than repeated field accesses:

for (@items) {
    push @$_, deaccent( $_->[$selected+2] ) . $_->[$selected+2];
};
@items = sort { $a->[-1] cmp $b->[-1] } @items;
[download]