On 22 Feb 2005 at 10:46, Schupp Roderich (extern) Com MD PD SWP 2 wrote:
> > I am impressed with pp and PAR, however I have come across a
> > problem concerning utf-8 data. The script uni2hex.pl
> > recognizes the utf-8 characters when run on rh9 using their
> > perl. The output is given in uni2hex.pl_output. The problem
> > is that after running pp the resulting executable ran in the
> > same directory in the same machine does not recognize the
> > utf-8 characters but splits them into there ansii parts. The
> > output from a.out is given in a.out_output.
> > ...
> > This may be because of something different about the setup of
> > perl on rh9, if so it would be useful to be able to include
> > this feature in executables made by pp.
>
> I can confirm that behaviour on Debian Sid (Perl 5.8.4),
> so it's not a Redhat specific problem. I had to set
> PERL_UNICODE=CSL in the environment and also set LANG
> to a utf8 locale in order for uni2hex.pl to produce the
> expected output, though. The pp generated executable would
> not recognize utf8 characters even with these settings.
>
> Cheers, Roderich
I'm running Win32 and I am not strong on Unicode in Perl. I haven't had any
luck
getting uni2hex.pl to run with utf8 behavior by default, no matter what I try
for
environment variables.
However, I have gotten the same results as uni2hex.pl_output by explicitly
stating utf8
in the code:
my $file = 'test_utf8_data';
open(TEST,"<:utf8", $file) or die('aaaaagh!!');
binmode STDOUT, ':raw :utf8';
while(<TEST>){
$cur = $_;
chomp $cur;
my $testl = length($cur);
print "$cur is $testl long\n";
my @letters = split //,"$cur";
for my $loop(0..$#letters){
my $try1 = $letters[$loop];
my $number1 = ord($try1);
my $hex = sprintf "%1x",$number1;
print "$try1 is $hex\n";
}
}
close(TEST);
It works for me on Win32 by itself or as a pp packed app, no environment
variables
needed. Note that I used ':raw' in:
binmode STDOUT, ':raw :utf8';
only to eliminate the Win32 default of crlf, for comparison to
uni2hex.pl_output.
If I could get this to default under Win32, perhaps I could figure out what's
missing
when pp packed.
Alan
