use warnings;
use CGI qw/:standard/;
use CGI::Carp 'fatalsToBrowser';
use POSIX;


print header, start_html('upload form');
my $filename = "text.txt";
binmode($filename);

print start_form(), table(
Tr(
    td("File: "),
    td(filefield(-name=>'upload',
                -size=>50,
                -maxlength=>80),
    ),
    ),
Tr(
    td(),
    td(submit('button','submit'),
)
)
),
end_form(),
hr;

if (param()) {
my $upload =  param('upload');

    while (<$filename>) {
       print;
    }


    open (OUTFILE,">>/home/myname/public_html/upload/text.txt") || die
+ $!;
    binmode(OUTFILE);
    
    while ($bytesread=read($filename,$buffer,1024)) {
       print OUTFILE $buffer;
       print "the thing is open";
    }
    
print "The file should be uploaded if everything went right, which it 
+probably didn't.\n";
}
[download]

May I suggest trying the documentation again? CGI.pm is exceedingly well documented. The examples you copied are highly incorrect. In particular:

my $filename = "text.txt";
binmode($filename);
[download]

Calling binmode on a string doesn't do anything useful -- it only does something interesting on filehandles.

while (<$filename>) {
       print;
}
[download]

Attempting to read from a string doesn't do anything useful -- it's a symbolic reference, but since you don't open a filehandle anywhere called 'text.txt', you won't get any data.

while ($bytesread=read($filename,$buffer,1024)) {
       print OUTFILE $buffer;
       print "the thing is open";
}
[download]

Again, $filename isn't a filehandle or even a symbolic reference to a filehandle. Even if it were, you'd have read everything from it in the while loop above. Besides that, mixing buffered reads with unbuffered generally mixes things up.

The CGI documentation is *really* good about explaining how to receive uploaded files. It's well-worth your time to figure out how to read it rather than trying to copy and paste several other scripts together. That's obviously not working here. Why not try something different?

I think you may be slightly confused on the whole files thing. A file is simply a bunch of bytes stored in a accessible little package on a drive - there's no 'magic' involved. A 'text' file is just a file which (a) will generally only contain nice 'readable' ASCII characters (plus some line endings) and (b) *may* indicate that it contains this by having a 'txt' or 'asc' or somesuch extension.

When a browser uploads a file to your server as part of a form, it simply reads the local bytes, munges them all into an HTTP-friendly packet and sends off the whole lot. What CGI.pm does for you is split off those bytes, save them to a temporary file and then give you access to a filehandle - basically a way of reading the bytes back in from the temporary file and doing somewhere else with them - writing back out to a file of your choosing, parsing, dividing-by-the-number-you-first-thought-of...whatever.

In order to have some idea of what the file actually contains - whether it's bunch of bytes that make up a picture, or the aforementioned nice'n'readable text file, you either need to examine the also-supplied-by-CGI.pm MIME type for the file, which will tell you whether it's a gif, an mp3, a text file or whatever. Generally, one does this in order to have an idea of the extension to then give the filename - never trust user-supplied filenames :)

Hope this helps

Ben