UTF-8 file reading and writing for PHP

I'm creating a page that:
- accepts user input in whatever language
- saves that input to a file
- reads the file and displays the original input

The following code successfully writes the user input to a file (when I
open the file, it's in the correct font), but I can't get PHP to read
the file and display the correct characters.

HTML --------------- Form
<FORM name=saveform method=post action="wiki.php">
File:
<TEXTAREA name=thetext rows=20 cols=30></TEXTAREA>
</TEXTAREA>
<INPUT type=submit>
<INPUT type=hidden name=action value="save">
</FORM>

PHP --------------- Sticks the data in a file
$message = $_REQUEST['thetext'];
echo $message; // This displays the correct stuff
$filename = "tmp/tmp.txt";
$fr = fopen($filename, "wb+");
// adding header
fwrite($fr, pack("CCC",0xef,0xbb,0xbf));
fputs($fr, $message);
fclose($fr);

PHP --------------- Read the data from the file
$thefile = file($filename);
array_shift($thefile); //To get rid of the BOM
$ret = "";
foreach ($thefile as $i => $line) {
$line = rtrim(utf8_decode($line));
$ret .= $line;
}
echo $ret; // This _doesn't_ display the correct stuff

Thanks for the reply...

My goal is to allow user input in UTF-8, in Arabic script, for example.
I then save what they input to a file. Then I'd like to retrieve and
print out the original stuff that they wrote.

I've tried various variations of utf8_encode() and utf8_decode() and
even without them, and every time, the resulting stuff is just ????? or
other weird characters.

Bent Stigsen wrote:
> HaggMan wrote:
> > I'm creating a page that:
> > - accepts user input in whatever language
> > - saves that input to a file
> > - reads the file and displays the original input
> >
> > The following code successfully writes the user input to a file (when I
> > open the file, it's in the correct font), but I can't get PHP to read
> > the file and display the correct characters.
> [snip]
> > PHP --------------- Sticks the data in a file
> > $message = $_REQUEST['thetext'];
> > echo $message; // This displays the correct stuff
> > $filename = "tmp/tmp.txt";
> > $fr = fopen($filename, "wb+");
> > // adding header
> > fwrite($fr, pack("CCC",0xef,0xbb,0xbf));
>
> Is it safe to assume the data to be UTF-8?
>
> If you just discard the byteordermark later, there's little reason to
> add it (if there ever was).
>
> > fputs($fr, $message);
> > fclose($fr);
> >
> > PHP --------------- Read the data from the file
> > $thefile = file($filename);
> > array_shift($thefile); //To get rid of the BOM
>
> BOM = 3 bytes
> $thefile = array of lines of text terminated by newline.
>
> > $ret = "";
> > foreach ($thefile as $i => $line) {
> > $line = rtrim(utf8_decode($line));
>
> I am not sure what to make of this. If you expect the browser to send
> data in utf-8, then I would assume you serve your pages in utf-8, then
> why convert the text to iso8859-1?
>
> > $ret .= $line;
> > }
> > echo $ret; // This _doesn't_ display the correct stuff
>
> Start with simple file_put_contents and file_get_contents.
>
>
> /Bent