PHP: Send email with attachment and UTF8 in filenames

Guttorm · 02-17-2012, 04:07 AM

Hi

I am trying to send email with a file attachment. Everything is UTF8. Here is some code:

PHP Code:



require_once 'Mail.php'; 
require_once 'Mail/mime.php'; 
require_once 'Mail/mail.php'; 
 
$to = "recipient@example.org"; 
$from = "sender@example.org"; 
$subject = "æøå-subject"; 
$message = "æøå-message"; 
$attachment_data = "æøå-filecontent"; 
$attachment_filename = utf8_decode("æøå-name.txt"); 
send_mail_utf8_with_attachment($to, $from, $subject, $message ,$attachment_data, $attachment_filename); 
 
function send_mail_utf8_with_attachment($to, $from, $subject, $message = "",$attachment_data="", $attachment_filename="") 
{ 
    $params = array(); 
    $params['head_charset'] = 'utf-8'; 
    $params['text_charset'] = 'utf-8'; 
    $params['html_charset'] = 'utf-8'; 
    $params['eol'] = "\n"; 
    $hdrs = array( 
              'From'    => $from, 
              'Subject' => $subject 
    ); 
    $mime = new Mail_mime($params); 
    $mime->setTXTBody($message); 
    if (strlen($attachment_data) && strlen($attachment_filename)) { 
        $mimeType = "application/octet-stream"; 
        $mime->addAttachment($attachment_data, $mimeType, $attachment_filename, false); 
    } 
    $body = $mime->get(); 
    $hdrs = $mime->headers($hdrs); 
    $mail =& Mail::factory('mail'); 
    return $mail->send($to, $hdrs, $body); 
}

The PHP file has UTF-8 encoding. Note the utf8_decode of $attachment_filename. Without it all email clients get it wrong. With it, some get it right (like GMail) but other still get it wrong, or replaces the filename with some default name. Also, I can't understand why the utf8_decode helps at all.

Everything else except the filename of the attachment is fine. If the filename contains just regular ascii, there is no problem.

I tried to read about it in the Pear documentation:

http://pear.php.net/manual/en/packag...attachment.php

But I don't understand it. Anyone have an idea what to do with $attachment_filename? Or what parameters to pass to tell that the filename is UTF8?

JSkywalker · 02-17-2012, 07:56 AM

Quote:

$subject = "æøå-subject";
$message = "æøå-message";
$attachment_data = "æøå-filecontent";

I think all headers in your mail-message are not allowed to have UTF-8 characters in them.

Guttorm · 02-17-2012, 08:34 AM

Well that's simply incorrect. As I wrote, everything except the filename of the attachment is fine. Also I can send and receive emails with UTF8 in filenames just fine. I just don't know how to do it with PHP.

Cedrik · 02-17-2012, 12:41 PM

Why not just try a test with all your utf8 strings decoded ?

PHP Code:



$from = "sender@example.org";
$subject = utf8_decode("æøå-subject");
$message = utf8_decode("æøå-message");
$attachment_data = utf8_decode("æøå-filecontent");
$attachment_filename = utf8_decode("æøå-name.txt");

JSkywalker · 02-17-2012, 01:02 PM

It indeed works, the subject becomes this:

Quote:

Subject: =?utf-8?Q?=C3=A6=C3=B8=C3=A5-subject?=

And the message + attachment are looking ok (when received with Thunderbird, and then 'view source', from 'Date:'-down....)

Quote:

Date: Fri, 17 Feb 2012 19:54:25 +0100 (CET)

--=_92896147cb2a02b45588d261d65a21fd
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain; charset=utf-8

=C3=A6=C3=B8=C3=A5-message
--=_92896147cb2a02b45588d261d65a21fd
Content-Transfer-Encoding: base64
Content-Type: application/octet-stream;
name*=US-ASCII''%E6%F8%E5-name.txt
Content-Disposition: attachment;
filename*=US-ASCII''%E6%F8%E5-name.txt

w6bDuMOlLWZpbGVjb250ZW50
--=_92896147cb2a02b45588d261d65a21fd--

But you also wrote:

Quote:

Without it all email clients get it wrong. With it, some get it right (like GMail) but other still get it wrong, or replaces the filename with some default name

Which client did get it wrong? (name, please ...)

Guttorm · 02-17-2012, 01:21 PM

Thanks. Good idea to check the source in the emails. The US_ASCII part seems incorrect, but I guess some email clients will understand that it's UTF-8 and some not?

I tried to use an email client called Firstclass that we use at work, and I sent an email to a guy who uses Outlook. Both got the filename in the attachment wrong. I'll try to investigate more on Monday when I'm at work. Did the strange letters look right with Thunderbird?

JSkywalker · 02-17-2012, 04:27 PM

I looked ath this message in WindowsLiveMail, and the attachment had the wrong name....(ATT.00xxxx.dat)

When i opened the attachment, it did not look like an UTF-8 formatted textfile, but it looked like a non-utf-8-text.

I dont know Firstclass stuff, but looking at their site, it did not look like they support a 'normal' IMAP-server, so i cannot test this