Hassan,

From the error it looks like the message has invalid/corrupted email headers. If this is the case, SpamFilter will always reject the email as it can't be processed. If you can have access to this email in a different way and forward it to us we can double-check to ensure there is nothing else that may be wrong.

Hi Roberto,
I sent the same mail that hassan is talking about to a gmail address and it arrives there.

Here are the headers

From: "XYZ" <info@xyzde>
To: "Atif Ghaffar" <atif.ghaffar@abccom>
Subject: =?iso-8859-1?Q?Insider-Geheimnis=20f=FCr=20die=20Verdoppelung=20oder=20Verdreifachung=20Ihres=20Umsatze=?s!?=

The subject without encoding is :
Insider-Geheimnis für die Verdoppelung oder Verdreifachung Ihres Umsatzes!

If I replace the ü with u then the message goes though.

If I keep the umlaut character and shorten the subject then the message goes through too.
For example:
Subject: Insider-Geheimnis für
Encoded: Subject: =?iso-8859-1?Q?Insider-Geheimnis=20f=FCr?=

Thanks for looking into it. We have many customers from different countries who use different umlauts which must be encoded.

best regards
--
Atif

Atif,

We're able to replicate this, and is obviously related to the encoding. Please allow us a short time to pinpoint the issue, I'll update this post as soon as we have a solution/fix.

Atif,

I'm afraid the problem looks like the subject has an incorrect ISO encoding. The source you posted to the thread is:

Subject: =?iso-8859-1?Q?Insider-Geheimnis=20f=FCr=20die=20Verdoppelung=20oder=20Verdreifachung=20Ihres=20Umsatze=?s!?=

However the text after the "=" sign must be two hexadecimal characters (0-9, a-f). In the sample above, it's ?s, which is invalid, and causes SpamFilter to reject the email due to an exception.

If in the Subject line above you replace the

Umsatze=?s!?=

with

Umsatzes!?=

then everything should be fine (the ?= indicates the end of the encoding).

So far this seems unrelated to the umlaut that appears at the beginning of the email.

Please however do let me know if you believe this is incorrect, or if the copy/paste operation in the forum does not result in the same email's source that is having the problem. Encodings are tricky, and it is possible I may be missing something here.

Here is the function that does the Quoted_Printable thing. which some print commands from me to show the result.
/**
* Encode a given string with the QUOTED_PRINTABLE mechanism
*
* @param string $str
* @param int $lineLength Defaults to {@link LINELENGTH}
* @param int $lineEnd Defaults to {@link LINEEND}
* @return string
*/
public static function encodeQuotedPrintable($str,
$lineLength = self::LINELENGTH,
$lineEnd = self::LINEEND){
$out = '';
print __LINE__ .":$str\n";

129:Insider-Geheimnis für die Verdoppelung oder Verdreifachung Ihres Umsatzes!


$str = str_replace('=', '=3D', $str);
print __LINE__ .":$str\n";

131:Insider-Geheimnis für die Verdoppelung oder Verdreifachung Ihres Umsatzes!


$str = str_replace(self::$qpKeys, self::$qpReplaceValues, $str);
print __LINE__ .":$str\n";

133:Insider-Geheimnis f=FCr die Verdoppelung oder Verdreifachung Ihres Umsatzes!


$str = rtrim($str);
print __LINE__ .":$str\n";

135:Insider-Geheimnis f=FCr die Verdoppelung oder Verdreifachung Ihres Umsatzes!



// Split encoded text into separate lines
while ($str) {
    $ptr = strlen($str);
    if ($ptr > $lineLength) {
    $ptr = $lineLength;
    }

    // Ensure we are not splitting across an encoded character
    $pos = strrpos(substr($str, 0, $ptr), '=');
    if ($pos !== false && $pos >= $ptr - 2) {
    $ptr = $pos;
    }

    // Check if there is a space at the end of the line and rewind
    if ($ptr > 0 && $str[$ptr - 1] == ' ') {
    --$ptr;
    }

    // Add string and continue
    $out .= substr($str, 0, $ptr) . '=' . $lineEnd;
    $str = substr($str, $ptr);
}

$out = rtrim($out, $lineEnd);
$out = rtrim($out, '=');
print __LINE__ .":$out\n";

162:Insider-Geheimnis f=FCr die Verdoppelung oder Verdreifachung Ihres Umsatze=
s!


return $out;
}

Roberto,
I have upgraded ZendFramework to version 1.7.6 and this problem now seems to be fixed.
Thanks a lot for debugging this with us.

best regards