bug in parser/parts/text_parser.php

[sebastienlhaire]

in method finish() charset conversion should not be activated when text is already utf-8.

my version:

    /**
     * Returns the ezcMailText part corresponding to the parsed message.
     *
     * @return ezcMailText
     */
    public function finish()
    {
        $charset = "us-ascii"; // RFC 2822 default
        if ( isset( $this->headers['Content-Type'] ) )
        {
            preg_match( '/\s*charset\s?=\s?"?([^;"\s]*);?/i',
                            $this->headers['Content-Type'],
                            $parameters );
            if ( count( $parameters ) > 0 )
            {
                $charset = strtolower( trim( $parameters[1], '"' ) );
            }
        }

        $encoding = strtolower( $this->headers['Content-Transfer-Encoding'] );
        if ( $encoding == ezcMail::QUOTED_PRINTABLE )
        {
            $this->text = quoted_printable_decode( $this->text );
        }
        else if ( $encoding == ezcMail::BASE64 )
        {
            $this->text = base64_decode( $this->text );
        }
        if ($charset != 'utf-8') {
            $this->text = ezcMailCharsetConverter::convertToUTF8($this->text, $charset);
        }

        $part = new ezcMailText( $this->text, 'utf-8', ezcMail::EIGHT_BIT, $charset );
        $part->subType = $this->subType;
        $part->setHeaders( $this->headers->getCaseSensitiveArray() );
        ezcMailPartParser::parsePartHeaders( $this->headers, $part );
        $part->size = strlen( $this->text );
        return $part;
    }

[derickr]

Do you have an example email that shows this problem? We are going to need that to make a patch and a test for this.

[sebastienlhaire]

Unfortunately I cannot remember which mail caused the problem. I have rolled back the code to your latest version and tried successfully to parse various mails. If I come across a mail which can reproduce the bug, I send it to you.

[derickr]

Is this still a live issue for you? If not, I will close it in the next few days.

[derickr]

Closing this out due to inactivity