To replicate: 1. Forward a message with Content-Type: multipart/alternative; that has one part text/plain and the other text/html 2. The received message garbles the content parts Specifically, for the text/html part that is forwarded: 1. "Content-Transfer-Encoding" is forced to 7bit irrespective of original encoding. 2. "Content-Type" charset info is dropped Also, since the text part may be modified and HTML part is not, it can be argued that the text/html part should have the additional Content-Disposition: attachment added with a suggested filename of "html message part.html" or something appropriate so that the multipart/mixed Content-Type does not leave a room for interpretation since I know at least Outlook prefers to show the original untouched HTML part by default and shows the modified text part as attachment (I believe RFC allows the MUA to decide how to interpret multipart/mixed as well).
Changes related to this bug have been committed. Please check latest CVS and update the bug accordingly. You can also get the patch from: http://www.colino.net/claws-mail/ 2010-12-22 [colin] 3.7.8cvs20 * src/compose.c Store charset if available when attaching text/* parts Fixes bug #2203, 'Problems in forwarding multipart/ alternative messages with HTML part'
Changes related to this bug have been committed. Please check latest CVS and update the bug accordingly. You can also get the patch from: http://www.colino.net/claws-mail/ 2010-12-23 [colin] 3.7.8cvs23 * src/compose.c Finish fixing bug #2203: When forwarding (inline) and there are some text/* parts with no name, set them a name and set their disposition to attachment, so that recipients with MUAs preferring HTML parts don't show them by default.
Thanks Colin!
Changes related to this bug have been committed. Please check latest CVS and update the bug accordingly. You can also get the patch from: http://www.claws-mail.org/tracker/ 2012-08-12 [colin] 3.8.1cvs29 * src/compose.c Make sure text/* attachments are not made inline, regardless of the compose mode. Finishes fixing bug #2203 after auto-save switches mode to COMPOSE_REEDIT. Thanks to Michael Gmelin for finding this.