> Ok. I've attached the message here. It has a Content-Type charset
> specified. It also specifies a Content-Transfer-Encoding header. Both were
> inserted by mutt. What courier did was not to change the headers, but to
> re-order them, and that caused the signature to become invalid. So, the
> question is, why did courier change the order of the headers when they were
> perfectly valid?

 Simple: it mustn't.

 sendmail had the same bug, and they fixed it in 8.9.3.

Reply via email to