Hi, I think there's a difference between nmh headers the user might typically add manually in a draft, e.g. Bcc, and those that are more the mechanics of something they drive another way, e.g. Attach and the whatnow prompt. (Will the new Forward grow a `forward' whatnow(1) command in time BTW?)
There seems to be two ideals. Not cooking up headers that may clash. Not letting our headers, however they're formed, leak. I was wondering about an illegal header prefix so that whatever's used we can ensure it doesn't leak. It could also be more terse for those that don't want to type `Nmh-'. from: tom to: dick .bcc: harry .forward: +inbox 42 314 1718 It would have to be post(8) that's checking for `dot' headers since it isn't send(1) that processes all of them; some remain for post, assuming that's the postproc. `Dot files' are familiar from ls(1) treating them as `hidden'. But! There are no illegal header field characters; RFC 2822 says anything from 33 to 126, except a colon, is OK. So we'd still be trampling, but having a shorter prefix to indicate it's internal, never to leak. (Though I agree "never" is doubtful, and there are alternate postprocs to mess up.) BTW, RFC 2822 defines Bcc, including that it can remain in the sent email with no values to indicate to the non-blind recipients that blinds were sent. So given the only illegal header is `:Forward: +inbox 42', and David's traceability argument of having the world knock on our door, I think I'm still in favour of an `Nmh-' prefix. -- Cheers, Ralph. https://plus.google.com/+RalphCorderoy _______________________________________________ Nmh-workers mailing list Nmh-workers@nongnu.org https://lists.nongnu.org/mailman/listinfo/nmh-workers