Re: [Nmh-workers] mh-format multiply function

[Ken Hornstein]

>But one of the the trade offs relevant to the existing or prospective nmh user
>is the extent to which nmh still supports "All power to the user". mh-format
>constitutes a real barrier to "All power to the user", for all but the most
>sophisticated of users. I also agree that any further significant effort on
>mh-format is probably ill advised.

I understand that concern; believe me, I do.  That's one reason we've been
talking about providing more examples for mh-format, so rather than having
to write one from scratch you could just use one of the examples (or
just take pieces from one of the examples).

>The option would be -exec procedure_name, or if you like, -eval
>procedure_name. If present, then procedure_name would be invoked for each
>message. procedure_name's stdout would completely replace each scan line. It
>would not redirect stderr. A non-zero exit status would be an error. For each
>component, comp, of a message, it would define an environment variable,
>NMH_FORMAT_comp, whose value was the content of that component.

I think that has most of the same security problems that the equivalent
function in mh-format has.  Also, there are some things that would make
it hard to re-implement the existing scan format (like %(mymbox)).
It's probably solvable, but I personally don't think it's worth it.

--Ken