perhaps better "repl -annotate" architecture?

[Alan J Rosenthal]

Greetings,

I'm finding the way "repl -annotate" works increasingly annoying for two
similar reasons: off-line mail writing on my laptop on the train, and
suspended e-mail messages (when you answer 'q' to "What now?" and resume
it later). In both cases, the connection to the original message being
replied to gets lost, because by the time the message is sent, later,
I'm no longer running the original 'repl' command.

It seems to me that the message draft itself should contain an indication of
what message this is a reply to, for the purposes of annotation. Rather than
the current system in which the state is not sufficiently connected to the
draft message.

I think that "repl -annotate" should add a new header field to the draft
message, let's call it "Acc" for the purposes of this message, e.g. the
line would say "Acc: inbox/12345". Then send would act on this header field
and do the annotation, just as it now acts on fields such as Fcc. That is,
the actual annotation would be done by send, not by repl (and by forw and by
dist). So you could resume a suspended message and still get the annotation.

I'm wondering
1) whether this is a good idea
2) if there's something already existing similar to the above
3) or if I implemented the above, would the/any maintainer of nmh eventually
maybe take my patches?
4) and any opinions or guidance on how this should be implemented (I'm
thinking an {acc} variable for the creation of the draft mesage but then
this would need to be added to all *comps files)

thanks,
ajr