[FEATURE] indexing arbitrary headers

[David Bremner]

Łukasz Stelmach <l.stelmach at samsung.com> writes:

> I'd like to ask for a new feature: indexing of arbitrary headers. Not
> all headers but a few selected by users.
>
> For example, I get a lot of mails from a Gerrit system. I'd like to keep
> them for a while and remove them when they are old enough. Although
> these messages are distinguishable in my inbox, they've got subjects
> like "Change in ..." and are from "nobody". But these may happen in
> other e-mails too. The only 100% sure way to distringuish those e-mails
> is to look for headers Gerrit adds.
>

Thanks for writing this up. Similar requests have been around for a
while, but it's good to have a note in nmbug.

There are two main blockers I'm aware of (other than someone writing and
reviewing the code).

1) We need to index all copies of a message (e.g. using
   something like the series at  id:20170507124012.30188-2-david at tethera.net)

2) we need to decide what happens when someone changes the list of
   indexed headers.

d