Should there be some guidance for the 'host' value when a SIEVE
script is running on the localhost?
I'm not sure what you're getting at. You're running on the "local" host pretty
much by definition, I'd say.
Are you asking about what to do when the name of the host you're running on
isn't known or has been given some silly default name like "localhost"? In such
cases I think it would be best to regard "host" as unsupported.
In the long run, should "ihave" replace "requires" since it subsumes
that functionality?
Well, it can't replace it completely since you have to require ihave...
I don't really know what the end game will be here. My plan is to move forward
with this as an experimental specification, not standards track. I'm reluctant
to push for standards-track status because past experience with mechanisms like
this has not been great - too many devils in the details.
I'm hoping that given knowledge of, say, how PostScript's use of conditionals
in comments didn't really pan out (implying that the mechanism needs to be well
integrated into the language), we can get the details right this time. But only
time will te, as ihave has beenll. And even if we get the details right,
there's no guarantee of uptake. The require approach is simpler, and maybe
nobody will care enough about the ability to use the same script in dissimilar
environments to want to deal with the added complexity.
Ned