The previous few messages are my perception of issues recently argued. They
include my bias -- that is, I am likely to forget about anything I don't
want to argue about. Worse, I'm doing this from memory.
I do not believe we had concensus on any of these issues and would like to
cover them in the next draft.
Please post any issues that I've missed (a seperate subject line would be
helpful, by the way).
(I would have sent this introduction on the first message, if only I'd
gotten the address on the first message right ...)
The require/support commands must be evaluated before the language is
parsed if they are to be used usefully. If you don't parse them, the
extension keywords look like syntax errors. This is broken. It is not
consistant with whichever part of the draft demanded the script be
syntactically correct before running it.
Rob's post looks to be right to me. I'd like to see some other mechanism.
What, I'm not sure, but something else; would real block commands ("{",
"}", or "BEGIN"/"END" if there are any pascal bigots on the list), fix
anything? (I think it helps, provided someone doesn't do something
stupid with control structures.) Can we do something more useful with
require (all requires must be stated in the first line of the script)?
(ugh)
(for those recent subscribers, the archive is under
<http://www.imc.org/ietf-mta-filters/>.)
--
Tim Showalter
tjs(_at_)andrew(_dot_)cmu(_dot_)edu