garrett_damore@tadpole.com wrote:

> but everyone else said "just implement it and get some real-world
> results".  So I'm doing that.

Well, in that case, you don't have to change whole attachment
before waiting comments. Changes to com.c, comvar.h, and one
example of attachments are enough to review real code.

> If you (or anyone else) comes up with a much better way to do this, then
> you  (or whoever else) is more than free to either change what I've done
> or ask core@ to reject my proposal with an alternate proposal.  I don't
> really care either way.

From my experience on NetBSD, rushing to commit without enough
period often caused unpleasure flame. That is my concern.
Actuall I won't object your COM_INIT_REGS so strongly if
some other people say it's acceptable. Fortunately I have
spare time to see/respond your mail in these two days,
but I don't think all people have time in such short period.

How do you think the way yamt@ did on his many changes recently?
It's quite reasonable and causes few unfortunate results, I think.
---
Izumi Tsutsui