Ok, this actually happens on my FreeBSD system, but maybe I'l just throw it
out here for fun-
(Don't wanna resub to the FreeBSD lists)

But my "slow vi" problem is only on bootup, when it's "recovering vi editor
sessions"-
This box is a 486-50, and works perfectly as a router/nat box, but goddamn-
Sometimes when it boots, it sits at "recovering vi editor sessions" for
about 5 minutes.
Other times, it takes 10 seconds (regardless of any vi usage as far as I
could tell)
I have done the usual homework and can not find out exactly what is
happening in this part of the bootup sequence-
Any ideas?
Or is this a FreeBSD-specific issue?
--

"Behold the illusion of control..."
----- Original Message -----
From: "Matthias Buelow" <mkb@mukappabeta.de>
To: "Henry Nelson" <henry@irm.nara.kindai.ac.jp>
Cc: <netbsd-help@netbsd.org>
Sent: Monday, April 02, 2001 6:33 PM
Subject: Re: best way to speed up "vi"


> Henry Nelson <henry@irm.nara.kindai.ac.jp> writes:
>
> >> how long is a "most irritating amount of time" circa?
> >> If it's longer than 3-4 seconds it really is irritating since
> >
> >Oh, yes, longer than 3-4 seconds.  I just loaded this file,
> >"1699 Apr  2 19:26 hub.mc," and it took over 7 seconds.  A larger
> >file like 20-40 Kb really takes a long time.
>
> then something's badly wrong, a p60 should be about 5x as fast
> as the vs3100 I was speaking about... especially when the time
> gets significantly longer with "larger" filesizes (it really
> shouldn't matter at all with 40K sized files.)
>
> >> Does "eval `/usr/bin/tset -sQ`" make any difference?
> >
> >Actually, it did: about 1 second faster!  Could I stick that in
> >my .login file?
>
> yes, of course (the NetBSD default dotfiles do it), what it does
> to give some speedup (on BSD) with starting termcap based programs
> is put the termcap entry for $TERM into the TERMCAP variable so that
> the termcap/curses code doesn't have to grovel through the termcap
> file everytime to find the terminal descriptions.
> To me it looks like you have some very pathetic disk transfer speeds...
>
> mkb