Apurdue.266
net.bugs.2bsd
utzoo!decvax!pur-ee!purdue!cak
Sat Apr 24 21:25:47 1982
Re: My experiences
Re: Extremeley slow ex/vi on 2.8

Boy, did you ever hit a sore spot there! At cincy one night (this is on
an 11/23, by the way), we started a vmstat 5 on one terminal and a vi
on another, with nothing else on the system. Typing a carriage return
caused about 180 overlay switches. It got progressively worse from there.
The 'smart' display update algorithm takes more real-time to compute
the best way to display the changes from, say, a ^D, than it would take
to repaint the whole scheme brute force.

Consultation with Carl Smith led us to the conclusion that the overlays
for the version 3 ex were tossed together one night, found to work
on a reasonably fast machine, and left untouched. Highly unsatisfactory.

I left cincy soon after this, but we never popularized ex for this
reason (almost all our users were new to Unix and didn't know any
better). Jim Reuter took over my job, and was going to look into this,
but I don't know what he's achieved. I'm sure he'll comment.

The Makefile for overlaid ex appeared early in this newsgroup, perhaps
Carl can retransmit it from the archives.

Bill Jolitz has let it be known that in 2.81, the overlay overhead should
be much less than currently (I have a number of 10% of current overhead
in mind, but I'm not at all sure that it's correct -- take it with a couple
of grains of salt).

Cheers,
Chris Kent


-----------------------------------------------------------------
 gopher://quux.org/ conversion by John Goerzen <jgoerzen@complete.org>
 of http://communication.ucsd.edu/A-News/


This Usenet Oldnews Archive
article may be copied and distributed freely, provided:

1. There is no money collected for the text(s) of the articles.

2. The following notice remains appended to each copy:

The Usenet Oldnews Archive: Compilation Copyright (C) 1981, 1996 
 Bruce Jones, Henry Spencer, David Wiseman.