Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!iuvax!uxc.cso.uiuc.edu!tank!eecae!cps3xx!usenet
From: usenet@cps3xx.UUCP (Usenet file owner)
Newsgroups: comp.emacs
Subject: MicroEmacs users read this
Message-ID: <3597@cps3xx.UUCP>
Date: 29 Jun 89 21:40:33 GMT
Reply-To: jhl@frith.egr.msu.edu ()
Organization: Michigan State University, Engineering, E. Lansing
Lines: 35

I have found the following non-bug problem in MicroEmacs 3.09e under
SCO XENIX 286 and SCO XENIX 386 which may cause others problems
under other OSs as well.

I found that when two separate copies of MicroEmacs were running
on the machine whether from two virtual terminals or two real
terminals etc, that output would come out in a bursty, jerky
fashion. I could type and then characters would come back at
me in chunks of 6-12 (or so) characters. After some source code 
reading, I found that I had  #define TYPEAH 1 in estruct.h. I
tried a compile with #define TYPEAH 0 and the problem disappeared.
Checking the output of prof(1) on the version with TYPEAH 1, I
found that 80% of the execution time was spent in read(). With
TYPEAH 0, this time was reduced to 3%.

I used to use MicroEmacs on an IBM RT Model 135 and didn't seem to
see this behavior happen (it may have and I just didn't notice).
But it definitely happens on my current work machine, 20MHz 386DX
with SCO XENIX 386 v2.3.1

So my suggestion to any UNIX users is #define TYPEAH 0. Perhaps 
this suggestion could be noted in the next release in the
compilation notes or as a comment in estruct.h. (What do you think
Dan?)


On a related note: when MicroEmacs 3.10 was first posted, people claimed
it didn't work under XENIX, has anyone come up with patches yet?
(before I try)


John H. Lawitzke           UUCP: Work: ...uunet!frith!dale1!jhl
Dale Computer Corp., R&D         Home  ...uunet!frith!ipecac!jhl
2367 Science Parkway       Internet:   jhl@frith.egr.msu.edu
Okemos, MI, 48864                             [35.8.8.108]