Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!uwm.edu!csd4.csd.uwm.edu!wls
From: wls@csd4.csd.uwm.edu (Bill Stapleton)
Newsgroups: comp.bugs.4bsd
Subject: script problem
Summary: Is script broken?  Is there a fix?
Keywords: script, pty, ^S/^Q, RAW/CBREAK modes, TIOCPKT
Message-ID: <1571@uwm.edu>
Date: 19 Dec 89 21:32:32 GMT
Sender: news@uwm.edu
Reply-To: wls@csd4.csd.uwm.edu (Bill Stapleton)
Organization: Computing Services, U of Wisc-Milwaukee
Lines: 25


[I posted this to comp.unix.wizards and got no response, so...]

I've been looking into a problem we have with the script program garbaging 
output to the terminal.  I haven't found a newer version, fix, or mention of
this problem, so here goes.  This is with BSD 4.3-tahoe and script 5.6:

The problem seems to stem from script running in RAW mode, expecting everything
else to take care of the flow-control.  Unfortunately, the response to XOFF is
then too slow, and a slow terminal is over-run, producing various unusable
displays.  I note that other programs, notably rlogin, use the TIOCPKT feature
of the pty to figure out when RAW mode is neccesary, running CBREAK (with its
own local flow-control) otherwise.  Is there a reason that script doesn't do
this?  (appologies if I've condensed this down past making any sense)

Does anybody else consider this a real problem?  Has anyone fixed it?
(This problem may be reproduced using a "slow" terminal which requires flow-
control to "keep up", running script, then running something that produces
lots of output quickly (ls, etc);  Alternately, simply notice how slow the
response to XOFF is while running through script)

--
Bill Stapleton
     wls@csd4.csd.uwm.edu
     uwmcsd4!wls