Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!rutgers!mit-eddie!Think!ames!cit-vax!tybalt.caltech.edu!walton
From: walton@tybalt.caltech.edu.UUCP
Newsgroups: comp.sys.amiga
Subject: Re: Shellv2.05 trashing disk?
Message-ID: <1640@cit-vax.Caltech.Edu>
Date: Thu, 29-Jan-87 13:59:39 EST
Article-I.D.: cit-vax.1640
Posted: Thu Jan 29 13:59:39 1987
Date-Received: Sat, 31-Jan-87 03:39:23 EST
References: <595@rayssd.RAY.COM>
Sender: news@cit-vax.Caltech.Edu
Reply-To: walton@tybalt.caltech.edu.UUCP (Steve Walton)
Distribution: na
Organization: Calfornia Institute of Technology
Lines: 19
Keywords: Shellv2.05
Summary: timeout value bug in WaitForInput()

In article <595@rayssd.RAY.COM> reh@rayssd.RAY.COM (Robert E. Hathaway) writes:
>
>	I recently installed Steve Drew's ShellV2.05 on my system.
>After using it for most of the night I rebooted my system one last time
>and got a "not a DOS disk message".  This was like deja vu having had a
>disk trashed with the shell V2.03.  After running diskdoctor I was told
>that track 40, sector 0 was bad.  This software is great but what a price.
>I am backing out to V2.04 and hoping the problem doesn't exist there and
>just hasn't manifested itself yet.

   Sounds like an old bug got mistakenly re-introduced.  To do the command
line editing, Drew's shell sets the console to RAW mode when taking input,
then back to CON mode while a command is running so as to get typeahead
capability.  But when he used WaitForInput(file-handle, 1L) to get the
typeahead after completion of the command, he got disk trashing.  Setting
the timeout on WaitForInput to any value other than 1L (2L or more) fixed
the problem.  A quick look at the source for Shell 2.05 shows at least one
WaitForInput() call with 1L as the timeout, in main().  Change it to 2L
and see if it fixes the problem.