Path: utzoo!mnetor!uunet!lll-winken!lll-lcc!lll-tis!ptsfa!ihnp4!ihlpg!tainter
From: tainter@ihlpg.ATT.COM (Tainter)
Newsgroups: comp.sys.atari.st
Subject: Re: Changing  to / in TOS
Message-ID: <4881@ihlpg.ATT.COM>
Date: 20 Feb 88 03:48:38 GMT
References: <1988Jan29.130047.18554@jarvis.csri.toronto.edu> <969@atari.UUCP> <432@nikhefh.hep.nl>
Organization: AT&T Bell Laboratories - Naperville, Illinois
Lines: 19

In article <432@nikhefh.hep.nl>, gert@nikhefh.hep.nl (Gert Poletiek) writes:
> In article <4804@ihlpg.ATT.COM> tainter@ihlpg.ATT.COM (Tainter) writes:
> >You know.  It strikes me that this should have been very simple to fix.
> >You just have the HD device driver do mediach() events.
> >--j.a.tainter

> Fffhew, great idea. Ever thought about disk caching? 
> I happen to have a 1.2 MegaByte disk cache installed, always. I would hate to
> see your 'solution' causing my cache to be invalidated. It is already a pain
> in the neck that GemDos returns media changed on every other access to a
> write protected floppy...  

I am not saying have every event cause a mediach() nor do I advocate
an at random() times cause a mediach() event.  Instead have the driver check
the system space available and when it gets low cause a mediach().  This
should then be an infrequent event and will in fact only occur when their
has been progression through the disk.  This type of situation would be
defeating your caches anyway.

> Gert Poletiek