Xref: utzoo comp.sys.amiga:28958 comp.sys.amiga.tech:3644
Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!eecae!cps3xx!usenet
From: usenet@cps3xx.UUCP (Usenet file owner)
Newsgroups: comp.sys.amiga,comp.sys.amiga.tech
Subject: Dos BUGS
Message-ID: <1746@cps3xx.UUCP>
Date: 7 Feb 89 22:06:08 GMT
Reply-To: porkka@frith.egr.msu.edu ()
Distribution: na
Organization: Michigan State University
Lines: 35

I ran into an interesting bug in AmigaDos the otherday when busy testing
a program of mine. It seems that if you do the following:
Make sure that run and type are resident.
cd to ram: and put a sizeable file called "testfile" in ram:
(The resident and RAM: requirements make things go faster. Speed is
of the essence so to speak for this bug to show itself)
then execute the following batch file

run type >con:0/0/100/100/FOO testfile
run type >con:0/0/100/100/FOO testfile
run type >con:0/0/100/100/FOO testfile
run type >con:0/0/100/100/FOO testfile
run type >con:0/0/100/100/FOO testfile
run type >con:0/0/100/100/FOO testfile
run type >con:0/0/100/100/FOO testfile
run type >con:0/0/100/100/FOO testfile
run type >con:0/0/100/100/FOO testfile
run type >con:0/0/100/100/FOO testfile
run type >con:0/0/100/100/FOO testfile

CTRL-\

Executeing that file produces one or both of the following:
One or more "Task Held" requesters
Bogus entries in the process table returned by STATUS

This happens with either CON: or NEWCON:

Is this a problem with DOS needing X amount of time to spawn a process
or is it a problem with CON: needing X amount of time to Create a new
instance? In either case shouldn't a semaphore be in there somewher
to prevent timeing from being a problem?


porkka@frith.egr.msu.edu