Message-Id: <m0tz3RR-000GPWC@gwar.ardi.com>
Date: Tue, 19 Mar 96 08:35 MST
From: mat AT ardi DOT com (Mat Hostetter)
To: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
Cc: djgpp-workers AT delorie DOT com
Subject: Re: Re:`stat' fails to detect nonexistent fi
In-Reply-To: <Pine.SUN.3.91.960319172910.20848N-100000@is>
References: <m0tz31v-000GPWC AT gwar DOT ardi DOT com>
	<Pine DOT SUN DOT 3 DOT 91 DOT 960319172910 DOT 20848N-100000 AT is>

>>>>> "Eli" == Eli Zaretskii <eliz AT is DOT elta DOT co DOT il> writes:

    >> For what it's worth, WinNT was recently biting us because it
    >> didn't support the standard INT 2F BIOS call to read CD-ROMs.
    >> It just silently returned after doing nothing...with the carry
    >> clear, which would normally indicate success.  Woohoo.

    Eli> Is this in `getmntent' or in the code you've written?

In code we wrote.  I hadn't even thought about getmntent.  Is it going
to do something bad also?

I worked around it by putting magic cookies in our transfer buffer
that should get overwritten by a successful call and seeing if they
survived the call.  If they survived the call, I wrote a different set
of magic cookies in the transfer buffer and retried the read.  If both
sets of cookies survived, it must be the case that the read actually
failed (as opposed to the CD sector happening to contain some of the
cookies).

-Mat