From: emil AT skatter DOT usask DOT ca (Emil Hallin)
Subject: B20: possible bug in mount on NT 4 (SP4 and SP3)
5 Nov 1998 05:58:03 -0800
Message-ID: <363F550B.6C4D77E2.cygnus.gnu-win32@skatter.usask.ca>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
To: gnu-win32 AT cygnus DOT com

There are some odd behaviors associated with the B20 mount utility.
These may be bugs:

1) mount points survive a control panel uninstall/machine
reboot/reinstall cycle. This is a little disconcerting, as I think the
uninstall should remove all the pieces if possible.

2) building a new version of EGCS (say egcs1.1b as the latest current
release) works perfectly provided that:
	"/" is on the same partition as %SYSTEMROOT%
	the source code is mounted under the same mount point

On my machine, I can make "configure" and "make bootstrap"  hang with
100% reproducibility by either uninstalling B20 and reinstalling it on a
different partition (that does not contain %SYSTEMROOT%) or by leaving
B20 on the %SYSTEMROOT% partition and moving the source code so it is
found beneath another mount point. I *think*, but haven't had time to
reproduce this, it is sufficient to leave the source code for EGCS where
it is, and just add a second mount point pointing to it. This also makes
"configure" and "make bootstrap" hang at arbitrary points. 

It does not appear to matter *where* %SYSTEMROOT% is; all that matters
is that "/" and %SYSTEMROOT% be on the same partition. Of course, I
don't have the time to check on the reproducibility of this too
thoroughly...I only did a complete reinstall of NT 4 once!

3) mount --reset does NOT work as advertized. On my machine, it simply
removes a mount point one at a time until they are all gone.

	emil
-
For help on using this list (especially unsubscribing), send a message to
"gnu-win32-request AT cygnus DOT com" with one line of text: "help".