>Number:         29563
>Category:       pkg
>Synopsis:       sysutils/fdgw over-precise message emitted
>Confidential:   no
>Severity:       non-critical
>Priority:       low
>Responsible:    pkg-manager
>State:          open
>Class:          change-request
>Submitter-Id:   net
>Arrival-Date:   Tue Mar 01 03:40:01 +0000 2005
>Originator:     Robert Elz
>Release:        NetBSD 2.0.1   (pkgsrc -current as of 4/5 hours before e-mail)
>Organization:
	Prince of Songkla University
>Environment:
System: NetBSD jade.coe.psu.ac.th 2.0.1 NetBSD 2.0.1 (JADE-1.7-20050129) #1: Sat Jan 29 15:40:38 ICT 2005 kre@lilac.noi.kre.to:/usr/obj/2.0/kernels/JADE-HP i386
Architecture: i386
Machine: i386
>Description:
	sysutils/fdgw tells me ...

	********************************************************
	Please note: The install stage of fdgw requires a kernel
		     with 'pseudo-device vnd 4' enabled.
	********************************************************

	I find that very hard to believe.   Does it really require
	exactly 4 vnd pseduo-devices?   (I find it hard to believe
	it really requires more than 1, but will it really break if
	there are 5 or 6 vnd's in the kernel?)

	[Aside: vnd should be changed to use the clone interface, if
	it hasn't already, in which case that message will become even
	more incorrect].

>How-To-Repeat:
	Do just about anything with pkgsrc in sysutils/fdgw.
	Wonder what should be done with a system that already has
	more than 4 vnds in its kernel for other reasons...
>Fix:
	Change the message to:

		Please note: The install stage of fdgw requires a kernel
			     with the vnd pseudo-device driver included.

	or perhaps, if this is more correct:

		Please note: The install stage of fdgw requires a kernel
			     with at least 4 vnd pseudo-devices included.

	And if appropriate, which it easily might be (I haven't
	actually ever wanted to install this thing, I just run
	"make checksum" on everything in the universe, and then
	check the messages)   also note something like

		The vnd0 pseudo-device must be available for use when
		this package is installed.

	That's assuming the install phase of this is actually using
	a vnd device, and (I'd guess, perhaps incorrectly) that it
	is going to have the specific /dev name it expects to use
	built into the script, and isn't going searching for a vnd
	device that is currently free...