Path: utzoo!utgpu!water!watmath!clyde!att!osu-cis!killer!vector!rpp386!jfh
From: jfh@rpp386.UUCP (John F. Haugh II)
Newsgroups: comp.sources.d
Subject: Re: newgetty question
Message-ID: <2525@rpp386.UUCP>
Date: 8 Jun 88 02:27:42 GMT
References: <1259@percival.UUCP>
Reply-To: jfh@rpp386.UUCP (The Beach Bum)
Organization: Big "D" Home for Wayward Hackers
Lines: 53

In article <1259@percival.UUCP> nerd@percival.UUCP (Michael Galassi) writes:
>Howdy,
>	I just finished trying to compile the newgetty program recently
>posted.  I got _rdchk, _defread and _defopen as undefined by ld.  I
>solved the _rdchk as follows but have no clue as to what the other two
>routines are supposed to do.  Anyone have a clue for me to go by?  As
>per usual,
>thanks in advance,
>		-michael

defopen() and defread() and routines for using defaults files.  the
idea being that OEM's would want to be able to configure their systems
without having to provide sources.  using a ``defaults'' file allows
different configurations using the same binary.

an example of a defaults file for dump(1) might be

tape=/dev/rfd0
disk=/dev/rroot

for a small pc, and then

tape=/dev/rmt/0m
disk=/dev/dsk/0s1

for a larger machine using the system v naming convention.  the
normal convention is for a program named 'xxx' to store its defaults
in the file '/etc/default/xxx'.

the call to defopen() for that program would be

int	defopen (char *);

	defopen ("xxx");

and the call to defread() to determine the tape device would be
something like

char	*defread (char *);
char	*mytape, *cp;

	if ((cp = defread ("tape")) == (char *) 0)
		cp = "/dev/rmt/0m";

	strdup (mytape, cp);

i seem to recall there is a bug (or was in the Microsoft Xenix
version i used to use) with these routines.  you may want to write
your own.  i did once and it took mere minutes.  don't forget the
strdup() call, the return value from defread() "points to static
data, whose contents are overwritten by each call."

- john.