Date: Mon, 13 Jan 1997 23:29:20 +0100 From: j@uriah.heep.sax.de (J Wunsch) To: freebsd-current@freefall.freebsd.org Subject: Re: a last word on last Message-ID: <Mutt.19970113232920.j@uriah.heep.sax.de> In-Reply-To: <199701130757.IAA05591@gilberto.physik.rwth-aachen.de>; from Christoph Kukulies on Jan 13, 1997 08:57:38 %2B0100 References: <Mutt.19970112215435.j@uriah.heep.sax.de> <199701130757.IAA05591@gilberto.physik.rwth-aachen.de>
next in thread | previous in thread | raw e-mail | index | archive | help
As Christoph Kukulies wrote: > > You can also use src/tools/3.0-upgrade/cvs-wtmp. > > I did that a couple of times and either cvs-wtmp barfed at > some entries or it didn't work for some other reasons. Have you tried the `force' switch of the most recent version? > I know that the wtmp/utmp issues have been discussed widely but > a clearing word on this issue would really be helpful. What is the idea > to cope with this problem? The idea for the final solution has been floating around here, and was commonly accepted: all the utmp/wtmp handling stuff should be isolated in libutil (it is already mostly), and no application should ever need to include <utmp.h> again (login(3) is still failing this API requirement). There's already one positive side-effect of this discussion: i finally (after more than 10 years of existance?) wrote the man pages for libutil. ;-) Once this has been done, and the major packages and third-party software actually uses it, we can do with utmp whatever we want, we can convert it into dBase format :), make it a record structure, or what else might be the most rationale format. Needless to say, somebody needs to do all this. I have the feeling that David Nugent (hi Dave :) feels compelled to work in this area... -- cheers, J"org joerg_wunsch@uriah.heep.sax.de -- http://www.sax.de/~joerg/ -- NIC: JW11-RIPE Never trust an operating system you don't have sources for. ;-)
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Mutt.19970113232920.j>