Strange Dependancies

1998-04-05 Thread Jason Gunthorpe
FYI, I have just added support in APT for constructs like: Package: sgml-tools Version: 1.0.5-1 Priority: optional Section: text Maintainer: Sven Rudolph <[EMAIL PROTECTED]> Depends: libc6, sp (>= 1.1.1-3), sgml-data (>= 0.02), sgml-base, gawk, perl (>= 5.004.04-4) Suggests: groff, late

Re: IMPORTANT: calling ldconfig in maintainer scripts

1998-04-05 Thread Guy Maor
Christian Schwarz <[EMAIL PROTECTED]> writes: > Any package installing shared libraries in a directory that's listed > in /etc/ld.so.conf has to call "ldconfig" in its postinst script, unless > this script is called with a "failed-*" or "abort-*" argument (in which > case ldconfig may _not

Re: libtool varying versions

1998-04-05 Thread jdassen
On Sun, Apr 05, 1998 at 04:35:09PM +0200, Christian Schwarz wrote: > On Tue, 17 Mar 1998 [EMAIL PROTECTED] wrote: > > IMO the new libtool's behaviour (forced -rpath and no inter-library > > dependencies) is seriously broken. The Debian maintainer should lobby > > with the upstream author to get if

Re: libtool varying versions

1998-04-05 Thread Christian Schwarz
On Tue, 17 Mar 1998 [EMAIL PROTECTED] wrote: > On Tue, Mar 17, 1998 at 08:30:08AM -0600, Marcelo E. Magallon wrote: > > On Tue, 17 Mar 1998 [EMAIL PROTECTED] wrote: > > > One of the irritating other things is that newer versions of libtool > > > force -rpath. > > > > Yes. And lintian generates a

Re: IMPORTANT: calling ldconfig in maintainer scripts

1998-04-05 Thread Rob Browning
Christian Schwarz <[EMAIL PROTECTED]> writes: > The result of this discussion surprises me. I remember a few other > discussions on this topic here and the result has always been: `the > packaging manual is completely correct--if everyone does it exactly as > it's documented, we won't have any pro

IMPORTANT: calling ldconfig in maintainer scripts

1998-04-05 Thread Christian Schwarz
(I'm sorry for the late reply. I was very busy the last days. I'll catch up with the other mails here on debian-policy soon.) The result of this discussion surprises me. I remember a few other discussions on this topic here and the result has always been: `the packaging manual is completely corre

Re: Is it a bug to call /usr/bin/X11/foo in the postinst?

1998-04-05 Thread Brian Mays
Upon further reflection and experimentation, I have decided to end this discussion by implementing rxvt's alternative method through /usr/X11R6/bin. The problem that I discussed in my previous post can be avoided by an aptly timed removal of the X11R6 alternative. Brian -- To UNSUBSCRIBE, email