Subject: RE: Revisiting the tk80 issue
To: Hubert Feyrer <email@example.com>
From: Tim Rightnour <firstname.lastname@example.org>
Date: 07/16/1998 02:04:06
On 16-Jul-98 Hubert Feyrer spoke unto us all:
# On Thu, 16 Jul 1998, Tim Rightnour wrote:
# > Opinions?
# Ewww, yet another variable.
# I'd rather wait for what comes out of xpkgwedge instead of adding more
# hacks here.
a) I think xpkgwedge is a good solution to imake. But it's optional, and
resistence will come if we suggest placing it in NetBSD.cf
b) I *really* don't consider this a hack. The idea is that:
1: You fix tk80 and dependents' brokeness
2: You provide a way for people to use X11, and not install it into the X
tree. Look at something like ghostscript. It has an x11 engine in it for
converting some docs to x11. 99% of this pkg has nothing to do with X. But
you cannot build it without it the way we have it set up, and it's utterly
rediculous to install this thing in /usr/X11R6. A variable like this allows a
pkg developer to DTRT with each pkg. If its a totally X11 pkg, then it gets an
USE_XBASE=yes. If it only has a tiny gui front end (postgresql comes to mind)
then we mark the dependecy on X11, and install it into /usr/pkg anyway.
There are alot of things in this category too.. TeTeX, mm (it has *nothing*
with a GUI, it uses xbdftopcf in the build!!)
If you want *everything* to go into /usr/pkg, then you install xpkgwedge and
the USE_XBASE stuff gets short-circuted. This gives us the most flexibility
of everything. If we mandate xpkgwedge for everyone, then use it to fix tk80,
we may end up pissing off people who actually want to install stuff into the
Tim Rightnour - email@example.com