Re: dpkg semi-hijack - an announcement (also, triggers)
I'm going to have to agree with this; I admin m68k buildds; I had to
preinstall most of the texex packages simply because it was taking hours
(and that is not an exaggeration) to install them all, and then remove
them all again. A trigger based system would help relieve this problem. I
had a similar slowdown installing texex packages on a slower powerpc
That being said, I neither agree nor disagree with the hijacking of dpkg,
but triggers are important, especially for slower architectures.
On Mon, 10 Mar 2008, Adam Borowski wrote:
Date: Mon, 10 Mar 2008 00:07:15 +0100
From: Adam Borowski <email@example.com>
Subject: Re: dpkg semi-hijack - an announcement (also, triggers)
Resent-Date: Sun, 9 Mar 2008 23:11:16 +0000 (UTC)
On Sun, Mar 09, 2008 at 02:17:15PM -0800, Mike Bird wrote:
On Sun March 9 2008 14:46:50 Cyril Brulebois wrote:
Is dpkg handling the boot sequence? Or are you confusing that with
I hope I'm not mis-stating Frans's position when I say that Frans
believes dpkg triggers are the best way to install dependency-based
initscripts. Otherwise the installer unnecessarily and repetitively
globally recalculates initscript dependencies for each package installed.
I expect dpkg triggers would also be valuable for things like mktexlsr
runs when working with texlive.
AOL. Try installing anything tex-related on a slow or mid-speed machine.
So, what about leaving the bikeshed painted in Ian's color, and starting
from that point? There are two strong technical arguments: 1. triggers
being a vital piece, and 2. diverging from Ubuntu is badly counter-
And then you can duke it out about NULL vs (char*)0 to your heart's content.
1KB // Microsoft corollary to Hanlon's razor:
// Never attribute to stupidity what can be
// adequately explained by malice.
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact firstname.lastname@example.org