Gentoo Archives: gentoo-user

From: Michal 'vorner' Vaner <vorner@×××.cz>
To: gentoo-user@l.g.o
Subject: Re: [gentoo-user] package.features
Date: Sat, 03 Mar 2007 20:47:41
Message-Id: 20070303204146.GC1729@tarantula.kolej.mff.cuni.cz
In Reply to: Re: [gentoo-user] package.features by "Bo Ørsted Andresen"
1 Hello,
2
3 On Sat, Mar 03, 2007 at 09:07:23PM +0100, Bo Ørsted Andresen wrote:
4 > On Saturday 03 March 2007 20:39:01 Michal 'vorner' Vaner wrote:
5 > > But there may be situations where this possibility could be used in some
6 > > other ways. You may want nostrip for few packages only (need of debug
7 > > symbols, and you do not want to have them for the whole system), or the
8 > > make check fails for one package you critically need just now and have
9 > > no other choice than to use it anyway.
10 > >
11 > > In my opinion, it should be possible to configure all the variables
12 > > according to the package name.
13 >
14 > The cases you've mentioned here (nostrip, splitdebug, test) are handled on the
15 > bash side of portage which means disabling them in /etc/portage/bashrc on a
16 > per package basis (or in /etc/portage/env/*) is possible.
17
18 Ups, seems like I overlooked this nice file at all. It seems to be a
19 really powerful tool :-). Sorry for not reading properly enough.
20
21 > FEATURES=userpriv, however, is handled on the python side and hence that won't
22 > work. In order to persuade someone to change this you'll probably need a
23 > better use case than -userpriv which ultimately should be either fixed or
24 > restricted in the ebuild.
25
26 You are probably right. And no one will want to add some odd file that
27 could configure any arbitrary variable according to a package just
28 because it would look cool and would be consistent :D. (Or maybe it can
29 be partly created by the bashrc thingie)
30
31 Anyway, thank you very much. At last, I know a bit more now, and won't
32 waste my time hacking something portage probably already has.
33
34 --
35 Work with computer has 2 phases. First, computer waits for the user to tell it what
36 to do, then the user waits for the computer to do it. Therefore, computer work
37 consists mostly of waiting.
38
39 Michal "vorner" Vaner