Gentoo Archives: gentoo-dev

From: "Michał Górny" <mgorny@g.o>
To: gentoo-dev@l.g.o
Cc: tommy@g.o
Subject: Re: [gentoo-dev] Re: [RFC] multilib-build.eclass and restricting unsupported ABIs
Date: Mon, 04 Mar 2013 22:48:49
Message-Id: 20130304234909.542a88b6@pomiocik.lan
In Reply to: Re: [gentoo-dev] Re: [RFC] multilib-build.eclass and restricting unsupported ABIs by Thomas Sachau
1 On Mon, 04 Mar 2013 23:21:36 +0100
2 Thomas Sachau <tommy@g.o> wrote:
3
4 > Michał Górny schrieb:
5 > > On Mon, 4 Mar 2013 11:02:40 +0100
6 > > Alexis Ballier <aballier@g.o> wrote:
7 > >
8 > >> you are called with ABI=sth argv[0] = your name
9 > >
10 > > I'm afraid that's the first potential point of failure. Relying
11 > > on argv[0] is a poor idea and means that any application calling exec()
12 > > with changed argv[0] on a wrapped binary will fail terribly.
13 >
14 > Nobody said, that one cannot create situations, where such a wrapper
15 > does fail, the point is more an easy and general solution for wrapping
16 > binaries without implementing different solutions for the same issue in
17 > every ebuild.
18
19 There's no such thing as 'easy and general solution'. You always
20 sacrifice something.
21
22 And in this case, you're creating a point of failure which is
23 completely custom to Gentoo and actually quite hard to track. Just to
24 support a specific package manager feature specific to Gentoo.
25
26 > If you have a better, yet still easy and general solution not requiring
27 > every ebuild to create something on its own, please write it instead of
28 > just complaining how bad the wrapper solution actually is.
29
30 The solution is called eclasses.
31
32 > > Yep, I intended to just have the additional variant of glxinfo directly
33 > > callable, with a name chosen specifically by the X11 team. Wrapper
34 > > would be more confusing than beneficial here IMO.
35 >
36 > Ah, you actually want each ebuild to have its own custom hack instead of
37 > one global solution usable everywhere?
38
39 Yes.
40
41 > >> To some extent that's what happened to python too :) As a python
42 > >> maintainer, you could share your thoughts on the topic. python slotting
43 > >> was intended to make switching between python versions easy but has
44 > >> been needing wrappers for the python binary.
45 > >
46 > > I'm doing just that. Any kind of wrapping is an increasing mess. I'm
47 > > still trying to find out good solutions for Python wrapping but there's
48 > > no such thing. It's always about choosing one evil over the other.
49 >
50 > So you are wrapping python, have not yet found anything better and still
51 > dont want to wrap abi-specific binaries, while you dont have a better
52 > solution at hand? Saying no to everything is easy, providing something
53 > better if you dont like a suggestion is the challenge.
54
55 Yes, it is easy and mess-free. Using a cheap hack is mess-full, and is
56 just asking for trouble which will eventually rise.
57
58 --
59 Best regards,
60 Michał Górny

Attachments

File name MIME type
signature.asc application/pgp-signature