Gentoo Archives: gentoo-dev

From: Pacho Ramos <pacho@g.o>
To: gentoo-dev@l.g.o
Subject: Re: [gentoo-dev] [RFC] Future of gentoo's stable and unstable trees: what are your thoughts?
Date: Tue, 25 Jul 2017 12:16:00
Message-Id: 1500984949.4544.80.camel@gentoo.org
In Reply to: Re: [gentoo-dev] [RFC] Future of gentoo's stable and unstable trees: what are your thoughts? by "Michał Górny"
1 El mar, 25-07-2017 a las 13:54 +0200, Michał Górny escribió:
2 > Dnia 25 lipca 2017 11:18:21 CEST, Pacho Ramos <pacho@g.o> napisał(a):
3 > > El mar, 25-07-2017 a las 08:18 +0200, Hans de Graaff escribió:
4 > > > On Mon, 2017-07-24 at 23:22 +0000, Peter Stuge wrote:
5 > > > >
6 > > > > I hold a perhaps radical view: I would like to simply remove
7 > >
8 > > stable.
9 > > > >
10 > > > > [snip]
11 > > > >
12 > > > > I consider dev time a precious resource.
13 > > >
14 > > > If we were to drop stable I would have to start maintaining my own
15 > > > stable lists to determine what would be ready to into production for
16 > >
17 > > my
18 > > > company. In production "works most of the time" and "good enough"
19 > > > simply aren't good enough.
20 > > >
21 > > > I estimate that would at least equal the amount of time I'm currently
22 > > > spending on Gentoo work, and consequently my contributions to Gentoo
23 > > > would dwindle to a halt. Most likely I would start looking at other
24 > > > solutions altogether.
25 > > >
26 > > > > More troubleshooting and fixing "hard" problems, less routine work.
27 > > >
28 > > > Except that some of that routine work is actually what I enjoy doing
29 > >
30 > > in
31 > > > Gentoo. I already get plenty of the other two in my day job.
32 > > >
33 > > > Hans
34 > >
35 > > If stable goes away I will simply switch to other distribution and
36 > > retire
37 >
38 > What's the "over my dead commit access" spirit? 
39 >
40
41 Jumping from trying to maintain stable tree to arches dead for ages to drop all
42 stable trees looks to me like a joke promoted by people that has never handled
43 any stabilization request and saw on them how running a pure "testing" system is
44 impossible on many conditions. It seems that some people think that if it fits
45 ok for them, it will fit for all others like we were all using Gentoo for doing
46 the same.
47
48 I could of course deal with things in my personal computer like, for example,
49 needing to run gcc-6 (current testing) and having tons of packages failing to
50 build, or run python-3.6 with only a few subset of packages, or running latest
51 ffmpeg with random packages going to fail with it, or many other issues that
52 anyone doing some stabilization work would have noticed. But, of course, I
53 cannot pretend that all the people using Gentoo systems for working or doing
54 something productive and that for now rely on me for maintaining or helping them
55 with the issues that could arise, will now be also forced to run systems that
56 are likely going to break in different and new ways every time they pretend to
57 update.
58
59 I am also really surprised to see how we can jump from some people that were
60 fighting in the past against we running automatic scripts that already existed
61 to fill stabilization bug reports and CC arches after timeouts, to a new
62 situation of "oh, testing tree is good enough for all the people". We will jump
63 for some people asking for things like doing deeper tests runs for packages
64 going to stable (at a level that was really unfeasible on a large scale) to a
65 situation in that nothing (even no compile test) will be checked at all.
66
67 Additionally, this will also cause new issues between people that were used to
68 run "testing" in the way they are running it now and they pushing to unmask
69 things faster and, others used to "stable", pushing to keep more things hard
70 masked until they are fixed. It's not the first time that we see that, for
71 example, a new glibc version is unmasked when maintainer feels it's ready to
72 allow people to catch the bugs before it going to be stable. If we have no
73 stable tree, that couldn't be done as we couldn't use "testing" for the purpose
74 of "lets unmask X package it give it more visibility and let people catch the
75 bugs". Then, either we keep breaking "testing" even knowing there is no stable,
76 or we will start getting lots of packages in package.mask leading to new issues
77 (like those packages having less visibility and fights between people thinking
78 that a mid breakage is ok and others that not).
79
80 Then, in my case it will be as simply as, if Gentoo becomes testing only, I
81 won't be able to use it for anything productive, only for "playing" with it...
82 and then, I won't see much sense on staying while I will need to use a different
83 distribution de facto for the work and any computer that is not the laptop I use
84 for committing and doing Gentoo dev work.

Replies