1 |
On Wed, Jul 8, 2015 at 11:07 AM, Tobias Klausmann <klausman@g.o> |
2 |
wrote: |
3 |
|
4 |
> Hi! |
5 |
> |
6 |
> This got a bit rambly, sorry 'bout that. |
7 |
> |
8 |
> tl;dr: Continuous Integration is a neat idea if you have the |
9 |
> Hardware. The off-mainstream arch teams don't. |
10 |
> |
11 |
|
12 |
Clearly because we cannot be perfect, we should not even try! |
13 |
Realistically aside from the major arches none of the other arches have |
14 |
large userbases. I think quality for the sake of quality is pretty silly; |
15 |
but if we can improve the quality of arches where hardware is readily |
16 |
available we should do so (regardless of minority architecture objections.) |
17 |
|
18 |
-A |
19 |
|
20 |
|
21 |
> |
22 |
> |
23 |
> On Tue, 07 Jul 2015, Ciaran McCreesh wrote: |
24 |
> > On Tue, 07 Jul 2015 08:04:47 +0800 |
25 |
> > Patrick Lauer <patrick@g.o> wrote: |
26 |
> > > I'll laugh about it next time I update OpenFOAM: |
27 |
> > > |
28 |
> > > Fri Jun 27 12:52:19 2014 >>> sci-libs/openfoam-2.3.0 |
29 |
> > > merge time: 1 hour, 5 minutes and 8 seconds. |
30 |
> > > or |
31 |
> > > |
32 |
> > > Sun Jun 29 20:36:09 2014 >>> sci-mathematics/nusmv-2.5.4 |
33 |
> > > merge time: 2 hours, 58 minutes. |
34 |
> > |
35 |
> > What's the problem? Better the build bot wastes that time once than a |
36 |
> > whole bunch of users do. |
37 |
> |
38 |
> Just be aware that that Just Won't Happen for the off-mainstream |
39 |
> archs. We just don't have the hardware to keep up with something |
40 |
> like that. Unless you want to employ several people reordering |
41 |
> the CI checks so a test for glibc doesn't make mutt, util-linux, |
42 |
> iptables and 55 other packages wait for hours. |
43 |
> |
44 |
> 'Cause invariably, if the CI system is delayed beyond the |
45 |
> expectation for the package itself, it will be ignored. |
46 |
> |
47 |
> It would also mean that the other purposes the devboxes are used |
48 |
> for will be delayed non-trivially. In the case of Alpha, that |
49 |
> would for example be upstream gcc testing, and stage3 building. I |
50 |
> am sure arches like PPC64, IA64 and so on have the same problem. |
51 |
> |
52 |
> > > That's without dependencies, so from a clean minimal starting point |
53 |
> > > (once you figure out what useflags are needed) you're looking at 12+ |
54 |
> > > hours of walltime to get that checked. (On a reasonably modern Xeon |
55 |
> > > with SSDs ...) |
56 |
> > |
57 |
> > Uh, no, because you use binary packages for the dependencies, and you |
58 |
> > use a package mangler that can figure out the use flags for you. |
59 |
> |
60 |
> Even with binary package usage (which is not as efficient as it |
61 |
> could be due to USE flags), testing Qt4 (as in compile it and run |
62 |
> the test suite, which is the bare minimum) took me close to 10 |
63 |
> hours the other day. Not including fixing fetch problems halfway |
64 |
> through. During this time, very little else got done on the one |
65 |
> devbox we have. |
66 |
> |
67 |
> > > So thanks for your intentional comedy, but let's be serious here. |
68 |
> > |
69 |
> > Exherbo's been doing all this for years, and it works rather well. The |
70 |
> > only comedy is that Gentoo doesn't even realise this is trivial to do |
71 |
> > nowadays. |
72 |
> |
73 |
> Exherbo is not supporting as many off-mainstream architectures as |
74 |
> we do, so the comparison is flawed at best. Yes, sure, let's run |
75 |
> CI for amd64 and maybe x86 and whoever else has the hardware. But |
76 |
> it will not speed up stabilization bugs regarding the long tail. |
77 |
> It might even lead to people completely ignoring the slow |
78 |
> architectures and the rift between amd64 and the rest becoming |
79 |
> yet wider. |
80 |
> |
81 |
> Don't get me wrong, abandoning Alpha and other architectures is |
82 |
> an option, but remember that for some of these, we are the last |
83 |
> distro that still provides install media and a decent (as in: |
84 |
> current and complete-ish) set of packages. The very fact that we |
85 |
> are able to let amd64 and the rest drift a bit is the reason why |
86 |
> we can still pull it off. |
87 |
> |
88 |
> I have considered dropping Alpha entirely (for myself, that is; |
89 |
> I'd resign as the arch team lead and stop working on it; the |
90 |
> devbox, which is soft-donated through me, would stay available). |
91 |
> But I would rather not, for various reasons, none of which are |
92 |
> technical. Then again, I'm not a Gentoo dev because it makes |
93 |
> sense rationally (it does, but it's not my reason). |
94 |
> |
95 |
> In essence, assuming we can "just scale" to make CI work is |
96 |
> ignoring the matter of the slower archs. And I suspect the "it |
97 |
> works on amd64, fuck everyone else" is not something we want to |
98 |
> pick as a motto. |
99 |
> |
100 |
> Regards, |
101 |
> Tobias` |
102 |
> |
103 |
> |
104 |
> |
105 |
> -- |
106 |
> Sent from aboard the Culture ship |
107 |
> GSV (System Class) Empiricist |
108 |
> |
109 |
> |