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