1 |
On Wed, Jul 8, 2015 at 10:11 PM, Steev Klimaszewski <steev@g.o> |
2 |
wrote: |
3 |
|
4 |
> On Wed, 2015-07-08 at 21:11 +0200, Tobias Klausmann wrote: |
5 |
> > Hi! |
6 |
> > |
7 |
> > On Wed, 08 Jul 2015, Ciaran McCreesh wrote: |
8 |
> > > On Wed, 8 Jul 2015 20:07:34 +0200 |
9 |
> > > Tobias Klausmann <klausman@g.o> wrote: |
10 |
> > > > In essence, assuming we can "just scale" to make CI work is |
11 |
> > > > ignoring the matter of the slower archs. And I suspect the "it |
12 |
> > > > works on amd64, fuck everyone else" is not something we want to |
13 |
> > > > pick as a motto. |
14 |
> > > |
15 |
> > > "It works on amd64 on a clean build system" is a heck of a lot better |
16 |
> > > than "it may or may not work on one developer's heavily modified |
17 |
> > > box"... The point of testing is not to catch all problems. It's just |
18 |
> > > there to catch some simple screwups, cheaply. |
19 |
> > |
20 |
> > Agreed. Still, in my experience, problems that are not seen on |
21 |
> > amd64 are the vast majority of timesinks. Usually, by the time |
22 |
> > non-amd64 shows up on a non-security bug, the Basic Bullshitâ„¢ has |
23 |
> > been sorted. And even if it isn't: Arches are usually quick to |
24 |
> > point it out and the rest of arches move on to the next bug. The |
25 |
> > truly arch-dependent bugs are what wastes my time: |
26 |
> > |
27 |
> > For example: |
28 |
> > |
29 |
> > - dependencies not being keyworded for arch or ~arch but only |
30 |
> > amd64/~amd64 |
31 |
> > - dependencies not even building on ~arch, but on ~amd64 |
32 |
> > - package assuming availability of gcc/ghc/Python-X.Y when |
33 |
> > arch/~arch only has something for <Y or <X |
34 |
> > - my favourite: assuming udev is at version X for arch/~arch when |
35 |
> > it has been broken there for roughly twelve kernel versions due |
36 |
> > to udev/systemd upstream not caring. The information is one |
37 |
> > equery-y command line away, but no, let's file that bug. |
38 |
> > |
39 |
> |
40 |
> That's a failing of the arch team or the committer then - or whomever |
41 |
> keyworded the package without testing the dependencies. That's why the |
42 |
> keywording bugs happen when dependencies change - and yes, occasionally |
43 |
> a dependency loses it's keywords and that may be where the issue comes |
44 |
> from, I'm not sure. This used to be watched very closely, but I believe |
45 |
> the tool being used broke at some point. |
46 |
> |
47 |
> What exactly do you mean, it's just one command line away but let's file |
48 |
> the bug - yes a bug SHOULD be filed, so that the people know of the |
49 |
> issue so it doesn't get repeated, over and over. |
50 |
> |
51 |
> |
52 |
> > Having every arch team chase the deps for every package is very, |
53 |
> > very frustrating, since it is so trivial a problem. We need a |
54 |
> > tool that answers the question: "I want to mov cat/pkg-1.2.3 to |
55 |
> > stable for arches a, b and c, what do I need?" for _every |
56 |
> > package_. Some groups seem to have tooling for this, but it is |
57 |
> > far from easily (obviously?) available, so every team gets to |
58 |
> > re-discover dependency hell. Ruby is especially bad since |
59 |
> > FEATURES=test make the depgraph explode (and sometimes circular). |
60 |
> > |
61 |
> > Regards, |
62 |
> > Tobias |
63 |
> > |
64 |
> > |
65 |
> |
66 |
> The only fear I have about CI, is that we turn into every other distro |
67 |
> out there where "it builds, ship it!" - I know I prefer to have our arm |
68 |
> team actually test the packages more than just doing FEATURES=test |
69 |
> (assuming the tests aren't broken on arm) emerge, although I know there |
70 |
> are some people on the team who feel that it's too much to actually test |
71 |
> all of the arm boards. For a long time, certain binary distros were |
72 |
> compiling anything and everything, and if it built for other arches it |
73 |
> was available. Even if there was no way it would possibly work on that |
74 |
> arch. |
75 |
> |
76 |
> |
77 |
So don't keyword or stabilize packages you don't test thoroughly. I see how |
78 |
CI *could* be used to do bad things; but I don't see anyone proposing it be |
79 |
used in that way; nor do you have to accept such suggestions (its your |
80 |
arch, after all.) |
81 |
|
82 |
-A |
83 |
|
84 |
|
85 |
> Regards, |
86 |
> Steev |
87 |
> |
88 |
> |
89 |
> |