1 |
On Fri, 24 Jan 2014 14:29:02 -0600 |
2 |
Steev Klimaszewski <steev@g.o> wrote: |
3 |
|
4 |
> On Fri, 2014-01-24 at 20:29 +0100, Tom Wijsman wrote: |
5 |
> > On Fri, 24 Jan 2014 12:10:30 -0600 |
6 |
> > Steev Klimaszewski <steev@g.o> wrote: |
7 |
> > |
8 |
> > > The problem isn't finding someone that has everything - we have |
9 |
> > > people that test on ARMv5, some that test on ARMv6, we have some |
10 |
> > > that test on ARMv7 - until ALL of them are tested, it doesn't get |
11 |
> > > stabled on ARM. So again, it just shuffles around the work, and |
12 |
> > > does nothing to address the actual problem which is manpower with |
13 |
> > > people that have the slower machines to finish their testing. |
14 |
> > > Unless you would like to suggest that we maybe just say fuck |
15 |
> > > anyone using a slow machine? |
16 |
> > |
17 |
> > Consider how packages would rarely get stabilized if we had to wait |
18 |
> > for all arches to test them first before adding any stable keyword |
19 |
> > at all. |
20 |
> > |
21 |
> |
22 |
> Theoretical, again, as always, and not even worth considering because |
23 |
> it doesn't reflect reality. |
24 |
|
25 |
We are talking about reality here, while expanding it on a larger scale; |
26 |
as to make it more clear how that wouldn't work out well in reality. |
27 |
|
28 |
> > Organize first, then get more manpower; otherwise we say the F-word |
29 |
> > to everyone with a faster machine. Joining arm with a slower |
30 |
> > configuration and have everyone waiting on you is a working |
31 |
> > condition to avoid; so, we could have the slower configuration |
32 |
> > stabilize at its own pace. |
33 |
> > |
34 |
> > Would we say F-word to 'em? No, we give them better working |
35 |
> > conditions. |
36 |
> > |
37 |
> |
38 |
> We're all adults here, you can say fuck. |
39 |
|
40 |
What about better working conditions? |
41 |
|
42 |
> For the record, the ARM team does just fine in stabling things in a |
43 |
> reasonable amount of time, so no, we aren't going to change our |
44 |
> working methods. |
45 |
|
46 |
That is a contradiction with what you have said earlier. |
47 |
|
48 |
So, is there an ACTUAL problem with the ARM team or not? |
49 |
|
50 |
In context, you have demonstrated a result of that problem here: |
51 |
|
52 |
@steev | and then you end up with stuff like git is broken on armv7 |
53 |
uclibc |
54 |
@TomWij | steev: Better to know that than to know that it is broken on |
55 |
arm uclibc. |
56 |
@steev | TomWij: it was known that it's broken, it was stabled anyway, |
57 |
which makes it worse |
58 |
@steev | because i commented and said DO NOT STABLE IT |
59 |
|
60 |
> The point of this email thread was we all need to stable faster, |
61 |
|
62 |
That is just a deduction from your point of view; as stated before, |
63 |
that's a possible deduction you can easily make now. But, deducing it |
64 |
every time without guaranteeing that the arch teams improve can lead |
65 |
to it worsening over time; which is something we need to look out for. |
66 |
|
67 |
> and slower arches need to just become unstable only, |
68 |
|
69 |
What about slower configurations keeping up slower arches? |
70 |
|
71 |
> and fuck them. And I'm saying everyone needs to step back because |
72 |
> stabling things faster and faster doesn't allow for proper testing. |
73 |
> |
74 |
> As QA, you should be focusing on making stable, actually stable, not |
75 |
> more bleeding edge. |
76 |
|
77 |
That's the task of the mentors, recruiters and the arch teams; the QA |
78 |
team is there to ensure quality, not performance. If performance limits |
79 |
us from stabilizing everything we want to see stabilized properly, that |
80 |
means that there is simply a lack of interest; then this thread as well |
81 |
as QA comes in to reorganize our efforts as well as policy to be more |
82 |
efficient and effective. |
83 |
|
84 |
The delays that can be witnessed in the actual problem you have brought |
85 |
forward is one example of what can be reorganized to not delay |
86 |
stabilization; another example of what I am working towards is to try |
87 |
to get more people by making our documents for new developers more |
88 |
accessible, some examples: |
89 |
|
90 |
- Revised |
91 |
https://wiki.gentoo.org/index.php?title=Contributing_to_Gentoo&oldid=11534 |
92 |
to |
93 |
https://wiki.gentoo.org/index.php?title=Contributing_to_Gentoo&oldid=83101 |
94 |
|
95 |
- Added reference in the devmanual to the developer handbook: |
96 |
https://github.com/gentoo/devmanual.gentoo.org/commit/ced738e2cf213f7001a4f39cd561983f71631582 |
97 |
|
98 |
- Start a guide that introduces how to write an ebuild from scratch: |
99 |
https://wiki.gentoo.org/wiki/Basic_guide_to_write_Gentoo_Ebuilds |
100 |
|
101 |
Yes, QA is doing something about that. To take this whole thread a step |
102 |
further, it is proposed as an agenda topic for the first QA meeting. |
103 |
|
104 |
https://wiki.gentoo.org/wiki/Project:Quality_Assurance/Meeting_Agenda |
105 |
|
106 |
-- |
107 |
With kind regards, |
108 |
|
109 |
Tom Wijsman (TomWij) |
110 |
Gentoo Developer |
111 |
|
112 |
E-mail address : TomWij@g.o |
113 |
GPG Public Key : 6D34E57D |
114 |
GPG Fingerprint : C165 AF18 AB4C 400B C3D2 ABF0 95B2 1FCD 6D34 E57D |