1 |
On 17/11/16 22:56, Rich Freeman wrote: |
2 |
> On Thu, Nov 17, 2016 at 4:37 AM, Michael Palimaka <kensington@g.o> wrote: |
3 |
>> On 17/11/16 20:16, Rich Freeman wrote: |
4 |
>>> On Thu, Nov 17, 2016 at 2:16 AM, Michael Palimaka <kensington@g.o> wrote: |
5 |
>>>> * A leaf package such as {{package|kde-apps/kcalc}} may not require any |
6 |
>>>> runtime testing at all |
7 |
>>> |
8 |
>>> I'm not really a big fan of this, but if we REALLY didn't want to do |
9 |
>>> any runtime testing on a package then we should have some way to tag |
10 |
>>> the package as such, and then have some way to do automated |
11 |
>>> build-test-only stabilization. If you aren't doing runtime testing |
12 |
>>> then manual stabilization adds zero value. |
13 |
>> |
14 |
>> How much value do you think we gain from runtime testing a package like |
15 |
>> kcalc as part of the stabilisation process, considering that it already |
16 |
>> sat in ~arch for at least 30 days? |
17 |
> |
18 |
> We ensure that it actually runs at all with non-~arch dependencies? |
19 |
> |
20 |
> The 30 days spent in ~arch tells you very little about whether the |
21 |
> package works with stable dependencies, since only those running mixed |
22 |
> keywords would be testing that. |
23 |
|
24 |
What is the *real* risk that kde-apps/kcalc builds against stable |
25 |
dev-libs/gmp but then starts producing funny numbers at runtime? |
26 |
|
27 |
Let's put it another way - assume we're stabilising a new version of |
28 |
dev-libs/gmp instead. Should we test already-stable kde-apps/kcalc |
29 |
first? What about the other hundred reverse dependencies? |
30 |
|
31 |
Just to be clear, I'm not advocating banning runtime testing. I just |
32 |
think that, considering the state of the stable tree, we should consider |
33 |
very careful in which situations we actually gain value from it. That's |
34 |
for another thread, however. I deliberately worded the document so that |
35 |
the final decision on the exact level of testing required (runtime or |
36 |
otherwise) is between the person filing the stabilisation request and |
37 |
the person actioning it. |
38 |
|
39 |
>> Also, based on conversations with various arch team members, my |
40 |
>> understanding is that a lot of stabilisations that happen right now are |
41 |
>> already build-only. |
42 |
>> |
43 |
> |
44 |
> Certainly this isn't the documented process and it is the first I've |
45 |
> heard of this. |
46 |
|
47 |
Indeed it is not documented, but then again not a lot about |
48 |
stabilisation is well documented. |
49 |
|
50 |
Consider bug #584468, a typical bulk stabilisation request from the |
51 |
GNOME team consisting of 188 packages of varying types. I very much |
52 |
doubt each arch team member sat down and tested file-roller to make sure |
53 |
it extracts archives with stable libarchive and looks OK with stable |
54 |
GTK, then makes sure gnucash-docs still looks pretty with stable |
55 |
docbook-xsl-stylesheets, and so on 186 more times. |
56 |
|
57 |
> However, I'd be interested in metrics on failures discovered in |
58 |
> runtime testing and so on, or missed with a lack of runtime testing. |
59 |
> I'll admit that a lot of runtime testing tends to be fairly shallow, |
60 |
> but I do think there is something to be said for doing some kind of |
61 |
> runtime testing. |
62 |
|
63 |
I'd be interested in metrics like this too, but I'm not sure how we'd |
64 |
collect such information. |
65 |
|
66 |
Runtime testing is important, and in an ideal world every package would |
67 |
receive detailed runtime testing before being moved to stable. Where we |
68 |
are now, I think forcing runtime testing in every case costs more than |
69 |
we gain. |
70 |
|
71 |
Building and running for as long as possible the next stable glibc has a |
72 |
comparatively low cost compared to the pain if it broke. Manually |
73 |
running 300 kde-apps/* packages every three months is a guaranteed waste |
74 |
of time. Most packages probably sit somewhere in between. |
75 |
|
76 |
> I think we need to think about why we actually have a stable branch. |
77 |
> Does it offer any value if all we do is build testing, when I'm sure |
78 |
> the maintainers at least build their packages in ~arch before they |
79 |
> commit them? |
80 |
|
81 |
The whole point of the 30 day waiting period is that issues can and do |
82 |
appear that the maintainer did not encounter when they bumped. |