1 |
On 2 October 2013 08:51, Peter Stuge <peter@×××××.se> wrote: |
2 |
|
3 |
> I agree, but I think the problem is basically that many people |
4 |
> consider it impossible for "newer" to ever be shitty. |
5 |
> |
6 |
> Even if they are intimately familiar with the details of a package |
7 |
> upstream they may still not be capable of determining what is shitty |
8 |
> in the context of a distribution. |
9 |
> |
10 |
> I guess most stabilization requesters as well as actual stabilizers |
11 |
> are even less familiar with upstream details, so determining what is |
12 |
> shitty and not is really hard.. :\ |
13 |
> |
14 |
|
15 |
|
16 |
The other really annoying thing you have to consider, is that most people |
17 |
out there are using all (~ARCH) or all (ARCH) keywording, not a mix of the |
18 |
two. |
19 |
|
20 |
^ This has the fun side effect of meaning packages that are (~ARCH) and |
21 |
have (ARCH) dependents, where the package that is currently (~ARCH) is |
22 |
pending stabilization, has likely had nobody test it at all except for |
23 |
arch testers. |
24 |
|
25 |
So if you're relying on the presence of filed bugs to give some sort of |
26 |
coverage metric, you're going to be out of luck from time to time. For |
27 |
instance, that fun bug where stabilising a version of libraw broke the |
28 |
things depending upon it that were already stable. |
29 |
|
30 |
Its ironic really, thats essentially a hidden bug that exists as a result |
31 |
of having two tiers of testing. |
32 |
|
33 |
https://bugs.gentoo.org/show_bug.cgi?id=458516 |
34 |
|
35 |
I've been long wishing there was a FEATURE I could turn on that would just |
36 |
report installs to a database somewhere, showing |
37 |
success/fail/succcess+tests/fail+tests , with dependencies, useflags, and |
38 |
other relevant context, so you'd at least have a dataset of *success* |
39 |
rates, and you could do more heavy testing on things where there were fail |
40 |
results, or an absense of results. |
41 |
|
42 |
CPAN has a comparable service that leverages end users reporting test runs |
43 |
on code while they install it, and some end users, given this power, go so |
44 |
out and set up mass automated testing boxes, the utility of which I find |
45 |
useful on a daily basis, because a user is far more likely to get an |
46 |
automated success/fail report sent to a server, and far *less* likely to |
47 |
want to set time aside to go through the rigmarole of filing a bug report. |
48 |
|
49 |
Some users are also inclined to just try a few versions either side, and |
50 |
never file a bug report, or try twiddling USE flags or disabling |
51 |
problematic FEATURES to find a version that works for them, and you may |
52 |
never see a bug report for that. |
53 |
|
54 |
An automated "X combination failed" report at very least lets you know a |
55 |
datapoint where a failure occurred. |
56 |
|
57 |
I'm not saying we should make any automated decision making *based* on |
58 |
those reports, but it would be far more useful to have a list of known |
59 |
failure cases up front than to ask a tester to try be lucky and find it by |
60 |
looking for it. |
61 |
|
62 |
( After all, bugs often arise when you're not looking for them, as opposed |
63 |
to when you are, and some bugs, when looked for, vanish ) |
64 |
|
65 |
-- |
66 |
Kent |