1 |
On Wed, 26 Apr 2017 15:35:47 -0400 |
2 |
Mike Gilbert <floppym@g.o> wrote: |
3 |
|
4 |
> On Wed, Apr 26, 2017 at 3:06 PM, Michał Górny <mgorny@g.o> |
5 |
> wrote: |
6 |
> > Even partially failing tests are better than none. |
7 |
> |
8 |
> While I personally agree with this, arch testers tend to give us a |
9 |
> hard time if the tests are failing when they process a keyword or |
10 |
> stable request. |
11 |
|
12 |
Usually if there are tests that fail for reasons other than packaging |
13 |
(like requirements on other distro stuff (I have encountered tests that |
14 |
run 'apt-get' or 'yum') or running daemons), I try to patch them |
15 |
out rather than have failing tests that can't reasonably be fixed. I |
16 |
will also try to fix buggy tests, if it can be done with reasonable |
17 |
effort. |