1 |
On Sat, 7 Jul 2012 17:46:49 -0400 |
2 |
Alexandre Rostovtsev <tetromino@g.o> wrote: |
3 |
|
4 |
> On Sat, Jul 7, 2012 at 11:21 AM, Michał Górny <mgorny@g.o> |
5 |
> wrote: |
6 |
> > To be honest, I think the first thing to do would be fixing the test |
7 |
> > suites to skip tests which fail due to internet connection being |
8 |
> > unavailable. Well, there would still be question how to reliably |
9 |
> > determine that... |
10 |
> |
11 |
> For some packages, e.g. geocode-glib, which is basically a library for |
12 |
> calling a particular web service from C code, running the test suite |
13 |
> without network access is almost pointless. (Unless, of course, you |
14 |
> feel like implementing a clone of that web service just to run the |
15 |
> test suite.) |
16 |
> |
17 |
> I don't like tests that need network access, but in a few cases, they |
18 |
> are the only to automatically verify that a package works. |
19 |
|
20 |
And 'skipped' tests simply mean that the test suite was unable to |
21 |
verify whether the package works for one reason or another. Well, other |
22 |
than build-time failures and a few possible runtime failures. |
23 |
|
24 |
You just have to ensure that it correctly notices the difference |
25 |
between 'no internet' and 'no matching API there'. Probably the domain |
26 |
resolution failure should be the borderline. |
27 |
|
28 |
Well, and I don't really mind having PROPERTIES about it. Some users |
29 |
may actually want to know that tests could do better with internet |
30 |
access. |
31 |
|
32 |
-- |
33 |
Best regards, |
34 |
Michał Górny |