1 |
On Sat, 5 Aug 2006 14:35:49 -0400 |
2 |
Mike Frysinger <vapier@g.o> wrote: |
3 |
|
4 |
> On Saturday 05 August 2006 06:57, Kevin F. Quinn wrote: |
5 |
> > On Sat, 5 Aug 2006 11:49:53 +0200 |
6 |
> > Danny van Dyk <kugelfang@g.o> wrote: |
7 |
> > > Please re-read the list of packages that fail tests: |
8 |
> > > * glibc |
9 |
> > > * autoconf |
10 |
> > > * gettext |
11 |
> > > * tar |
12 |
> > > That makes _4_ system packages. Before I would consider making |
13 |
> > > FEATURES=test a default, I would add least want the system set to |
14 |
> > > actually merge with it. |
15 |
> > |
16 |
> > So you're happy to let users install these packages without them |
17 |
> > knowing the tests would fail? |
18 |
> |
19 |
> before i added binutils-2.17, i ran `make check` on it for about 25 |
20 |
> targets ... of those, about 10 failed ... |
21 |
> |
22 |
> i checked with upstream and others reproduced it ... i dont know |
23 |
> about you, but i dont have the skills to go in and fix the failures |
24 |
> for all of those architectures |
25 |
|
26 |
Agreed, however you could rig src_test() to either skip the tests on |
27 |
those arches, or run them without die()ing and ewarn about the known |
28 |
failures on those arches, or just leave them in ~arch (or even |
29 |
masked) for the arches where they fail, depending on the impact of the |
30 |
failures. That sort of thing is well within our role of package |
31 |
management. |
32 |
|
33 |
> while i like the idea of all packages being able to pass |
34 |
> FEATURES=test, somethings it just aint gonna happen with Gentoo's |
35 |
> available skill set -mike |
36 |
|
37 |
-- |
38 |
Kevin F. Quinn |