1 |
On Thu, 16 Sep 2010 22:42:02 +0200 |
2 |
Thomas Sachau <tommy@g.o> wrote: |
3 |
|
4 |
> Controling the test USE flag alone without the test FEATURE is useless, since it wont run the |
5 |
> src_test phase. |
6 |
|
7 |
...then don't do that? :P |
8 |
|
9 |
> And being able to disable the test USE flag with FEATURES=test will result in |
10 |
> missing deps or build-system args. |
11 |
|
12 |
Add something like this to the beginning of the default src_test: |
13 |
|
14 |
hasq test $USE || return 0 |
15 |
|
16 |
I'm sure there's a reason it won't work I'm not thinking of, but that's |
17 |
the jist of the idea. |
18 |
|
19 |
> Can you tell me any reason, why you want to expose and control |
20 |
> the test USE flag independently of FEATURES=test? |
21 |
|
22 |
So it can be controlled on a per-package basis? Circular dependencies, |
23 |
unwanted dependencies, excessive testsuites, perpetually failing testsuites |
24 |
that no one bothers to fix... You can mask the flag but it's inconsistant |
25 |
and unintuitive. FWIW I currently have 22 packages that have the test flag |
26 |
masked in package.use.mask, and probably a dozen more I just haven't gotten |
27 |
around to masking. |
28 |
|
29 |
In any case, this is straying off topic. I believe the current policy is |
30 |
this: if your package's test suite requires extra dependencies, add a test |
31 |
USE flag and make them conditional on that. I don't think there's any reason |
32 |
it should change. |
33 |
|
34 |
|
35 |
-- |
36 |
fonts, gcc-porting, we hold our breath, we spin around the world |
37 |
toolchain, wxwidgets you and me cling to the outside of the earth |
38 |
@ gentoo.org EFFD 380E 047A 4B51 D2BD C64F 8AA8 8346 F9A4 0662 |