1 |
On 03/05/2012 08:50 AM, Mike Gilbert wrote: |
2 |
> When I run into test failures in www-client/chromium, the normal |
3 |
> procedure is to file a bug and filter the failing tests until the |
4 |
> problem can be fixed (usually upstream). |
5 |
> |
6 |
> Ideally, I would like to be able to do the following: |
7 |
> |
8 |
> 1. Start building chromium using emerge or ebuild. |
9 |
> |
10 |
> 2. If src_test dies, modify the ebuild to filter the failing tests. |
11 |
> |
12 |
> 3. Re-run the test phase without rebuilding the source code. |
13 |
> |
14 |
> If I do not run ebuild clean, portage re-uses the environment from the |
15 |
> previous run, which does not include my changes to the ebuild. |
16 |
> |
17 |
> Is there an easy way to have portage source the modified ebuild |
18 |
> instead of just using the previous environment? Wiping out the |
19 |
> environment file isn't really an option because we set some global |
20 |
> variables in pkg_setup. |
21 |
|
22 |
FEATURES=noauto makes it source the ebuild again, mixing it in with the |
23 |
previous environment. You can enable it temporarily by exporting the |
24 |
FEATURES=noauto environment setting to the command: |
25 |
|
26 |
env FEATURES=noauto ebuild foo.ebuild test |
27 |
-- |
28 |
Thanks, |
29 |
Zac |