| 1 |
On Mon, Mar 5, 2012 at 12:08 PM, Zac Medico <zmedico@g.o> wrote: |
| 2 |
> On 03/05/2012 08:50 AM, Mike Gilbert wrote: |
| 3 |
>> When I run into test failures in www-client/chromium, the normal |
| 4 |
>> procedure is to file a bug and filter the failing tests until the |
| 5 |
>> problem can be fixed (usually upstream). |
| 6 |
>> |
| 7 |
>> Ideally, I would like to be able to do the following: |
| 8 |
>> |
| 9 |
>> 1. Start building chromium using emerge or ebuild. |
| 10 |
>> |
| 11 |
>> 2. If src_test dies, modify the ebuild to filter the failing tests. |
| 12 |
>> |
| 13 |
>> 3. Re-run the test phase without rebuilding the source code. |
| 14 |
>> |
| 15 |
>> If I do not run ebuild clean, portage re-uses the environment from the |
| 16 |
>> previous run, which does not include my changes to the ebuild. |
| 17 |
>> |
| 18 |
>> Is there an easy way to have portage source the modified ebuild |
| 19 |
>> instead of just using the previous environment? Wiping out the |
| 20 |
>> environment file isn't really an option because we set some global |
| 21 |
>> variables in pkg_setup. |
| 22 |
> |
| 23 |
> FEATURES=noauto makes it source the ebuild again, mixing it in with the |
| 24 |
> previous environment. You can enable it temporarily by exporting the |
| 25 |
> FEATURES=noauto environment setting to the command: |
| 26 |
> |
| 27 |
> env FEATURES=noauto ebuild foo.ebuild test |
| 28 |
> -- |
| 29 |
> Thanks, |
| 30 |
> Zac |
| 31 |
> |
| 32 |
|
| 33 |
That sounds like exactly what I was looking for. Thanks! |
Archives