That would be step 6, after any issues have been resolved after the push to
gerrit.

Nathan.
On Sep 5, 2012 8:24 PM, "Jason Edgecombe" <[email protected]> wrote:

> On 09/05/2012 03:37 PM, Derrick Brashear wrote:
>
>> On Thu, Jun 14, 2012 at 1:20 PM, Troy Benjegerdes <[email protected]>
>> wrote:
>>
>>> On Mon, Apr 16, 2012 at 07:28:23PM -0400, Derrick Brashear wrote:
>>>
>>>> 1) make the tests be built as part of "make all"; push to gerrit. it
>>>> may fail. if so,
>>>>
>>> now builds (current iteration of 7225)
>>
>> so i guess after the predecessor patchsets are merged, we can move on to
>> step 4.
>>
>>  2) attempt make the tests actually build on all platforms; rebase that
>>>> change before the one described in 1, and push both to gerrit.
>>>> 3) iterate on 2 until things build everywhere
>>>> 4) a simple way to then work on getting the tests working everywhere
>>>> without breaking every slave: submit a change which makes "make check"
>>>> happen as part
>>>> of "make all"; mark it not to be merged, but buildbot will happily
>>>> explode wherever a buildslave would be screwed.
>>>> 5) fix it or decide we don't care or each platform, iterating over 4.
>>>> 6) when we know which platforms are done, turn it on for those
>>>> platforms.
>>>>
>>> I wanted to follow up and see how we were doing on this list..
>>> Is there anything I could be doing that might help?
>>>
>>
>>
>>  Should I add the "make check" step as a build step on some, but not all,
> of the buildslaves? That would not require an "all or nothing" approach. If
> you want make check on some of the buildslaves. Just let me know which ones.
>
> Jason
>
> Jason
> ______________________________**_________________
> OpenAFS-devel mailing list
> [email protected]
> https://lists.openafs.org/**mailman/listinfo/openafs-devel<https://lists.openafs.org/mailman/listinfo/openafs-devel>
>

Reply via email to