Re: adding a snippet to a patch

[David Nalesnik]

On Wed, Jan 11, 2017 at 3:27 AM, James <address@hidden> wrote:
> Hello,
>
>
> On 11/01/17 08:08, Mark Knoop wrote:
>>
>> At 15:51 on 09 Jan 2017, David Nalesnik wrote:
>>>
>>> On Mon, Jan 9, 2017 at 2:00 PM, David Nalesnik
>>> <address@hidden> wrote:
>>>>
>>>> Hi,
>>>>
>>>> I'm adding a snippet to a patch dealing with hairpins, and I'm not
>>>> sure what I need to do.
>>>>
>>>> I've added the snippet to Documentation/snippets/new.
>>>>
>>>> Do I then need to run scripts/auxiliar/makelsr.py and add the
>>>> resulting files to my patch for upload to Rietveld?
>
>  For testing/code review uploading a patch that also contains a makelsr.py
> run can sometimes end up affecting a lot of other files that are not really
> part of the review.
>
> So, what others have done in the past which I think is more convenient, is
> to upload the patch to Rietveld *with* the snippet change but *without* the
> makelsr.py run.
>
> In the 'olden' days when I had a script to download and run the tests (i.e.
> without any need for any kind of intervention from me other than to check
> the reg test output for anything unusual) I would in the cases of snippets
> in patches, run the tests manually and include a makelsr.py as part of the
> workflow for testing the patch.
>
> This is one of the benefits of having flexible patch testing methods by the
> way.
>
> Then once the patch was approved the general consensus was to apply the
> patch 'as is', then, in a separate commit the makelsr.py, that way it was
> easier to backout something that one massive patch+makelsr.py checkin.
>
> So if you prefer to do that - let the patch test include the makelsr.py -
> then just make sure that is stated in the Tracker (in case the patch tester
> - which is usually me) doesn't notice a snippet being added/modified.
>
> I hope that helps.
>

Thank you, James.  I will follow your recommended procedure when it
comes time to put my newer patch set up.

David