[openstack-dev] [neutron] -2'ing all patches on every gate breakage

Hayes, Graham graham.hayes at hpe.com
Tue Apr 5 11:53:56 UTC 2016


On 04/04/2016 23:20, Carl Baldwin wrote:
> On Mon, Apr 4, 2016 at 3:22 PM, Doug Wiegley
> <dougwig at parksidesoftware.com> wrote:
>> I don’t know, -1 really means, “there is something wrong, the submitter
>> should fix it and clear the slate.”  Whereas -2 has two meanings.  The first
>> is “procedural block”, and the second is “f*** you.”
>>
>> I really don’t see a reason not to use the procedural block as a procedural
>> block. Are you not trusting the other cores to remove them or something?
>> It’s literally what it’s there for.
>
> I'm not complaining.  I've had plenty of these -2s and I understanding
> the reason behind it.  But, I thought I'd chime in.
>
> I interpret a -2 on a patch as a procedural block because of something
> related to the patch.  It is awkward as a procedural block when it is
> being applied due to circumstances that have nothing to do with the
> patch itself and the only person who can remove the block is the
> person who applied it in the first place.  That person might get
> distracted, leave for the week-end, go on vacation, etc.
>
> Would it be nice if the project itself had an easy procedural block?
> A single switch that turns off entering the gate queue for the entire
> project?  Wouldn't it also be nice if the switch could be toggled by
> any one of a group responsible for it?  I think it would be nice but
> I'm not sure how it could be easily implemented.
>
> Carl

See my response earlier. if it is too "over engineered" there is
variations that could work (e.g. a script with shared creds / some sort
of side repo that triggers a jenkins job to run it) - I just suggested
a bot as it seemed easier.

On a side note, I have always thought that our overloaded use of -2 to
procedurally block things was a bit weird. A new label seems like a
much better idea.



More information about the OpenStack-dev mailing list