[all][gate][stable] Pinning tox<4 in stable branch testing

Clark Boylan cboylan at sapwetik.org
Fri Jan 6 16:36:36 UTC 2023



On Fri, Jan 6, 2023, at 6:49 AM, Jeremy Stanley wrote:
> On 2023-01-06 08:42:45 -0500 (-0500), Brian Rosmaita wrote:
> [...]
>> The zuul ensure_tox role only ensures that tox is present.  The
>> ensure_tox_version var has a slightly misleading name in that it is only
>> used when the role decides it needs to install tox, and then it uses the
>> value of that var; it doesn't ensure that the available tox is that version.
> [...]
>
> And even if it did, it would only be able to do so at the point at
> which it was invoked, so if something reinstalled/upgraded tox after
> that point it would still be different than whatever you set in
> ensure_tox_version. Still, it might be possible to make
> ensure_tox_version enforce a specific version if set which would
> catch circumstances like the one you observed where something
> running in the job before the ensure-tox role is invoked installed a
> different version globally.

Doing so would just mask that you've improperly installed tox elsewhere (a bug itself). I think the current behavior is correct because it isn't letting us get away with that. Elsewhere we have tools (devstack) that install tempest at least 3 redundant times for every tempest job. Its a waste of effort and we shouldn't make it easier for ourselves to fall into those traps.

> -- 
> Jeremy Stanley
>
> Attachments:
> * signature.asc



More information about the openstack-discuss mailing list