I am not sure if this vote is still proceeding, but based on Ephraim's
feedback on the discussion thread, I am voting

-1 (binding)
Until that discussion is resolved.


On Mon, Jul 27, 2026 at 1:25 PM Sumit Maheshwari <[email protected]>
wrote:

> +1 (binding)
>
> On Mon, Jul 27, 2026 at 11:40 AM Dev-iL <[email protected]> wrote:
>
> > +0.75 (non -binding): +1 on the need, +0.5 on edge cases considered.
> >
> > One concern I have is how to deal with dag versions that should be
> > ineligible for pinning. For example, suppose version V of the dag assumes
> > dependency X, and V+1 assumes X+1. If the deployment is updated to X+1,
> > what's to stop the user from switching to the now-incompatible (either
> > explicitly - parse time error, or implicitly - runtime error) version V?
> >
> > I don't have a good solution for this. A conceptually simple solution
> would
> > be a dag-level metadata that contains critical dependency version pins,
> but
> > parsing and enforcing those seems to increase the scope of the AIP by a
> > lot.
> >
> > On Mon, 27 Jul 2026, 8:14 Piyush Maheshwari, <[email protected]
> >
> > wrote:
> >
> > > Hi everyone,
> > >
> > > I would like to call a vote on AIP-109 DAG Version Pinning:
> > >
> > >
> >
> https://cwiki.apache.org/confluence/display/AIRFLOW/AIP-109+DAG+Version+Pinning
> > >
> > > The previous discussion thread can be found here:
> > > https://lists.apache.org/thread/6h811nmzjrgfhj1b0kwqtjjlhvc5jrvr
> > >
> > > I'd like to thank everyone who took the time to comment on the AIP and
> > > participate in the discussion threads. Your feedback has significantly
> > > improved the proposal.
> > >
> > > The vote will remain open for 5 days and will close on *Monday, 3
> August
> > > 2026, at 05:00 GMT*.
> > >
> > > Everyone is encouraged to vote, although only votes from PMC members
> and
> > > committers are considered binding.
> > >
> > > Please vote accordingly:
> > >
> > > [ ] +1 Approve
> > > [ ] +0 no opinion
> > > [ ] -1 disapprove with the reason
> > >
> > > Regards,
> > > Piyush
> > >
> >
>

Reply via email to