+1 (binding) On Wed, Oct 7, 2026 at 1:57 PM Yufei Gu <[email protected]> wrote:
> +1(binding) > Yufei > > On Wed, Oct 7, 2026 at 1:55 PM Péter Váry <[email protected]> > wrote: > > > > +1 > > > > Matt Topol <[email protected]> ezt írta (időpont: 2026. okt. 7., > Sze, 22:30): > >> > >> +1 (binding) > >> > >> On Wed, Oct 7, 2026, 4:03 PM Neelesh Salian <[email protected]> > wrote: > >>> > >>> +1 (non-binding) > >>> > >>> On Thu, Oct 8, 2026 at 03:23 Steven Wu <[email protected]> wrote: > >>>> > >>>> Hi, > >>>> > >>>> I propose requiring each new snapshot in a v4 table to have a > timestamp-ms strictly greater than its parent snapshot’s timestamp on the > same branch. Root snapshots are unconstrained, and v1–v3 behavior remains > unchanged. > >>>> > >>>> This ensures snapshot timestamps align with snapshot order. This > change would resolve time travel issues caused by out-of-order snapshot > timestamps. This can also enable a row timestamp metadata column inherited > from the snapshot timestamp (a separate follow-up v4 spec change). > >>>> > >>>> Spec PR: #16294 > >>>> Reference implementation: #16293 > >>>> > >>>> Please vote in the next 72 hours: > >>>> > >>>> [ ] +1 Adopt the requirement > >>>> [ ] +0 No preference > >>>> [ ] -1 Do not adopt the requirement (please explain) > >>>> > >>>> Thanks, Steven > >>>> > >>>> > >>>> Discussion thread: > https://lists.apache.org/thread/cln7xtjy9kw2096gzqlg5c4gg8b9td10 >
