Re: Subversion 1.15.0-rc2 up for testing/signing

On Mon, Jul 6, 2026 at 12:48 PM Evgeny Kotkov via dev <
[email protected]> wrote:

> Evgeny Kotkov <[email protected]> writes:
>
> > That leaves (2) as the only remaining item.
> >
> > From the RM perspective, I'd lean toward reverting the XML schema
> validation
> > from the 1.15.x branch so we can roll RC3, and finalizing the validation
> > work on trunk on its own schedule.
>
> The responses on this topic have been somewhat divided so far.
>
> I think there are factors that support staying on the course where we
> backport
> the new version of the change (r1935776):
>
> - The original change with venv is already a part of 1.15.x.  That puts it
> on
>   a different track than, for example, adding a new feature.
>
> - The change has a reduced blast radius: it doesn't touch the core code and
>   is focused mostly on the tests.  (This part is a bit arguable, since it
>   also changes the build system, but I still think the argument applies.)
>
> - Even if we aim for a revert, it would still have to go through STATUS,
>   possibly with a backport branch of its own.  So there's also some
> community
>   effort required.



I'd like to add one more factor to support the new version of the change:
avoiding complexity and confusion.

Since we're already planning to keep the new version for the future (1.16),
1.15 hasn't been released yet, and 1.14 didn't have it at all, I think it
is better to release 1.15 with the new version.

Otherwise we'll be stuck supporting both the old and the new versions for a
long time.

Weighing the risk/benefit of backporting it to 1.15 now versus having to
support both versions for an indefinite amount of time, I think the risk is
small and the reduced maintenance burden is worth it.


With that in mind, I'm inclined to proceed with the following plan: wait
> until
> the r1935776 change is merged to 1.15.x, and roll RC3 shortly after that.
> I also plan to cast my own vote on the nomination, if time permits.
>
> If this plan doesn't work out (for example, if we run into issues with the
> currently proposed nomination), I'll make an alternative nomination that
> cuts the change from 1.15.x, retargeting it to 1.16.
>
> Thoughts?
>
>
> Thanks,
> Evgeny Kotkov
>

This plan sounds reasonable to me.

Thanks,
Nathan