Bug#985028: Bug#981413: age: New upstream version (1.0.0~beta6)
On Thu, Mar 18, 2021 at 12:53 AM Filippo Valsorda <filippo@ml.filippo.io> wrote:
>
> 2021-03-17 17:33 GMT+01:00 Shengjing Zhu <zhsj@debian.org>:
>
> On Thu, Mar 18, 2021 at 12:06 AM Filippo Valsorda <filippo@ml.filippo.io> wrote:
> >
> > 2021-03-17 16:19 GMT+01:00 nicoo <nicoo@debian.org>:
> > > This is currently blocked on golang-filippo-edwards25519 getting into
> > > testing, which didn't happen because of the whole issue with NEW requiring
> > > binary uploads and transition requiring source-only ones.
> >
> > Thank you for packaging age and for picking up this upload!
> >
> > I dropped the golang-filippo-edwards25519 dependency in v1.0.0-rc.1 to
> > make it easier to get that version into Bullseye.
> >
> > Let me know if it's fine to add it back in a later v1.0, or if it would
> > block bugfixes from making it into Bullseye during the cycle, and I
> > should wait for v1.1.
> >
>
> It depends on what version of age wants to be included in Bullseye.
> As per Bullseye freeze policy[1], golang-filippo-edwards25519 is out
> of luck to be included in Bullseye(No new package since 2/12). But it
> will be included in the next release of cause.
> Currently age v1.0.0-rc1 seems fine for Bullseye, it will migrated
> from Unstable to Bullseye after 20 days.
>
>
> v1.0.0-rc.1 is definitely ok for the Bullseye release.
>
> I was wondering if I should delay the golang-filippo-edwards25519
> dependency to v1.1.0 to allow v1.0.x bugfixes into later Bullseye point
> releases, or if the bar for those is so high that it doesn't matter
> anyway.
IMO, it doesn't matter too much.
But if you can maintain a stable branch which has minimal diff to
v1.0.0-rc1(no new dependency, only backporting important patches),
that would be best for Bullseye.
Otherwise, we can cherry-pick patches for Bullseye ourselves. An usual
life cycle for Debian release is about 5 years(with LTS), so it's sure
much easier if upstream can maintain a stable branch.
--
Shengjing Zhu
Reply to: