Bug#1032120: Upload of new upstream version before fix has migrated to testing (Was: Re: Bug#1032120: tiledb: uses atomic operations, but is not linked to libatomic)
[Release team in CC] Am Wed, Mar 15, 2023 at 08:45:21AM +0530 schrieb Nilesh Patra: > I am removing myself from the uploaders field/maintenance of tiledb-py. Feel > free to update it once tiledb is ready for migration. The repository > lives at python team namespace. I admit I'm strongly in favour of following release policy properly instead of putting work on other DDs who need to check rdepends. I became aware of this bug since genomicsdb received a testing removal warning. Thus I insist that the correct fix for this violation of the freeze policy is to upload 2.15.0-2+really_2.14.1 (or at your preference using an epoch) as I wrote before in this bug log[1]) @Dirk, I would be happy if you could simply say sorry guys for creating this noise and follow freeze policy. Kind regards Andreas. [1] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1032120#37 -- http://fam-tille.de
Bug#1032120: Upload of new upstream version before fix has migrated to testing (Was: Re: Bug#1032120: tiledb: uses atomic operations, but is not linked to libatomic)
Hi, I am removing myself from the uploaders field/maintenance of tiledb-py. Feel free to update it once tiledb is ready for migration. The repository lives at python team namespace. -- Best, Nilesh signature.asc Description: PGP signature
Bug#1032120: Upload of new upstream version before fix has migrated to testing (Was: Re: Bug#1032120: tiledb: uses atomic operations, but is not linked to libatomic)
On Sun, Mar 12, 2023 at 07:27:46PM +0530, Nilesh Patra wrote: > On 3/12/23 19:00, Dirk Eddelbuettel wrote: > > > > Why not wait a week on 2.15.0-1 which now 'Too young, only 4 of 10 days > > old'? > > It won't migrate as told earlier too. By tomorrow it would have 'blocked by > freeze' written over there. And it has that written over there now. Another better option can be to add autopkgtests to tiledb, as this is not a key package and hence can migrate w/o freeze block. Let me know what you think. -- Best, Nilesh signature.asc Description: PGP signature
Bug#1032120: Upload of new upstream version before fix has migrated to testing (Was: Re: Bug#1032120: tiledb: uses atomic operations, but is not linked to libatomic)
Tl;dr: Because It will not migrate and eventually get kicked out of testing and next stable on April, 13 regardless of what the tracker says now. On 3/12/23 19:00, Dirk Eddelbuettel wrote: Why not wait a week on 2.15.0-1 which now 'Too young, only 4 of 10 days old'? It won't migrate as told earlier too. By tomorrow it would have 'blocked by freeze' written over there. Hard freeze starts tomorrow, and it'd need an unblock request which wouldn't be agreed upon as only changes like these[1] are allowed in hard freeze. The version in testing has an RC bug, and hence this will be taken out of testing if we do not revert the version to a previous one. [1]: https://release.debian.org/bullseye/freeze_policy.html#appropriate Best, Nilesh
Bug#1032120: Upload of new upstream version before fix has migrated to testing (Was: Re: Bug#1032120: tiledb: uses atomic operations, but is not linked to libatomic)
Why not wait a week on 2.15.0-1 which now 'Too young, only 4 of 10 days old'? Dirk -- dirk.eddelbuettel.com | @eddelbuettel | e...@debian.org
Bug#1032120: Upload of new upstream version before fix has migrated to testing (Was: Re: Bug#1032120: tiledb: uses atomic operations, but is not linked to libatomic)
Dirk - Ping on this? If you do not have the time, let me know. I'll do a NMU for tiledb to revert to prev version and also file an unblock request. On 3/10/23 20:53, Nilesh Patra wrote: Quoting Andreas Tille: your recent upload of tiledb 2.15.0-1 into unstable is quite unfortunate in freeze policy. You should have waited at least until 2.14.1-2 fixing this bug to migrate to testing. You would need to ask release team for migration of 2.15.0 which will probably be refused The changes between 2.14 and 2.15 are non-trivial and are likely to be rejected by release team at the moment, considering there are no autopkgtests either. even if you would be able to fix the regression in autopkgtest[1]. This is because the version of tiledb-py is not compatible with tiledb 2.15.0. Uploading a new version of tiledb-py would fix this, but: - It'd be useless unless tiledb 2.15.0 migrates. - It seems to need pyarrow for tests, and other functionalities, which is un-packaged. So current situation is: * tiledb 2.14.1-1 in testing (and next stable) is marked for auto-removal. * consequently, tiledb-py is also marked for autoremoval. * tiledb 2.15.0-1 can't migrate because of freeze policy. What a freaking mess! Dirk, we could have avoided all of it if you had quite literally ** waited ** for a few hours to let tiledb migrate on the 8th of March. I'm sad to see such mess-ups happening at this time as I had put quite some efforts on tiledb-py myself too. The solution right now is what Andreas said: My recommendation would be to upload a version 2.15.0-2+really_2.14.1 Dirk, please fix this so tiledb can make a (valid) stable release. [1]: https://tracker.debian.org/pkg/tiledb