Well, I see a lot of religion in that thread, and I haven't much use for religion. There's a reason Ubuntu Universe and PPAs exist, after all. Like you, I'd rather see those reasons understood and turned into a robust solution.
But... GitLab's a really, really, really bad example. I have exactly 0 interest in installing a fixed version of GitLab in the first place. It develops truly valuable new features and the freeze policy is counterproductive to my goals in adopting it. GitLab is a VC startup able to support its own packaging choices and they aren't about to put real hard cash at risk because of someone else's packaging religion. Of course, singularity-container is largely maintained by a VC startup and has a similar technical issue with respect to Golang dependencies.
My view is that Debian should help these projects by allowing them to upload to unstable and gain the benefits of the build infrastructure / error detection while encouraging the use of the 3rd party repo structure:
https://wiki.debian.org/DebianRepository/UseThirdParty
https://packages.debian.org/stretch/neurodebian
https://packages.debian.org/stretch/neurodebian-archive-keyring
In the meantime the religious wars over volatile can be fought and decided. I'd encourage something like this for singularity-container.
Tom
On 1/4/19, 12:56 PM, "Carsten Aulbert" <carsten.aulbert@aei.mpg.de> wrote:
Hi Michael, Tom, all
On 1/4/19 6:05 PM, Michael Banck wrote:
> AFAIK the backports maintainers don't like that or are outright refusing
> to do this.
There was a quite long, partially "heated" discussion on the backports
list over the holidays about this topic but with gitlab as the main,
initial focus[1].
I don't know what the final verdict will be in the future, but I hope
something like what was discussed there will become a robust reality for
software where upstream is unwilling/unable to provide small patches for
security updates.
Cheers
Carsten
[1] https://lists.debian.org/debian-backports/2018/12/msg00081.html
--
Dr. Carsten Aulbert, Max Planck Institute for Gravitational Physics,
Callinstraße 38, 30167 Hannover, Germany
Phone: +49 511 762 17185
Attachment:
smime.p7s
Description: S/MIME cryptographic signature