Thanks all for the feedbacks,
I am 100% behind the idea that, the end user does not care about which
fedora base image is used (as long as he trust that patch and security
fixes are applied). If I want to use the nginx container for example
as a user I am more interested in the version of nginx than the
version of fedora.
I also feel that having to specify the version fedora during the pull
is not a very nice user experience. (ie docker pull
Something more user friendly would be docker pull
My feeling is that we are not making a great use of the container tags.
Coming back to the Fedora Container Release, the current tooling
automates the REbuilds of containers. So if an image was never built
against the f28 branch the tooling will not be able to rebuild the
image. Also currently when a new branch is created ( for example f28 )
maintainer still need to update their FROM line to point to the
correct base image.
I think we need to come up with exactly what we want to achieve with
the container release and once we have a better idea changing or
adapting the tooling will be the next logical step.