Do not maintain multiple image IDs

[apyrgio]

Currently, Dangerzone ships with two files in its share/ directory that are not part of this repo:

• share/container.tar.gz, which is the image tarball that accompanies every new Dangerzone release.
• share/image-id.txt, which is a file that holds all the image IDs that match the contents of this tarball.

The purpose of the image ID file is to specify what is the expected image ID in the local Podman/Docker installation. If not found, Dangerzone will load the image tarball, which takes a while. The problem with image IDs though is that they have stopped being reliable, now that Docker has introduced its own containerd image store (see #933). For this reason, as part of the release we have to get two image IDs, one for the regular image store that Docker and Podman use, and one for the containerd image store. This is a pain for several reasons:

1. It's a release task that is not part of our CI, and is therefore not tested.
2. It's easy to forget and ship an image-id.txt file with just one image ID.
3. It can't be automated easily, as you would either require two separate container runtimes, or manually changing the image store of the Docker runtime.
4. Depending on the order of image creations, we may be bitten by this bug: Podman v3 cannot load container images built with BuildKit #988

For this reason, it would be best to use a different type of image ID, such as image tags. We can tag our Dangerzone images with a unique tag, include it in the image-id.txt file, and look for the required tag in docker images / podman images, which we already do. This change should not affect the developer's and user's workflow, but it will improve the release situation a lot.