GitLab

Seems harmless, but trips our script due to errexit

https://github.com/containers/podman/issues/4314

This takes a lot of time, so the person making changes will just need
to push it to the registry themselves!

Alpine edge has a go version too new for u-root.

The patch will be sent to podman's upstream after proving we are happy
with it.

The only untested code-path should be the module loading, but I am
unsatisfied by it and will be replaced soon anyway :)

Up until now, chained containers would execute one after the other as
long as they would have the exit code 0. Let's change that to ignoring
the exit code, and allow the user to set b2c.pipefail to restore the
previous behaviour.

This is to mimic what shell scripts are doing, and thus be less of a
surprise to users.
Suggested-by: Charlie Turner <cturner@igalia.com>

This information is shared to the container using environment
variables.

Closes: #1

By using eval, we enable the use of ;, &&, and ||, which might be
useful to some people.

By setting the pull policy to `always`, we make sure the image is up to
date, before executing it.

By syncing, we make sure that the layers get stored on the disk, before
we potentially crash.

Fixes: 44808c4f

If a container failed to start/stop, they cannot be pruned. Let's just
remove all the containers in cleanup since none are useful at this
point,

The network can take a while to be up, so let's do as much as we can
ahead of blocking on it!

Fixes: 4f812b88

Fixes: 30e083d8

The default is for the machine to shut down.

The first release is out :)

This allows people to generate their initramfs without having to
generate the building container themselves (slow).