Use gemfile.d for upstream CI

[ekohl]

The gemspec shouldn't contain environment variables, because it's also used in other contexts where it could break.

In common CI we read gemfile.d/*.rb so this is the best place for bundler overrides.

Fixes: 3eedb81 ("feat(actions): switch to upstream CI templates (#113)")

I'm not 100% sure this works, but if it does then I'll document it as a recommended approach.

[laugmanuel]

This approach looks simpler/cleaner - let's see what the pipeline says.

[ekohl]

I failed rubocop. Now updated

[laugmanuel]

Now failing due to theforeman/foreman-js#477 / theforeman/foreman#10116.
Will this patches be backported to Foreman 3.7 / 3.8 / 3.9?

[ekohl]

Will this patches be backported to Foreman 3.7 / 3.8 / 3.9?

We literally just now discussed it for 3.9, but 3.7 & 3.8 are EOL and I don't think we'll update those anymore. There certainly won't be any more releases.

[ekohl]

@laugmanuel since current master will also fail on the same thing, should this be merged?

[laugmanuel]

Will this patches be backported to Foreman 3.7 / 3.8 / 3.9?

We literally just now discussed it for 3.9, but 3.7 & 3.8 are EOL and I don't think we'll update those anymore. There certainly won't be any more releases.

I guess we need to drop tests for that Foreman version then. I'll prepare a PR for that.
What's the plan for 3.9 and newer?

@laugmanuel since current master will also fail on the same thing, should this be merged?

Let's wait until old versions are dropped from CI.

[ekohl]

What's the plan for 3.9 and newer?

Do cherry picks of theforeman/foreman@7008091, which is being worked on.

[laugmanuel]

Thanks @ekohl