Bashtard v2.1.0

Its been about another year since I made a post about Bashtard, its 2.0.0 release. Today marks the 2.1.0 release, meaning I’ve gone almost a year without breaking backwards compatibility. To me, this is a very good sign.

The release today isn’t as big as the 2.0.0 release was, mostly because most of the functionality I want to have is already present. Some new features were added over the year, though. The most important one is variable references. This allows re-use of a variable’s value for another variable. Its quite simplistic in how it works, due to the nature of Bashtard being written in Bash and trying to keep things rather simple and lightweight. It does however get the job done for most use-cases I had for it.

Another feature that was added in this release is the zap subcommand. It is a convenience command more than anything, simply removing an existing registry entry without going through playbook_del(). The use case for this is mostly for testing new playbooks. I found that while writing a playbook I’d often remove the playbook entry from the registry to re-run the playbook_add() function to see if it works exactly as desired, and I wanted this to be more convenient. In theory this new zap subcommand is also useful for dealing with entries of broken or removed playbooks.

For a future release I’d like to add import and export subcommands, for making it easier to handle external playbooks. The import subcommand is intended to take an URL and optionally a name to simply add an existing repository as a submodule in the playbooks.d directory.

The export subcommand should take a name as it exists in the playbooks.d directory and turn it into a git submodule, so it can be pushed to its own repository. The intent is that you can just make a regular playbook for use within Bashtard, and if you decide “hey, this could actually be useful for others in its current state”, you can simply export it, and push it to a repository for other people to pull from.

Additionally, I would like to remove the backup subcommand from Bashtard, as I feel it adds a level of bloat and scope-creep which simply should not be there. While this would result in a 3.0.0 release of Bashtard, keeping it just for backwards compatibility seems a little silly to me.

I’m on the fence for the ssh subcommand, as it seems more closely aligned to Bashtard being used to manage several systems, and ssh can be very useful to check something across your entire network. Considering the recent SSHd vulnerability, it was quite easy to run a single command across all my machines and get an overview of which machines were affected.

Let’s see what the rest of this year brings in Bashtard changes!