FAQ.md

FAQ

## How does Kirby calculate a coverage?

Here are 3 points to understand Kirby.

1. After each task execution, Kirby runs serverspec.

   Simply put, when Kirby is not installed, ansible works like this:

   

   When Kirby is installed, ansible and Kirby work like this:

   

2. Determines whether a task is tested, by comparing serverspec's results.

   When serverspec runs, it returns a result like this:

   Failures:
   
     1) File "./dir1" should be directory
        Failure/Error: it { should be_directory }
          expected `File "./dir1".directory?` to return true, got false
        # ./spec/localhost/sample_spec.rb:4:in `block (2 levels) in <top (required)>'
   
   Finished in 0.79599 seconds (files took 0.79919 seconds to load)
   1 example, 1 failure
   
   Failed examples:
   
   rspec ./spec/localhost/sample_spec.rb:4 # File "./dir1" should be directory

   Kirby checks the number of failed tests (1 in this case).

   If the number of failed tests is smaller than the last one, the task is tested. Otherwise, not tested.

3. Calculate a coverage

   After a playbook is executed, Kirby calculates a coverage using the simple formula:

   the number of tested tasks / the number of tasks

   If Kirby removes tasks from the coverage, these tasks are not considered.

## It takes longer time to run a playbook.

It takes longer time, since Kirby runs serverspec many times.

Here are several ways to shorten running time:

1. Use changed_when

   Kirby does not run serverspec if the task's result is not changed, so use changed_when to override the task's result if possible.

2. Parallelize Serverspec

   It is easier than it sounds. There are useful gems like parallel_tests.

   Here is the blog post by the serverspec's author about it (it's Japanese, but Google translate works well).

Please share your idea to improve this!

## How can I disable Kirby?

To disable:

export KIRBY_ENABLE=no

To enable:

export KIRBY_ENABLE=yes or unset KIRBY_ENABLE.