Testing your Mustache implementation against this specification should be relatively simple. If you have a readily available testing framework on your platform, your task may be even simpler.

In general, the process for each .yml file is as follows:

1. Use a YAML parser to load the file.

2. For each test in the 'tests' array:

   1. Ensure that each element of the 'partials' hash (if it exists) is stored in a place where the interpreter will look for it.

   2. If your implementation will not support lambdas, feel free to skip over the optional '~lambdas.yml' file.

   2.1. If your implementation will support lambdas, ensure that each member of 'data' tagged with '!code' is properly processed into a language- specific lambda reference.

   • e.g. Given this YAML data hash:

     { x: !code { ruby: 'proc { "x" }', perl: 'sub { "x" }' } }

     a Ruby-based Mustache implementation would process it such that it was equivalent to this Ruby hash:

     { 'x' => proc { "x" } }

   • If your implementation language does not currently have lambda examples in the spec, feel free to implement them and send a pull request.

   • The JSON version of the spec represents these tagged values as a hash with a '__tag__' key of 'code'.

     1. Render the template (stored in the 'template' key) with the given 'data' hash.

     2. Compare the results of your rendering against the 'expected' value; any differences should be reported, along with any useful debugging information.

        • Of note, the 'desc' key contains a rough one-line description of the behavior being tested -- this is most useful in conjunction with the file name and test 'name'.