Instead of interrupting task evaluation when condition is not met,
allow the whole task to be evaluated (including condition and
evaluation) so that the interpreter will get all tasks actions (whether
condition is met or not) and be able to query the condition.
* Modify Interpreter#process_task: test if task condition is met before
applying the actions;
* Implement condition handling in Task and Task::DSL;
* Implement Condition and Condition::DSL (useless as they are, but
needed to implement later test keywords as part of the condition DSL.
* Move recipe processing code in the worker;
* Refactor CLI and use the the worker;
* Implement Recipe#tasks and remove tasks application during evaluation,
tasks are now applied by the worker after all evaluations are done.
* Report invalid action usages from tasks;
* Implement backtrace cleaning in CLI;
* Extract error class declarations in a new errors file;
* Replace raise with fail keyword in task DSL class.
