Transparent Requisites

Transparent requisites allow requisites to be defined on a function by function basis. This means that a given function can always requires any instance of another function, in the background. This makes it easy for state authors to ensure that executions are always executed in the correct order without the end user needing to define those orders.

It is easy to do, at the top of your system module just define the TREQ dict, this dict defines what functions will require what other functions:

TREQ = {
    "present": {
        "require": [
            "test.other.present",
        ]
    },
}

This stanza will look for the function named present inside of the module that it is defined in, then it will add require : - test.other.present for every instance found of test.other.present in the current run. If test.other.present is never used, then no requisites are set. Any requisite can be used, and multiple requisites can be used.

Transparent Requisite In Contracts

Transparent requisites can be added to Idem contracts. All the states that implement contract automatically inherit transparent requisite defined in the contract.

Unique Transparent Requisite

Another type of transparent requisite is unique. A function can be declared unique to prevent concurrent executions. The unique transparent requisite is significant in case of a parallel execution (default). TREQ dict at the top of the system module define unique with a list of functions within the module.

TREQ = {
    "unique": [
        "test.create",
        "test.delete",
    ]
}

In the example above, the instances of test.create within the current run will be invoked serially, and all the instances of test.delete will be invoked serially. Instance of test.create and test.delete can be invoked in parallel. It is achieved by selecting a single instance of the unique function, and setting the other instances of the same function as dependent on it. During the next run, a new instance will be selected. The unique requisite is re-evaluated in each run.