$\mu$Puppet: A Declarative Subset of the Puppet Configuration Language

TL;DR

This paper formalizes a subset of the Puppet configuration language with operational semantics, facilitating understanding, debugging, and potential static analysis, while providing a foundation for future language improvements.

Contribution

It introduces $$Puppet, a formal semantics for a key subset of Puppet, enabling better analysis and understanding of the language's complex features.

Findings

Formal semantics clarifies complex language features

Provides a basis for static typechecking and debugging

Supports implementation in Haskell

Abstract

Puppet is a popular declarative framework for specifying and managing complex system configurations. The Puppet framework includes a domain-specific language with several advanced features inspired by object-oriented programming, including user-defined resource types, 'classes' with a form of inheritance, and dependency management. Like most real-world languages, the language has evolved in an ad hoc fashion, resulting in a design with numerous features, some of which are complex, hard to understand, and difficult to use correctly. We present an operational semantics for $\mu$Puppet, a representative subset of the Puppet language that covers the distinctive features of Puppet, while excluding features that are either deprecated or work-in-progress. Formalising the semantics sheds light on difficult parts of the language, identifies opportunities for future improvements, and provides a foundation for future analysis or debugging techniques, such as static typechecking or provenance tracking. Our semantics leads straightforwardly to a reference implementation in Haskell. We also discuss some of Puppet's idiosyncrasies, particularly its handling of classes and scope, and present an initial corpus of test cases supported by our formal semantics.