Safeness of Make-based Incremental Recompilation

Niels Jørgensen

    Research output: Working paperResearch

    121 Downloads (Pure)

    Abstract

    The make program is widely used in the software industry to reduce compilation time in large projects.make skips source files that would have compiled to the same result as in the previous build. (Or so it is hoped, at least.) The crucial issue of safeness of omitting a full build-from-scratch is addressed by defining a semantic model for make. The model is in some ways similar to models proposed for logic programming languages, because make files, similarly to logic programs, have no global variables and execution is query driven. Safeness is shown to hold if a set of criteria are satisfied, including soundness, fairness, and completeness of make file rules. The safeness result is useful for the make file programmer because these criteria, while stated formally,are also intuitively reasonable, and may form a basis for a kind of checklist for make file rules. The rigorous semantic definition for make may also be helpful in the construction of tools for automatic make file generation.
    Original languageEnglish
    Place of PublicationFrederiksberg
    PublisherDepartment of Informatics INF, Copenhagen Business School
    Number of pages7
    Publication statusPublished - 1999
    SeriesWorking Paper / Institut for Informatik. Handelshøjskolen i København
    Number1999-12

    Cite this