Design of a Modular Platform for Static Analysis

Abstract : We present the design and implementation of Mopsa, a platform that simplifies the construction of semantic static analyzers by abstract interpretation. Mopsa computes sound program invariants and reports run-time errors, undefined behaviors, and uncaught exceptions. Mopsa differs from existing platforms by its highly modular and extensible design: semantic abstractions of numeric values, pointers, objects, control, as well as syntax-driven iterators, are defined in small, reusable domains with loose coupling, that can be combined and reused to a greater extent than in previous work. Moreover, Mopsa aims at supporting several languages (currently, subsets of both C and Python) while sharing abstraction components as much as possible. Mopsa is a work in progress, and not yet capable of analyzing full programs; nevertheless, we report early experimental results on verification benchmarks.
Document type :
Conference papers
Liste complète des métadonnées

https://hal.sorbonne-universite.fr/hal-01870001
Contributor : Abdelraouf Ouadjaout <>
Submitted on : Friday, September 7, 2018 - 9:58:06 AM
Last modification on : Wednesday, March 27, 2019 - 1:34:22 AM
Document(s) archivé(s) le : Saturday, December 8, 2018 - 1:05:06 PM

File

mine-al-tapas18.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01870001, version 1

Citation

Antoine Miné, Abdelraouf Ouadjaout, Matthieu Journault. Design of a Modular Platform for Static Analysis. The Ninth Workshop on Tools for Automatic Program Analysis (TAPAS'18), Aug 2018, Fribourg-en-Brisgau, Germany. 〈hal-01870001〉

Share

Metrics

Record views

79

Files downloads

51