Automated reasoning for reflective programs

Horsfall, Benjamin (2014) Automated reasoning for reflective programs. Doctoral thesis (PhD), University of Sussex.

[img]
Preview
PDF - Published Version
Download (1MB) | Preview

Abstract

Reflective programming allows one to construct programs that manipulate or examine their behaviour or structure at runtime. One of the benefits is the ability to create generic code that is able to adapt to being incorporated into different larger programs, without modifications to suit each concrete setting. Due to the runtime nature of reflection, static verification is difficult and has been largely ignored or only weakly supported. This work focusses on supporting verification for cases where generic code that uses reflection is to be used in a “closed” program where the structure of the program is known in advance.

This thesis first describes extensions to a verification system and semi-automated tool that was developed to reason about heap-manipulating programs which may store executable code on the heap. These extensions enable the tool to support a wider range of programs on account of the ability to provide stronger specifications. The system’s underlying logic is an extension of separation logic that includes nested Hoare-triples which describe behaviour of stored code. Using this verification tool, with the crucial enhancements in this work, a specified reflective library has been created.

The resulting work presents an approach where metadata is stored on the heap such that the reflective library can be implemented using primitive commands and then specified and verified, rather than developing new proof rules for the reflective operations. The supported
reflective functions characterise a subset of Java’s reflection library and the specifications guarantee both memory safety and a degree of functional correctness. To demonstrate the application of the developed solution two case studies are carried out, each of which focuses on different reflection features.

The contribution to knowledge is a first look at how to support semi-automated static verification of reflective programs with meaningful specifications.

Item Type: Thesis (Doctoral)
Schools and Departments: School of Engineering and Informatics > Informatics
Subjects: Q Science > QA Mathematics > QA0076 Computer software
Depositing User: Library Cataloguing
Date Deposited: 29 Oct 2014 06:54
Last Modified: 25 Sep 2015 12:09
URI: http://sro.sussex.ac.uk/id/eprint/49871

View download statistics for this item

📧 Request an update