An inference and checking framework for context-sensitive pluggable types
Loading...
Authors
Huang, Wei
Issue Date
2014-05
Type
Electronic thesis
Thesis
Thesis
Language
ENG
Keywords
Computer science
Alternative Title
Abstract
This thesis presents several instantiations of interesting pluggable type systems: (1) a context-sensitive type system ReIm for reference immutability and an efficient quadratic type inference analysis, (2) the first effective type inference analysis for the classical Ownership Types, (3) a novel quadratic type inference analysis for Universe Types, (4) a context-sensitive type system SFlow/Integrity for detecting information flow vulnerabilities in Java web applications and a novel, worst-case cubic inference analysis, and (5) the dual type system SFlow/Confidentiality for detecting privacy leaks in Android apps and the corresponding inference analysis; the analysis scales well and detects leaks in apps from the Google Play Store and in known malware.
Description
May 2014
School of Science
School of Science
Full Citation
Publisher
Rensselaer Polytechnic Institute, Troy, NY