METHOD FOR DETECTING SOFTWARE ERRORS AND VULNERABILITIES

Patent №

US 7,577,889

Granted

2009-08-18

Filed 2007

Owner

AT&T CORP.

+2 more

Lab

AI components

1

kr

Assignment

Recorded

Dataset

AIPD

2023_r1 edition

Application

11873125

The present embodiments provide methods for detecting errors and vulnerabilities in software without access to its source code. The method entails extracting functions from dynamically linked applications, determining error return codes and error numbers for the extracted functions, and forming a wrapper using the extracted functions. The wrapper is used to intercept some of the function calls, wherein the intercepted functions are manipulated to fail. The embodiments further relate to an article of manufacture to detect errors. The article of manufacture utilizes an extraction module to extract functions from dynamically linked applications, software to compile error return codes and error numbers for functions, and a wrapper module to form a wrapper with the extracted functions.

AI classification

Knowledge representation0.80
Natural language0.09
Planning0.04
AI hardware0.03
Machine learning0.02
Vision0.01
Evolutionary computation0.00
Speech0.00

Ownership

AT&T CORP.

assignment · 199700564

AT&T PROPERTIES, LLC

assignment · 226860939

AT&T INTELLECTUAL PROPERTY II, L.P.

assignment · 226910702

Assignors

FETZER, CHRISTOF, XIAO, ZHEN, XIAO, ZHEN, FETZER, CHRISTOF

On an employer assignment, the assignors are typically the inventors.

© 2026 NYSGPT2525 LLC