AUTOMATIC RISK ANALYSIS OF SOFTWARE

摘要

Techniques are described herein that are capable of performing automatic risk analysis of software. An automatic risk analyzer may determine correlations between binaries of the software and source files of the software, between the source files and functions (i.e., sub-routines) of the software, between changes to source code of the software and defects, new features, and/or feature enhancements in the software, between attributes of the software (e.g., attributes associated with the changes to the source code), etc. at various time instances. The automatic risk analyzer may compare the correlations corresponding to a first time instance and the correlations corresponding to a second time instance to determine any of a variety of temporal relationships regarding underlying factors, including but not limited to new feature(s), feature enhancement(s), defect(s), churn rate, code coverage, code complexity, etc. The automatic risk analyzer may determine risk(s) for the software based on the temporal relationships.

主权项

1. A method comprising:
correlating a plurality of source code files of a software program to a plurality of binaries that are included in a compiled representation of the software program to provide a plurality of first correlations; for each time period of a plurality of time periods, mapping one or more changes that occur with respect to one or more source code files of the plurality of source code files during the respective time period to one or more portions of the compiled representation of the software program that represent at least one risk for the software program during execution to provide one or more mappings based at least in part on the plurality of first correlations; determining one or more attributes of the software program; for each time period of the plurality of time periods, determining one or more intermediate relationships between the one or more mappings for that time period and the one or more attributes; correlating the one or more intermediate relationships that correspond to a first time period of the plurality of time periods to the one or more intermediate relationships that correspond to a second time period of the plurality of time periods to provide one or more temporal relationships; and determining one or more likelihoods that one or more respective subsets of a plurality of factors are to interfere with execution of the software program based at least in part on the one or more temporal relationships.