srcSlice: very efficient and scalable forward static slicing
Hakam W. Alomari, Michael L. Collard, Jonathan I. Maletic, Nouh Alhindawi, Omar Meqdadi
Abstract
Hakam W. Alomari, Michael L. Collard, Jonathan I. Maletic, Nouh Alhindawi, Omar Meqdadi
Abstract
ABSTRACT A highly efficient lightweight forward static slicing approach is presented and evaluated. The approach does not compute the program/system dependence graph but instead dependence and control information is computed as needed while computing the slice on a variable. The result is a list of line numbers, dependent variables, aliases, and function calls that are part of the slice for all variables (both local and global) for the entire system. The method is implemented as a tool, calledsrcSlice, on top ofsrcML, an XML representation of source code. The approach is highly scalable and can generate the slices for all variables of the Linux kernel in approximately 20 min on a typical desktop. Benchmark results are compared with theCodeSurferslicing tool from GrammaTech Inc., and the approach compares well with regard to accuracy of slices. Copyright © 2014 John Wiley & Sons, Ltd.
OpenAlex reports 36 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
ABSTRACT A highly efficient lightweight forward static slicing approach is presented and evaluated. The approach does not compute the program/system dependence graph but instead dependence and control information is computed as needed while computing the slice on a variable. The result is a list of line numbers, dependent variables, aliases, and function calls that are part of the slice for all variables (both local and global) for the entire system. The method is implemented as a tool, calledsrcSlice, on top ofsrcML, an XML representation of source code. The approach is highly scalable and can generate the slices for all variables of the Linux kernel in approximately 20 min on a typical desktop. Benchmark results are compared with theCodeSurferslicing tool from GrammaTech Inc., and the approach compares well with regard to accuracy of slices. Copyright © 2014 John Wiley & Sons, Ltd.
Key concepts: Program slicing, Slicing, Computer science, Scalability, XML, Benchmark (surveying), Parallel computing, Kernel (algebra)