DocumentCode :
1885070
Title :
Design recovery and maintenance of build systems
Author :
Adams, Bram ; Tromp, Herman ; De Schutter, Kris ; De Meuter, Wolfgang
Author_Institution :
Ghent Univ., Ghent
fYear :
2007
fDate :
2-5 Oct. 2007
Firstpage :
114
Lastpage :
123
Abstract :
The build system forms an indispensable part of any software project. It needs to evolve in parallel with the source code in order to build, test and install the software. Unfortunately, little tool support exists to help maintainers gain insight into the build system, much less to refactor it. In this paper, we therefore present the design and implementation of a re(verse)-engineering framework for build systems named MAKAO. At its heart the framework makes the build´s dependency graph available in a tangible way. Aside from visualisation, this enables powerful querying of all build-related data, as well as various filtering techniques to define views on the build architecture. If desired, all this gathered information can be put to use to write aspects for refactoring the build. Afterwards, validation rules can help in assessing failure or success. We applied our implementation on an industrial C system and the Linux 2.6.16.18 kernel, with good results.
Keywords :
C language; Linux; data flow graphs; program testing; program verification; reverse engineering; software architecture; software maintenance; system recovery; systems analysis; Linux 2.6.16.18 kernel; build system architecture; build system design recovery; build system maintenance; build system refactoring; build-related data querying; dependency graph; filtering technique; industrial C system; reverse-engineering framework; software installation; software testing; system failure; validation rule; Data visualization; Filtering; Heart; Kernel; Linux; Software maintenance; Software systems; Software testing; Software tools; Technological innovation;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Software Maintenance, 2007. ICSM 2007. IEEE International Conference on
Conference_Location :
Paris
ISSN :
1063-6773
Print_ISBN :
978-1-4244-1256-3
Electronic_ISBN :
1063-6773
Type :
conf
DOI :
10.1109/ICSM.2007.4362624
Filename :
4362624
Link To Document :
بازگشت