To identify errors in early stage of software development ensures lower price of development process and higher quality of the final software release. Automated testing procedures in EMI performed during ARC middleware development are done by two independent tool sets. First is the ETICS build system and second is the ARC automated build and functionality test system.
Description of the work
Automated builds enable us to track errors which could break integrity of the final code. For each code change committed to ARC subversion the tool checks status of following build steps: Autogen, Configure, Make, Make Check, Make Dist, and Build. The check is done for several platforms (including Mac OS X) and subtrees of ARC sources codes.
The functionality tests are periodically running carefully chosen test scenaria on ARC “nightlies” or on specific code revisions. The goal of these test scenaria is to check a basic functionality of ARC clients and services. During these tests we additionally check integration of Clients and Servers. The test scenaria, build and functionality tests results are stored in DB and so they can easily be visualized and exported in any form using simple php or python scripts. The request for new test scenario is done by a simple web form.
We present automatic tools which are used to detect build and functional errors in early stage of software development process.