After releasing OXID eShop 4.1.4, Our development team wrote an updated
quality report. Let’s go through some important points from the report, which are clearer now that we have a longer time frame to look at.
First of all, we can see a small rise over time in cyclomatic code complexity and code convention errors diagrams, together with a small dip in the code coverage diagram in the same time frame. From January until April the quality suffered slightly. Coincidentally, most of the OXID eShop 4.1 features were developed in this same time frame. Of course this isn’t actually a coincidence: During feature development time new code comes into the product which has not yet matured fully.
Actually, these metrics showed us early on which parts of the shop we needed to pay more attention to. You can see the result clearly in the diagrams;
you see the code complexity and convention errors drop again and coverage rises again. We have reached a new high in Unit Test code coverage: 93.34%.
The longer we keep track of quality like this, the more we learn about our own development process and its quality. We now get a sense of the quality even before running the metrics statistics. My guess is that during the main feature development time frame for 4.2 we will lose one or two percentage points of code coverage again, and they will rise again during the last iterations (remember we do XP) of development for 4.2 and first maintenance releases. Let’s check that out when the next quality report is available, shall we?
I’m glad to see that community contributors are helping to improve quality. The number of reported bugs grew to over 100. Thus the team worked mostly on fixing bugs. During this period the total number of lines of eShop source code remained approximately the same. Of course, the amount of source code from Unit Tests kept growing. Now we’re close to 4,000 unit tests. Just 52 more until we can celebrate the 4,000th test!
For more details, please download the full OXID eShop Quality Report.