rmech
Introduction
Well you have seen the film, read the book and ignored the nightly generated reports. Static code review looks at the rather large amount of Sakai code 800,000 lines or so and performs checks for known bug types, if properly tuned this approach can work as a neutral and accurate form of feedback and a learning process. However, as my old pap used to say on numerous occasions for numerous reasons. "You can get a horse to water, but you can't make it drink". This process is only worth effort if it regularly used. Now that's the question.
Overview
Three reviewing tools and a couple of other utilities are run against the trunk of Sakai once an evening. The outputted information is inputted into a database from which a website is generated.
Motivation
- University of Amsterdam wishes to pilot OSP
- Sakai's code base is large, larger, growing
- Sakai has a reasonable number of sub projects and teams
- Moore's law implies that within a couple of years the process can sit on anyone's desktop and run promptly.
- Static code tools are improving
- Static code reviews are observer neutral (but, pattern dependent)
- The process may help with communication, bug patterns awareness, general quality of code, an understanding of the relative quality of sub project.
- Have almost taught the system to make a reasonable cup of decayed tea...
Issues
What issues?
The reports contain too much information, false positives and what the heck is that bug anyway.
Details
- WorkFlow
- Purpose of each script file
- Installation instructions
- The one Ant script to rule them all
- Internationalization dashboard details
Resources
Static analysis tools applied