You should put version control as much as possible. The pinnacle of that which does not change has zero value, understanding later that you cannot restore the state you need is fatal. Storage is very cheap, text files are very small, most SVM-s deadweight, which are small in size and speed.
Agile practices also explicitly include this. In addition to the source code and project files, consider building, deploying scripts, database schemas and layouts, configurations, documentation (!), Test files, dependent libraries, todos, the project website ... Of course, sometimes it can be difficult to place a virtual image machines running version (binary), conventional snapshots may be more applicable for this. But if you ever ask yourself, I have to put this under version control, the answer is almost always yes, you should. (Following this practice is also easier to make version control decisions.)
sibidiba
source share