Update the example dataset with more edge cases
The example dataset does not exercise lots of edge cases; mostly because it is very annoying to update it to add one edge case.
Here is the wishlist for the next big update
- dangling nodes
- multi-arcs
- snp->dir and snp->cnt edges
- dir->rev edges
- very high and low timestamps
- permissions
- partial visits
- non-UTF8 file names and branch names