Like find(1), with less features :)
Breadth-first listing work well with fuzzy files finders and similar tools (see
scripts/tmux-menu for an example).
I wanted to try to compare a dynamically shrinking zipper tree with a simple
queue and see if it could performs with a smaller memory usage on real workloads.
That's because I have the habit of exploring the whole $HOME directory and
often run lows on memory.
Try with:
./bench_mem <DIR>
Unfortunately, the expected memory gains are meager for the moment. The maximum residency is about the same as the queue implementation, with twice the running time. Collapsing straight branches could probably help the tree win!
Also, an excuse to play with QuickCheck and criterion!
- option to limit depth?