* greps for manpages with nix-locate * fetches the nar archives non-recursively * extracts manpages * converts to html This involved implementing a multiprocessing limit and lru file cache using bash...