[OT] The horizon of a stream

[bearophile]

This versions compares hash values first, avoiding some slower string compares, because in D1 strings don't store their hash value:

import d.all, d.string;

void main() {
    int[Record!(hash_t, string)] seen;
    int horizon;

    foreach (nline, line; xfile("data.txt")) {
        auto clean_line = line.chomp();
        auto hash_clean_line = hash(clean_line);
        auto pos_ptr = record(hash_clean_line, clean_line) in seen;
        if (pos_ptr) {
            if ((nline - *pos_ptr) >  horizon)
                 horizon = nline - *pos_ptr;
        } else
            seen[record(hash_clean_line, clean_line.dup)] = nline;
    }

    putr(horizon);
}

Maybe it's not fully correct yet, so I'll test it some more...

Bye,
bearophile