Run-Length Encoding
<count><char> pairs, demoing the p/moul/x/daily/rle library.
| input | encoded | size | round-trips |
|---|---|---|---|
aaaaaaaaaabbbbbbbbbb |
10a10b |
30% | ✅ |
aaabbc |
3a2b1c |
100% | ✅ |
abcdef |
1a1b1c1d1e1f |
200% | ✅ |
Over 100% means the encoding made the data bigger. RLE only wins on runny input, and
abcdefis the honest counter-example.
Append text to the path to encode it — digits are rejected, since they would be ambiguous with a run count.