Each repository has a "master hash table" that assigns hash-keys to buckets. Unfortunately, I didn't implement growing of that table yet (it's either slow or complicated), and for that reason the default size of that hash is 2MB, just to make sure that we don't produce too many clashes when a repository becomes really large (Like in all hash tables, it's a tradeoff efficiency vs. memory usage here).
<div><br></div><div>However, all the tables for 21 repositories combined have a constant size of 42MB, which is not really worrying I think. Instead of investing work here, it would be more interesting to find out what is producing the memory peaks (in the area of a gigabyte).</div>
<div><br></div><div>Greetings, David</div>