08:26:10 <m-relay> <r​brunner7:monero.social> I am still having fun with the central method to come up with new peers to connect to, `bool node_server<t_payload_net_handler>::make_new_connection_from_peerlist`
08:27:02 <m-relay> <r​brunner7:monero.social> It has an outer try loop that looks like it originally tried to find a new peer up to 10 times, with counting in `try_count`
08:27:51 <m-relay> <r​brunner7:monero.social> There is second counter, `rand_count`, once limited to the number of available known white or gray peers
08:30:12 <m-relay> <r​brunner7:monero.social> It seems to me like something funny happened with moneromooo's epochal prune commit 7 years ago. Setting `max_random_index` to 0 has the probably unintended side effect that now it tries only up to 3 times, not 10 times.
08:30:40 <m-relay> <r​brunner7:monero.social> This is the file with the method in question in that commit: https://github.com/moneromooo-monero/bitmonero/blob/b750fb27b0f20e9443827732b69a504a76036430/src/p2p/net_node.inl
08:31:56 <m-relay> <r​brunner7:monero.social> I think I will go with "don't fix if it's not broken" and in my full rewrite of that method will also try only up to 3 times.
08:33:30 <m-relay> <r​brunner7:monero.social> The code line with the "funny" conditions where the mentioned outer loop starts is this one: https://github.com/moneromooo-monero/bitmonero/blob/b750fb27b0f20e9443827732b69a504a76036430/src/p2p/net_node.inl#L1126