[1.0-beta1] P2P: Fix stuck in peer syncing state

[heifner]

Test failure: https://github.com/AntelopeIO/spring/actions/runs/9114438126/job/25059003107?pr=149

The producing node (Node03) was stuck in LIB catchup. Since blocks can be processed inside controller during a fork and not call sync_master->sync_recv_block with blk_applied==true.

This is similar to the issue addressed in #27. In the #27, a handshake with the wrong information was sent. In this case no handshake was sent at all as it become synced up during a fork switch. I believe this issue was masked by #147 because it was incorrectly reporting blocks as applied. Without this improper reporting of applied blocks a fork switch that applies up to synced would not call sync_master->sync_recv_block with blk_applied==true and not move node out of LIB catchup.

Now also send out handshakes if we determine we are synced and there is a pending handshake to send.

Related to #27
Related to #147

Resolves #150

[ericpassmore]

Note:start
group: IF
category: BUG
summary: Add handshakes to cover additional P2P scenarios, including during a fork and when there are pending handshakes after a sync complete.
Note:end