Solve the problem of data inconsistency in the collator upon restart #972
Description
When starting a node and attempting to synchronize while restoring the message queue, a situation arises where the diff is committed to queue number N, but there is no state for this block at that moment.
tycho/collator/src/manager/mod.rs
Line 3407 in a111cf4
Because of this, the queue is restored from an earlier block, and a panic occurs when apply_diff to a diff that is below the commit pointer