Analyze milestones as soon as they arrive

[GalRogozinski]

Description

As soon as we hear on a milestone, add it to the milestone to be analyzed queue. If we hear of a new milestone candidate we analyze it right away. To avoid having to read it from the DB afterwards.

At the moment, we load all the milestones that we want to analyze from the DB.

Motivation

Save on DB reads and synchronize your node more quickly

Requirements

• We hear of a milestone candidate and we start analyzing it right away.
• The latestSolidMilestoneTracker keeps on solidifying according to index order.
• We bootstrap from a DB and we already have the milestones in the DB we bootstrap from, but we don't have the state - to make sure we read the milestones from the DB.
• We check milestone candidates for whether they are in the DB already (by index), and if we already validated them, we discard the TX.
• The list (queue) of milestone hashes waiting to solidify (needs to be capped at Y milestones.)
• The latestMilestoneTracker only needs a list of milestones, no need for a map. (Also capped at Y milestones.)
• We should maybe have a milestoneCandidate map in the milestoneSolidifier so that we can get the milestones in the right order.
• We perform simple syntactic checks - size of the bundle, index in obsoleteTag, and if fail we skip the signature validation
• We only store milestone hashes in the queues

Follow-up

Another follow up issue will be discarding the milestones that we have already loaded.

[GalRogozinski]

We need to pay attention to --rescan