Prysm: Update weak subjectivity checkpoint epoch calculation

[terencechain]

In mainnet config, SafetyDecay is defined is 1/10 which is 0 for uint64. The spec uses 10% for float which we would love to avoid at all cost. Working with Aditya on this

https://github.com/prysmaticlabs/prysm/blob/master/shared/params/mainnet_config.go#L135

Thanks @prestonvanloon for pointing this out

[terencechain]

As it turns out the calculation of the weak subjectivity period in the hackmd is not accurate since it doesn't account for inactivity leaks, withdrawal delays, etc. Specifying an accurate equation is underway & @adiasg will share as soon as he has something

https://notes.ethereum.org/@adiasg/weak-subjectvity-eth2