Fix INACTIVITY_PENALTY_QUOTIENT

[hwwhww]

#1693 alternative

Issue

The amount of inactivity penalty was adjusted to half since we were applying penalty for missing FFG target and source. But now we only apply it for missing target, so INACTIVITY_PENALTY_QUOTIENT should have been 2**24.

How did I solve it

Update INACTIVITY_PENALTY_QUOTIENT from 2**25 to 2**24.

Note that the status-quo 2**25 doesn't break security:
#1693 (comment)

@vbuterin: Though it doesn't break security at all if it's kept at 2**25; it would just mean the leak takes 41% longer, so if this ends up being the only change we need to do then not bothering with it is fine too.

IMO this PR makes the current eth2.0-specs explainers (e.g., Serenity Design Rationale) stay consistent leaking rate. Also, we can avoid making INACTIVITY_PENALTY_QUOTIENT configuration explainer complicated in #1693.

It's not required for testnet, but it would be nice to include this configuration adjustment in the final v1.0.0 ™️ spec.

[djrtwo]

Maybe we should maintain a v0.12 branch off of dev that will incorporate the BLS IETF changes and anything minor like this. I'd like to keep dev (or another branch) ready to release v0.11.x releases that have updated network specs other backward compatible changes

[hwwhww]

@djrtwo sounds good! I added v012x branch for the next minor version upgrade.

[djrtwo]

great!