[NIT-2489] Honest Strategy Revamp to Consider Path Weights

[rauljordan]

Fixes: NIT-2489

Key reading:

• https://www.notion.so/arbitrum/Reimbursing-Mini-Stakes-In-Band-in-BoLD-f7d27046d6304c2ab6933a152b83057d

Background:
This PR changes the honest strategy to match the expected specification linked above. The core idea is that edge trackers should only attempt to bisect or open subchallenges if the weight of timers to their closest root ancestor is < 1 challenge period.

This means we also need to change the rules for when an edge tracker should "despawn". The changes are as follows:

If the edge is an essential root => despawn once it is confirmed.
else if not an essential root  => 
	if edge is small step and length one => despawn once confirmed by OSP
	else: despawn once the closest essential ancestor root is confirmed

Terminology remapping:

• Path weight: the sum of the local timers along a path of edges up to some root edge
• Proof node: any lowest level subchallenge edge of length 1

This PR also adds an e2e test that checks that all essential root edges are confirmed, beyond just the challenge completing.

[eljobe]

In the description of the PR, Raul suggests that we should also be checking in the e2e_test.go that the essential root edges are confirmed in addition to winning the challenge, can you take a stab at that as well?

Mostly, this PR is looking really good though. I especially like the simplification of getting rid of the min heap.

[rauljordan]

Note @eljobe , the big diff is due to updating MODULE.bazel.lock. The real diff is still around +1000

[eljobe]

LGTM