Staking OCW should not create solutions that are too big

[kianenigma]

Currently, the network size is such that the solutions to the NPoS easily fit into a block, both in terms of weight and size (TODO: script to report these easily). Sooner or later, we might reach a point where the solution might not fit into a block. For this, we have permanent solutions:

1. limit the total number of nominators
2. DNPoS

But one way or another, if we don't have these limitations, the OCW must be capable of generating a solution that is sane and fits in the block.

The danger is that a malicious actor might generate a lot of nominators, so that the OCW will always generate a solution that is too big to fit. Then, they can submit a mediocre solution on chain which will be accepted. Sadly, we won't fallback to on-chain phragmen either.

This can also be mitigated by implementing the PJR-check #7009.

[kianenigma]

indeed the numbers don't look so good.

 Staking election closed at 0x2f6a864e45f19926f7007ce87a9ad099d72f1a5eb19108cde4754e9971ae5d70 (1489532) with compute Authority
✅ Found a correct submit_election_solution_unsigned for era 97 => score [25000481475547417, 5872904777311892206, 177816180753394825113507963012696032]
⌚️ Weight = {"normal":845352150000,"operational":0} (0.422676075). Len = 29926 (0.005707931518554687)
🤑 Staking election closed at 0xae81482a6af5d08548c5509e1e1a65069ab6d8c8c2388443a855bad6e8f7f58d (1475154) with compute Authority
✅ Found a correct submit_election_solution_unsigned for era 96 => score [24997769671168078, 5866384217728530269, 177542812156025309450630273681874439]
⌚️ Weight = {"normal":822277790000,"operational":0} (0.411138895). Len = 29092 (0.005548858642578125)
🤑 Staking election closed at 0xbcda72530fd4175e0eb7758863d08f803b2d99ab3da7af2ccbc1cc51a9c44a2f (1460778) with compute Authority
✅ Found a correct submit_election_solution_unsigned for era 95 => score [24997769671168078, 5863414995248924815, 177455339295779061904389502061269235]
⌚️ Weight = {"normal":812748350000,"operational":0} (0.406374175). Len = 28770 (0.0054874420166015625)
🤑 Staking election closed at 0xb406be670736d72e4463a5dea39eda4cb4ed2f2e20bc028050faacadefd5eac9 (1446405) with compute Authority
✅ Found a correct submit_election_solution_unsigned for era 94 => score [24997769671168078, 5852675987488101749, 176571312187658874785317838860683927]
⌚️ Weight = {"normal":790574520000,"operational":0} (0.39528726). Len = 27988 (0.005338287353515625)
✅ Found a correct submit_election_solution_unsigned for era 94 => score [24997769671168078, 5852675987488101734, 176698110589451241890854882825526804]
⌚️ Weight = {"normal":790169520000,"operational":0} (0.39508476). Len = 27988 (0.005338287353515625)
🤑 Staking election closed at 0x7580bf1f3d72ed0cf7619d78edf4f3ed7625715d283e045131395ef3b8c8eb8f (1432032) with compute Authority
✅ Found a correct submit_election_solution_unsigned for era 93 => score [24997769671168078, 5866038993804244583, 177584523397207320526973676763852425]
⌚️ Weight = {"normal":779721550000,"operational":0} (0.389860775). Len = 27638 (0.005271530151367188)
🤑 Staking election closed at 0xbb67e5e696478d75ecbf50503a6c7a7fc40b13af49af97f9998a75ee0595e829 (1417661) with compute Authority
✅ Found a correct submit_election_solution_unsigned for era 92 => score [24997844671168078, 5845856672974488538, 176370876332636841837159090610554660]
⌚️ Weight = {"normal":754999800000,"operational":0} (0.3774999). Len = 26764 (0.005104827880859375)
🤑 Staking election closed at 0xe845406ddb329d79b536b5e463d0154429d3283ebc6bbf8d70809a0cc7a08a98 (1403288) with compute Authority
✅ Found a correct submit_election_solution_unsigned for era 91 => score [24997844671168078, 5803735389327106897, 173498808704042621035612220506441491]
⌚️ Weight = {"normal":733272730000,"operational":0} (0.366636365). Len = 26012 (0.004961395263671875)
🤑 Staking election closed at 0x771f7f5a979905271ded8e934dd682eb182d84f4a275c6a163d95145d21e406f (1388916) with compute Authority
✅ Found a correct submit_election_solution_unsigned for era 90 => score [24724162758068208, 5777134167239130414, 171849679471658644726809718241071848]
⌚️ Weight = {"normal":706667190000,"operational":0} (0.353333595). Len = 25064 (0.00478057861328125)
✅ Found a correct submit_election_solution_unsigned for era 90 => score [24724162758068208, 5777134167239130414, 171977199150919428644954835067107668]
⌚️ Weight = {"normal":706667190000,"operational":0} (0.353333595). Len = 25072 (0.0047821044921875)

[kianenigma]

Okay, I've done a bit more thinking about this, and I am not sure of it is worth fixing.

1. Say an OCW computes a large solution. The process to reduce the size of the solution is undefined. But regardless, whatever we do, we'll end up reducing the edge backing stakes of the winners and thus the solution will surely have a worse score than it initially had. Moreover, once we have PJR-test, this will almost surely cause to not be PJR. Thus, this whole effort is moot.

2. Once we have two phase submission, this will be mitigated entirely. With two phase submission, the results will we stored with a deposit, so the entire weight of this call will be removed.