Benchmark drop_assets of TrappistDropAssets

[bernardoaraujor]

follow up to #109

Motivation

drop_assets of TrappistDropAssets currently returns weight 0 to the Executor, which is innacurate.

it does a few storage checks (which are not weightless), then it calls drop_assets implementation of pallet-xcm (which is also still unbenchmarked and returns 0)
ref:

• Real weight for DropAssets. polkadot-sdk#954
• https://github.com/paritytech/polkadot/blob/88013730166ba90745ae7c9eb3e0c1be1513c7cc/xcm/pallet-xcm/src/lib.rs#L1308

so we need to benchmark everything inside for asset in multi_assets { ... } and add that to whatever pallet-xcm's drop_assets returns.

trappist/primitives/xcm/src/lib.rs

Lines 114 to 119 in 098634e

	// TODO: put real weight of execution up until this point here
	let mut weight = 0;
	if !trap.is_empty() {
	weight += XcmPallet::drop_assets(origin, trap.into());
	}

[stiiifff]

@pawanbisht62 Any ETA ?

[pawanbisht62]

Hi @stiiifff, as of now I'm working on bridges-task-force. I guess I won't be able to complete this next week either.

[pawanbisht62]

We cannot benchmark this as it is. We have to benchmark it in the pallet where TrappistDropAssets will be used.

Reason:

1. FRAME benchmarking only works for pallets.
2. TrappistDropAssets type is generic. So it cannot be generically benchmarked since the results depend on the concrete generic types

So can we close or hold this one for now?

cc: @stiiifff @bernardoaraujor