Improve 1346 to make it more efficient

https://github.com/crytic/slither/pull/1346 is a nice addition, which gives us fundamental features to do structure packing optimisations.

However it has a couple of drawback. For example packing storage variables might lead to more costly execution because of the additional opcode cost.

We could improve the detector to add a couple of heuristics. One example that comes to mind is to use https://eips.ethereum.org/EIPS/eip-2929#sstore-changes, and to pack together variables that are read within the same transaction.

We could make it a tool instead of a detector, and propose different scenario as a result, ie:
- To optimize the run of `contract.function1` pack together ..
- To optimize the run of `contract.function2` pack together ..

Having a tool would allow us additional fine tuning and flags.

Start by reviewing pull request 1346 and the EIP-2929 SSTORE changes referenced in the issue. Examine the existing storage-packing detector and decide whether the work should remain a detector or become a tool. Done means the analysis accounts for transaction-local reads and can report packing scenarios for contract functions, with suitable flags or heuristics.

Written by the indexing model from the issue text.