Gas Optimizations: PR #1 (closed #209 to include changes in this one)

[Antony-SS]

Changes are limited to Verifier.sol, Pairing.sol, MiMC.sol, and MerkleTree.sol

I took advantage of the following methods for gas saving:

• ++i instead of i++ everywhere in all files. Pre-increment is cheaper because it does not store a copy of i's value.
• Replaced memory with calldata where applicable in Verifier.sol. Calldata is an immutable, temporary location where calldata can be stored that behaves similarly to memory. It stops unecessary copies being made, thus saving gas.
• Marking variables constants where possible. Constants are turned to bytecode at compile time and save the costly sload operation when accessed from that point forwards. All changes to do with this are in MerkleTree.sol
• Caching storage variables that are accessed more than once in a given function. Accessing storage variables is costly. It is better to save them into memory if they are going to be used more than once. All changes to do with this are in the MerkleTree.sol file
• Adding indexed to event with value types. Small trick that is slightly more gas efficient
• Adding unchecked to increment variables in for loops where we know they won't overflow. This saves the EVM from having to do overflow checks everytime it increments.

For specifics on the gas savings see this Notion Doc: https://antony-ss.notion.site/Antony-s-Starlight-Gas-Investigation-PR-1-4ee119fe89d24e56b41f3cc87d9b686d?pvs=4

There are some changes to the .gitignore b/c I added a test-contracts folder with all the contracts I was testing. You can choose whether or not to include this in the PR.

Addtionally, in order to get Starlight to run on my computer I had to update my "https://github.com/types" dependency, which is reflected in the changes. I would reccomend including this change.

[Antony-SS]

Reject the change to cast to a uint64. Accidentally left this in from when I was testing treeWidth as a uint64. Make sure to keep it as a constant though.