Outdated: Large constants via MLOAD.

[DanielVF]

Description

This PR makes an approximately 5% bytecode size reduction for contracts compiled with the normal defaults of optimizer + runs 200 vs without this PR.

There are also gas savings as well. (note that the total gas used in these transactions includes other contracts, these are real mainnet contracts, with just one selected contract swapped out)

(Some benchmarking results available here: https://danielvf.github.io/purplebench/site/compilers/solc-0.8.36-memorymasking-03.html)

Bitmasking for stack cleanup currently makes up a double digit portion of solidity code (see https://ethereum-magicians.org/t/do-we-need-a-bitmasking-opcode-23-3-of-aave-v4-core-bytecode-is-bitmasking/28323). The current methods used by solidity either use up a lot of bytecode for raw push bytes, or use 9 bytes, 18 gas for computing a large constant.

By adding a 32 bytes of 0xFF after the zero constant memory slot, we can create any byte aligned, right aligned constant by just selecting a memory region that contains the right amounts of 0x00's and 0xFF's.

000000000000FFFFFFFFFFFF
        |----------|
     |----------|

This makes the cost of most large constants, not just for masks be reduced to 4 bytes, 9 gas, with a PUSH1 MLOAD AND

This PR is yolo code intended to demonstrate that the concept works. It is not intended for merging as is.

Setting optimizer runs to a high value forces using raw large constants and this no longer provides benefits.

Checklist

• I have read the contributing guidelines and the review checklist

AI Disclosure

• No AI tools were used
• AI tools were used (details below)

Code written with codex 5.5 on xhigh.

[DanielVF]

Note that I'll have a much better draft PR up today.

[DanielVF]

Closing this, will be posting a much cleaner one.