feat: Use runtime loops for brillig array initialization

[sirasistant]

Description

Problem*

Brillig bytecode size can blowup if your function needs to create a big constant array. This is because the array is initialized with compile-time issued instructions (N STORE instructions where N is the number of items in the array)

Summary*

This PR will initialize the array with a runtime loop if it's bigger than a minimum size, all the elements are equal and contains no nested arrays.

Additional Context

This optimization improves bytecode size for some aztec public functions affected by this kind of blowup:

Transpiling function _approve_bridge_and_exit_input_asset_to_L1 on contract Uniswap with size 3165 => 2427
Transpiling function claim_public on contract GasToken with size 3415 => 2395
Transpiling function claim_public on contract TokenBridge with size 3457 => 2437
Transpiling function consume_message_from_arbitrary_sender_public on contract Test with size 1893 => 1135
Transpiling function consume_mint_public_message on contract Test with size 2600 => 1580
Transpiling function emit_unencrypted on contract Test with size 343 => 301
Transpiling function emit_unencrypted_log on contract AvmTest with size 930 => 774

Documentation*

Check one:

• No documentation needed.
• Documentation included in this PR.
• [For Experimental Features] Documentation to be submitted in a separate PR.

PR Checklist*

• I have tested the changes locally.
• I have formatted the changes with Prettier and/or cargo fmt on default settings.

[sirasistant]

Tested here AztecProtocol/aztec-packages#7044