Remove SSTORE2 chunking now that we removed the contract size cap in geth

[RogerPodacter]

Note

Switches content storage to a single SSTORE2 pointer via SSTORE2Unlimited, removes chunking, and updates/expands tests to cover various sizes and deduplication.

• Contracts:
  • Storage Simplification: Replace chunked storage with SSTORE2Unlimited.
    • Ethscriptions.sol: switch contentPointersBySha: bytes32 => address[] to contentPointerBySha: bytes32 => address; update read/write to use SSTORE2Unlimited.read/write.
    • Remove libraries/SSTORE2ChunkedStorageLib.sol; add libraries/SSTORE2Unlimited.sol (single-contract, unlimited-size storage).
• Tests:
  • Update EthscriptionsWithTestFunctions.sol: remove chunk helpers; add hasContent, getContentPointer, readContent.
  • Rewrite EthscriptionsJson.t.sol to validate single-pointer storage and JSON outputs (no chunk assumptions); optimize loops.
  • Add SSTORE2ContentSizes.t.sol: comprehensive size tests (empty → 2MB), retrieval, and deduplication gas checks.
• Config:
  • foundry.toml: set gas_limit to 18446744073709551615.

^{Written by Cursor Bugbot for commit 356f2c7. This will update automatically on new commits. Configure here.}

[Copilot]

Pull Request Overview

This PR refactors the content storage mechanism from a chunked SSTORE2 approach to a single-contract SSTORE2Unlimited implementation that can handle arbitrarily large content (up to 4GB).

Key changes:

• Replaces the SSTORE2ChunkedStorageLib with a new SSTORE2Unlimited library that stores all content in a single contract deployment
• Simplifies the storage model from mapping(bytes32 => address[]) to mapping(bytes32 => address) for content pointers
• Adds comprehensive test suite for various content sizes from 0 bytes to 2MB

Reviewed Changes

Copilot reviewed 7 out of 7 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
contracts/src/libraries/SSTORE2Unlimited.sol	New library implementing unlimited-size SSTORE2 storage using a single contract deployment
contracts/src/libraries/SSTORE2ChunkedStorageLib.sol	Removed old chunked storage library
contracts/src/Ethscriptions.sol	Updated to use single pointer instead of array of pointers for content storage
contracts/test/EthscriptionsWithTestFunctions.sol	Updated test helper functions from chunk-based to single-pointer model
contracts/test/SSTORE2ContentSizes.t.sol	New comprehensive test suite for various content sizes
contracts/test/EthscriptionsJson.t.sol	Updated tests from chunk verification to single-pointer verification; improved loop efficiency
contracts/foundry.toml	Increased gas limit to maximum value for testing large content

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Potential issue when reading content with a zero address pointer. If contentPointerBySha[ethscription.contentSha] is address(0) (which can happen for empty content or if content was never stored), calling SSTORE2Unlimited.read(pointer) will attempt to read from address(0). In the read function at line 69, this will compute pointer.code.length - DATA_OFFSET which for address(0) becomes 0 - 1, causing an underflow. Consider adding a check: if (pointer == address(0)) return new bytes(0); before the read call.

Suggested change

	address pointer = contentPointerBySha[ethscription.contentSha];
	address pointer = contentPointerBySha[ethscription.contentSha];
	if (pointer == address(0)) {
	return new bytes(0);
	}

[Copilot]

Potential array out-of-bounds access for empty content. The _verifyContent function is called when size > 0 (line 142), but it doesn't guard against empty arrays. If an empty content somehow reaches this function, accessing index 0 or length - 1 will revert. While the caller checks size > 0, adding a size check at the start of this function would make it more defensive.

[Copilot]

Missing unchecked block for loop increment. Since i is bounded by size which is at most 1024 in this branch, and the loop is already incrementing with i++, wrapping the increment in an unchecked block would save gas and be consistent with the pattern used elsewhere in the codebase (e.g., lines 217-219, 437-439).

Suggested change

	for (uint i = 0; i < size; i++) {
	assertEq(retrievedContent[i], originalContent[i], "Byte mismatch");
	for (uint i = 0; i < size;) {
	assertEq(retrievedContent[i], originalContent[i], "Byte mismatch");
	unchecked { ++i; }

[Copilot]

Missing unchecked block for loop increment. The loop counter i is bounded by the fixed-size array length (5 elements), so overflow is impossible. Wrapping the increment in an unchecked block would save gas and be consistent with the pattern used elsewhere in the file (e.g., lines 217-219).

Suggested change

	for (uint256 i = 0; i < sizes.length; i++) {
	_testDeduplication(sizes[i], string.concat("Size: ", vm.toString(sizes[i])));
	for (uint256 i = 0; i < sizes.length;) {
	_testDeduplication(sizes[i], string.concat("Size: ", vm.toString(sizes[i])));
	unchecked { ++i; }

[cursor]

Bug: Overflow on zero address in SSTORE2Unlimited.read breaking empty content handling

The SSTORE2Unlimited.read function underflows when called with address(0), as pointer.code.length - DATA_OFFSET becomes 0 - 1. This causes getEthscriptionContent to revert for existing ethscriptions with empty content, because contentPointerBySha returns address(0) for these. This breaks backward compatibility, as the old system gracefully returned empty bytes for such cases.

Additional Locations (1)

• contracts/src/libraries/SSTORE2Unlimited.sol#L66-L70