fix: allow keccak256() and uint() in compile-time context for layout base expressions by JPier34 · Pull Request #16439 · argotorg/solidity

JPier34 · 2026-02-03T14:22:35Z

Summary

Implements evaluation of keccak256() and explicit uint() conversion in a compile-time context. This allows using keccak256() in storage layout base expressions:

contract C layout at uint(keccak256("my.contract.id")) {}

Fixes #16421. Implements Stage 1 of #16420.

Motivation

Layout base expressions commonly use either ERC-7201 (handled in PR #15968) or plain hashing with keccak256(). This change supports the latter. Since keccak256() returns bytes32 while the layout base must be an integer, we also evaluate explicit uint() conversions at compile time so that uint(keccak256("...")) is valid.

Changes

ConstantEvaluator.cpp/h
- Evaluate uint(...) type conversion at compile time when the target type is uint256.
- Evaluate builtin keccak256(...) at compile time when the argument is a string or hex literal.
Tests
- literal_cast.sol now expects success for uint(42).
- New test layout_keccak256_string_literal.sol for layout at uint(keccak256("...")).

Backwards compatibility

Fully backwards-compatible; no change to runtime behaviour.

…ressions

github-actions · 2026-02-03T14:22:46Z

Thank you for your contribution to the Solidity compiler! A team member will follow up shortly.

If you haven't read our contributing guidelines and our review checklist before, please do it now, this makes the reviewing process and accepting your contribution smoother.

If you have any questions or need our help, feel free to post them in the PR or talk to us directly on the #solidity-dev channel on Matrix.

Copilot

Pull request overview

This PR implements compile-time evaluation of keccak256() and explicit uint() type conversions, enabling storage layout base expressions like contract C layout at uint(keccak256("my.contract.id")) {}. This addresses issue #16421 and implements Stage 1 of issue #16420.

Changes:

Adds compile-time evaluation support for uint256 type conversions from rational values
Adds compile-time evaluation support for keccak256() with string and hex string literals
Updates test expectations to reflect the new compile-time capabilities

Reviewed changes

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

File	Description
libsolidity/analysis/ConstantEvaluator.h	Adds `endVisit` method declaration for `FunctionCall` nodes
libsolidity/analysis/ConstantEvaluator.cpp	Implements compile-time evaluation logic for `uint()` conversions and `keccak256()` calls, with proper range checking and error handling
test/libsolidity/syntaxTests/storageLayoutSpecifier/literal_cast.sol	Removes error expectation since `uint(42)` is now supported in compile-time context
test/libsolidity/syntaxTests/storageLayoutSpecifier/layout_keccak256_string_literal.sol	Adds new test case demonstrating the main use case: `uint(keccak256("..."))` in layout expressions

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

libsolidity/analysis/ConstantEvaluator.cpp

test/libsolidity/syntaxTests/storageLayoutSpecifier/layout_keccak256_string_literal.sol

libsolidity/analysis/ConstantEvaluator.cpp

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

matheusaaguiar

Hi @JPier34, Thank you for the initiative and the work you’ve put into this.
While I think it’s a solid starting point and it is in the right direction, I feel that the implementation currently aligns more with a proof of concept/prototype.
It is missing several tests and I encountered some issues like the conversion as the length of an array. I still need to check to make sure if that is a problem with the implementation or something to do with the order the AST nodes are type checked.

Additionally, I started wondering about support for byte conversions/string constants for keccak. Constant strings can now be evaluated as part of the introduction of builtin erc7201 in the last pre-release. This requires some design decisions by the team.

Considering that, we have decided that it’s best for us to complete this work internally, so I will be closing this PR.
We truly appreciate your effort and willingness to contribute, and we hope you’ll consider collaborating again in the future.
You can participate in the design discussions and get more involved (and hopefully more straightforward first tasks/issues) by attending our Solidity Language Design Calls. Check the agenda for day/time and topics.

matheusaaguiar · 2026-03-12T00:59:58Z

libsolidity/analysis/ConstantEvaluator.cpp

+{
+	// Stage 1 (#16420): explicit uint() in comptime (enables layout at uint(keccak256(...)) #16421).
+	if (_functionCall.annotation().kind.set() && *_functionCall.annotation().kind == FunctionCallKind::TypeConversion &&
+		_functionCall.arguments().size() == 1)


I think that conversions always have exactly one argument, so this would be better as an assertion inside the if block.

matheusaaguiar · 2026-03-12T01:05:19Z

libsolidity/analysis/ConstantEvaluator.cpp

+	if (_functionCall.annotation().kind.set() && *_functionCall.annotation().kind == FunctionCallKind::TypeConversion &&
+		_functionCall.arguments().size() == 1)
+	{
+		Type const* resultType = _functionCall.annotation().type;


Not necessarily wrong, but I am wondering whether functionType::returnParametersTypes() is more precise.

matheusaaguiar · 2026-03-12T01:08:32Z

libsolidity/analysis/ConstantEvaluator.cpp

+					);
+					return;
+				}
+				bigint intValue = value.numerator() / value.denominator();


Suggested change

bigint intValue = value.numerator() / value.denominator();

bigint intValue = value.numerator();

matheusaaguiar · 2026-03-12T01:16:01Z

libsolidity/analysis/ConstantEvaluator.cpp

+
+	// #16421: keccak256() in comptime (other common layout base beside erc7201 from PR 15968).
+	auto const* builtin = dynamic_cast<MagicVariableDeclaration const*>(ASTNode::referencedDeclaration(_functionCall.expression()));
+	if (!builtin || _functionCall.arguments().size() != 1)


Suggested change

if (!builtin || _functionCall.arguments().size() != 1)

if (!builtin)

Surely keccak256 gets exactly one argument, but I guess that at least for readability/clarity of purpose, best to check the kind of the function as done in the next lines.

matheusaaguiar · 2026-03-12T01:17:11Z

libsolidity/analysis/ConstantEvaluator.cpp

+	if (!builtin || _functionCall.arguments().size() != 1)
+		return;
+	auto const* functionType = builtin->functionType(true);
+	if (!functionType || functionType->kind() != FunctionType::Kind::KECCAK256)


Suggested change

if (!functionType || functionType->kind() != FunctionType::Kind::KECCAK256)

if (functionType->kind() != FunctionType::Kind::KECCAK256)

Should probably just assert functionType before.

matheusaaguiar · 2026-03-12T01:24:39Z

libsolidity/analysis/ConstantEvaluator.cpp

+
+	solAssert(functionType->returnParameterTypes().size() == 1, "");
+	Type const* bytes32Type = functionType->returnParameterTypes().front();
+	u256 const val = u256(util::h256::Arith(hash));


I think it would be better to keep the value as our internal h256.

Also this would probably require extending further TypedRational (which already was refactored into TypedValue in the last version).

test/libsolidity/syntaxTests/storageLayoutSpecifier/layout_keccak256_string_literal.sol

JPier34 · 2026-03-12T09:21:40Z

I see. Thank you for commenting and appreciating my effort! Look forward to cooperate with you again in the future!

Add keccak256() and uint() evaluation in comptime for layout base exp…

6c65501

…ressions

github-actions bot added the external contribution ⭐ label Feb 3, 2026

r0qs requested a review from Copilot February 16, 2026 15:01

Copilot started reviewing on behalf of r0qs February 16, 2026 15:01 View session

rodiazet assigned rodiazet and matheusaaguiar Feb 16, 2026

rodiazet requested a review from matheusaaguiar February 16, 2026 15:03

rodiazet unassigned rodiazet and matheusaaguiar Feb 16, 2026

Copilot AI reviewed Feb 16, 2026

View reviewed changes

JPier34 and others added 4 commits February 18, 2026 09:25

Update libsolidity/analysis/ConstantEvaluator.cpp

15d38a3

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

Apply suggestion from @Copilot

cffa3b7

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

Handle keccak256(empty hex literal) in compile-time evaluation

e28c2b9

Merge branch 'develop' into feature/keccak256-comptime-layout

d9c0dc9

matheusaaguiar reviewed Mar 12, 2026

View reviewed changes

matheusaaguiar closed this Mar 12, 2026

JPier34 deleted the feature/keccak256-comptime-layout branch March 12, 2026 09:21

msooseth mentioned this pull request Mar 24, 2026

fix: resolve isoltest crash on interactive update #16438

Open

	bigint intValue = value.numerator() / value.denominator();
	bigint intValue = value.numerator();

	if (!builtin \|\| _functionCall.arguments().size() != 1)
	if (!builtin)

	if (!functionType \|\| functionType->kind() != FunctionType::Kind::KECCAK256)
	if (functionType->kind() != FunctionType::Kind::KECCAK256)

Conversation

JPier34 commented Feb 3, 2026

Summary

Motivation

Changes

Backwards compatibility

Uh oh!

github-actions bot commented Feb 3, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

matheusaaguiar left a comment

Choose a reason for hiding this comment

Uh oh!

matheusaaguiar Mar 12, 2026

Choose a reason for hiding this comment

Uh oh!

matheusaaguiar Mar 12, 2026

Choose a reason for hiding this comment

Uh oh!

matheusaaguiar Mar 12, 2026

Choose a reason for hiding this comment

Uh oh!

matheusaaguiar Mar 12, 2026

Choose a reason for hiding this comment

Uh oh!

matheusaaguiar Mar 12, 2026

Choose a reason for hiding this comment

Uh oh!

matheusaaguiar Mar 12, 2026

Choose a reason for hiding this comment

Uh oh!

matheusaaguiar Mar 12, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

JPier34 commented Mar 12, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants