WIP: Permission Claims - Voting Mechanics [skip-ci] #1390

rackstar · 2025-05-19T08:38:26Z

Description

Closes: #1387

Describe the changes made in this PR. A link to the issue is enough if it contains all the relevant information.

Testing

Explain how you tested your changes to ensure they work as expected.

Checklist

Performed a self-review of my own code
Made corresponding changes to the documentation

rackstar · 2025-05-19T13:33:51Z

contracts/modules/assessment/PermissionedAssessment.sol

+    // Get the tally again after the vote
+    (acceptCount, denyCount) = _getVoteTally(claimId, assessorGroup);


we can avoid calling _getVoteTally a second time by manually adjusting the vote counts:

if its an update

remove the previous vote from the tally

increment the tally with the current vote

if its a new vote

increment the tally with the current vote

but when I implemented it, it looked less readable. so opted calling _getVotedTally a second time for better readability (but small additional cost in gas)

But _getVotedTally seems heavy in gas as it goes through the whole enumerable set, then check maps for votes. I thing we should try to not call it at all when casting regular votes (left comment below about that, one option is to split castVote in few functions as we need total vote count only in special cases).

…aram

duxxud · 2025-05-21T11:46:36Z

contracts/modules/assessment/PermissionedAssessment.sol

+
+  /* ========== STATE VARIABLES ========== */
+
+  mapping(uint32 assessorGroupId => EnumerableSet.UintSet assessorGroup) private _assessorGroups;


I would keep the shorter name, just to be groups as before, and there can be a short comment to explain? This seems like too much of word assessor everywhere so it's less readable

Also I think we can omit naming value in the map if it is just a singular of the variable name

duxxud · 2025-05-21T11:51:15Z

contracts/modules/assessment/PermissionedAssessment.sol

+    } else {
+      // Otherwise, check if we need to extend the voting period
+      uint32 nextDay = uint32(block.timestamp + SILENT_ENDING_PERIOD);
+      // If the poll ends in less than 24h from the latest vote, extend it to 24h


there shouldn't be any voting extensions.

duxxud · 2025-05-21T12:00:27Z

contracts/modules/assessment/PermissionedAssessment.sol

+        else if (vote == Vote.DENY) denyCount++;
+
+        // Unchecked increment to save gas - cannot overflow as assessor group size is a relatively small number
+        unchecked { ++i; }


this looks like a micro optimization, not needed here.

duxxud · 2025-05-21T12:01:20Z

contracts/modules/assessment/PermissionedAssessment.sol

+  /// @param denyCount Number of deny votes
+  /// @param assessmentEnd Timestamp when the assessment voting period ends
+  /// @return true if the assessment is decided, false otherwise
+  function _isAssessmentDecided(uint256 acceptCount, uint256 denyCount, uint32 assessmentEnd) internal view returns (bool) {


seems like this can be a part of return of _getVoteTally function, not needed as a separate one?

duxxud · 2025-05-21T12:11:14Z

contracts/modules/assessment/PermissionedAssessment.sol

+    denyCount = 0;
+
+    for (uint i = 0; i < assessorGroupLength;) {
+        uint256 assessorMemberId = assessorGroup.at(i);


I think it may be faster to get first the whole array to memory by using values(), and then iterate through the memory array. But we can test this later also (as I'm writing this, I'm also thinking on optimizing the whole EnumberableSet because now our memberIds will be all uint32 and we don't need to use the whole uint256 slot for each of it).

yes, I did consider the memory approach as well but since we are accessing each item only once it should be the same gas cost. so went with storage method as it seemed more readable.

and yes, might be worth doing our own custom EnumberableSet that supports uint32. we should go with the memory approach if that is the case

contracts/modules/assessment/PermissionedAssessment.sol

duxxud · 2025-05-21T12:18:41Z

contracts/modules/assessment/PermissionedAssessment.sol

+  /// @param productTypeId Type of product the claim is for
+  /// @dev Only callable by internal contracts
+  /// @dev Reverts if an assessment already exists for the given claimId
+  function startAssessment(bytes32 claimId, uint16 productTypeId) external onlyInternal {


This interface changes how it works from Claims, as we had before assessmentId which was returned and saved in Claims. Now it's more coupled, maybe change is for better, but requires a lot of changes in Claims also then.

duxxud · 2025-05-21T12:35:41Z

contracts/interfaces/IPermissionedAssessment.sol

+  struct Ballot {
+    bytes32 ipfsHash;
+    Vote vote;
+    uint32 timestamp;


do we need this timestamp onchain?

yes, we could probably get the timestamp from the event but since it also packs into slot 1, I've added it in for convenience.

duxxud · 2025-05-21T12:50:12Z

contracts/modules/assessment/PermissionedAssessment.sol

+    EnumerableSet.UintSet storage assessorGroup = _assessorGroups[assessment.assessorGroupId];
+    uint256 assessorGroupLength = assessorGroup.length();
+
+    // Only allow voting if the poll is not yet decided (no votes, a draw or voting period hasn't ended)


It may be better to have both 'early closing' and 'draw decider' as separate functions, and then only check here if voting period ended and keep castVote very simple.

duxxud · 2025-05-22T09:32:17Z

contracts/interfaces/IPermissionedAssessment.sol

+    uint32 end;
+    uint32 assessorGroupId;
+    uint32 finalizedAt; // 0, not closed yet else timestamp of closure
+    AssessmentResult result; // decision snapshot


if we have finalizedAt, this can be just bool accepted?

duxxud · 2025-05-22T09:34:47Z

contracts/modules/assessment/PermissionedAssessment.sol

+    emit VoteCast(claimId, msg.sender, assessorMemberId, vote, ipfsHash);
+  }
+
+  function closeAssessment(uint256 claimId) external {


should it revert in all cases when not closeable?

duxxud · 2025-05-22T09:38:02Z

contracts/interfaces/IPermissionedAssessment.sol

+
+  struct Assessment {
+    uint32 start;
+    uint32 end;


we probably also don't need end anymore, because it shouldn't be changeable and can be derived as start + VOTING_PERIOD

rackstar self-assigned this May 19, 2025

rackstar changed the title ~~WIP: Permission Claims - Voting Mechanics~~ WIP: Permission Claims - Voting Mechanics [skip-ci] May 19, 2025

rackstar force-pushed the feat/permissioned-assessment-voting-mechanics branch from 092a03f to 5f132e8 Compare May 19, 2025 08:43

rackstar marked this pull request as draft May 19, 2025 08:43

rackstar force-pushed the feat/permissioned-assessment-voting-mechanics branch 7 times, most recently from 23adb13 to 0351136 Compare May 19, 2025 10:18

feat: add PermissionedAssessment voting mechanics --WIP-- [skip-ci]

9fbba73

rackstar force-pushed the feat/permissioned-assessment-voting-mechanics branch from 0351136 to 9fbba73 Compare May 19, 2025 13:27

rackstar commented May 19, 2025

View reviewed changes

rackstar force-pushed the feat/permissioned-assessment-voting-mechanics branch 2 times, most recently from 5b084ac to 787a114 Compare May 19, 2025 14:16

feat: add validation when reading mapping values

4d86c71

rackstar force-pushed the feat/permissioned-assessment-voting-mechanics branch from 787a114 to 4d86c71 Compare May 19, 2025 14:51

rackstar added 2 commits May 19, 2025 18:01

refactor: modify _getTally to have assessment instead of claimId as p…

87bd2ed

…aram

perf: reduce SLOAD for gas optimisation

1c692e9

rackstar force-pushed the feat/permissioned-assessment-voting-mechanics branch from f0d4421 to 1c692e9 Compare May 19, 2025 15:46

rackstar added 4 commits May 19, 2025 19:21

refactor: simplify _isAssessmentDecided params

42fc4d7

refactor: optimise Ballot struct slot packing

255c2a5

refactor: remove revert on some of the validations

4c68543

feat: add assessor address to VoteCast event

eacfa1d

duxxud reviewed May 21, 2025

View reviewed changes

rackstar added 3 commits May 22, 2025 10:45

fix: change claimId to uint256

9b7e236

feat: add finalizedAt and result fields to Assessment struct

8598739

feat: add closeAssessment and simplify castVote function

8bfcaba

duxxud reviewed May 22, 2025

View reviewed changes

rackstar added 3 commits May 23, 2025 15:31

feat: add snapshot acceptVotes and denyVotes on Assessment struct

8192cb9

feat: prevent vote cast if voting period ends and not a draw

a02db5a

feat: add isReadyToCloseAfterVote helper function

c2901d4

		// Get the tally again after the vote
		(acceptCount, denyCount) = _getVoteTally(claimId, assessorGroup);


		/* ========== STATE VARIABLES ========== */

		mapping(uint32 assessorGroupId => EnumerableSet.UintSet assessorGroup) private _assessorGroups;

WIP: Permission Claims - Voting Mechanics [skip-ci] #1390

Are you sure you want to change the base?

WIP: Permission Claims - Voting Mechanics [skip-ci] #1390

Uh oh!

Conversation

rackstar commented May 19, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Testing

Checklist

Uh oh!

rackstar May 19, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

rackstar commented May 19, 2025 •

edited

Loading

rackstar May 19, 2025 •

edited

Loading