[Rodrigoue9]

Summary

What was wrong

Previously, VeroContract allowed callers with the TaskManager role to specify arbitrary, sparse task_id values when registering tasks. This exposed the protocol to task ID collisions, out-of-order IDs, sequence gaps, and potential ID resurrection attacks if a purged task ID were re-registered after removal from storage.

What was done & Alternatives Considered

This pull request resolves Issue #13 by introducing a strictly monotonic sequential counter stored in instance storage (DataKey::TaskCounter). Task registration now enforces that each incoming task_id strictly equals current_counter + 1. The counter starts at 0 (meaning the first registered task is ID 1) and advances strictly by 1 per task. The counter is preserved across task lifecycle transitions (including purge_task), permanently preventing ID reuse or resurrection.

• Alternative considered: Auto-generating and assigning task_id inside register_task without requiring caller to pass it. Rejected because altering the signature register_task(admin, task_id, min_votes) would break public ABI compatibility, existing BatchCall::RegisterTask enum variants, and external client integrations. Enforcing sequential IDs against current_counter + 1 satisfies the monotonic guarantee while preserving backward-compatible signatures.
• Alternative considered: Allowing gaps as long as task_id > last_id. Rejected because the issue acceptance criteria strictly specified "IDs increment by exactly 1" to ensure dense contiguous sequencing.

What to look at

• Storage permanence: Reviewers should verify that purge_task removes the task data but intentionally leaves DataKey::TaskCounter untouched, ensuring purged IDs cannot be revived.
• Batch atomicity: In register_tasks, sequence checks are performed up-front before the reentrancy lock, and state changes (including DataKey::TaskCounter) are applied atomically. A gap or out-of-order ID anywhere in a batch aborts the entire registration without updating the counter.
• Public getter: Added public entrypoint get_task_counter(env: Env) -> u64 to allow clients and frontends to determine the next expected task ID before submitting transactions.
• Test updates: Existing integration tests in archive_task_auth.rs, consensus_delegation.rs, integration.rs, and multi_contract_isolation.rs that previously used arbitrary mock IDs (e.g. 42, 100) were updated to use sequential IDs (1, 2, 3) to comply with the monotonic invariant.

How verified

• Added 10 dedicated integration tests in tests/monotonic_task_id.rs covering counter initialization, sequential increments, gap rejection, zero ID rejection, duplicate/lower ID rejection, full lifecycle transitions, purge non-reuse, multi-contract isolation, batch atomic rollback, and RBAC/pause validation.
• Added unit test test_register_tasks_batch_monotonic in src/task.rs.
• Ran full test suite across the workspace: 216 tests passed.
• Ran cargo fmt --all -- --check: 100% compliant.
• Ran cargo clippy --workspace --all-targets -- -D clippy::all: 0 warnings, 0 errors.

Related Issue

Closes #13

Changes Made

• Storage Layout (src/contracts/storage_layout.rs):
  • Added DataKey::TaskCounter variant in instance storage tracking the highest registered task sequence ID.
• Task Management (src/task.rs):
  • Implemented pub fn get_task_counter(env: &Env) -> u64 reading DataKey::TaskCounter (defaulting to 0).
  • Updated register_tasks to validate that incoming task IDs strictly increment by 1 from current_counter.
  • Atomically commits the updated counter under the reentrancy lock.
  • Added unit test test_register_tasks_batch_monotonic verifying sequential batches, internal gap rejection, and out-of-order rejection.
• Contract Entrypoint (src/contracts/proxy_entry/entry_tasks.rs):
  • Added documented public entrypoint pub fn get_task_counter(env: Env) -> u64.
• Integration Test Suite (tests/monotonic_task_id.rs):
  • Created 10 integration tests validating counter lifecycle, edge cases, purge non-reuse, multi-contract isolation, and rollback semantics.
• Existing Test Mock IDs:
  • Updated mock task IDs in tests/archive_task_auth.rs, tests/consensus_delegation.rs, tests/integration.rs, and tests/multi_contract_isolation.rs to start sequentially from ID 1.

Testing Performed

• Ran cargo test --test monotonic_task_id: 10/10 tests passed.
• Ran cargo test --lib task::tests: passed.
• Ran full test suite across the workspace: 216 tests passed.
• Ran cargo fmt --all -- --check: formatting 100% compliant.
• Ran cargo clippy --workspace --all-targets -- -D clippy::all: 0 warnings, 0 errors.

Checklist

• I have linked the related issue with Closes #13.
• I have tested my changes locally.
• I have run cargo fmt --check.
• I have run cargo clippy -- -D warnings where applicable.
• I have updated relevant documentation.