Oracle auth check is O(n) linear scan on every proof submission — degrades with multi-oracle deployments

[cybermax4200]

Why this matters now:
The engine supports a multi-oracle roster (added as a feature). Every call to submit_proof, approve_proof, reject_proof, dispute_proof, and resolve_dispute calls require_oracle → is_registered_oracle → read_oracles → iterates the full Vec<Address>. With even 10 oracles this is a 10-entry Vec scan on every proof operation. This is a compute cost and CPU-instruction budget problem that grows linearly with oracle count, directly threatening the test_approve_proof_budget budget assertion that must stay under 50% of the Soroban CPU limit.

Problem / What:
storage::is_registered_oracle calls read_oracles (a full Vec read from instance storage) then calls .iter().any(|o| o == *addr). This is O(n) in oracle count for every auth check. The fix is to mirror the Sponsor pattern already used in task-registry: store oracle membership as DataKey::OracleFlag(Address) -> bool in instance storage (O(1) lookup), maintaining the DataKey::Oracles Vec only for the get_oracles() enumeration endpoint. Both structures must be kept in sync on add_oracle, remove_oracle, and set_oracle.

Key Challenges:

• add_oracle, remove_oracle, and set_oracle must maintain both the OracleFlag map and the Oracles Vec atomically (Soroban has no transactions, but both writes happen in the same contract invocation).
• set_oracle replaces the entire roster — must clear all existing OracleFlag entries before setting the new one; requires iterating the old Vec exactly once during the replacement.
• Budget test test_approve_proof_budget must still pass (and should improve).

Acceptance Criteria:

• is_registered_oracle(addr) performs exactly 1 instance storage read regardless of oracle count.
• add_oracle, remove_oracle, set_oracle keep OracleFlag(Address) and Oracles Vec consistent.
• get_oracles() still returns the full list.
• All existing oracle tests pass.
• test_approve_proof_budget CPU budget assertion still passes.

Relevant files/functions:

• contracts/reward-engine/src/storage.rs — is_registered_oracle, push_oracle, remove_oracle_from_list, write_oracles, read_oracles, DataKey
• contracts/reward-engine/src/verification.rs — require_oracle, add_oracle, remove_oracle, set_oracle

Out of scope: Changing oracle authorization semantics, TTL management.

[midenotch]

Hi Maintainer,
Interested in working on this issue. I'll implement the requested changes and ensure everything is properly tested before submission.

Payout wallet (Stellar): GBLT6QSRPF7YGSCOWEOGBXKS2DRKOPBLF6QHJAD6QCNCWWQ54E2NJ4AB
GrantFox OSS campaign — smart escrow payout on merge.

[Anichris-koded]

Hi, my name is Winner, and I’m a professional software engineer experienced in blockchain performance, oracle infrastructure, and smart contract optimization. I can replace the O(n) authorization scan performed on every proof submission with a more efficient lookup strategy that scales reliably as oracle providers increase. My approach focuses on reducing computational overhead, preserving strict authorization guarantees, improving scalability, and adding comprehensive performance and security tests.

[Salauayo]

Hello Maintainer! I’d like to be assigned this issue. I’ve reviewed the O(n) oracle membership lookup and the required OracleFlag(Address) optimization. I understand the need to keep the flag map and roster Vec fully consistent across add_oracle, remove_oracle, and set_oracle, while preserving existing behavior and improving the CPU budget. I’m ready to implement and test this end to end.

[Santia2004]

Hi! I would like to work on this issue for the GrantFox campaign. I have experience with this stack, understand the requirements, and I'm ready to submit a clean, well-tested PR once assigned. Could you please assign it to me?

[CodeMayor]

Hi! I'd like to work on this issue. I have experience working with Rust and smart-contract storage patterns, and I can implement the oracle membership optimization while preserving the existing authorization semantics. I'll introduce the OracleFlag(Address) storage mapping for constant-time membership checks, keep the existing oracle list for enumeration, and update add_oracle, remove_oracle, and set_oracle so both representations remain consistent. I'll also add or update the relevant tests and verify that the oracle suite and test_approve_proof_budget continue to pass, with the change improving rather than weakening the Soroban CPU budget usage.

[grantfox-oss]

🦊 GrantFox — @Salauayo has been assigned to this issue as part of the Third Campaign campaign!

Next steps:

1. Open a Pull Request referencing this issue (e.g., Closes #75)
2. Your PR will be reviewed by the ecotask-network maintainers

Good luck! Track your progress on GrantFox.