Tech debt: Avoid store→fetch round-trip in contract validation

[sanity]

Overview

During the fix for #2214, we identified an architectural inefficiency where verify_and_store_contract() stores a contract then immediately fetches it back for validation.

Current Flow

// verify_and_store_contract() in runtime.rs

// Line 1076: We HAVE the contract
if let Some(contract) = trying_contract.take() {
    // Line 1092: Store it (moves ownership)
    self.runtime.contract_store.store_contract(contract)?;
}

// Line 1105: Immediately fetch it back for validation
let result = self.runtime.validate_state(&trying_key, ...)?;

The validate_state() call needs the contract bytes to compile a WASM module via prepare_contract_call() → fetch_contract().

The Issue

We have the contract data in hand, give it to the store, then immediately ask the store to give it back. This required adding wait() after cache insert (#2215) because stretto uses eventual consistency and the data wasn't visible yet.

Potential Solutions

1. Pass contract to validate_state directly:

   self.runtime.validate_state_with_contract(&trying_key, &contract, ...)?;
   self.runtime.contract_store.store_contract(contract)?;

2. Have store_contract also compile and cache the WASM module

3. Clone contract code before storing and use directly for compilation

Impact

• Performance: Minimal - wait() is fast (just flushes buffer)
• Correctness: Current fix works correctly
• Code clarity: The round-trip is architecturally awkward

Priority

Low - the current fix is correct and not performance-impacting. This is cleanup for code clarity.

Related

• Fixed in fix: wait for stretto cache insert before validation #2215
• Root cause issue GET request fails to find locally cached contract after different client's PUT #2214

[AI-assisted - Claude]

[github-actions]

🤖 Auto-labeled

Applied labels:

• T-enhancement (95% confidence)
• E-easy (80% confidence)
• A-contracts (95% confidence)

Reasoning: This is a code-cleanup/optimization (tech debt) to avoid an unnecessary store→fetch round-trip in contract validation and improve architectural clarity, which fits T-enhancement. The change appears small and straightforward (options include passing the in-memory contract into validation or minor store API changes), so E-easy is appropriate. The work is centered on the contract runtime/WASM compilation path, so A-contracts is the relevant area.

Previous labels: none

If these labels are incorrect, please update them. This helps improve the auto-labeling system.