refactor(gas): adjust FUEL_DENOM_RATE from 1000 to 20 based on benchmarks

[d1r1]

Summary

Adds fluent-testnet feature flag to maintain backward compatibility with existing testnet deployment while preparing infrastructure for correct FUEL_DENOM_RATE based on benchmarks.

Changes

• Added fluent-testnet feature flag across entire workspace (all crates, contracts, examples)
• Implemented conditional compilation for FUEL_DENOM_RATE in crates/types/src/lib.rs:
  • With fluent-testnet feature: 1000 (legacy testnet rate)
  • Without feature: 20 (correct rate based on benchmarks)
• Updated build.rs to propagate feature flag when compiling contracts
• Modified e2e tests to use hardcoded 1000 divisor (testnet compatibility mode)
• Added cfg-if workspace dependency for future conditional logic improvements
• Reverted revm-rwasm dependency to v82-patched branch

Rationale

Original 1000:1 ratio was a placeholder. Benchmarks show WASM instructions are approximately 20x faster than EVM instructions. However, existing testnet deployment uses rate 1000 and must remain stable. Feature flag enables:

• Testnet continues using legacy rate 1000 without redeployment
• Future mainnet/devnet can use correct rate 20
• Smooth transition without breaking existing deployments

Usage

# For testnet deployment (legacy rate 1000)
cargo build --features fluent-testnet

# For future mainnet/devnet (correct rate 20)
cargo build

Testing

• All e2e and builtin tests pass with fluent-testnet feature enabled (default in e2e)
• Feature flag propagates correctly through workspace and contracts
• Contract compilation respects feature flag via build.rs

Migration Path

1. Current (testnet): Deploy with --features fluent-testnet
2. Future (mainnet): Deploy without feature flag to use correct rate
3. Next steps: Separate PR will update gas calculation to use ceil rounding for costs

Notes

• This PR only adds feature flag infrastructure - no behavioral changes yet
• Gas calculation improvements (ceil vs floor) will be in follow-up PR
• All system fuel constants will be updated in future PR when switching to rate 20

[github-actions]

Criterion results (vs baseline)

running 72 tests


ERC20 Transfer Comparison/1_Original_EVM_ERC20
                        time:   [7.3482 µs 7.3491 µs 7.3501 µs]
Found 32 outliers among 1000 measurements (3.20%)
  16 (1.60%) high mild
  16 (1.60%) high severe
ERC20 Transfer Comparison/2_Emulated_EVM_ERC20
                        time:   [3.0182 ms 3.0190 ms 3.0198 ms]
Found 100 outliers among 1000 measurements (10.00%)
  1 (0.10%) low mild
  58 (5.80%) high mild
  41 (4.10%) high severe
ERC20 Transfer Comparison/3_rWasm_Contract_ERC20
                        time:   [2.4529 ms 2.4560 ms 2.4593 ms]
Found 91 outliers among 1000 measurements (9.10%)
  3 (0.30%) low severe
  4 (0.40%) low mild
  6 (0.60%) high mild
  78 (7.80%) high severe
ERC20 Transfer Comparison/4_Precompiled_UniversalToken
                        time:   [623.46 µs 624.30 µs 625.59 µs]
Found 2 outliers among 1000 measurements (0.20%)
  2 (0.20%) high severe

Greeting Contract Comparison/Original_Greeting
                        time:   [3.6306 µs 3.6322 µs 3.6341 µs]
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) high mild
Greeting Contract Comparison/EVM_Greeting
                        time:   [460.78 µs 460.89 µs 461.00 µs]
Found 10 outliers among 100 measurements (10.00%)
  3 (3.00%) low severe
  4 (4.00%) low mild
  3 (3.00%) high mild
Greeting Contract Comparison/WASM_Greeting
                        time:   [6.9533 µs 6.9553 µs 6.9574 µs]
Found 3 outliers among 100 measurements (3.00%)
  1 (1.00%) high mild
  2 (2.00%) high severe

Heads-up: runner perf is noisy; treat deltas as a smoke check.