chore: add From and TryFrom conversions between dto and foreign chain inspector types

[DSharifi]

closes #2085

[claude]

Code Review

Found critical issues that need addressing before merge:

1. ⚠️ Asymmetric Conversion - Data Loss Risk

File: crates/foreign-chain-inspector/src/contract_interface_conversions.rs:41-51

The TryFrom<dtos::EvmFinality> for EthereumFinality conversion rejects EvmFinality::Latest even though EthereumFinality::Latest exists (line 36). This creates an asymmetric conversion:

• ✅ EthereumFinality::Latest → dtos::EvmFinality::Latest works
• ❌ dtos::EvmFinality::Latest → EthereumFinality::Latest fails

Impact: If the contract sends a finality level of Latest, follower nodes will fail to process verify_foreign_tx requests.

Fix: Either:

• Support bidirectional conversion by handling Latest case, OR
• Document why Latest is intentionally unsupported in the reverse direction

2. ⚠️ Integer Overflow Risk (Platform-Dependent)

File: crates/foreign-chain-inspector/src/contract_interface_conversions.rs:150, 162

log_index: log_index as u64,    // line 150 - usize → u64
log_index: log_index as usize,  // line 162 - u64 → usize

Issue: On 32-bit platforms, casting u64 → usize can silently overflow when log_index > 2^32.

Impact:

• Wrong log extracted from transaction on 32-bit systems
• Silent data corruption (no error, just wrong index)

Fix:

// Line 162
log_index: log_index.try_into()
    .map_err(|_| UnsupportedVariantError { 
        context: "log_index exceeds platform usize" 
    })?,

3. 💭 Future Maintainability Concern

All TryFrom implementations use catch-all patterns (_ => Err(...)) for #[non_exhaustive] enums. When new variants are added to DTOs:

• ✅ Compilation succeeds (no warnings)
• ❌ Runtime conversions fail silently
• ⚠️ No visibility that conversions need updating

Recommendation: Add explicit variant matches instead of wildcards where possible to get compiler warnings on missing cases.

---

⚠️ Issues Found - Please address the integer overflow and asymmetric conversion issues before merge.

[DSharifi]

@claude

[netrome]

Lgtm, a few nits

[gilcu3]

LGTM!

[gilcu3]

I am a bit surprised that the Log type was not defined in this crate, therefore conversion cannot be done idiomatically.