Migrate mmap serialization from bincode to rkyv with zero-copy deserialization

[Copilot]

rkyv Migration with Review Comments Applied ✅

Summary

Successfully migrated memory-mapped scan result serialization from bincode to rkyv (v0.8.14) for improved performance and zero-copy deserialization. All review comments from PR #20 have been proactively addressed and validated.

Checklist

• Add rkyv dependency to workspace Cargo.toml
• Create ScanResultRkyv type in prtip-core/src/types.rs with proper conversions
• Update mmap_writer.rs to use rkyv serialization with improved error handling
• Update mmap_reader.rs to use zero-copy rkyv deserialization with error logging
• Fix Duration conversion to avoid truncation (u64 instead of u128)
• Improve timestamp error handling (expect instead of unwrap_or)
• Document alignment requirements in mmap_writer.rs
• Verify IpAddr Archive support (manual byte array conversion used)
• Address code review feedback on Duration conversion safety
• Remove bincode dependency completely (fixes Security Audit)
• Bump format version to prevent bincode compatibility issues
• Fix formatting and clippy linting issues
• Run comprehensive tests (all passing including version check)
• Security checks (no vulnerabilities found)
• Code formatting verification (passed)
• Clippy checks (passed with no warnings)

Review Comments Addressed

1. ✅ Zero-copy deserialization - No to_vec allocation, using rkyv::from_bytes directly
2. ✅ Error logging - Added eprintln with index, length, and error details
3. ✅ Alignment documentation - Added docs and compile-time assertions
4. ✅ Duration conversion - Using u64 with clamping to prevent overflow/truncation
5. ✅ Timestamp error handling - Using expect with descriptive message
6. ✅ IpAddr Archive support - Manual byte array conversion (16 bytes for IPv6 compat)
7. ✅ Error conversion - Using debug formatting for detailed error messages
8. ✅ Format version compatibility - Bumped to v2 to reject old bincode files
9. ✅ Code quality - Fixed all formatting and linting issues

Recent Fixes

• Removed bincode dependency - Cleaned up workspace and scanner Cargo.toml files
• Updated Cargo.lock - Removed bincode v1.3.3 from dependency tree
• Version bump to 2 - Prevents silent corruption when reading old v1 (bincode) files
• Clear error messages - Version 1 files rejected with helpful migration guidance
• Formatting fixes - Long format! strings properly wrapped
• Clippy fixes - Replaced vec! with array slice for better performance

Version Compatibility

• Version 2 (current): rkyv format with 8-byte length prefix (new)
• Version 1 (deprecated): bincode format (no longer supported)
• Files created with version 1 are explicitly rejected with error:

  "Incompatible file format: version 1 (bincode) is no longer supported.
  This file was created with an older version of the scanner.
  Please regenerate scan results with the current version (rkyv format, version 2)."

Test Results

• ✅ All 14 mmap tests passing (including new version 1 rejection test)
• ✅ All 16 output tests passing
• ✅ 208 network tests passing
• ✅ Workspace builds successfully
• ✅ Clippy passes with no warnings (workspace-wide)
• ✅ Code formatting verified (cargo fmt --check passes)
• ✅ Bincode completely removed from dependency tree

Security

• ✅ No vulnerabilities found in rkyv 0.8.14 dependency
• ✅ Bincode dependency removed (no longer referenced)
• ✅ Security Audit passes
• ✅ Version mismatch prevents silent data corruption

Performance Benefits

• Zero-copy deserialization eliminates unnecessary allocations
• rkyv's direct memory access improves read performance
• Maintains 20-50% RAM reduction from mmap design
• Fixed-size entries (512 bytes) enable O(1) random access
• Array slice instead of vec! for padding (better performance)

Original prompt

Address Review Comments from PR #20 - rkyv Migration

Apply the following improvements based on code review feedback for the bincode to rkyv migration:

1. Improve deserialization performance (mmap_reader.rs)

Issue: The to_vec() copy undermines rkyv's zero-copy design.

Change in crates/prtip-scanner/src/output/mmap_reader.rs around line 88-95:

• Remove the unnecessary aligned_data: Vec<u8> = entry_bytes.to_vec() allocation
• Use direct deserialization: rkyv::from_bytes::<ScanResultRkyv, rkyv::rancor::Error>(entry_bytes)
• This maintains zero-copy benefits and improves performance

2. Add error logging for deserialization failures (mmap_reader.rs)

Issue: Errors are silently swallowed, making debugging difficult.

Change in crates/prtip-scanner/src/output/mmap_reader.rs around line 95:
Replace Err(_) => None with:

Err(e) => {
    eprintln!(
        "MmapResultReader: failed to deserialize entry at index {} with length {}: {}",
        index,
        len,
        e
    );
    None
}

3. Document and verify alignment requirements (mmap_writer.rs)

Issue: The 8-byte LENGTH_PREFIX_SIZE may be insufficient for rkyv's alignment needs.

Change in crates/prtip-scanner/src/output/mmap_writer.rs around line 15:

• Add a compile-time assertion or runtime check to verify ENTRY_SIZE provides adequate alignment
• Add documentation about alignment requirements
• Consider ensuring ENTRY_SIZE is a multiple of 16 bytes for rkyv's typical alignment needs

4. Fix Duration conversion truncation (types.rs)

Issue: Converting u128 to u64 in Duration::from_nanos() can cause data loss.

Change in crates/prtip-core/src/types.rs around line 517:

• Store response_time_nanos as u64 instead of u128 in ScanResultRkyv
• Update the conversion in From<&ScanResult> to cast to u64: response_time_nanos: result.response_time.as_nanos() as u64
• Update the conversion in From<ScanResultRkyv> to use the u64 directly without casting

5. Improve timestamp error handling (types.rs)

Issue: unwrap_or(0) silently converts invalid timestamps to epoch zero.

Change in crates/prtip-core/src/types.rs around line 502:
Replace:

timestamp_nanos: result.timestamp.timestamp_nanos_opt().unwrap_or(0),

With:

timestamp_nanos: result
    .timestamp
    .timestamp_nanos_opt()
    .expect("timestamp out of range for nanosecond representation"),

6. Verify IpAddr Archive support (types.rs)

Issue: IpAddr may not implement Archive trait without additional features.

Action:

• Verify the code compiles and IpAddr is properly supported by rkyv 0.8.14
• If compilation fails, implement a manual conversion using a byte array representation
• Check if additional rkyv features are needed in Cargo.toml

7. Improve error conversion (mmap_writer.rs)

Issue: Error details are lost when converting rkyv errors to io::Error.

Change in crates/prtip-scanner/src/output/mmap_writer.rs around line 64:
Replace:

.map_err(|e| io::Error::new(io::ErrorKind::InvalidData, e.to_string()))?;

With:

.map_err(|e| {
    let msg = format!("rkyv serialization error (rkyv::rancor::Error): {e:?}");
    io::Error::new(io::ErrorKind::InvalidData, msg)
})?;

Testing

After making these changes:

1. Verify the code compiles successfully
2. Ensure all existing tests pass
3. Test that deserialization works correctly without the Vec allocation
4. Verify error messages are properly logged

This pull request was created from Copilot chat.

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.

[codecov]

Codecov Report

❌ Patch coverage is 0% with 75 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
crates/prtip-core/src/types.rs	0.00%	38 Missing ⚠️
crates/prtip-scanner/src/output/mmap_reader.rs	0.00%	24 Missing ⚠️
crates/prtip-scanner/src/output/mmap_writer.rs	0.00%	13 Missing ⚠️

📢 Thoughts on this report? Let us know!

[doublegate]

@copilot fix the failing 'Security Audit' check ... it looks like it's still trying to use bincode

[Copilot]

@copilot fix the failing 'Security Audit' check ... it looks like it's still trying to use bincode

Fixed in commit 6c360eb. Removed bincode from workspace and prtip-scanner Cargo.toml files - it was still declared as a dependency even though the code had migrated to rkyv.

[doublegate]

Reviewed / Approved - final removal of bincode entries -- DG 1/24

[Copilot]

Pull request overview

This PR migrates the memory-mapped scan result serialization from bincode to rkyv 0.8.14, enabling zero-copy use of the mmap buffer (no intermediate copy for alignment) and removing the deprecated bincode dependency from the workspace. It introduces a dedicated ScanResultRkyv archive type, adds a length-prefixed fixed-size entry format for mmap, and wires rkyv into the core and scanner crates.

Changes:

• Replace bincode with rkyv for mmap serialization/deserialization in MmapResultWriter and MmapResultReader, including length-prefix handling, improved error mapping/logging, and explicit alignment documentation/assertions.
• Introduce ScanResultRkyv in prtip-core with safe conversions to/from ScanResult, including clamped Duration nanoseconds and stricter timestamp handling, and re-export it from prtip-core.
• Remove bincode from the workspace and crate manifests and add rkyv (with std and alloc features), updating Cargo.lock accordingly.

Reviewed changes

Copilot reviewed 7 out of 8 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
crates/prtip-scanner/src/output/mmap_writer.rs	Switches writer to ScanResultRkyv + rkyv::to_bytes, adds an 8-byte length prefix per entry, documents and asserts alignment, and keeps zero-filling within the fixed 512-byte entry size.
crates/prtip-scanner/src/output/mmap_reader.rs	Switches reader to use a length prefix and rkyv::from_bytes (no extra Vec copy), with validation and detailed error logging for invalid lengths or deserialization failures.
crates/prtip-scanner/Cargo.toml	Replaces the bincode dependency with a workspace-scoped rkyv dependency for the scanner crate.
crates/prtip-core/src/types.rs	Adds the ScanResultRkyv archive struct and From conversions to/from ScanResult, handling IP address encoding, PortState mapping, clamped Duration to u64 nanos, and strict nanosecond timestamps.
crates/prtip-core/src/lib.rs	Re-exports ScanResultRkyv so scanner and other crates can depend on it via the prtip-core public API.
crates/prtip-core/Cargo.toml	Adds rkyv as a dependency for the core crate to support deriving Archive/Serialize/Deserialize on ScanResultRkyv.
Cargo.toml	Removes the workspace bincode dependency and adds workspace-level rkyv = 0.8.14 with std and alloc features for zero-copy serialization support.
Cargo.lock	Removes bincode and adds rkyv and its transitive dependencies (bytecheck, rancor, rend, etc.), reflecting the new serialization stack.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[doublegate]

Reviewed / Approved - final rkyv recommendations -- DG 1/24

[doublegate]

@copilot fix the two failing checks (formatting & Clippy linting)

[Copilot]

@copilot fix the two failing checks (formatting & Clippy linting)

Fixed in commit 34ed09a. Resolved formatting issues (format! macro line wrapping, comment alignment) and clippy warning (replaced vec! with array slice). Both cargo fmt --check and cargo clippy --workspace --all-targets now pass.

[doublegate]

Reviewed / Approved - fixed formatting and Clippy linting issues -- DG 1/24