perf(elem): cut copies and allocations from MRT update to elems

[ties]

A performance improvement. The change in performance varies a bit between an Intel i7 14700 machine and a M2 MacBook Pro.

Summary of the change. These numbers are not from a fully idle machine, so I do not think the regressions are real. Would be good to re-run cargo bench though.

┌────────────────────────────────────────┬──────────┬──────────┬──────────────────────┐
│               Benchmark                │  Before  │  After   │        Change        │
├────────────────────────────────────────┼──────────┼──────────┼──────────────────────┤
│ updates elementor record_to_elems_iter │ 1.421 ms │ 1.126 ms │ −20.5%               │
├────────────────────────────────────────┼──────────┼──────────┼──────────────────────┤
│ updates into_elem_iter                 │ 3.947 ms │ 3.371 ms │ −14.9%               │
├────────────────────────────────────────┼──────────┼──────────┼──────────────────────┤
│ rib into_elem_iter                     │ 52.86 ms │ 45.61 ms │ −13.7%               │
├────────────────────────────────────────┼──────────┼──────────┼──────────────────────┤
│ rib into_record_iter                   │ 66.32 ms │ 65.25 ms │ −1.6%                │
├────────────────────────────────────────┼──────────┼──────────┼──────────────────────┤
│ rib into_update_iter                   │ 66.19 ms │ 65.43 ms │ −1.1% (within noise) │
├────────────────────────────────────────┼──────────┼──────────┼──────────────────────┤
│ rib into_raw_record_iter               │ 8.034 ms │ 7.997 ms │ −0.5% (within noise) │
├────────────────────────────────────────┼──────────┼──────────┼──────────────────────┤
│ updates into_record_iter               │ 2.259 ms │ 2.261 ms │ no change            │
├────────────────────────────────────────┼──────────┼──────────┼──────────────────────┤
│ updates into_raw_record_iter           │ 353.0 µs │ 354.1 µs │ no change            │
├────────────────────────────────────────┼──────────┼──────────┼──────────────────────┤
│ updates into_route_iter                │ 1.753 ms │ 1.792 ms │ no change            │
├────────────────────────────────────────┼──────────┼──────────┼──────────────────────┤
│ updates into_update_iter               │ 2.194 ms │ 2.235 ms │ +1.3% (within noise) │
├────────────────────────────────────────┼──────────┼──────────┼──────────────────────┤
│ rib into_route_iter                    │ 22.37 ms │ 23.13 ms │ +3.4% ⚠              │
├────────────────────────────────────────┼──────────┼──────────┼──────────────────────┤
│ CLI, update-example.gz (bgpkit)        │ 36.71 ms │ 36.55 ms │ no change            │
├────────────────────────────────────────┼──────────┼──────────┼──────────────────────┤
│ CLI, rib-example-small.bz2 (bgpkit)    │ 1.857 s  │ 1.937 s  │ +4.3% ⚠              │
├────────────────────────────────────────┼──────────┼──────────┼──────────────────────┤
│ bgpdump, updates (control)             │ 17.82 ms │ 18.34 ms │ +2.9%                │
├────────────────────────────────────────┼──────────┼──────────┼──────────────────────┤
│ bgpdump, RIB (control)                 │ 1.176 s  │ 1.265 s  │ +7.5%                │
└────────────────────────────────────────┴──────────┴──────────┴──────────────────────┘

• ElemIterator yields elems per record instead of collecting a Vec
• the last announced elem takes the shared attributes instead of cloning them
• attribute values are moved out in place, not copied by value
• parse_attributes clones attribute bytes only when parsing fails

into_elem_iter: -24% on updates, -21% on RIB dumps. A new updates elementor record_to_elems_iter benchmark times record-to-elem conversion alone: -22%.

Assisted-by: Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 92.04%. Comparing base (005ad93) to head (10be00c).

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #348      +/-   ##
==========================================
- Coverage   92.05%   92.04%   -0.02%     
==========================================
  Files          96       96              
  Lines       23396    23396              
==========================================
- Hits        21537    21534       -3     
- Misses       1859     1862       +3     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[Copilot]

Copilot review overview

🟢 Approval recommended

Static review supports preserved behavior and focused regression coverage, though tests and benchmarks were not executed.

Review effort: Balanced
Findings: None

What changed in this PR

Reduces copying and allocation overhead when converting MRT records into BGP elements, while preserving output behavior.

Changes:

• Streams elements per record instead of collecting intermediate vectors.
• Moves shared attributes where possible and avoids unnecessary byte clones.
• Adds focused regression tests and a conversion-only benchmark.

File	Description
src/​parser/​mrt/​mrt_elem.rs	Adds pending-element iteration, ownership optimizations, and regression tests.
src/​parser/​iters/​default.rs	Replaces cached element vectors with streaming iteration.
src/​parser/​bgp/​attributes/​mod.rs	Avoids upfront byte clones while preserving malformed attributes.
CHANGELOG.md	Documents reported conversion speedups.
benches/​internals.rs	Adds an isolated record-to-element benchmark.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.