Duplicate Code: sequence field codec mark encode/decode loops (V2 vs V3)

🔍 Duplicate Code Detected: Sequence field codec mark encode/decode loops

Analysis of commit 0350e0deb1b4ad2e1a060a7c65fded411de6882c

Assignee: @copilot

Summary

The sequence-field codecs for format V2 and V3 contain a largely duplicated encode/decode loop that walks the marks array and maps common fields (count, effect, cellId, changes). This duplication is >10 lines and increases maintenance risk when evolving the wire format or adding mark fields.

Duplication Details

Pattern: Encode/decode loops for Changeset marks

• Severity: Medium

• Occurrences: 2

• Locations:

  • packages/dds/tree/src/feature-libraries/sequence-field/sequenceFieldCodecV2.ts (lines 283-345)
  • packages/dds/tree/src/feature-libraries/sequence-field/sequenceFieldCodecV3.ts (lines 93-141)

• Code Sample (duplicated structure):

  V2 (excerpt):

  const jsonMarks: Encoded.Changeset(NodeChangeSchema) = [];
  for (const mark of changeset) {
    const encodedMark: Encoded.Mark(NodeChangeSchema) = { count: mark.count };
    if (!isNoopMark(mark)) {
      encodedMark.effect = markEffectCodec.encode(mark, context.baseContext);
    }
    if (mark.cellId !== undefined) {
      encodedMark.cellId = changeAtomIdCodec.encode(mark.cellId, context.baseContext);
    }
    if (mark.changes !== undefined) {
      encodedMark.changes = context.encodeNode(mark.changes);
    }
    jsonMarks.push(encodedMark);
  }
  

  V3 (excerpt):

  const jsonMarks: Encoded.Changeset(NodeChangeSchema) = [];
  for (const mark of changeset) {
    const encodedMark: Encoded.Mark(NodeChangeSchema) = { count: mark.count };
    if (!isNoopMark(mark)) {
      encodedMark.effect = markEffectCodec.encode(mark, context.baseContext);
    }
    if (mark.cellId !== undefined) {
      encodedMark.cellId = atomIdCodec.encode(mark.cellId, context.baseContext);
    }
    if (mark.changes !== undefined) {
      encodedMark.changes = context.encodeNode(mark.changes);
    }
    jsonMarks.push(encodedMark);
  }
  

Impact Analysis

• Maintainability: Changes to shared mark encoding/decoding rules must be duplicated across V2 and V3, increasing the chance of drift.
• Bug Risk: Fixes in one version (e.g., handling of optional fields or future mark metadata) can be missed in the other.
• Code Bloat: Duplicated boilerplate obscures the actual differences between V2 and V3.

Refactoring Recommendations

1. Extract shared mark-loop helpers

   • Extract to something like packages/dds/tree/src/feature-libraries/sequence-field/sequenceFieldCodecHelpers.ts (or export helpers from sequenceFieldCodecV2.ts).
   • Suggested split:
     • encodeMarks(changeset, context, { effectCodec, atomIdCodec })
     • decodeMarks(encoded, context, { effectCodec, atomIdCodec, postprocessDecodedMark? })
   • Estimated effort: Medium (2–4 hours)
   • Benefits: Single point of truth for common fields; V2 can keep its rename-normalization postprocess while V3 can omit it.

2. Make version differences explicit

   • Keep V2/V3-specific logic limited to effect codec differences and any version-specific postprocessing.
   • Benefits: Easier to reason about format deltas when adding new versions.

Implementation Checklist

• Extract shared encode/decode loops into helpers
• Update V2 and V3 codecs to use helpers
• Ensure V2’s rename normalization behavior is preserved (postprocess hook)
• Run existing unit tests for sequence-field codecs

Analysis Metadata

• Analyzed Files: 2
• Detection Method: Serena semantic code analysis + manual verification
• Commit: 0350e0deb1b4ad2e1a060a7c65fded411de6882c
• Analysis Date: 2026-03-26T21:46:55.817Z

Generated by Duplicate Code Detector · ◷

To install this agentic workflow, run

gh aw add github/gh-aw/.github/workflows/duplicate-code-detector.md@94662b1dee8ce96c876ba9f33b3ab8be32de82a4

Start with the encode/decode loops in packages/dds/tree/src/feature-libraries/sequence-field/sequenceFieldCodecV2.ts and sequenceFieldCodecV3.ts, then inspect the existing sequence-field codec unit tests. Extract shared helpers while keeping the V2 rename-normalization postprocess and version-specific codecs distinct. Done means both codecs use the helpers and the existing codec tests pass.