wow-viewer Shared I/O Implementation

Prompt file imported from akspa0/parp-tools (.codex/prompts/wow-viewer-shared-io-implementation.md). Copyright stays with the author.

Implement the next narrow wow-viewer shared-format or shared-I/O slice without losing the current ownership and validation rules.

Read First

1. gillijimproject_refactor/memory-bank/activeContext.md
2. gillijimproject_refactor/memory-bank/progress.md
3. gillijimproject_refactor/plans/wow_viewer_shared_io_library_plan_2026-03-26.md
4. wow-viewer/README.md
5. AGENTS.md

Goal

Move one more reusable non-PM4 format capability into wow-viewer/src/core/WowViewer.Core or wow-viewer/src/core/WowViewer.Core.IO with explicit validation and without reintroducing tool-local parser ownership. Typical slices include ADT root or split ADT family seams, WDT summaries, WMO top-level readers, or early BLP or DBC or DB2 detection or summary work.

Current Working Rules

• WowViewer.Core and WowViewer.Core.IO are the intended shared home for non-PM4 file detection, chunk reading, and future format ownership.
• WowViewer.Tool.Inspect and WowViewer.Tool.Converter should stay thin consumers of those shared seams.
• The default direction is library-first wow-viewer work, not broader MdxViewer integration.
• Shared detection or summary behavior is real progress, but it is not the same thing as deep payload parsing or write support.

Current WMO Continuation Note

• The current validated WMO root-light seam is standard and Alpha MOLT semantic summary in shared Core.IO, with real proof for Alpha ironforge.wmo.MPQ and standard 0.6.0 world/wmo/khazmodan/cities/ironforge/ironforge.wmo.
• Already-settled facts for the standard 48-byte layout: bytes 2..3 are a raw non-zero headerFlagsWord on real Ironforge (0x0101), offsets 24..39 are quaternion rotation, and offsets 40 and 44 are attenuation start and end.
• If the user says to continue this work without naming a narrower seam, default to one of these next WMO follow-ups instead of revisiting resolved offsets:
  1. prove whether headerFlagsWord varies across additional real standard roots
  2. add a per-light inspect dump for standard MOLT entries
• Do not claim per-bit meaning for headerFlagsWord until more than Ironforge has been checked.

Active Shared Surface To Build On

• FourCC, ChunkHeader, ChunkHeaderReader
• ChunkedFileReader
• map contracts: MapChunkIds, MapFileKind, MapChunkLocation, MapFileSummary
• MapFileSummaryReader
• file detection contracts: WowFileKind, WowFileDetection
• WowFileDetector
• WowViewer.Tool.Inspect verb map inspect
• WowViewer.Tool.Converter verb detect
• regression floor in wow-viewer/tests/WowViewer.Core.Tests

Non-Negotiable Constraints

• Do not turn a narrow shared-I/O slice into a broad migration-plan rewrite.
• Do not add new tool-local heuristics when the seam should live in Core or Core.IO.
• Do not claim full parsing or write support when the slice only proves detection or top-level summary.
• Prefer one concrete shared seam with proof over several partial abstractions.

What The Work Must Produce

1. The exact seam to add or change.
2. The files that should own it in wow-viewer.
3. The validation required for that seam.
4. Any current-production touch point that remains reference-only.
5. The continuity files that must be updated after the slice lands.

Deliverables

Return all items:

1. the next shared I/O slice to implement
2. why that slice is the right next step
3. exact files to change
4. exact validation to run
5. what should stay out of scope for this slice
6. which memory or prompt surfaces must be updated afterward

First Output

Start with:

1. the current shared I/O boundary you are assuming
2. the single next seam you would extract or add
3. the narrowest proof that would show the slice is real
4. what you are explicitly not claiming yet