# Sprint 150 Plan: Cross-Language Conflict Resolution and Merge Engine

## Context

Sprint 150 provides deterministic text/AST divergence detection and merge
policies implemented once in core and specialized through adapter hints.

---

## Goals

1. Detect divergence consistently across language adapters
2. Implement deterministic merge policies (`text_wins`, `ast_wins`, `three_way`)
3. Preserve merge provenance and replayability for audit/debugging
4. Expose conflict impact previews before apply

---

## Steps

### Step 1739: Generic text-AST divergence schema (10 tests)
### Step 1740: Adapter conflict-region mapping model (10 tests)
### Step 1741: Cross-language merge policy engine model (10 tests)
### Step 1742: Merge impact and provenance model (10 tests)
### Step 1743: `whetstone_detect_text_ast_conflicts_generic` MCP tool (8 tests)
### Step 1744: `whetstone_preview_text_ast_merge_generic` MCP tool (8 tests)
### Step 1745: `whetstone_apply_text_ast_merge_generic` MCP tool (8 tests)
### Step 1746: Cross-language merge decision packet model (8 tests)
### Step 1747: Conflict/merge qualification report artifact (8 tests)
### Step 1748: Sprint 150 integration summary + full-program regression (8 tests)

---

## Architecture Gate

- Merge outcomes must be deterministic under identical inputs
- Adapter hints must be explicit and non-authoritative over core invariants
- Apply operations must emit structured provenance for every resolved region
- Max 600 lines per header