From b8e377278a5017c4e0aeb702a5ba539cf9fbbae8 Mon Sep 17 00:00:00 2001
From: Meta Vens <rwvens@hotmail.com>
Date: Sun, 30 Aug 2026 23:21:47 +0200
Subject: [PATCH] Canonicalize source checkout line endings

---
 packages/rag/src/source-document-archive.ts | 6 +++++-
 scripts/smoke-source-rag-commit-boundary.ts | 8 ++++----
 2 files changed, 9 insertions(+), 5 deletions(-)

diff --git a/packages/rag/src/source-document-archive.ts b/packages/rag/src/source-document-archive.ts
index cc58fd9..cb1d049 100644
--- a/packages/rag/src/source-document-archive.ts
+++ b/packages/rag/src/source-document-archive.ts
@@ -433,11 +433,15 @@ function comparableFields(document: ArchivedSourceDocumentRecord): Record<string
     path: document.path,
     language: document.language,
     title: document.title,
-    content: document.content,
+    content: canonicalizeLineEndings(document.content),
     metadata: document.metadata ?? {},
   };
 }
 
+function canonicalizeLineEndings(content: string): string {
+  return content.replace(/\r\n?/g, "\n");
+}
+
 function compareDocuments(
   left: ArchivedSourceDocumentRecord,
   right: ArchivedSourceDocumentRecord,
diff --git a/scripts/smoke-source-rag-commit-boundary.ts b/scripts/smoke-source-rag-commit-boundary.ts
index 12d8c43..b2d976e 100644
--- a/scripts/smoke-source-rag-commit-boundary.ts
+++ b/scripts/smoke-source-rag-commit-boundary.ts
@@ -25,16 +25,16 @@ async function main(): Promise<void> {
 
   try {
     const archive = new FileSourceDocumentArchiveRepository(archivePath);
-    const original = document("original body", "2026-06-03T00:00:00.000Z");
+    const original = document("original\r\nbody", "2026-06-03T00:00:00.000Z");
     const updated = document("updated body", "2026-08-30T00:00:00.000Z");
     await archive.syncRepoDocuments("Fixture", [original]);
 
     const timestampOnlyPlan = await archive.planRepoDocuments(
       "Fixture",
-      [document("original body", "2026-08-30T00:00:00.000Z")],
+      [document("original\nbody", "2026-08-30T00:00:00.000Z")],
     );
-    assertEqual(timestampOnlyPlan.changedSourceIds.length, 0, "checkout timestamp changed document identity");
-    assertEqual(timestampOnlyPlan.unchanged, 1, "checkout timestamp did not remain observational");
+    assertEqual(timestampOnlyPlan.changedSourceIds.length, 0, "checkout metadata changed document identity");
+    assertEqual(timestampOnlyPlan.unchanged, 1, "checkout timestamp and line endings did not remain observational");
 
     const failingPipeline = new SourceRagPipeline(
       archive,