refactor(diffUtils): remove SEARCH/REPLACE handling and normalize diff content processing

Author: hannesrudolph

webview-ui/src/components/chat/ChatRow.tsx

@@ -399,15 +399,11 @@ export const ChatRowContent = ({
 	// Inline diff stats for edit/apply_diff/insert/search-replace/newFile asks
 	const diffTextForStats = useMemo(() => {
 		if (!tool) return ""
-		// For appliedDiff, backend provides unified diff; do not fallback/normalize
-		if ((tool as any).tool === "appliedDiff") {
-			return ((tool as any).content as string) || ""
-		}
 		return (
 			extractUnifiedDiff({
 				toolName: tool.tool as string,
 				path: tool.path,
-				diff: (tool as any).content,
+				diff: (tool as any).content ?? (tool as any).diff,
 				content: (tool as any).diff,
 			}) || ""
 		)
@@ -420,14 +416,10 @@ export const ChatRowContent = ({
 	// Clean diff content for display (normalize to unified diff)
 	const cleanDiffContent = useMemo(() => {
 		if (!tool) return undefined
-		// For appliedDiff, show backend's unified diff directly
-		if ((tool as any).tool === "appliedDiff") {
-			return ((tool as any).content as string) || undefined
-		}
 		const unified = extractUnifiedDiff({
 			toolName: tool.tool as string,
 			path: tool.path,
-			diff: (tool as any).content,
+			diff: (tool as any).content ?? (tool as any).diff,
 			content: (tool as any).diff,
 		})
 		return unified || undefined

webview-ui/src/utils/diffUtils.ts

@@ -1,7 +1,6 @@
 /**
  * Frontend-only normalization helper.
  * - If a unified diff already exists, return it.
- * - If a Roo SEARCH/REPLACE block is provided, convert to unified diff.
  * - If it's a new file with raw content, synthesize a unified diff with all lines as additions.
  * - Otherwise, pass through raw content (DiffView will no-op if not unified).
  */
@@ -17,17 +16,14 @@ export function extractUnifiedDiff(params: {
 	if (!raw) return ""
 
 	raw = stripCData(raw)
+	// Remove diff noise lines like "\ No newline at end of file"
+	raw = raw.replace(/(^|\n)[ \t]*(?:\\ )?No newline at end of file[ \t]*(?=\n|$)/gi, "$1")
 
 	// Explicit new file: build a unified diff from raw content
 	if ((params.toolName || "").toLowerCase() === "newfilecreated") {
 		return convertNewFileToUnifiedDiff(raw, filePath)
 	}
 
-	// SEARCH/REPLACE blocks → unified
-	if (isSearchReplace(raw)) {
-		return convertSearchReplaceToUnifiedDiff(raw, filePath)
-	}
-
 	// Already unified?
 	if (isUnifiedDiff(raw)) {
 		return raw
@@ -44,13 +40,6 @@ function isUnifiedDiff(s: string): boolean {
 	return hasHunk || hasHeaders
 }
 
-/** Detects Roo SEARCH/REPLACE multi-block format */
-function isSearchReplace(s: string): boolean {
-	return (
-		/(^|\n)<<<<<<< ?SEARCH|(^|\n)<<<<<<<? ?SEARCH|(^|\n)<<<<<<< SEARCH>/.test(s) || /(^|\n)<<<<<<< ?SEARCH/.test(s)
-	)
-}
-
 /** Remove CDATA markers and any HTML-encoded variants */
 function stripCData(s: string): string {
 	return (
@@ -61,71 +50,10 @@ function stripCData(s: string): string {
 	)
 }
 
-/**
- * Convert Roo SEARCH/REPLACE blocks into unified diff using the diff library.
- * Matches optional metadata lines and optional '-------' separator.
- */
-export function convertSearchReplaceToUnifiedDiff(content: string, filePath?: string): string {
-	// Backend-compatible regex: captures :start_line: and :end_line:, optional '-------', and SEARCH/REPLACE bodies
-	const blockRegex =
-		/(?:^|\n)(?<!\\)<<<<<<< SEARCH>?\s*\n((?::start_line:\s*(\d+)\s*\n))?((?::end_line:\s*(\d+)\s*\n))?((?<!\\)-------\s*\n)?([\s\S]*?)(?:\n)?(?:(?<=\n)(?<!\\)=======\s*\n)([\s\S]*?)(?:\n)?(?:(?<=\n)(?<!\\)>>>>>>> REPLACE)(?=\n|$)/g
-
-	const fileName = filePath || "file"
-	let hasBlocks = false
-	let headerEmitted = false
-	let unified = ""
-
-	// Helper to normalize EOLs and get stable line arrays without trailing empty caused by final newline
-	const toStableLines = (s: string): string[] => {
-		const norm = s.replace(/\r\n/g, "\n")
-		if (norm === "") return []
-		const parts = norm.split("\n")
-		return parts[parts.length - 1] === "" ? parts.slice(0, -1) : parts
-	}
-
-	let match: RegExpExecArray | null
-	while ((match = blockRegex.exec(content)) !== null) {
-		hasBlocks = true
-
-		// 1: full start_line line, 2: start_line number, 3: full end_line line, 4: end_line number
-		const startLine = match[2] ? parseInt(match[2], 10) : 1
-		const endLine = match[4] ? parseInt(match[4], 10) : undefined
-
-		// 6: SEARCH body, 7: REPLACE body
-		const searchBody = match[6] ?? ""
-		const replaceBody = match[7] ?? ""
-
-		const searchLines = toStableLines(searchBody)
-		const replaceLines = toStableLines(replaceBody)
-
-		// Old/new hunk metadata. If end_line is present, prefer it for oldLines; otherwise count SEARCH lines.
-		const oldStart = startLine
-		const oldLines = endLine !== undefined ? Math.max(0, endLine - startLine + 1) : searchLines.length
-		const newStart = startLine
-		const newLines = replaceLines.length
-
-		// Emit file headers once so parsePatch can recognize a complete unified diff
-		if (!headerEmitted) {
-			unified += `--- a/${fileName}\n`
-			unified += `+++ b/${fileName}\n`
-			headerEmitted = true
-		}
-
-		// Hunk header
-		unified += `@@ -${oldStart},${oldLines} +${newStart},${newLines} @@\n`
-
-		// We don't have surrounding context here; emit deletions then additions to visualize the change
-		for (const line of searchLines) unified += `-${line}\n`
-		for (const line of replaceLines) unified += `+${line}\n`
-	}
-
-	return hasBlocks ? unified : content
-}
-
 /** Build a unified diff for a brand new file (all content lines are additions).
  * Trailing newline is ignored for line counting and emission.
  */
-export function convertNewFileToUnifiedDiff(content: string, filePath?: string): string {
+function convertNewFileToUnifiedDiff(content: string, filePath?: string): string {
 	const fileName = filePath || "file"
 	// Normalize EOLs to keep counts consistent
 	const normalized = content.replace(/\r\n/g, "\n")