Search results for "UTF-8 with BOM" files shifted on first line by a character (microsoft#66189)

Victorique Ko · roblourens · commit 8c2cef95b3b3 · 2019-01-10T09:04:56.000-08:00
* Fix microsoft#66188 * Strip utf-8 bom from matchText * Index can't be negative
diff --git a/src/vs/workbench/services/search/node/ripgrepTextSearchEngine.ts b/src/vs/workbench/services/search/node/ripgrepTextSearchEngine.ts
@@ -237,7 +237,12 @@ export class RipgrepParser extends EventEmitter {
 
 	private createTextSearchMatch(data: IRgMatch, uri: vscode.Uri): vscode.TextSearchMatch {
 		const lineNumber = data.line_number - 1;
-		const fullText = bytesOrTextToString(data.lines);
+		let isBOMStripped = false;
+		let fullText = bytesOrTextToString(data.lines);
+		if (lineNumber === 0 && startsWithUTF8BOM(fullText)) {
+			isBOMStripped = true;
+			fullText = stripUTF8BOM(fullText);
+		}
 		const fullTextBytes = Buffer.from(fullText);
 
 		let prevMatchEnd = 0;
@@ -255,6 +260,11 @@ export class RipgrepParser extends EventEmitter {
 			}
 
 			let matchText = bytesOrTextToString(match.match);
+			if (lineNumber === 0 && i === 0 && isBOMStripped) {
+				matchText = stripUTF8BOM(matchText);
+				match.start = match.start <= 3 ? 0 : match.start - 3;
+				match.end = match.end <= 3 ? 0 : match.end - 3;
+			}
 			const inBetweenChars = fullTextBytes.slice(prevMatchEnd, match.start).toString().length;
 			let startCol = prevMatchEndCol + inBetweenChars;
 
@@ -265,12 +275,6 @@ export class RipgrepParser extends EventEmitter {
 				stats.lastLineLength :
 				stats.lastLineLength + startCol;
 
-			if (lineNumber === 0 && i === 0 && startsWithUTF8BOM(matchText)) {
-				matchText = stripUTF8BOM(matchText);
-				startCol -= 3;
-				endCol -= 3;
-			}
-
 			prevMatchEnd = match.end;
 			prevMatchEndCol = endCol;
 			prevMatchEndLine = endLineNumber;
