From 19364d791e4eef557fa39965ea16e45da9c95722 Mon Sep 17 00:00:00 2001
From: Connor Peet <connor@peet.io>
Date: Thu, 1 Feb 2024 17:08:19 -0800
Subject: [PATCH] fix: regex in binary files by using ascii for regex matches

Fixes #406
---
 src/searchRequest.ts   | 13 ++++++-------
 src/test/range.test.ts |  2 +-
 2 files changed, 7 insertions(+), 8 deletions(-)

diff --git a/src/searchRequest.ts b/src/searchRequest.ts
index 7044548..0461f56 100644
--- a/src/searchRequest.ts
+++ b/src/searchRequest.ts
@@ -1,12 +1,11 @@
 // Copyright (c) Microsoft Corporation.
 // Licensed under the MIT license.
 
-import { HexDocument } from "./hexDocument";
-import { LiteralSearchQuery, RegExpSearchQuery, SearchResult, SearchResultsWithProgress } from "../shared/protocol";
 import { Disposable } from "vscode";
-import { utf8Length } from "./util";
-import { caseInsensitiveEquivalency, LiteralSearch, Wildcard } from "./literalSearch";
+import { LiteralSearchQuery, RegExpSearchQuery, SearchResult, SearchResultsWithProgress } from "../shared/protocol";
 import { Uint8ArrayMap } from "../shared/util/uint8ArrayMap";
+import { HexDocument } from "./hexDocument";
+import { LiteralSearch, Wildcard, caseInsensitiveEquivalency } from "./literalSearch";
 
 /** Type that defines a search request created from the {@link SearchProvider} */
 export interface ISearchRequest extends Disposable {
@@ -148,7 +147,7 @@ export class RegexSearchRequest implements ISearchRequest {
 		let strStart = 0;
 
 		const { re, document } = this;
-		const decoder = new TextDecoder();
+		const decoder = new TextDecoder("ascii");
 		const encoder = new TextEncoder();
 		const collector = new ResultsCollector(await document.size(), this.cap);
 
@@ -162,8 +161,8 @@ export class RegexSearchRequest implements ISearchRequest {
 
 			let lastReIndex = 0;
 			for (const match of str.matchAll(re)) {
-				const start = strStart + utf8Length(str.slice(0, match.index!));
-				const length = utf8Length(match[0]);
+				const start = strStart + str.slice(0, match.index!).length;
+				const length = match[0].length;
 				collector.push(encoder.encode(match[0]), start, start + length);
 				lastReIndex = match.index! + match[0].length;
 			}
diff --git a/src/test/range.test.ts b/src/test/range.test.ts
index 505dfbc..4420491 100644
--- a/src/test/range.test.ts
+++ b/src/test/range.test.ts
@@ -6,7 +6,7 @@ import { getRangeSelectionsFromStack, Range } from "../../shared/util/range";
 
 describe("Range", () => {
 
-	describe.only("getRangeSelectionsFromStack", () => {
+	describe("getRangeSelectionsFromStack", () => {
 		it("works for a single", () => {
 			expect(getRangeSelectionsFromStack([new Range(10, 20)])).to.deep.equal([new Range(10, 20)]);
 		});