From 3d941f2fdf08190aa347f9015a737d83718b747b Mon Sep 17 00:00:00 2001
From: Ken Aoki <ken3574.25@gmail.com>
Date: Sat, 23 Jan 2021 11:42:16 +0000
Subject: [PATCH] add BOM skip logic

---
 src/common/input/TextDecoder.test.ts | 13 +++++++++++++
 src/common/input/TextDecoder.ts      |  4 ++--
 2 files changed, 15 insertions(+), 2 deletions(-)

diff --git a/src/common/input/TextDecoder.test.ts b/src/common/input/TextDecoder.test.ts
index e9299021..da1760a2 100644
--- a/src/common/input/TextDecoder.test.ts
+++ b/src/common/input/TextDecoder.test.ts
@@ -232,6 +232,19 @@ describe('text encodings', () => {
         }
         assert(decoded, 'Ä€𝄞Ö𝄞€Ü𝄞€');
       });
+
+      it('BOMs (3 byte sequences) - advance by 2', () => {
+        const decoder = new Utf8ToUtf32();
+        const target = new Uint32Array(5);
+        const utf8Data = fromByteString('\xef\xbb\xbf\xef\xbb\xbf');
+        let decoded = '';
+        for (let i = 0; i < utf8Data.length; i += 2) {
+          const written = decoder.decode(utf8Data.slice(i, i + 2), target);
+          decoded += toString(target, written);
+        }
+        assert.equal(decoded, '');
+      });
+
       it('test break after 3 bytes - issue #2495', () => {
         const decoder = new Utf8ToUtf32();
         const target = new Uint32Array(5);
diff --git a/src/common/input/TextDecoder.ts b/src/common/input/TextDecoder.ts
index 9df26f59..715e9197 100644
--- a/src/common/input/TextDecoder.ts
+++ b/src/common/input/TextDecoder.ts
@@ -192,8 +192,8 @@ export class Utf8ToUtf32 {
             target[size++] = cp;
           }
         } else if (type === 3) {
-          if (cp < 0x0800 || (cp >= 0xD800 && cp <= 0xDFFF)) {
-            // illegal codepoint
+          if (cp < 0x0800 || (cp >= 0xD800 && cp <= 0xDFFF) || cp === 0xFEFF) {
+            // illegal codepoint or BOM
           } else {
             target[size++] = cp;
           }