Fix detection of short comments <!-->

domenic · domenic · commit 28f680f5c332 · 2020-01-05T12:10:25.000-05:00
Closes #4.
diff --git a/lib/html-encoding-sniffer.js b/lib/html-encoding-sniffer.js
@@ -27,8 +27,8 @@ function prescanMetaCharset(buffer) {
     let c = buffer[i];
     if (c === 0x3C) {
       // "<"
-      let c1 = buffer[i + 1];
-      let c2 = buffer[i + 2];
+      const c1 = buffer[i + 1];
+      const c2 = buffer[i + 2];
       const c3 = buffer[i + 3];
       const c4 = buffer[i + 4];
       const c5 = buffer[i + 5];
@@ -37,11 +37,10 @@ function prescanMetaCharset(buffer) {
         i += 4;
         for (; i < l; i++) {
           c = buffer[i];
-          c1 = buffer[i + 1];
-          c2 = buffer[i + 2];
+          const cMinus1 = buffer[i - 1];
+          const cMinus2 = buffer[i - 2];
           // --> (comment end)
-          if (c === 0x2D && c1 === 0x2D && c2 === 0x3E) {
-            i += 2;
+          if (c === 0x3E && cMinus1 === 0x2D && cMinus2 === 0x2D) {
             break;
           }
         }
diff --git a/test/no-bom-charset-short-comment.html b/test/no-bom-charset-short-comment.html
@@ -0,0 +1,7 @@
+<!DOCTYPE html>
+<html>
+    <head>
+    <!--><meta charset="iso-8859-2">
+    </head>
+    <body></body>
+</html>
diff --git a/test/tests.js b/test/tests.js
@@ -205,6 +205,34 @@ describe("A file with no BOM and a ><meta charset>", () => {
   });
 });
 
+describe("A file with no BOM and a <meta charset> preceeded by a short comment <!-->", () => {
+  const buffer = read("no-bom-charset-short-comment.html");
+
+  it("should sniff as the charset value, given no options", () => {
+    const sniffedEncoding = htmlEncodingSniffer(buffer);
+
+    assert.strictEqual(sniffedEncoding, "ISO-8859-2");
+  });
+
+  it("should sniff as the transport layer encoding, given that", () => {
+    const sniffedEncoding = htmlEncodingSniffer(buffer, {
+      transportLayerEncodingLabel: "windows-1251",
+      defaultEncoding: "ISO-8859-16"
+    });
+
+    assert.strictEqual(sniffedEncoding, "windows-1251");
+  });
+
+
+  it("should sniff as the charset value, given only a default encoding", () => {
+    const sniffedEncoding = htmlEncodingSniffer(buffer, {
+      defaultEncoding: "ISO-8859-16"
+    });
+
+    assert.strictEqual(sniffedEncoding, "ISO-8859-2");
+  });
+});
+
 for (const utf16Encoding of ["utf-16be", "utf-16", "utf-16le"]) {
   describe(`A file with a BOM and a <meta charset> of ${utf16Encoding}`, () => {
     const buffer = read(`no-bom-charset-${utf16Encoding}.html`);
