buffer: simplify atob approach of disregarding whitespace

* Use approach of counting non-ASCII whitespace characters instead of removing the characters via string replacement * Additional atob validation tests have been added
nodejs · Apr 9, 2022 · 1c109ce · 1c109ce
1 parent 2ad1a00
commit 1c109ce
Show file tree

Hide file tree

Showing 2 changed files with 29 additions and 25 deletions.
diff --git a/lib/buffer.js b/lib/buffer.js
@@ -1232,12 +1232,14 @@ function btoa(input) {
   return buf.toString('base64');
 }
 
-// Refs: https://infra.spec.whatwg.org/#forgiving-base64-decode
-const kForgivingBase64AllowedChars = [
+const asciiWhitespaceCharacters = [
   // ASCII whitespace
   // Refs: https://infra.spec.whatwg.org/#ascii-whitespace
   0x09, 0x0A, 0x0C, 0x0D, 0x20,
+];
 
+// Refs: https://infra.spec.whatwg.org/#forgiving-base64-decode
+const kForgivingBase64AllowedChars = [
   // Uppercase letters
   ...ArrayFrom({ length: 26 }, (_, i) => StringPrototypeCharCodeAt('A') + i),
 
@@ -1260,32 +1262,29 @@ function atob(input) {
     throw new ERR_MISSING_ARGS('input');
   }
 
-  if (input === undefined || input === false || typeof input === 'number') {
-    throw lazyDOMException(
-      'The string to be decoded is not correctly encoded.',
-      'ValidationError');
-  }
+  input = `${input}`;
+  let nonAsciiWhitespaceCharCount = 0;
 
-  // Remove all ASCII whitespace from data.
-  //
-  // See #1 - https://infra.spec.whatwg.org/#forgiving-base64
-  input = `${input}`.replace(/\s/g, '');
+  for (let n = 0; n < input.length; n++) {
+    const char = StringPrototypeCharCodeAt(input, n);
+
+    if (ArrayPrototypeIncludes(kForgivingBase64AllowedChars, char)) {
+      nonAsciiWhitespaceCharCount++;
+    } else if (!ArrayPrototypeIncludes(asciiWhitespaceCharacters, char)) {
+      throw lazyDOMException('Invalid character', 'InvalidCharacterError');
+    }
+  }
 
   // If data's code point length divides by 4 leaving a remainder of 1, then
   // return failure.
   //
   // See #3 - https://infra.spec.whatwg.org/#forgiving-base64
-  if (input.length % 4 === 1) {
+  if (nonAsciiWhitespaceCharCount % 4 === 1) {
     throw lazyDOMException(
       'The string to be decoded is not correctly encoded.',
       'ValidationError');
   }
 
-  for (let n = 0; n < input.length; n++) {
-    if (!ArrayPrototypeIncludes(kForgivingBase64AllowedChars,
-                                StringPrototypeCharCodeAt(input, n)))
-      throw lazyDOMException('Invalid character', 'InvalidCharacterError');
-  }
   return Buffer.from(input, 'base64').toString('latin1');
 }
 

diff --git a/test/parallel/test-btoa-atob.js b/test/parallel/test-btoa-atob.js
@@ -16,11 +16,16 @@ throws(() => buffer.btoa(), /TypeError/);
 strictEqual(atob(' '), '');
 strictEqual(atob('  YW\tJ\njZA=\r= '), 'abcd');
 
-throws(() => buffer.atob(undefined), /ValidationError/);
-throws(() => buffer.atob(false), /ValidationError/);
-throws(() => buffer.atob(1), /ValidationError/);
-throws(() => buffer.atob(0), /ValidationError/);
-throws(() => buffer.atob('a'), /ValidationError/);
-throws(() => buffer.atob('a '), /ValidationError/);
-throws(() => buffer.atob(' a'), /ValidationError/);
-throws(() => buffer.atob('aaaaa'), /ValidationError/);
+strictEqual(atob(null), '\x9Eée');
+strictEqual(atob(NaN), '5£');
+strictEqual(atob(Infinity), '"wâ\x9E+r');
+strictEqual(atob(true), '¶»\x9E');
+strictEqual(atob(1234), '×mø');
+strictEqual(atob([]), '');
+strictEqual(atob({ toString: () => '' }), '');
+strictEqual(atob({ [Symbol.toPrimitive]: () => '' }), '');
+
+throws(() => atob(), /ERR_MISSING_ARGS/);
+throws(() => atob(Symbol()), /TypeError/);
+[undefined, false, () => {}, 0, 1, 0n, 1n, -Infinity, [1], {}].forEach((value) =>
+  throws(() => atob(value), { constructor: DOMException }));