22ff2b6b93
fix: Node.js atob input validation
89 lines
2.5 KiB
Diff
89 lines
2.5 KiB
Diff
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
|
|
From: Shelley Vohr <shelley.vohr@gmail.com>
|
|
Date: Tue, 23 Aug 2022 11:13:45 +0200
|
|
Subject: buffer: fix `atob` input validation
|
|
|
|
This patch combines:
|
|
|
|
* https://github.com/nodejs/node/pull/42539
|
|
* https://github.com/nodejs/node/pull/42662
|
|
|
|
To bring the Node.js implementation of atob into alignment with the
|
|
WHATWG spec.
|
|
|
|
diff --git a/lib/buffer.js b/lib/buffer.js
|
|
index 57d6cddbaa2e6bdd846a667897588dea18daeb42..7602d4049e9bb1c09440bc3af09ad5ad9c768308 100644
|
|
--- a/lib/buffer.js
|
|
+++ b/lib/buffer.js
|
|
@@ -23,8 +23,10 @@
|
|
|
|
const {
|
|
Array,
|
|
+ ArrayFrom,
|
|
ArrayIsArray,
|
|
ArrayPrototypeForEach,
|
|
+ ArrayPrototypeIndexOf,
|
|
MathFloor,
|
|
MathMin,
|
|
MathTrunc,
|
|
@@ -1231,8 +1233,25 @@ function btoa(input) {
|
|
return buf.toString('base64');
|
|
}
|
|
|
|
-const kBase64Digits =
|
|
- 'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/=';
|
|
+// Refs: https://infra.spec.whatwg.org/#forgiving-base64-decode
|
|
+const kForgivingBase64AllowedChars = [
|
|
+ // ASCII whitespace
|
|
+ // Refs: https://infra.spec.whatwg.org/#ascii-whitespace
|
|
+ 0x09, 0x0A, 0x0C, 0x0D, 0x20,
|
|
+
|
|
+ // Uppercase letters
|
|
+ ...ArrayFrom({ length: 26 }, (_, i) => StringPrototypeCharCodeAt('A') + i),
|
|
+
|
|
+ // Lowercase letters
|
|
+ ...ArrayFrom({ length: 26 }, (_, i) => StringPrototypeCharCodeAt('a') + i),
|
|
+
|
|
+ // Decimal digits
|
|
+ ...ArrayFrom({ length: 10 }, (_, i) => StringPrototypeCharCodeAt('0') + i),
|
|
+
|
|
+ 0x2B, // +
|
|
+ 0x2F, // /
|
|
+ 0x3D, // =
|
|
+];
|
|
|
|
function atob(input) {
|
|
// The implementation here has not been performance optimized in any way and
|
|
@@ -1241,11 +1260,31 @@ function atob(input) {
|
|
if (arguments.length === 0) {
|
|
throw new ERR_MISSING_ARGS('input');
|
|
}
|
|
+
|
|
input = `${input}`;
|
|
+ let nonAsciiWhitespaceCharCount = 0;
|
|
+
|
|
for (let n = 0; n < input.length; n++) {
|
|
- if (!kBase64Digits.includes(input[n]))
|
|
+ const index = ArrayPrototypeIndexOf(
|
|
+ kForgivingBase64AllowedChars,
|
|
+ StringPrototypeCharCodeAt(input, n));
|
|
+
|
|
+ if (index > 4) {
|
|
+ // The first 5 elements of `kForgivingBase64AllowedChars` are
|
|
+ // ASCII whitespace char codes.
|
|
+ nonAsciiWhitespaceCharCount++;
|
|
+ } else if (index === -1) {
|
|
throw lazyDOMException('Invalid character', 'InvalidCharacterError');
|
|
+ }
|
|
}
|
|
+
|
|
+ // See #3 - https://infra.spec.whatwg.org/#forgiving-base64
|
|
+ if (nonAsciiWhitespaceCharCount % 4 === 1) {
|
|
+ throw lazyDOMException(
|
|
+ 'The string to be decoded is not correctly encoded.',
|
|
+ 'InvalidCharacterError');
|
|
+ }
|
|
+
|
|
return Buffer.from(input, 'base64').toString('latin1');
|
|
}
|
|
|