[CODEC-270] Base32 and Base64 still allow decoding some invalid trailing characters

• Type: Bug
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 1.13
• Fix Version/s: 1.14
• Assignee: Alex Herbert
• Reporter: Alex Herbert
• Created: 24/Nov/19 22:46
• Updated: 21/Jan/20 13:25
• Resolved: 28/Nov/19 10:54

Description

Both Base32 and Base64 check that the final bits from the trailing digit that will be discarded are zero.

The test for the trailing bits in the final digits in Base64 is:

private long validateCharacter(final int numBitsToDrop, final Context context) {
    if ((context.ibitWorkArea & numBitsToDrop) != 0) {

It should be:

private long validateCharacter(final int numBitsToDrop, final Context context) {
    int mask = (1 << numBitsToDrop) - 1;
    if ((context.ibitWorkArea & mask) != 0) {

Likewise in Base32.

The following base64 is illegal but is still decoded:

AB==

A : 000000
B : 000001

byte = 00000000 + 0001 discarded 

Here the check for the 4 trailing bits to drop in this case checks only bit 3 and ignores bit 1 which is set.

Same for Base32, this is illegal:

AB======

A : 00000
B : 00001

byte = 00000000 + 01 discarded

But the check for the 2 trailing bits to drop in this case checks bit 2 and ignores bit 1 which is set.

Note: The test cases using "AC" has bit 2 set and so is flagged as invalid.

Issue Links

• is related to: CODEC-134 Base32 would decode some invalid Base32 encoded string into arbitrary value
• relates to: CODEC-279 Base64.decode fails on Java11 for certain valid base 64 encoded String
• relates to: CODEC-280 Base32/64 to allow optional strict/lenient decoding
• links to: GitHub Pull Request #29