Decode: 4 Chars to 3 Bytes

Reverse of encoding. For each 4-char chunk:

1. Map each char to its 6-bit value (or 0 if =).
2. Combine into a 24-bit integer.
3. Split back into 3 bytes.
4. Trim trailing bytes per the = count.

Validation:

• Input length must be multiple of 4 (after padding)
• Only valid alphabet chars (or =)
• = only allowed at the end (1 or 2 of them)
• Stricter validators check that the trailing 6-bit value's "ignored" bits are zero (since they padded with zero bytes)

Most implementations are lenient: accept whitespace (newlines especially — PEM files have line breaks every 64 or 76 chars), ignore unknown chars, etc.