cloudtrust / fpe Goto Github PK

• Second implementation review by a cryptographer

I happened to notice that EPFL's LASEC has the following papers on their website:

Breaking The FF3 Format-Preserving Encryption Standard Over Small Domains (CRYPTO 2017)
Breaking the FF3 Format Preserving Encryption (ESC 2017)

See https://lasec.epfl.ch/

I do not know if these attacks have implications for your implementation and/or applications.

@bsoniam code review suggests a few improvements/bug fixes.

1. In FF3 (ff3Encrypter, method CryptBlocks), we test that len(dst) < len(numeralString). It is wrong, it should be len(dst) < len(src).
2. Add seed when we use rand.
3. In FF1 and FF3 method CryptBlocks: Change test of dst size (len(dst) < len(src)) to check strict equality (len(dst) == len(src)).
4. Remove magic numbers from tests.
5. In function BytesToNumeralString, add math.Ceil for the size of the array.
6. Using reflect.DeepEqual to compare int slices is overkill, make simple function that do the comparison.
7. Fix imports