[#105] Fixes decryption of data with size 0 bytes and 1 byte #108
Conversation
| @@ -302,6 +303,33 @@ mod tests { | |||
| assert_eq!(*decrypted_plaintext, plaintext[..]); | |||
| } | |||
|
|
|||
| // Even very small documents of 0 and 1 bytes should roundtrip between AesEncryptedValue and bytes | |||
| #[test] | |||
| fn test_roundtrip_aesencryptedvalue_zero_one_bytes() -> Result<(), IronOxideErr> { | |||
There was a problem hiding this comment.
This test name seems like it could be improved or the test changed because you're not really doing a "roundtrip" test right? I would expect a roundtrip to do an encrypt and decrypt and assert you got the same value out.
There was a problem hiding this comment.
My intention is that the AesEncryptedValue is roundtripping to/from a bytes representation. This is narrower than a full AES encrypt/decrypt. I only included the encrypt step to avoid referencing a the AES_IV_LEN and AES_GCM_TAG_LEN consts. That was probably a bad idea. I have reworked this test to remove encrypt so hopefully it's more clear what's being tested.
| updated: doc_meta.0.updated, | ||
| decrypted_data: decrypted_doc.to_vec(), | ||
| } | ||
| })?, |
There was a problem hiding this comment.
Just curious, is this a more Rust-preferred way of doing this?
There was a problem hiding this comment.
I would say yes, so long as you are just calling into() on an error and not inspecting the error at all.
See #105
Decryption had a check to explicitly disallow "empty" encrypted data. The check was also mistakenly disallowing 1 byte. Both restrictions were relaxed and tests introduced.