RFR: 8331008: Implement JEP 478: Key Derivation Function API (Preview) [v15]

[Weijun Wang]

On Mon, 26 Aug 2024 18:15:30 GMT, Kevin Driver <kdriver at openjdk.org> wrote:

>> src/java.base/share/classes/javax/crypto/KDF.java line 317:
>> 
>>> 315:      *     if no {@code Provider} supports a {@code KDFSpi} implementation for
>>> 316:      *     the specified algorithm
>>> 317:      * @throws InvalidAlgorithmParameterException
>> 
>> No IAPE is thrown in the current implementation.
>
> This is declared for implementation use. Are you arguing that no implementation may require it?

This is `KDF` not `KDFSpi`, the implementation is in this file. I just meant this exception should be thrown when the condition described here is met.

>> src/java.base/share/classes/javax/crypto/KDFSpi.java line 54:
>> 
>>> 52:  * {@code KDFParameters} may throw an {@code InvalidAlgorithmParameterException}
>>> 53:  * upon receiving a {@code null} value. Furthermore, implementations
>>> 54:  * may supply default values for {@code KDFParameters}, mutating the
>> 
>> The `mutating` word is suspicious. The object is very likely to be immutable. Just say a different object should be returned in the next sentence.
>
> This depends on the implementation. It may be just as inaccurate to say "a different object".

OK, maybe you can just say the "actual" object. I still think `mutating` is strange.

-------------

PR Review Comment: https://git.openjdk.org/jdk/pull/20301#discussion_r1731682691
PR Review Comment: https://git.openjdk.org/jdk/pull/20301#discussion_r1731684535