On Wed, 15 Nov 2023 18:56:11 GMT, Sean Mullan <mul...@openjdk.org> wrote:
>> Add parameters to `MessageDigest` and introduce `SHAKE256-LEN` with an
>> integer parameter.
>
> src/java.base/share/classes/java/security/MessageDigest.java line 436:
>
>> 434: * {@code MessageDigestSpi} implementation for the
>> 435: * specified algorithm and parameters
>> 436: *
>
> Is there a reason why you didn't throw IAPE when params are bad?
This is following the `SecureRandom` style where there are both `getInstance`
methods that has or has not the parameters argument. If you think IAPE is
better I can also throw it. `CertStore` does this way and it has a dedicated
method to extract the cause of the NSAE thrown by `GetInstance.getInstance` and
rethrow it. See
https://github.com/openjdk/jdk/blob/9486969bd3cb084c89a7255de0c664c980d1e661/src/java.base/share/classes/java/security/cert/CertStore.java#L249
-------------
PR Review Comment: https://git.openjdk.org/jdk/pull/16660#discussion_r1394778408
