JDK8 Code review request for 7199939 and 7199941

Vincent Ryan vincent.x.ryan at oracle.com
Tue Sep 25 14:21:07 UTC 2012


Understood.
Fixes look good.

On 24/09/2012 21:27, Valerie (Yu-Ching) Peng wrote:
> Vinnie,
>
> Thanks for the comment!
> The reason that I use "<" instead of "<=" on line 150 is because that
> (1024, 160) is supported by the DSAGenParameterSpec class and thus I
> didn't view this as legacy value. Either init(...) calls should work for
> 1024, so I'd like to use the newer form of init(...) for 1024 to
> separate it from the legacy values.
> I noticed that some comments in ParameterCache class is a bit outdated,
> so I updated webrev again just to update the comments to mention the
> default/pre-computed parameters for (2048, 224) and (2048, 256).
> http://cr.openjdk.java.net/~valeriep/7199939/webrev.01/
> Thanks,
> Valerie
>
> On 09/24/12 02:39, Vincent Ryan wrote:
>> Fixes look good.
>> Just a minor comment at l.150 of ParameterCache: the test could use
>> '<='??
>>
>> On 22 Sep 2012, at 00:45, Valerie (Yu-Ching) Peng wrote:
>>
>>> Hi, Vinnie,
>>>
>>> Can you please review these following 2 JDK8 fixes? They are straight forward regressions found by SQE tests introduced as byproducts of the NSA Suite B changes.
>>> Since they are already caught by existing SQE tests, so I didn't add more regression tests.
>>>
>>> 7199939  <http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=7199939>: DSA 576 ad 640 bit keys fail when initializing for No precomputed parameters
>>> Webrev:http://cr.openjdk.java.net/~valeriep/7199939/webrev.00/
>>> Fix: Fixed the impl of DSAKeyPairGenerator.initialize(int, SecureRandom) so it will attempt to generate new parameters (instead
>>> of error out) when no precomputed parameters are available. Also fixed ParameterCache.getNewDSAParameterSpec(...) method to handle legacy DSA key sizes.
>>>
>>>
>>>
>>> 7199941  <http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=7199941>: test about AES/ECB mode fails
>>> Webrev:http://cr.openjdk.java.net/~valeriep/7199941/webrev.00/
>>> Fix: Moved the initialization of the field 'blockMode' back to the right code block.
>>>
>>> The failed SQE tests now pass after the code changes.
>>>
>>> Thanks!
>>> Valerie
>>
>




More information about the security-dev mailing list