[TestBug] RFR : JDK-8192909 - Invalid username or password in HashedPasswordFileTest.java

[Harsha Wardhana B]

Hi Daniel,


On Tuesday 05 December 2017 03:42 PM, Daniel Fuchs wrote:
> Hi Harsha,
>
> Looks good.
Thanks for the review.
>
> nit:
>
>  366                 if(random.nextBoolean()) {
>  367                     String[] tokens = line.split("\\s+");
>  368                     if ((tokens.length == 4 || tokens.length == 
> 3)) {
>
> inverting the two if () (testing for the applicability of the line
> first) would probably give a better chance that an existing
> password is replaced, unless most lines are applicable.
All the lines will be applicable. The password file will be hashed 
before the above lines are executed. An Assert statement at line 353 
makes sure of that. Hence no point inverting the two if().
>
> best regards,
>
> -- daniel
Regards
Harsha
>
> n 04/12/2017 18:27, Harsha Wardhana B wrote:
>> Hi All,
>>
>> Please review and provide comments for fix for,
>>
>> issue: https://bugs.openjdk.java.net/browse/JDK-8192909
>>
>> having webrev at,
>>
>> webrev : http://cr.openjdk.java.net/~hb/8192909/webrev.00/
>>
>> Fix details: The test was failing intermittently because of duplicate 
>> entries for role in input password file. The duplicate entries get 
>> over-written by JMX agent, but the client was testing with stale 
>> entries for duplicated role. Also, the test now uses a single random 
>> number generator from test package (Utils.getRandomInstance) instead 
>> of two.
>>
>> Regards
>>
>> Harsha
>>
>