RFR (XS): 7102300: performance warnings cause results diff failure in Test6890943

[Vladimir Kozlov]

On 2/17/13 2:29 AM, Leonid Mesnik wrote:
> Hi
>
> Test prints only lines started with 'Case #'. It would be better filter
> only them so any VM warnings are skipped.

We want the test to fail if it print any other warnings because they are 
not expected.

>
> Also it would be nice to fails if java returns none-zero code and add
> more informative summary of failure.

If java exits abnormally the output will not match and we can look on it 
to investigate the problem.

I think current fix solved the problem with the test. In a future we can 
rewrite the test to have "golden" output in java test (String[]) and 
compare output in java without using shell script (my dream :) ).

Thanks,
Vladimir

>
> Leonid
>
> On 02/15/2013 10:51 PM, David Chase wrote:
>> http://cr.openjdk.java.net/~drchase/7102300/webrev.00/
>>
>> bug:
>> compiler/5091921/Test6890943.java sometimes tickles a hashtable
>> performance warning, unrelated to what it is supposed to be testing.
>> This warning causes a diff against golden results, which is classified
>> as a failure.
>>
>> fix:
>> (1) Strip lines matching the performance warning from the output
>> before diff.
>> (2) Add -XX+IgnoreUnrecognizedVMOptions to avoid other
>> failure-generating warnings.
>>
>> testing:
>> JTreg, JPRT.
>>
>
>