RFR(S) : 8203238: [TESTBUG] rewrite MemOptions shell test in Java

[Igor Ignatyev]

> On Apr 22, 2020, at 2:58 AM, Kim Barrett <kim.barrett at oracle.com> wrote:
> 
>> On Mar 25, 2020, at 1:42 PM, Igor Ignatyev <igor.ignatyev at oracle.com> wrote:
>> 
>> http://cr.openjdk.java.net/~iignatyev//8203238/webrev.00
>>> 330 lines changed: 91 ins; 236 del; 3 mod; 
>> 
>> Hi all,
>> 
>> could you please review this small patch which rewrites MemOptions shell test?
>> 
>> while porting the test, I noticed that available memory checks aren't required, and the test successfully passes even w/o them, so the java version of the test doesn't check available memory and only @requires 64 bits vm. given the test doesn't require lots of time/resources to execute, I've also removed it from exclusiveAccess. MemStat class was made static inner class of MemOptionsTest for the sake of readability and brevity.
>> 
>> webrev: http://cr.openjdk.java.net/~iignatyev//8203238/webrev.00
>> testing: the changed tests multiple tests on {linux, windows, mac} w/ {SerialGC,ZGC,G1GC,ParallelGC}
>> JBS: https://bugs.openjdk.java.net/browse/JDK-8203238
>> 
>> NB the shell version of the test had a bug which prevent its execution. an incorrect operator (:=) was used at L#23,23, which led to bogus 'java' variable at L#44 and non zero exit code at L#48, so the test passes w/ 'Skipping the test; a 64-bit VM is required.' message on all platforms. so this patch effectively resurrects the test.
>> 
>> Thanks,
>> -- Igor
> 
> Looks good.
> 
> The resolution of the discussion around Misha's comment about the 4G
> min memory all seems fine.
> 
> I'm wondering why this has `@library /vmTestbase`. Everything here
> looks like vanilla Java + /test/lib.
it's just an artifact from conversion of vm testbase tests to jtreg, I've check that the test passed w/ it; so removed. 
> 
> There should be a blank line before the MemStat class definition.
added.
> 
> I don't need another webrev for the above.

thanks for your review, pushed.


-- Igor