RFR(S) : 8242314 : use reproducible random in vmTestbase shared code

[Kim Barrett]

> On Apr 22, 2020, at 2:03 PM, Igor Ignatyev <igor.ignatyev at oracle.com> wrote:
> 
> http://cr.openjdk.java.net/~iignatyev/8242314/webrev.00/ 
>> 122 lines changed: 71 ins; 19 del; 32 mod;
> 
> 
> Hi all,
> 
> could you please review this small patch which updates shared vmTestbase  test code to use j.t.l.Utils.getRandomInstance() as a random number generator or as a source of seed values? this patch lays the ground work for updates of :vmTestbase_* test groups which are to be done by remaining subtasks of 8241623[1], such split make it possible to work on/review/test testgroup-specific changes in concurrently.  
> 
> unfortunately, such extra dependency on /test/lib classes broke the shaky balance and caused some tests to fail due to "split testlibrary" problem (see 8188828[2] for the definition), the solution to that is to remove all dependencies on /test/lib classes from ExecDriver class, so the patch also includes "inlining" of several jdk.test.lib.Utils and Platform methods/fields into ExecDriver.
> 
> webrev: http://cr.openjdk.java.net/~iignatyev/8242314/webrev.00/ 
> testing: all :vmTestbase_* test groups on {linux,windows,macosx}-x64 and solaris-sparcv9
> JBS: https://bugs.openjdk.java.net/browse/JDK-8242314
> 
> [1] https://bugs.openjdk.java.net/browse/JDK-8241623
> [2] https://bugs.openjdk.java.net/browse/JDK-8188828
> 
> Thanks,
> -- Igor

The test changes look fine.

But I don't feel like I understand the "split library" problem well
enough to review the ExecDriver changes.  I will say that I don't like
the solution of copying code, and hope there is a better way.  Is
there a bug for the problem?  It seems like it could crop up again,
perhaps repeatedly, in which case the code copying becomes increasingly
bad.  And it might be a good idea to leave some breadcrumbs associated
with this code copying workaround, so an eventual better fix could also
cleanup the duplication.