Add numberOfElements parameter to getArrayletLayout()

[dmitripivkine]

0-length arrays have Discontiguous layout.
Currently detection of 0-length arrays based on dataSizeInBytes. There is not good enough for 0-stride arrays obviously (have dataSizeInBytes equal 0). All 0-stride arrays should be treated as Contiguous with exception of 0-length 0-stride should be Discontiguous. Passing of numberOfElements helps to solve this problem, dataSizeInBytes is not required to be passed any more.

[dmitripivkine]

In #17994 fix for #17994 we have assumed 0-length 0-stride array can be treated as Contiguous (as any other 0-stride array). This appeared to be incorrect in the current code. 0-length 0-stride array has to be treated as Discontiguous exceptionally (as any other 0-length array).
There is only functional change in this PR, rest of it is code refactoring.

[amicic]

jenkins sanity aix,win jdk17

[dmitripivkine]

@hangshao0 @hzongaro @a7ehuo FYI
I don't think it should change or break anything, it relates to 0-length 0-stride arrays only. The difference between Contiguous and Discontiguous arrays is an object header format in this case.

[amicic]

jenkins test sanity aix,win jdk17

[dmitripivkine]

Failure in Windows build relates to infrastructure. Re-launching.

[dmitripivkine]

jenkins test sanity win jdk17

[dmitripivkine]

I don't see other errors in the Windows build except

18:04:33  [WS-CLEANUP] Deleting project workspace...
18:04:33  [WS-CLEANUP] Deferred wipeout is disabled by the job configuration...
18:05:02  ERROR: Cannot delete workspace :Unable to delete 'F:\Users\jenkins\workspace\Build_JDK17_x86-64_windows_Personal\openssl\NUL'. Tried 3 times (of a maximum of 3) waiting 0.1 sec between attempts.

Looks like it is reason for failure for me.

[hzongaro]

@dmitripivkine @amicic, would this change have caused this error in the Windows PR builds:

18:00:28  f:\users\jenkins\workspace\build_jdk17_x86-64_windows_personal\openj9\runtime\gc_glue_java\ArrayletObjectModel.hpp(1152): error C2220: the following warning is treated as an error
18:00:28  f:\users\jenkins\workspace\build_jdk17_x86-64_windows_personal\openj9\runtime\gc_glue_java\ArrayletObjectModel.hpp(1152): warning C4244: '=': conversion from 'uintptr_t' to 'uint32_t', possible loss of data

[amicic]

it indeed is (ArrayletObjectModel.hpp(1152)). I'll revert it - it's not an urgent change, anyway