The code has not properly handled the read of a Field or Array with both
ungridded / undistributed dimensions and > 1 DE per PET. It appears that
this is due to an incorrect ordering of dealing with these two
situations in IO::read. This commit changes that ordering to resolve
this issue.

Note that I also needed to reassign temp_array_p to temp_array_undist_p
before doing the redistStore and redist; without this, the redistStore
call generated an error with this trace-back:

20230724 170325.547 ERROR            PET3 ESMCI_Array.C:6238 ESMCI::Array::tRedistStore() Invalid argument  - srcArray and dstArray must provide identical number of exclusive elements
20230724 170325.547 ERROR            PET3 ESMCI_Array.C:6018 ESMCI::Array::redistStore() Invalid argument  - Internal subroutine call returned Error
20230724 170325.547 ERROR            PET3 ESMCI_IO.C:387 ESMCI::IO::read() Invalid argument  - Internal subroutine call returned Error
20230724 170325.547 ERROR            PET3 ESMCI_IO.C:282 ESMCI::IO::read() Invalid argument  - Internal subroutine call returned Error
20230724 170325.548 ERROR            PET3 ESMCI_IO_F.C:210 c_esmc_ioread() Unable to read from file  - Internal subroutine call returned Error
20230724 170325.548 ERROR            PET3 ESMF_IO.F90:397 ESMF_IOAddArray() Unable to read from file  - Internal subroutine call returned Error
20230724 170325.548 ERROR            PET3 ESMF_FieldPr.F90:336 ESMF_FieldRead Unable to read from file  - Internal subroutine call returned Error
20230724 170325.548 ERROR            PET3 esmApp.F90:185 Unable to read from file  - Passing error in return code

Resolves #150

This tests a case that was failing until the work on this branch: each
of these situations was tested individually, but the combination wasn't
tested and wasn't working.

See #169 (comment)