(close #1987) Replace Intrinsic Operators with IntrinsicCalls

[sergisiso]

This allows to delete lots of code (Nary Operations, named argument operations) and improve other code (use is_pure, is_elemental, is_inquiry) for intrinsics as if they are regular calls instead of having special cases.

This turned out to be a big PR so I stopped here, but I would like to follow up with:
#2102 Which may mean moving the intrinsic enum from IntrinsicCall to a new IntrinsicSymbol
#2302 To also list the mandatory intrinsic arguments. (which I could do it this PR if the reviewer is fine with a longer one)

Also closes: #1366, #537, #1161, #1837 and #414 ? and PR #2222

[codecov]

Codecov Report

Patch coverage: 100.00% and project coverage change: -0.01% ⚠️

Comparison is base (b4ceb72) 99.85% compared to head (3a8afd3) 99.84%.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2298      +/-   ##
==========================================
- Coverage   99.85%   99.84%   -0.01%     
==========================================
  Files         342      342              
  Lines       46144    46069      -75     
==========================================
- Hits        46075    45998      -77     
- Misses         69       71       +2     

Files Changed	Coverage Δ
src/psyclone/domain/lfric/__init__.py	100.00% <ø> (ø)
src/psyclone/psyir/nodes/__init__.py	100.00% <ø> (ø)
src/psyclone/psyir/nodes/ranges.py	100.00% <ø> (ø)
...clone/psyir/transformations/intrinsics/__init__.py	100.00% <ø> (ø)
...psyir/transformations/intrinsics/sum2code_trans.py	100.00% <ø> (ø)
...ests/psyir/frontend/fparser2_alloc_handler_test.py	100.00% <ø> (ø)
...mmon/transformations/kernel_module_inline_trans.py	100.00% <100.00%> (ø)
...main/gocean/transformations/gocean_opencl_trans.py	100.00% <100.00%> (ø)
src/psyclone/domain/lfric/lfric_builtins.py	100.00% <100.00%> (ø)
src/psyclone/domain/lfric/lfric_loop_bounds.py	100.00% <100.00%> (ø)
... and 46 more

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[arporter]

The NEMO integration tests (ecmwf build) failed with:

NEMOGCM_V40/cfgs/SPITZ12_openmp_cpu/WORK/obs_fbm.f90(1024): error #6410: This name has not been declared as an array or a function.   [CDLTMP]
      cdltmp(idx) = ACHAR(IACHAR(cdltmp(idx)) + 32)
----------^
NEMOGCM_V40/cfgs/SPITZ12_openmp_cpu/WORK/obs_fbm.f90(1024): warning #7319: This argument's data type is incompatible with this intrinsic procedure; procedure assumed EXTERNAL.   [IACHAR]
      cdltmp(idx) = ACHAR(IACHAR(cdltmp(idx)) + 32)
-------------------------------------^
NEMOGCM_V40/cfgs/SPITZ12_openmp_cpu/WORK/obs_fbm.f90(1024): error #6404: This name does not have a type, and must have an explicit type.   [IACHAR]
      cdltmp(idx) = ACHAR(IACHAR(cdltmp(idx)) + 32)
------------------------------^
NEMOGCM_V40/cfgs/SPITZ12_openmp_cpu/WORK/obs_fbm.f90(1024): error #6303: The assignment operation or the binary expression operation is invalid for the data types of the two operands.   [ACHAR]
      cdltmp(idx) = ACHAR(IACHAR(cdltmp(idx)) + 32)
------------------------^

Please could you take a look.

[sergisiso]

The integration test issue comes from expanding the ranges into a loop for:
cdltmp(1:1) = ACHAR(IACHAR(cdltmp(1 : 1)) + 32)
because it now knows the intrinsics are elemental. But intel compiler is fine with the (1:1) range but not the:

do idx = 1, 1, 1
    cdltmp(idx) = ...

[sergisiso]

The issue with NEMO integration test is now fixed, I decided to filter the file and leave the fix to another PR because this one is already quite big and I noticed that the omp_gpu_trans had this file already filtered, so both scripts are now doing this same.
Also note that this PR provides an improvement from 22s to 19.7s to the NEMO OpenACC kernels.

[sergisiso]

@arporter This is ready for another review. I submitted an integration test job that came out green and I have fixed a couple of codecov issues to make the patch 100%.

[arporter]

Well done on fixing things and thanks for improving the code by adding the dangling Assignment node. I just have a couple of small queries about that and then this can be merged.

[sergisiso]

@arporter This is ready for another review. I have addressed one of the comments and just replied to the other, let me know if you disagree with the second.

[arporter]

All requested changes have been made.
Integration tests were all green (20 hrs ago).
Will proceed to merge.