Resolves building issues with Intel compilers (ifx/icx)

[mgs-52]

Adding/Replacing missing prototypes. Remove unsupported/bogus compiler switches.

TYPE: Bug fix

KEYWORDS: Missing Prototypes

SOURCE: Alexander Knyazev

DESCRIPTION OF CHANGES:
Problem: Most C code written long ago. No prototypes were used.

Solution:
Add missing prototypes.

LIST OF MODIFIED FILES:

external/io_grib1/grib1_util/read_grib.c
external/io_grib1/grib1_util/write_grib.c
external/io_grib1/grib1_util/read_grib.h
external/io_grib1/MEL_grib1/FTP_getfile.c
external/io_grib1/MEL_grib1/gribfuncs.h
external/io_grib1/MEL_grib1/gribgetbds.c
external/io_grib1/MEL_grib1/upd_child_errmsg.c
external/io_grib1/MEL_grib1/gribhdr2file.c
external/io_grib1/MEL_grib1/init_enc_struct.c
external/io_grib1/MEL_grib1/init_gribhdr.c
external/io_grib1/MEL_grib1/init_struct.c
external/io_grib1/MEL_grib1/ld_enc_input.c
external/io_grib1/MEL_grib1/ld_enc_lookup.c
external/io_grib1/MEL_grib1/map_lvl.c
external/io_grib1/MEL_grib1/map_parm.c
external/io_grib1/MEL_grib1/prt_inp_struct.c
external/io_grib1/MEL_grib1/prt_badmsg.c
external/io_grib1/MEL_grib1/set_bytes.c
external/io_grib1/MEL_grib1/ld_grib_origctrs.c
external/io_grib1/gribmap.c
external/io_grib1/grib1_routines.c
external/io_grib1/gribmap.h
external/io_grib1/trim.c
external/io_grib1/trim.h
external/io_grib_share/get_region_center.c
external/io_grib_share/open_file.c
external/RSL_LITE/gen_comms.c
external/RSL_LITE/c_code.c
external/RSL_LITE/buf_for_proc.c
external/RSL_LITE/rsl_malloc.c
external/RSL_LITE/rsl_bcast.c
external/RSL_LITE/task_for_point.c
external/RSL_LITE/period.c
external/RSL_LITE/swap.c
external/RSL_LITE/cycle.c
external/RSL_LITE/rsl_lite.h
tools/protos.h
tools/reg_parse.c
tools/misc.c
tools/symtab_gen.c
tools/gen_config.c
tools/sym.c
tools/gen_streams.c
tools/gen_irr_diag.c
tools/sym.h
frame/pack_utils.c
arch/configure.defaults

TESTS CONDUCTED:

1. Compilation with Intel oneAPI compiler works with these changes.
2. The Jenkins tests are all passing.

RELEASE NOTE: This PR fixed some c code by adding and replacing missing prototypes to allow Intel oneAPI compiler to work. Remove unsupported/bogus compiler switches.

[weiwangncar]

Jenkins test results:

Test Type              | Expected  | Received |  Failed
= = = = = = = = = = = = = = = = = = = = = = = =  = = = =
Number of Tests        : 23           24
Number of Builds       : 60           57
Number of Simulations  : 158           150        0
Number of Comparisons  : 95           86        0

Failed Simulations are: 
None
Which comparisons are not bit-for-bit: 
None

[islas]

@mgs-52 This should target develop

[weiwangncar]

@islas @mgduda I just compiled this code on Derecho. With these modifications, one can compile WRF using the default modules. Would it be ok if we move this PR for release-4.5.2 release?

[islas]

Yes, this should go into the next release. My own testing with the latest Intel compilers from their download page also works with these updates

The base branch was changed.

[weiwangncar]

@mgs-52 Can you let us know your affiliation, so that we can fill the SOURCE section of the PR? Thanks.

[islas]

@mgs-52 Any updates? I think affiliation/SOURCE is the last change needed

[mgs-52]

Well. Technically I'm working in Intel. However this change I did on my own so fill free to Include Intel or not :).

[weiwangncar]

Well. Technically I'm working in Intel. However this change I did on my own so fill free to Include Intel or not :).

How about your name? Would you want to include your real name? With that, we don't have to add Intel.

[mgs-52]

Alexander (Alex) Knyazev.

[islas]

@mgduda When you get a chance, could you take a look at this? Ignoring whitespace changes most of the edits are rearrangement of function order and fixing bad function signatures (implicit returns and Fortran-esque argument typing)

[ahheo]

@mgs-52 May you please do the same job for the codes under WRF/chem (especially those of KPP) as well? I guess WRF-chem users may also get trouble with ifx/icx. Actually, I have done the code modification dealing with the issues of missing of prototypes and function order, and successfully built the WRF with WRF_CHEM=1 and WRF_KPP=1 using the newest Intel compilers. Well, as this is of the same issue that this pull is dealing with, I therefore suggest to complete the same job taking into account WRF-chem before the pull to be merged. What do you think @weiwangncar @islas?

[weiwangncar]

@ahheo Can you point out the files you modified for ifx/icx in chem/ directory? I think it is also ok to do another PR to address WRF-Chem.

[ahheo]

@weiwangncar I have done another PR #1942 as you suggested.

[smileMchen]

@mgs-52 Alexander Knyazev,
Would you please tell where you are working? We would like to have this information for release note. Thanks.

[mgs-52]

@mgs-52 Alexander Knyazev, Would you please tell where you are working? We would like to have this information for release note. Thanks.

Intel, but I as stated above this doesn't need to be included since I did it on my own.

[weiwangncar]

@mgs-52 Will these icx compiler flags, '-Wno-implicit-function-declaration -Wno-incompatible-function-pointer-types', help with any of the old c code changed in this PR? See PR-1972 for WRFDA compilation change. Just curious.

[mgs-52]

@mgs-52 Will these icx compiler flags, '-Wno-implicit-function-declaration -Wno-incompatible-function-pointer-types', help with any of the old c code changed in this PR? See PR-1972 for WRFDA compilation change. Just curious.
They do, but I'd say it's a clutch: better make the code "compatible" with the ("C") standard rather than instruct compiler to ignore "incompatibilities".