use dynamic memory allocation for all builds, and remove _DA suffix

[jbathegit]

Fixes #77

This is a prerequisite for task #78, so that we eventually only have to build one version of the library. As part of this update, a number of *.F routines no longer require any preprocessing and therefore have now been moved to *.f or *.f90 names.

The problem I'm having is that all of the Python tests are failing, but that's a known issue that pre-dates this PR (see #135) and is therefore unrelated to these changes.

@aerorahul could you please help me to figure out why the Python tests are no longer working?

[kgerheiser]

I would recommend adding --output-on-failure to ctest so you can see what the errors are. I'll put in a PR.

[jbathegit]

Thanks @kgerheiser for the suggestion and PR!

So the failure(s) are now saying:
Traceback (most recent call last):
File "/home/runner/work/NCEPLIBS-bufr/NCEPLIBS-bufr/bufr/python/test/test_checkpoint.py", line 2, in
import ncepbufr
ModuleNotFoundError: No module named 'ncepbufr'

[jbathegit]

I'm wondering why did that work previously, but not now?

[edwardhartnett]

Wow this is a considerable simplification! Nice work. When we get down to only one library it will be even simpler...

[jbathegit]

I did some more digging and found two lines in python/setup.py.in that were still pointing to the _4_DA build. I've reverted those back to _4 now and we'll see if that fixes it...

[jbathegit]

So now the main.yml tasks are working, but the main_new.yml tasks (for the code coverage) aren't working, which is the exact opposite of what we saw before. Here's what the upload-artifact step is showing:

Run actions/upload-artifact@v2
with:
name: test-coverage
path: bufr/build/*.html
if-no-files-found: warn
env:
FC: gfortran-10
CC: gcc-10
With the provided path, there will be 268 files uploaded
Error: Unexpected response. Unable to upload chunk to https://pipelines.actions.githubusercontent.com/QCZondvhs1qS4vPTx8Z7shARLaFCUqoYbCWNYULz8QxSUVcZmF/_apis/resources/Containers/14479178?itemPath=test-coverage%2Ftest-coverage.bvers.f.04afe2841f19af69af3b8c9be888e136.html

Begin Diagnostic HTTP information

Status Code: 408
Status Message: Request Timeout
Header Information: {
"content-type": "application/json",
"strict-transport-security": "max-age=2592000",
"x-tfs-processid": "15646812-9d62-492b-a2de-b88807a1ea5f",
"activityid": "966583a7-d7d3-44bc-a21a-dc9582061619",
"x-tfs-session": "966583a7-d7d3-44bc-a21a-dc9582061619",
"x-vss-e2eid": "966583a7-d7d3-44bc-a21a-dc9582061619",
"x-vss-senderdeploymentid": "13a19993-c6bc-326c-afb4-32c5519f46f0",
"x-tfs-serviceerror": "The+request+has+been+canceled%3a+Client+disconnected..",
"x-cache": "CONFIG_NOCACHE",
"x-msedge-ref": "Ref A: 0E7D4EAC6BC44479B95EE8A1FC09D61F Ref B: PHL30EDGE0208 Ref C: 2021-08-18T17:46:28Z",
"date": "Wed, 18 Aug 2021 17:46:28 GMT",
"content-length": "0"
}

End Diagnostic HTTP information

Warning: Aborting upload for /home/runner/work/NCEPLIBS-bufr/NCEPLIBS-bufr/bufr/build/test-coverage.bvers.f.04afe2841f19af69af3b8c9be888e136.html due to failure
Error: aborting artifact upload
Total size of all the files uploaded is 69704 bytes
Finished uploading artifact test-coverage. Reported size is 69704 bytes. There were 255 items that failed to upload
Error: An error was encountered when uploading test-coverage. There were 255 items that failed to upload.

@edwardhartnett or @aerorahul , any ideas? If we can figure this out then maybe we can also close #135 ;-)

[jbathegit]

Interestingly, the code coverage task for the push process just finished successfully, even though it failed previously on the pull_request process. Does that make any sense? Maybe just a temporary connection glitch?

If you want I can try rerunning the checks(?)

[jbathegit]

So I went ahead and re-ran the failed check, and it now runs to completion successfully. Go figure!?

[jbathegit]

Hi @edwardhartnett and @aerorahul. Could at least one of you please review this, and either approve or let me know if you have any further suggestions for improvement? I've also added @jack-woollen in case he may be able to help.

Just FYI, I've got a separate update ready which should take care of the few remaining GNU compiler warnings for this library, but I'd like to get this PR squared away first.

[aerorahul]

Looks good to me, although I did not review each file carefully.

In a future PR, why not just move all .f to .f90 and where appropriate .F to .F90 or .f90 if there is no preprocessor involved.

No one is using g77 to compile this code anymore.