Jacob Williams

Note: added this in f7539e5ef0c455bb800a72e34792decaf988ea9c but it doesn't seem to work right for the jpl code, so I disabled it for now. Will try to fix.

The options could be: * `null_to_integer_mode` * 0 (default) - raise an exception - what is does now * 1 - user specified int value for nulls * `null_to_integer_value` -...

Yeah, as you might be able to guess, the reason I did this was so adding a new argument didn't require me to make the same change in 10 different...

ah ok. Since it's just a workaround for a minor annoyance, I'm open to changing it. I'd want to run all the unit tests though (I need to see if...

Note: see also: https://github.com/fortran-lang/setup-fortran/issues/12

FYI: The `18.1.5` version at [conda-forge](https://anaconda.org/conda-forge/flang) is [giving me](https://github.com/jacobwilliams/json-fortran/actions/runs/9151296172/job/25157115399): > error: loc("/home/runner/work/json-fortran/json-fortran/src/json_value_module.F90":15:5): /home/conda/feedstock_root/build_artifacts/flang-split_1715212541315/work/flang/lib/Lower/CallInterface.cpp:948: not yet implemented: support for polymorphic types LLVM ERROR: aborting

FYI: I've not been able to successfully get it to work from those scripts... seems like they run but then flang-new isn't in the path when it's done... (I tried...

I haven't really had a chance to look at it, and don't really have time right now. If you (or anybody) want to take a look and let me know...

I'd need to think about it some more. I guess there are two inefficiencies currently: * storing the independent arrays multiple times (x,y,...). this is a RAM issue. * searching...