Charlie Zender

Thank you for pointing this out. I'd rather repair the option to build libnco_c++, though default that to turn-off, and remove the check on the undefined variable you point out....

@pedro-vicente Can you help with or close this issue?

Thank you so much for figuring this out! This also helps explain why Appveyor fails to build NCO 4.8.0-alphaXX against netCDF 4.6.1, a problem that has been driving me crazy....

Re-opening issue to improve the Appveyor build. As the Windows output shows, `NC_HAVE_META_H` is now redundantly defined in every file, e.g.,: ``` nco_att_utl.c C:\Miniconda36-x64\envs\TEST\Library\include\netcdf.h(1993): warning C4005: 'NC_HAVE_META_H': macro redefinition C:\Miniconda36-x64\envs\TEST\Library\include\netcdf.h(1993):...

So this clumsy introduction of tokens by Unidata in the 4.6.x series creates a messy Appveyor log for all non-4.6.0 builds, which we wish to avoid. Please change `CMakeLists.txt` to...

Improved accuracy is good news. Please document the timing difference for this grid between master and your branch. Why did you create this as an issue? There are already numerous...

Please note below the formula used for the trigonometric cross-product and any special conditions you placed in the implementation. I want to ensure that the highest precision version of the...

Two questions: 1. Did you ever try to improve ill-conditioned cross-products that occur due to subtracting two nearly equal products, e.g., when a1*b2 ~ a2*b1 in computation of c0 above?...

Henry, The trigonometric cross-product function above could lead to ill-conditioned results for small angles. This occurs frequently at high resolution since lon1 ~ lon2 so lon1-lon2 ~ 0 and cos(~0)...

With Panoply I plot 2-D contour plots so I can see the geographic distribution of RMS biases which are on a lat/lon grid. There is no averaging. Please make trig...