I will write 100 times "I won't forget to update the tests anymore" (when swapping MeteoData::parameters around). I used the opportunity to imrpove the outputs of the 2D_interpolations test.

Fixed a stupid bug in Unsworth generator (it was always generating data...), improved the documentation of POT_RADIATION and tried to play with a cloudiness factor (not working yet).

A new generator for ISWR or RSWR has been implemented, providing clear sky, potential radiation. Overall, writing the code and its documentation and testing it took less than half an hour!

A buf in some unsworth implementation in Atmosphere has been fixed (check with nodata). The order of the meteodata fields has been changed because this directly impacts the order of the processing of the fields (the goal being that fields that are often checked with a combination of other fields get processed in the end).

The unsworth_dilley ILWR data generator has been implemented with the usage of the last cloudiness during the night as well as the usage of a generic snow albedo (or grass albedo) to convert RSWR to ISWR (if necessary) coupled with a threshold on the snow height.

The METEO_TIMSERIES and STATION_TIMESERIES have been renamed in METEO_SET and STATIONS_SET in order to better reflect their content. 

The concept of data generators is introduced. A new section has been created [Generators] to define how to generate data for each parameter when everything else failed. Soon, you will be able to generate ILWR directly by declaring it in your io.ini!

The soap files for GSN have been re-generated with gsoap 2.8.7 and some explanations on how to do this added in a txt file

Since compiling the meteoio examples is usually problematic for the users, trying to get a proper (but still as simple as possible) CMakeLists.txt. The FindMeteoIO module has been updated to also look for meteoio in "${SRC}/../../lib" so the examples can find the local meteoio.

the parameters mapping has been extended to cover access to the imis stations through GSN. Some units thransforms were not being performs.

Improving the help text (the timer now has at least 1 us resolution and up to the 1 ns, depending on the platform). Fixed a wrong commit for meteo_reading.

A rounding bug has been found in the Date class: when using a Date object in a loop where it gets incremented, it accumulates errors and after ~70000 iterations, the error would get bigger than one second, making date comparisons with a fixed date impossible. This has been solved by always rounding the internal gmt Julian date to the closest second, after each assignment (from the constructors, the setDate() calls or any arithmetic operation). Moreover, the rounding methods have been re-written to be more efficient as well as other roundings (specially when computing the date decomposition from Julian). 

The seek() call performing a binary search has also been simplified, in order to remove unnecessary tests, put some similar tests together and make the code clearer. 

Some constification and better vector usage took place in ResamplingAlgorithms and Meteo1DInterpolator.

Putting the code into shape for major changes in the spatial interpolations as well as pushing the minor version nunmber up

Lots of vectors copies were made using push_back(). This is definitely not efficient! This has been replaced by vec_o = vec_i, we will see which impact it has on performances...

Porting Nander's implementation of Thomas algorithm (solver for tridiagonal matrix) to the matrix class

Still tracking a stupid bug somewhere between meteoio, snowpack and alpine3d... Trying to remove some unecessary stuff

fixing some of the warnings found on Hera (sorry for not noticing before). All the size_t / unsigned int stuff will have to be addressed later

The A3DIO plugin was not properly reading data, this has been fixed. Some warnings have been turned off for osX (signedness and 32/64 bits because of size_t since we will have to redo all our indices to make they consistently rely on size_t)

Trying to get rid of a pointer bug in A3DIO: the ifstream object does not need to be a class member (ie reducing its scope), simplifying some calls...

My previous commit was terrible... it simply made it impossible to define optional keys. This has now been fixed but required a change in the API. So, don't use Config::nothrow but IOUtils::nothrow instead.

The "nothrow" option was not properly forwarded in Config, leading to unnecessary throw/catch. The sun.cc example was set to executable.

Fixing the indentation in libsmet and fixing the (very annoying) warnings in MathOptim.h. The cubic root optimizations have been tested and work as advertized (~3 times faster, largest relative error 1.2e-4).

This is a fix for Snowpack issue 291 : on Windows, when the file would stop right after [DATA} (ie without even a new line), the data section pointer would be invalid (set to -1) but a getline on it would not set eof and therefore the reading loop would never end... So if the data section pointer is -1, we simply return (there is obviously nothing to read).

In order to try to be closer to the spirit of the STL, now the empty strings/vectors/lists are tested with .empty(), the first/last elements of vectors accessed with front()/back(), the empty strings are not contructed with string("") but with string(), etc This makes a benchmark based on meteo_reading ~4% faster and this makes the code suprisingly easier to read. 

Some extra constification took place as well as some minor code cleanup (optimizing the comparison between two vectors in order to perform less tests, etc).

The Date class can now properly generate and parse full ISO timestamps, including time zone information. The full specification is now supported by convertString() with a helper method in the Date class. Ouputing an ISO string with its time zone is done by requesting the ISO_TZ format.

Now, when libsmet attemps to write a station that does not contain any data, it will complain but will not throw an exception (otherwise when working with large numbers of stations, the odds of getting an exception are too high). An implementation for fast cubic root is now part of Optim, but still has to be tested. Some code cleanup (using .empty(), using copy constructor instead of construct+copy) occured in IOUtils.

Truncating the path+file down to file only for exceptions was not working on Windows (since the path delimiter is '\' instead of '/')

Lots of small fixes after running cppcheck on Windows: unused variables, use ++i instead of i++ for complex types (ie mostly iterators for us) for more performance, a few variables that could be passed as const&, use vector.empty() instead of vector.size==0 (for clarity and potentially performances).

The plugins used to keep a reference to a config object as provided by IOManager. But this was triggering some warnings (reference to an object that might disappear outside of the constructor), therefore it has been replaced by a const copy (it is not yet 100% sure if this is really better, but at least the compilers won't complain).

When a user was provinding IMISIO with a non-existing station name, the error messages were quite cryptic. In order to properly tell the user what happened, the code has been slightly reorganized. Some code clean up also took place in IMISIO.