A few extra comments have been added to better understand the flow of data. The "merge" feature was not properly handling all situations, this should now be fixed.

The EXCLUDE/KEEP/MERGE keys (as well as the associated stationIDs) were case sensitive and this was making trouble. This is now case-insensitive. A warning flag has been changed for Clang to prevent reporting fake warnings (the flag does not make sense at all and only exists for compatibility reasons).

The MERGE feature has been much improved: the syntax now follows the (more logical) EXCLUDE/KEEP syntax: 
*WFJ::MERGE = WFJ2 WFJ1
The documentation and debug toString() have been updated.

A new "raw data editing" method has been implemented: merge_by_name. Stations sharing a common name can be merged together with the MERGE_BY_NAME key in the [Input] section. The documentation has been edited to bring visibility to the "raw data editing" features.

Micro optimizations for Sun.cc, converted an exception into a warning in ImisIO and fixing case in message in Smetio

Swapping the order of the tests hoping this could solve the issue of the "SMET plugin not found" when testing Snowpack (ie maybe the coverage compile breaks some of the plugins)

This was quite a nasty bug: in some cases, a point that could not be gridified would be re-tried, resulting in a valid point although the vector as a whole signaled that it had invalid entries. Way too many hours for that...

A bug/inconsistency has been fixed in the Winstral algorithm: when pixels are excluded because their air temperature is too high for snow, they should not contribute to the min/max erosion/deposition! A new version is now in dveelopment that takes into account the local wind field as well as a threshold on the wind speed. This is not documented as it is still highly experiemental.

Better documentation for the USER interpolation method (grids are read from GRID2DPATH or a provided sub-dir) and better info message

Fixed a bug in gridify(): cells exactly on the upper/right border where not marked as invalid (thanks Aurélien!)

Code cleanup and fix for IMIS: now the ENET stations receive the number 1 for the automatic stations instead of 0.

Moved <unistd.h> within an ifdef, removed flto for gcc on Windows (since this is currently very, very buggy on Windows).

By simplifying MeteoData, the memory consumption of a benchmark such as data_converter is reduced by almost 20%

very minor code cleanup, some comments were wrong and have been fixed, a wrong debug message has been fixed (in toString())

the USER interpolation method should now work as expected. A new method "getParameterIndex()" has been written for MeteoGrids for a future use.

Clang tries to check if doxygen comments are correct, but it does not understand quite a big chunk of them... This is quite stupid, so turning off these warnings

Better output format and move a few lines from a wrong location to a right one... (ie fixed a bug in setProj)

Better handling of the i,j,k index in Coords objects, better strategy for handling invalid POI (the invalid points can be marked as invalid but kept in the vector)

Big cleanup on the Coords: all the coordinate systems support methods have been moved to a new CoordsAlgorithms static class. The "printlatlon()" method has been replaced by toString(Coords::LATLON).

Add a few string formats options to Coords for easier use in error messages; minor code cleanup in Date; add a new argument to gridify() in order to be able to retrieve the list of rejected coordinates.

Making the grid operators "const" when possible. Type casting in the << and >> operators of the Arrays to please clang...