Finally!! The virtual stations seem to be in proper order! Now the spatial interpolations are only called every vstations_refresh_rate and otherwise the data is temporally resampled.

Finally, the proper handling of virtual stations is taking shape: using two tsmanagers, it is possible to temporally resample the data between spatially interpolated time steps. The strategy to decide which timesteps should be temporally resampled is still buggy, this will have to be improved (currently, it resamples and then spatially interpolates at time when it should resample the spatial interpolations instead).

The Buffer class has been expanded with a new push() call and the documentation moved around in IOUtils

The Date::mod() method had some serious rounding errors that could accumulate. This was degrading the precision. The code has been replaced bya more robust implementation that should offer seconds precision (which is what the class offers).

The data_converter code has been cleaned up and made more similar to the snowpack operational "data_qa" tool. An invalid check (< when <= was needed) has been fixed in Meteo1DInterpolator (this was leading to segfaults when a buffer only contained 1 data point).

Code cleanup, documentation improvement (link to WFJ dataset), convenience method in StationData (getAltitude).

A tricky bug was found (thanks Leonardo for giving me a test case showing the bug!) when copying data from the buffer back into a vector: the "insert" call from the STL does the copy in the range [start, end[ so we were always missing the last element. Most of the time, this was not an issue since we had such a large buffer, but with small buffers, one could see that the last data point was missing... So now, the "end" iterator is always "last_required+1".

Following an issue with an wrongly chosen coordinate system, a few improvements have been developed: the documentation has been improved, UTM zones can be provided as {num}{letter} or the opposite, and a bug has been fixed in NetCDFIO (in some cases, the opened file would not be closed).

For better consistency, now using the grid's size() method instead of getNx()*getNy(). The debug messages in libinterpol2D have been removed after performing the checks (ie that our implementation of Winstral does conserve mass).

More includes cleanup and code cleanup in the plugins (getting rid of the "stationindex" thing that was pretty much deprecated in all plugins)

Several bugs have been found in the FULL_MERGE merging strategy. Merging half-hourly data into a vector of 3 hours data was not working at all, now it should work!

The getPointsBetween() methods have been removed from DEMObject since they were not really used and not really reliable either... Instead the sky view factor has been implemented in getCellSkyViewFactor()

A small optimization has been added for the + and * arrays operators (+0 or *1 does not have to be computed!). Code cleanup in CNRMIO.

Fixed two bugs: one member of DEMObject was not properly serialized (for MPI) and the CNRM plugin was not properly declared.

For convenience, the "size()" method has been defined. It returns the allocated number of element so it can be used when iterating over all the cells of an array/grid.

A new method for the spatial interpolations of ISWR has been implemented. Helper methods have been implemented in SunObject and DEMObject (possibility to reset the altitude without recomputing the whole solar position, horizon computation from one point toward a given direction). 
The smet_extract script now writes its output to stdout instead of to a file since the previous behavior was mostly an anoyance...

A little bit of includes cleanup, code cleanup in Sun.cc and added the atmospheric corrections logic (from Alpine3D) to the Sun object so it can easily be reused.

The MERGE feature was still quite buggy, this should now work much better (ie properly handling the "extra" parameters). A warning has been fixed in libsmet. Better handling of julian_date in smet_extract (converting to ISO timestamp in the output).

The "<" and ">" operators have been defined for MeteoData in order to make it easier to sort vectors of MeteoData. The FULL_MERGE merging strategy has been implemented alongside with some code cleanup/restructuring and some code optimizations.

When merging data from different stations, it is now possible to specify (with the Input::MERGE_STRATEGY key) how the merge should be done. The FULL_MERGE is currently not implemented.

Trying to make the CMakeLists clearer and more modular: the compiler options are now set through a macro that will be shared with the other projects. New targets appeared: Intel and Cray, but this is still a work in progress.

Improved an error message and added a note in the documentation that libproj is currently not thread-safe (as well as a little code cleanup)

A bug has been fixed in CosmoXMLIO: it was reading all the data with a time-shift (ie always reading the file *after* the timestep that it should have read).

A few documentation glitches have been fixed in IOInterface and Date and an invalid default value for the time zones fixed in SMETIO (instead of taking 0 as intended, it was always nodata). The missing files for the new OSHD plugin are now committed, although the plugin is NOT READY to be used yet.

A resampling method has been added to the Grid2DObject so it is possible to change the cellsize. The support for the '*' wildcard has been added to the KEEP and EXCLUDE keys.

The MOD operator has been added to Date (it takes a number of seconds in argument). This is used by the new VSTATIONS_REFRESH_RATE key to avoid calling the spatial interpolations too often in the virtual stations. More checks have been implemented in the virtual stations and dowscaling for added safety. 

A major bug has been found when reading 3D grids in ARPS: the indexes when not in the right order.

A bug has been found in NetCDFIO: it should look for the GRID2D key, not the METEO key! In Coords, the lat/lon are checked for validity in setLatLon() in order to catch grossly out of range coordinates (as was found on the Snowpack operational system for a few stations).

A bug has been fixed: there was a confusion between the mean adiabatic lapse rate and the dry one. Small code cleanup in DEMObject (with a better toString() method), ARPSIO and ProcShade.

Now GSN uses the virtual sensor name as station_id, making it much more logical when using exclude/keep/merge statements. The NetCDF plugin does not require anymore to be configured for the inputs in order to produce output grids. Otherwise, code and documentation cleanup.

Code cleanup: more consistency in the handling of keep/exclude/merge. Better, more explicit error messages.

A bug in the MeteoData::mergeTimeSeries() method has been fixed that was related to the handling of extra parameters. The SHADE filter can now compute masks directly from the DEM, but this has required to propagate the Config object all the way down to the filter. Some minor code cleanup and documentation fixes.

Please notice that the SHADE filter is currently only appropriate for a single station! The most probable use case is to merge different stations into a new location (or to downscale to a new location) and properly mask the radiation by the surrounding terrain.

The horizon computation in the DEMObject has been fully redone, copying what is done in Alpine3D. This should be more accurate and much more efficient.