The commit (in Gitlab) 982dc871 introduced a bug with colormaps -- I found that the fill() function was not populating my data array correctly, and after digging into the code, it appears as though the requested fill value is being passed as the second argument to the memset() function. The C library definition of this function though is expecting the second argument to be an "int" (though interpreted as a byte), but the value passed is a "double", and the internal storage for mData is an array of doubles. So using memset() for this function is going to effectively set the data to garbage. When I revert to the for() loop prior to this commit, the fill() function works properly again.