#2423 Local weather: setprop() passed a NaN in local_weather/weather_tiles.nas, line 2795

When testing FG stability with the UFO I found a (probably non-fatal) error in local weather.

  846.47 [ALRT]:nasal      Nasal runtime error: setprop() passed a NaN
  846.47 [ALRT]:nasal        at /home/mike/src/git/flightgear/fgdata/Nasal/local_weather/weather_tiles.nas, line 2795
  846.47 [ALRT]:nasal        called from: /home/mike/src/git/flightgear/fgdata/Nasal/local_weather/weather_tile_management.nas, line 77

It is likely caused by this line:

                var gust_relative_strength = (gust_strength - windspeed)/windspeed;

if windspeed is very small.

[James Turner]

Good catch. Will ping Thorsten for comment.

[James Turner]

Over to ThorstenR for his opinion.

[xDraconian]

Related to ticket 2425.
Resolved on branch next

[Anonymous]

The bug still appears on next, same line numbers:

  835.74 [ALRT]:nasal      Nasal runtime error: setprop() passed a NaN
  835.74 [ALRT]:nasal        at /home/mike/src/git/flightgear/fgdata/Nasal/local_weather/weather_tiles.nas, line 2795
  835.74 [ALRT]:nasal        called from: /home/mike/src/git/flightgear/fgdata/Nasal/local_weather/weather_tile_management.nas, line 77

[Anonymous]

Why does it rely on windspeed being non-zero in var gust_relative_strength = (gust_strength - windspeed)/windspeed;? Why not be more defensive and have if(abs(windspeed) < smallnumber) gust_relative_strength=0.0?

[Gijs]

@trenk, I think setting the gust_relative_strength to zero when the windspeed is zero should be safe? As far I as I understand it, further down the route the actual wind is calculated as follows, which would output to zero anyway for a 0kt base wind:

windspeed_multiplier =  (1.0 + (random_factor * gust_relative_strength * alt_fact));
windspeed_current = windspeed_current *  windspeed_multiplier

This does of course leave the question whether a METAR reporting zero wind with non-zero gusts is valid. But FlightGear should not crash on that.