Closed lesquoyb closed 10 months ago
After some investigation I found the part that is probably at fault:
following the stack trace for the NullPointerException in the first message, in the creation of the experiment's population, in GamaPopulation.java
line 280, the function orderAttributes
returns a list composed only of null values, though the code in that function is close to black magic to me so I really don't understand what's going on or what should be happening. @AlexisDrogoul, @hqnghi88 any pointers ?
Can you try this in GamaServerExperimentController, line 165
Seems to work perfectly, I'll push that fix
OK, all the thing there is just to take into account the new value of parameters, which are restructured to already done before here:
I didn't see any problem after some times using it, so closing the issue
Describe the bug In gama-server, it is not possible anymore to control a simulation that has been reloaded, because the experiment id seem to have changed with the reload and there's no indication of what is the new experiment id. Before the recent changes in server we would use the same experiment id as the initial simulation to control the reloaded one.
To Reproduce Steps to reproduce the behavior:
Expected behavior Either we can use the old experiment id as previously, or the new one is sent as a return to the reload command
Additional context This happens on server only mode (no GUI). Additionally it may be caused by an internal error, the reload command send a
CommandExecutedSuccessfully
message, but I noticed errors in the eclipse console:the model I use for example is only an experiment so I don't think it is at fault here:
Update: In GUI, provided we pause long enough between load and reload to avoid exceptions, it works, but the experiment_id is actually the experiment name instead of a number as it used to be. So the problem may come from a mismatch between those two systems