This code will give different results in desktop/headless than in NetLogo Web:
to in-radius-order-check
clear-all
resize-world -4 4 -4 4
random-seed 123456
crt 30 [ setxy random-xcor random-ycor ]
ask one-of turtles [
show [who] of turtles in-radius 3
]
end
Or, in a form suitable for the Tortoise.txt language tests (results from NetLogo 6.1.1):
InRadiusGivesSameOrder
globals [ check-list ]
O> resize-world -4 4 -4 4
O> random-seed 123456
O> crt 30 [ set color one-of [red yellow white] setxy random-xcor random-ycor ]
O> set check-list []
O> ask one-of turtles [ ask other turtles in-radius 3 [ set check-list lput who check-list ] ]
check-list => [23 21 18 7 11 2]
The likely issue is desktop optimizations for the in-radius prim causing the backing collection order to be different before using of or ask to access the shuffler to get the agents out. But it's strange that this has not caused an issue before in the docking tests for models that use in-radius extensively, like Fireflies or Flocking.
This code will give different results in desktop/headless than in NetLogo Web:
Or, in a form suitable for the Tortoise.txt language tests (results from NetLogo 6.1.1):
The likely issue is desktop optimizations for the
in-radius
prim causing the backing collection order to be different before usingof
orask
to access the shuffler to get the agents out. But it's strange that this has not caused an issue before in the docking tests for models that usein-radius
extensively, like Fireflies or Flocking.