new GlobalStiffnessTimeStepper can not be deactivated anymore in O.engines

[Christian Jakob]

I tried to deactivate GlobalStiffnessTimeStepper in O.engines with

GlobalStiffnessTimeStepper.dead=True

or with

GlobalStiffnessTimeStepper.active=False

, but this commands will not deactivate it.

To solve the problem one just have not to declare it in O.engines.

[Bruno Chareyre (bruno-chareyre)]

I can't confirm this bug. If on my side I turn dead=1 or active=0, then the timestep is not updated any longer.
I turn the bug invalid. Re-open with an axample script if needed.

Do you really type "GlobalStiffnessTimeStepper.dead=True"?
It may be your problem: "className.attr=... " is meaningless in this context.
You want "objectName.attr=..."

[Christian Jakob (jakob-ifgt)]

I provide more information on this topic, maybe its really just a script syntax problem ...

O.engines=[
 ForceResetter(),
 InsertionSortCollider([Bo1_Sphere_Aabb(),Bo1_Facet_Aabb(),Bo1_Box_Aabb()]),
 InteractionLoop(
  [Ig2_Sphere_Sphere_ScGeom(),Ig2_Facet_Sphere_ScGeom(),Ig2_Box_Sphere_ScGeom()],
  [Ip2_FrictMat_FrictMat_MindlinCapillaryPhys(betan=0.9,betas=0.9,label='ContactModel')],
  [Law2_ScGeom_MindlinPhys_Mindlin(neverErase=False,label='Mindlin')]
  ),
 Law2_ScGeom_CapillaryPhys_Capillarity(CapillaryPressure=suction,fusionDetection=True,label='CapPhys'),
 GlobalStiffnessTimeStepper(active=1,timeStepUpdateInterval=1,timestepSafetyCoefficient=0.8, defaultDt=0.9*utils.PWaveTimeStep(),label='TimeStepper'),
 NewtonIntegrator(gravity=(0,0,-9.81),damping=local_damping_pre,label='integrator')
]

#Now I want to deactivate GlobalStiffnessTimeStepper (activate later)

TimeStepper.dead=True

#Set preferred time step:

O.dt=1e-8

When I look at time step in yade GUI, time stepper is still active and it has time step > 1e-8 !
Did I do something wrong? The script was working well with an older version of yade (before density scaling was introduced). Now it makes calculation instable (explosion), because time step is too high.

[Bruno Chareyre (bruno-chareyre)]

We can't test anything with this piece of code. Please send a working script.

Anyway, naming an object "TimeStepper" is ringing a bell: the object as the name of a class.
Who knows if python will interpret it at the class or the object in TimeStepper.dead=True?...

[Bruno Chareyre (bruno-chareyre)]

Never name an object with an uppercase, if you want to avoid this kind of problem.

[Christian Jakob (jakob-ifgt)]

Aaahh, thank you very much Bruno.
Problem solved with using label = 'ts' instead of label = 'TimeStepper'.

> Never name an object with an uppercase, if you want to avoid this kind of problem.

ok, got it.