yade-dev team mailing list archive

[Václav S(milauer <eudoxos@xxxxxxxx>]

> I upgraded yade from r1386 to last revision, and I see a bug has been 
> introduced (sorry, I can't read all commits and I don't know when it 
> could have been introduced). To see the bug : generate a default 
> TriaxialTest and run,  walls will go through the sample without moving 
> spheres. From saved xml, I see that no interactions are detected (listed 
> interactions all have isReal=false).
> The same procedure gives a good result with r1386.
>   
Ah, that's been a long time. I am investigating that now, though I am
not aware of any change that could have introduced that.
> Another boring thing is that the timestep computation will always switch 
> off alone with the QtGUI, so that I need to click "auto" all the time 
> when I start-stop-load simulations. Is it possible to revert to old 
> behaviour?
>   
The point was that if you save the simulation to XML, you should get the
exact same state when you reload and run, including timestepper
settings. If you activate timestepper in the preprocessor, it will be
turned on when you load the xml.

Vaclav
_______________________________________________
yade-dev mailing list
yade-dev@xxxxxxxxxxxxxxxx
https://lists.berlios.de/mailman/listinfo/yade-dev