[CIG-SHORT] Unexpected output behavior

[Brad Aagaard]

Scott,

Yes, the settings in new_output.cfg will overwrite any corresponding 
settings in step20.cfg.

HOWEVER, the object hierarchy changes when step20.cfg has

output = [domain,subdomain]

and new_output.cfg has

output = [points]

The settings in step20.cfg for "domain" and "subdomain" are no longer 
valid, because output is now [points] and not [domain,subdomain].

ADDITIONALLY, in step20.cfg we use the more specific object name 
[timedependent.implicit] where as in new_output.cfg you used the more 
general [problem.formulation]. Using the more specific object name takes 
precedence (this is subtle and it took me a few minutes to figure this out).

Fixing the second issue is easy. Dealing with the first issue generally 
requires splitting up your .cfg files so you only add to the hierarchy 
and don't subtract from it when you add more .cfg files on the command line.

Regards,
Brad


On 07/30/2014 12:45 PM, Scott Henderson wrote:
> Hello,
>
> I was just trying to modify the output of the example 3d/hex8/step20.cfg to only include points at the surface. It seems like if I create another .cfg file for this purpose, running the following command should work (since .cfg settings take priority from right to left):
>
> pylith step20.cfg new_output.cfg
>
> However, resetting output = [domain, subdomain], to output=[points] doesn’t work. I’m not sure if I’ve missed something obvious, or this is intentional? But it seems a bit counterintuitive. What does work is setting output=[domain,subdomain,points] and explicitly suppressing all outputs defined in step20. My two output.cfg files are attached.
>
> Cheers,
> Scott
>
>
>
> _______________________________________________
> CIG-SHORT mailing list
> CIG-SHORT at geodynamics.org
> http://lists.geodynamics.org/cgi-bin/mailman/listinfo/cig-short
>