Derek Gaston wrote:

> As for the Jacobian in parallel... let's call not_implemented().  I  
> would rather get an error than get completely different behavior and  
> not know why.

I hate to butt in (well, I try to suppress my enjoyment of butting in, 
anyway), but:

Would you handle the missing capability in two places?  First, if you 
put not_implemented() for more than one processor in the library code, 
that will protect anyone else trying to use the new code under 
conditions it's not ready for yet - like you say, better to get an 
outright error than something that's subtlely wrong.  And then second, 
if you put a cerr message plus "return 0" in the example C file itself, 
that'll prevent the example from ever hitting the "die after throwing an 
exception" code in not_implemented().  We try to have the examples 
return 0 (success) for cases where it's known they won't work - even ex0 
tests for ENABLE_AMR and just exits with a nice explanation if AMR is 
turned off.  If this is going to be a new official example (which I 
think it should be) I'd like to know that my habitual "make 
run_examples" with LIBMESH_RUN='mpirun -np 4' won't be terminated by it.
---
Roy

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Libmesh-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/libmesh-devel

Reply via email to