ODE NAN INF derivs

[Kevin Hobbs]

Brian Gough wrote:
> Kevin Hobbs writes:
>  > Hmm looks like I also have to undo the step by zeroing the other 
>  > derivatives and setting time back.
> 
> Yes, you would need to save the previous state.
> Arguably the framework should do this though.
> 


For my current project saving the state is easy as it only has a few 
variables.  But later I may only have room in memory for one set of 
variables and derivatives.  I am sure this is the case for others but 
their equations probably don't have the valid range problem like the 
ones I inherited.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 252 bytes
Desc: not available
URL: <http://sourceware.org/pipermail/gsl-discuss/attachments/20040413/f020c83c/attachment.sig>