Re: [deal.II] Bug? - FETools::extrapolate changes behavior in 8.5

2017-04-13 Thread winnifried

Dear Wolfgang, 
sure, I have posted a corresponding example.

Best 
Winni

Am Mittwoch, 12. April 2017 17:54:58 UTC+2 schrieb Wolfgang Bangerth:
>
>
> Hi Winni, 
>
> > when adjusting my own code to deal.ii version 8.5. I noticed a change in 
> > the behavior of the FETools::extrapolate method. 
> > Since I could not find this in the list of changes, I would like to let 
> > you know as this might have been unintended. 
> > 
> > Description: 
> > 
> > Until 8.4.1 FETools::extrapolate has been working with 
> > dealii::BlockVectors while in 8.5 the method returns an error 
> > stating that the size 0 does not match the number of unknowns. 
> > 
> > Details: 
> > 
> > I have traced back its origin (i.e. my input data is correctly scaled). 
> > 
> > The error message is created in line 86 of fe_tools_interpolate.cc 
> > and is apparently caused by the internal vector u3 
> > created in line 1586 of 
> > source/fe/fe_tools_extrapolate.cc 
> > This vector is not resized in line 1587 by the method 
> >   internal::reinit_distributed 
> > while this was working in version 8.4.1 
>
> This looks like a bug. Do you think you could come up with a tiny piece 
> of code (doesn't have to do anything useful other than setting up the 
> data structures and calling the function in question) that demonstrates 
> the problem? That would be very valuable. If so, please open an issue 
> here: 
>https://github.com/dealii/dealii/issues/new 
>
> Thanks 
>   W. 
>
> -- 
>  
> Wolfgang Bangerth  email: bang...@colostate.edu 
>  
> www: http://www.math.colostate.edu/~bangerth/ 
>

-- 
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see 
https://groups.google.com/d/forum/dealii?hl=en
--- 
You received this message because you are subscribed to the Google Groups 
"deal.II User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to dealii+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.


Re: [deal.II] Bug? - FETools::extrapolate changes behavior in 8.5

2017-04-12 Thread Wolfgang Bangerth


Hi Winni,


when adjusting my own code to deal.ii version 8.5. I noticed a change in
the behavior of the FETools::extrapolate method.
Since I could not find this in the list of changes, I would like to let
you know as this might have been unintended.

Description:

Until 8.4.1 FETools::extrapolate has been working with
dealii::BlockVectors while in 8.5 the method returns an error
stating that the size 0 does not match the number of unknowns.

Details:

I have traced back its origin (i.e. my input data is correctly scaled).

The error message is created in line 86 of fe_tools_interpolate.cc
and is apparently caused by the internal vector u3
created in line 1586 of
source/fe/fe_tools_extrapolate.cc
This vector is not resized in line 1587 by the method
  internal::reinit_distributed
while this was working in version 8.4.1


This looks like a bug. Do you think you could come up with a tiny piece 
of code (doesn't have to do anything useful other than setting up the 
data structures and calling the function in question) that demonstrates 
the problem? That would be very valuable. If so, please open an issue here:

  https://github.com/dealii/dealii/issues/new

Thanks
 W.

--

Wolfgang Bangerth  email: bange...@colostate.edu
   www: http://www.math.colostate.edu/~bangerth/

--
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see 
https://groups.google.com/d/forum/dealii?hl=en
--- 
You received this message because you are subscribed to the Google Groups "deal.II User Group" group.

To unsubscribe from this group and stop receiving emails from it, send an email 
to dealii+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.