dolfin team mailing list archive

[Anders Logg <logg@xxxxxxxxx>]

On Mon, Dec 29, 2008 at 08:15:17AM +0000, Garth N. Wells wrote:
> 
> 
> Anders Logg wrote:
> > On Sun, Dec 28, 2008 at 10:51:16AM +0100, Johan Hake wrote:
> >> On Saturday 27 December 2008 23:32:37 Anders Logg wrote:
> >>> The three broken demos (advection-diffusion, and the two nonlinear
> >>> ones) have now been fixed, but three "new" demos are now broken:
> >>>
> >>>   demo/mesh/intersection/python
> >> This one uses the plot instance directly. When running the tests plotting is 
> >> disabled, and the plot() command returns None, causing an attribute error.
> >>
> >>>   demo/function/eval/python
> >> Fine for me.
> >>
> >>>   demo/function/projection-interpolation/python
> >> I think this one is the only that is really broken. I have looked into it but 
> >> failed to find the actuall error. It bail out at the end of the demo if the 
> >> interpolate function has been called. I suspect that it is some memory that 
> >> has not been released.
> >>
> >> Johan
> >>
> >>> If anyone knows why please fix it.
> > 
> > All demos (C++ and Python) run now after a fix to the plotting and
> > running instant-clean.
> > 
> > Are there any outstanding issues that must be fixed before a release
> > of 0.9.0?
> > 
> 
> Assembler::assemble_system needs to be tested. A problem was reported 
> some time ago. We should also look to break up the function 
> Assembler::assemble_system before a release.
> 
> Garth

Yes, this is desirable. But could it wait until after the release?
It will take some effort/time to split up assemble_system and it's not
like I'm very eager to do it now.

It might be enough to test it?

-- 
Anders

Attachment: signature.asc
Description: Digital signature