yade-dev team mailing list archive

[Jérôme Duriez <jerome.duriez@xxxxxxxxxxxxxxx>]

Hello,

Thanks for comments, second commit (*) should solve your remarks 1 and 5.

For 2, I will avoid to answer saying that I do not use REC_STRESS, so I 
do not know what should be the best name... (your "C" corresponds to 
"Concrete" ? or something completely different ?)

For 3, indeed the full matrix is not available (see also below)

For 4, I get your idea, but at the moment I do not know how to store 
matrices in vtkDoubleArray... I even have the feeling, according to 
vtkDoubleArray.h that this is not possible ? (And at least, there is no 
other example than scalar and tuple in current code, no ?)

(*) : 
https://github.com/yade/trunk/commit/ee881f8cae6551ae8549de0132a9f74d5ab55576


Le 16/07/2014 14:04, Bruno Chareyre a écrit :
> Thanks Jerome, it was the good way to proceed I think. I have a few
> comments/suggestions:
> 1- Please don't use french names for variables, even in the c++ code
> (valPropres).
> 2- We should now remame REC_STRESS -> REC_CSTRESS, no?
> 3- it seems the most used data (matrix in x,y,z) is still not available,
> isn't it possible to include it also the output in the form of a matrix
> (maybe with a separate flag are all together if BSTRESS)?
> 4- You can output eigen values and directions in the form of vector3 +
> matrix, it would save a lot of lines of code, then in paraview you can
> always select individual components.
> 5- For braces, we tend to use this when possible:
>
> section {
>      bla;}
>
> instead of
>
> section
> {
>      bla;
> }
>
> Combining 3 and 4, your 65 lines (466-530) could fit in maybe 10 lines,
> which will be a lot easier to browse and read.
>
> Bruno
>
>
>
>
> On 16/07/14 13:47, Jérôme Duriez wrote:
> > Question #251712 on Yade changed:
> > https://answers.launchpad.net/yade/+question/251712
> >
> > Jérôme Duriez posted a new comment:
> > Hi,
> >
> > See
> > https://github.com/yade/trunk/commit/939e3d771bd6978da2bb16833a535eebd4eb6043
> > for this commit.
> >
> > It includes a new recorder (yet another one..) in VTKRecorder:
> > "bstresses" to generate scalar data for principal stresses, and
> > arraydata for principal directions, in "sphere...s" VTK files. See the
> > doc.
> >
> > Tell me if something is wrong...

--
Jérôme Duriez
Post-Doctorant UJF
Laboratoire 3SR
Bureau E139 - 04.56.52.86.30