gmsh-bug in WIN98/NT?

Andre Haufe haufe at schnurzel.net
Mon Mar 26 18:41:17 CEST 2001


Hi Christophe,

> Hello,
> 
> I've just loaded your sample mesh file and my version of gmsh correctly
> displays the 4 points in the mesh (that's all what should be displayed
> for line elements in the current version). What we already experienced
> in the OpenGL implementation on Windows is a bug which causes points to
> be displayed as single pixels (instead of 5x5 pixel rectangles). What
> happens when you simply display one of the examples in the tutorial? Do
> you see the points in this case?
no, it wouldn't display any points, if I try to switch the point
numbering on it crashes sometimes, but will eventually not display 
any information.
 
> 
> Yes, it's also one of the problems for us: the Windows version is still
> very new (we all have a Unix background, and we still discover Windows
> particularities quite often...), and exits too often without leaving a
> message (that's a bad Unix habit, where you always have the terminal
> window to see what happened, especially for syntax errors in the input
> files which cause the input data to become invalid). Nevertheless, all
> messages should be logged in a file called ".gmsh-errors". Is this file
> correctly created in your case?
It is created but but it does not contain the information what went
wrong.
I guess windows will just not flush the buffer before the application
crashes.

I now use the .geo-Format which works well so far. What are the
limitations
when I also put postprocessing informtion into this file according to
this
sample:

---snip--- 
 /* CARAT Design-Output for Gmsh  */
 
 /* Characteristic Length  */
 mtr = 1.0 ;
 cm  = 0.01 ;
 lz  = mtr ;
Point(   1)={      .000000,      .000000,      .000000 ,lz} ;
Point(   2)={     1.000000,      .000000,      .000000 ,lz} ;
Point(   3)={     1.000000,     1.000000,      .000000 ,lz} ;
Point(   4)={      .000000,     1.000000,      .000000 ,lz} ;
Point(   5)={      .000000,     1.000000,     1.000000 ,lz} ;
Point(   6)={     1.000000,     1.000000,     1.000000 ,lz} ;
 
Line(   1)={    1,    2 } ;
Line(   2)={    2,    3 } ;
Line(   3)={    3,    4 } ;
Line(   4)={    4,    1 } ;
Line(   5)={    4,    5 } ;
Line(   6)={    5,    6 } ;
Line(   7)={    6,    3 } ;
 
Line Loop(   8)={   1,   2,   3,   4 } ;
Plane Surface(   9)={   8 } ;
Line Loop(  10)={   3,   5,   6,   7 } ;
Plane Surface(  11)={  10 } ;


View "gagatest" 
{
ST   (  .000000,   .000000,  .000000,  1.000000,      .000000,     
.000000 , 1.000000,     1.000000,      .000000 ) {1.0, 2.0, 0.5} ;
ST   ( 0.000000,   .000000,  .000000 , 1.000000,     1.000000,     
.000000 , .000000,     1.000000,      .000000) {2.0, 0.5, 4} ;
SP    ( 1.000000,      .000000,      .000000 ) {2} ;
SP    ( 1.000000,     1.000000,      .000000 ) {3} ;
SP    (  .000000,     1.000000,      .000000 ) {3} ;
SP    (  .000000,     1.000000,     1.000000 ) {4} ;
SP    ( 1.000000,     1.000000,     1.000000 ) {1} ;

} ;
---snip--- 

Is it just slow? Or why should I use the Post-Processing format?

Thanks!

Andre
 
> 
> --
> Christophe Geuzaine
> 
> Tel: 32 (0) 4 366 37 10    http://geuz.org
> Fax: 32 (0) 4 366 29 10    mailto:Christophe.Geuzaine at ulg.ac.be

-- 
================================
Andre Haufe
Visiting Scholar at the
Department of Civil Engineering
University of Calgary
2500 University Drive N.W.
Calgary, Alberta, T2N 1N4
CANADA

http://www.ucalgary.ca
haufe at schnurzel.net
http://www.schnurzel.net
================================