| Author | Message | |||
     Adam Thorn (adamthorn) New member Username: adamthorn Post Number: 1 Registered: 02-2008 |
Here's a simple test script I've written, which as you can see just solves the Laplace equation in a box with a layer of charge in the middle (defined via a natural bounday condition). The first time I run the script, it runs fine. If I then go back to the edit window and make a small change (for example, I halve the number in that natural(u) condition, or change the sign of it) and rerun the script then FlexPDE locks up. The "Status" window has the output from the initial run, followed by "--DONE--", followed by "Freeing Memory", but the program is totally unresponsive (none of the buttons, menus etc do anything) I'm running version 5.0.20, Professional version, on a 64 bit Ubuntu machine.
| |||
| Adam Thorn (adamthorn) New member Username: adamthorn Post Number: 2 Registered: 02-2008 |
An additional observation: the problem seems to be with the report(surf_integral...) line. If I comment this out, FlexPDE behaves properly. | |||
| Marek Nelson (mgnelson) Moderator Username: mgnelson Post Number: 22 Registered: 07-2007 |
This is a little strange. We will have to look into what is going on. | |||
| Robert G. Nelson (rgnelson) Moderator Username: rgnelson Post Number: 1064 Registered: 06-2003 |
Turns out not to be too mysterious after all. You have a misplaced closing parenthesis. You have passed three arguments to the NORMAL function, which FlexPDE expects to be the three components of a vector. In fact, they are a grad vector and two strings. This causes some undiagnosed symbolic processing errors and bad memory pointers, which crash. We will add diagnostics to detect this error, but in the meantime, you can simply move your closing paren: report(surf_integral(NORMAL(eps*grad(u)), 'middle', 'domain')) |
