Thank you for your email. Do not worry ... vacations are always important.
I hope you enjoyed them.
In the meantime I tested a solution provided by Dr. David Powell, who
suggested to change the arguments of the call to the method make_mesh()
included in maxwell-pml.py:
mesh = make_mesh(a=numpy.array((-1.0,-1.0)), b=numpy.array((1.0,1.0)),
and this worked for me. However, I've checked the git versions using
your instructions. This is the result:
534c36b,8 months ago,Switch from setuptools to distribute.
b0b70ef,3 months ago,Upgrade numpy to 1.3.0.
d232ecb,9 weeks ago,Minor changes.
1bc6017,2 months ago,Mapper: Reduce IfPositive to AlgebraicLeaf.
46f2ea5,4 months ago,Minor doc updates.
bd9fa47,3 months ago,Revert "Upgrade tetgen to 1.4.3."
Taking advantage of this email, I plan to write a small tutorial or
example, including explanations, of a typical python script for a
Maxwell-pml problem. I am reading your examples and "demos" for
meshpy, etc. I plan to begin with a small structure/problem, i.e.
a 2D ring resonator. My question is: are your demo/test files updated
to the last version of the packages in the repository or it is
recommended to download a stable version to explore the demo/test files?
Thank you in advance for your support.
On 19:48 Sat 19 Jun , Andreas Kloeckner wrote:
First of all, I'd like to apologize for the long silence from my side--I
was gone on some (much-needed) vacation, which left me with very little
time to do work-related things. Aside from a few short trips, my
response time should be much better throughout the rest of the summer.
On Wed, 2 Jun 2010 00:16:23 +0200 (CEST), rgomezal(a)unex.es wrote:
I've been reading the code and doing the
tests. When I began to test the
file maxwell-pml.py I found this problem:
rgomezal@localhost ~ $ python maxwell-pml.py
Traceback (most recent call last):
File "maxwell-pml.py", line 222, in <module>
File "maxwell-pml.py", line 74, in main
mesh = make_mesh(a=numpy.array((-1,-1)), b=numpy.array((1,1)),
File "maxwell-pml.py", line 37, in make_mesh
line 102, in set
line 62, in set_points
self.points[i] = pt
TypeError: No registered converter was able to produce a C++ rvalue of
type double from this Python object of type numpy.int32
Like David, I cannot reproduce this, having tried on a 32- and a 64-bit
machine. Can you post the output of
./repotool for-all -v git log -1 --pretty=format:%h,%ar,%s%n --no-color
to check what versions you are using? PyUblas would be especially
relevant, I think.