Closed JC3 closed 8 years ago
Open the .geo file in gmsh first. This will regenerate the mesh file (.msh). (The .msh files aren't included in the repository because they're so huge.)
Once a .msh file has been created for the geometry that you're interested in, you don't need to generate it again (unless you want to change the meshing parameters, of course), and you can use the .pro file from that point on.
Let me know if this works for you.
Thanks. When I do this:
It generates oak_hole.msh, which appears to be empty, consisting only of:
$MeshFormat
2.2 0 8
$EndMeshFormat
$Nodes
0
$EndNodes
$Elements
0
$EndElements
It does the same with all of the other .geo files.
I do see a model in the viewport that I can interact with, if it matters:
It hasn't actually constructed the mesh yet (the gmsh UI isn't the most intuitive thing in the world). Try this:
This should cause the mesh to be generated (you'll see "Meshing 1D..." in the status line). Once it has been generated, which takes about a minute on my machine, GetDP will start processing the data.
Thanks. Sadly, that had the same failed results. "Meshing 1D" is actually the same point that the original set of steps fail at, so it must be a getdp problem not a gmsh problem. In both cases the log finishes with:
Info : Done reading 'C:\Programs\woodemc\lap_water-master\models\oak_hole.geo'
Info : Meshing 1D...
Info : Meshing curve 101 (Line)
ANN: ERROR------->Requesting more near neighbors than data points<-------------ERROR
Maybe something in gmsh/getdp changed in a version since you originally made the models. Do you recall what versions of the programs you used when you had it working last?
Hmm. Meshing is a gmsh process. I'm using a slightly older version of gmsh; let me try with the newer version and see if I can replicate the results.
I was able to reproduce the problem with gmsh 2.11.0. Of course, I don't know why it's happening, but I'll investigate...
It looks like the error message is coming from some code in this library: [https://www.cs.umd.edu/~mount/ANN/], but that's all I know at the moment.
For the time being, I suggest reverting to earlier versions of gmsh/getdp. A combination that works for me is gmsh 2.9.3 and getdp 2.5.1. You can download installers for the former from [http://gmsh.info/bin/Windows/] and the latter from [http://getdp.info/bin/Windows/]. (As long as you install to a different location, you don't need to uninstall the other versions.)
There's a bit of a trick to getting gmsh to "forget" its connection to the existing version of getdp:
Thanks. No need for you to spend your time debugging this futher. From the gmsh mail list just now:
From: Christophe Geuzaine Date: Wed, Feb 3, 2016 at 11:58 AM Subject: Re: [Gmsh] Gmsh exits immediately after run, "Requesting more near neighbors than data points"
Jason - Thanks for the report; this is indeed a regression, that I just fixed in SVN r22568.
Christophe
So at some point presumably a stable build will have this working again. In the mean time I'll try it with the versions you mentioned.
I'm closing the issue. Thanks for all your help, and for the models in the first place. I found it via http://blog.lostartpress.com/2014/03/17/water-vs-wood-grudge-match/ via http://woodworking.stackexchange.com/a/3234. Very enlightening.
Good to hear that Cristophe has it figured out.
Feel free to contact me directly if you have any questions: steve@donacobi.us. About drying wood, that is, not gmsh/getdp. ;-)
-Steve
I have posted this to the gmsh mailing list but am also posting it here.
Using:
When I do the following (this is a fresh install, first time running this software):
Gmsh runs for about 2-3 seconds then silently exits (verified gmsh.exe terminated immediately with task manager). The last message printed to the terminal (I set message verbosity to "100" arbitrarily) is:
ANN: ERROR------->Requesting more near neighbors than data points<-------------ERROR
Partial gmsh debug log is attached: log.txt
How should I proceed?