Re: Mesh, Octave and Freefem++

[fotios]

Στις 2013-06-01 14:27, ο/η c. έγραψε:
> It looks
> On 1 Jun 2013, at 14:22, address@hidden wrote:
>
> > Message: 1
> > Date: Sat, 01 Jun 2013 13:06:20 +0200
> > From: fotios <address@hidden>
> > To: address@hidden
> > Subject: Mesh, Octave and Freefem++
> > Message-ID: <address@hidden>
> > Content-Type: text/plain; charset="iso-8859-1"; Format="flowed"
> >
> > Once I wrote some simple interface to generate a mesh through Octave
> > using Freefem++ [1], after reordering my code I ended up with the
> > attached (non documented) functions and here is a simple example
> Fotios,
>
> This sounds like a valuable addition to the msh package.
> Would you mind if I adapt it and commit it there?
>
>
> c.

Not at all! Free to do as you desire. There are some things to comment 
there though. In buildmesh.m line 53 till end (or sth like that) change 
code to the following (so that the loop is avoided)

[p, t, e] = readmesh ();
p = p(:, 1:2);
t = t(:, 1:3);
e = e(:, 1:3);
p1 = p(t(:, 1), :);
p2 = p(t(:, 2), :);
p3 = p(t(:, 3), :);
d12 = sqrt (sum ((p1 - p2).^2, 2));
d23 = sqrt (sum ((p2 - p3).^2, 2));
d31 = sqrt (sum ((p3 - p1).^2, 2));

hmax = max ([max(d12), max(d23), max(d31)]); # or hmax = max ([d12; d23; 
d31])

Moreover, in case there is an error rm ff2O*.* will never be riched and 
next mesh creation will yield an error since the ff2O_bound file will 
not be empty (it is opened with a+ option) . . . my (non-intelligent) 
solution to that is to have a function cleanff2O.m that must be called 
before the first call to boundary.m . . . free to do it in a clever way 
;)). Depends on Freefem++ with executable Freefem++ (maybe make a check 
test for that regarding different platforms?) . . . works on my win8 
machine with latest freefem. Enjoy!

/Fotis