[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [ff3d-users] Newbie question
From: |
Stephane Del Pino |
Subject: |
Re: [ff3d-users] Newbie question |
Date: |
Wed, 21 Jul 2004 13:35:16 +0200 |
Luigi Bardelli wrote:
>
> Hi Stephane,
> thank you, now it works! I have another question (maybe related):
> writing something like:
>
> double A=0.123456789; // constant for
> solve(var) in MESH
> {
> pde(var)
> // dx(dx(var))=0; // ok
> // A*dx(dx(var))=0; // not ok
> // -dx(dx(var))=0; // ok
> // dx(dx(var))*A=0; // not ok
> // dx(dx(A*var))=0; // not ok
> // -dx(dx(A*var))=0; // not ok
> // -dx(A*dx(var))=0; // ok !
> // dx(A*dx(var))=0; // ok !
> }
>
> sometimes gives errors like:
> line 103:syntax error, unexpected "unkown variable": "var" unexpected
>
> If A is a pde unknown the error changes in
> line 103:syntax error, unexpected '*', expecting '=' or '-' or '+': "*"
> unexpected
>
> Is it normal?
Yes !
ff3d is not very clever and does not know that if A is constant
A*dx(dx(u)) = dx(A*dx(u))
In fact, if A is not constant, you need to write it dx(A*dx(u)) to use
the Green formula in order to establish the variationnal formula.
I think this answers the question...
> Another (minor) remark: in docs/Makefile.am the line
> dvips `basename $< .tex`.dvi
> should be changed in something like
> dvips `basename $< .tex`.dvi -o `basename $< .tex`.ps
> otherwise the manual will be printed on the default printer every time you
> compile the lib (I have already two copies...)
Ooops.
I thought this was fixed. Maybe it stayed in my developing version. I
will look at it this evening...
Stephane.