dolfin team mailing list archive

["Martin Sandve Alnæs" <martinal@xxxxxxxxx>]

On Tue, Jan 20, 2009 at 7:18 PM, Anders Logg <logg@xxxxxxxxx> wrote:
> On Tue, Jan 20, 2009 at 04:55:51PM +0100, Mehdi Nikbakht wrote:
>> Hi,
>>
>> I am trying to know how the new wrappers for Dolfin are generated.
>>
>> Suppose this simple form.
>>
>> element = FiniteElement("Lagrange", "triangle", 1)
>>
>> v = TestFunction(element)
>>
>> L = v*dx
>>
>> If we compile this form file, the wrappers would contain these classes.
>>
>>
>> // DOLFIN wrappers
>>
>> class testLinearFormFunctionSpace0 : public dolfin::FunctionSpace
>> {
>> ---
>> };
>>
>> class testTestSpace : public dolfin::FunctionSpace
>> {
>> ---
>> };
>>
>> class testTrialSpace : public dolfin::FunctionSpace
>> {
>> ---
>> };
>>
>> class testFunctionSpace : public dolfin::FunctionSpace
>> {
>> ---
>> };
>>
>> class testLinearForm : public dolfin::Form
>> {
>> ---
>> };
>>
>>
>> For the first four classes derived from dolfin::FunctionSpace, we have exactly
>> the same body.
>> Aren't they redundant?
>>
>> For example, There is no TrialFunction in the form, but we have a class called
>> testTrialSpace in Wrapper.
>>
>> What are the difference between testLinearFormFunctionSpace0 and
>> testFunctionSpace classes?
>>
>>
>> Thank you in advance,
>> Mehdi
>
> Yes, some redundant spaces may be created, but it doesn't matter.
> Use whichever one you prefer.
>
> For example, there will always be classes named
>
>  PrefixTestSpace    (same as PrefixFunctionSpace0)
>  PrefixTrialSpace   (same as PrefixFunctionSpace1)
>
> and possibly also classes for coefficient function spaces.
>
> If all these should happen to be the same, then another class is generated:
>
>  PrefixFunctionSpace
>
> --
> Anders

In the dolfin wrappers generated by the dolfin_utils/ module,
typedefs are used instead for these additional function space
names, producing a bit smaller code, but as you say it doesn't
really matter.

Martin