Re: Polymorphic GOData

[Emmanuel Pacaud <emmanuel pacaud lapp in2p3 fr>]

Hi,

Le mercredi 22 avril 2009 à 19:38 -0400, Morten Welinder a écrit : 
> get_n_sizes seems to me is returning the number
> of coordinates, i.e., what we would normally call
> the dimension.  "get_n_dimensions", perhaps.

At first, I have used get_n_dimensions, but changed it to get_n_sizes
for consistency with the sizes array. But I can put back
get_n_dimensions.

> go_data_get_string refers to "positions".  They seem
> to be what would otherwise be called "coordinates".

Ok for coordinates.

> Why it is allowed to specify fewer coordinates than
> the underlying class requires?  Same question with
> more coordinates.

The idea was to let behave a vector like a matrix with 1 row, or to
access the first row of a matrix like a vector. But it's probably not a
good idea. And that would make the code simpler if we don't allow this
behaviour.

> Why does _data_scalar_get_n_sizes return 1?  I would
> have expected 0, i.e., one less than _vector_.

A mistake.

> Why the initial "_" on lots of the names?  They are static
> and thus not visible outside anyway.

There was a conflict between a virtual function implementation and an
already existing function (go_data_matrix_get_values). I've opted for
the removal of the go prefix for all virtual functions. But I can also
change that.

        Thanks for the review.

                Emmanuel.

-- 
Emmanuel PACAUD
Groupe Virgo
LAPP - Laboratoire d'Annecy-le-Vieux de physique des particules
9, chemin de Bellevue - BP 110
74941 Annecy-le-Vieux CEDEX
France
Tél: (+33)4 50 09 55 92 ou 17 89 - Fax: (+33)4 50 27 94 95