Re: [Pytables-users] PyTables 1.1-beta3

[Francesc A. <fa...@ca...>]

Hi Antonio,

On Saturday 18 June 2005 19:28, Antonio Valentino wrote:
> a really good job ;)

Well, part of it was possible thanks to your efforts ;)

> the test fails on my laptop (see the attatched file).
> It seems that the "itemsize" in the _createEArray method is never
> initialized. This should fix the problem:
>
> --- hdf5Extension.pyx (Id 1015 2005-06-17 17:55:14Z)
> +++ hdf5Extension.pyx
> @@ -1319,6 +1319,7 @@
>      complib =3D PyString_AsString(self.filters.complib)
>      version =3D PyString_AsString(self._v_version)
>      class_ =3D PyString_AsString(self._c_classId)
> +    itemsize =3D atom.itemsize
>      fill_value =3D <void *>malloc(<size_t> itemsize)
>      if(fill_value):
>        for i from  0 <=3D i < itemsize:

Of course. The strange thing is that this bug was not exposed on my
Debian systems (neither sarge nor sid). Fortunately, your Fedora
system revealed it. Well, thanks anyway.

Ad=E9u!

=2D-=20
=46rancesc