Perl Backend - UPDATE

[reinhard.e.voglmaier@gsk.com]

This just a little update on how work on
the Perl Backend is proceeding.
In the meantime I worked out how the backend
is plugged in the backend logic. I also tested a little bit possible solutions
of passing structure into Perl objects. Well, the solution I would prefer
in this moment is to have a Perl Object defined in Perl and call the constructor
of this object from C. The data contained in the C-struct ( struct Operation
being the interesting one ) will be moved into Perl using the stack. This
does mean that C pushes the data onto the stack , meanwhile Perl pops it
off. Same mecchanism the data is sent back. This also illustrates one of
the issues embedding a language into the backend. You have to move the
data from the C language space into the space of the other language, since
the way of holding the data will not be the same. One could also move binary
data, but it has to be seen if this would more performant.

The first interesting data is the connection
id . This will first be a proof of concept of the way of passing data and
using object in the perl space. The new object in Perl space will be the
connection object. So one object is created at the start of the server,
that's the PerlBackend object, The other types of objects will be created
with every incoming connection ( bind ).

some comments ??? I would also be interested
in knowing which data would be useful to have in Perl space.

cheers

reinhard