C->Scilab: Now I'm lost and need some directions

[Antoine Monmayrant]

Hi all,

I am fighting to call some C code from within Scilab, or more precisely 
to add a primitive to Scilab using some C code I wrote.
I have a working C code and I managed to link it to Scilab by using 
"ilib_for_link".
It works, no problem.
Now I would like to bullet-proof it, that is to get it to work like a 
real scilab primitive.
For example, at the moment, if I try to open a file that do not exist 
(from within my C code), I just kill my Scilab session.
This is due to the way the error is treated in my C code.
I would prefer to get a normal Scilab error: "!--error 241, File does 
not exist".
In the end, I would like to handle all the usual scilab error (wrong 
number of arguments, wrong type, ...) gracefully.
What should I do?
I found bits and pieces here and there about how to link, what the 
interface should be like, how to use call, ... but I can't put the 
pieces together to get a big picture of what I should do.
I found "intersci" but it is obsolete. I found some articles but they 
are outdated ( http://www.saphir-control.fr/articles/lm15/article.html ).
Does anyone know a proper tutorial or *complete* example ( c_source, 
interface_source, build and example of use) on the subject?

Thank you in advance,

Antoine