On 4/2/07, Alex Queiroz <[EMAIL PROTECTED]> wrote:
I'm talking about writing code to be require'd by a Chicken app,
in pure C. A "pure C module." :-)
OK. Then, back to your statement:
I guess there are so many ffi choices because it must be a pain
to write pure C extensions, judged by what I've seen from chicken's
output.
No, I don't think it follows that there are many FFI choices because
writing pure C extensions is a pain.
But I think that's inevitable given how Chicken uses the machine stack.
I'm no expert, but here's my take. The C output of the Chicken
compiler may look unusual, and Cheney-on-the-MTA design could be
considered a perversion of the "intent" of the machine stack. But as
Baker wrote, the point of the design was "to compile Scheme by
converting it into continuation-passing style (CPS), and then compile
the resulting lambda expressions into individual C functions.
Arguments are passed as normal C arguments, and function calls are
normal C calls." In other words, mapping Scheme onto C stack
semantics. The only freaky bit is the "nothing ever returns" clause,
but the rest is good old C.
Don't judge the complexity of pure C modules by the scary appearance
of CSC-generated, CPS-mangled C code. Instead, write a pure C module
or two, in a natural style, and find out for yourself that it's not
that hard at all.
Graham
_______________________________________________
Chicken-users mailing list
[email protected]
http://lists.nongnu.org/mailman/listinfo/chicken-users
