Eduardo Cavazos <[email protected]> writes:

> Hello,
>
> I've had some xlib bindings laying around for a while, but they weren't
> packaged as a proper R6RS library. I recently revisited this. They're
> now available as a library '(psilab xlib ffi)' in this repository:
>
>     http://github.com/dharmatech/psilab
>
> I say this message is semi off-topic because that code uses the Ypsilon
> FFI. However, if Ikarus's '(ypsilon-compat)' library is someday extended
> to include the full API, hopefully the xlib library will work on Ikarus
> as well.
>
> [...]
>
> In 'cons-wm', I'm also using Ypsilon's library '(ypsilon process)', so
> an eventual port to Ikarus will also need to abstract over that.
>
I thought I'd mention at this place that spells[0] offers an FFI (in
`(spells foreign)' that abstracts over Ikarus and Ypsilon (PLT is in the
pipelin), but it is quite low-level. A useful higher-level API targetted
at "static" (hand-written) bindings could be built on top of it -- I
didn't do so since I use it mainly for sbank, which, besides a tiny core
set of procedures, does "dynamic" bindings (i.e. those where the ABI
information is consulted at runtime or expand-time, instead of
binding-writing-time).

Spells also provides a process interface, `(spells process)', which also
works on Ypsilon and Ikarus (again, PLT is in-the-works).

</advertisement> ;-)

Regards, Rotty
-- 
Andreas Rottmann -- <http://rotty.yi.org/>

Reply via email to