Jeremy Smith wrote:
>It wasn't a slot object
>- it was a structure!
Using C-PTR etc. will convert a foreign struct into a Lisp object, and not reference foreign memory anymore afterwards. If you don't like that or prefer a style closer to C, use (c-pointer surface) instead and work with pointers, dereference them etc. as needed.
Both usages have pros and cons. You choose.
For the SDL event type, I'd think the conversion to Lisp looks nicer. OTOH, you may not want to convert...
FWIW, I've updated my UFFI wrappers (added 3 comments)
and added 2 patches to the CL-SDL sourceforge patches section. http://sourceforge.net/tracker/?group_id=44704&atid=440619
It shows how I managed to get cl-sdl running with CLISP.
I'd like to hear if it also works on Win32 (either cygwin or native).
W.r.t. the state of UFFI+CLISP, there are still some design issues were UFFI is ill-defined w.r.t. nested structures. As a result, UFFI still does not natively support CLISP. My patches are just good enough to use a few libraries, which all exhibit a very regular behaviour, but they don't pass the whole UFFI testsuite.