Re: [Chicken-users] Re: bytevectors
On Tue, Feb 17, 2009 at 11:33 PM, Eduardo Cavazos wrote: > Here's a screenshot of the fruits of today's labors: > > http://proteus.freeshell.org/abstracting-screenshots/game1-turn6-chicken.png > > That was produced by a pure Scheme implementation of the ContextFreeArt > semantics. More info about ContextFree: > >http://www.contextfreeart.org/ > > The code that rendered it also runs on Larceny and Ypsilon. > > If anybody is curious to see the code behind the program and the framework > supporting it: > >http://github.com/dharmatech/abstracting/tree/master > Wow - very impressive. Well done, Eduardo. The contextfreeart site has a few examples that are really mind-boggling. I have to find some time to play with it. cheers, felix ___ Chicken-users mailing list Chicken-users@nongnu.org http://lists.nongnu.org/mailman/listinfo/chicken-users
Re: [Chicken-users] Re: bytevectors
Eduardo Cavazos wrote: -- (use srfi-4) (define (make-bytevector n) (u8vector->blob/shared (make-u8vector n))) (define bytevector-length blob-size) (define (bytevector-ieee-double-native-ref bv i) (f64vector-ref (blob->f64vector/shared bv) (/ i 8))) (define (bytevector-ieee-double-native-set! bv i val) (f64vector-set! (blob->f64vector/shared bv) (/ i 8) val)) -- felix winkelmann wrote: That looks good. And it should also be the most efficient solution. Cool! Here's a screenshot of the fruits of today's labors: http://proteus.freeshell.org/abstracting-screenshots/game1-turn6-chicken.png That was produced by a pure Scheme implementation of the ContextFreeArt semantics. More info about ContextFree: http://www.contextfreeart.org/ The code that rendered it also runs on Larceny and Ypsilon. If anybody is curious to see the code behind the program and the framework supporting it: http://github.com/dharmatech/abstracting/tree/master Ed ___ Chicken-users mailing list Chicken-users@nongnu.org http://lists.nongnu.org/mailman/listinfo/chicken-users
Re: [Chicken-users] Re: bytevectors
On Tue, Feb 17, 2009 at 10:51 PM, Eduardo Cavazos wrote: > > -- > (use srfi-4) > > (define (make-bytevector n) > (u8vector->blob/shared (make-u8vector n))) > > (define bytevector-length blob-size) > > (define (bytevector-ieee-double-native-ref bv i) > (f64vector-ref (blob->f64vector/shared bv) (/ i 8))) > > (define (bytevector-ieee-double-native-set! bv i val) > (f64vector-set! (blob->f64vector/shared bv) (/ i 8) val)) > -- > That looks good. And it should also be the most efficient solution. chers, felix ___ Chicken-users mailing list Chicken-users@nongnu.org http://lists.nongnu.org/mailman/listinfo/chicken-users
[Chicken-users] Re: bytevectors
Eduardo Cavazos wrote: Hmmm... After poking around some, I see that I can probably cook up a compatability layer. Something along these lines: This is what I ended up using. I loaded into Chicken and it's handling some "R6RS" code of mine. These procedures are in tight loops. Chicken is keeping up with Larceny and Ypsilon so I'm guessing the overhead introduced isn't too bad. ;-) -- (use srfi-4) (define (make-bytevector n) (u8vector->blob/shared (make-u8vector n))) (define bytevector-length blob-size) (define (bytevector-ieee-double-native-ref bv i) (f64vector-ref (blob->f64vector/shared bv) (/ i 8))) (define (bytevector-ieee-double-native-set! bv i val) (f64vector-set! (blob->f64vector/shared bv) (/ i 8) val)) -- Ed ___ Chicken-users mailing list Chicken-users@nongnu.org http://lists.nongnu.org/mailman/listinfo/chicken-users
[Chicken-users] Re: bytevectors
Eduardo Cavazos wrote: So... do you think chicken can support the bytevectors api? :-) Hmmm... After poking around some, I see that I can probably cook up a compatability layer. Something along these lines: -- (use srfi-4) (define (make-bytevector n) (make-blob n)) (define (bytevector-ieee-double-native-ref blob i) (f64vector-ref (blob->f64vector/shared blob) i)) (define (bytevector-ieee-double-native-set! blob i val) (f64vector-set! (blob->f64vector/shared blob) i val)) -- Hopefully somebody can advise on the most efficient approach. Let's suppose I take this approach, where I use Chicken blobs in place of R6RS bytevectors. There's one problem that exists. Some of the procedures in the OpenGL egg want f64vector objects. For example, this results in an error: (let ((bv (make-blob (* 16 8 (glGetDoublev GL_MODELVIEW_MATRIX bv) bv) Again for comparison, the OpenGL libraries in Larceny, Ypsilon, and Ikarus allow a bytevector to be passed to 'glGetDoublev'. I'm willing to use a customized gl egg for my purposes. This is from 'gl.scm' in the opengl egg: void glGetDoublev( GLenum pname, GLdouble *params ); What could I change that to such that blobs are allowed as the second argument? Ed ___ Chicken-users mailing list Chicken-users@nongnu.org http://lists.nongnu.org/mailman/listinfo/chicken-users