[CM] hard times getting clm-4 working

[Matti Koskinen]

On Jan 14, 2010, at 2:39 PM, Bill Schottstaedt wrote:

> I don't think you need to use -m32. 
snow leopard gcc compiles 64-bit code without -m32 , which can't be used in 32-bit lisp
> 
> clisp expands loop sometimes to code that 
> includes loop-finish, and then routinely changes the expansion of loop-finish, 
> breaking my optimizer.  I'll have to change the run-time loops to use "do", I suppose. 
> 
> sbcl (32 bit) works fine for me:
> 
> /Users/bil/clm/ sbcl
> This is SBCL 1.0.34, an implementation of ANSI Common Lisp.
> ..
> * (load "all.lisp")
> ..
> * (compile-file "singer.ins")
> ; Writing "/Users/bil/clm/clm_SINGER.c"
> ; Compiling "/Users/bil/clm/clm_SINGER.c"
> ..
> * (load "singer")
> 
> T
> * (with-sound ()  (singer 0 .1 '((.4 ehh.shp test.glt 523.0 .8 0.0 .01) (.6 oo.shp test.glt 523.0 .7 .1 
> 01))))
> 
> "test.aiff"
> * 
> 

to me, now go. Are you using 10.6 osx?

> I don't have clozure.  macroexpand-all is defined in walk.lisp,
> so perhaps you can just comment out that definition using #-openmcl.
> 
> The genetic_algs.lisp I find is an 8 year old 220b project using
> the singer -- it's not something I maintain as a part of the
> CLM package.  I don't know what goal-spectr or spectr.ins are.
> If you want help with random stuff like this, you'll need
> to at least send the code (or a url), and show the actual errors.
> 
sorry, this is the page
https://ccrma.stanford.edu/~nharlow/220b/proj/spectr.ins

after removing clm-print i get no segfault error and changing the loop clause to dotimes it compiles under clisp. But trying it, it mangles the filename:
 (with-sound(:play nil)(goal-spectr "/Users/mjkoskin/sounds/test1.aiff" "goal-spectra.lisp" "goal-spectra" 256))

can't open esU/m/srsokj/niknuost/sd1tsefia.: No such file or directoryesU/m/srsokj/niknuost/sd1tsefia. chans: -1can't open esU/m/srsokj/niknuost/sd1tsefia.: No such file or directoryesU/m/srsokj/niknuost/sd1tsefia.

???

-matti