Hi, Am 28.06.2009 um 07:53 schrieb Handkea fumosa:
The recur arg in question is (+ (* G__12815 G__12815) (* G__12817 G__12816 G__12816) G__12819) all of whose operands are doubles. This seems buggy.
Yes, that seems ominous. I'm not doing much number crunching (read: I don't need tight loops with primitives), so I can't really tell, why it doesn't work.
(If you're asking why the meaningless names, it's because this is actually adapted from the output of macroexpand-1. I have a macro to turn symbolic mathematical expressions into optimized iteration loops. It's not ready for prime-time yet; besides this problem, it doesn't do common subexpression elimination, so for instance there's two of (* G__12815 G__12815) in there.
I don't know, how your macro looks like, but maybe you can use the # notation: c-r# will give you something like c-r__123. However there are limitations for # notation (needs to be in the same `). In that case you can specify a prefix to gensym. (gensym "c-r__") will also give you c-r__123. This should help making things in macros clearer. Sincerely Meikel
smime.p7s
Description: S/MIME cryptographic signature
