#lisp - Tue 15 May 2007 between 06:29 and 07:14

NY Lost Funds



btbngrtrebor_win: yep, sbcl is much preferred :)
oh, hang on, i didn't uncomment the debug declaration. now i see the error.
trebor_winbtbngr: so what do you think? is it due to boxing/unboxing data?
Xofpkhuong: are you here?
(also I will emit my perennial query for nikodemus...)
btbngrtrebor_win: it's basically complaining that it has to box that value in that place in the code
trebor_win, just a performance warning, that's all.
trebor_winbtbngr: just for understanding right. is sbcl complaining about "1.0" or "det-a"?
btbngrtrebor_win: i don't think it's making it all that clear tbh
XofI suspect it's talking about the return value from (/ 1.0 det-a)
the point is that since sbcl knows that 1.0 and det-a are floats, it can inline the division, and just use the floating point registers for that
however, mul-mat-scal is a function about which it knows nothing (in principle you could redefine it later, for instance) and so it has to turn the float-in-a-register into something which can be passed as an argument to an arbitrary lisp function
can you paste the exact things that the compiler says?
trebor_winXof: ; in: DEFUN INV-MAT3D ; (MUL-MAT-SCAL MAT-TEMP (/ 1.0 DET-A)) ; ; note: doing float to pointer coercion (cost 13)
Xof<http://www.sbcl.org/manual/The-Parts-of-a-Compiler-Diagnostic.html> this might help understand a bit more
trebor_winXof: thanks.
actiontrebor_win is reading <http://www.sbcl.org/manual/The-Parts-of-a-Compiler-Diagnostic.html>
trebor_winis reading <http://www.sbcl.org/manual/The-Parts-of-a-Compiler-Diagnostic.html>
XofI'm a little surprised that you don't get source path information in that message. I wonder why
michaelwlichtblau: has CLOAK fallen off the net?
trebor_winXof: i did not cite correctly, the compiler notes started with: ; compiling file "d:\\Robert\\dki\\Sourcen\\workdir\\lisp\\percolation.lisp" (written 14 MAY 2007 05:42:47 PM):
Xofthat's not the kind of source path I meant
trebor_winXof: this one ; file: d:\Robert\dki\Sourcen\workdir\lisp\percolation.lisp ?
Xofnono, the "processing path" in the page I cited
knoboIf I have a pane (clim) how can I get the stream to that pane?
trebor_winXof: hm. i see it now, too. there is no "-->" in the note.
Fadedoes sbcl generally work well on windows?
trebor_winFade: i am a newbie - i do not know ;)
Fade: but it works with slime/emacs.
FadeI've never used lisp on anything but unix style systems.
trebor_winFade: and save-lisp-and-die works in sbcl 1.0.2 (it did not quite well in earlier versions)
actionFade nods
Fadenods
rr--there is no default key binding for slime-load-file, is there
prxqC-c C-l, iirc
btbngrrr: C-h m is handy at times like that :)
trebor_winis there a preferred order for defgeneric/defmethod/defclass? defclass -> defgeneric -> defmethod?
rr--omg you are right ... a key key binding, not mentioned in the slime manual
tcrrr--: The slime manual is not very up-to-date.
btbngrtrebor_win: that would be the ordering, yes.
tcrrr--: If you know the function name, you can generally look for a key binding of that function via `C-h w'.

Page: 2 9 16 23 

IrcArchive

NY Lost Funds