[Ecls-list] About type propagation...

[Juan Jose Garcia-Ripoll]

On Wed, Aug 5, 2009 at 8:41 AM, Tobias C. Rittweiler<tcr at freebits.de> wrote:
> It's a bit suspicious. Did you try SBCL on it? Doesn't it signal a
> warning in this case? (For (CHAR-CODE (WHEN (ODDP X) #\X)) it does.)

The point is that SBCL is able to detect the code invariant, so the
code is correct when you have both type propagation + analysis of
branches and conditionals, but not with type propagation alone.

Juanjo

-- 
Instituto de Física Fundamental, CSIC
c/ Serrano, 113b, Madrid 28006 (Spain)
http://juanjose.garciaripoll.googlepages.com