[Gcl-devel] Re: safety 3

[Camm Maguire]

Greetings!  Wonder if it is legit to move *push-events* up to the
non-existent safety 4  level, in case someone needs it for backward
compatibility.

Take care,

Robert Boyer <address@hidden> writes:

> > safety 3 has come up recently.  This mode is actually slower than running
> > interpreted in many cases.
> 
> That's one of the funniest things I have heard in a long time.  It's like
> hearing of a compression algorithm actually producing a longer file.
> 
> This may be a half baked reply:  Compiled code should never run slower than
> interpreted code, excepting of course in deliberately weird situations.
> There is *no point* in having compiled code if it is not faster.  A possible
> and simple fix is simply to save, and later run, the interpreted code, after
> expanding macros, somehow disguising the code as "compiled" to save the
> appearances, e.g., compiled-function-p.  A pretty general assumption for
> Lisps is that compiled code always has had all of its macros expanded out.
> 
> Bob
> 
> 
> 
> 
> 

-- 
Camm Maguire                                            address@hidden
==========================================================================
"The earth is but one country, and mankind its citizens."  --  Baha'u'llah