Re: [Gcl-devel] Re: SLIME

[Eric Merritt]

--- Camm Maguire <address@hidden> wrote:
> Public domain is completely GPL compatible.

 Then it shouldn't be an problem at all to include
these streams into gcl.
 
> > > 3) Any progress on the sibling call front?  
> > 
> 
> OK, I've been testing a bit, as you've likely seen
> from the mails.
> The patch is ready to go in if/when 1) I can trace
> down the ANSI C
> stack overflow error reporting difficulty and 2)
> some kind of sane
> default behavior is chosen.  For example, inlining
> functions, turned
> on in -O3, apparently interferes with sibcall
> optimization.  So should
> we default to -O3 -fno-inline-functions?  Even if we
> discover a
> superior way to write the C code to disable an
> inlining in a particular
> call instance, how do we decide what is best? 

 Shouldn't inlining give the same semantics as
sibcalls? If one functions is inlined into another
function then there shouldn't be any stack issues
right? or am I missing something?

> Perhaps the simplest is
> to implement the no-line lisp declaration directly
> as a gcc function
> attribute where gcc is available, contenting
> ourselves with per
> function rather than per call inlining decisions?

That sounds like a plan to me.
 
> 
> Agreed.  But apparently we can't have everything at
> once. 
> 

 Isn't that always the case? ;)


        
                
__________________________________
Do you Yahoo!?
Win a $20,000 Career Makeover at Yahoo! HotJobs  
http://hotjobs.sweepstakes.yahoo.com/careermakeover