[Armedbear-devel] Tail recursion

[Erik Huelsmann]

Hi Vibhu,

Thanks for sharing your thoughts. Indeed, rewriting function bodies in the
case of tail recursion is an option. However, do we know which percentage
of TCO calls can be optimized by using deftro? So far, I have not spent any
energy on TCO because my idea is that the benefit is just too small -- in
other words, I expect that the number of cases that would benefit from TCO
without getting optimization is just too big...

Do you have any idea/ numbers?


Regards,


Erik.


On Thu, Nov 20, 2014 at 8:54 PM, Vibhu Mohindra <vibhu.mohindra at gmail.com>
wrote:

> One idea I have is to use deftro (described below) instead of defun
> for tail recursive functions. On TCO Lisps define deftro to just
> be defun. On others define it to rewrite the body. It won't optimise
> all tail calls, just those are self-recursive.
>
> http://www.wispym.com/notes/tech/lisp/tco.txt
>
> I haven't used this myself. But I might start, because it's not
> ugly in the following sense. Writing a tail recursive function and
> depending implicitly on TCO is non-portable. deftro is a portability
> layer.
>
>
> _______________________________________________
> Armedbear-devel mailing list
> Armedbear-devel at common-lisp.net
> http://mailman.common-lisp.net/cgi-bin/mailman/listinfo/armedbear-devel
>



-- 
Bye,

Erik.

http://efficito.com -- Hosted accounting and ERP.
Robust and Flexible. No vendor lock-in.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.common-lisp.net/pipermail/armedbear-devel/attachments/20141123/c58cf7fc/attachment.html>
-------------- next part --------------
_______________________________________________
Armedbear-devel mailing list
Armedbear-devel at common-lisp.net
http://mailman.common-lisp.net/cgi-bin/mailman/listinfo/armedbear-devel