On Jun 3, 2009, at 4:40 PM, Michele Simionato wrote:
On Wed, Jun 3, 2009 at 3:32 PM, Abdulaziz Ghuloum
<[email protected]> wrote:
http://funcall.blogspot.com/2009/05/you-knew-id-say-something-
part-iv.html
Ah, I did not know that. So Ikarus is using a ring of buffers?
Technically, a ring of rings: outer rings for nontail calls and inner
rings for tail calls within the nontail calls.
That makes sense and it looks like a good approach.
It does look like it. The only drawback that I'm seeing is that making
deeply-nested nontail calls recycles earlier traces, and when you return
from deep recursion, you end up with little or no traces. Being able to
set the size of the rings might be sufficient (though I highly doubt
it).
Aziz,,,
