[Ecls-list] [PATCH] Don't do unnecessary work: clrhash on empty hash

[Juan Jose Garcia-Ripoll]

On Mon, Aug 4, 2008 at 7:10 PM, Erik Huelsmann <ehuels at gmail.com> wrote:
> Well,  it worked miracles for SBCL (reducing build time by 10%). I
> have no idea about the effect it has on ECL performance, but the patch
> below adds the equivalent check as recently implemented on SBCL for
> ECL.

Thanks a lot for the patch. I only had to make a minor change, which
is to factor out the memory cleaning loop, because it is also used by
make-hash-table and it has to be executed even if hash.entries = 0
(initially).

Juanjo

-- 
Instituto de Física Fundamental
CSIC, Serrano, 113, Madrid 28040 (Spain)
http://juanjose.garciaripoll.googlepages.com