Re: [i]array_displace_check()

[Sam S. <sd...@gn...>]

> * In message <159...@ho...>
> * On the subject of "Re: [i]array_displace_check()"
> * Sent on Fri, 31 Jan 2003 22:32:44 +0100 (CET)
> * Honorable Bruno Haible <ha...@il...> writes:
>
> Sam writes (to the wrong mailing list):
> > iarray_displace_check() differes from array_displace_check() by only
> > one check in the beginning.
> > shouldn't the former be dropped and replaced with the latter?!
> 
> I don't believe that deoptimization of a carefully handcrafted code is
> a reasonable thing to do.

replacing a call to iarray_displace_check with a call to
array_displace_check adds one check - simple_arrayp.

> > (the potential performance hit from the extra check might be made up for
> > by the smaller code size :=)
> 
> I don't think so. 2 function calls instead of 1 function call for the
> simple task of preparing the access to an array: it _does_ matter for
> performance.

there will still be only one function call.
iarray_displace_check() should be removed altogether and all calls
replaced with array_displace_check().

> Indirect arrays are frequent enough that this case is worth not being
> pessimized. (Used by vector-push-extend, for I/O buffers, in streams,
> etc.)

ok.


-- 
Sam Steingold (http://www.podval.org/~sds) running RedHat8 GNU/Linux
<http://www.camera.org> <http://www.iris.org.il> <http://www.memri.org/>
<http://www.mideasttruth.com/> <http://www.palestine-central.com/links.html>
Two wrongs don't make a right, but three rights make a left.