[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Chicken-users] improved inlining

From: felix winkelmann
Subject: Re: [Chicken-users] improved inlining
Date: Thu, 30 Oct 2008 10:06:22 +0100

On Thu, Oct 30, 2008 at 10:00 AM, Alaric Snell-Pym
<address@hidden> wrote:
> How does this compare with the integrations mechanism? Will it one
> day subsume it entirely with a fully generic system that doesn't have
> to have special knowledge of a list of Scheme primitives?

This would be the ideal, yes. The current method of integrating "intrinsic"
functions is quite old and somewhat crude and ad-hoc. Handling all
this with cross-module inlining would be the most elegant way.
Currently there are conflicts: for example, some optimization of rest-
arg list handling gets confused by inlining of primitives like "null?", etc,
as these primitives are sometimes replaced with faster operations, if
they access a rest-arg. For this reason local-mode inlining is currently
not enabled for core libraries (but for the compiler and other tools).


reply via email to

[Prev in Thread] Current Thread [Next in Thread]