This is the mail archive of the
kawa@sourceware.cygnus.com
mailing list for the Kawa project.
Re: two bugs and a buggy patch
- To: Marco Vezzoli <marco dot vezzoli at st dot com>
- Subject: Re: two bugs and a buggy patch
- From: Per Bothner <per at bothner dot com>
- Date: 12 Apr 2000 00:02:39 -0700
- Cc: kawa at sourceware dot cygnus dot com
- References: <m2zor1p12h.fsf@kelso.bothner.com> <38F418B6.5229809E@st.com>
Marco Vezzoli <marco.vezzoli@st.com> writes:
> I read the SRFI9; I find it really useful.
> I tried to use the first two macros with kawa 1.6.66 and they worked
> well.
> Do you think kawa need some adaptation / optimization of the sample
> code?
> Or you simply want to add it in the next snapshot?
Well, I don't want to use an implementation on top of the record
feature since that less efficient than it can be. Specifically, I
want to generate the new class at compile time, not run-time, and I
don't want to use reflection to implement it. That is the field
accessor and modifier functions should be compiler-generated Procedure
that do not use reflection. That would make them much faster.
I think the thing to do is get define-class working, and
then impement SRFI-19 as macros that expand to define-class.
> I don't find this module in the SLib release I use at work (2c6);
> strangely enough they appear both in the change log and in a patch file
> packed in the same tarball.
Ah. I now notice in the Slib ChangeLog:
Thu Jul 18 17:37:14 1996 Aubrey Jaffer <jaffer@jacal.bertronics>
* rbtest.scm rbtree.scm: removed for lack of copying permissions.
I don't think we want to include in Kawa in that case.
--
--Per Bothner
per@bothner.com http://www.bothner.com/~per/