On 26 Aug 2007, at 23:40, Vyacheslav Akhmechet wrote:
> On 8/26/07, Christophe Rhodes <csr21@...> wrote:
>> Did you not get my message on closer, in a reply to this same issue?
> Ah, sorry, I didn't realize you were from sbcl-devel. I thought there
> were no responses from the devs so I just sent a patch.
>> The problem is that class-prototype really needs to return a direct
> Like 42? :) BTW, does that mean that CLISP and CMUCL which return
> "approximations" of direct instances do not conform to the spec? Did
> the SBCL team remove this from CMUCL because of a stricter
> interpretation of the spec, or was it added to CMUCL after the fork?
The idea that there could be a more or less strict interpretation of
the CLOS MOP specification is a misunderstanding. The spec is simply
underspecified, so an implementation can essentially do whatever it
wants, including returning 42 for any built in class.
Granted, it is a somewhat surprising behavior. ;-)
Pascal Costanza, mailto:pc@..., http://p-cos.net
Vrije Universiteit Brussel, Programming Technology Lab
Pleinlaan 2, B-1050 Brussel, Belgium