From: Fernando Mato Mira
Subject: CMUCL GC (was: cautios question (about languages))
Date:
Message-ID: <37A70371.E5BDADD1@iname.com>
Christian Lynbech wrote:
> >>>>> "Fernando" == Fernando Mato Mira <········@iname.com> writes:
>
> Also rscheme does not support the full functionality of tinyclos (or
> derivatives such as STk). It lacks at least multidispatch, and perhaps
> also multiinheritance (though I am not quite sure about that).
One more thing to wait for then.
How difficult would it be to replace the CMUCL GC by something else ?
From: Raymond Toy
Subject: Re: CMUCL GC (was: cautios question (about languages))
Date:
Message-ID: <4nyafs4wxk.fsf@rtp.ericsson.se>
>>>>> "Fernando" == Fernando Mato Mira <········@iname.com> writes:
Fernando> Christian Lynbech wrote:
>> >>>>> "Fernando" == Fernando Mato Mira <········@iname.com> writes:
>>
>> Also rscheme does not support the full functionality of tinyclos (or
>> derivatives such as STk). It lacks at least multidispatch, and perhaps
>> also multiinheritance (though I am not quite sure about that).
Fernando> One more thing to wait for then.
Fernando> How difficult would it be to replace the CMUCL GC by something else ?
Pretty hard. CMUCL uses a stop-and-copy (?) for all ports. The x86
port has a conservative generational GC available. Porting this to
the other ports should not be extremely difficult.
Ray