[Openmcl-devel] MCL no longer on features?

[Gary Byers]

This issue came up on this mailing list in September.  My recollection of
the discussion at the time (see the last paragraph of

<http://clozure.com/mailman/htdig/openmcl-devel/2005-September/003036.html>

and the responses to that message)

is:

a) in the long run, it'd be good to remove :MCL from OpenMCL's features;
the implementations have diverged enough that one winds up having to
say:

   #+(and mcl (not openmcl)) (something-specific-to-commercial-mcl)
   #+openmcl                 (something-specific-to-openmcl)

but that this would break:

   #+mcl                     (something-for-either-commercial-mcl-or-openmcl
                               traditionally)
making it mean

   #+mcl                     (something-specific-to-commercial-mcl
                               breaking-with-tradition)

(and that there were also a lot of mysterious, legacy things on
*features* whose meaning had long been forgotten.)

b) breaking that case is probably the right thing, but it should be
done with some advance notice.

Part of the reason that mysterious legacy things stay on *features*
long after most people have forgotten what they mean is concern
about breaking existing code.  I don't remember exactly where I
saw it, but I have seen code that was conditionalized for :CORAL
within the last few years.  (The code itself was ~20 years old;
I mean that I saw it relatively recently; the code that I'm
thinking of hadn't been actively maintained in a long time, but
was still useful to someone and was still floating around the
net.)

I think that if some code is broken now (#+MCL no longer implies
OpenMCL), it's best in the long run.  The change does need to
be better publicized than it has been, and if people think that
it's too disruptive and not worthwhile to make the change in the
next release that's seems like a reasonable argument.  (I don't
think that the 1.0 release notes mentioned this; maybe it'd be
best to schedule the change for 1.2 and mention the issue in the
1.1 release notes.)




On Tue, 14 Feb 2006, David L. Rager wrote:

> Hola Gary K,
>
> Yep - we did the same for our program.
>
> GL,
> David
>
> -----Original Message-----
> From: Gary King [mailto:gwking at metabang.com]
> Sent: Tuesday, February 14, 2006 1:29 PM
> To: David L. Rager
> Cc: 'openmcl-devel Devel'
> Subject: Re: [Openmcl-devel] MCL no longer on features?
>
> Hi David,
>
> I assumed it was intentional but wanted to be sure because <whine>it
> breaks my legacy code.</whine>.
>
> I wanted to check before I update my code for the future...
>
> thanks,
>
> On Feb 14, 2006, at 12:26 PM, David L. Rager wrote:
>
>> Hi Gary,
>>
>> I'll let Gary B answer if he wants, but in the meantime, I'm 85-90%
>> certain
>> it was intentional.  I forget what the reason is, but it might have
>> something to do with legacy code.
>>
>> David
>>
>>> -----Original Message-----
>>> From: openmcl-devel-bounces at clozure.com [mailto:openmcl-devel-
>>> bounces at clozure.com] On Behalf Of Gary King
>>> Sent: Tuesday, February 14, 2006 10:39 AM
>>> To: openmcl-devel Devel
>>> Subject: [Openmcl-devel] MCL no longer on features?
>>>
>>> I build MCL from bleeding edge sources late last week and found:
>>>
>>>> Welcome to OpenMCL Version 1.1-pre-060125 (DarwinPPC32)!
>>>> ? (find :mcl *features*)
>>>> NIL
>>>> ?
>>>
>>> Is this intentional?
>>>
>>> Thanks,
>>> --
>>> Gary Warren King
>>> metabang.com
>>> http://www.metabang.com/
>>>
>>>
>>> _______________________________________________
>>> Openmcl-devel mailing list
>>> Openmcl-devel at clozure.com
>>> http://clozure.com/mailman/listinfo/openmcl-devel
>>
>
>
> -- 
> Gary Warren King
> metabang.com
> http://www.metabang.com/
>
>
> _______________________________________________
> Openmcl-devel mailing list
> Openmcl-devel at clozure.com
> http://clozure.com/mailman/listinfo/openmcl-devel
>
>