Re: TV:TV-SLOT-NUMBER

[Eric Swenson]

> On Nov 8, 2024, at 2:17 PM, Alfred M. Szmidt <ams@gnu.org> wrote:
> 
> 
> 
>> On Nov 8, 2024, at 12:52 PM, Alfred M. Szmidt <ams@gnu.org> wrote:
>> 
>> 
>> 
>>> On Nov 8, 2024, at 10:47, Alfred M. Szmidt <ams@gnu.org> wrote:
>>> 
>>> 
>>> Where is this symbol assigned a value?  I don’t see any LISP files
>>> that assign this value, and one LISP file (WINDOW; COLD LISP) that
>>> references it.  I see an old system-98-47.lisp patch file that
>>> initializes it and declares it a constant, but that won’t be used.
>>> 
>>> WINDOW; COLD LISP includes:
>>> 
>>> (DEFUN INITIALIZE-WIRED-KBD-BUFFER ()
>>>   (COND ((= PROCESSOR-TYPE-CODE LAMBDA-TYPE-CODE)
>>>                      ^^^^^^
>>> 
>>> This is Lambda specific code.  Lambda support started in or around System 
>>> 98.
>>> 
>>>      (%NUBUS-WRITE TV:TV-SLOT-NUMBER 4
>>>            (LOGAND (LOGNOT 40) (COMPILER:%NUBUS-READ TV:TV-SLOT-NUMBER 
>>> 4)))))
>>> …
>>> 
>>> Is it possible that we aren’t ever invoking that function?  The compiler 
>>> warns that TV:TV-SLOT-NUMBER is assumed special.
>> 
>>  Must not be since variable isn’t bound.
>> 
>> What? %NUBUS-WRITE is Lambda, CADR does not have it defined, just like
>> TV-SLOT-NUMBER.
>> 
> 
>   Ok, then a simple (declare (special tv:tv-slot-number)) will get
>   rid of the compiler warning, and it doesn’t matter whether it is
>   bound since that code will never get executed on a non-lambda.
> 
> TV-SLOT-NUMBER is not a special variable, declaring it as such is
> wrong.  If anything, #+lambda the code, IF-FOR-LAMBDA, or
> SELECT-PROCESSOR.
> 
> In either case .. that form is nuked in window;cold#187

I thought all DEFCONSTS were considered special.  They are not in scope, but in 
global scope.  I’ve always declared them special.

In any case, as you said, the issue is moot if you’ve removed the form that 
referenced it.