Re: [PATCH] functional-or returned #f for (functional-or #f #t #f). Fixed.

[David Kastrup]

Patrick McCarty <address@hidden> writes:

> On Tue, Nov 17, 2009 at 11:24 PM, David Kastrup <address@hidden> wrote:
>> Patrick McCarty <address@hidden> writes:
>>
>>> On 2009-11-16, David Kastrup wrote:
>>>>
>>>> It still suffers from not doing short-circuit evaluation.  For
>>>> readability and efficiency, I'd really prefer replacing
>>>> '(apply functional-or' with '(any'
>>>
>>> It looks like Neil removed the "is-harmonic" binding from this
>>> procedure a couple of days ago, since it was unused.
>>>
>>> So, nothing really needs to be changed after all.  :-)
>>
>> Still I feel more comfortable with this patch in place.  I am not
>> entirely sure whether it would not be simpler to just scratch both
>> functional-or and functional-and altogether since they are basically
>> only useful in situations where one would better use a different mapping
>> function with short-circuit evaluation.
>
> Oh, I didn't realize you wanted to change the `functional-or' itself.
> I don't think we should remove the two procedures entirely, because
> users might have written Scheme extension code that uses them.

They couldn't.  Wrong namespace.

lilypond -e '(functional-or 1)'
GNU LilyPond 2.13.8
<unnamed port>:1:8: In expression (functional-or 1):
<unnamed port>:1:8: Unbound variable: functional-or

> Can you send a formal patch so I can see how you want these procedures
> implemented?

I really don't want to have them implemented at all since you are better
off using a different mapping function that uses short-circuit
evaluation rather than mapping over functional-or.

-- 
David Kastrup