Re: master 6ebce84ff2b: Use t for non-nil default values in boolean defcustom declarations

Re: master 6ebce84ff2b: Use t for non-nil default values in boolean defcustom declarations

[Michael Albinus]

Mattias Engdegård <mattiase@acm.org> writes:

Hi Mattias,

>     Use t for non-nil default values in boolean defcustom declarations
>
> +(defcustom viper-ms-style-os-p
> +  (not (not (memq system-type '(ms-dos windows-nt))))

I'm just curious. Is there an advantage in using the (not (not ...))
pattern? I would use (and ... t) instead, but perhaps your pattern is
more performant? Or is it just a personal style?

Best regards, Michael.

Re: master 6ebce84ff2b: Use t for non-nil default values in boolean defcustom declarations

[Mattias Engdegård]

29 apr. 2023 kl. 09.39 skrev Michael Albinus <michael.albinus@gmx.de>:

> I'm just curious. Is there an advantage in using the (not (not ...))
> pattern? I would use (and ... t) instead, but perhaps your pattern is
> more performant? Or is it just a personal style?

I have no particular preference for either. If (and X t) is generally better liked, then it's fine to change to that.

Right now (not (not X)) is often but not always slightly shorter and faster but it doesn't matter in these cases and we could easily make both generate identical code.

Re: master 6ebce84ff2b: Use t for non-nil default values in boolean defcustom declarations

[Michael Albinus]

Mattias Engdegård <mattiase@acm.org> writes:

Hi Mattias,

>> I'm just curious. Is there an advantage in using the (not (not ...))
>> pattern? I would use (and ... t) instead, but perhaps your pattern is
>> more performant? Or is it just a personal style?
>
> I have no particular preference for either. If (and X t) is generally better liked, then it's fine to change to that.

Thanks for the reply. No, I don't believe it is needed to change your
personal style :-)

Best regards, Michael.

Re: master 6ebce84ff2b: Use t for non-nil default values in boolean defcustom declarations

[Andreas Schwab]

On Apr 29 2023, Michael Albinus wrote:

> Mattias Engdegård <mattiase@acm.org> writes:
>
> Hi Mattias,
>
>>     Use t for non-nil default values in boolean defcustom declarations
>>
>> +(defcustom viper-ms-style-os-p
>> +  (not (not (memq system-type '(ms-dos windows-nt))))
>
> I'm just curious. Is there an advantage in using the (not (not ...))
> pattern?

Logically, (not (null ...)) would fit better.

(not is just an alias of null, so both are the same, of course).

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 7578 EB47 D4E5 4D69 2510  2552 DF73 E780 A9DA AEC1
"And now for something completely different."

Re: master 6ebce84ff2b: Use t for non-nil default values in boolean defcustom declarations

[Robert Pluim]

>>>>> On Sat, 29 Apr 2023 13:23:42 +0200, Andreas Schwab <schwab@linux-m68k.org> said:

    Andreas> On Apr 29 2023, Michael Albinus wrote:
    >> Mattias Engdegård <mattiase@acm.org> writes:
    >> 
    >> Hi Mattias,
    >> 
    >>> Use t for non-nil default values in boolean defcustom declarations
    >>> 
    >>> +(defcustom viper-ms-style-os-p
    >>> +  (not (not (memq system-type '(ms-dos windows-nt))))
    >> 
    >> I'm just curious. Is there an advantage in using the (not (not ...))
    >> pattern?

    Andreas> Logically, (not (null ...)) would fit better.

    Andreas> (not is just an alias of null, so both are the same, of course).

Yes, but (not (not ...)) maps nicely to '!!' in C

Iʼm very curious if this change will cause any issues :-)

Robert
-- 

Re: master 6ebce84ff2b: Use t for non-nil default values in boolean defcustom declarations

[Andreas Schwab]

On Mai 02 2023, Robert Pluim wrote:

>>>>>> On Sat, 29 Apr 2023 13:23:42 +0200, Andreas Schwab <schwab@linux-m68k.org> said:
>
>     Andreas> On Apr 29 2023, Michael Albinus wrote:
>     >> Mattias Engdegård <mattiase@acm.org> writes:
>     >> 
>     >> Hi Mattias,
>     >> 
>     >>> Use t for non-nil default values in boolean defcustom declarations
>     >>> 
>     >>> +(defcustom viper-ms-style-os-p
>     >>> +  (not (not (memq system-type '(ms-dos windows-nt))))
>     >> 
>     >> I'm just curious. Is there an advantage in using the (not (not ...))
>     >> pattern?
>
>     Andreas> Logically, (not (null ...)) would fit better.
>
>     Andreas> (not is just an alias of null, so both are the same, of course).
>
> Yes, but (not (not ...)) maps nicely to '!!' in C

C doesn't have lists, and memq returns a list.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 7578 EB47 D4E5 4D69 2510  2552 DF73 E780 A9DA AEC1
"And now for something completely different."