Re: macaddr 64 bit (EUI-64) datatype support

On 10/12/16, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> Vitaly Burovoy <vitaly(dot)burovoy(at)gmail(dot)com> writes:
>> I'm sorry for the offtopic, but does anyone know a reason why a
>> condition in mac.c
>
>>> if ((a < 0) || (a > 255) || (b < 0) || (b > 255) ||
>>> (c < 0) || (c > 255) || (d < 0) || (d > 255) ||
>>> (e < 0) || (e > 255) || (f < 0) || (f > 255))
>
>> can not be rewritten as:
>
>>> if (((a | b | c | d | e | f) < 0) ||
>>> ((a | b | c | d | e | f) > 255))
>

> Well, it's ugly and

> it adds a bunch of assumptions about arithmetic
> behavior that we don't particularly need to make.

It explains the reason, thank you. I'm just not familiar with other
architectures where it is not the same as in X86/X86-64.

> If this were some
> amazingly hot hot-spot then maybe it would be worth making the code
> unreadable to save a few nanoseconds, but I doubt that it is.

> (Anyway, you've not shown that there actually is any benefit ...)
I don't think it has a speed benefit, especially comparing with the sscanf call.
But personally for me the second variant does not seem ugly, just "no
one bit in all variables is out of a byte" (looks better with
comparison with "0xff" as sscanf operates with "%2x").
Sorry for my bad taste and for a noise.

--
Best regards,
Vitaly Burovoy