Re: table sort: IEEE floating-point number NaN?

Yes, by definition NaN is always different from everything (including NaN
itself, unless you use an exact identity test, e.g. with ===, instead of
normal equality test such as when 0 == 0.0 with different int/float/double
types).

Le sam. 25 avr. 2026 à 10:51, sur-behoffski <[email protected]> a
écrit :

> (Interesting discussion elided for briefness.)
>
> Does the floating-point number NaN (assume IEEE754 etc.) always
> compare as false to any other value, even another NaN?
>
> If so, this could confuse some partitioning algorithms.
>
> b
>
> --
> You received this message because you are subscribed to the Google Groups
> "lua-l" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> To view this discussion visit
> https://groups.google.com/d/msgid/lua-l/3b8f83f3-8a7a-44f9-8934-ef416f505988%40grouse.com.au
> .
>

-- 
You received this message because you are subscribed to the Google Groups "lua-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
To view this discussion visit https://groups.google.com/d/msgid/lua-l/CAGa7JC1oHuRKoHdFPXf_K090p%3D_nJ3tFJi8_ZUV5eg2sT_eF%3DA%40mail.gmail.com.