Re: iolist_size/1 in guard tests

So guard BIFs are supposed to be O(1) not just side effect free?

On Tue, Jan 4, 2022 at 12:36 PM Loïc Hoguin <[email protected]> wrote:

> Yes but that's considered a historical error.
>
> On 04/01/2022 18:38, Mark Jones wrote:
> > But length/1 is allowed in guard tests and it is O(N).
> >
> > Cheers,
> > Mark
> >
> > On Tue, Jan 4, 2022 at 2:53 AM Aleksander Nycz
> > <[email protected] <mailto:[email protected]>> wrote:
> >
> >     Probably
> >
> >     iolist_size/1 -> O(N)
> >
> >     byte_size/1 -> O(1)
> >
> >
> >     Regards
> >     Aleksander
> >
> >
> >     W dniu 04.01.2022 o 09:49, Mark Jones pisze:
> >      > Is there a reason iolist_size/1 is not a guard BIF like
> byte_size/1
> >      > and bit_size/1?
> >      >
> >      > Cheers,
> >      > Mark
> >
>
> --
> Loïc Hoguin
> https://ninenines.eu
>