Re: [libgpiod] [RFC PATCH] bindings: python: allow specifying infinite timeout

[Nicolas Frattaroli <frattaroli.nicolas@xxxxxxxxx>]

On Dienstag, 23. Mai 2023 15:04:54 CEST Bartosz Golaszewski wrote:
> On Tue, May 23, 2023 at 2:38 PM Nicolas Frattaroli
> <frattaroli.nicolas@xxxxxxxxx> wrote:
> >
> > On Dienstag, 23. Mai 2023 14:33:12 CEST Bartosz Golaszewski wrote:
> > >
> > > ...
> > >
> > > As has been said elsewhere - the pythonic way is to interpret None as
> > > indefinite timeout. It's just that it would change the current
> > > behavior. The question is - should we interpret the current behavior
> > > as "undefined" and change it, or "defined but not documented" and
> > > consider it part of the API.
> > >
> > > Bart
> > >
> >
> > As an alternate suggestion, we could change the default function argument
> > to 0.0 and remove the None -> 0 code. That way, people who were calling
> > the function with no arguments still get the same behaviour, and the
> > only break is for users who explicitly passed None.
> >
> 
> Honestly, if we were to change the behavior, then I'd prefer to do it
> right and not use any half-measures.
> 
> Bart
> 

That's fair, I don't mind either way. I don't think the behavioural
change would be that big of an issue either way; the obviousness of a
wait on events function that returns immediately by default is
debatable in my eyes, as the only use for using it like that I could
ever come up with is to peek at whether there are any events queued
up based on its return value.

What I do think is a bad solution now is the float("inf") stuff.
While it's fun to assign meaning to special floating point values,
I think it strictly makes the API worse for the sake of stability.

Kind regards,
Nicolas Frattaroli