Re: lua32: Any easy fix for 2038 problem of os.time?

[Flyer31 Test <flyer31@...>]

Yes, maybe you are right ... so I will just accept that time_t is
32bit unsigned as anyway predfined in Keil ARMCC now... I will test,
but if this is working, then presumably I will be too lazy to
implement this 8sec version... .

On Wed, Aug 11, 2021 at 9:52 AM Oliver Kroth <oliver.kroth@nec-i.de> wrote:
>
> You don't need any more than 32 bits until about 2100.
> All you need is to look at time_t as an usigned 32bit value.
>
>
> Am 10.08.21 um 20:59 schrieb Flyer31 Test:
> > instead for looking for ±68 years around
> > 1970-01-01, the definition will be changed to up to +136 years since
> > 1970-01-01 (and nothing before that). I.e. we will have another 68 years
> > until the 32 bit counter will spill over.