Darren Reed wrote:
> FWIW, I think we're approaching the time when we are going to need to
> start thinking about transitioning time_t to being uint64_t - even for
> 32bit platforms (as a point of interest, earlier this year we went past
> the point in time where a 32bit time_t can hold the start and end dates
> of a 30 year mortgage.)
>
> Currently we have it as 'long' (in <sys/types.h>), so we get 32bit
> or 64bit depending on the CPU architecture.
>
> The fallout of that should be things like stat64(2), etc...
>
> Darren
>
I agree whole heartedly. I think the "fix" presented for touch, which
is to deliver 64-bit only, is incomplete, since it doesn't help 32-bit
kernels which are still interesting.
This may be a bigger project to fix, but I think at some point we're
going to have to do it, because I don't think we can wait the 10 years
it will take for us to finally EOF 32-bit platforms.
-- Garrett
