Re: [PATCH v3 18/24] y2038: Add support for 64-bit time on legacy ABIs

[Stafford Horne <shorne@gmail.com>]

On Wed, Oct 13, 2021 at 09:45:22AM -0700, Paul Eggert wrote:
> On 10/13/21 4:44 AM, Stafford Horne via Libc-alpha wrote:
> 
> > > +@item
> > > +@code{_TIME_BITS=64} and port from the outset uses 64-bit
> > > +@code{time_t} and word size equals to @w{32 bits} (e.g. ARC, RV32)
> > > +- no action
> > 
> > OpenRISC falls into this category.  We have 64-bit support but word size is
> > 32 bits.
> 
> If OpenRISC has always had 64-bit time_t, the -D_TIME_BITS=64 option should
> have no effect on compilation. I.e., time_t used to be 64 bits, and it'll
> still be 64 bits regardless of whether you use that option.
> 
> > On OpenRISC we have __TIMESIZE == 64, so __USE_TIME_BITS64 does not get
> > defined.
> 
> That sounds OK. If time_t has always been 64 bits, there's no need for the
> 32- vs 64-bit time_t complexity.
> 
> The diagnostics you gave don't seem to have anything to do with time_t.
> However, if you're getting them with -D_TIME_BITS=64 but not getting them
> without that option, that suggests there's a glitch somewhere that needs to
> be tracked down more carefully.

Thanks, between this and Joseph's mail it makes sense where I was going wrong.

-Stafford