Re: [TCLCORE] TIP-723 "timer" ready for review

[Marc C. <cul...@gm...>]

The problem I see with "realtime" is that that word has a completely
different meaning in linux: see:
 https://wiki.archlinux.org/title/Realtime_kernel
<https://wiki.archlinux.org/title/Realtime_kernel>

The meaning of wallclock is very clear.  It would also reduce ambiguity to
use "real" instead of "realtime" since that is what the linux time command
uses (see man time) which claims to be following POSIX standard 1003.2.

- Marc

- Marc

On Fri, Dec 5, 2025 at 4:21 AM Harald Oehlmann <har...@el...>
wrote:

> Am 04.12.2025 um 11:02 schrieb Christian Werner:
> > On 12/04/2025 10:39 AM, Harald Oehlmann wrote:
> >
> >> ...
> >> - will have a new type "wallclock".
> >>
> >> -> To native speakers: Is this a good term ?
> >
> > Speaking for POSIX their termini are CLOCK_REALTIME (= wall clock) vs.
> > CLOCK_MONOTONIC.
> > Maybe "realtime" and "monotonic" then?
> >
> > Cheers,
> > Christian
> This is a brilliant idea. In my IBM PC from 1986, there was a chip
> called "RTC "Real time clock" with a battery on the top.
> What about this:
>
> clock microseconds ?-realtime|-monotonic? (default: -realtime for
> backward compatibility)
>
> % timer info $id
> cmd monotonic 12345
> cmd realtime 45678
> cmd idle
>
> where the number is the invocation time point in the given timer.
>
> % after info $id
> cmd timer (for both, monotonic and realtime)
> cmd idle
>
> I will also replace the word "wall clock" by "realtime" in the TIP.
>
> Thanks again,
> Harald
> _______________________________________________
> Tcl-Core mailing list
> Tcl...@li...
> https://lists.sourceforge.net/lists/listinfo/tcl-core
>