Hyrum K Wright wrote on Thu, May 19, 2011 at 16:51:27 +0200:
> 2011/5/19 Branko Čibej <br...@e-reka.si>:
> > On 19.05.2011 15:38, Greg Stein wrote:
> >> 2011/5/19 Branko Čibej <br...@e-reka.si>:
> >>> On 19.05.2011 11:53, Stefan Sperling wrote:
> >>>> On Wed, May 18, 2011 at 09:38:55PM +0200, Branko Čibej wrote:
> >>>>> Why? That doesn't make sense. Second of all, all these wordy aliases are
> >>>>> just shorthands for real timestamps anyway -- by your reasoning, you
> >>>>> could eliminate all of them.
> >>>> There is otherwise no way to express dates relative to the current time.
> >>> So instead of introducing a subset of the silliness that was in CVS, why
> >>> then don't you invent an unambiguous format that /can/ express dates
> >>> relative to the current time?
> >>>
> >>> For example, you might support: svn -r {-1.12:13:56}, meaning one day,
> >>> twelve hours, 13 minutes and 56 seconds ago.
> >> "one day ago" is certainly easier than "-1"
> >>
> >> I don't see this as "silliness" but an easy way to express certain
> >> times. So what if it doesn't do everything? It doesn't the easy stuff
> >> just fine. It hasn't made the medium or hard stuff any more difficult.
> >
> > So someone who's not a native English speaker (or a fair imitation like
> > myself) will have to go looking at the docs ... it is silliness. We
> > don't parse anything but ISO dates, and now suddenly we'll parse whole
> > essays just to get the equivalent of that "-1 day". Sigh.
>
> Or just not use the feature?
>
> (It is, after all, completely undocumented for a reason.)
>
Lack of time on stsp's side?
