SysTime.fromISOString

[Russel Winder via Digitalmars-d-learn]

On Sat, 2017-05-13 at 05:03 -0700, Jonathan M Davis via Digitalmars-d-
learn wrote:
> 
[…]
> SysTime st;
> try
> {
>     // If it's YYYY-MM-DDTHH:MM:SS.FFFFFFFZ
>     // If it's YYYY-MM-DDTHH:MM:SS.FFFFFFZ
>     // If it's YYYY-MM-DDTHH:MM:SS.FFFFFZ
>     // If it's YYYY-MM-DDTHH:MM:SS.FFFFZ
>     // If it's YYYY-MM-DDTHH:MM:SS.FFFZ
>     // If it's YYYY-MM-DDTHH:MM:SS.FFZ
>     // If it's YYYY-MM-DDTHH:MM:SS.FZ
>     // or      YYYY-MM-DDTHH:MM:SSZ
>     // or      YYYY-MM-DDTHH:MM:SS.FFFFFFF
>     // or      YYYY-MM-DDTHH:MM:SS.FFFFFF
>     // or      YYYY-MM-DDTHH:MM:SS.FFFFF
>     // or      YYYY-MM-DDTHH:MM:SS.FFFF
>     // or      YYYY-MM-DDTHH:MM:SS.FFF
>     // or      YYYY-MM-DDTHH:MM:SS.FF
>     // or      YYYY-MM-DDTHH:MM:SS.F
>     // or      YYYY-MM-DDTHH:MM:SS
>     st = SysTime.fromISOExtString(str);
> }
> catch(DateTimeException)
> {
>     try
>     {
>         // If it's just YYYY-MM-DD
>         st = SysTime(Date.fromISOExtString(str));
>     }

A variation on this is where I have gone, so that you are not telling
me I am doing it wrong, means it is probably the least worst solution.
I do a cascade of attempting to construct using SysTime.fromISOString,
SysTime.fromISOExtString, DateTime.fromISOString,
DateTime.fromISOExtString, Date.fromISOString, Date.fromISAExtString
(creating a SysTime in the latter four cases obviously).

>     catch(DateTimeException)
>     {
>         // If it's just HH:MM:SS
>         st = SysTime(Date.init, TimeOfDay.fromISOExtString(str));
>     }
> }

I haven't got to this stage as yet, but I suspect I am going to have
to.

> But that would be downright bizarre in the case where it was just
> HH:MM:SS,
> because then you'd end up with a SysTime on January 1st, 1 A.D. And
> while
> getting just the date is less awkward, it's still kind off to then
> operate
> on that as a SysTime. The time would always be at midnight.

I think you have to use the current date/time and extract the Date to
initialise. Times have to be in extended format to avoid the YYYY HHMM
conflict you pointed out.

> Really, just parsing an "ISO formatted string" and allowing either
> the date
> or time to be optional doesn't make much sense to me. That implies
> that you
> don't really know what it's for, since the meaning of something that
> is just
> a date and the meaning of something that is just a time are
> completely
> different.

I disagree. There is now and some temporal location in the past.
Separating out durations makes that easy. Using a time means sometime
today before now or is an error. Using a date/time means before today
or is an error. This is not a general case, it is a specific need for
the use of two date/times, and it seems all viable to me. Actually I
can probably separate date/times and times as well as separating out
durations before now to make things easier.

> Regardless, std.datetime was designed with the idea that you knew
> whether
> you were dealing with a date, a time, or both a date and a time when
> dealing
> with ISO formatted strings.

I'll get round it. Now I have "from the horses mouth" the guidance
around my questions, I am fairly confident of where to go to achieve
the goal. It will not be pretty, but it will be doable without having
to create a magic parsing function.

Thanks for your exchanges on this, most constructive and most helpful.

-- 
Russel.
=============================================================================
Dr Russel Winder      t: +44 20 7585 2200   voip: sip:russel.winder at ekiga.net
41 Buckmaster Road    m: +44 7770 465 077   xmpp: russel at winder.org.uk
London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: This is a digitally signed message part
URL: <http://lists.puremagic.com/pipermail/digitalmars-d-learn/attachments/20170513/4f793091/attachment.sig>