Subj : Unix point?
To   : mark lewis
From : Maurice Kinal
Date : Fri Jan 14 2022 10:18:15

Hey mark!

 ml> so none of the Type 3 or Type 10 formats?

None.  What for?  From what I've witnessed over the last two decades is that there is no official acceptance of type 2+ headers which from this angle looks to be the most used.  Besides why tackle a new idea if we can't get the even the simplest things right such as telling the difference between a number and a string.

Just saying ... :-)

 ml> until you convert it to a number for the necessary math when
 ml> needed ;)

Sure.  I'd appreciate seeing your routine for taking care of that, esecially considering given the potentially corrupted (eastern) utc offset.  Myself I will continue ignoring it ... after I send this reply that is.

 ml> their string representation to some sort of numeric
 ml> representation so they can be added or subtracted from a time
 ml> value to give the time in another location...

You don't even have to do that if you use strftime() and a suitable offset.

  TZ=UTC date --rfc-3339=sec --date="14 Jan 22  10:48:08 -0500"
  2022-01-14 15:48:08+00:00

Doesn't that look much better?  For the record coreutils' date app is 100% strftime compatible but strftime() can be readily compiled in c for example.  The strftime function can be found in time.h.

Piece of cake when you follow real standards instead of poorly hacked and definetly corrupted "standard" such as demonstrated by fts-4008.002.

I will deal with CHRS when the appropriate time comes and given last year's performance I have serious doubts it will ever happen, nevermind in my lifetime.

Life is good,
Maurice

.... Weorða ðe selfne godum dædum, ðenden ðin God recce.
    Bring honour to yourself with good deeds, while God guides you.
--- GNU bash, version 5.1.8(1)-release (x86_64-moosile-linux-gnu)
 * Origin: One of us @ (1:153/7001.2989)

.