> On 1970-01-01 01:00, Jan van den Broek wrote:

> --***********

> What's up with your clock?

I suspect what's going on here is that some (*cough*tbird*cough*)
software is choking on the time zone spec in his date field. A _strict_
interpretation of RFC2822/5322 doesn't include arbitrary time zone spec
strings. You're allowed +/- four digits, or some military zone spec, or
a short list of mostly US-centric TLAs.

Yes, being that strict would violate the principle of "liberal in what
you accept". Yes, it'd be stupid not to accept common time zone
strings, if not pretty much any zone TLA there. Yes, it's software, and
_most_ of it is stupid in some form or another. But the fact that not
everyone sees this suggests there's a variable like different software
in play.

There _is_ a Reply-By: header defined[1] for use with X.400 systems, but
I'd never heard of it until I did a little searching just now, and I
sincerely hope that news readers aren't abusing its purpose somehow.

De

[1] RFC 2156

On 2023-03-20 23:01, Dennis Boone wrote:
> > On 1970-01-01 01:00, Jan van den Broek wrote:
>
> > --***********
>
> > What's up with your clock?
>
> I suspect what's going on here is that some (*cough*tbird*cough*)
> software is choking on the time zone spec in his date field.

Date: Sat, 18 Mar 2023 22:03:51 CET

> A _strict_
> interpretation of RFC2822/5322 doesn't include arbitrary time zone spec
> strings. You're allowed +/- four digits, or some military zone spec, or
> a short list of mostly US-centric TLAs.
>
> Yes, being that strict would violate the principle of "liberal in what
> you accept". Yes, it'd be stupid not to accept common time zone
> strings, if not pretty much any zone TLA there. Yes, it's software, and
> _most_ of it is stupid in some form or another. But the fact that not
> everyone sees this suggests there's a variable like different software
> in play.

So it resets to the start of time (zeroes?), not that it uses the
"Reply-By" string. It could also have used:

Injection-Date: Sat, 18 Mar 2023 22:03:49 +0100

> There _is_ a Reply-By: header defined[1] for use with X.400 systems, but
> I'd never heard of it until I did a little searching just now, and I
> sincerely hope that news readers aren't abusing its purpose somehow.
>
> De
>
>
> [1] RFC 2156

--
Cheers, Carlos.

> So it resets to the start of time (zeroes?), not that it uses the
> "Reply-By" string. It could also have used:

More like the parse failed so it never filled it in, probably,
but yes.

> Injection-Date: Sat, 18 Mar 2023 22:03:49 +0100

I doubt many newsreaders go fishing for any field other than Date
when they set up a quote like that.

De

On 2023-03-21 03:47, Dennis Boone wrote:
> > So it resets to the start of time (zeroes?), not that it uses the
> > "Reply-By" string. It could also have used:
>
> More like the parse failed so it never filled it in, probably,
> but yes.

That's what I meant. Fail parse, not filled, so the variable has
whatever value it gets on creation as new by whatever compiler they use,
or the programmer set it intentionally to "zero" when the code creates
it. The 1970 date marks variable not filled. Assuming that 1970... the
lowest value it can hold.

>
> > Injection-Date: Sat, 18 Mar 2023 22:03:49 +0100
>
> I doubt many newsreaders go fishing for any field other than Date
> when they set up a quote like that.

Right.

--
Cheers, Carlos.