What is the protocal for GMT offset in SMTP (e-mail) header time-stamp?

[Mail Man]

As far as I know, Windows (perhaps any version of windows) is configured
with a single GMT offset as defined by the Time Zone properties in the
system's clock / calendar interface or setup window.

For the Eastern Time Zone (US / Canada) that is set as GMT - 5 hours.

The DST rules for that zone specifies that a bias of + 1 hour start the
second sunday in march and end the first sunday in november.

But does that mean that those systems will alter their time-stamping
systems such that they change both their clock-time (add 1 hour at 2 am)
*AND* change their time-zone bias (from GMT -5 to GMT -4) ???

As far as I can tell, windows systems do not change their zone bias from
-5 to -4.

The tzedit.exe program (which allows the DST rules to edited or
modified) does not seem to allow for a change in the GMT bias. Windows
systems that perform the DST time-change correctly don't seem to alter
their GMT bias from -5 to -4.

I believe that any windows e-mail client or mail server that creates or
handles SMTP messages will (or does) continue to time-stamp messages
headers with GMT -05:00 instead of GMT -04:00. This causes an apparent
1-hour future time-shift of the messages.

Am I right in this observation?

Does the Windows operating system change it's GMT offset in response to
DST? When application programs request the current date and time on a
Windows system, does the OS report back a GMT offset of -5, or -4 during
the DST period?

[Sam]

Post by Mail Man
As far as I know, Windows (perhaps any version of windows) is configured
with a single GMT offset as defined by the Time Zone properties in the
system's clock / calendar interface or setup window.
For the Eastern Time Zone (US / Canada) that is set as GMT - 5 hours.
The DST rules for that zone specifies that a bias of + 1 hour start the
second sunday in march and end the first sunday in november.
But does that mean that those systems will alter their time-stamping
systems such that they change both their clock-time (add 1 hour at 2 am)
*AND* change their time-zone bias (from GMT -5 to GMT -4) ???
As far as I can tell, windows systems do not change their zone bias from
-5 to -4.

I'm not really familiar with what goes on in the Windows world, but on Unix
this is all very simple. Unix's internal system clock runs on GMT, and the
timezone setting is simply the offset from GMT.

Therefore, the current local time is computed simply by applying the current
GMT offset to the system clock. The fact that the real system clock runs on
GMT is not something that anybody cares about. Whenever the current time, or
some other time, is shown, the appropriate offset from GMT gets applied to
the time value, and the result is presented for consumption by human
eyeballs.

Changing a server's timezone involves merely changing an environment
variable. There's no change applied to the system clock. It always runs on
GMT, only the effective offset changes, a separate setting. On Unix, one
process may run with its effective timezone being Eastern time, and a
different process running at the same time have its effective timezone
environment variable specifying Pacific time. The processes would, at the
same time, always see their local time being three hours apart, but there's
still only one system clock, on GMT.

The "DST rules", as you call them, merely specify when the current effective
timezone offset changes. Nothing out of the ordinary happens when, to us
humans, the current timezone offset changes. The internal system clock
continues to run on GMT, and only the effective GMT offset changes.

The timezone rules are a little bit more flexible than just "on such and
such Sunday of such and such month". Timezone rules change occasionally --
as they did a couple of years ago, so that, in the United States, the
so-called "Standard Time" is now in effect for a lesser part of the year
than the, presumably, non-standard time, but that's a different kind of
insanity.

To simplify, the time zone rules are a boring list of, basically, the
following data for the America/New_York timezone (a.k.a. Eastern Time):



On 01 Nov 2009 06:00:00 GMT, GMT offset -0500 is in effect
On 14 Mar 2010 07:00:00 GMT, GMT offset -0400 is in effect
On 07 Nov 2010 06:00:00 GMT, GMT offset -0500 is in effect



And so on, starting at the dawn of time, and ending when our sun blows up.
That, to oversimplify, is how this single timezone get defined. You can find
more information here: http://en.wikipedia.org/wiki/Tz_database

Therefore, on precisely 14 Mar 2010 06:59:59 GMT, GMT offset -0500 was in
effect, therefore the local time was 14 Mar 2010 01:59:59 -0500.

Nothing unusual happened a second later. The system clock advanced to 14 Mar
2010 07:00:00 GMT, and that's all. At this time, GMT offset -0400 is in
effect, and the local time is 14 Mar 2010 03:00:00 -0400.

Later this year, specifically on 07 Nov 2010 05:59:59 GMT, the GMT offset
will still be -0400, and the local time will be 07 Nov 2010 01:59:59 -0400.

One second later will be 07 Nov 2010 06:00:00 GMT, the GMT offset in effect
will be -0500, and the local time will be 07 Nov 2010 01:00:00 -0500.

Internally, time is always kept in GMT, and gets translated to local time
whenever it needs to be shown for human consumption. The opposite process is
also occasionally needed: taking the local time and mapping it back to GMT.

Post by Mail Man
I believe that any windows e-mail client or mail server that creates or
handles SMTP messages will (or does) continue to time-stamp messages
headers with GMT -05:00 instead of GMT -04:00. This causes an apparent
1-hour future time-shift of the messages.
Am I right in this observation?

Probably not. I do recall there's a knob that specifies whether daylight
time is in effect, or not.

A sane operating system merely gives the application the current system
clock time in GMT, and a library function that converts between GMT and
local time, using whatever timezone the application wants to use, which is
the system default timezone, most of the time. The local time information
also includes the offset from GMT in effect, which goes directly into the
±NNNN portion of the timestamp.

My mail applications simply take the current system time, convert it to
localtime, and the results are shoved directly into the Date: header,
verbatim. They fit right in. No confusion whatsoever.

Similarly, when my application consume someone else's Date: header, such as
a Date: header in someone else's email, with some arbitrary timezone:

Well, whatever the timezone is, note that the Date: header conveniently
includes the actual ±HHMM offset, in the sender's timezone (ignoring, for
the moment, some ancient Date: headers that references explicit US timezone
names). So, based on the literal date, time, and offset given directly in
the Date: header, applying the offset to the literal date and time gives me
the GMT time. Then, the same localtime function converts it back into my
application's local time. So, whatever time was the sender's local time, my
application always displays the corresponding local time for the recipient.

[Mail Man]

Post by Sam

Post by Mail Man
As far as I can tell, windows systems do not change their zone
bias from -5 to -4.

I'm not really familiar with what goes on in the Windows world,

Which makes you or your position somewhat rare in the IT world

Post by Sam
but on Unix this is all very simple.

Any given machine (unix or otherwise) must know two things:

a) it's "normal" GMT offset, as dictated by it's timezone

b) it's daylight-savings-time strategy - if any.

If some form of DST kicks in for a given computer, the simple thing for
it to do is to just change it's GMT offset. When it needs to generate a
human-useful time output, it simply applies the GMT offset to it's
internal clock (which is always running at GMTime).

When a machine generates a time-stamp for e-mail, then it stamps it with
the current local time, and adds the GMT offset to complete the stamp.
That way, when any other machine decodes the stamp, it can reference it
back to GMT, then it can display it in terms of the it's own time -
which it knows with respect to GMT.

Post by Sam
Changing a server's timezone involves merely changing an environment
variable. There's no change applied to the system clock.

I'm not so sure that Windows systems change their GMT offset during DST.

Post by Sam

Post by Mail Man
I believe that any windows e-mail client or mail server that
creates or handles SMTP messages will (or does) continue to
time-stamp messages headers with GMT -05:00 instead of GMT
-04:00. This causes an apparent 1-hour future time-shift of
the messages.
Am I right in this observation?

Probably not. I do recall there's a knob that specifies whether
daylight time is in effect, or not.

I'm seeing lots of problems with Microsoft Outlook )on various versions
of Windows) and how it's displaying e-mail recieved times ever since EST
started.

[Sam]

Post by Mail Man

Post by Sam

Post by Mail Man
As far as I can tell, windows systems do not change their zone
bias from -5 to -4.

I'm not really familiar with what goes on in the Windows world,

Which makes you or your position somewhat rare in the IT world

That's debatable. There are far more Unix and Linux servers in the entire
world -- including your Usenet server, news.aioe.org, than there are Windows
machine.

Post by Mail Man

Post by Sam
Changing a server's timezone involves merely changing an environment
variable. There's no change applied to the system clock.

I'm not so sure that Windows systems change their GMT offset during DST.

I'm fairly certain that there is some dim concept that a different GMT
offser would apply during DST. Otherwise, Windows is even more brain damaged
than I suspected.

Post by Mail Man

Post by Sam

Post by Mail Man
Am I right in this observation?

Probably not. I do recall there's a knob that specifies whether
daylight time is in effect, or not.

I'm seeing lots of problems with Microsoft Outlook )on various versions
of Windows) and how it's displaying e-mail recieved times ever since EST
started.

Sorry to hear that.

[Mail Man]

Post by Sam

Post by Mail Man

Post by Sam
I'm not really familiar with what goes on in the Windows world,

Which makes you or your position somewhat rare in the IT world

That's debatable.

It's highly likely that your position would not exist except for the
fact that there exists many people using Windows and those windows
systems interact with servers running Linux / Unix.

You may find linux/unix superior in many ways over windows (and I don't
argue that) but it's undisputable that many of those linux/unix servers
exist to serve people and machines running windows. Hence an IT
administrator or sys-admin usually must know about or deal with the
correct inter-operatiblity of these various platforms.

Post by Sam
There are far more Unix and Linux servers in the entire world --
including your Usenet server, news.aioe.org, than there are
Windows machine.

Are you stating that there are more Unix/Linux servers than there are
machines running Windows, or than there are servers running Windows?

[Sam]

Post by Mail Man

Post by Sam

Post by Mail Man

Post by Sam
I'm not really familiar with what goes on in the Windows world,

Which makes you or your position somewhat rare in the IT world

That's debatable.

It's highly likely that your position would not exist except for the
fact that there exists many people using Windows and those windows
systems interact with servers running Linux / Unix.

My position does not involve Windows on either the server side or the client
side.

Post by Mail Man
You may find linux/unix superior in many ways over windows (and I don't
argue that) but it's undisputable that many of those linux/unix servers
exist to serve people and machines running windows. Hence an IT
administrator or sys-admin usually must know about or deal with the
correct inter-operatiblity of these various platforms.

That may be true for some "IT administrators or sys-admins". However it is
not true in every case. Furthermore, there are plenty of jobs in IT in
addition to administrators and sys-admins. Any assumption that I'm either an
IT administrator or an sysadmin would reveal a fundamental lack of
understanding.

Post by Mail Man

Post by Sam
There are far more Unix and Linux servers in the entire world --
including your Usenet server, news.aioe.org, than there are
Windows machine.

Are you stating that there are more Unix/Linux servers than there are
machines running Windows, or than there are servers running Windows?

A raw number comparison is generally meaningless, because of the entirely
different natures of both ecosystems. It is fairly common, for example, to
have a single multihomed Linux server run a squid cache for locally
originated web traffic, run Apache to serve public web pages, a mail server
for handling incoming and outgoing mail, and bind, to provide DNS resolution
for the LAN. Maybe even Asterisk, for voice-over-IP.

Good luck trying to get Exchange, IIS, Active Directory, and some web
caching proxy running on a single Windows server that also functions as a
multihomed router.

[Fritz Wuehler]

Post by Mail Man

Post by Sam
I'm not really familiar with what goes on in the Windows world,

Which makes you or your position somewhat rare in the IT world

Your point was? At the top of your head, Windows-boy!

The point is Windows isn't important. Nobody serious about computing uses
it for anything. Oh, for games it's quite fun. But not for getting real
work done...

[Mail Man]

Post by Fritz Wuehler

Post by Mail Man

Post by Sam
I'm not really familiar with what goes on in the Windows world,

Which makes you or your position somewhat rare in the IT world

Your point was? At the top of your head, Windows-boy!

Read my last reply to Sam for an explanation of my point.

Post by Fritz Wuehler
The point is Windows isn't important.

What kind of joke statement is that?

If it wasn't for millions of PC's running windows, then most servers
wouldn't need to exist. I can say that, without needing to admit that
windows is a horribly-designed operating system.

Post by Fritz Wuehler
Nobody serious about computing uses it for anything.

I don't know too many people that use their computers for "computing".

They use their computers to accomplish tasks, such as document and
graphic creation / editing, message communication, multimedia viewing,
entertainment, etc.

Post by Fritz Wuehler
Oh, for games it's quite fun. But not for getting real
work done...

Yea - I guess that's why Linux and open office is the predominant OS in
commercial, institutional and SOHO environments.

[Sam]

Post by Mail Man

Post by Fritz Wuehler
Oh, for games it's quite fun. But not for getting real
work done...

Yea - I guess that's why Linux and open office is the predominant OS in
commercial, institutional and SOHO environments.

Yes, it is. Both Windows and Solaris are considered legacy platforms in
contemporary data centers.

[Mail Man]

Post by Sam

Post by Mail Man
Yea - I guess that's why Linux and open office is the predominant
OS in commercial, institutional and SOHO environments.

Yes, it is. Both Windows and Solaris are considered legacy platforms
in contemporary data centers.

Since when were we talking about data centers?

[Sam]

Post by Mail Man

Post by Sam

Post by Mail Man
Yea - I guess that's why Linux and open office is the predominant
OS in commercial, institutional and SOHO environments.

Yes, it is. Both Windows and Solaris are considered legacy platforms
in contemporary data centers.

Since when were we talking about data centers?

Since you mentioned commercial, institutional, and SOHO environments.

[Frank Slootweg]

Post by Mail Man

Post by Fritz Wuehler

Post by Mail Man

Post by Sam
I'm not really familiar with what goes on in the Windows world,

Which makes you or your position somewhat rare in the IT world

Your point was? At the top of your head, Windows-boy!

Read my last reply to Sam for an explanation of my point.

Post by Fritz Wuehler
The point is Windows isn't important.

What kind of joke statement is that?
If it wasn't for millions of PC's running windows, then most servers
wouldn't need to exist.

*Historically*, it's probably the other way around! Without UNIX
servers, there wouldn't have been "millions of PC's running windows",
because there would never have been any Internet, mail and web worth
mentioning.

[...]

[Mail Man]

Post by Frank Slootweg

Post by Mail Man

Post by Fritz Wuehler
The point is Windows isn't important.

What kind of joke statement is that?
If it wasn't for millions of PC's running windows, then most
servers wouldn't need to exist.

*Historically*, it's probably the other way around! Without UNIX
servers, there wouldn't have been "millions of PC's running
windows"

The development and introduction of the consumer-grade PC back in the
late 1970's and early 1980's did not require or depend on the existance
of unix servers.

Consumer PC's of that era were communicating with each other via BBS,
fido, etc, and did not require the interconnectivity of the (then)
embryonic TCP/IP-based internet to exist.

Post by Frank Slootweg
because there would never have been any Internet, mail and web
worth mentioning.

Modem-based communication for PC's (as previously mentioned) was already
in place by the very early 1980's, and by today would have been
developed along lines we can only imagine had the internet not existed
or have been cross-platform-deployed as widely as it is now.

It is a complete fallacy to say that Unix led to servers which led to
the internet which led to windows-based consumer PC's.

One thing that can be said is that consumer-grade windows PC's has led
to a wide(r) distribution and availablility of DSL service, and a
corresponding price drop for anyone connecting a machine to the internet
(even someone running a linux/unix server).

And today, this statement is certainly true:

If it wasn't for millions of PC's running windows, then most
(linux/unix) servers wouldn't need to exist.

[Frank Slootweg]

Post by Mail Man

Post by Frank Slootweg

Post by Mail Man

Post by Fritz Wuehler
The point is Windows isn't important.

What kind of joke statement is that?
If it wasn't for millions of PC's running windows, then most
servers wouldn't need to exist.

*Historically*, it's probably the other way around! Without UNIX
servers, there wouldn't have been "millions of PC's running
windows"

The development and introduction of the consumer-grade PC back in the
late 1970's and early 1980's did not require or depend on the existance
of unix servers.

There were *some* *consumer*-grade PCs in the late 1970's and early
1980's. *Most* PCs were not really "personal" let alone consumer/private,
but let's not quibble about that, because that wasn't the point I was
making. And note that we were talking about PC's running *Windows*. If
we combine the two - Windows *and* consumer-grade - , there were very,
very few in that era.

[...]

Post by Mail Man
It is a complete fallacy to say that Unix led to servers which led to
the internet which led to windows-based consumer PC's.

You said "millions of PC's running windows" and obviously meant
"hundreds of millions" (billions?). *That* is the point I'm arguing, not
your backpedalled version.

[...]

Post by Mail Man
If it wasn't for millions of PC's running windows, then most
(linux/unix) servers wouldn't need to exist.

In *numbers*, correct. In *percentage*, wrong.

[David F. Skoll]

Post by Mail Man
If it wasn't for millions of PC's running windows, then most
(linux/unix) servers wouldn't need to exist.

Meh. Windows was marginally acceptable in the pre-Internet era.
However, the system has had its day and should gracefully die,
yielding to better-designed systems. Ironically, even though
UNIX-like systems are older than Windows, they have proven
better-designed, more nimble, more secure, and more flexible than the
lumbering beast Redmond sees fit to foist on the world every few
years.

-- David.

[Jonathan de Boyne Pollard]

Post by Mail Man
If it wasn't for millions of PC's running windows, then most
(linux/unix) servers wouldn't need to exist.

Meh. Windows was marginally acceptable in the pre-Internet era.
However, the system has had its day and should gracefully die,
yielding to better-designed systems. Ironically, even though UNIX-like
systems are older than Windows, they have proven better-designed, more
nimble, more secure, and more flexible than the lumbering beast
Redmond sees fit to foist on the world every few years.

What versions of Microsoft Windows are you thinking both pre-dated the
Internet and were marginally acceptable?

[Tim Streater]

Post by Jonathan de Boyne Pollard

Post by Mail Man
If it wasn't for millions of PC's running windows, then most
(linux/unix) servers wouldn't need to exist.

Meh. Windows was marginally acceptable in the pre-Internet era.
However, the system has had its day and should gracefully die,
yielding to better-designed systems. Ironically, even though UNIX-like
systems are older than Windows, they have proven better-designed, more
nimble, more secure, and more flexible than the lumbering beast
Redmond sees fit to foist on the world every few years.

What versions of Microsoft Windows are you thinking both pre-dated the
Internet and were marginally acceptable?

None, in fact.

W95 was just about usable and you really wanted W98.

Meanwhile I'd been using Word V4 for Mac back in the late 80s, some ten
years earlier.

[Dave]

Post by Tim Streater

Post by Jonathan de Boyne Pollard

Post by Mail Man
If it wasn't for millions of PC's running windows, then most
(linux/unix) servers wouldn't need to exist.

Meh. Windows was marginally acceptable in the pre-Internet era.
However, the system has had its day and should gracefully die,
yielding to better-designed systems. Ironically, even though UNIX-like
systems are older than Windows, they have proven better-designed, more
nimble, more secure, and more flexible than the lumbering beast
Redmond sees fit to foist on the world every few years.

What versions of Microsoft Windows are you thinking both pre-dated the
Internet and were marginally acceptable?

I would say NT 3.5...

Post by Tim Streater
None, in fact.
W95 was just about usable and you really wanted W98.
Meanwhile I'd been using Word V4 for Mac back in the late 80s, some ten
years earlier.
--
Tim
"That excessive bail ought not to be required, nor excessive fines imposed,
nor cruel and unusual punishments inflicted" -- Bill of Rights 1689

[David F. Skoll]

Post by Jonathan de Boyne Pollard

Meh. Windows was marginally acceptable in the pre-Internet era.

What versions of Microsoft Windows are you thinking both pre-dated the
Internet and were marginally acceptable?

No version of Windows predated the Internet, but by the "pre-Internet era",
I meant before the Internet became generally popular with average consumers.
I should have specified that more precisely; I meant before about 1995.

And Windows 3.1 was marginally acceptable if you were used to DOS.

Regards,

David.

[Patrick Scheible]

Post by David F. Skoll
And Windows 3.1 was marginally acceptable if you were used to DOS.

With third-party additions, Windows 3.1 could act as a terminal to a
real computer.

-- Patrick

[Joe Pfeiffer]

Post by Patrick Scheible

Post by David F. Skoll
And Windows 3.1 was marginally acceptable if you were used to DOS.

With third-party additions, Windows 3.1 could act as a terminal to a
real computer.

That was what came on my Compaq Concerto (25MHz 486!), which was my
first laptop (I was *really* late in getting into laptops). I lasted
about a day before I started looking into Unix-like alternatives -- I
still don't remember why I picked Linux over FreeBSD, but I've been very
happy ever since.

[Jonathan de Boyne Pollard]

Post by Frank Slootweg

Post by Mail Man

The point is Windows isn't important. Nobody serious about
computing uses it for anything. Oh, for games it's quite fun. But
not for getting real work done...

What kind of joke statement is that?
If it wasn't for millions of PC's running windows, then most servers
wouldn't need to exist.

*Historically*, it's probably the other way around! Without UNIX
servers, there wouldn't have been "millions of PC's running windows",
because there would never have been any Internet, mail and web worth
mentioning.

And ... cue Mark Crispin! (-:

[Frank Slootweg]

Post by Frank Slootweg

Post by Mail Man

The point is Windows isn't important. Nobody serious about
computing uses it for anything. Oh, for games it's quite fun. But
not for getting real work done...

What kind of joke statement is that?
If it wasn't for millions of PC's running windows, then most servers
wouldn't need to exist.

*Historically*, it's probably the other way around! Without UNIX
servers, there wouldn't have been "millions of PC's running windows",
because there would never have been any Internet, mail and web worth
mentioning.

Come on, Jonathan! I was talking about *history*, not newbies! :-) G&D

[Frank Slootweg]

[Newsgroups: restored]

Post by Frank Slootweg

Post by Mail Man

Post by Fritz Wuehler
The point is Windows isn't important.

What kind of joke statement is that?
If it wasn't for millions of PC's running windows, then most servers
wouldn't need to exist.

*Historically*, it's probably the other way around! Without UNIX
servers, there wouldn't have been "millions of PC's running windows",
because there would never have been any Internet, mail and web worth
mentioning.

There were millions of PCs running Windows before there was a World
Wide Web.

Straw man. I said "Internet, mail and web", i.e. I mentioned the web
*last*. (MS-)Windows does *not* pre-date the Internet, let alone mail.

[...]

[Mail Man]

Post by Frank Slootweg

There were millions of PCs running Windows before there was a
World Wide Web.

Straw man. I said "Internet, mail and web", i.e. I mentioned the
web *last*. (MS-)Windows does *not* pre-date the Internet, let
alone mail.

The internet and SMTP e-mail had nothing to do with PC's becoming
popular for personal and home / soho uses.

PC's running dos, Win-3.x and Win-95 became popular without having an
internet connection nor e-mail available to their owners.

As I said, and which you keep ignoring, there were other messaging
systems available for PC's in the late 1970's and during the 1980's that
had nothing to do with the internet or systems running unix.

You realize this, because you refuse to label those systems from the
late 1970's and early 80's as "personal computers" - for reasons which
you can't coherently defend.

[Sam]

Post by Mail Man

Post by Frank Slootweg

There were millions of PCs running Windows before there was a
World Wide Web.

Straw man. I said "Internet, mail and web", i.e. I mentioned the
web *last*. (MS-)Windows does *not* pre-date the Internet, let
alone mail.

The internet and SMTP e-mail had nothing to do with PC's becoming
popular for personal and home / soho uses.

Indeed. Without the Internet and SMTP e-mail, PCs would be just as popular
as they are now.

Post by Mail Man
PC's running dos, Win-3.x and Win-95 became popular without having an
internet connection nor e-mail available to their owners.

Gee, I guess that Trumpet Winsock was a figment of my imagination.

Post by Mail Man
As I said, and which you keep ignoring, there were other messaging
systems available for PC's in the late 1970's and during the 1980's that
had nothing to do with the internet or systems running unix.

Yes, I forgot -- AOL and Compuserve, in their early days, ran on DOS.

[Frank Slootweg]

Post by Mail Man

Post by Frank Slootweg

There were millions of PCs running Windows before there was a
World Wide Web.

Straw man. I said "Internet, mail and web", i.e. I mentioned the
web *last*. (MS-)Windows does *not* pre-date the Internet, let
alone mail.

The internet and SMTP e-mail had nothing to do with PC's becoming
popular for personal and home / soho uses.

And that's *another* straw man.

Post by Mail Man
PC's running dos, Win-3.x and Win-95 became popular without having an
internet connection nor e-mail available to their owners.
As I said, and which you keep ignoring, there were other messaging
systems available for PC's in the late 1970's and during the 1980's that
had nothing to do with the internet or systems running unix.
You realize this, because you refuse to label those systems from the
late 1970's and early 80's as "personal computers" - for reasons which
you can't coherently defend.

And that's *yet* another straw man.

But more to the point: Exactly which part of "let's not quibble about
that, because that wasn't the point I was making" wasn't coherent enough
for you?

You keep side-stepping that my response was to your "If it wasn't for
millions of PC's running windows, then most servers wouldn't need to
exist.", i.e *millions* and *Windows* (not DOS).

So please stick to one issue at a time, perhaps then there can be some
progress.

[Jonathan de Boyne Pollard]

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
<title></title>
</head>
<body bgcolor="#ffffff" text="#000000">
<blockquote cite="mid:***@Man.com" type="cite">
<blockquote type="cite">
<p wrap="">I'm not really familiar with what goes on in the Windows
world, but on Unix this is all very simple. [... Unix explained ...]<br>
</p>
</blockquote>
<p wrap="">Any given machine (unix or otherwise) must know two things:<br>
a) it's "normal" GMT offset, as dictated by it's timezone<br>
b) it's daylight-savings-time strategy - if any.<br>
</p>
</blockquote>
<p>No, it mustn't.&nbsp; Sam just gave you an explanation of the way that
Unix works, which doesn't involve a machine-wide "GMT offset" nor a
machine-wide "DST strategy" at all.&nbsp; Timezone offsets and DST rules are
purely <em>process-local</em> things, that can vary from process to
process, per the setting of an environment variable.<br>
</p>
<p>Windows and Unix have very different paradigms, here, with Windows
still firmly entrenched in <a
href="http://homepage.ntlworld.com./jonathan.deboynepollard/FGA/keeping-time-in-dos.html">the
DOS Think way of keeping time</a>, where these things are machine-wide
and affect the hardware RTC.&nbsp; This is of course why Sam's response
really has little bearing on what happens in Win32 MUAs, albeit that
xyr overall point &#8212; that MUAs just call the API functions for obtaining
local time and the current local time offset from UTC, use whatever
they're given, and the date header comes out fine &#8212; still applies.</p>
</body>
</html>

[David Schwartz]

On Apr 1, 11:16 am, Jonathan de Boyne Pollard <J.deBoynePollard-

No, it mustn't.  Sam just gave you an explanation of the way that Unix works,
which doesn't involve a machine-wide "GMT offset" nor a machine-wide
"DST strategy" at all.  Timezone offsets and DST rules are purely
process-local things, that can vary from process to process, per the
setting of an environment variable.

Just to clarify, pretty much every UNIX system *does* have a machine-
wide GMT offset and a machine-wide DST strategy. They just use them
only in very rare cases where there is no other option. They consider
this a horrible, awful thing implemented only because some things can
be done no other way -- for example, implementing FAT32 in the kernel.

From Linux (example code from linux/time.h and fs/udf/udftime.c):

struct timezone {
int tz_minuteswest; /* minutes west of Greenwich */
int tz_dsttime; /* type of dst correction */
};
[...]
extern struct timezone sys_tz;
[...]
offset = -sys_tz.tz_minuteswest;
[...]
ts.tv_sec += offset * 60;

And from netfilter/xt_time.c:
if (minutes < 0) /* east of Greenwich */
printk(KERN_INFO KBUILD_MODNAME
": kernel timezone is +%02d%02d\n",
-minutes / 60, -minutes % 60);
else /* west of Greenwich */
printk(KERN_INFO KBUILD_MODNAME
": kernel timezone is -%02d%02d\n",
minutes / 60, minutes % 60);

DS

[Jonathan de Boyne Pollard]

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
<blockquote
cite="mid:***@commodore.email-scan.com"
type="cite">
<p>I'm not really familiar with what goes on in the Windows world,
but on Unix this is all very simple. </p>
</blockquote>
<p><a
href="http://homepage.ntlworld.com./jonathan.deboynepollard/FGA/keeping-time-in-dos.html">Here's
something</a> that may familiarize you to an extent.&nbsp; The MSDN Library
covers the system APIs such as <a
href="http://msdn.microsoft.com/en-us/library/ms724421%28VS.85%29.aspx"><code>GetTimeZoneInformation()</code></a>
and <a
href="http://msdn.microsoft.com/en-us/library/ms724338%28VS.85%29.aspx"><code>GetLocalTime()</code></a>,
and <a
href="http://msdn.microsoft.com/en-us/library/ms724493%28VS.85%29.aspx">the
Win32 concept of local time</a>.<br>
</p>
</body>
</html>