Re: problem with time-stamps on GNU/Linux and Windows

Re: problem with time-stamps on GNU/Linux and Windows

[Jason Rumney]

On Sep 4, 7:02 pm, Seweryn Kokot <s.ko...@po.opole.pl> wrote:
> Hello,
>
> Could you explain to me why I get slightly different time-stamps under
> Windows and Linux?

These localized weekday names come from the system. No amount of
configuration within Emacs will make them the same in Windows as they
are in GNU/Linux.

Re: problem with time-stamps on GNU/Linux and Windows

[ken]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1



On 09/04/2008 09:00 AM Jason Rumney wrote:
> On Sep 4, 7:02 pm, Seweryn Kokot <s.ko...@po.opole.pl> wrote:
>> Hello,
>>
>> Could you explain to me why I get slightly different time-stamps under
>> Windows and Linux?
> 
> These localized weekday names come from the system. No amount of
> configuration within Emacs will make them the same in Windows as they
> are in GNU/Linux.

I'm not quite understanding your problem.  And I don't often use
Windows.  But I would think that emacs should fetch the same couple
words (i.e., two bytes) representing the time regardless of which OS it
is running on.  But I'm guessing that the problem isn't the accuracy of
the time, but rather the human-readable output derived from those words.

I trust that you've read and understand the help on
"format-time-string"...  i.e.,

C-h f format-time-string [Return]


hth,
ken

- --
The significant problems we face cannot be solved at the
same level of thinking we were at when we created them.
	-- Albert Einstein
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)

iD8DBQFIwBpk8CeNiFrQkecRAqiUAJ9KqePfBDQBzKwH35JAEYpprqpYRQCfTBMl
SRmyj/ePHpuTgTJ3Ha1BXXg=
=HQqT
-----END PGP SIGNATURE-----

Re: problem with time-stamps on GNU/Linux and Windows

[Eli Zaretskii]

> Date: Thu, 04 Sep 2008 13:27:00 -0400
> From: ken <gebser@mousecar.com>
> Cc: Help-gnu-emacs@gnu.org
> 
> >> Could you explain to me why I get slightly different time-stamps under
> >> Windows and Linux?
> > 
> > These localized weekday names come from the system. No amount of
> > configuration within Emacs will make them the same in Windows as they
> > are in GNU/Linux.
> 
> I'm not quite understanding your problem.  And I don't often use
> Windows.  But I would think that emacs should fetch the same couple
> words (i.e., two bytes) representing the time regardless of which OS it
> is running on.

The problem is not the time, but the abbreviated name of the second
day of the week ("Tue" in English).  These abbreviated names come from
a call to a library function, which are different on Windows and on
GNU/Linux, so they return different strings.

> But I'm guessing that the problem isn't the accuracy of
> the time, but rather the human-readable output derived from those words.

Emacs does not derive the names from those words, it simply returns
whatever the library functions hand it.

Re: problem with time-stamps on GNU/Linux and Windows

[Seweryn Kokot]

Eli Zaretskii <eliz@gnu.org> writes:

>> >> Could you explain to me why I get slightly different time-stamps under
>> >> Windows and Linux?
>> > 
>> > These localized weekday names come from the system. No amount of
>> > configuration within Emacs will make them the same in Windows as they
>> > are in GNU/Linux.
>> 
>> I'm not quite understanding your problem.  And I don't often use
>> Windows.  But I would think that emacs should fetch the same couple
>> words (i.e., two bytes) representing the time regardless of which OS it
>> is running on.
>
> The problem is not the time, but the abbreviated name of the second
> day of the week ("Tue" in English).  These abbreviated names come from
> a call to a library function, which are different on Windows and on
> GNU/Linux, so they return different strings.
>
>> But I'm guessing that the problem isn't the accuracy of
>> the time, but rather the human-readable output derived from those words.
>
> Emacs does not derive the names from those words, it simply returns
> whatever the library functions hand it.

Well, (current-time-string) function returns the same strings on both systems,
but in English, for example: "Thu Sep  4 20:16:51 2008". BTW why this
function gives abbreviations in English and not the locale's ones?

Whereas this function

(defun my-insert-time-stamp ()
(interactive)
(insert (format-time-string "%a %b %d %02H:%02M:%02S %Y")))

on Windows gives
(my-insert-time-stamp)
Cz wrz 04 20:13:20 2008

and on GNU/Linux:
czw wrz 04 20:16:29 2008

I raise this problem because org-mode has some problems parsing these
inconsistent abbreviations.

Is it possible to get the result of (my-insert-time-stamp) in English
like in the case of the (current-time-string) function?

regards,
Seweryn

Re: problem with time-stamps on GNU/Linux and Windows

[ken]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1



On 09/04/2008 02:37 PM Seweryn Kokot wrote:
> Eli Zaretskii <eliz@gnu.org> writes:
> 
>>>>> Could you explain to me why I get slightly different time-stamps under
>>>>> Windows and Linux?
>>>> These localized weekday names come from the system. No amount of
>>>> configuration within Emacs will make them the same in Windows as they
>>>> are in GNU/Linux.
>>> I'm not quite understanding your problem.  And I don't often use
>>> Windows.  But I would think that emacs should fetch the same couple
>>> words (i.e., two bytes) representing the time regardless of which OS it
>>> is running on.
>> The problem is not the time, but the abbreviated name of the second
>> day of the week ("Tue" in English).  These abbreviated names come from
>> a call to a library function, which are different on Windows and on
>> GNU/Linux, so they return different strings.
>>
>>> But I'm guessing that the problem isn't the accuracy of
>>> the time, but rather the human-readable output derived from those words.
>> Emacs does not derive the names from those words, it simply returns
>> whatever the library functions hand it.
> 
> Well, (current-time-string) function returns the same strings on both systems,
> but in English, for example: "Thu Sep  4 20:16:51 2008". BTW why this
> function gives abbreviations in English and not the locale's ones?
> 
> Whereas this function
> 
> (defun my-insert-time-stamp ()
> (interactive)
> (insert (format-time-string "%a %b %d %02H:%02M:%02S %Y")))
> 
> on Windows gives
> (my-insert-time-stamp)
> Cz wrz 04 20:13:20 2008
> 
> and on GNU/Linux:
> czw wrz 04 20:16:29 2008
> 
> I raise this problem because org-mode has some problems parsing these
> inconsistent abbreviations.
> 
> Is it possible to get the result of (my-insert-time-stamp) in English
> like in the case of the (current-time-string) function?
> 
> regards,
> Seweryn

All I can say is that we're working with open source code.  It took a
little bit of time, but I just tracked down at least one place where the
dayname strings are defined.  On my system it's

(defvar calendar-day-name-array
  ["Sunday" "Monday" "Tuesday" "Wednesday" "Thursday" "Friday" "Saturday"]
  "Array of capitalized strings giving, in order, the day names.")

(and apparently other places) in calendar.el.

I.e., just go in and change the names of the days to what you want.


Tell us if that does it... I'd be interested.

hth,
ken
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)

iD8DBQFIwDFW8CeNiFrQkecRAu+hAJ9H0pbac0fQI3/hxuBgZBbEQfDiNQCfabfj
FSODnxVVIKKRKMaWvHTIcn8=
=ymLD
-----END PGP SIGNATURE-----

Re: problem with time-stamps on GNU/Linux and Windows

[Seweryn Kokot]

ken <gebser@mousecar.com> writes:

> All I can say is that we're working with open source code.  It took a
> little bit of time, but I just tracked down at least one place where the
> dayname strings are defined.  On my system it's
>
> (defvar calendar-day-name-array
>   ["Sunday" "Monday" "Tuesday" "Wednesday" "Thursday" "Friday" "Saturday"]
>   "Array of capitalized strings giving, in order, the day names.")
>
> (and apparently other places) in calendar.el.
>
> I.e., just go in and change the names of the days to what you want.
>
>
> Tell us if that does it... I'd be interested.
>
> hth,
> ken

I have tried this on Windows and nothing changed. I even tried to customize
`calendar-day-abbrev-array' variable but still format-time-string %a gives
"Cz" for "Thu", also `calendar-abbrev-length' is set to 3. 

Seweryn

Re: problem with time-stamps on GNU/Linux and Windows

[Giorgos Keramidas]

On Thu, 04 Sep 2008 20:37:26 +0200, Seweryn Kokot <s.kokot@po.opole.pl> wrote:
> Well, (current-time-string) function returns the same strings on both
> systems, but in English, for example: "Thu Sep 4 20:16:51 2008". BTW
> why this function gives abbreviations in English and not the locale's
> ones?

Because this way the output is predictably formatted, and programs can
read it with a well-known, predefined format.  The `format-time-string'
function can be used for locale-aware results.

> Whereas this function
>
> (defun my-insert-time-stamp ()
>   (interactive)
>   (insert (format-time-string "%a %b %d %02H:%02M:%02S %Y")))
>
> on Windows gives
> (my-insert-time-stamp)
> Cz wrz 04 20:13:20 2008
>
> and on GNU/Linux:
> czw wrz 04 20:16:29 2008
>
> I raise this problem because org-mode has some problems parsing these
> inconsistent abbreviations.
>
> Is it possible to get the result of (my-insert-time-stamp) in English
> like in the case of the (current-time-string) function?

One way to get predictable timestamps in `my-insert-time-stamp' is to
temporarily alter the value of `system-time-locale':

    (defun my-insert-time-stamp ()
      (interactive)
      (insert (let ((system-time-locale "C"))
                (format-time-string "%a %b %d %02H:%02M:%02S %Y"))))

This will always use the "C" locale when displaying time, and the output
inserted by `my-insert-time-stamp' will be similar to:

    "Fri Sep 05 03:34:47 2008"

To make the timestamps inserted independent of the local timezone, it
may also be worth to call `format-time-string' with a third argument of
`T' to output the timestamp in UTC:

    (defun my-insert-time-stamp ()
      (interactive)
      (insert (let ((system-time-locale "C"))
                (format-time-string "%a %b %d %02H:%02M:%02S %Y" (current-time) t))))

Re: problem with time-stamps on GNU/Linux and Windows

[Seweryn Kokot]

ken <gebser@mousecar.com> writes:

> On 09/04/2008 09:00 AM Jason Rumney wrote:
>> On Sep 4, 7:02 pm, Seweryn Kokot <s.ko...@po.opole.pl> wrote:
>>> Hello,
>>>
>>> Could you explain to me why I get slightly different time-stamps under
>>> Windows and Linux?
>> 
>> These localized weekday names come from the system. No amount of
>> configuration within Emacs will make them the same in Windows as they
>> are in GNU/Linux.
>
> I'm not quite understanding your problem.  And I don't often use
> Windows.  But I would think that emacs should fetch the same couple
> words (i.e., two bytes) representing the time regardless of which OS it
> is running on.  But I'm guessing that the problem isn't the accuracy of
> the time, but rather the human-readable output derived from those words.
>
> I trust that you've read and understand the help on
> "format-time-string"...  i.e.,
>
> C-h f format-time-string [Return]

I had read it before. I referred to %a from this help buffer. And the
problem is with the abbreviated name of the day of week. Under Linux I
get three-letter abbreviations (e.g. "wto" - "Tue") and under Windows -
two-letter abbrev (e.g. "Wt").

What I want is the three-letter abbreviation.

regards,
Seweryn

problem with time-stamps on GNU/Linux and Windows

[Seweryn Kokot]

Hello,

Could you explain to me why I get slightly different time-stamps under
Windows and Linux? 

This is what I get in Linux

2008-08-19 wto 07:26

and under Windows I get

2008-08-05 Wt 11:44

Namely the problem is with %a - abbreviated day of week.

Here are the settings under Linux:

(locale-info 'codeset): "ISO-8859-2"
locale-coding-system: iso-latin-2

and under Windows I have

locale-coding-system: iso-latin-2
system-time-locale: iso-latin-2
system-message-locale: iso-latin-2
(locale-info 'codeset): "cp1250"

Under Linux iso-latin-2 is set by default and on Windows I have to put
the following in .emacs file

(setq locale-coding-system 'iso-latin-2)
(setq system-message-locale 'iso-latin-2)
(setq system-time-locale 'iso-latin-2)

otherwise these values are cp1250, nil and nil

I'm wondering why setting on windows (setq locale-coding-system 'iso-latin-2) has
no impact on (locale-info 'codeset) which still gives "cp1250"?

Any idea what should I do to get the same results on both operating systems?

Thanks in advance,
regards,
Seweryn