Here it's again:

//------------------------
import std.date;
import std.stdio;

void main()
{
writefln("UTC  :%s", toTimeString(getUTCtime()));
writefln("Local:%s", toTimeString(UTCtoLocalTime(getUTCtime())));
}
//------------------------

:::: gdc 0.17, Mac OS X 10.4.6

$ ./test
UTC  :06:20:52 GMT+0500
Local:11:20:52 GMT+0500

$ date
Mon May  8 20:21:01 ECT 2006


:::: DMD 0.155, Windows XP SP 2

D:\>test
UTC  :01:27:19 GMT+0000
Local:01:27:19 GMT+0000

D:\>time
La hora actual es: 20:27:32,08


:::: DMD 0.155, Mandrake Linux 10.1

$ ./test
UTC  :20:34:23 GMT-0500
Local:15:34:23 GMT-0500

$ date
lun may  8 20:34:27 ECT 2006


Walter, at least take a look at UTC on Linux: that's the correct local time, which most certainly gives a clue of where the solution might be.

In all my systems, the timezone is GMT -0500 aka Bogotá, Lima, Quito, aka ECT.


-- 
Carlos Santander Bernal

Carlos Santander wrote:
> Here it's again:
> 
> //------------------------
> import std.date;
> import std.stdio;
> 
> void main()
> {
> writefln("UTC  :%s", toTimeString(getUTCtime()));
> writefln("Local:%s", toTimeString(UTCtoLocalTime(getUTCtime())));
> }
> //------------------------
> 
> :::: gdc 0.17, Mac OS X 10.4.6
> 
> $ ./test
> UTC  :06:20:52 GMT+0500
> Local:11:20:52 GMT+0500
> 
> $ date
> Mon May  8 20:21:01 ECT 2006
> 
> 
> :::: DMD 0.155, Windows XP SP 2
> 
> D:\>test
> UTC  :01:27:19 GMT+0000
> Local:01:27:19 GMT+0000
> 
> D:\>time
> La hora actual es: 20:27:32,08
> 
> 
> :::: DMD 0.155, Mandrake Linux 10.1
> 
> $ ./test
> UTC  :20:34:23 GMT-0500
> Local:15:34:23 GMT-0500
> 
> $ date
> lun may  8 20:34:27 ECT 2006
> 
> 
> Walter, at least take a look at UTC on Linux: that's the correct local time, which most certainly gives a clue of where the solution might be.
> 
> In all my systems, the timezone is GMT -0500 aka Bogotá, Lima, Quito, aka ECT.
> 
> 

You might want to use mango.locale. http://dsource.org/projects/mango/browser/trunk/mango/locale?rev=883

-- 
Kyle Furlong // Physics Undergrad, UCSB

"D is going wherever the D community wants it to go." - Walter Bright

I am getting AssertError when running your example on both Linux and Windows:
$ ./std_date_01.exe
UTC  :12:48:08 GMT+0200
Local:14:48:08 GMT+0200
Error: AssertError Failure std_date_01(8)

DMD version:   0.156
compiled with: dmd std_date_01.d

Kyle Furlong escribió:
> Carlos Santander wrote:
>> Here it's again:
>>
>> //------------------------
>> import std.date;
>> import std.stdio;
>>
>> void main()
>> {
>> writefln("UTC  :%s", toTimeString(getUTCtime()));
>> writefln("Local:%s", toTimeString(UTCtoLocalTime(getUTCtime())));
>> }
>> //------------------------
>>
>> :::: gdc 0.17, Mac OS X 10.4.6
>>
>> $ ./test
>> UTC  :06:20:52 GMT+0500
>> Local:11:20:52 GMT+0500
>>
>> $ date
>> Mon May  8 20:21:01 ECT 2006
>>
>>
>> :::: DMD 0.155, Windows XP SP 2
>>
>> D:\>test
>> UTC  :01:27:19 GMT+0000
>> Local:01:27:19 GMT+0000
>>
>> D:\>time
>> La hora actual es: 20:27:32,08
>>
>>
>> :::: DMD 0.155, Mandrake Linux 10.1
>>
>> $ ./test
>> UTC  :20:34:23 GMT-0500
>> Local:15:34:23 GMT-0500
>>
>> $ date
>> lun may  8 20:34:27 ECT 2006
>>
>>
>> Walter, at least take a look at UTC on Linux: that's the correct local time, which most certainly gives a clue of where the solution might be.
>>
>> In all my systems, the timezone is GMT -0500 aka Bogotá, Lima, Quito, aka ECT.
>>
>>
> 
> You might want to use mango.locale. http://dsource.org/projects/mango/browser/trunk/mango/locale?rev=883
> 

I know there're other options, but that's no reason for Phobos to be wrong.

-- 
Carlos Santander Bernal

Dejan Lekic escribió:
> I am getting AssertError when running your example on both Linux and Windows:
> $ ./std_date_01.exe
> UTC  :12:48:08 GMT+0200
> Local:14:48:08 GMT+0200
> Error: AssertError Failure std_date_01(8)
> 
> DMD version:   0.156
> compiled with: dmd std_date_01.d

Are you sure you didn't change it to "int main()" and forgot to add a return?

-- 
Carlos Santander Bernal

Carlos Santander wrote:
> Here it's again:
> 
> //------------------------
> import std.date;
> import std.stdio;
> 
> void main()
> {
> writefln("UTC  :%s", toTimeString(getUTCtime()));
> writefln("Local:%s", toTimeString(UTCtoLocalTime(getUTCtime())));
> }
> //------------------------
> 
> :::: gdc 0.17, Mac OS X 10.4.6
> 
> $ ./test
> UTC  :06:20:52 GMT+0500
> Local:11:20:52 GMT+0500
> 
> $ date
> Mon May  8 20:21:01 ECT 2006
> 
> 
> :::: DMD 0.155, Windows XP SP 2
> 
> D:\>test
> UTC  :01:27:19 GMT+0000
> Local:01:27:19 GMT+0000
> 
> D:\>time
> La hora actual es: 20:27:32,08
> 
> 
> :::: DMD 0.155, Mandrake Linux 10.1
> 
> $ ./test
> UTC  :20:34:23 GMT-0500
> Local:15:34:23 GMT-0500
> 
> $ date
> lun may  8 20:34:27 ECT 2006
> 
> 
> Walter, at least take a look at UTC on Linux: that's the correct local time, which most certainly gives a clue of where the solution might be.
> 
> In all my systems, the timezone is GMT -0500 aka Bogotá, Lima, Quito, aka ECT.

So you're saying the linux results are correct?

Walter Bright escribió:
> Carlos Santander wrote:
>> Here it's again:
>>
>> //------------------------
>> import std.date;
>> import std.stdio;
>>
>> void main()
>> {
>> writefln("UTC  :%s", toTimeString(getUTCtime()));
>> writefln("Local:%s", toTimeString(UTCtoLocalTime(getUTCtime())));
>> }
>> //------------------------
>>
>> :::: gdc 0.17, Mac OS X 10.4.6
>>
>> $ ./test
>> UTC  :06:20:52 GMT+0500
>> Local:11:20:52 GMT+0500
>>
>> $ date
>> Mon May  8 20:21:01 ECT 2006
>>
>>
>> :::: DMD 0.155, Windows XP SP 2
>>
>> D:\>test
>> UTC  :01:27:19 GMT+0000
>> Local:01:27:19 GMT+0000
>>
>> D:\>time
>> La hora actual es: 20:27:32,08
>>
>>
>> :::: DMD 0.155, Mandrake Linux 10.1
>>
>> $ ./test
>> UTC  :20:34:23 GMT-0500
>> Local:15:34:23 GMT-0500
>>
>> $ date
>> lun may  8 20:34:27 ECT 2006
>>
>>
>> Walter, at least take a look at UTC on Linux: that's the correct local time, which most certainly gives a clue of where the solution might be.
>>
>> In all my systems, the timezone is GMT -0500 aka Bogotá, Lima, Quito, aka ECT.
> 
> So you're saying the linux results are correct?

No, I'm saying the UTC result on linux is the correct local result.

-- 
Carlos Santander Bernal

Carlos Santander wrote:
> No, I'm saying the UTC result on linux is the correct local result.


It's hard for me to debug this, because it works in my timezone. Want to give it a try in yours?

Carlos Santander wrote:
> Walter Bright escribió:
>> Carlos Santander wrote:
>>> Here it's again:
>>>
>>> //------------------------
>>> import std.date;
>>> import std.stdio;
>>>
>>> void main()
>>> {
>>> writefln("UTC  :%s", toTimeString(getUTCtime()));
>>> writefln("Local:%s", toTimeString(UTCtoLocalTime(getUTCtime())));
>>> }
>>> //------------------------
>>>
>>> :::: gdc 0.17, Mac OS X 10.4.6
>>>
>>> $ ./test
>>> UTC  :06:20:52 GMT+0500
>>> Local:11:20:52 GMT+0500
>>>
>>> $ date
>>> Mon May  8 20:21:01 ECT 2006
>>>
>>>
>>> :::: DMD 0.155, Windows XP SP 2
>>>
>>> D:\>test
>>> UTC  :01:27:19 GMT+0000
>>> Local:01:27:19 GMT+0000
>>>
>>> D:\>time
>>> La hora actual es: 20:27:32,08
>>>
>>>
>>> :::: DMD 0.155, Mandrake Linux 10.1
>>>
>>> $ ./test
>>> UTC  :20:34:23 GMT-0500
>>> Local:15:34:23 GMT-0500
>>>
>>> $ date
>>> lun may  8 20:34:27 ECT 2006
>>>
>>>
>>> Walter, at least take a look at UTC on Linux: that's the correct local time, which most certainly gives a clue of where the solution might be.
>>>
>>> In all my systems, the timezone is GMT -0500 aka Bogotá, Lima, Quito, aka ECT.
>>
>> So you're saying the linux results are correct?
> 
> No, I'm saying the UTC result on linux is the correct local result.
> 

It's GMT + 2 + Daylight saving Time here now:

$ ./test
UTC  :23:27:47 GMT+0300
Local:02:27:47 GMT+0300
$ date
pe 12.5.2006 23:27:58 +0300

The date values have been this way (=wrong) as long as I remember. 'UTC'
in D is the current local time and 'Local' is UTC + 2 * (Local - UTC).

-- 
Jari-Matti

Walter Bright escribió:
> Carlos Santander wrote:
>> No, I'm saying the UTC result on linux is the correct local result.
> 
> 
> It's hard for me to debug this, because it works in my timezone. Want to give it a try in yours?

I'm completely confused about this whole thing, and I think the problem is documentation.

-- On linux (DMD):

If I do toUTCString(getUTCtime()) I correctly get:

Mon, 15 May 2006 17:58:57 UTC

But when I do toTimeString(getUTCtime()) I get:

12:58:57 GMT-0500

So I think toTimeString must be wrong, my logic being that I asked it to return the current time as UTC time. But what it returned was the time for my local timezone. However, it's never explicitly stated that it should behave that way.

-- On mac (GDC):

It's the same as linux, but the timezone seems to be doing the opposite:

UTC  :Mon, 15 May 2006 18:07:44 UTC
Local:04:07:44 GMT+0500

So it performed an addition instead of a difference.

-- On windows (DMD):

toTimeString(getUTCtime()) returns:

18:13:14 GMT+0000

And toUTCString(getUTCtime()):

Mon, 15 May 2006 18:13:14 UTC

So, on windows, these are ok, I think. At least they behave as I expected them to. If I'm right with this, then toTimeString on linux is definitely wrong.

Now, toTimeString(UTCtoLocalTime(getUTCtime())) also returns:

18:13:14 GMT+0000

So, unless toTimeString always prints UTC time (which I don't think it should), this is obviously wrong, and the problem has to be in the timezone conversion.

Part of the solution might be here: http://www.digitalmars.com/drn-bin/wwwnews?D/20470, but if GetTimeZoneInformation() fails, what if instead you use GetLocalTime() to get the local time, and since getUTCtime() is ok, then you could use those two values to get the timezone (UTCtime - localtime). Does that make sense?

-- Finally....

It's easy, in fact, for you to test it too: change your timezone to different parts of the world when debugging this and see what happens. It's really easy to do in both linux and windows.

Sorry if that sounded harsh, but I think right now std.date doesn't work in a usable way outside your own timezone and maybe a couple more, and that is a shame for D, especially since these problems have been around for so long.

-- 
Carlos Santander Bernal