Recording of CTime Object
Serge Potteck -- serge@lune.cst.cnes.fr
Friday, August 23, 1996
Environment : VC++ 1.51, Windows 3.11.
What is put on the disk by
CTime time;
ar<
Gerry Sweeney -- gerry@hornbill.com
Tuesday, August 27, 1996
Environment : VC++ 1.51, Windows 3.11.
Serge,
I am not to sure if I understand your question but here goes. When you
archive a CTime object, a 32 bit number of type 'time_t' which is basically
a 'long'. This holds the number of seconds since January 1, 1970. See the
help on the 'C' runtime library function 'time()'. As for the different
computers, A MAC and a PC for example use different processors. (MAC use
Motorola, PC use Intel). These processors store there numbers in memory in
different order. This is known as big-endian and little-endian. Because the
serialize() function in MFC writes a binary copy of a 'time_t' variable as a
sequence of 32 bits this file would not work between computers with
processors using different endian formats. One way around this is to convert
the CTime object to a string and then store it. For example
void CBoo::Serialize(CArchive ar)
{
if(ar.IsStoring())
{
// ar << time;
char buff[34];
CString tmp = _ltoa(tim.GetTime(), buff, 10);
ar << tmp;
}
else
{
// ar >> time;
CString tmp;
time = (time_t)_atol((const char *)tmp);
}
}
This would store the time as a string representation of the time value. I am
not sure of the 'time()' runtime functions on different platforms have the
same base time. You could store the time in a format like 'ddmmyyyyhhmmss'
and use sscanf to get the information back. This would be pretty portable.
I do hope this is the sort of thing you were looking for. If I have
misunderstood your question please forgive me
Gerry Sweeney
----------
From: owner-mfc-l
To: mfc-l
Subject: Recording of CTime Object
Date: 23 August 1996 17:18
Environment : VC++ 1.51, Windows 3.11.
What is put on the disk by
CTime time;
ar<
pjn -- pjn@indigo.ie
Tuesday, September 03, 1996
On Fri, 23 Aug 1996 17:18:13 -0400, you wrote:
>Environment : VC++ 1.51, Windows 3.11.
>
>What is put on the disk by
>
> CTime time;
> ar<
>depends on the processor you use.
>
>I could find two classes of processors understanding
>eachother perfectly. The oldest computer
>seems to belong to the same class, but I found
>a very recent Pentium belonging to that class too.
>It doesn't depend on the code you make (8086 or 386).
>
>With a parametre in the .ini file, it would
>be easy to bypass the problem, if the shift
>were constant. But it seems not to be, and
>to depend on the application.
>
>I am very sorry, because that problem is
>probably well known. But would
>anybody accept to help me ?
>
>Thanks
>
>Serge Potteck
>National space center
>France
>
>
>
I was told by a friend of mine who used VC 1.0 that the single member
variable that CTime wrapped changed from an unsigned long to a signed
long from VC 1.0 to VC 1.5. If this is the case, then this might be
the cause of the problem. Other than that, I have not came across
other problems with CTime in this area.
'''
@ @
+========================ooO-(_)-Ooo=================================+
| PJ Naughter |
| |
| Software Developer Email: pjn@indigo.ie |
| Softech Telecom Tel: +353-1-2958384 |
| Fax: +353-1-2956290 |
| Author of DTime - A Collection URL: http://indigo.ie/~pjn |
| of Date & Time classes for MFC |
| |
| Addr: 7 Woodford, Brewery Road, Blackrock, |
| Co. Dublin, Republic of Ireland |
+====================================================================+
Gerry Sweeney -- gerry@hornbill.com
Friday, September 06, 1996
Environment : VC++ 1.51, Windows 3.11.
I do not know if that is the case, but even if it is it should not matter.
We have not reached a time that would effect this. We are on appx
0x3*******. The most significant bit is used to indicate a (-)negative
number so It should make no difference.
Gerry
----------
From: owner-mfc-l
To: mfc-l
Subject: Re: Recording of CTime Object
Date: 03 September 1996 20:28
On Fri, 23 Aug 1996 17:18:13 -0400, you wrote:
>Environment : VC++ 1.51, Windows 3.11.
>
>What is put on the disk by
>
> CTime time;
> ar<
>depends on the processor you use.
>
>I could find two classes of processors understanding
>eachother perfectly. The oldest computer
>seems to belong to the same class, but I found
>a very recent Pentium belonging to that class too.
>It doesn't depend on the code you make (8086 or 386).
>
>With a parametre in the .ini file, it would
>be easy to bypass the problem, if the shift
>were constant. But it seems not to be, and
>to depend on the application.
>
>I am very sorry, because that problem is
>probably well known. But would
>anybody accept to help me ?
>
>Thanks
>
>Serge Potteck
>National space center
>France
>
>
>
I was told by a friend of mine who used VC 1.0 that the single member
variable that CTime wrapped changed from an unsigned long to a signed
long from VC 1.0 to VC 1.5. If this is the case, then this might be
the cause of the problem. Other than that, I have not came across
other problems with CTime in this area.
'''
@ @
+========================ooO-(_)-Ooo=================================+
| PJ Naughter |
| |
| Software Developer Email: pjn@indigo.ie |
| Softech Telecom Tel: +353-1-2958384 |
| Fax: +353-1-2956290 |
| Author of DTime - A Collection URL: http://indigo.ie/~pjn |
| of Date & Time classes for MFC |
| |
| Addr: 7 Woodford, Brewery Road, Blackrock, |
| Co. Dublin, Republic of Ireland |
+====================================================================+
| Вернуться в корень Архива
|