The argument to ftime is a
struct timeb *
hence :
::ftime (& tb);
and not
::ftime (* tb);
3. In line 120 of PatternLayout.cpp, the compiler fires an error for "ambiguous operator <<"
I guess we need to recast the t variable to a (int), but I'm not sure of that
Sorry to bother y'all with this Micro$oft stuff
:-))
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Thanks for your input. log4cpp is supposed to support MSVC6, so your comments aren't a 'bother' at all.
I've adopted 1 and 2. For point 3, casting to an int will lead to cycling of the stamp within a month. This may not be a problem on Win32, but Unix daemons tend to run longer than that :-).
According to MS itself it is a bug in the MSVC C++ library, see: http://support.microsoft.com/default.aspx?scid=kb;EN-US;q168440
I've included their workaround.
Please try the fixes and let me know if they are OK. You can get the updated files from CVS, or if that is too difficult I can send you a fresh tar ball. Just ask...
Regards,
Bastiaan
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Here are some compile-time errors I get when trying the build the (0.2.6 rc 4) project :
1. int64_t is undefined under Micro$oft.
But __int64 is.
I propose to change the declaration in config-win32.h to
#ifndef LOG4CPP_HAVE_INT64_T
#define LOG4CPP_HAVE_INT64_T
#define int64_t __int64
#endif
2. There's an error in line 37 of TimeStamp.cpp
The argument to ftime is a
struct timeb *
hence :
::ftime (& tb);
and not
::ftime (* tb);
3. In line 120 of PatternLayout.cpp, the compiler fires an error for "ambiguous operator <<"
I guess we need to recast the t variable to a (int), but I'm not sure of that
Sorry to bother y'all with this Micro$oft stuff
:-))
Hi Yves,
Thanks for your input. log4cpp is supposed to support MSVC6, so your comments aren't a 'bother' at all.
I've adopted 1 and 2. For point 3, casting to an int will lead to cycling of the stamp within a month. This may not be a problem on Win32, but Unix daemons tend to run longer than that :-).
According to MS itself it is a bug in the MSVC C++ library, see: http://support.microsoft.com/default.aspx?scid=kb;EN-US;q168440
I've included their workaround.
Please try the fixes and let me know if they are OK. You can get the updated files from CVS, or if that is too difficult I can send you a fresh tar ball. Just ask...
Regards,
Bastiaan
Bastiaan,
Your new version is <b>great</b>. Fixed all that.
I put a
#define LOG4CPP_USE_CLEANUP
in config-win32.h
This worked perfectly. BoundChecker doesn't blink.
Now, I'll try to <i>use</i> your tool ...
Have a great day
Yves