[Libpqxx-general] Problem with the tests using Visual Studio 2008 and libpqxx version 3.0

[Jeroen Vermeulen]

Andrew Maclean wrote:

> Things seem to build Ok in debug mode but I am still having trouble in
> release mode. If I get some time I will look further into it. I have
> switched to using the subversion repository.

I can't find anything that would explain a difference between debug and 
release mode, assuming you're not getting stale object files or 
something like that.  Maybe TestStaticRelease and/or TestDllRelease 
haven't been created?

If you're using the svn tree on a Windows system, do you run the 
tools/template2mak.py directly to generate these Makefiles?  Or do you 
use something like Cygwin and autogen.sh?


> CMake will let you have a file like configure.h.in and depending upon
> the cmake process it will configure a file called configure.h. Look at
> http://www.cmake.org/Wiki/CMake, in particular
> http://www.cmake.org/Wiki/CMake_HowToDoPlatformChecks. If you are
> interested I have an example of this usage.

The biggest problem is that, realistically, I don't think I'll have time 
to figure it out!


> This is a constant problem on Windows (the drive names) if you look at
> the CMake scripts, you can see that you can set an environment
> variable LIBPQXX_ROOT to point to the root of the directory. Similarly
> for FindPostgresql.cmake you can set the environment variable
> POSTGRESQL_ROOT. If these are set, FindLibpqxx and FindPostgresql will
> use these in preference to searching. A lot of the FindXXX scripts on
> the modules subdirectory of CMake do this too.

Is adding the right pg_config script to the execution path not an option 
with Windows?


Jeroen