Hi Giles,
That’s quite weird actually! Maybe your Qt4 installation is messed up… On my machine:
ls /usr/local/Trolltech/Qt-4.3.0/mkspecs
aix-g++ freebsd-g++40 irix-cc linux-icc-64 netbsd-g++ tru64-cxx
aix-g++-64 freebsd-icc irix-cc-64 linux-kcc openbsd-g++ tru64-g++
aix-xlc hpux-acc irix-g++ linux-lsb-g++ qconfig.pri unixware-cc
aix-xlc-64 hpux-acc-64 irix-g++-64 linux-pgcc qws unixware-g++
common hpux-acc-o64 linux-cxx lynxos-g++ sco-cc win32-g++
darwin-g++ hpux-g++ linux-ecc-64 macx-g++ sco-g++
default hpux-g++-64 linux-g++ macx-icc solaris-cc
features hpuxi-acc linux-g++-32 macx-pbuilder solaris-cc-64
freebsd-g++ hpuxi-acc-64 linux-g++-64 macx-xcode solaris-g++
freebsd-g++34 hurd-g++ linux-icc macx-xlc solaris-g++-64
Where default is a symlink to linux-g++ - that would explain why I do not need to set QMAKESPEC. I think that qmake determines the QTDIR from the location of the qmake binary and then just looks at mkspecs/default.
What I find strange is that you added a symlink to qt3’s linux-g++-64 in your /usr/share/qt4/mkspecs. Don’t you have a bunch of dirs there, just like I do (see above)?
If you want to checkout the latest revision from the dev branch:
bzr branch http://bzr.calitko.org/calitko
then you could use bzr pull to pull only the new revisions.
About the compile error, the code is:
p->destroyedMapper.setMapping (session, reinterpret_cast
I did a search on the Internet and it turns out that the pointer is 64bit but int is still 32 bit (that seems to be the LP64 model).
A solution would be to cast to long rather then to an int. That will work also on 32bit platforms, where long is 32bit. However, I read that under Win64 long is still 32bit - one should use long long instead. I personally feel I’ll need to read more…
Maybe somebody already has experience with the 64 bit issues and would know how to portably solve that! Having a look at Qt4’s source code could also be an idea.
Regards,
Peter