On Fri, Aug 19, 2016 at 03:22:52PM +0200, Kasper F. Brandt wrote: > On Fri, Aug 19, 2016 at 12:53 PM, Oswald Buddenhagen > <[email protected]> wrote: > > On Thu, Aug 18, 2016 at 04:26:30PM +0200, Kasper F. Brandt wrote: > > > In d8be811[0] qt_build_extra was made to skip the inclusion of > > > EXTRA_DEFINES, EXTRA_INCLUDEPATH and EXTRA_LIBS into resp. > > > DEFINES, INCLUDEPATH and INCLUDEPATH. This means that if you e.g. try to > > > specify the include path to zlib with -I <path/to/zlib/headers> to > > > configure then the compilation of bootstrap fails in qbytearray.cpp > > > because it can't find zconf.h. > > > > > if it fails to find the .h now, it will fail to find the .so/.a later > > on as well. > > > I'm building against a static zlib build (on windows actually), and I can > tell that it works just fine if I revert that commit. > nothing follows from that, as it's consistent with what i said.
> > it could have worked before only if you mixed the host paths into the > > target paths, and made sure that the target would not actually use them > > (i.e., -qt-zlib, and having the paths not contain any of the other > > libraries qt could find), which is rather fragile. > > > It only applies to host builds so I don't understand this. > The target and the host systems are the same. > that is actually the relevant data point. fix uploaded. > > > Is there some other way you are supposed to specify the include path, > > > or is this an oversight? > > > > > i think the commit message is pretty clear in this regard. > > > > > make -D/-I/-L/-l/-R not affect bootstrapped tools > > No? > the commit message is longer than that. _______________________________________________ Development mailing list [email protected] http://lists.qt-project.org/mailman/listinfo/development
