On Mon, Feb 23, 2015 at 05:35:49AM -0800, Dirk Hohndel wrote: > On Mon, Feb 23, 2015 at 08:53:46AM +0100, Patrick Valsecchi wrote: > > Dirk, > > > > You introduced weird circular dependencies with you commit 5bf86f6042 that > > makes fixing the CMakeLists.txt impossible. > > I'll admit that I don't use cmake to build Subsurface. Tomaz added that to > do unit testing but since I have yet to get it to even work for me without > overwriting files in the source directory (I reported this here and others > told me that it works for them, it sure as heck doesn't work for me). > So to me at this stage cmake is not something I worry about. The way I > build Subsurface for all our supported platforms is qmake. THAT I worry > about.
Having said that... I have 4 more hours of flight time with really rotten internet, so I might as well figure this out and get it fixed... Open heart surgery at 33,000ft :-) > > Now qthelper.cpp (in subsurface_corelib) needs UserSurvey which is provided > > by subsurface_interface which in turn needs subsurface_corelib. > > Yeah, that's Tomaz' separation into libraries that he needed for testing. > > > I don't know how we can fix that... either move reverseGeoLookup somewhere > > else (tough since parse-xml.c is using it) or remove the split in different > > libs in the CMakeLists.txt (should be OK, but I don't know why we had it in > > a first place). > > It's easy to simply move the new geo lookup stuff into a different file, > qthelper.cpp at this point contains a random mix of leftovers :-) I'll clean that up (a little). /D _______________________________________________ subsurface mailing list [email protected] http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface
