On Wednesday 07 June 2006 17:13, William A. Hoffman wrote: > At 10:26 AM 6/7/2006, David Faure wrote: > >On Monday 22 May 2006 10:34, David Faure wrote: > >> I have seen this problem multiple times, but I'm not sure something can be > >> done about it... > >> Basically the first time "make" is run, if it triggers a cmake run, then > >> the new dependencies > >> are not taken into account. They are only when running make again. > > > >BTW I'm seeing the same when a header file is moved. > > What was the full command for make given at this point?
Just "make". (which triggered a re-configure+re-generate step before compiling) > >make[2]: *** No rule to make target > >`/d/kde/src/4/kdelibs/kutils/kplugininfo.h', needed by > >`kutils/CMakeFiles/kutils.dir/kpluginselector.o'. Stop. > >make[1]: *** [kutils/CMakeFiles/kutils.dir/all] Error 2 > >make: *** [all] Error 2 > >makeobj[0]: Leaving directory `/d/kde/build/4/kdelibs' > > > >-obelix- dfaure 16:06 /d/kde/build/4/kdelibs>cd kutils > >-obelix- dfaure 16:21 /d/kde/build/4/kdelibs/kutils>make > >makeobj[0]: Entering directory `/d/kde/build/4/kdelibs/kutils' > >Dependee "/d/kde/src/4/kdelibs/kutils/kplugininfo.h" does not exist for > >depender > > This is what should happen, and does happen when I have a small test case. > It seems like the first run of cmake is not running the cmake > --check-build-system > for some reason. Isn't the whole set of configure checks, plus "Configuring done" plus "Generating done" exactly what --check-build-system does? So it seems to me that it did do that.... Any patch (e.g. for additional debug output) that I could test? I'm using CMake CVS, CMake-2-4 branch [just updated -> no change]. -- David Faure, [EMAIL PROTECTED], sponsored by Trolltech to work on KDE, Konqueror (http://www.konqueror.org), and KOffice (http://www.koffice.org). _______________________________________________ Kde-buildsystem mailing list [email protected] https://mail.kde.org/mailman/listinfo/kde-buildsystem
