On Friday 2008-11-28 20:05, Ralf Wildenhues wrote: >> 14:04 yaguchi:../test/obj > make distcheck V=2 >> [...] >> make[1]: Entering directory `/dev/shm/test/obj/foo-0/_build' >> depbase=`echo foo.o | sed 's|[^/]*$|.deps/&|;s|\.o$||'`;\ >> gcc -DPACKAGE_NAME=\"foo\" -DPACKAGE_TARNAME=\"foo\" >> -DPACKAGE_VERSION=\"0\" -DPACKAGE_STRING=\"foo\ 0\" -DPACKAGE_BUGREPORT=\"\" >> -DPACKAGE=\"foo\" -DVERSION=\"0\" -I. -I.. -g -O2 -MT foo.o -MD -MP -MF >> -$depbase.Tpo -c -o foo.o ../foo.c &&\ >> mv -f $depbase.Tpo $depbase.Po >> ../foo.c:1:17: error: foo.h: No such file or directory >> make[1]: *** [foo.o] Error 1 >> make[1]: Leaving directory `/dev/shm/test/obj/foo-0/_build' >> make: *** [distcheck] Error 2 >> >> That is because it uses "/dev/shm/test/obj/foo-0/_build/../.." as >> $srcdir (which however is the $builddir) instead of >> "/dev/shm/test/obj/foo-0/_build/.." > >This seems to be a misunderstanding. What distcheck does here is the >following: >- first create a distribution tarball, >- then, extract this tarball into a subdirectory foo-0/, >- then, configure the sources in that subdirectory, in a VPATH setup, > i.e., using a relative build directory named _build/. > >The most likely reason for the failure is that foo.h is not packaged.
Ah that's it. Sigh, I so really prefer doing a git-archive HEAD | tar -C /tmp/whereever -xf- && run autogen.sh in whereever && pack up the stuff over distcheck atm :-P Jan
