Daniel Veillard wrote: > On Mon, Feb 22, 2010 at 05:01:33PM -0700, Eric Blake wrote: >> Note that this commit mirrors the latest upstream gnulib, but does NOT >> update our gnulib checkout to the latest, so our version of bootstrap >> is now newer than .gnulib/build-aux/bootstrap. >> >> This also fixes a bug in the .pot files, regarding the copyright holder. >> >> * bootstrap: Update to version in .gnulib/build-aux. >> * bootstrap.conf (MSGID_BUGS_ADDRESS, COPYRIGHT_HOLDER, SKIP_PO) >> (gnulib_mk, ACLOCAL, bootstrap_epilogue): Provide overrides. >> * autogen.sh (autoreconf): Avoid redundant autoreconf if bootstrap >> was run. >> * po/Makevars: Delete, now that bootstrap creates it. > > This seems to break for me. po/Makevars.template doesn't define > top_builddir , but that's needed in a few places in po/Makefile.in.in > and as a result I end up with > > make[2]: Entering directory `/u/veillard/libvirt/po' > make[2]: *** No rule to make target `/config.status', needed by > `Makefile'. Stop. > > when trying to build the normal way (./autogen.sh --system ; make) > > could you have a look at this ? Hi DV, So far, I've been able to reproduce that only in a build directory in which .gnulib is locally modified (i.e., when .gnulib is listed in "git diff" output as being locally modified). Once I undo that local modification (say via "git submodule update") "make" then reports that I must run autogen.sh, as expected. And when I do that, it *does* run bootstrap and recreate po/Makevars. -- libvir-list mailing list libvir-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/libvir-list