On Thu, May 10, 2018 at 09:29:57PM +0200, Matthijs Kooijman wrote: > Hi Guido, > > I noticed one more related thing, also possibly a bug. It seems that the > arguments to --git-pbuilder-options are further processed. I tried > running this command: > > gbp buildpackage --git-pbuilder-options=--extrapackage=/foo/a.deb\ > --extrapackage=/foo/b.deb\ --bindmounts=/foo > > Which resulted in this pbuilder command: > > I: forking: pbuilder build --debbuildopts --debbuildopts > --bindmounts /foo --buildplace /var/cache/pbuilder/build/cow.31796 > --buildresult /tmp > --extrapackages /foo/b.deb --no-targz --internal-chrootexec > 'chroot /var/cache/pbuilder/build/cow.31796 cow-shell' > /tmp/openttd-opengfx_0.5.4-2.dsc > > > Interesting is that the --bindmounts and --extrapackages are no longer > adjacent, and that only the second --extrapackages is present. This > behaviour is even more surprising to me, and I'd think this would > certainly constitute a bug (at least the latter).
There are several levels of quiting going on, gbp itself does not do much but git-pbuilder and pbuilder do. You can set GIT_PBUILDER_DEBUG env var to see how git-pbuilder passes arguments to pbuilder. Only then can you tell who mixed up the arguments. Cheers, -- Guido