On Tue, Jan 08, 2019 at 06:19:32PM +0100, Mattia Rizzolo wrote: > On Thu, Jan 03, 2019 at 06:00:18PM +0100, Guido Günther wrote: > > > It's quite strange that I cannot reproduce this problem anymore > > > on the commit in question. > > > > > > @Mattia: Are you able to reproduce this problem? > > Weird, indeed I can't reproduce this right now :\ > So upsetting, consider how many headaches it gave me over the years. > > > That does not work you want to use gbp export-orig to export the > > tarballs and pass the subtarball names with --component (it's symmetric > > to the import) since you otherwise lack the needed trees. Check the tb > > package. So far I don't see a bug here. > > > > gbp export-orig --component contrib > > Could you please explain what `gbp export-orig` does? > I never heard of that command, and till now I always called > `pristine-tar checkout ../foobar_ver.tar.xz`. And to be honest I expect > that to work regardless of what wrapper was used to do the initial > `pristine-tar commit`. > > What particulare tree are you expecting? > The only thing I can think of is a tree with the main tarball without > the additonal tarball committed, are you having `export-orig` generate > such a thing each time somehow (how would you get the same sha1??)
You need a tree that has the component's tarball dirs dropped: https://github.com/agx/git-buildpackage/blob/master/gbp/scripts/export_orig.py#L77 gbp-buildpackage uses the same code when generating tarballs. Cheers, -- Guido