On Mon, Dec 7, 2015 at 10:58 AM, Mike Gilbert <[email protected]> wrote: > On Mon, Dec 7, 2015 at 1:26 PM, Matt Turner <[email protected]> wrote: >> On Mon, Dec 7, 2015 at 10:19 AM, Robin H. Johnson <[email protected]> wrote: >>> On Mon, Dec 07, 2015 at 05:16:01AM +0000, Doug Goldstein wrote: >>>> commit: 37c1ec7ef31a51206421f03a2df489ec7d7ca244 >>>> Author: Doug Goldstein <cardoe <AT> gentoo <DOT> org> >>>> AuthorDate: Mon Dec 7 05:14:50 2015 +0000 >>>> Commit: Doug Goldstein <cardoe <AT> gentoo <DOT> org> >>>> CommitDate: Mon Dec 7 05:15:22 2015 +0000 >>>> URL: https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=37c1ec7e >>>> >>>> app-emulation/runc: version bump >>>> >>>> bump to latest upstream release. >>>> >>>> Signed-off-by: Doug Goldstein <cardoe <AT> gentoo.org> >>>> >>>> app-emulation/runc/runc-0.0.5.ebuild | 52 >>>> ++++++++++++++++++++++++++++++++++++ >>>> 1 file changed, 52 insertions(+) >>> You did not update the Manifest or use repoman, and thus blocked the >>> rsync tree. >>> >>> I fixed it now. >> >> This is happening with some frequency. Could we install a server-side >> git hook that prevents pushes if an ebuild is added without a Manifest >> change? It wouldn't have to verify anything more than that to catch >> nearly all of these problems. >> > > That would not work for ebuilds which do not have any distfiles > associated with them.
Good point. It *would* have to be somewhat more involved then...
