On Wed, Feb 11, 2015 at 7:53 AM, Doug Hellmann <[email protected]> wrote:
> > > On Tue, Feb 10, 2015, at 07:12 PM, Joe Gordon wrote: > > Hi, > > > > As you know a few of us have been spending way too much time digging > > stable/juno out of the ditch its currently in. And just when we thought > > we > > were in the clear a new library was released without a requirements cap > > in > > stable global-requirements and broke stable/juno grenade. Everytime this > > happens we risk breaking everything. While there is a good long term fix > > in > > progress (pin all of stable/juno > > https://review.openstack.org/#/c/147451/), > > this will take a bit of time to get right and land. > > > > The good news is there is a nice easy interim solution. Before releasing > > a > > new library go to stable/juno and stable/icehouse global requirements and > > check if $library has a version cap, if not add one. And once that lands > > go > > ahead and release your library. For example: > > https://review.openstack.org/#/c/154715/2 > > > __________________________________________________________________________ > > OpenStack Development Mailing List (not for usage questions) > > Unsubscribe: > > [email protected]?subject:unsubscribe > > http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev > > The Oslo team has several libraries we're holding for release until this > is resolved. We do have projects blocked on those releases, though, so > if Joe asks you for help with anything related to stable branch > maintenance, please make it a priority so we can get the caps in place. > > We have landed the patch to cap all stable/juno requirements that are installed in a tempest-dsvm-neutron-full job. So we should be out of the woods for now (unless you are a project that uses one of the still uncapped requirements). https://review.openstack.org/#/c/147451/ Implications: * Until Dean's patches to install CLI tools (python-*clients) inside of venvs, we are not testing master clients with stable/juno. * An indirect dependency can change and still break us, but hopefully this won't happen. > Doug > > __________________________________________________________________________ > OpenStack Development Mailing List (not for usage questions) > Unsubscribe: [email protected]?subject:unsubscribe > http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev >
__________________________________________________________________________ OpenStack Development Mailing List (not for usage questions) Unsubscribe: [email protected]?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
