On 2015-12-15 15:11:24 +1100 (+1100), Tony Breeds wrote: > I'm not really sure what could have gone wrong but heres the context. [...] If you look at the job log for the last two commits in the repo's history, you'll see we have a bit of a publication race when multiple changes are approved at roughly the same time. If the first[1] change takes slightly longer to complete than the second[2] then the resulting publication represents the older of the two. This will "correct" itself the next time someone approves another change to that repo. [1] http://logs.openstack.org/19/198ca4b325d7e9d0ae26d98c9bc4811800077f53/post/project-team-guide-publish/31488b2/console.html.gz [2] http://logs.openstack.org/b1/b15b9a7aa07a213a44d9ba605d693ddbda297be5/post/project-team-guide-publish/608d8d1/console.html.gz -- Jeremy Stanley