On Wed, Sep 3, 2014 at 7:41 AM, Dave <snoopd...@gmail.com> wrote: > On Wed, Sep 3, 2014 at 7:29 AM, John D. Ament <john.d.am...@gmail.com> > wrote: > > > On Wed, Sep 3, 2014 at 7:24 AM, Dave <snoopd...@gmail.com> wrote: > > > > > On Wed, Sep 3, 2014 at 1:21 AM, Branko Čibej <br...@apache.org> wrote: > > > > > > > On 03.09.2014 05:03, Jake Farrell wrote: > > > > > Hi John > > > > > I requested that Dave add the RC tag to better keep track of > multiple > > > > > release candidates and make it easier for testing and not mixing > any > > > > > previous version up accidentally. This is very common and currently > > > done > > > > in > > > > > many TLP's including Thrift, Mesos, and Cassandra to name a few. > > > > > > > > Agreed. And it is (or should be) normal process to start a new vote > for > > > > every new set of release bits. Even if 1.0.0 is just a repackaging of > > > > 1.0.0-rc4, there may be bugs in the packaging process itself (such as > > > > we've seen in this thread, when files that were not intended to be > > > > released were bundled in the release artefacts), so the PPMC should > > test > > > > and vote again. > > > > > > > > > But, if 1.0.0 is not a repackage but instead the *exact* same set of > > files > > > as 1.0.0-rc4 but each has been renamed to remove the rc4 designation > then > > > voting again is redundant and unnecessary. Do you agree? > > > > > > > How could you possibly do that without updating the source code? > Certainly > > there's something, even if just a pom file, that mentions the version # > and > > should refer back to an exact git tag/revision (in your case). > > > > > No need for that. All you do is this, for example: > > svn mv apache-usergrid-incubating-1.0.0-rc5-source.tar.gz > apache-usergrid-incubating-1.0.0-source.tar.gz >
So - just to confirm. If I look at apache-usergrid-incubating-1.0.0-rc5-source.tar.gz the contents in the pom.xml files will read the version "1.0.0-rc5" and the git tag will also read "1.0.0-rc5" then after you do an svn mv the pom files will be updated to read "1.0.0" ? If you look at Cassandra, this is how they do it. See [1] and [2]. > > - Dave > [1]: https://github.com/apache/cassandra/releases [2]: https://github.com/apache/cassandra/blob/cassandra-2.1.0-rc6/build.xml#L28
