On Apr 2, 2015, at 9:51 AM, Karthik Kambatla <ka...@cloudera.com> wrote: >> >> a) remove CHANGES.TXT from trunk >> > > Removing this from trunk makes it particularly hard to cherry-pick changes > from trunk to branch-2. I would gate this on the removal of CHANGES.txt on > branch-2 as well, at least until we have some non-future releases off > branch-2.
Removing changes.txt, relnotes.py, etc from branch-2 would be an incompatible change. Pushing aside the questions of that document’s quality (hint: lots of outright lying and missing several hundred jiras), it's effectively an interface in used by quite a few folks. But in reality, I suspect the opposite: removing changes.txt just from trunk will make cherry picks easier. If you don’t have to update trunk’s changes.txt, you can cherry-pick with no worries about conflict merges on changes.txt in other branches. Then just update changes.txt in branch-2 manually as you would have done pre- this change anyway. >> b) pre-populate x amount of Hadoop 2.x release data into trunk so that the >> auto-indexer can pick it up >> c) update the HowToRelease information with, well, how to do releases >> based upon these new capabilities >> > > There is a create-release script that likely needs updating. > Yup. I knew about it (I happen to have it running in another window as I type this), but from what I can see, it’s completely undocumented. :( As I update HowToRelease, I’m going to puts some notes in it about this script.