Update:
It seems "git log" diff compare has some bugs that cherry-picks cannot be
recognized properly sometimes.
Worked with Vinod to directly compare JIRA tickets and git log result. We
finally found 20 JIRAs (list below) which fix version is 2.8.0 and not in
branch-2.8. I've backported all of t
In fact, 3 out of those 4 bugs are part of a bigger umbrella refactor effort.
So I don’t see any problem as of now.
+Vinod
> On Dec 17, 2015, at 11:29 AM, Vinod Kumar Vavilapalli
> wrote:
>
> Wangda,
>
> I see just 13 tickets *only in 2.9.0*
> (https://issues.apache.org/jira/issues/?jql=pro
Wangda,
I see just 13 tickets *only in 2.9.0*
(https://issues.apache.org/jira/issues/?jql=project%20in%20%28YARN%2C%20MAPREDUCE%2C%20HDFS%2C%20HADOOP%29%20AND%20resolution%20%3D%20Fixed%20and%20fixVersion%20%3D%202.9.0%20ORDER%20BY%20priority%20DESC).
If you leave the new features / improvements
Tx for your comments.
We decided to cut the branch mid last month with an intention of locking down
feature set and have realistic progress towards getting out an RC.
We had elaborate discussions on the feature set for 2.8.0 last couple of
months. As much as is possible, I’ll avoid resetting al
Sounds like branch-2.8 was cut off prematurely.
What is the point of forking off if the release is not imminent.
We don't want this thing branching like a banyan again, with each commit
going into 5 branches.
I think it would be easier to retire branch-2.8 for now, and reset it to
branch-2.9 when
Not entirely sure we have enough info to warrant a hard-reset. Yet.
Do we know how many of these are new features / improvement that we shouldn’t
put any more into 2.8.0 and how many are bugs that we need to? If you don’t
have time, I’ll look at it right after getting out 2.7.2 RC1.
Thanks
+Vin
Thanks Wangda for reporting issue. I'm also +1 for recreating a new
branch-2.8 based on current branch-2.8 too. After that, we can revert
2.9.0-only patches on the branch. I will help you if necessary, so
please let me know.
Thanks,
- Tsuyoshi
On Thu, Dec 17, 2015 at 10:51 AM, Jing Zhao wrote:
>
Thanks for reporting the issue, Wangda. Recreating a new branch-2.8 based
on the current branch-2 makes sense to me.
Thanks,
-Jing
On Wed, Dec 16, 2015 at 5:38 PM, Wangda Tan wrote:
> Hi folks,
>
> I found there're lots of commits are in branch-2 only.
> Ran "git log branch-2.8..branch-2".
>
>
Hi folks,
I found there're lots of commits are in branch-2 only.
Ran "git log branch-2.8..branch-2".
There're 35 commits for YARN, 41 commits for HDFS and 12 commits for
COMMON. Only several of them are placed in 2.9.0 section in CHANGES.txt.
I think we can either hard reset branch-2.8 to branch
