If you are proposing to keep all important changes in release notes as
usual:
Sure, add more to the migration guide, too. It can't hurt, going forward.
But many release-note changes have not much more to say about migration.
If you're proposing to not mention most things in the release notes,
Yea we can't update the 3.0.0 migration guide now, but AFAIK we do mention
most of the breaking changes there, except for the ML module. I think we
can still put all the ML breaking changes in the release notes this time.
But in the future, shall we put breaking changes in the migration guide? It
Congrats and thanks, Holden!
Bests,
Takeshi
On Thu, Jun 11, 2020 at 11:16 AM Dongjoon Hyun
wrote:
> Thank you so much, Holden! :)
>
> On Wed, Jun 10, 2020 at 6:59 PM Hyukjin Kwon wrote:
>
>> Yay!
>>
>> 2020년 6월 11일 (목) 오전 10:38, Holden Karau 님이 작성:
>>
>>> We are happy to announce the
Thank you so much, Holden! :)
On Wed, Jun 10, 2020 at 6:59 PM Hyukjin Kwon wrote:
> Yay!
>
> 2020년 6월 11일 (목) 오전 10:38, Holden Karau 님이 작성:
>
>> We are happy to announce the availability of Spark 2.4.6!
>>
>> Spark 2.4.6 is a maintenance release containing stability, correctness,
>> and
Yay!
2020년 6월 11일 (목) 오전 10:38, Holden Karau 님이 작성:
> We are happy to announce the availability of Spark 2.4.6!
>
> Spark 2.4.6 is a maintenance release containing stability, correctness,
> and security fixes.
> This release is based on the branch-2.4 maintenance branch of Spark. We
> strongly
I think the proposal doesn't mean to don't add the JIRAs with release-notes
into the release notes (?).
People will still label the JIRAs when the change is significant or
breaking whether it's a bug or not, and they will be in the release notes.
I guess the proposal TL;DR is:
- If that's a
We are happy to announce the availability of Spark 2.4.6!
Spark 2.4.6 is a maintenance release containing stability, correctness, and
security fixes.
This release is based on the branch-2.4 maintenance branch of Spark. We
strongly recommend all 2.4 users to upgrade to this stable release.
To
This seems like a change to current practice, as breaking changes are
marked for release notes with release-notes, etc:
https://spark.apache.org/contributing.html
My only concrete concern, is this seems to imply (?) that many JIRAs with
release-notes and Docs text are not going to get included in
My 2 cents:
Since we have a migration guide, I think people who hit problems when
upgrading Spark will read it. We should mention all the breaking changes
there, except for trivial ones like obvious bug fixes. Even if there is no
meaningful migration to guide for things like removing a deprecated
We have scenario where we wanted to generate groupy on specific set of
dimensions. Grouping sets solves our purpose, but its not supported in spark
streaming. We cannot go for cube as it generates lot of combination which
will be useless for us.
So wanted to know if there is any specific reason
10 matches
Mail list logo
