[ANNOUNCE] Spark branch-1.5

[Reynold Xin <rx...@databricks.com>]

Hi Devs,

Just an announcement that I've cut Spark's branch-1.5 to form the basis of
the 1.5 release. Other than a few stragglers, this represents the end of
active feature development for Spark 1.5. *If committers are merging any
features (outside of alpha modules), please shoot me an email so I can help
coordinate. Any new commits will need to be explicitly merged into
branch-1.5*.

In the next few days, we should come up with testing plans for the release
and create umbrella JIRAs for testing various components and changes. I
plan to cut a preview package for 1.5 towards the end of this week or early
next week.


- R

Re: [ANNOUNCE] Spark branch-1.5

[Michael Armbrust <mi...@databricks.com>]

>
> Would it be reasonable to start un-targeting non-bug non-blocker
>> issues? like, would anyone yell if I started doing that? that would
>> leave ~100 JIRAs, which still seems like more than can actually go
>> into the release. And anyone can re-target as desired.
>>
>
I think the maintainers of the various components should take care of
this.  Reynold and I just did a pass over SQL and I think that by Friday
there should only be blocker bugs / documentation remaining.

Re: [ANNOUNCE] Spark branch-1.5

[Joseph Bradley <jo...@databricks.com>]

I agree that it's high time to start changing/removing target versions,
especially if component maintainers have a good idea of what is not needed
for 1.5.  I'll start doing that on ML.

On Mon, Aug 3, 2015 at 12:05 PM, Sean Owen <so...@cloudera.com> wrote:

> Are these about the right rules of engagement for now until the
> release candidate?
>
> - Don't merge new features or improvements into 1.5 unless they're
> Important and Have Been Discussed
> - Docs and tests are OK to merge into 1.5
> - Bug fixes can be merged into 1.5, with increasing conservativeness
> as the release candidate approaches
>
> FWIW there are now 331 JIRAs targeted at 1.5.0.
>
> Would it be reasonable to start un-targeting non-bug non-blocker
> issues? like, would anyone yell if I started doing that? that would
> leave ~100 JIRAs, which still seems like more than can actually go
> into the release. And anyone can re-target as desired.
>
> I'm interested with using this to communicate about release planning
> so we can actually see how things are moving along and decide if 1.5
> has to be pushed back or not; otherwise it seems pretty unpredictable
> what's coming, going in, and when the process stops and outputs a
> release.
>
>
> On Mon, Aug 3, 2015 at 7:11 PM, Reynold Xin <rx...@databricks.com> wrote:
> > Hi Devs,
> >
> > Just an announcement that I've cut Spark's branch-1.5 to form the basis
> of
> > the 1.5 release. Other than a few stragglers, this represents the end of
> > active feature development for Spark 1.5. If committers are merging any
> > features (outside of alpha modules), please shoot me an email so I can
> help
> > coordinate. Any new commits will need to be explicitly merged into
> > branch-1.5.
> >
> > In the next few days, we should come up with testing plans for the
> release
> > and create umbrella JIRAs for testing various components and changes. I
> plan
> > to cut a preview package for 1.5 towards the end of this week or early
> next
> > week.
> >
> >
> > - R
> >
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@spark.apache.org
> For additional commands, e-mail: dev-help@spark.apache.org
>
>

Re: [ANNOUNCE] Spark branch-1.5

[Sean Owen <so...@cloudera.com>]

Are these about the right rules of engagement for now until the
release candidate?

- Don't merge new features or improvements into 1.5 unless they're
Important and Have Been Discussed
- Docs and tests are OK to merge into 1.5
- Bug fixes can be merged into 1.5, with increasing conservativeness
as the release candidate approaches

FWIW there are now 331 JIRAs targeted at 1.5.0.

Would it be reasonable to start un-targeting non-bug non-blocker
issues? like, would anyone yell if I started doing that? that would
leave ~100 JIRAs, which still seems like more than can actually go
into the release. And anyone can re-target as desired.

I'm interested with using this to communicate about release planning
so we can actually see how things are moving along and decide if 1.5
has to be pushed back or not; otherwise it seems pretty unpredictable
what's coming, going in, and when the process stops and outputs a
release.


On Mon, Aug 3, 2015 at 7:11 PM, Reynold Xin <rx...@databricks.com> wrote:
> Hi Devs,
>
> Just an announcement that I've cut Spark's branch-1.5 to form the basis of
> the 1.5 release. Other than a few stragglers, this represents the end of
> active feature development for Spark 1.5. If committers are merging any
> features (outside of alpha modules), please shoot me an email so I can help
> coordinate. Any new commits will need to be explicitly merged into
> branch-1.5.
>
> In the next few days, we should come up with testing plans for the release
> and create umbrella JIRAs for testing various components and changes. I plan
> to cut a preview package for 1.5 towards the end of this week or early next
> week.
>
>
> - R
>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@spark.apache.org
For additional commands, e-mail: dev-help@spark.apache.org