I volunteer to be release manager, since no one else has stepped forward.

I've been trying to get
https://issues.apache.org/jira/browse/CALCITE-466 complete before 1.0,
but it just isn't happening fast enough[1]. So, I propose to drop
CALCITE-466 from the goals for 1.0.

Are there any remaining blockers for 1.0?

If there are no blockers, I'll update the change history, create a new
snapshot for Hive and others to try, start running through the release
process[2], and we should have a release-candidate in a day or two.

Julian

[1] More details about the CALCITE-466 task, if you're interested: One
of the pre-requisites was to make the interpreter handle all the SQL
in the test suite (pretty much the whole of SQL), and user-defined
aggregates, windowed aggregation and outer joins are not working yet.
There are still about 100 test failures. It's impossible to go from
there to a stable release candidate in a few days. I'll check in my
changes to the interpreter, but the interpreter will not be on the
main code path; enumerable convention will continue to be used to
evaluate most of our test queries.

[2] 
https://github.com/apache/incubator-calcite/blob/master/doc/HOWTO.md#making-a-release-for-calcite-committers


On Tue, Jan 20, 2015 at 11:05 AM, Brock Noland <[email protected]> wrote:
> Hi,
>
> From the Hive side, we don't expose Calcite to users so we should be
> able to handle any breakage on our end.
>
> Brock
>
> On Sun, Jan 18, 2015 at 7:54 PM, Jacques Nadeau <[email protected]> wrote:
>> As long as everyone is comfortable with potential API breakage in 1.1, I'm
>> +1 for a release as well.
>>
>> On Wed, Jan 14, 2015 at 3:29 PM, Julian Hyde <[email protected]> wrote:
>>
>>> I agree. I would just like to get a release out. We’ve been talking about
>>> 1.0 for several months, so no one should be surprised.
>>>
>>> If anyone is struck with pangs of methodological guilt over the fact that
>>> we haven’t documented the public API, they are welcome to document it.
>>>
>>> Julian
>>>
>>>
>>>
>>> On Jan 14, 2015, at 3:22 PM, Vladimir Sitnikov <
>>> [email protected]> wrote:
>>>
>>> > If we polish really hard and make 1.0 perfect, then we would have nothing
>>> > to do for 2.0 )
>>> >
>>> > I would prefer release sooner, so we can get feedback.
>>> >
>>> > Otherwise I have no preference.
>>> >
>>> > Vladimir
>>>
>>>

Reply via email to