You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@kafka.apache.org by Bill Bejeck <bb...@gmail.com> on 2017/09/01 21:53:14 UTC
Re: [VOTE] KIP-182 - Reduce Streams DSL overloads and allow easier
use of custom storage engines
+1
On Thu, Aug 31, 2017 at 10:20 AM, Damian Guy <da...@gmail.com> wrote:
> Thanks everyone for voting! Unfortunately i've had to make a bit of an
> update based on some issues found during implementation.
> The main changes are:
> BytesStoreSupplier -> StoreSupplier
> Addition of:
> WindowBytesStoreSupplier, KeyValueBytesStoreSupplier,
> SessionBytesStoreSupplier that will restrict store types to <Bytes, byte[]>
> 3 new overloads added to Materialized to enable developers to create a
> Materialized of the appropriate type, i..e, WindowStore etc
> Update DSL where Materialized is used such that the stores have generic
> types of <Bytes, byte[]>
> Some minor changes to the arguments to Store#persistentWindowStore and
> Store#persistentSessionStore
>
> Please take a look and recast the votes.
>
> Thanks for your time,
> Damian
>
> On Fri, 25 Aug 2017 at 17:05 Matthias J. Sax <ma...@confluent.io>
> wrote:
>
> > Thanks Damian. Great KIP!
> >
> > +1
> >
> >
> > -Matthias
> >
> > On 8/25/17 6:45 AM, Damian Guy wrote:
> > > Hi,
> > >
> > > I've just realised we need to add two methods to StateStoreBuilder or
> it
> > > isn't going to work:
> > >
> > > Map<String, String> logConfig();
> > > boolean loggingEnabled();
> > >
> > > These are needed when we are building the topology and determining
> > > changelog topic names and configs.
> > >
> > >
> > > I've also update the KIP to add
> > >
> > > StreamBuilder#stream(String topic)
> > >
> > > StreamBuilder#stream(String topic, Consumed options)
> > >
> > >
> > > Thanks
> > >
> > >
> > > On Thu, 24 Aug 2017 at 22:11 Sriram Subramanian <ra...@confluent.io>
> > wrote:
> > >
> > >> +1
> > >>
> > >> On Thu, Aug 24, 2017 at 10:20 AM, Guozhang Wang <wa...@gmail.com>
> > >> wrote:
> > >>
> > >>> +1. Thanks Damian!
> > >>>
> > >>> On Thu, Aug 24, 2017 at 9:47 AM, Bill Bejeck <bb...@gmail.com>
> > wrote:
> > >>>
> > >>>> Thanks for the KIP!
> > >>>>
> > >>>> +1
> > >>>>
> > >>>> Thanks,
> > >>>> Bill
> > >>>>
> > >>>> On Thu, Aug 24, 2017 at 12:25 PM, Damian Guy <da...@gmail.com>
> > >>> wrote:
> > >>>>
> > >>>>> Hi,
> > >>>>>
> > >>>>> I'd like to kick off the voting thread for KIP-182:
> > >>>>> https://cwiki.apache.org/confluence/display/KAFKA/KIP-
> > >>>>> 182%3A+Reduce+Streams+DSL+overloads+and+allow+easier+
> > >>>>> use+of+custom+storage+engines
> > >>>>>
> > >>>>> Thanks,
> > >>>>> Damian
> > >>>>>
> > >>>>
> > >>>
> > >>>
> > >>>
> > >>> --
> > >>> -- Guozhang
> > >>>
> > >>
> > >
> >
> >
>
Re: [VOTE] KIP-182 - Reduce Streams DSL overloads and allow easier
use of custom storage engines
Posted by Damian Guy <da...@gmail.com>.
Hi All,
There has been one further update to the KIP. WindowedKStream has been
renamed to TimeWindowedKStream.
Thanks,
Damian
On Tue, 19 Sep 2017 at 12:21 Damian Guy <da...@gmail.com> wrote:
> All, a small addition to the KIP. During implementation we realized it
> would be good to add the below two methods to StreamBuilder
>
> public synchronized <K, V> GlobalKTable<K, V> globalTable(final String
> topic, final Materialized<K, V, KeyValueStore<Bytes, byte[]>> materialized)
> public synchronized <K, V> KTable<K, V> table(final String topic, final
> Materialized<K, V, KeyValueStore<Bytes, byte[]>> materialized)
>
> Thanks,
> Damian
>
> On Fri, 15 Sep 2017 at 12:37 Damian Guy <da...@gmail.com> wrote:
>
>> Sounds good to me.
>>
>> On Thu, 14 Sep 2017 at 19:55 Guozhang Wang <wa...@gmail.com> wrote:
>>
>>> I'd suggest we remove both to and through together in KIP-182, since for
>>> operator "KTable#to" is as confusing as to "KTable#through" which
>>> overwhelms its benefit as a syntax sugar. I think the extra step
>>> "toStream"
>>> is actually better to remind the caller that it is sending its changelog
>>> stream to topic, plus it is not that much characters.
>>>
>>>
>>> Guozhang
>>>
>>> On Wed, Sep 13, 2017 at 12:40 AM, Damian Guy <da...@gmail.com>
>>> wrote:
>>>
>>> > Hi Guozhang,
>>> >
>>> > I had an offline discussion with Matthias and Bill about it. It is
>>> thought
>>> > that `to` offers some benefit, i.e., syntactic sugar, so perhaps no
>>> harm in
>>> > keeping it. However, `through` less so, seeing as we can materialize
>>> stores
>>> > via `filter`, `map` etc, so one of the main benefits of `through` no
>>> longer
>>> > exists. WDYT?
>>> >
>>> > Thanks,
>>> > Damian
>>> >
>>> > On Tue, 12 Sep 2017 at 18:17 Guozhang Wang <wa...@gmail.com> wrote:
>>> >
>>> > > Hi Damian,
>>> > >
>>> > > Why we are deprecating KTable.through while keeping KTable.to?
>>> Should we
>>> > > either keep both of them or deprecate both of them in favor or
>>> > > KTable.toStream if people agree that it is confusing to users?
>>> > >
>>> > >
>>> > > Guozhang
>>> > >
>>> > >
>>> > > On Tue, Sep 12, 2017 at 1:18 AM, Damian Guy <da...@gmail.com>
>>> > wrote:
>>> > >
>>> > > > Hi All,
>>> > > >
>>> > > > A minor update to the KIP, i needed to add KTable.to(Produced) for
>>> > > > consistency. KTable.through will be deprecated in favour of using
>>> > > > KTable.toStream().through()
>>> > > >
>>> > > > Thanks,
>>> > > > Damian
>>> > > >
>>> > > > On Thu, 7 Sep 2017 at 08:52 Damian Guy <da...@gmail.com>
>>> wrote:
>>> > > >
>>> > > > > Thanks all. The vote is now closed and the KIP has been accepted
>>> > with:
>>> > > > > 2 non binding votes - bill and matthias
>>> > > > > 3 binding - Damian, Guozhang, Sriram
>>> > > > >
>>> > > > > Regards,
>>> > > > > Damian
>>> > > > >
>>> > > > > On Tue, 5 Sep 2017 at 22:24 Sriram Subramanian <ram@confluent.io
>>> >
>>> > > wrote:
>>> > > > >
>>> > > > >> +1
>>> > > > >>
>>> > > > >> On Tue, Sep 5, 2017 at 1:33 PM, Guozhang Wang <
>>> wangguoz@gmail.com>
>>> > > > wrote:
>>> > > > >>
>>> > > > >> > +1
>>> > > > >> >
>>> > > > >> > On Fri, Sep 1, 2017 at 3:45 PM, Matthias J. Sax <
>>> > > > matthias@confluent.io>
>>> > > > >> > wrote:
>>> > > > >> >
>>> > > > >> > > +1
>>> > > > >> > >
>>> > > > >> > > On 9/1/17 2:53 PM, Bill Bejeck wrote:
>>> > > > >> > > > +1
>>> > > > >> > > >
>>> > > > >> > > > On Thu, Aug 31, 2017 at 10:20 AM, Damian Guy <
>>> > > > damian.guy@gmail.com>
>>> > > > >> > > wrote:
>>> > > > >> > > >
>>> > > > >> > > >> Thanks everyone for voting! Unfortunately i've had to
>>> make a
>>> > > bit
>>> > > > >> of an
>>> > > > >> > > >> update based on some issues found during implementation.
>>> > > > >> > > >> The main changes are:
>>> > > > >> > > >> BytesStoreSupplier -> StoreSupplier
>>> > > > >> > > >> Addition of:
>>> > > > >> > > >> WindowBytesStoreSupplier, KeyValueBytesStoreSupplier,
>>> > > > >> > > >> SessionBytesStoreSupplier that will restrict store types
>>> to
>>> > > > <Bytes,
>>> > > > >> > > byte[]>
>>> > > > >> > > >> 3 new overloads added to Materialized to enable
>>> developers to
>>> > > > >> create a
>>> > > > >> > > >> Materialized of the appropriate type, i..e, WindowStore
>>> etc
>>> > > > >> > > >> Update DSL where Materialized is used such that the
>>> stores
>>> > have
>>> > > > >> > generic
>>> > > > >> > > >> types of <Bytes, byte[]>
>>> > > > >> > > >> Some minor changes to the arguments to
>>> > > > Store#persistentWindowStore
>>> > > > >> and
>>> > > > >> > > >> Store#persistentSessionStore
>>> > > > >> > > >>
>>> > > > >> > > >> Please take a look and recast the votes.
>>> > > > >> > > >>
>>> > > > >> > > >> Thanks for your time,
>>> > > > >> > > >> Damian
>>> > > > >> > > >>
>>> > > > >> > > >> On Fri, 25 Aug 2017 at 17:05 Matthias J. Sax <
>>> > > > >> matthias@confluent.io>
>>> > > > >> > > >> wrote:
>>> > > > >> > > >>
>>> > > > >> > > >>> Thanks Damian. Great KIP!
>>> > > > >> > > >>>
>>> > > > >> > > >>> +1
>>> > > > >> > > >>>
>>> > > > >> > > >>>
>>> > > > >> > > >>> -Matthias
>>> > > > >> > > >>>
>>> > > > >> > > >>> On 8/25/17 6:45 AM, Damian Guy wrote:
>>> > > > >> > > >>>> Hi,
>>> > > > >> > > >>>>
>>> > > > >> > > >>>> I've just realised we need to add two methods to
>>> > > > >> StateStoreBuilder
>>> > > > >> > or
>>> > > > >> > > >> it
>>> > > > >> > > >>>> isn't going to work:
>>> > > > >> > > >>>>
>>> > > > >> > > >>>> Map<String, String> logConfig();
>>> > > > >> > > >>>> boolean loggingEnabled();
>>> > > > >> > > >>>>
>>> > > > >> > > >>>> These are needed when we are building the topology and
>>> > > > >> determining
>>> > > > >> > > >>>> changelog topic names and configs.
>>> > > > >> > > >>>>
>>> > > > >> > > >>>>
>>> > > > >> > > >>>> I've also update the KIP to add
>>> > > > >> > > >>>>
>>> > > > >> > > >>>> StreamBuilder#stream(String topic)
>>> > > > >> > > >>>>
>>> > > > >> > > >>>> StreamBuilder#stream(String topic, Consumed options)
>>> > > > >> > > >>>>
>>> > > > >> > > >>>>
>>> > > > >> > > >>>> Thanks
>>> > > > >> > > >>>>
>>> > > > >> > > >>>>
>>> > > > >> > > >>>> On Thu, 24 Aug 2017 at 22:11 Sriram Subramanian <
>>> > > > >> ram@confluent.io>
>>> > > > >> > > >>> wrote:
>>> > > > >> > > >>>>
>>> > > > >> > > >>>>> +1
>>> > > > >> > > >>>>>
>>> > > > >> > > >>>>> On Thu, Aug 24, 2017 at 10:20 AM, Guozhang Wang <
>>> > > > >> > wangguoz@gmail.com>
>>> > > > >> > > >>>>> wrote:
>>> > > > >> > > >>>>>
>>> > > > >> > > >>>>>> +1. Thanks Damian!
>>> > > > >> > > >>>>>>
>>> > > > >> > > >>>>>> On Thu, Aug 24, 2017 at 9:47 AM, Bill Bejeck <
>>> > > > >> bbejeck@gmail.com>
>>> > > > >> > > >>> wrote:
>>> > > > >> > > >>>>>>
>>> > > > >> > > >>>>>>> Thanks for the KIP!
>>> > > > >> > > >>>>>>>
>>> > > > >> > > >>>>>>> +1
>>> > > > >> > > >>>>>>>
>>> > > > >> > > >>>>>>> Thanks,
>>> > > > >> > > >>>>>>> Bill
>>> > > > >> > > >>>>>>>
>>> > > > >> > > >>>>>>> On Thu, Aug 24, 2017 at 12:25 PM, Damian Guy <
>>> > > > >> > damian.guy@gmail.com
>>> > > > >> > > >
>>> > > > >> > > >>>>>> wrote:
>>> > > > >> > > >>>>>>>
>>> > > > >> > > >>>>>>>> Hi,
>>> > > > >> > > >>>>>>>>
>>> > > > >> > > >>>>>>>> I'd like to kick off the voting thread for KIP-182:
>>> > > > >> > > >>>>>>>>
>>> https://cwiki.apache.org/confluence/display/KAFKA/KIP-
>>> > > > >> > > >>>>>>>>
>>> 182%3A+Reduce+Streams+DSL+overloads+and+allow+easier+
>>> > > > >> > > >>>>>>>> use+of+custom+storage+engines
>>> > > > >> > > >>>>>>>>
>>> > > > >> > > >>>>>>>> Thanks,
>>> > > > >> > > >>>>>>>> Damian
>>> > > > >> > > >>>>>>>>
>>> > > > >> > > >>>>>>>
>>> > > > >> > > >>>>>>
>>> > > > >> > > >>>>>>
>>> > > > >> > > >>>>>>
>>> > > > >> > > >>>>>> --
>>> > > > >> > > >>>>>> -- Guozhang
>>> > > > >> > > >>>>>>
>>> > > > >> > > >>>>>
>>> > > > >> > > >>>>
>>> > > > >> > > >>>
>>> > > > >> > > >>>
>>> > > > >> > > >>
>>> > > > >> > > >
>>> > > > >> > >
>>> > > > >> > >
>>> > > > >> >
>>> > > > >> >
>>> > > > >> > --
>>> > > > >> > -- Guozhang
>>> > > > >> >
>>> > > > >>
>>> > > > >
>>> > > >
>>> > >
>>> > >
>>> > >
>>> > > --
>>> > > -- Guozhang
>>> > >
>>> >
>>>
>>>
>>>
>>> --
>>> -- Guozhang
>>>
>>
Re: [VOTE] KIP-182 - Reduce Streams DSL overloads and allow easier
use of custom storage engines
Posted by Damian Guy <da...@gmail.com>.
All, a small addition to the KIP. During implementation we realized it
would be good to add the below two methods to StreamBuilder
public synchronized <K, V> GlobalKTable<K, V> globalTable(final String
topic, final Materialized<K, V, KeyValueStore<Bytes, byte[]>> materialized)
public synchronized <K, V> KTable<K, V> table(final String topic, final
Materialized<K, V, KeyValueStore<Bytes, byte[]>> materialized)
Thanks,
Damian
On Fri, 15 Sep 2017 at 12:37 Damian Guy <da...@gmail.com> wrote:
> Sounds good to me.
>
> On Thu, 14 Sep 2017 at 19:55 Guozhang Wang <wa...@gmail.com> wrote:
>
>> I'd suggest we remove both to and through together in KIP-182, since for
>> operator "KTable#to" is as confusing as to "KTable#through" which
>> overwhelms its benefit as a syntax sugar. I think the extra step
>> "toStream"
>> is actually better to remind the caller that it is sending its changelog
>> stream to topic, plus it is not that much characters.
>>
>>
>> Guozhang
>>
>> On Wed, Sep 13, 2017 at 12:40 AM, Damian Guy <da...@gmail.com>
>> wrote:
>>
>> > Hi Guozhang,
>> >
>> > I had an offline discussion with Matthias and Bill about it. It is
>> thought
>> > that `to` offers some benefit, i.e., syntactic sugar, so perhaps no
>> harm in
>> > keeping it. However, `through` less so, seeing as we can materialize
>> stores
>> > via `filter`, `map` etc, so one of the main benefits of `through` no
>> longer
>> > exists. WDYT?
>> >
>> > Thanks,
>> > Damian
>> >
>> > On Tue, 12 Sep 2017 at 18:17 Guozhang Wang <wa...@gmail.com> wrote:
>> >
>> > > Hi Damian,
>> > >
>> > > Why we are deprecating KTable.through while keeping KTable.to? Should
>> we
>> > > either keep both of them or deprecate both of them in favor or
>> > > KTable.toStream if people agree that it is confusing to users?
>> > >
>> > >
>> > > Guozhang
>> > >
>> > >
>> > > On Tue, Sep 12, 2017 at 1:18 AM, Damian Guy <da...@gmail.com>
>> > wrote:
>> > >
>> > > > Hi All,
>> > > >
>> > > > A minor update to the KIP, i needed to add KTable.to(Produced) for
>> > > > consistency. KTable.through will be deprecated in favour of using
>> > > > KTable.toStream().through()
>> > > >
>> > > > Thanks,
>> > > > Damian
>> > > >
>> > > > On Thu, 7 Sep 2017 at 08:52 Damian Guy <da...@gmail.com>
>> wrote:
>> > > >
>> > > > > Thanks all. The vote is now closed and the KIP has been accepted
>> > with:
>> > > > > 2 non binding votes - bill and matthias
>> > > > > 3 binding - Damian, Guozhang, Sriram
>> > > > >
>> > > > > Regards,
>> > > > > Damian
>> > > > >
>> > > > > On Tue, 5 Sep 2017 at 22:24 Sriram Subramanian <ra...@confluent.io>
>> > > wrote:
>> > > > >
>> > > > >> +1
>> > > > >>
>> > > > >> On Tue, Sep 5, 2017 at 1:33 PM, Guozhang Wang <
>> wangguoz@gmail.com>
>> > > > wrote:
>> > > > >>
>> > > > >> > +1
>> > > > >> >
>> > > > >> > On Fri, Sep 1, 2017 at 3:45 PM, Matthias J. Sax <
>> > > > matthias@confluent.io>
>> > > > >> > wrote:
>> > > > >> >
>> > > > >> > > +1
>> > > > >> > >
>> > > > >> > > On 9/1/17 2:53 PM, Bill Bejeck wrote:
>> > > > >> > > > +1
>> > > > >> > > >
>> > > > >> > > > On Thu, Aug 31, 2017 at 10:20 AM, Damian Guy <
>> > > > damian.guy@gmail.com>
>> > > > >> > > wrote:
>> > > > >> > > >
>> > > > >> > > >> Thanks everyone for voting! Unfortunately i've had to
>> make a
>> > > bit
>> > > > >> of an
>> > > > >> > > >> update based on some issues found during implementation.
>> > > > >> > > >> The main changes are:
>> > > > >> > > >> BytesStoreSupplier -> StoreSupplier
>> > > > >> > > >> Addition of:
>> > > > >> > > >> WindowBytesStoreSupplier, KeyValueBytesStoreSupplier,
>> > > > >> > > >> SessionBytesStoreSupplier that will restrict store types
>> to
>> > > > <Bytes,
>> > > > >> > > byte[]>
>> > > > >> > > >> 3 new overloads added to Materialized to enable
>> developers to
>> > > > >> create a
>> > > > >> > > >> Materialized of the appropriate type, i..e, WindowStore
>> etc
>> > > > >> > > >> Update DSL where Materialized is used such that the stores
>> > have
>> > > > >> > generic
>> > > > >> > > >> types of <Bytes, byte[]>
>> > > > >> > > >> Some minor changes to the arguments to
>> > > > Store#persistentWindowStore
>> > > > >> and
>> > > > >> > > >> Store#persistentSessionStore
>> > > > >> > > >>
>> > > > >> > > >> Please take a look and recast the votes.
>> > > > >> > > >>
>> > > > >> > > >> Thanks for your time,
>> > > > >> > > >> Damian
>> > > > >> > > >>
>> > > > >> > > >> On Fri, 25 Aug 2017 at 17:05 Matthias J. Sax <
>> > > > >> matthias@confluent.io>
>> > > > >> > > >> wrote:
>> > > > >> > > >>
>> > > > >> > > >>> Thanks Damian. Great KIP!
>> > > > >> > > >>>
>> > > > >> > > >>> +1
>> > > > >> > > >>>
>> > > > >> > > >>>
>> > > > >> > > >>> -Matthias
>> > > > >> > > >>>
>> > > > >> > > >>> On 8/25/17 6:45 AM, Damian Guy wrote:
>> > > > >> > > >>>> Hi,
>> > > > >> > > >>>>
>> > > > >> > > >>>> I've just realised we need to add two methods to
>> > > > >> StateStoreBuilder
>> > > > >> > or
>> > > > >> > > >> it
>> > > > >> > > >>>> isn't going to work:
>> > > > >> > > >>>>
>> > > > >> > > >>>> Map<String, String> logConfig();
>> > > > >> > > >>>> boolean loggingEnabled();
>> > > > >> > > >>>>
>> > > > >> > > >>>> These are needed when we are building the topology and
>> > > > >> determining
>> > > > >> > > >>>> changelog topic names and configs.
>> > > > >> > > >>>>
>> > > > >> > > >>>>
>> > > > >> > > >>>> I've also update the KIP to add
>> > > > >> > > >>>>
>> > > > >> > > >>>> StreamBuilder#stream(String topic)
>> > > > >> > > >>>>
>> > > > >> > > >>>> StreamBuilder#stream(String topic, Consumed options)
>> > > > >> > > >>>>
>> > > > >> > > >>>>
>> > > > >> > > >>>> Thanks
>> > > > >> > > >>>>
>> > > > >> > > >>>>
>> > > > >> > > >>>> On Thu, 24 Aug 2017 at 22:11 Sriram Subramanian <
>> > > > >> ram@confluent.io>
>> > > > >> > > >>> wrote:
>> > > > >> > > >>>>
>> > > > >> > > >>>>> +1
>> > > > >> > > >>>>>
>> > > > >> > > >>>>> On Thu, Aug 24, 2017 at 10:20 AM, Guozhang Wang <
>> > > > >> > wangguoz@gmail.com>
>> > > > >> > > >>>>> wrote:
>> > > > >> > > >>>>>
>> > > > >> > > >>>>>> +1. Thanks Damian!
>> > > > >> > > >>>>>>
>> > > > >> > > >>>>>> On Thu, Aug 24, 2017 at 9:47 AM, Bill Bejeck <
>> > > > >> bbejeck@gmail.com>
>> > > > >> > > >>> wrote:
>> > > > >> > > >>>>>>
>> > > > >> > > >>>>>>> Thanks for the KIP!
>> > > > >> > > >>>>>>>
>> > > > >> > > >>>>>>> +1
>> > > > >> > > >>>>>>>
>> > > > >> > > >>>>>>> Thanks,
>> > > > >> > > >>>>>>> Bill
>> > > > >> > > >>>>>>>
>> > > > >> > > >>>>>>> On Thu, Aug 24, 2017 at 12:25 PM, Damian Guy <
>> > > > >> > damian.guy@gmail.com
>> > > > >> > > >
>> > > > >> > > >>>>>> wrote:
>> > > > >> > > >>>>>>>
>> > > > >> > > >>>>>>>> Hi,
>> > > > >> > > >>>>>>>>
>> > > > >> > > >>>>>>>> I'd like to kick off the voting thread for KIP-182:
>> > > > >> > > >>>>>>>>
>> https://cwiki.apache.org/confluence/display/KAFKA/KIP-
>> > > > >> > > >>>>>>>>
>> 182%3A+Reduce+Streams+DSL+overloads+and+allow+easier+
>> > > > >> > > >>>>>>>> use+of+custom+storage+engines
>> > > > >> > > >>>>>>>>
>> > > > >> > > >>>>>>>> Thanks,
>> > > > >> > > >>>>>>>> Damian
>> > > > >> > > >>>>>>>>
>> > > > >> > > >>>>>>>
>> > > > >> > > >>>>>>
>> > > > >> > > >>>>>>
>> > > > >> > > >>>>>>
>> > > > >> > > >>>>>> --
>> > > > >> > > >>>>>> -- Guozhang
>> > > > >> > > >>>>>>
>> > > > >> > > >>>>>
>> > > > >> > > >>>>
>> > > > >> > > >>>
>> > > > >> > > >>>
>> > > > >> > > >>
>> > > > >> > > >
>> > > > >> > >
>> > > > >> > >
>> > > > >> >
>> > > > >> >
>> > > > >> > --
>> > > > >> > -- Guozhang
>> > > > >> >
>> > > > >>
>> > > > >
>> > > >
>> > >
>> > >
>> > >
>> > > --
>> > > -- Guozhang
>> > >
>> >
>>
>>
>>
>> --
>> -- Guozhang
>>
>
Re: [VOTE] KIP-182 - Reduce Streams DSL overloads and allow easier
use of custom storage engines
Posted by Damian Guy <da...@gmail.com>.
Sounds good to me.
On Thu, 14 Sep 2017 at 19:55 Guozhang Wang <wa...@gmail.com> wrote:
> I'd suggest we remove both to and through together in KIP-182, since for
> operator "KTable#to" is as confusing as to "KTable#through" which
> overwhelms its benefit as a syntax sugar. I think the extra step "toStream"
> is actually better to remind the caller that it is sending its changelog
> stream to topic, plus it is not that much characters.
>
>
> Guozhang
>
> On Wed, Sep 13, 2017 at 12:40 AM, Damian Guy <da...@gmail.com> wrote:
>
> > Hi Guozhang,
> >
> > I had an offline discussion with Matthias and Bill about it. It is
> thought
> > that `to` offers some benefit, i.e., syntactic sugar, so perhaps no harm
> in
> > keeping it. However, `through` less so, seeing as we can materialize
> stores
> > via `filter`, `map` etc, so one of the main benefits of `through` no
> longer
> > exists. WDYT?
> >
> > Thanks,
> > Damian
> >
> > On Tue, 12 Sep 2017 at 18:17 Guozhang Wang <wa...@gmail.com> wrote:
> >
> > > Hi Damian,
> > >
> > > Why we are deprecating KTable.through while keeping KTable.to? Should
> we
> > > either keep both of them or deprecate both of them in favor or
> > > KTable.toStream if people agree that it is confusing to users?
> > >
> > >
> > > Guozhang
> > >
> > >
> > > On Tue, Sep 12, 2017 at 1:18 AM, Damian Guy <da...@gmail.com>
> > wrote:
> > >
> > > > Hi All,
> > > >
> > > > A minor update to the KIP, i needed to add KTable.to(Produced) for
> > > > consistency. KTable.through will be deprecated in favour of using
> > > > KTable.toStream().through()
> > > >
> > > > Thanks,
> > > > Damian
> > > >
> > > > On Thu, 7 Sep 2017 at 08:52 Damian Guy <da...@gmail.com> wrote:
> > > >
> > > > > Thanks all. The vote is now closed and the KIP has been accepted
> > with:
> > > > > 2 non binding votes - bill and matthias
> > > > > 3 binding - Damian, Guozhang, Sriram
> > > > >
> > > > > Regards,
> > > > > Damian
> > > > >
> > > > > On Tue, 5 Sep 2017 at 22:24 Sriram Subramanian <ra...@confluent.io>
> > > wrote:
> > > > >
> > > > >> +1
> > > > >>
> > > > >> On Tue, Sep 5, 2017 at 1:33 PM, Guozhang Wang <wangguoz@gmail.com
> >
> > > > wrote:
> > > > >>
> > > > >> > +1
> > > > >> >
> > > > >> > On Fri, Sep 1, 2017 at 3:45 PM, Matthias J. Sax <
> > > > matthias@confluent.io>
> > > > >> > wrote:
> > > > >> >
> > > > >> > > +1
> > > > >> > >
> > > > >> > > On 9/1/17 2:53 PM, Bill Bejeck wrote:
> > > > >> > > > +1
> > > > >> > > >
> > > > >> > > > On Thu, Aug 31, 2017 at 10:20 AM, Damian Guy <
> > > > damian.guy@gmail.com>
> > > > >> > > wrote:
> > > > >> > > >
> > > > >> > > >> Thanks everyone for voting! Unfortunately i've had to make
> a
> > > bit
> > > > >> of an
> > > > >> > > >> update based on some issues found during implementation.
> > > > >> > > >> The main changes are:
> > > > >> > > >> BytesStoreSupplier -> StoreSupplier
> > > > >> > > >> Addition of:
> > > > >> > > >> WindowBytesStoreSupplier, KeyValueBytesStoreSupplier,
> > > > >> > > >> SessionBytesStoreSupplier that will restrict store types to
> > > > <Bytes,
> > > > >> > > byte[]>
> > > > >> > > >> 3 new overloads added to Materialized to enable developers
> to
> > > > >> create a
> > > > >> > > >> Materialized of the appropriate type, i..e, WindowStore etc
> > > > >> > > >> Update DSL where Materialized is used such that the stores
> > have
> > > > >> > generic
> > > > >> > > >> types of <Bytes, byte[]>
> > > > >> > > >> Some minor changes to the arguments to
> > > > Store#persistentWindowStore
> > > > >> and
> > > > >> > > >> Store#persistentSessionStore
> > > > >> > > >>
> > > > >> > > >> Please take a look and recast the votes.
> > > > >> > > >>
> > > > >> > > >> Thanks for your time,
> > > > >> > > >> Damian
> > > > >> > > >>
> > > > >> > > >> On Fri, 25 Aug 2017 at 17:05 Matthias J. Sax <
> > > > >> matthias@confluent.io>
> > > > >> > > >> wrote:
> > > > >> > > >>
> > > > >> > > >>> Thanks Damian. Great KIP!
> > > > >> > > >>>
> > > > >> > > >>> +1
> > > > >> > > >>>
> > > > >> > > >>>
> > > > >> > > >>> -Matthias
> > > > >> > > >>>
> > > > >> > > >>> On 8/25/17 6:45 AM, Damian Guy wrote:
> > > > >> > > >>>> Hi,
> > > > >> > > >>>>
> > > > >> > > >>>> I've just realised we need to add two methods to
> > > > >> StateStoreBuilder
> > > > >> > or
> > > > >> > > >> it
> > > > >> > > >>>> isn't going to work:
> > > > >> > > >>>>
> > > > >> > > >>>> Map<String, String> logConfig();
> > > > >> > > >>>> boolean loggingEnabled();
> > > > >> > > >>>>
> > > > >> > > >>>> These are needed when we are building the topology and
> > > > >> determining
> > > > >> > > >>>> changelog topic names and configs.
> > > > >> > > >>>>
> > > > >> > > >>>>
> > > > >> > > >>>> I've also update the KIP to add
> > > > >> > > >>>>
> > > > >> > > >>>> StreamBuilder#stream(String topic)
> > > > >> > > >>>>
> > > > >> > > >>>> StreamBuilder#stream(String topic, Consumed options)
> > > > >> > > >>>>
> > > > >> > > >>>>
> > > > >> > > >>>> Thanks
> > > > >> > > >>>>
> > > > >> > > >>>>
> > > > >> > > >>>> On Thu, 24 Aug 2017 at 22:11 Sriram Subramanian <
> > > > >> ram@confluent.io>
> > > > >> > > >>> wrote:
> > > > >> > > >>>>
> > > > >> > > >>>>> +1
> > > > >> > > >>>>>
> > > > >> > > >>>>> On Thu, Aug 24, 2017 at 10:20 AM, Guozhang Wang <
> > > > >> > wangguoz@gmail.com>
> > > > >> > > >>>>> wrote:
> > > > >> > > >>>>>
> > > > >> > > >>>>>> +1. Thanks Damian!
> > > > >> > > >>>>>>
> > > > >> > > >>>>>> On Thu, Aug 24, 2017 at 9:47 AM, Bill Bejeck <
> > > > >> bbejeck@gmail.com>
> > > > >> > > >>> wrote:
> > > > >> > > >>>>>>
> > > > >> > > >>>>>>> Thanks for the KIP!
> > > > >> > > >>>>>>>
> > > > >> > > >>>>>>> +1
> > > > >> > > >>>>>>>
> > > > >> > > >>>>>>> Thanks,
> > > > >> > > >>>>>>> Bill
> > > > >> > > >>>>>>>
> > > > >> > > >>>>>>> On Thu, Aug 24, 2017 at 12:25 PM, Damian Guy <
> > > > >> > damian.guy@gmail.com
> > > > >> > > >
> > > > >> > > >>>>>> wrote:
> > > > >> > > >>>>>>>
> > > > >> > > >>>>>>>> Hi,
> > > > >> > > >>>>>>>>
> > > > >> > > >>>>>>>> I'd like to kick off the voting thread for KIP-182:
> > > > >> > > >>>>>>>>
> https://cwiki.apache.org/confluence/display/KAFKA/KIP-
> > > > >> > > >>>>>>>> 182%3A+Reduce+Streams+DSL+overloads+and+allow+easier+
> > > > >> > > >>>>>>>> use+of+custom+storage+engines
> > > > >> > > >>>>>>>>
> > > > >> > > >>>>>>>> Thanks,
> > > > >> > > >>>>>>>> Damian
> > > > >> > > >>>>>>>>
> > > > >> > > >>>>>>>
> > > > >> > > >>>>>>
> > > > >> > > >>>>>>
> > > > >> > > >>>>>>
> > > > >> > > >>>>>> --
> > > > >> > > >>>>>> -- Guozhang
> > > > >> > > >>>>>>
> > > > >> > > >>>>>
> > > > >> > > >>>>
> > > > >> > > >>>
> > > > >> > > >>>
> > > > >> > > >>
> > > > >> > > >
> > > > >> > >
> > > > >> > >
> > > > >> >
> > > > >> >
> > > > >> > --
> > > > >> > -- Guozhang
> > > > >> >
> > > > >>
> > > > >
> > > >
> > >
> > >
> > >
> > > --
> > > -- Guozhang
> > >
> >
>
>
>
> --
> -- Guozhang
>
Re: [VOTE] KIP-182 - Reduce Streams DSL overloads and allow easier
use of custom storage engines
Posted by Guozhang Wang <wa...@gmail.com>.
I'd suggest we remove both to and through together in KIP-182, since for
operator "KTable#to" is as confusing as to "KTable#through" which
overwhelms its benefit as a syntax sugar. I think the extra step "toStream"
is actually better to remind the caller that it is sending its changelog
stream to topic, plus it is not that much characters.
Guozhang
On Wed, Sep 13, 2017 at 12:40 AM, Damian Guy <da...@gmail.com> wrote:
> Hi Guozhang,
>
> I had an offline discussion with Matthias and Bill about it. It is thought
> that `to` offers some benefit, i.e., syntactic sugar, so perhaps no harm in
> keeping it. However, `through` less so, seeing as we can materialize stores
> via `filter`, `map` etc, so one of the main benefits of `through` no longer
> exists. WDYT?
>
> Thanks,
> Damian
>
> On Tue, 12 Sep 2017 at 18:17 Guozhang Wang <wa...@gmail.com> wrote:
>
> > Hi Damian,
> >
> > Why we are deprecating KTable.through while keeping KTable.to? Should we
> > either keep both of them or deprecate both of them in favor or
> > KTable.toStream if people agree that it is confusing to users?
> >
> >
> > Guozhang
> >
> >
> > On Tue, Sep 12, 2017 at 1:18 AM, Damian Guy <da...@gmail.com>
> wrote:
> >
> > > Hi All,
> > >
> > > A minor update to the KIP, i needed to add KTable.to(Produced) for
> > > consistency. KTable.through will be deprecated in favour of using
> > > KTable.toStream().through()
> > >
> > > Thanks,
> > > Damian
> > >
> > > On Thu, 7 Sep 2017 at 08:52 Damian Guy <da...@gmail.com> wrote:
> > >
> > > > Thanks all. The vote is now closed and the KIP has been accepted
> with:
> > > > 2 non binding votes - bill and matthias
> > > > 3 binding - Damian, Guozhang, Sriram
> > > >
> > > > Regards,
> > > > Damian
> > > >
> > > > On Tue, 5 Sep 2017 at 22:24 Sriram Subramanian <ra...@confluent.io>
> > wrote:
> > > >
> > > >> +1
> > > >>
> > > >> On Tue, Sep 5, 2017 at 1:33 PM, Guozhang Wang <wa...@gmail.com>
> > > wrote:
> > > >>
> > > >> > +1
> > > >> >
> > > >> > On Fri, Sep 1, 2017 at 3:45 PM, Matthias J. Sax <
> > > matthias@confluent.io>
> > > >> > wrote:
> > > >> >
> > > >> > > +1
> > > >> > >
> > > >> > > On 9/1/17 2:53 PM, Bill Bejeck wrote:
> > > >> > > > +1
> > > >> > > >
> > > >> > > > On Thu, Aug 31, 2017 at 10:20 AM, Damian Guy <
> > > damian.guy@gmail.com>
> > > >> > > wrote:
> > > >> > > >
> > > >> > > >> Thanks everyone for voting! Unfortunately i've had to make a
> > bit
> > > >> of an
> > > >> > > >> update based on some issues found during implementation.
> > > >> > > >> The main changes are:
> > > >> > > >> BytesStoreSupplier -> StoreSupplier
> > > >> > > >> Addition of:
> > > >> > > >> WindowBytesStoreSupplier, KeyValueBytesStoreSupplier,
> > > >> > > >> SessionBytesStoreSupplier that will restrict store types to
> > > <Bytes,
> > > >> > > byte[]>
> > > >> > > >> 3 new overloads added to Materialized to enable developers to
> > > >> create a
> > > >> > > >> Materialized of the appropriate type, i..e, WindowStore etc
> > > >> > > >> Update DSL where Materialized is used such that the stores
> have
> > > >> > generic
> > > >> > > >> types of <Bytes, byte[]>
> > > >> > > >> Some minor changes to the arguments to
> > > Store#persistentWindowStore
> > > >> and
> > > >> > > >> Store#persistentSessionStore
> > > >> > > >>
> > > >> > > >> Please take a look and recast the votes.
> > > >> > > >>
> > > >> > > >> Thanks for your time,
> > > >> > > >> Damian
> > > >> > > >>
> > > >> > > >> On Fri, 25 Aug 2017 at 17:05 Matthias J. Sax <
> > > >> matthias@confluent.io>
> > > >> > > >> wrote:
> > > >> > > >>
> > > >> > > >>> Thanks Damian. Great KIP!
> > > >> > > >>>
> > > >> > > >>> +1
> > > >> > > >>>
> > > >> > > >>>
> > > >> > > >>> -Matthias
> > > >> > > >>>
> > > >> > > >>> On 8/25/17 6:45 AM, Damian Guy wrote:
> > > >> > > >>>> Hi,
> > > >> > > >>>>
> > > >> > > >>>> I've just realised we need to add two methods to
> > > >> StateStoreBuilder
> > > >> > or
> > > >> > > >> it
> > > >> > > >>>> isn't going to work:
> > > >> > > >>>>
> > > >> > > >>>> Map<String, String> logConfig();
> > > >> > > >>>> boolean loggingEnabled();
> > > >> > > >>>>
> > > >> > > >>>> These are needed when we are building the topology and
> > > >> determining
> > > >> > > >>>> changelog topic names and configs.
> > > >> > > >>>>
> > > >> > > >>>>
> > > >> > > >>>> I've also update the KIP to add
> > > >> > > >>>>
> > > >> > > >>>> StreamBuilder#stream(String topic)
> > > >> > > >>>>
> > > >> > > >>>> StreamBuilder#stream(String topic, Consumed options)
> > > >> > > >>>>
> > > >> > > >>>>
> > > >> > > >>>> Thanks
> > > >> > > >>>>
> > > >> > > >>>>
> > > >> > > >>>> On Thu, 24 Aug 2017 at 22:11 Sriram Subramanian <
> > > >> ram@confluent.io>
> > > >> > > >>> wrote:
> > > >> > > >>>>
> > > >> > > >>>>> +1
> > > >> > > >>>>>
> > > >> > > >>>>> On Thu, Aug 24, 2017 at 10:20 AM, Guozhang Wang <
> > > >> > wangguoz@gmail.com>
> > > >> > > >>>>> wrote:
> > > >> > > >>>>>
> > > >> > > >>>>>> +1. Thanks Damian!
> > > >> > > >>>>>>
> > > >> > > >>>>>> On Thu, Aug 24, 2017 at 9:47 AM, Bill Bejeck <
> > > >> bbejeck@gmail.com>
> > > >> > > >>> wrote:
> > > >> > > >>>>>>
> > > >> > > >>>>>>> Thanks for the KIP!
> > > >> > > >>>>>>>
> > > >> > > >>>>>>> +1
> > > >> > > >>>>>>>
> > > >> > > >>>>>>> Thanks,
> > > >> > > >>>>>>> Bill
> > > >> > > >>>>>>>
> > > >> > > >>>>>>> On Thu, Aug 24, 2017 at 12:25 PM, Damian Guy <
> > > >> > damian.guy@gmail.com
> > > >> > > >
> > > >> > > >>>>>> wrote:
> > > >> > > >>>>>>>
> > > >> > > >>>>>>>> Hi,
> > > >> > > >>>>>>>>
> > > >> > > >>>>>>>> I'd like to kick off the voting thread for KIP-182:
> > > >> > > >>>>>>>> https://cwiki.apache.org/confluence/display/KAFKA/KIP-
> > > >> > > >>>>>>>> 182%3A+Reduce+Streams+DSL+overloads+and+allow+easier+
> > > >> > > >>>>>>>> use+of+custom+storage+engines
> > > >> > > >>>>>>>>
> > > >> > > >>>>>>>> Thanks,
> > > >> > > >>>>>>>> Damian
> > > >> > > >>>>>>>>
> > > >> > > >>>>>>>
> > > >> > > >>>>>>
> > > >> > > >>>>>>
> > > >> > > >>>>>>
> > > >> > > >>>>>> --
> > > >> > > >>>>>> -- Guozhang
> > > >> > > >>>>>>
> > > >> > > >>>>>
> > > >> > > >>>>
> > > >> > > >>>
> > > >> > > >>>
> > > >> > > >>
> > > >> > > >
> > > >> > >
> > > >> > >
> > > >> >
> > > >> >
> > > >> > --
> > > >> > -- Guozhang
> > > >> >
> > > >>
> > > >
> > >
> >
> >
> >
> > --
> > -- Guozhang
> >
>
--
-- Guozhang
Re: [VOTE] KIP-182 - Reduce Streams DSL overloads and allow easier
use of custom storage engines
Posted by Damian Guy <da...@gmail.com>.
Hi Guozhang,
I had an offline discussion with Matthias and Bill about it. It is thought
that `to` offers some benefit, i.e., syntactic sugar, so perhaps no harm in
keeping it. However, `through` less so, seeing as we can materialize stores
via `filter`, `map` etc, so one of the main benefits of `through` no longer
exists. WDYT?
Thanks,
Damian
On Tue, 12 Sep 2017 at 18:17 Guozhang Wang <wa...@gmail.com> wrote:
> Hi Damian,
>
> Why we are deprecating KTable.through while keeping KTable.to? Should we
> either keep both of them or deprecate both of them in favor or
> KTable.toStream if people agree that it is confusing to users?
>
>
> Guozhang
>
>
> On Tue, Sep 12, 2017 at 1:18 AM, Damian Guy <da...@gmail.com> wrote:
>
> > Hi All,
> >
> > A minor update to the KIP, i needed to add KTable.to(Produced) for
> > consistency. KTable.through will be deprecated in favour of using
> > KTable.toStream().through()
> >
> > Thanks,
> > Damian
> >
> > On Thu, 7 Sep 2017 at 08:52 Damian Guy <da...@gmail.com> wrote:
> >
> > > Thanks all. The vote is now closed and the KIP has been accepted with:
> > > 2 non binding votes - bill and matthias
> > > 3 binding - Damian, Guozhang, Sriram
> > >
> > > Regards,
> > > Damian
> > >
> > > On Tue, 5 Sep 2017 at 22:24 Sriram Subramanian <ra...@confluent.io>
> wrote:
> > >
> > >> +1
> > >>
> > >> On Tue, Sep 5, 2017 at 1:33 PM, Guozhang Wang <wa...@gmail.com>
> > wrote:
> > >>
> > >> > +1
> > >> >
> > >> > On Fri, Sep 1, 2017 at 3:45 PM, Matthias J. Sax <
> > matthias@confluent.io>
> > >> > wrote:
> > >> >
> > >> > > +1
> > >> > >
> > >> > > On 9/1/17 2:53 PM, Bill Bejeck wrote:
> > >> > > > +1
> > >> > > >
> > >> > > > On Thu, Aug 31, 2017 at 10:20 AM, Damian Guy <
> > damian.guy@gmail.com>
> > >> > > wrote:
> > >> > > >
> > >> > > >> Thanks everyone for voting! Unfortunately i've had to make a
> bit
> > >> of an
> > >> > > >> update based on some issues found during implementation.
> > >> > > >> The main changes are:
> > >> > > >> BytesStoreSupplier -> StoreSupplier
> > >> > > >> Addition of:
> > >> > > >> WindowBytesStoreSupplier, KeyValueBytesStoreSupplier,
> > >> > > >> SessionBytesStoreSupplier that will restrict store types to
> > <Bytes,
> > >> > > byte[]>
> > >> > > >> 3 new overloads added to Materialized to enable developers to
> > >> create a
> > >> > > >> Materialized of the appropriate type, i..e, WindowStore etc
> > >> > > >> Update DSL where Materialized is used such that the stores have
> > >> > generic
> > >> > > >> types of <Bytes, byte[]>
> > >> > > >> Some minor changes to the arguments to
> > Store#persistentWindowStore
> > >> and
> > >> > > >> Store#persistentSessionStore
> > >> > > >>
> > >> > > >> Please take a look and recast the votes.
> > >> > > >>
> > >> > > >> Thanks for your time,
> > >> > > >> Damian
> > >> > > >>
> > >> > > >> On Fri, 25 Aug 2017 at 17:05 Matthias J. Sax <
> > >> matthias@confluent.io>
> > >> > > >> wrote:
> > >> > > >>
> > >> > > >>> Thanks Damian. Great KIP!
> > >> > > >>>
> > >> > > >>> +1
> > >> > > >>>
> > >> > > >>>
> > >> > > >>> -Matthias
> > >> > > >>>
> > >> > > >>> On 8/25/17 6:45 AM, Damian Guy wrote:
> > >> > > >>>> Hi,
> > >> > > >>>>
> > >> > > >>>> I've just realised we need to add two methods to
> > >> StateStoreBuilder
> > >> > or
> > >> > > >> it
> > >> > > >>>> isn't going to work:
> > >> > > >>>>
> > >> > > >>>> Map<String, String> logConfig();
> > >> > > >>>> boolean loggingEnabled();
> > >> > > >>>>
> > >> > > >>>> These are needed when we are building the topology and
> > >> determining
> > >> > > >>>> changelog topic names and configs.
> > >> > > >>>>
> > >> > > >>>>
> > >> > > >>>> I've also update the KIP to add
> > >> > > >>>>
> > >> > > >>>> StreamBuilder#stream(String topic)
> > >> > > >>>>
> > >> > > >>>> StreamBuilder#stream(String topic, Consumed options)
> > >> > > >>>>
> > >> > > >>>>
> > >> > > >>>> Thanks
> > >> > > >>>>
> > >> > > >>>>
> > >> > > >>>> On Thu, 24 Aug 2017 at 22:11 Sriram Subramanian <
> > >> ram@confluent.io>
> > >> > > >>> wrote:
> > >> > > >>>>
> > >> > > >>>>> +1
> > >> > > >>>>>
> > >> > > >>>>> On Thu, Aug 24, 2017 at 10:20 AM, Guozhang Wang <
> > >> > wangguoz@gmail.com>
> > >> > > >>>>> wrote:
> > >> > > >>>>>
> > >> > > >>>>>> +1. Thanks Damian!
> > >> > > >>>>>>
> > >> > > >>>>>> On Thu, Aug 24, 2017 at 9:47 AM, Bill Bejeck <
> > >> bbejeck@gmail.com>
> > >> > > >>> wrote:
> > >> > > >>>>>>
> > >> > > >>>>>>> Thanks for the KIP!
> > >> > > >>>>>>>
> > >> > > >>>>>>> +1
> > >> > > >>>>>>>
> > >> > > >>>>>>> Thanks,
> > >> > > >>>>>>> Bill
> > >> > > >>>>>>>
> > >> > > >>>>>>> On Thu, Aug 24, 2017 at 12:25 PM, Damian Guy <
> > >> > damian.guy@gmail.com
> > >> > > >
> > >> > > >>>>>> wrote:
> > >> > > >>>>>>>
> > >> > > >>>>>>>> Hi,
> > >> > > >>>>>>>>
> > >> > > >>>>>>>> I'd like to kick off the voting thread for KIP-182:
> > >> > > >>>>>>>> https://cwiki.apache.org/confluence/display/KAFKA/KIP-
> > >> > > >>>>>>>> 182%3A+Reduce+Streams+DSL+overloads+and+allow+easier+
> > >> > > >>>>>>>> use+of+custom+storage+engines
> > >> > > >>>>>>>>
> > >> > > >>>>>>>> Thanks,
> > >> > > >>>>>>>> Damian
> > >> > > >>>>>>>>
> > >> > > >>>>>>>
> > >> > > >>>>>>
> > >> > > >>>>>>
> > >> > > >>>>>>
> > >> > > >>>>>> --
> > >> > > >>>>>> -- Guozhang
> > >> > > >>>>>>
> > >> > > >>>>>
> > >> > > >>>>
> > >> > > >>>
> > >> > > >>>
> > >> > > >>
> > >> > > >
> > >> > >
> > >> > >
> > >> >
> > >> >
> > >> > --
> > >> > -- Guozhang
> > >> >
> > >>
> > >
> >
>
>
>
> --
> -- Guozhang
>
Re: [VOTE] KIP-182 - Reduce Streams DSL overloads and allow easier
use of custom storage engines
Posted by Guozhang Wang <wa...@gmail.com>.
Hi Damian,
Why we are deprecating KTable.through while keeping KTable.to? Should we
either keep both of them or deprecate both of them in favor or
KTable.toStream if people agree that it is confusing to users?
Guozhang
On Tue, Sep 12, 2017 at 1:18 AM, Damian Guy <da...@gmail.com> wrote:
> Hi All,
>
> A minor update to the KIP, i needed to add KTable.to(Produced) for
> consistency. KTable.through will be deprecated in favour of using
> KTable.toStream().through()
>
> Thanks,
> Damian
>
> On Thu, 7 Sep 2017 at 08:52 Damian Guy <da...@gmail.com> wrote:
>
> > Thanks all. The vote is now closed and the KIP has been accepted with:
> > 2 non binding votes - bill and matthias
> > 3 binding - Damian, Guozhang, Sriram
> >
> > Regards,
> > Damian
> >
> > On Tue, 5 Sep 2017 at 22:24 Sriram Subramanian <ra...@confluent.io> wrote:
> >
> >> +1
> >>
> >> On Tue, Sep 5, 2017 at 1:33 PM, Guozhang Wang <wa...@gmail.com>
> wrote:
> >>
> >> > +1
> >> >
> >> > On Fri, Sep 1, 2017 at 3:45 PM, Matthias J. Sax <
> matthias@confluent.io>
> >> > wrote:
> >> >
> >> > > +1
> >> > >
> >> > > On 9/1/17 2:53 PM, Bill Bejeck wrote:
> >> > > > +1
> >> > > >
> >> > > > On Thu, Aug 31, 2017 at 10:20 AM, Damian Guy <
> damian.guy@gmail.com>
> >> > > wrote:
> >> > > >
> >> > > >> Thanks everyone for voting! Unfortunately i've had to make a bit
> >> of an
> >> > > >> update based on some issues found during implementation.
> >> > > >> The main changes are:
> >> > > >> BytesStoreSupplier -> StoreSupplier
> >> > > >> Addition of:
> >> > > >> WindowBytesStoreSupplier, KeyValueBytesStoreSupplier,
> >> > > >> SessionBytesStoreSupplier that will restrict store types to
> <Bytes,
> >> > > byte[]>
> >> > > >> 3 new overloads added to Materialized to enable developers to
> >> create a
> >> > > >> Materialized of the appropriate type, i..e, WindowStore etc
> >> > > >> Update DSL where Materialized is used such that the stores have
> >> > generic
> >> > > >> types of <Bytes, byte[]>
> >> > > >> Some minor changes to the arguments to
> Store#persistentWindowStore
> >> and
> >> > > >> Store#persistentSessionStore
> >> > > >>
> >> > > >> Please take a look and recast the votes.
> >> > > >>
> >> > > >> Thanks for your time,
> >> > > >> Damian
> >> > > >>
> >> > > >> On Fri, 25 Aug 2017 at 17:05 Matthias J. Sax <
> >> matthias@confluent.io>
> >> > > >> wrote:
> >> > > >>
> >> > > >>> Thanks Damian. Great KIP!
> >> > > >>>
> >> > > >>> +1
> >> > > >>>
> >> > > >>>
> >> > > >>> -Matthias
> >> > > >>>
> >> > > >>> On 8/25/17 6:45 AM, Damian Guy wrote:
> >> > > >>>> Hi,
> >> > > >>>>
> >> > > >>>> I've just realised we need to add two methods to
> >> StateStoreBuilder
> >> > or
> >> > > >> it
> >> > > >>>> isn't going to work:
> >> > > >>>>
> >> > > >>>> Map<String, String> logConfig();
> >> > > >>>> boolean loggingEnabled();
> >> > > >>>>
> >> > > >>>> These are needed when we are building the topology and
> >> determining
> >> > > >>>> changelog topic names and configs.
> >> > > >>>>
> >> > > >>>>
> >> > > >>>> I've also update the KIP to add
> >> > > >>>>
> >> > > >>>> StreamBuilder#stream(String topic)
> >> > > >>>>
> >> > > >>>> StreamBuilder#stream(String topic, Consumed options)
> >> > > >>>>
> >> > > >>>>
> >> > > >>>> Thanks
> >> > > >>>>
> >> > > >>>>
> >> > > >>>> On Thu, 24 Aug 2017 at 22:11 Sriram Subramanian <
> >> ram@confluent.io>
> >> > > >>> wrote:
> >> > > >>>>
> >> > > >>>>> +1
> >> > > >>>>>
> >> > > >>>>> On Thu, Aug 24, 2017 at 10:20 AM, Guozhang Wang <
> >> > wangguoz@gmail.com>
> >> > > >>>>> wrote:
> >> > > >>>>>
> >> > > >>>>>> +1. Thanks Damian!
> >> > > >>>>>>
> >> > > >>>>>> On Thu, Aug 24, 2017 at 9:47 AM, Bill Bejeck <
> >> bbejeck@gmail.com>
> >> > > >>> wrote:
> >> > > >>>>>>
> >> > > >>>>>>> Thanks for the KIP!
> >> > > >>>>>>>
> >> > > >>>>>>> +1
> >> > > >>>>>>>
> >> > > >>>>>>> Thanks,
> >> > > >>>>>>> Bill
> >> > > >>>>>>>
> >> > > >>>>>>> On Thu, Aug 24, 2017 at 12:25 PM, Damian Guy <
> >> > damian.guy@gmail.com
> >> > > >
> >> > > >>>>>> wrote:
> >> > > >>>>>>>
> >> > > >>>>>>>> Hi,
> >> > > >>>>>>>>
> >> > > >>>>>>>> I'd like to kick off the voting thread for KIP-182:
> >> > > >>>>>>>> https://cwiki.apache.org/confluence/display/KAFKA/KIP-
> >> > > >>>>>>>> 182%3A+Reduce+Streams+DSL+overloads+and+allow+easier+
> >> > > >>>>>>>> use+of+custom+storage+engines
> >> > > >>>>>>>>
> >> > > >>>>>>>> Thanks,
> >> > > >>>>>>>> Damian
> >> > > >>>>>>>>
> >> > > >>>>>>>
> >> > > >>>>>>
> >> > > >>>>>>
> >> > > >>>>>>
> >> > > >>>>>> --
> >> > > >>>>>> -- Guozhang
> >> > > >>>>>>
> >> > > >>>>>
> >> > > >>>>
> >> > > >>>
> >> > > >>>
> >> > > >>
> >> > > >
> >> > >
> >> > >
> >> >
> >> >
> >> > --
> >> > -- Guozhang
> >> >
> >>
> >
>
--
-- Guozhang
Re: [VOTE] KIP-182 - Reduce Streams DSL overloads and allow easier
use of custom storage engines
Posted by Damian Guy <da...@gmail.com>.
Hi All,
A minor update to the KIP, i needed to add KTable.to(Produced) for
consistency. KTable.through will be deprecated in favour of using
KTable.toStream().through()
Thanks,
Damian
On Thu, 7 Sep 2017 at 08:52 Damian Guy <da...@gmail.com> wrote:
> Thanks all. The vote is now closed and the KIP has been accepted with:
> 2 non binding votes - bill and matthias
> 3 binding - Damian, Guozhang, Sriram
>
> Regards,
> Damian
>
> On Tue, 5 Sep 2017 at 22:24 Sriram Subramanian <ra...@confluent.io> wrote:
>
>> +1
>>
>> On Tue, Sep 5, 2017 at 1:33 PM, Guozhang Wang <wa...@gmail.com> wrote:
>>
>> > +1
>> >
>> > On Fri, Sep 1, 2017 at 3:45 PM, Matthias J. Sax <ma...@confluent.io>
>> > wrote:
>> >
>> > > +1
>> > >
>> > > On 9/1/17 2:53 PM, Bill Bejeck wrote:
>> > > > +1
>> > > >
>> > > > On Thu, Aug 31, 2017 at 10:20 AM, Damian Guy <da...@gmail.com>
>> > > wrote:
>> > > >
>> > > >> Thanks everyone for voting! Unfortunately i've had to make a bit
>> of an
>> > > >> update based on some issues found during implementation.
>> > > >> The main changes are:
>> > > >> BytesStoreSupplier -> StoreSupplier
>> > > >> Addition of:
>> > > >> WindowBytesStoreSupplier, KeyValueBytesStoreSupplier,
>> > > >> SessionBytesStoreSupplier that will restrict store types to <Bytes,
>> > > byte[]>
>> > > >> 3 new overloads added to Materialized to enable developers to
>> create a
>> > > >> Materialized of the appropriate type, i..e, WindowStore etc
>> > > >> Update DSL where Materialized is used such that the stores have
>> > generic
>> > > >> types of <Bytes, byte[]>
>> > > >> Some minor changes to the arguments to Store#persistentWindowStore
>> and
>> > > >> Store#persistentSessionStore
>> > > >>
>> > > >> Please take a look and recast the votes.
>> > > >>
>> > > >> Thanks for your time,
>> > > >> Damian
>> > > >>
>> > > >> On Fri, 25 Aug 2017 at 17:05 Matthias J. Sax <
>> matthias@confluent.io>
>> > > >> wrote:
>> > > >>
>> > > >>> Thanks Damian. Great KIP!
>> > > >>>
>> > > >>> +1
>> > > >>>
>> > > >>>
>> > > >>> -Matthias
>> > > >>>
>> > > >>> On 8/25/17 6:45 AM, Damian Guy wrote:
>> > > >>>> Hi,
>> > > >>>>
>> > > >>>> I've just realised we need to add two methods to
>> StateStoreBuilder
>> > or
>> > > >> it
>> > > >>>> isn't going to work:
>> > > >>>>
>> > > >>>> Map<String, String> logConfig();
>> > > >>>> boolean loggingEnabled();
>> > > >>>>
>> > > >>>> These are needed when we are building the topology and
>> determining
>> > > >>>> changelog topic names and configs.
>> > > >>>>
>> > > >>>>
>> > > >>>> I've also update the KIP to add
>> > > >>>>
>> > > >>>> StreamBuilder#stream(String topic)
>> > > >>>>
>> > > >>>> StreamBuilder#stream(String topic, Consumed options)
>> > > >>>>
>> > > >>>>
>> > > >>>> Thanks
>> > > >>>>
>> > > >>>>
>> > > >>>> On Thu, 24 Aug 2017 at 22:11 Sriram Subramanian <
>> ram@confluent.io>
>> > > >>> wrote:
>> > > >>>>
>> > > >>>>> +1
>> > > >>>>>
>> > > >>>>> On Thu, Aug 24, 2017 at 10:20 AM, Guozhang Wang <
>> > wangguoz@gmail.com>
>> > > >>>>> wrote:
>> > > >>>>>
>> > > >>>>>> +1. Thanks Damian!
>> > > >>>>>>
>> > > >>>>>> On Thu, Aug 24, 2017 at 9:47 AM, Bill Bejeck <
>> bbejeck@gmail.com>
>> > > >>> wrote:
>> > > >>>>>>
>> > > >>>>>>> Thanks for the KIP!
>> > > >>>>>>>
>> > > >>>>>>> +1
>> > > >>>>>>>
>> > > >>>>>>> Thanks,
>> > > >>>>>>> Bill
>> > > >>>>>>>
>> > > >>>>>>> On Thu, Aug 24, 2017 at 12:25 PM, Damian Guy <
>> > damian.guy@gmail.com
>> > > >
>> > > >>>>>> wrote:
>> > > >>>>>>>
>> > > >>>>>>>> Hi,
>> > > >>>>>>>>
>> > > >>>>>>>> I'd like to kick off the voting thread for KIP-182:
>> > > >>>>>>>> https://cwiki.apache.org/confluence/display/KAFKA/KIP-
>> > > >>>>>>>> 182%3A+Reduce+Streams+DSL+overloads+and+allow+easier+
>> > > >>>>>>>> use+of+custom+storage+engines
>> > > >>>>>>>>
>> > > >>>>>>>> Thanks,
>> > > >>>>>>>> Damian
>> > > >>>>>>>>
>> > > >>>>>>>
>> > > >>>>>>
>> > > >>>>>>
>> > > >>>>>>
>> > > >>>>>> --
>> > > >>>>>> -- Guozhang
>> > > >>>>>>
>> > > >>>>>
>> > > >>>>
>> > > >>>
>> > > >>>
>> > > >>
>> > > >
>> > >
>> > >
>> >
>> >
>> > --
>> > -- Guozhang
>> >
>>
>
Re: [VOTE] KIP-182 - Reduce Streams DSL overloads and allow easier
use of custom storage engines
Posted by Damian Guy <da...@gmail.com>.
Thanks all. The vote is now closed and the KIP has been accepted with:
2 non binding votes - bill and matthias
3 binding - Damian, Guozhang, Sriram
Regards,
Damian
On Tue, 5 Sep 2017 at 22:24 Sriram Subramanian <ra...@confluent.io> wrote:
> +1
>
> On Tue, Sep 5, 2017 at 1:33 PM, Guozhang Wang <wa...@gmail.com> wrote:
>
> > +1
> >
> > On Fri, Sep 1, 2017 at 3:45 PM, Matthias J. Sax <ma...@confluent.io>
> > wrote:
> >
> > > +1
> > >
> > > On 9/1/17 2:53 PM, Bill Bejeck wrote:
> > > > +1
> > > >
> > > > On Thu, Aug 31, 2017 at 10:20 AM, Damian Guy <da...@gmail.com>
> > > wrote:
> > > >
> > > >> Thanks everyone for voting! Unfortunately i've had to make a bit of
> an
> > > >> update based on some issues found during implementation.
> > > >> The main changes are:
> > > >> BytesStoreSupplier -> StoreSupplier
> > > >> Addition of:
> > > >> WindowBytesStoreSupplier, KeyValueBytesStoreSupplier,
> > > >> SessionBytesStoreSupplier that will restrict store types to <Bytes,
> > > byte[]>
> > > >> 3 new overloads added to Materialized to enable developers to
> create a
> > > >> Materialized of the appropriate type, i..e, WindowStore etc
> > > >> Update DSL where Materialized is used such that the stores have
> > generic
> > > >> types of <Bytes, byte[]>
> > > >> Some minor changes to the arguments to Store#persistentWindowStore
> and
> > > >> Store#persistentSessionStore
> > > >>
> > > >> Please take a look and recast the votes.
> > > >>
> > > >> Thanks for your time,
> > > >> Damian
> > > >>
> > > >> On Fri, 25 Aug 2017 at 17:05 Matthias J. Sax <matthias@confluent.io
> >
> > > >> wrote:
> > > >>
> > > >>> Thanks Damian. Great KIP!
> > > >>>
> > > >>> +1
> > > >>>
> > > >>>
> > > >>> -Matthias
> > > >>>
> > > >>> On 8/25/17 6:45 AM, Damian Guy wrote:
> > > >>>> Hi,
> > > >>>>
> > > >>>> I've just realised we need to add two methods to StateStoreBuilder
> > or
> > > >> it
> > > >>>> isn't going to work:
> > > >>>>
> > > >>>> Map<String, String> logConfig();
> > > >>>> boolean loggingEnabled();
> > > >>>>
> > > >>>> These are needed when we are building the topology and determining
> > > >>>> changelog topic names and configs.
> > > >>>>
> > > >>>>
> > > >>>> I've also update the KIP to add
> > > >>>>
> > > >>>> StreamBuilder#stream(String topic)
> > > >>>>
> > > >>>> StreamBuilder#stream(String topic, Consumed options)
> > > >>>>
> > > >>>>
> > > >>>> Thanks
> > > >>>>
> > > >>>>
> > > >>>> On Thu, 24 Aug 2017 at 22:11 Sriram Subramanian <ram@confluent.io
> >
> > > >>> wrote:
> > > >>>>
> > > >>>>> +1
> > > >>>>>
> > > >>>>> On Thu, Aug 24, 2017 at 10:20 AM, Guozhang Wang <
> > wangguoz@gmail.com>
> > > >>>>> wrote:
> > > >>>>>
> > > >>>>>> +1. Thanks Damian!
> > > >>>>>>
> > > >>>>>> On Thu, Aug 24, 2017 at 9:47 AM, Bill Bejeck <bbejeck@gmail.com
> >
> > > >>> wrote:
> > > >>>>>>
> > > >>>>>>> Thanks for the KIP!
> > > >>>>>>>
> > > >>>>>>> +1
> > > >>>>>>>
> > > >>>>>>> Thanks,
> > > >>>>>>> Bill
> > > >>>>>>>
> > > >>>>>>> On Thu, Aug 24, 2017 at 12:25 PM, Damian Guy <
> > damian.guy@gmail.com
> > > >
> > > >>>>>> wrote:
> > > >>>>>>>
> > > >>>>>>>> Hi,
> > > >>>>>>>>
> > > >>>>>>>> I'd like to kick off the voting thread for KIP-182:
> > > >>>>>>>> https://cwiki.apache.org/confluence/display/KAFKA/KIP-
> > > >>>>>>>> 182%3A+Reduce+Streams+DSL+overloads+and+allow+easier+
> > > >>>>>>>> use+of+custom+storage+engines
> > > >>>>>>>>
> > > >>>>>>>> Thanks,
> > > >>>>>>>> Damian
> > > >>>>>>>>
> > > >>>>>>>
> > > >>>>>>
> > > >>>>>>
> > > >>>>>>
> > > >>>>>> --
> > > >>>>>> -- Guozhang
> > > >>>>>>
> > > >>>>>
> > > >>>>
> > > >>>
> > > >>>
> > > >>
> > > >
> > >
> > >
> >
> >
> > --
> > -- Guozhang
> >
>
Re: [VOTE] KIP-182 - Reduce Streams DSL overloads and allow easier
use of custom storage engines
Posted by Sriram Subramanian <ra...@confluent.io>.
+1
On Tue, Sep 5, 2017 at 1:33 PM, Guozhang Wang <wa...@gmail.com> wrote:
> +1
>
> On Fri, Sep 1, 2017 at 3:45 PM, Matthias J. Sax <ma...@confluent.io>
> wrote:
>
> > +1
> >
> > On 9/1/17 2:53 PM, Bill Bejeck wrote:
> > > +1
> > >
> > > On Thu, Aug 31, 2017 at 10:20 AM, Damian Guy <da...@gmail.com>
> > wrote:
> > >
> > >> Thanks everyone for voting! Unfortunately i've had to make a bit of an
> > >> update based on some issues found during implementation.
> > >> The main changes are:
> > >> BytesStoreSupplier -> StoreSupplier
> > >> Addition of:
> > >> WindowBytesStoreSupplier, KeyValueBytesStoreSupplier,
> > >> SessionBytesStoreSupplier that will restrict store types to <Bytes,
> > byte[]>
> > >> 3 new overloads added to Materialized to enable developers to create a
> > >> Materialized of the appropriate type, i..e, WindowStore etc
> > >> Update DSL where Materialized is used such that the stores have
> generic
> > >> types of <Bytes, byte[]>
> > >> Some minor changes to the arguments to Store#persistentWindowStore and
> > >> Store#persistentSessionStore
> > >>
> > >> Please take a look and recast the votes.
> > >>
> > >> Thanks for your time,
> > >> Damian
> > >>
> > >> On Fri, 25 Aug 2017 at 17:05 Matthias J. Sax <ma...@confluent.io>
> > >> wrote:
> > >>
> > >>> Thanks Damian. Great KIP!
> > >>>
> > >>> +1
> > >>>
> > >>>
> > >>> -Matthias
> > >>>
> > >>> On 8/25/17 6:45 AM, Damian Guy wrote:
> > >>>> Hi,
> > >>>>
> > >>>> I've just realised we need to add two methods to StateStoreBuilder
> or
> > >> it
> > >>>> isn't going to work:
> > >>>>
> > >>>> Map<String, String> logConfig();
> > >>>> boolean loggingEnabled();
> > >>>>
> > >>>> These are needed when we are building the topology and determining
> > >>>> changelog topic names and configs.
> > >>>>
> > >>>>
> > >>>> I've also update the KIP to add
> > >>>>
> > >>>> StreamBuilder#stream(String topic)
> > >>>>
> > >>>> StreamBuilder#stream(String topic, Consumed options)
> > >>>>
> > >>>>
> > >>>> Thanks
> > >>>>
> > >>>>
> > >>>> On Thu, 24 Aug 2017 at 22:11 Sriram Subramanian <ra...@confluent.io>
> > >>> wrote:
> > >>>>
> > >>>>> +1
> > >>>>>
> > >>>>> On Thu, Aug 24, 2017 at 10:20 AM, Guozhang Wang <
> wangguoz@gmail.com>
> > >>>>> wrote:
> > >>>>>
> > >>>>>> +1. Thanks Damian!
> > >>>>>>
> > >>>>>> On Thu, Aug 24, 2017 at 9:47 AM, Bill Bejeck <bb...@gmail.com>
> > >>> wrote:
> > >>>>>>
> > >>>>>>> Thanks for the KIP!
> > >>>>>>>
> > >>>>>>> +1
> > >>>>>>>
> > >>>>>>> Thanks,
> > >>>>>>> Bill
> > >>>>>>>
> > >>>>>>> On Thu, Aug 24, 2017 at 12:25 PM, Damian Guy <
> damian.guy@gmail.com
> > >
> > >>>>>> wrote:
> > >>>>>>>
> > >>>>>>>> Hi,
> > >>>>>>>>
> > >>>>>>>> I'd like to kick off the voting thread for KIP-182:
> > >>>>>>>> https://cwiki.apache.org/confluence/display/KAFKA/KIP-
> > >>>>>>>> 182%3A+Reduce+Streams+DSL+overloads+and+allow+easier+
> > >>>>>>>> use+of+custom+storage+engines
> > >>>>>>>>
> > >>>>>>>> Thanks,
> > >>>>>>>> Damian
> > >>>>>>>>
> > >>>>>>>
> > >>>>>>
> > >>>>>>
> > >>>>>>
> > >>>>>> --
> > >>>>>> -- Guozhang
> > >>>>>>
> > >>>>>
> > >>>>
> > >>>
> > >>>
> > >>
> > >
> >
> >
>
>
> --
> -- Guozhang
>
Re: [VOTE] KIP-182 - Reduce Streams DSL overloads and allow easier
use of custom storage engines
Posted by Guozhang Wang <wa...@gmail.com>.
+1
On Fri, Sep 1, 2017 at 3:45 PM, Matthias J. Sax <ma...@confluent.io>
wrote:
> +1
>
> On 9/1/17 2:53 PM, Bill Bejeck wrote:
> > +1
> >
> > On Thu, Aug 31, 2017 at 10:20 AM, Damian Guy <da...@gmail.com>
> wrote:
> >
> >> Thanks everyone for voting! Unfortunately i've had to make a bit of an
> >> update based on some issues found during implementation.
> >> The main changes are:
> >> BytesStoreSupplier -> StoreSupplier
> >> Addition of:
> >> WindowBytesStoreSupplier, KeyValueBytesStoreSupplier,
> >> SessionBytesStoreSupplier that will restrict store types to <Bytes,
> byte[]>
> >> 3 new overloads added to Materialized to enable developers to create a
> >> Materialized of the appropriate type, i..e, WindowStore etc
> >> Update DSL where Materialized is used such that the stores have generic
> >> types of <Bytes, byte[]>
> >> Some minor changes to the arguments to Store#persistentWindowStore and
> >> Store#persistentSessionStore
> >>
> >> Please take a look and recast the votes.
> >>
> >> Thanks for your time,
> >> Damian
> >>
> >> On Fri, 25 Aug 2017 at 17:05 Matthias J. Sax <ma...@confluent.io>
> >> wrote:
> >>
> >>> Thanks Damian. Great KIP!
> >>>
> >>> +1
> >>>
> >>>
> >>> -Matthias
> >>>
> >>> On 8/25/17 6:45 AM, Damian Guy wrote:
> >>>> Hi,
> >>>>
> >>>> I've just realised we need to add two methods to StateStoreBuilder or
> >> it
> >>>> isn't going to work:
> >>>>
> >>>> Map<String, String> logConfig();
> >>>> boolean loggingEnabled();
> >>>>
> >>>> These are needed when we are building the topology and determining
> >>>> changelog topic names and configs.
> >>>>
> >>>>
> >>>> I've also update the KIP to add
> >>>>
> >>>> StreamBuilder#stream(String topic)
> >>>>
> >>>> StreamBuilder#stream(String topic, Consumed options)
> >>>>
> >>>>
> >>>> Thanks
> >>>>
> >>>>
> >>>> On Thu, 24 Aug 2017 at 22:11 Sriram Subramanian <ra...@confluent.io>
> >>> wrote:
> >>>>
> >>>>> +1
> >>>>>
> >>>>> On Thu, Aug 24, 2017 at 10:20 AM, Guozhang Wang <wa...@gmail.com>
> >>>>> wrote:
> >>>>>
> >>>>>> +1. Thanks Damian!
> >>>>>>
> >>>>>> On Thu, Aug 24, 2017 at 9:47 AM, Bill Bejeck <bb...@gmail.com>
> >>> wrote:
> >>>>>>
> >>>>>>> Thanks for the KIP!
> >>>>>>>
> >>>>>>> +1
> >>>>>>>
> >>>>>>> Thanks,
> >>>>>>> Bill
> >>>>>>>
> >>>>>>> On Thu, Aug 24, 2017 at 12:25 PM, Damian Guy <damian.guy@gmail.com
> >
> >>>>>> wrote:
> >>>>>>>
> >>>>>>>> Hi,
> >>>>>>>>
> >>>>>>>> I'd like to kick off the voting thread for KIP-182:
> >>>>>>>> https://cwiki.apache.org/confluence/display/KAFKA/KIP-
> >>>>>>>> 182%3A+Reduce+Streams+DSL+overloads+and+allow+easier+
> >>>>>>>> use+of+custom+storage+engines
> >>>>>>>>
> >>>>>>>> Thanks,
> >>>>>>>> Damian
> >>>>>>>>
> >>>>>>>
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>> --
> >>>>>> -- Guozhang
> >>>>>>
> >>>>>
> >>>>
> >>>
> >>>
> >>
> >
>
>
--
-- Guozhang
Re: [VOTE] KIP-182 - Reduce Streams DSL overloads and allow easier
use of custom storage engines
Posted by "Matthias J. Sax" <ma...@confluent.io>.
+1
On 9/1/17 2:53 PM, Bill Bejeck wrote:
> +1
>
> On Thu, Aug 31, 2017 at 10:20 AM, Damian Guy <da...@gmail.com> wrote:
>
>> Thanks everyone for voting! Unfortunately i've had to make a bit of an
>> update based on some issues found during implementation.
>> The main changes are:
>> BytesStoreSupplier -> StoreSupplier
>> Addition of:
>> WindowBytesStoreSupplier, KeyValueBytesStoreSupplier,
>> SessionBytesStoreSupplier that will restrict store types to <Bytes, byte[]>
>> 3 new overloads added to Materialized to enable developers to create a
>> Materialized of the appropriate type, i..e, WindowStore etc
>> Update DSL where Materialized is used such that the stores have generic
>> types of <Bytes, byte[]>
>> Some minor changes to the arguments to Store#persistentWindowStore and
>> Store#persistentSessionStore
>>
>> Please take a look and recast the votes.
>>
>> Thanks for your time,
>> Damian
>>
>> On Fri, 25 Aug 2017 at 17:05 Matthias J. Sax <ma...@confluent.io>
>> wrote:
>>
>>> Thanks Damian. Great KIP!
>>>
>>> +1
>>>
>>>
>>> -Matthias
>>>
>>> On 8/25/17 6:45 AM, Damian Guy wrote:
>>>> Hi,
>>>>
>>>> I've just realised we need to add two methods to StateStoreBuilder or
>> it
>>>> isn't going to work:
>>>>
>>>> Map<String, String> logConfig();
>>>> boolean loggingEnabled();
>>>>
>>>> These are needed when we are building the topology and determining
>>>> changelog topic names and configs.
>>>>
>>>>
>>>> I've also update the KIP to add
>>>>
>>>> StreamBuilder#stream(String topic)
>>>>
>>>> StreamBuilder#stream(String topic, Consumed options)
>>>>
>>>>
>>>> Thanks
>>>>
>>>>
>>>> On Thu, 24 Aug 2017 at 22:11 Sriram Subramanian <ra...@confluent.io>
>>> wrote:
>>>>
>>>>> +1
>>>>>
>>>>> On Thu, Aug 24, 2017 at 10:20 AM, Guozhang Wang <wa...@gmail.com>
>>>>> wrote:
>>>>>
>>>>>> +1. Thanks Damian!
>>>>>>
>>>>>> On Thu, Aug 24, 2017 at 9:47 AM, Bill Bejeck <bb...@gmail.com>
>>> wrote:
>>>>>>
>>>>>>> Thanks for the KIP!
>>>>>>>
>>>>>>> +1
>>>>>>>
>>>>>>> Thanks,
>>>>>>> Bill
>>>>>>>
>>>>>>> On Thu, Aug 24, 2017 at 12:25 PM, Damian Guy <da...@gmail.com>
>>>>>> wrote:
>>>>>>>
>>>>>>>> Hi,
>>>>>>>>
>>>>>>>> I'd like to kick off the voting thread for KIP-182:
>>>>>>>> https://cwiki.apache.org/confluence/display/KAFKA/KIP-
>>>>>>>> 182%3A+Reduce+Streams+DSL+overloads+and+allow+easier+
>>>>>>>> use+of+custom+storage+engines
>>>>>>>>
>>>>>>>> Thanks,
>>>>>>>> Damian
>>>>>>>>
>>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> --
>>>>>> -- Guozhang
>>>>>>
>>>>>
>>>>
>>>
>>>
>>
>