1. I think it would need a KIP discussion, yes. 2. What sub-tasks are you considering to break KAFKA-4436?
Guozhang On Thu, Jan 19, 2017 at 5:52 PM, Dongjin Lee <dong...@apache.org> wrote: > Guozhang and all, > > No problem. I just took KAFKA-2898. Let's concentrate on KAFKA-4436: > > 1. Does this issue require KIP? If it is true, I will prepare the KIP. > 2. Should this issue be moved into Umbrella task, with separated sub-tasks? > > Best, > Dongjin > > On Fri, Jan 20, 2017 at 2:22 AM, Guozhang Wang <wangg...@gmail.com> wrote: > >> Dongjin, >> >> Thanks for your interests. I looked at these three tickets, and I think we >> can tackle them incrementally instead of getting them in one shot, that is >> because: >> >> 1. Although I see some dependencies between 4436 and 2898, I think they >> can >> still be tackled separately (though like you said, probably better >> tackling >> 2898 then consider doing 4436 which would likely require a KIP discussion >> and we may want to do that for all clients, not only Streams). >> >> 2. Vahid seems already actively working on 4531, which I think is a bit >> orthogonal to other two. >> >> 3. We prefer to work on smaller tasks, which generates smaller patches to >> review etc. >> >> >> So how about start picking on 2898 first, which I think can be done in >> parallel with 4531? >> >> Guozhang >> >> >> >> >> On Wed, Jan 18, 2017 at 12:31 AM, Dongjin Lee <dong...@apache.org> wrote: >> >> > Hello. This mail is related to KAFKA-4436[^1], which Matthias left a >> > comment recently. I started to investigate this issue from last weekend, >> > and I thought that it would be much better to take your advice. >> > >> > What I would like to talk about this issue is following two things: >> > >> > 1. This issue is much larger than expected. To support >> > StreamsConfig#builder, we have to implement builders for ProducerConfig >> and >> > ConsumerConfig first. We can start this job only after implementing and >> > validating all the methods of ProducerConfig and ConsumerConfig. >> > 2. Some other config validation issues are related to this issue. I >> already >> > found following two, but it seems like there are more related ones: >> > - KAFKA-2898: Kafka producer config incorrectly warns that >> serializer.* >> > isn't a known config[^2] >> > - KAFKA-4531: Rationalise client configuration validation[^3] >> > >> > So, I suggest: >> > >> > 1. Change KAFKA-4436 into umbrella issue and add ProducerConfig, >> > ConsumerConfig and StreamsConfig as separate sub-issues. >> > 2. Assign above issues to me. I will take all the issues - and if you >> find >> > any validation issues like KAFKA-2898 or KAFKA-4531, don't hesitate to >> > leave a comment or send me a message. I will also handle them with the >> > config builders. >> > 3. If needed, I will submit the KIP for this issue. >> > >> > Thanks for reading. I am looking forward to your comments. >> > >> > Best, >> > Dongjin >> > >> > [^1]: https://issues.apache.org/jira/browse/KAFKA-4436 >> > [^2]: https://issues.apache.org/jira/browse/KAFKA-2898 >> > [^3]: https://issues.apache.org/jira/browse/KAFKA-4531 >> > >> > -- >> > *Dongjin Lee* >> > >> > >> > *Software developer in Line+.So interested in massive-scale machine >> > learning.facebook: www.facebook.com/dongjin.lee.kr >> > <http://www.facebook.com/dongjin.lee.kr>linkedin: >> > kr.linkedin.com/in/dongjinleekr >> > <http://kr.linkedin.com/in/dongjinleekr>github: >> > <http://goog_969573159/>github.com/dongjinleekr >> > <http://github.com/dongjinleekr>twitter: www.twitter.com/dongjinleekr >> > <http://www.twitter.com/dongjinleekr>* >> > >> >> >> >> -- >> -- Guozhang >> > > > > -- > *Dongjin Lee* > > > *Software developer in Line+.So interested in massive-scale machine > learning.facebook: www.facebook.com/dongjin.lee.kr > <http://www.facebook.com/dongjin.lee.kr>linkedin: > kr.linkedin.com/in/dongjinleekr > <http://kr.linkedin.com/in/dongjinleekr>github: > <http://goog_969573159/>github.com/dongjinleekr > <http://github.com/dongjinleekr>twitter: www.twitter.com/dongjinleekr > <http://www.twitter.com/dongjinleekr>* > -- -- Guozhang
