Re: Move Row, RowInputFormat to core package

2016-11-28 Thread Timo Walther
la module? -Original Message- From: Aljoscha Krettek [mailto:aljos...@apache.org] Sent: Monday, November 28, 2016 3:00 PM To: dev@flink.apache.org Subject: Re: Move Row, RowInputFormat to core package If we move it to core, we have to untangle it from Scala, as Timo said. The reason is that we wou

RE: Move Row, RowInputFormat to core package

2016-11-28 Thread Anton Solovev
What do you think about moving "Row" not into core module, but into Scala module? -Original Message- From: Aljoscha Krettek [mailto:aljos...@apache.org] Sent: Monday, November 28, 2016 3:00 PM To: dev@flink.apache.org Subject: Re: Move Row, RowInputFormat to core package If

Re: Move Row, RowInputFormat to core package

2016-11-28 Thread Aljoscha Krettek
le > (https://github.com/apache/flink/compare/master...tonycox:FLINK-2186-x) > (https://travis-ci.org/tonycox/flink/builds/178846355) > > -Original Message- > From: Flavio Pompermaier [mailto:pomperma...@okkam.it] > Sent: Friday, November 25, 2016 5:59 PM > To: de

RE: Move Row, RowInputFormat to core package

2016-11-25 Thread Anton Solovev
om/apache/flink/compare/master...tonycox:FLINK-2186-x) (https://travis-ci.org/tonycox/flink/builds/178846355) -Original Message- From: Flavio Pompermaier [mailto:pomperma...@okkam.it] Sent: Friday, November 25, 2016 5:59 PM To: dev@flink.apache.org Subject: Re: Move Row, RowInputFormat to core package

Re: Move Row, RowInputFormat to core package

2016-11-25 Thread Flavio Pompermaier
Fully agree with Timo :) On Fri, Nov 25, 2016 at 2:30 PM, Timo Walther wrote: > Hi Anton, > > I would also support the idea of moving Row and RowTypeInfo to Flink core. > I think there are many real-world use cases where a variable-length record > that supports null values is required. However,

Re: Move Row, RowInputFormat to core package

2016-11-25 Thread Timo Walther
Hi Anton, I would also support the idea of moving Row and RowTypeInfo to Flink core. I think there are many real-world use cases where a variable-length record that supports null values is required. However, I think that those classes needs to be reworked before. They should not depend on Sca