>Also, I'll not add the import statements to ALL examples, only to those
>where people might copy paste them.
That sounds ok to me for now.
On Thu, Nov 19, 2015 at 6:38 PM, Ufuk Celebi <u...@apache.org> wrote:
> I think it's confusing to only have a subset of import statements provided.
> But then again, the missing ones will be resolved without confusion
> (hopefully) ;) We can go with this and see what feedback we get.
>
> (Just doing it for some examples sounds reasonable.)
>
> – Ufuk
>
> On Thu, Nov 19, 2015 at 5:29 PM, Robert Metzger <rmetz...@apache.org> wrote:
>
>> Thank you for the feedback.
>>
>> I was also spending some time thinking about automating this, but I don't
>> have the time right now to bring the required infrastructure in place.
>>
>> For now, I'll just add import statements for classes with the potential of
>> confusion (in particular between the Scala and Java API, Hadoop/Flink
>> classes, ..)
>> Also, I'll not add the import statements to ALL examples, only to those
>> where people might copy paste them.
>> Please -1 me if you are against this, otherwise, I'll soon open a PR.
>>
>>
>> On Wed, Nov 18, 2015 at 5:04 PM, Nick Dimiduk <ndimi...@gmail.com> wrote:
>>
>> > In HBase we keep an hbase-examples module with working code. Snippets
>> from
>> > that module are pasted into docs and referenced. Yes, we do see
>> divergence,
>> > especially when refactor tools are involved. I once looked into a doc
>> tool
>> > for automatically extracting snippets from source code, but that turned
>> > into a rat-hole and didn't pursue it further. Maybe tooling has improved
>> > since then?
>> >
>> > On Wednesday, November 18, 2015, Maximilian Michels <m...@apache.org>
>> > wrote:
>> >
>> > > Hi Robert.
>> > >
>> > > Good suggestion. Generally, it would be nice to have complete code
>> > > examples available in the documentation. Even better, a way to only
>> > > show excerpts of the complete example with the option of copying the
>> > > complete working example.
>> > >
>> > > For instance:
>> > >
>> > > public Example {
>> > > public static void main(String[] args) {
>> > > ExecutionEnvironment env = ...
>> > >
>> > > // BEGIN: example
>> > > env.fromElements(1,2,3,4)
>> > > env.map(element -> element * 2)
>> > > // END: example
>> > >
>> > > env.print();
>> > > }
>> > > }
>> > >
>> > > This still poses the problem that we need to run those examples in an
>> > > automated way to ensure they are actually working.
>> > >
>> > > Cheers,
>> > > Max
>> > >
>> > > On Wed, Nov 18, 2015 at 12:09 PM, Robert Metzger <rmetz...@apache.org
>> > > <javascript:;>> wrote:
>> > > > Hi,
>> > > >
>> > > > I helped somebody yesterday on SO [1] who had issues with the Scala
>> API
>> > > > because he was importing the classes from the Java API.
>> > > > Somebody else complained about this issue as well in the comments
>> below
>> > > the
>> > > > documentation [2], and I think both users are right: Its an
>> unnecessary
>> > > > obstacle when learning Flink that users have to figure out which
>> class
>> > to
>> > > > import.
>> > > >
>> > > > How about adding import statements to the examples?
>> > > > Is there a nicer way of solving the problem?
>> > > >
>> > > >
>> > > > Regards,
>> > > > Robert
>> > > >
>> > > >
>> > > >
>> > > > [1] http://stackoverflow.com/a/33721528/568695
>> > > > [2]:
>> > > >
>> > >
>> >
>> https://ci.apache.org/projects/flink/flink-docs-release-0.10/apis/streaming_guide.html#comment-2365998014
>> > >
>> >
>>
