During the Scala API rewrite thread a minor naming issue popped up:
The scala examples are located at the org.apache.flink.examples.scala
package, while the java ones in the org.apache.flink.example.java. (Both
are located in the flink-examples project)
Kostas agreed that a single convention should be desirable (to rule them
all :)) and voted for examples.
Well, the one to rule them all approach will not work:
- Spargel and streaming have their own examples subpackages,
org.apache.flink.{spargel.java, streaming}.examples. As these are in the
respective projects this is fine.
- flink-jdbc, flink-avro, flink-hbase contain example packages, but all
of them actually contain single examples. As these do not require
additional examples and are currently active parts of the repo I would not
touch them.
- Flink-hadoop-compatibility contains 4 different packages containing
examples. The 4 packages logically refer to different Hadoop API-s. Each
one is named <api_name>.example. This can also stay as it is or can be
extracted to a common examples package.
Personally I would only change the org.apache.flink.example.java package to
org.apache.flink.examples.java. If there are no objections I'm renaming the
package on Aljoscha's scala-rework branch which is soon to be merged to the
main repo and contains considerable update on the flink-examples project.
Cheers,
Marton
