GitHub user Ziemin opened a pull request:
https://github.com/apache/incubator-predictionio/pull/288
[PIO-30] Set up a cross build for Scala 2.10 (Spark 1.6.2) and Scala â¦
I treat this PR as an RFC to get some insights on what could be improved
and what would be the best for the community. I don't think this is ready for
merge and it also should not be a part of the upcoming release, but it's an
example of working crossbuild setup.
### The good
The key changes include:
* `build.sbt` - here I set two versions of scala for the cross build. With
regards to the current version appropriate versions of spark, akka and hadoop
are chosen. To run an sbt command for both setups at the same time, you simply
need to type, e.g sbt +test. To choose scala 2.11.5 specifically you just have
to write e.g. sbt ++2.10.5 scalastyle. Using sbt in a default way will always
resort to 2.11.8.
*
`data/src/main/scala-2.10/org/apache/predictionio/data/SparkVersionDependent.scala`
and
`data/src/main/scala-2.11/org/apache/predictionio/data/SparkVersionDependent.scala`
- are the only examples of version dependent code. They are solely for
providing a proper type of an object for Spark sql related actions. Sbt is
smart enough to include version specific source paths like these.
* `make_distribution.sh` - in order to create an archive for Scala 2.10.5
one has to provide it with an argument (./make_distribution.sh 2.10.5). By
default it will use 2.11.8.
* **integration tests** - The docker image is updated, I pushed it with a
tag spark_2.0.0 not to interfere with the current build. It contains both
versions of Spark and on startup sets up environment according to dependencies
that predictionIO was built with. It uses a simple Java program
`tests/docker-files/BuildInfoPrinter.java` linked with the assembly of
PredictionIO to acquire necessary information. Travis CI makes use of the setup
and runs 8 parallel builds, the number doubled because of introducing two
different scala versions.
### The bad
Updating Spark caused some troubles
* The classpath has to be extended to run the unit tests successfully for
the data sub-package. (see `build.sbt`)
* Column names have to be handled differently for Postgres in
`JDBCPevents`, as Spark surrounds them with "..." what breaks the current
schema in this case
* `tests/pio_tests/utils.py` - has a special Spark pass through argument to
set `spark.sql.warehouse.dir`, because the defaults cause runtime exceptions.
See ->
[here](https://mail-archives.apache.org/mod_mbox/spark-user/201608.mbox/%3ccamassd+efz+uscmnzvkfp00qbr9ynv8lrfhvz9lrmnwh2vk...@mail.gmail.com%3E)
### The ugly
I haven't updated the install script. I think that there are too many
places containing version strings and related dependencies. This is the same
for setting up and downloading a proper version of spark in the tests. I think
that we should come up with a cleaner way of choosing a profile that would be
consistent between different scripts and easier to maintain. Currently bumping
version of scala or spark in one place involves modifying many files and is
very error prone. Ideally there should be a one place with a description of a
profile that all the other scripts could use for the set up.
I hope that this PR will lead to some discussion and we will finally devise
a better solution.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/Ziemin/incubator-predictionio crossbuild
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/incubator-predictionio/pull/288.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #288
----
commit 85c06aa3cae53ca8158db34e3f5c6788e06b38cf
Author: Marcin ZiemiÅski <[email protected]>
Date: 2016-08-09T21:20:54Z
[PIO-30] Set up a cross build for Scala 2.10 (Spark 1.6.2) and Scala 2.11
(Spark 2.0.0).
Changes also include updating travis integration tests, which run now
eight parallel builds.
----
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
