GitHub user andrewor14 opened a pull request:
https://github.com/apache/spark/pull/5685
[SPARK-7120][SPARK-7121][WIP] Closure cleaner nesting + documentation
For instance, in SparkContext, I tried to do the following:
{code}
def scope[T](body: => T): T = body // no-op
def myCoolMethod(path: String): RDD[String] = scope {
parallelize(1 to 10).map { _ => path }
}
{code}
and I got an exception complaining that SparkContext is not serializable.
The issue here is that the inner closure is getting its path from the outer
closure (the scope), but the outer closure actually references the SparkContext
object itself to get the `parallelize` method.
Note, however, that the inner closure doesn't actually need the
SparkContext; it just needs a field from the outer closure. If we modify
ClosureCleaner to clean the outer closure recursively while using the fields
accessed by the inner closure, then we can serialize the inner closure.
This is blocking my effort on a separate task. This is WIP because I plan
to add tests for this later.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/andrewor14/spark closure-cleaner
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/spark/pull/5685.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 #5685
----
commit 86f78237b7623e4efa06c5feb053e0c304979c73
Author: Andrew Or <[email protected]>
Date: 2015-04-24T10:05:58Z
Implement transitive cleaning + add missing documentation
See in-code comments for more detail on what this means.
commit 2390a608ed74a9703d3763d040421dccb51242ec
Author: Andrew Or <[email protected]>
Date: 2015-04-24T10:08:11Z
Feature flag this new behavior
... in case anything breaks, we should be able to resort to old
behavior.
----
---
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.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
