Github user srowen commented on a diff in the pull request:
https://github.com/apache/spark/pull/325#discussion_r11471958
--- Diff: core/src/main/scala/org/apache/spark/Logging.scala ---
@@ -135,4 +136,6 @@ trait Logging {
private object Logging {
@volatile private var initialized = false
val initLock = new Object()
+ SLF4JBridgeHandler.removeHandlersForRootLogger()
--- End diff --
Completely right, was just commenting that this was the motivating case,
SLF4J -> JUL, rather than JUL -> something else. It would fail at runtime
if you removed this dependency, but that is necessary. I agree with not
depending directly on this class.
On Thu, Apr 10, 2014 at 7:37 AM, Patrick Wendell
<[email protected]>wrote:
> In core/src/main/scala/org/apache/spark/Logging.scala:
>
> > @@ -135,4 +136,6 @@ trait Logging {
> > private object Logging {
> > @volatile private var initialized = false
> > val initLock = new Object()
> > + SLF4JBridgeHandler.removeHandlersForRootLogger()
>
> @srowen <https://github.com/srowen> the case I'm concerned about is the
> following. Someone wants to map SLF4J to JUL, so they'd need to remove the
> "jul-to-slf4j" dependency otherwise they'll get a routing loop. But then
> spark won't compile because this code depends on the
SLF4JBridgeHandlerwhich exists only in the "jul-to-slf4j" jar. Am I missing
something, or
> won't this cause a compile error for a user in that situation?
>
> â
> Reply to this email directly or view it on
GitHub<https://github.com/apache/spark/pull/325/files#r11471909>
> .
>
---
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.
---
