Conversation
…dd options for changing them. Both caches dump based on time
There was a problem hiding this comment.
seems like a magic number. Config? val?
High memory water mark when to flush and when to crush down a value
There was a problem hiding this comment.
Shouldn't this use the pool?
There was a problem hiding this comment.
I'm not sure how the pool could use the pool to shutdown? it would be recursive right?
There was a problem hiding this comment.
Is it not safe to call shutdown from a thread that is created by the pool? If not, how do we ensure that?
Not a big deal, but if you don't want to tackle, add an issue, a link to it here and a comment explaining the concern about correctness and using the pool to shut itself down.
There was a problem hiding this comment.
You could call shutdown, but this currently waits to ensure its shutdown before returning -- if we drop that part (awaitShutdown). Then we could do it inside the pool?
There was a problem hiding this comment.
What about: Runtime.getRuntime().availableProcessors() as the default pool size?
http://stackoverflow.com/questions/4759570/finding-number-of-cores-in-java
|
@johnynek poke |
There was a problem hiding this comment.
more comments here. So this branch happens when there IS a value for this key, right? Not there is confusing to me.
There was a problem hiding this comment.
I don't like these. Can we replace with JavaConverters._ and use the .asJava .asScala syntax. Implicit conversions of this kind bite us.
There was a problem hiding this comment.
(breakOut) magic is supposedly faster here:
import scala.collection.breakOut
flatMap { k => ... }(breakOut)
http://stackoverflow.com/questions/1715681/scala-2-8-breakout
Pluggable caches in FinalFlatMap
|
Thanks for pushing this through. Next up, see if we can get it upping Ashu's throughput? |
2 participants
Added an synchronous cache, pulled cache logic out thats used by the FFM. Allow ticks to trigger flushing.