Skip to content

josh gruenberg
joshng

Apr 22, 2016
joshng commented on pull request apache/kafka#1229
@joshng

This could be meaningfully optimized by checking for hashCode equality first (particularly if hashCode is computed in the constructor).

Apr 22, 2016
joshng commented on pull request apache/kafka#1229
@joshng

(I thought I'd posted this comment before, but I can't find it now, so I'll repost just in case; sorry if I'm repeating myself): Is it always the c…

Apr 19, 2016
joshng commented on pull request apache/kafka#1229
@joshng

It may not a problem given the current usage, but thought I'd point out that the comment on this particular method conflicts with its implementatio…

Apr 19, 2016
joshng commented on pull request apache/kafka#1229
@joshng

I don't think the performance would be any worse than ByteBuffer: it should really be faster, both because caching the hashCodes might be a win, an…

Apr 16, 2016
joshng commented on pull request apache/kafka#1229
@joshng

could set checked = true only in the case where it was false to begin with

Apr 16, 2016
joshng commented on pull request apache/kafka#1229
@joshng

@ijuma I think this is a tough call... It's true, Java/Scala collections don't do this... and bugs crop up very frequently as a result. (Observe: t…

Apr 16, 2016
joshng commented on pull request apache/kafka#1229
@joshng

Hmm, this test wouldn't fail if the exception wasn't thrown :-)

Apr 16, 2016
joshng commented on pull request apache/kafka#1229
@joshng

I agree: rather than using mutable and error-prone ByteBuffers, I'd prefer to see a simple custom-built immutable type here. This could also permit…

Apr 16, 2016
joshng commented on pull request apache/kafka#1229
@joshng

IIRC, I think the intent of using TreeSets here was to avoid the overhead of allocating ByteBuffers for each key? This might be a win, but I note t…

Apr 16, 2016
joshng commented on pull request apache/kafka#1229
@joshng

I haven't fully digested the way these things are being used, but this seems like an odd/fragile way to use ByteBuffers. Why not assume that the ca…

Apr 16, 2016
joshng commented on pull request apache/kafka#1229
@joshng

Doesn't this mutate the position of the buf? (I hate ByteBuffers for this reason :-P)

Apr 16, 2016
joshng commented on pull request apache/kafka#1229
@joshng

Given the frequency of execution in this area, it be worthwhile to lift some low-level lexical-comparison optimizations from here: https://github.c…

Apr 15, 2016
@joshng
  • 61d84f5
    use ByteBuffer for byte[] keys in maps/sets
Apr 4, 2016
@joshng
  • 28be078
    use ByteBuffer for byte[] keys in maps/sets
joshng created branch bytebuffer-keys at joshng/kafka
Apr 4, 2016
Mar 24, 2016
Mar 23, 2016
Feb 19, 2016
joshng pushed to master at joshng/papaya
@joshng
  • 3b8fb81
    pruning dependencies: delete AysncTrace, introduce papaya-proxy
Feb 19, 2016
@joshng
Feb 19, 2016
@joshng
  • @joshng 7eb0dd2
    Replace master branch with page content via GitHub
Feb 19, 2016
joshng created repository joshng/joshng.github.io
Feb 19, 2016
Feb 17, 2016
@joshng
  • @joshng 5561f5b
    Maven artifacts for 0-SNAPSHOT
Feb 15, 2016
joshng pushed to master at joshng/papaya
@joshng
  • @joshng bf53b4d
    introduce multi-module pom
Feb 15, 2016
joshng pushed to master at joshng/papaya
@joshng
Feb 11, 2016
@joshng
  • @joshng 0d4a81d
    Maven artifacts for 0-SNAPSHOT
Feb 11, 2016
@joshng
  • @joshng 3d5bd42
    Maven artifacts for 0-SNAPSHOT
Feb 11, 2016
joshng pushed to master at joshng/papaya
@joshng
Feb 11, 2016
@joshng
  • @joshng 6a87390
    Maven artifacts for 0-SNAPSHOT
Feb 11, 2016
@joshng
  • @joshng f5fe421
    Maven artifacts for 0-SNAPSHOT
Something went wrong with that request. Please try again.