johnou

4 ms·

1. 1.▲
   by johnou 9y ago

   I'd appreciate it if you could nag one of your colleagues to pick it up ;)
2. 2.▲
   by johnou 9y ago

   It's also about giving back to the open source community.
3. 3.▲
   by johnou 9y ago

   In this case I feel changing the LRUCache to support generics was an integral part of the changeset to prevent possible regressions with cache usage (as I was changing the cache key from String to another type). Without it, it would be poss
4. 4.▲
   by johnou 9y ago

   But then you are stuck with the maintenance burden..
5. 5.▲
   by johnou 9y ago

   I added generics to ensure type safety, it is not technically diamond brackets (they came in jdk7 but mysql connector targets jdk6).
6. 6.▲
   by johnou 9y ago

   I didn't benchmark the difference between cache usage with precalculating the hash code in a constructor vs calculation in the hash code itself but string concatenation then taking the hash code of that was the real performance killer.
7. 7.▲
   by johnou 9y ago

   The link is right down the bottom but the GitHub page also contains the benchmark code and results which seemed more interesting https://bugs.mysql.com/bug.php?id=88021
8. 8.▲
   by johnou 9y ago

   And 250% throughput increase with their cache when using their compound cache key which is not strictly related to server-side prepared statements.
9. 9.▲
   Fixing high GC pressure when MySQL driver uses server-side prepared statements(github.com)

   71 points by johnou 9y ago|35 comments
10. 10.▲
    by johnou 10y ago

    Thanks for the tip!
11. 11.▲
    Show HN: Extend and automate JIRA workflows using Groovy

    2 points by johnou 10y ago|2 comments