I gave a (techinical) keynote at the ICIS conference in Shanghai a year or two ago. It was a fun talk and I was recently asked for the notes so I thought I'd post them.
When I was at Google I spent some of my 20% time simply surfing through the Java code base looking for race conditions. It was a lot of fun and often I was patching code that was stuff that hadn't been changed in forever or was completely out of my normal stomping grounds.
After submitting the code changes with the comment 'Fix a race condition', I can remember at least one code reviewer responding, "Approved. Cool. But, what race condition?"
This talk covers a bunch of the common cases I encountered.
LeapToMultithreading.pdf
1 comment:
There often is a best synchronization solution, but IT'S not often worth finding it.
page 37
Post a Comment