> I find defer pretty expensive, I don't use them that liberally anymore, only when they're really called for. I use defer a lot so I decided to test your stat

> I find defer pretty expensive, I don't use them that liberally anymore, only when they're really called for. I use defer a lot so I decided to test your statement, and you were right. The testcase is not representative of the functionality in github.com/sebcat/store. here it is: https://pastie.se/e4352d50 https://pastie.se/e4352d50 BenchmarkNoDefer 10000000 177 ns/op BenchmarkDefer 2000000 633 ns/op ok foo 3.874s