Closed akyoto closed 4 years ago
After stumbling upon this comment about Iris I am no longer sure if the use of separate code paths for static routes and dynamic routes is accepted in your benchmark.
Likewise, I am using sync.Pool
for contexts which you - rightfully - criticized, because it's not really 0B.
If the implications of these decisions are unacceptable for your benchmark suite, please let me know.
I was curious to see how it would perform with the latest performance improvements and the numbers are vastly different with v1.3.41 (much better).
About 20-25% better routing performance and 80% less memory usage.
This PR aims to be minimal and only update
go.mod
andgo.sum
for aerogo/aero.