GopherCon 2021 - Day 1
summary
Go 1.18 brings generics: type parameters, type sets with the tilde token, and type inference so you can write a single binary-tree or mean function usable across types. Profiling talk shows real measurements: CPU/mutex profilers add little cost but the runtime tracer can be expensive, so test profiler overhead in production. Real-world wins include a Capital One migration that processed ~841 TB and retired a mainframe, a search service cutting memory use ~5x (500→150 GB cluster), and a $1,800/month saving by replacing Redis with disk-backed SSDB.
AI-written summary · about
chapters (11)
- 0:00Pre-Show
- 19:51Welcome - Day 1
- 29:15Robert Greisemer & Ian Lance Taylor - Generics!
- 1:09:17Felix Geisendörfer - Go Profiling and Observability from Scratch
- 1:58:22Kaylyn Gibilterra - Go at Mainframe Scale
- 2:53:23Lightning Talks
- 3:44:05Connor Hicks - The What's, Why's, and How's of Running WebAssembly in Go
- 4:22:18Daniel Selans - Reliability Nirvana
- 5:24:48Go Time Live - Go Beyond Work
- 6:13:27Akhil Indurti - Writing a Static Analyzer for Go Code
- 6:50:27Linus Lee - When Toy Languages Grow Up
