Topic Radar (X + Ecosystem Signals)
Topic Radar (X + Ecosystem Signals)
Overview
This chapter captures themes that repeatedly surface in Go engineering discourse on X and in adjacent blogs — used as a backlog radar for this deep-dives part. Signals change; the curriculum should track problems, not hype.
Diagram: Radar loop
flow:
[Repro]
|
v
[Chapter]
[Chapter]
|
v
[Link]
Hot Themes Observed
Runtime & scheduler
- GMP explainers and GOMAXPROCS capacity math
- Go 1.26 removal of P-level syscall state / cgo path simplification
- Go 1.27 generic methods, json/v2 GA, goroutineleak profile
- Channels as runtime queues (sudog), not magic
- GC pauses as product latency (Green Tea GC chatter)
Correctness under concurrency
- Context timeouts as non-optional
- Goroutine leak checklists
- Race detector as culture, not CI checkbox
testing/synctestreplacingsleeptests
Performance craft
- Escape analysis + pprof as literacy
sync.Pooland zero-alloc append patterns- Zero-copy
io.Copy/ sendfile limits with TLS time.Afterabuse in hot selects
Production engineering
- Default
http.Clientconsidered harmful - Structured logs with request IDs
- Mutex held across I/O
- Container CPU limits vs GOMAXPROCS
Learning meta
- “Features ≠ engineering”
- Read
src/runtimeworkshops - Interview lists that mix runtime + design judgment
Mapped Into This Part
| Signal | Chapter |
|---|---|
| GMP / syscall P changes | 201, 228 |
| Channel internals posts | 203 |
| synctest runtime (e.g. internals-for-interns.com) | 223 |
| Zero-copy / sendfile | 224 |
| Production top mistakes | 237 |
| GC / pprof | 208, 222 |
| Pool / allocs | 219 |
| Reading source | 240 |
Still-Fertile Backlog (Future)
encoding/json/v2migration cookbooks now that 1.27 shipped stable defaults- Full Swiss-map implementation walk with diagrams per release
- Runtime metrics export (
/debug→ OTel) cookbook - FIPS / boringcrypto build graphs for regulated shops
- Delve scripting for runtime bugs
- Port of
runtimeexecution tracer event schema reference - Arena alternatives pattern library (sync.Pool arenas, slab allocators)
- CPU profile vs wall-clock eBPF cross-check on Linux
How To Extend This Book
When a topic trends:
- Write a minimal repro that shows the symptom.
- Add a chapter under
20-go-deep-dives/with experiment + cross-links. - Link from this radar and from index.
- Avoid duplicating Part 07/18 tutorials — keep the why/runtime angle.
Experiment
# personal radar loop
rg -n 'TODO|FIXME|race|timeout' --glob '*.go' | headConvert one production footgun into a regression test this week.
Try next: Subscribe to Go release notes + one runtime blog; every release, tick experiments in GOEXPERIMENT.