With goimports gone (commit 856c6a8f) the gate is ~37s, of which ~44s of CPU work is spread across steps that do not depend on each other: test, test:tools, test:ui, ui:check, vet, lint, build. Taskfile's check task lists them as sequential 'task:' calls, and only proto/docs/gen depend on ordering. Declaring deps instead of ordering (or running the independent set as parallel deps) should save ~20s on a 16-core box.
Also: ui:deps (npm install) is declared by both test:ui and ui:check, so it runs twice. The ui:check summary already names this as a known gap.
Measured while removing goimports from the gate's formatter set, which dropped the gate from 167s to ~37s.
Close reason
Closed on GitHub; syncing bead state
Imported from beads archie-core-ia38 · GitHub #947 · type task · priority P3 · created 2026-09-19 · closed 2026-09-26
With goimports gone (commit 856c6a8f) the gate is ~37s, of which ~44s of CPU work is spread across steps that do not depend on each other: test, test:tools, test:ui, ui:check, vet, lint, build. Taskfile's check task lists them as sequential 'task:' calls, and only proto/docs/gen depend on ordering. Declaring deps instead of ordering (or running the independent set as parallel deps) should save ~20s on a 16-core box.
Also: ui:deps (npm install) is declared by both test:ui and ui:check, so it runs twice. The ui:check summary already names this as a known gap.
Measured while removing goimports from the gate's formatter set, which dropped the gate from 167s to ~37s.
### Close reason
Closed on GitHub; syncing bead state
---
_Imported from beads `archie-core-ia38` · GitHub #947 · type task · priority P3 · created 2026-09-19 · closed 2026-09-26_
sam
added the lowtask labels 2026-10-27 13:40:07 +00:00
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
With goimports gone (commit 856c6a8f) the gate is ~37s, of which ~44s of CPU work is spread across steps that do not depend on each other: test, test:tools, test:ui, ui:check, vet, lint, build. Taskfile's check task lists them as sequential 'task:' calls, and only proto/docs/gen depend on ordering. Declaring deps instead of ordering (or running the independent set as parallel deps) should save ~20s on a 16-core box.
Also: ui:deps (npm install) is declared by both test:ui and ui:check, so it runs twice. The ui:check summary already names this as a known gap.
Measured while removing goimports from the gate's formatter set, which dropped the gate from 167s to ~37s.
Close reason
Closed on GitHub; syncing bead state
Imported from beads
archie-core-ia38· GitHub #947 · type task · priority P3 · created 2026-09-19 · closed 2026-09-26