Plan: move Skip CI off Intel macOS runners
Nadie ha tomado este issue todavía.
Evaluación
- Dificultad
- 5/5
- Tiempo estimado
- Más de una semana
- Aptitud para principiantes
- 35/100
- Tipo de issue
- Refactorización
- Claridad
- Bastante claro
- Estado de actividad
- Activo
- Stack tecnológico
- android, github-actions, swift, yaml
- Área
- build-system, ci-cd, devops
Línea de trabajo
Start with the linked work in #22, skiptools/skip-ffi#15, skiptools/skipstone#289, and skiptools/skipstone#290, then read the relevant skip-framework.yml workflow and the PRs' runner and test-coverage notes. The issue is a multi-repository migration plan, with some work blocked; check each step's status and runner constraints before taking one on. Done means each migration preserves the tests it previously ran and the stated blocked work is resolved before changing the default runner.
Escrito por el modelo de indexación a partir del texto del issue.
Descripción
This tracks moving Skip's CI off Intel macOS runners. The PRs below link here; each one stands on its own and keeps every test it ran before.
Why
- Speed. The same work runs much faster off Intel. skip-ffi's macOS job: 24–40 min on
macos-26-intel, 8 min on armmacos-26. skipstoneandroid-ci's slowest path: 161–163 min, 75–106 min with this plan. - Queues. skiptools repos share one pool of concurrent macOS jobs. Intel jobs waited 10–56 min (skip-ffi) and up to 7.9 h (skipstone) before starting. Every macOS minute saved shortens that queue for every repo.
- Direction. GitHub announced the end of x86_64 macOS runners (then dated fall 2027), as Apple has dropped the architecture.
macos-26-intelhas since shipped with no end date, so this is a direction rather than a deadline.
Constraints
- No Android emulator on arm macOS runners. No nested virtualization:
HVF error: HV_UNSUPPORTEDonmacos-15andmacos-26, API 28 and 34 (probe). Emulator tests run onubuntu-24.04(KVM). - Not everything builds on Linux yet. Fuse (skip-android-bridge), and the host
swift testof skip-ui (CoreGraphics) and skip-fuse-ui (duplicate static products). Those emulator tests stay on Intel for now.
Steps
Merge order matters for the first two.
- #22:
skip-framework.ymlskips its emulator steps on arm64. Its guard readenv.RUNNER_ARCH, which is always empty, so no caller could move its macOS job to arm. Needs av1tag update. - skiptools/skip-ffi#15:
macos-26+ Ubuntu, after #22 is inv1. Its emulator tests already ran on Ubuntu; the Intel duplicate goes. Run: 11.5 min, set by the Ubuntu job. - skiptools/skipstone#289 (superseded by skiptools/skipstone#300):
android-cionmacos-26+ Ubuntu + onemacos-26-intelleg for the Fuse emulator test. 4 Intel jobs → 2; macOS minutes per run 524 → 217–310. - skiptools/skipstone#290: the Linux skip binary segfaulted transpiling SkipUI (musl's 128 KiB thread stacks). With 8 MiB,
skip android buildworks on Linux: skip-ui 32 s vs 91 s on Intel, skip-fuse-ui 57 s vs 180 s. - Blocked: skip-ui and skip-fuse-ui on
['macos-26', 'ubuntu-24.04'], until their hostswift testbuilds on Linux. Measured so far: on armmacos-26, their iOS, export and Robolectric work takes 20 / 43 min; the Intel job, which also runs the emulator tests, takes 47 / 89 min. - Later: change
skip-framework.yml's defaultruns-on(at least 43 callers), once the above has run for a while.
How coverage was checked
For each PR, the last green upstream run against a fork run of the branch: executed tests per kind (host swift test, Robolectric, iOS, emulator) and per runner, counted from the job logs. Every test that ran before still runs; where one changed runner (emulator tests now only on Ubuntu), the PR says so.
🤖 Generated with Claude Code
- Lenguaje dominante
- Sin datos de lenguaje
- Estrellas
- 0
- Forks
- 7
- Merge medio
- 29 min
- PR fusionados (30 d)
- 1
Preparar el entorno
- Sin Dockerfile ni archivo de Docker Compose
- Sin plantilla de pull request
- Leer la guía de contribución
Primeros pasos
- Lee el issue completo y luego la guía de contribución del proyecto.
- Comenta en el issue que vas a ocuparte — evita que dos personas hagan lo mismo.
- Haz un fork del repositorio y trabaja en una rama.
- Abre un pull request que haga referencia al número del issue.
Issues similares
-
prio:low type:bug
Dificultad 2/5 1-3 horas Aptitud para principiantes 72/100
NatronGitHub/Natron#1638 ·
Los mantenedores suelen responder en 1 día
-
Dificultad 2/5 1-3 horas Aptitud para principiantes 72/100
Los mantenedores suelen responder en 1 día
-
ci good first issue tests
Dificultad 2/5 1-3 horas Aptitud para principiantes 85/100
Haseeb-MernStack/commune#133 ·
Los mantenedores suelen responder en 1 día
-
difficulty: beginner dx good first issue help wanted web
Dificultad 2/5 1-3 horas Aptitud para principiantes 82/100
Yormee-103/PayLoop#158 ·
Los mantenedores suelen responder en 1 día
-
Dificultad 2/5 Menos de una hora Aptitud para principiantes 70/100