feat(update): 页面一键更新 + 守护启动脚本

系统配置页新增「版本与更新」卡片,顶栏版本号旁在有新版时亮出提示胶囊。
更新链路:下载发布包 → 比对 SHA256SUMS → -h 冒烟测试 → 暂存 artex.new
→ 以退出码 75 退出 → 守护脚本重新拉起并完成换装,页面轮询 /api/health
等版本号变化后自动刷新。

新增 start.sh / start.bat 作为正式启动入口,只按退出码决定是否重新拉起
(0 停止 / 75 应用新版 / 其余崩溃退避)。校验、冒烟、换装、回滚全部收在
selfupdate 包里而不写进脚本:这些逻辑在 sh 和 bat 上要写两套,而它们恰恰
最不能出错——换上跑不起来的二进制后守护进程会反复拉起它,只能上机器手工救。
Docker 的 ENTRYPOINT 一并改为该脚本,并由它转发 SIGTERM,否则 docker stop
只发给 PID 1,artex 做不了优雅关闭会被 SIGKILL 硬杀。

失败保护两道闸门互补:无法 exec 由换装前的冒烟挡住,能执行但初始化崩溃
由启动计数兜住(连续 3 次失败自动回滚 artex.old,坏的留作 artex.failed)。

GitHub 查询结果后端缓存 30 分钟(失败 2 分钟)。未认证 API 是每 IP 每小时
60 次,且这 60 次是服务器 IP 上所有用户共用的,顶栏每次整页加载都查一次,
不缓存会很快耗光配额导致真想更新时反而查不动。手动「检查更新」带 force=1
绕过缓存。

安全:只认 GitHub 域名且强制 HTTPS(逐跳校验重定向),发布源写死不可配置;
升级文件挂在可执行文件所在目录而非 CWD(服务化运行时 CWD 可能是 /);
开发构建版本号不可比较,禁用一键更新。

走完整 zip 而非裸二进制,因为现有 Release 的 SHA256SUMS 只覆盖 zip,
这样无需改 CI,也兼容已发布的历史版本。

已知限制:Docker 下只换程序不换镜像(工具链不同步、重建容器会退回镜像
版本);不同步发布包里的 skills/;更新即重启会中断运行中的任务。
This commit is contained in:
Autumn-27 committed 2026-09-11 12:39:59 -04:00
1 parent 557317491b
commit 44156a0a4c
22 files changed
+2619 -16

No files matched your search

+37 -1
View File
@@ -18,6 +18,7 @@ import (
"github.com/Autumn-27/artex/agent"
"github.com/Autumn-27/artex/config"
"github.com/Autumn-27/artex/selfupdate"
"github.com/Autumn-27/artex/server"
)
@@ -41,7 +42,15 @@ func printBanner(addr string) {
version, runtime.GOOS, runtime.GOARCH, runtime.Version(), addr)
}
// main only maps run's result onto the process exit code. The exit code is part
// of the update protocol — the supervising start script reads it to decide
// whether to relaunch us (see selfupdate.ExitRestart) — so the body has to live
// in a function that can *return* rather than os.Exit past its own defers.
func main() {
os.Exit(run())
}
func run() int {
var (
addr = flag.String("addr", ":8787", "HTTP listen address")
dataDir = flag.String("data", filepath.Join(config.BaseDir(), "data"), "data directory for SQLite stores (default: data/ next to the executable)")
@@ -59,6 +68,16 @@ func main() {
// page can show a live log stream. Do this first, to catch startup logs too.
server.StartLogCapture()
// Self-update bootstrap: swap in a staged binary, or count a post-swap boot
// attempt and roll back if the new build keeps dying. Must run before we open
// the stores or bind a port — this may end with "exit and let the start script
// relaunch me", and there is no point paying for either first.
action, upState := selfupdate.Bootstrap()
server.SetBootUpdateState(upState)
if action == selfupdate.Restart {
return selfupdate.ExitRestart
}
// surface which config file the binary reads (absolute, so `go run`'s relative
// "config.json" — resolved against the CWD — is unambiguous).
cfgPath := config.Path()
@@ -82,6 +101,12 @@ func main() {
}
defer mgr.Close()
// Surviving this long means a freshly swapped-in build actually works, so drop
// the upgrade marker and stop counting attempts. Until it fires, every boot
// increments the count and a build that keeps dying gets rolled back.
settle := time.AfterFunc(selfupdate.SettleDelay, selfupdate.Settle)
defer settle.Stop()
skillDir := config.SkillDir()
if abs, err := filepath.Abs(skillDir); err == nil {
skillDir = abs
@@ -101,11 +126,22 @@ func main() {
}
}()
<-ctx.Done()
// Two ways out: a signal (normal stop → exit 0, the start script stops looping)
// or a staged update / rollback (→ exit 75, the script relaunches us and the
// bootstrap above installs the new build).
code := 0
select {
case <-ctx.Done():
case <-server.RestartRequested():
code = selfupdate.ExitRestart
shutdown(agent.AbortShutdown)
}
log.Println("shutting down...")
shutdownCtx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
defer cancel()
_ = httpSrv.Shutdown(shutdownCtx)
return code
}
// shutdownContext deliberately does not derive from signalCtx. If it did, the