Files
AskaEth b1f9b11d6e
CI / Backend (build / vet / test / cross-compile) (push) Has been cancelled
CI / Frontend (tsc + vite build) (push) Has been cancelled
test(flake): 4 条负载敏感用例确定性化(只改测试,生产代码零改动)
- #1 TestSpawnTimeoutFromToolDefault:墙钟竞态(Spawn→Wait 的调度抖动 vs 120ms 工具时限)⇒ 用「托底子代理占住唯一槽位 + gate 放行」把 running 变成由门控保证;并新增 DeadlineExceeded 断言。

- #2 TestParallelLimit:真实成因**不是**「同进程迭代状态泄漏」(任务书推测错了)—— 是 transitionRunning 与 runner 内 active++ 之间的**抢占窗口**(复现:-count=200 六次红、三进程并行 1/9)⇒ 改等 runner 内部 activeCount()>=2 这一**稳定态**,并新增「整轮排空峰值==2」。

- #3 TestBackthink_List_NewestFirst:真实成因**不是**时间戳并列(SQL 是 ORDER BY id DESC)—— 是**后台心跳 writer 在两次 List 之间插入新行**(决定性证据:心跳周期缩到 1ms ⇒ 21/200 红,且第二次读到的 id 更大)⇒ 不再 Start() 心跳,改同步 think 写 3 条。

- #4 TestClient_OnConnectErrorTriggersRedial:固定 60ms 睡眠 vs 每周期耗时不定的墙钟竞态(改动前基线 5/9 红)⇒ 改为等第 2 次拨号事件;顺带改正与内容不符的旧注释。

- 证据:逐条 -count=20 绿;整树连跑 3 次 59 ok/0 FAIL;**三进程并行整树 9/9 绿(改动前基线 5/9 红)**;GOMAXPROCS=1 ×4 进程加压;-count=200/500/500/300;-race 三包绿。

- 主会话独立复验:隔离 worktree build/vet 绿、四条 -count=20 全绿、两进程并行整树双 59 ok/0 FAIL;**变异独立复现**(TestParallelLimit 的 MaxParallel 2→3 ⇒ max concurrent runs = 3, want 2 红)。

- 顺带发现第 5 条既有 flake:internal/runtime/exec/TestTerminal_Timeout_Killed(改动前基线 2/9 红);以及 audit/00-总览.md:321/372 对 #2/#3 成因的记载有误(DevDocs 由主会话更正)。
2026-10-04 18:51:25 +08:00

169 lines
4.7 KiB
Go

package wsutil
import (
"context"
"io"
"net/http"
"net/http/httptest"
"strings"
"sync"
"sync/atomic"
"testing"
"time"
"github.com/gorilla/websocket"
"git.yeij.top/AskaEth/Cyrene_v2/internal/logger"
)
var wsu = websocket.Upgrader{}
// TestClient_DialReadReconnect exercises the connect hook, the read loop, and the
// reconnect path: the server accepts a connection, sends one frame, drops it;
// the client re-reads a second frame on the next connection, then the test cancels
// the context and the client exits cleanly.
func TestClient_DialReadReconnect(t *testing.T) {
var conns atomic.Int64
var mu sync.Mutex
var got []string
received := make(chan string, 2)
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
conn, err := wsu.Upgrade(w, r, nil)
if err != nil {
t.Errorf("upgrade: %v", err)
return
}
defer conn.Close()
n := conns.Add(1)
msg := "msg-" + string(rune('0'+n))
_ = conn.WriteMessage(websocket.TextMessage, []byte(msg))
if n < 2 {
return // close the first connection to trigger a reconnect
}
// Hold the second connection open until the client cancels the context.
for {
if _, _, err := conn.ReadMessage(); err != nil {
return
}
}
}))
defer srv.Close()
url := "ws://" + strings.TrimPrefix(srv.URL, "http://") + "/testws"
client := New(
func(context.Context) (string, error) { return url, nil },
Handler{
OnConnect: func(ctx context.Context, c *Conn) error { return nil },
OnMessage: func(ctx context.Context, p []byte) error {
mu.Lock()
got = append(got, string(p))
mu.Unlock()
received <- string(p)
return nil
},
},
WithBackoff(NewBackoff(10*time.Millisecond, 50*time.Millisecond)),
WithLogger(logger.New(logger.Options{Out: io.Discard})),
)
ctx, cancel := context.WithCancel(context.Background())
done := make(chan error, 1)
go func() { done <- client.Run(ctx) }()
for i := 0; i < 2; i++ {
select {
case m := <-received:
if m != "msg-" + string(rune('1'+i)) {
t.Errorf("message[%d] = %q", i, m)
}
case <-time.After(3 * time.Second):
t.Fatalf("timed out waiting for message %d", i+1)
}
}
cancel()
select {
case err := <-done:
if err == nil {
t.Error("Run should return a non-nil error on cancel")
}
case <-time.After(3 * time.Second):
t.Fatal("Run did not exit after cancel")
}
mu.Lock()
defer mu.Unlock()
if len(got) != 2 {
t.Fatalf("received %d messages, want 2", len(got))
}
if conns.Load() < 2 {
t.Errorf("connection count = %d, want >= 2 (reconnect)", conns.Load())
}
}
// TestClient_OnConnectErrorTriggersRedial 断言连接钩子失败会触发重拨:等到第二次
// 拨号事件发生后才取消上下文,而不是先睡固定的 60ms 再看拨了几次——后者在并行
// 负载下可能只够拨一次,等于把断言交给机器负载决定。
func TestClient_OnConnectErrorTriggersRedial(t *testing.T) {
const (
// redialTarget 是断言「发生过重拨」所需的拨号次数:1 次首拨 + 1 次重拨。
redialTarget = 2
// dialEvents 是拨号事件通道的容量,保证 Run 在测试取消之前不会阻塞在投递上。
dialEvents = 8
)
var dials atomic.Int64
// 每次拨号投递一个事件;测试等满 redialTarget 次再取消,因此断言的是
// 「重拨确实发生了」这件事本身,与墙钟无关。
dialed := make(chan struct{}, dialEvents)
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
conn, err := wsu.Upgrade(w, r, nil)
if err != nil {
return
}
defer conn.Close()
for {
if _, _, err := conn.ReadMessage(); err != nil {
return
}
}
}))
defer srv.Close()
url := "ws://" + strings.TrimPrefix(srv.URL, "http://") + "/testws"
// OnConnect 恒错,因此每次会话都被客户端拒绝、拨号反复发生。
client := New(
func(context.Context) (string, error) {
dials.Add(1)
dialed <- struct{}{}
return url, nil
},
Handler{
OnConnect: func(ctx context.Context, c *Conn) error { return context.DeadlineExceeded },
},
WithBackoff(NewBackoff(5*time.Millisecond, 20*time.Millisecond)),
WithLogger(logger.New(logger.Options{Out: io.Discard})),
)
ctx, cancel := context.WithCancel(context.Background())
defer cancel()
done := make(chan error, 1)
go func() { done <- client.Run(ctx) }()
for i := 0; i < redialTarget; i++ {
select {
case <-dialed:
case <-time.After(3 * time.Second):
t.Fatalf("timed out waiting for dial %d", i+1)
}
}
cancel()
select {
case <-done:
case <-time.After(2 * time.Second):
t.Fatal("Run did not exit")
}
if dials.Load() < redialTarget {
t.Errorf("dials = %d, want >= %d (reconnect on failed connect hook)", dials.Load(), redialTarget)
}
}