test: cover dedup hint, reparse guard, store gate and CoW contract

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-09-21 09:47:06 +00:00
parent 5f9fbb9923
commit b867f4f19c
3 changed files with 222 additions and 0 deletions
@@ -1,6 +1,7 @@
 package materialindex_test

 import (
+	"context"
 	"strings"
 	"testing"

@@ -26,6 +27,23 @@ func TestBuildTOCAndFormatRAG(t *testing.T) {
 	}
 }

+func TestStoreGatedHelpers(t *testing.T) {
+	// nil pool（内存降级模式）：store 能力必须为 false，防止无状态机时打真实 API。
+	p := materialindex.NewPipeline(nil, nil)
+	if p.HasStore() {
+		t.Fatalf("nil pool must report HasStore=false")
+	}
+	ctx := context.Background()
+	if got := p.GetParseInfos(ctx, []string{"abc"}); len(got) != 0 {
+		t.Fatalf("GetParseInfos without store must be empty, got %v", got)
+	}
+	if p.ResetForReparse(ctx, "abc") {
+		t.Fatalf("ResetForReparse without store must be false")
+	}
+	// EnqueueAsync 无 store 时必须静默跳过（不 panic、不触网）。
+	p.EnqueueAsync("doc.pdf", "abc", []byte("%PDF-1.4"), 1)
+}
+
 func TestFormatRAG_Empty(t *testing.T) {
 	out := materialindex.FormatRAGMarkdown("x.pdf", nil)
 	if !strings.Contains(out, "未命中") {
@@ -168,3 +168,126 @@ func TestGetProjectAssetRaw_RoundTrip(t *testing.T) {
 		t.Fatalf("expected non-200 for mismatched project, got %d", wrongResp.StatusCode)
 	}
 }
+
+func uploadAssetBytes(t *testing.T, url, token string, filename string, payload []byte) projectAssetResponse {
+	t.Helper()
+	var body bytes.Buffer
+	w := multipart.NewWriter(&body)
+	declared := "application/octet-stream"
+	if strings.HasSuffix(strings.ToLower(filename), ".pdf") {
+		declared = "application/pdf"
+	}
+	part, _ := w.CreatePart(map[string][]string{
+		"Content-Disposition": {`form-data; name="file"; filename="` + filename + `"`},
+		"Content-Type":        {declared},
+	})
+	_, _ = part.Write(payload)
+	_ = w.Close()
+	req, _ := http.NewRequest(http.MethodPost, url, &body)
+	req.Header.Set("Content-Type", w.FormDataContentType())
+	withBearerAuth(req, token)
+	resp, err := http.DefaultClient.Do(req)
+	if err != nil {
+		t.Fatalf("upload %s: %v", filename, err)
+	}
+	defer resp.Body.Close()
+	if resp.StatusCode != http.StatusOK {
+		raw, _ := io.ReadAll(resp.Body)
+		t.Fatalf("upload %s status=%d body=%s", filename, resp.StatusCode, raw)
+	}
+	var out projectAssetResponse
+	if err := json.NewDecoder(resp.Body).Decode(&out); err != nil {
+		t.Fatalf("decode upload: %v", err)
+	}
+	return out
+}
+
+// 相同内容第二次上传：响应标注 deduplicated，列表接口同样标注非首条引用。
+func TestUploadProjectAsset_DedupHint(t *testing.T) {
+	cfg := &config.Config{Env: "test", DataDir: t.TempDir()}
+	svc := bizdev.MustNewService(t, cfg)
+	srv := NewServer(cfg, svc)
+	ts := httptest.NewServer(srv.Routes())
+	defer ts.Close()
+	token := loginTestAdmin(t, srv)
+
+	payload := []byte("%PDF-1.4 dedup hint shared content")
+	first := uploadAssetBytes(t, ts.URL+"/api/v1/projects/7/assets", token, "a.pdf", payload)
+	if first.Deduplicated {
+		t.Fatalf("first upload must not be deduplicated")
+	}
+	second := uploadAssetBytes(t, ts.URL+"/api/v1/projects/7/assets", token, "b.pdf", payload)
+	if !second.Deduplicated {
+		t.Fatalf("second identical upload must carry deduplicated=true, got %+v", second)
+	}
+	if second.SHA256 != first.SHA256 {
+		t.Fatalf("dedup must share the same blob sha")
+	}
+
+	listReq, _ := http.NewRequest(http.MethodGet, ts.URL+"/api/v1/projects/7/assets", nil)
+	withBearerAuth(listReq, token)
+	listResp, err := http.DefaultClient.Do(listReq)
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer listResp.Body.Close()
+	var assets []projectAssetResponse
+	if err := json.NewDecoder(listResp.Body).Decode(&assets); err != nil {
+		t.Fatal(err)
+	}
+	var flagged int
+	for _, a := range assets {
+		if a.Deduplicated {
+			flagged++
+		}
+		if a.AssetID == first.AssetID && a.Deduplicated {
+			t.Fatalf("first reference must stay unflagged")
+		}
+	}
+	if flagged != 1 {
+		t.Fatalf("expected exactly one dedup-flagged row in list, got %+v", assets)
+	}
+}
+
+// 重新解析：内存模式（无 PG 状态存储）下必须 503，且不触发任何解析。
+func TestReparseProjectAsset_NoStore503(t *testing.T) {
+	cfg := &config.Config{Env: "test", DataDir: t.TempDir()}
+	svc := bizdev.MustNewService(t, cfg)
+	srv := NewServer(cfg, svc)
+	ts := httptest.NewServer(srv.Routes())
+	defer ts.Close()
+	token := loginTestAdmin(t, srv)
+
+	up := uploadAssetBytes(t, ts.URL+"/api/v1/projects/7/assets", token, "doc.pdf", []byte("%PDF-1.4 fake"))
+	req, _ := http.NewRequest(http.MethodPost,
+		ts.URL+"/api/v1/projects/7/assets/"+strconv.FormatInt(up.AssetID, 10)+"/reparse", nil)
+	withBearerAuth(req, token)
+	resp, err := http.DefaultClient.Do(req)
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer resp.Body.Close()
+	if resp.StatusCode != http.StatusServiceUnavailable {
+		t.Fatalf("reparse without store want 503 got %d", resp.StatusCode)
+	}
+}
+
+// 不存在的资产 reparse 必须 404。
+func TestReparseProjectAsset_UnknownAsset404(t *testing.T) {
+	cfg := &config.Config{Env: "test", DataDir: t.TempDir()}
+	srv := NewServer(cfg, bizdev.MustNewService(t, cfg))
+	ts := httptest.NewServer(srv.Routes())
+	defer ts.Close()
+	token := loginTestAdmin(t, srv)
+
+	req, _ := http.NewRequest(http.MethodPost, ts.URL+"/api/v1/projects/7/assets/999999/reparse", nil)
+	withBearerAuth(req, token)
+	resp, err := http.DefaultClient.Do(req)
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer resp.Body.Close()
+	if resp.StatusCode != http.StatusNotFound {
+		t.Fatalf("reparse unknown asset want 404 got %d", resp.StatusCode)
+	}
+}
@@ -201,3 +201,84 @@ func TestCAS_TimeDecayCalculation(t *testing.T) {
 		t.Errorf("expected 14-day old chunk to decay below 0.5, got %.3f", hits[1].TimeDecay)
 	}
 }
+
+// TestCAS_ImmutabilityCoWContract 固化 CoW（写时复制）契约：
+//  1. 同内容重复写入只增引用计数，不写第二份物理文件，且旧 blob 字节保持不变；
+//  2. 可编辑文档产生新内容时按新 hash 落新 blob（写入即复制），旧版本 blob
+//     不被覆写，历史引用仍可读出原始字节。
+//
+// 该测试是防回归守卫：未来任何「就地改写 blob 文件」的实现都会在此失败。
+func TestCAS_ImmutabilityCoWContract(t *testing.T) {
+	tempDir, err := os.MkdirTemp("", "cas_cow_test_*")
+	if err != nil {
+		t.Fatalf("mkdir temp failed: %v", err)
+	}
+	defer os.RemoveAll(tempDir)
+
+	cas, err := NewCAS(tempDir, nil, nil)
+	if err != nil {
+		t.Fatalf("new cas failed: %v", err)
+	}
+	ctx := context.Background()
+
+	v1 := []byte("V1：产品需求说明书初始版本内容")
+	refA, err := cas.StoreFile(ctx, 900, 1, "prd.docx", "material", v1, "application/octet-stream")
+	if err != nil {
+		t.Fatalf("store v1 failed: %v", err)
+	}
+	blobPath := filepath.Join(tempDir, "blobs", refA.BlobSHA256[:2], refA.BlobSHA256[2:4], refA.BlobSHA256+".bin")
+	v1Bytes, err := os.ReadFile(blobPath)
+	if err != nil {
+		t.Fatalf("v1 blob file missing: %v", err)
+	}
+
+	// 1. 相同内容再传：命中去重，物理文件数量与字节内容不变。
+	refB, err := cas.StoreFile(ctx, 900, 2, "prd_copy.docx", "material", v1, "application/octet-stream")
+	if err != nil {
+		t.Fatalf("store duplicate failed: %v", err)
+	}
+	if !refB.IsDeduplicated || refB.BlobSHA256 != refA.BlobSHA256 {
+		t.Fatalf("expected dedup hit on identical content")
+	}
+	afterDup, err := os.ReadFile(blobPath)
+	if err != nil || string(afterDup) != string(v1Bytes) {
+		t.Fatalf("duplicate upload mutated existing blob bytes (CoW violation)")
+	}
+
+	// 2. 编辑后写入（内容变化）：必须产生新 blob；旧 blob 不被覆写。
+	v2 := []byte("V2：产品需求说明书——评审后修改条款")
+	refC, err := cas.StoreFile(ctx, 900, 1, "prd.docx", "material", v2, "application/octet-stream")
+	if err != nil {
+		t.Fatalf("store v2 failed: %v", err)
+	}
+	if refC.BlobSHA256 == refA.BlobSHA256 {
+		t.Fatalf("modified content must get a new blob hash")
+	}
+	if refC.IsDeduplicated {
+		t.Fatalf("modified content must not be marked dedup")
+	}
+	stillV1, err := os.ReadFile(blobPath)
+	if err != nil || string(stillV1) != string(v1Bytes) {
+		t.Fatalf("write of new version overwrote old blob (CoW violation)")
+	}
+	gotV1, _, err := cas.ReadRefBytes(ctx, refA.ID)
+	if err != nil || string(gotV1) != string(v1) {
+		t.Fatalf("historical ref must still read original V1 bytes, got %q err=%v", gotV1, err)
+	}
+	gotV2, _, err := cas.ReadRefBytes(ctx, refC.ID)
+	if err != nil || string(gotV2) != string(v2) {
+		t.Fatalf("new ref must read V2 bytes, got %q err=%v", gotV2, err)
+	}
+
+	// 3. 物理盘上恰好两份 blob（V1 一份共享 + V2 一份）。
+	count := 0
+	_ = filepath.Walk(filepath.Join(tempDir, "blobs"), func(path string, info os.FileInfo, err error) error {
+		if err == nil && !info.IsDir() && filepath.Ext(path) == ".bin" {
+			count++
+		}
+		return nil
+	})
+	if count != 2 {
+		t.Fatalf("expected exactly 2 physical blobs (CoW), got %d", count)
+	}
+}