Fix submodule parsing

modules/git/commit.go

@@ -9,7 +9,6 @@ import (
 	"bytes"
 	"context"
 	"errors"
-	"fmt"
 	"io"
 	"os/exec"
 	"strconv"
@@ -375,32 +374,13 @@ func (c *Commit) GetSubModules() (*ObjectCache, error) {
 	if err != nil {
 		return nil, err
 	}
-
 	defer rd.Close()
-	scanner := bufio.NewScanner(rd)
-	c.submoduleCache = newObjectCache()
-	var ismodule bool
-	var path string
-	for scanner.Scan() {
-		if strings.HasPrefix(scanner.Text(), "[submodule") {
-			ismodule = true
-			continue
-		}
-		if ismodule {
-			fields := strings.Split(scanner.Text(), "=")
-			k := strings.TrimSpace(fields[0])
-			if k == "path" {
-				path = strings.TrimSpace(fields[1])
-			} else if k == "url" {
-				c.submoduleCache.Set(path, &SubModule{path, strings.TrimSpace(fields[1])})
-				ismodule = false
-			}
-		}
-	}
-	if err = scanner.Err(); err != nil {
-		return nil, fmt.Errorf("GetSubModules scan: %w", err)
-	}
 
+	r := io.LimitReader(rd, 100*1024) // limit to 100KB
+	c.submoduleCache, err = parseSubmoduleContent(r)
+	if err != nil {
+		return nil, err
+	}
 	return c.submoduleCache, nil
 }
 

modules/git/commit_test.go

@@ -362,3 +362,48 @@ func Test_GetCommitBranchStart(t *testing.T) {
 	assert.NotEmpty(t, startCommitID)
 	assert.EqualValues(t, "9c9aef8dd84e02bc7ec12641deb4c930a7c30185", startCommitID)
 }
+
+func Test_parseSubmoduleContent(t *testing.T) {
+	submoduleFiles := []struct {
+		fileContent    string
+		expectedPath   string
+		expectedURL    string
+		expectedBranch string
+	}{
+		{
+			fileContent: `[submodule "jakarta-servlet"]
+url = ../../ALP-pool/jakarta-servlet
+path = jakarta-servlet`,
+			expectedPath:   "jakarta-servlet",
+			expectedURL:    "../../ALP-pool/jakarta-servlet",
+			expectedBranch: "",
+		},
+		{
+			fileContent: `[submodule "jakarta-servlet"]
+path = jakarta-servlet
+url = ../../ALP-pool/jakarta-servlet`,
+			expectedPath:   "jakarta-servlet",
+			expectedURL:    "../../ALP-pool/jakarta-servlet",
+			expectedBranch: "",
+		},
+		{
+			fileContent: `[submodule "jakarta-servlet"]
+path = jakarta-servlet
+url = ../../ALP-pool/jakarta-servlet
+branch = stable`,
+			expectedPath:   "jakarta-servlet",
+			expectedURL:    "../../ALP-pool/jakarta-servlet",
+			expectedBranch: "stable",
+		},
+	}
+	for _, kase := range submoduleFiles {
+		submodule, err := parseSubmoduleContent(strings.NewReader(kase.fileContent))
+		assert.NoError(t, err)
+		v, ok := submodule.Get(kase.expectedPath)
+		assert.True(t, ok)
+		subModule := v.(*SubModule)
+		assert.Equal(t, kase.expectedPath, subModule.Path)
+		assert.Equal(t, kase.expectedURL, subModule.URL)
+		assert.Equal(t, kase.expectedBranch, subModule.Branch)
+	}
+}

modules/git/submodule.go

@@ -5,7 +5,9 @@
 package git
 
 import (
+	"bufio"
 	"fmt"
+	"io"
 	"net"
 	"net/url"
 	"path"
@@ -17,8 +19,74 @@ var scpSyntax = regexp.MustCompile(`^([a-zA-Z0-9_]+@)?([a-zA-Z0-9._-]+):(.*)$`)
 
 // SubModule submodule is a reference on git repository
 type SubModule struct {
-	Name string
-	URL  string
+	Path   string
+	URL    string
+	Branch string
+}
+
+// parseSubmoduleContent this is not a complete parse for gitmodules file, it only
+// parses the url and path of submodules
+func parseSubmoduleContent(r io.Reader) (*ObjectCache, error) {
+	var path, url, branch string
+	var state int // 0: find section, 1: find path and url
+	subModules := newObjectCache()
+	scanner := bufio.NewScanner(r)
+	for scanner.Scan() {
+		line := strings.TrimSpace(scanner.Text())
+
+		// Skip empty lines and comments
+		if line == "" || strings.HasPrefix(line, "#") || strings.HasPrefix(line, ";") {
+			continue
+		}
+
+		// Section header [section]
+		if strings.HasPrefix(line, "[submodule") && strings.HasSuffix(line, "]") {
+			if path != "" && url != "" {
+				subModules.Set(path, &SubModule{
+					Path:   path,
+					URL:    url,
+					Branch: branch,
+				})
+			}
+			state = 1
+			path = ""
+			url = ""
+			branch = ""
+			continue
+		}
+
+		if state != 1 {
+			continue
+		}
+
+		parts := strings.SplitN(line, "=", 2)
+		if len(parts) != 2 {
+			continue
+		}
+		key := strings.TrimSpace(parts[0])
+		value := strings.TrimSpace(parts[1])
+		switch key {
+		case "path":
+			path = value
+		case "url":
+			url = value
+		case "branch":
+			branch = value
+		}
+	}
+
+	if err := scanner.Err(); err != nil {
+		return nil, fmt.Errorf("error reading file: %w", err)
+	}
+	if path != "" && url != "" {
+		subModules.Set(path, &SubModule{
+			Path:   path,
+			URL:    url,
+			Branch: branch,
+		})
+	}
+
+	return subModules, nil
 }
 
 // SubModuleFile represents a file with submodule type.