From 64824290912b6300ede2b2f95ff77d55dde9859b Mon Sep 17 00:00:00 2001
From: Lunny Xiao <xiaolunwen@gmail.com>
Date: Tue, 19 Nov 2024 11:44:21 -0800
Subject: [PATCH] Fix submodule parsing

(cherry picked from commit 33850a83fe4ebd23a762a7aac81614c42e303bfa)

This really is just the cherry pick of 407b6e6dfc7ee9ebb8a16c7f1a786e4c24d0516e
which is the first commit of the pull request, the one with the
change. The rest of the changes is a refactor that is unrelated to the
bug fix.

Conflicts:
	modules/git/commit_test.go
  trivial context conflict
---
 modules/git/commit.go      | 45 ++++++++++++++++++--------------------
 modules/git/commit_test.go | 30 +++++++++++++++++++++++++
 2 files changed, 51 insertions(+), 24 deletions(-)

diff --git a/modules/git/commit.go b/modules/git/commit.go
index b5ae2e0e52..0f3b749227 100644
--- a/modules/git/commit.go
+++ b/modules/git/commit.go
@@ -17,6 +17,8 @@ import (
 
 	"code.gitea.io/gitea/modules/log"
 	"code.gitea.io/gitea/modules/util"
+
+	"github.com/go-git/go-git/v5/config"
 )
 
 // Commit represents a git commit.
@@ -365,37 +367,32 @@ func (c *Commit) GetSubModules() (*ObjectCache, error) {
 		return nil, err
 	}
 
-	rd, err := entry.Blob().DataAsync()
+	content, err := entry.Blob().GetBlobContent(10 * 1024)
 	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
-			}
-		}
+	c.submoduleCache, err = parseSubmoduleContent([]byte(content))
+	if err != nil {
+		return nil, err
 	}
-	if err = scanner.Err(); err != nil {
-		return nil, fmt.Errorf("GetSubModules scan: %w", err)
+	return c.submoduleCache, nil
+}
+
+func parseSubmoduleContent(bs []byte) (*ObjectCache, error) {
+	cfg := config.NewModules()
+	if err := cfg.Unmarshal(bs); err != nil {
+		return nil, err
+	}
+	submoduleCache := newObjectCache()
+	if len(cfg.Submodules) == 0 {
+		return nil, fmt.Errorf("no submodules found")
+	}
+	for _, subModule := range cfg.Submodules {
+		submoduleCache.Set(subModule.Path, subModule.URL)
 	}
 
-	return c.submoduleCache, nil
+	return submoduleCache, nil
 }
 
 // GetSubModule get the sub module according entryname
diff --git a/modules/git/commit_test.go b/modules/git/commit_test.go
index af85bfe093..6bb7d776f5 100644
--- a/modules/git/commit_test.go
+++ b/modules/git/commit_test.go
@@ -369,3 +369,33 @@ func TestParseCommitRenames(t *testing.T) {
 		assert.Equal(t, testcase.renames, renames)
 	}
 }
+
+func Test_parseSubmoduleContent(t *testing.T) {
+	submoduleFiles := []struct {
+		fileContent  string
+		expectedPath string
+		expectedURL  string
+	}{
+		{
+			fileContent: `[submodule "jakarta-servlet"]
+url = ../../ALP-pool/jakarta-servlet
+path = jakarta-servlet`,
+			expectedPath: "jakarta-servlet",
+			expectedURL:  "../../ALP-pool/jakarta-servlet",
+		},
+		{
+			fileContent: `[submodule "jakarta-servlet"]
+path = jakarta-servlet
+url = ../../ALP-pool/jakarta-servlet`,
+			expectedPath: "jakarta-servlet",
+			expectedURL:  "../../ALP-pool/jakarta-servlet",
+		},
+	}
+	for _, kase := range submoduleFiles {
+		submodule, err := parseSubmoduleContent([]byte(kase.fileContent))
+		require.NoError(t, err)
+		v, ok := submodule.Get(kase.expectedPath)
+		assert.True(t, ok)
+		assert.Equal(t, kase.expectedURL, v)
+	}
+}