rewrite parse submodule

Author: lunny

modules/git/commit.go

@@ -9,16 +9,13 @@ import (
 	"bytes"
 	"context"
 	"errors"
-	"fmt"
 	"io"
 	"os/exec"
 	"strconv"
 	"strings"
 
 	"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.
@@ -373,34 +370,20 @@ func (c *Commit) GetSubModules() (*ObjectCache, error) {
 		return nil, err
 	}
 
-	content, err := entry.Blob().GetBlobContent(10 * 1024)
+	rd, err := entry.Blob().DataAsync()
 	if err != nil {
 		return nil, err
 	}
+	defer rd.Close()
 
-	c.submoduleCache, err = parseSubmoduleContent([]byte(content))
+	r := io.LimitReader(rd, 100*1024) // limit to 100KB
+	c.submoduleCache, err = parseSubmoduleContent(r)
 	if err != nil {
 		return nil, 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 submoduleCache, nil
-}
-
 // GetSubModule get the sub module according entryname
 func (c *Commit) GetSubModule(entryname string) (*SubModule, error) {
 	modules, err := c.GetSubModules()

modules/git/commit_test.go

@@ -365,30 +365,45 @@ func Test_GetCommitBranchStart(t *testing.T) {
 
 func Test_parseSubmoduleContent(t *testing.T) {
 	submoduleFiles := []struct {
-		fileContent  string
-		expectedPath string
-		expectedURL  string
+		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",
+			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",
+			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([]byte(kase.fileContent))
+		submodule, err := parseSubmoduleContent(strings.NewReader(kase.fileContent))
 		assert.NoError(t, err)
 		v, ok := submodule.Get(kase.expectedPath)
 		assert.True(t, ok)
-		assert.Equal(t, kase.expectedURL, v)
+		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 +19,72 @@ 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, "]") {
+			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.