Refactor "Content" for file uploading (#25851)

Before: the concept "Content string" is used everywhere. It has some
problems:

1. Sometimes it means "base64 encoded content", sometimes it means "raw
binary content"
2. It doesn't work with large files, eg: uploading a 1G LFS file would
make Gitea process OOM

This PR does the refactoring: use "ContentReader" / "ContentBase64"
instead of "Content"

This PR is not breaking because the key in API JSON is still "content":
`` ContentBase64 string `json:"content"` ``

Author: wxiaoguang

modules/structs/repo_file.go

@@ -26,7 +26,7 @@ type CreateFileOptions struct {
 	FileOptions
 	// content must be base64 encoded
 	// required: true
-	Content string `json:"content"`
+	ContentBase64 string `json:"content"`
 }
 
 // Branch returns branch name
@@ -54,7 +54,7 @@ type UpdateFileOptions struct {
 	DeleteFileOptions
 	// content must be base64 encoded
 	// required: true
-	Content string `json:"content"`
+	ContentBase64 string `json:"content"`
 	// from_path (optional) is the path of the original file which will be moved/renamed to the path in the URL
 	FromPath string `json:"from_path" binding:"MaxSize(500)"`
 }
@@ -74,7 +74,7 @@ type ChangeFileOperation struct {
 	// required: true
 	Path string `json:"path" binding:"Required;MaxSize(500)"`
 	// new or updated file content, must be base64 encoded
-	Content string `json:"content"`
+	ContentBase64 string `json:"content"`
 	// sha is the SHA for the file that already exists, required for update or delete
 	SHA string `json:"sha"`
 	// old path of the file to move

routers/api/v1/repo/file.go

@@ -408,6 +408,14 @@ func canReadFiles(r *context.Repository) bool {
 	return r.Permission.CanRead(unit.TypeCode)
 }
 
+func base64Reader(s string) (io.Reader, error) {
+	b, err := base64.StdEncoding.DecodeString(s)
+	if err != nil {
+		return nil, err
+	}
+	return bytes.NewReader(b), nil
+}
+
 // ChangeFiles handles API call for modifying multiple files
 func ChangeFiles(ctx *context.APIContext) {
 	// swagger:operation POST /repos/{owner}/{repo}/contents repository repoChangeFiles
@@ -449,14 +457,19 @@ func ChangeFiles(ctx *context.APIContext) {
 		apiOpts.BranchName = ctx.Repo.Repository.DefaultBranch
 	}
 
-	files := []*files_service.ChangeRepoFile{}
+	var files []*files_service.ChangeRepoFile
 	for _, file := range apiOpts.Files {
+		contentReader, err := base64Reader(file.ContentBase64)
+		if err != nil {
+			ctx.Error(http.StatusUnprocessableEntity, "Invalid base64 content", err)
+			return
+		}
 		changeRepoFile := &files_service.ChangeRepoFile{
-			Operation:    file.Operation,
-			TreePath:     file.Path,
-			FromTreePath: file.FromPath,
-			Content:      file.Content,
-			SHA:          file.SHA,
+			Operation:     file.Operation,
+			TreePath:      file.Path,
+			FromTreePath:  file.FromPath,
+			ContentReader: contentReader,
+			SHA:           file.SHA,
 		}
 		files = append(files, changeRepoFile)
 	}
@@ -544,12 +557,18 @@ func CreateFile(ctx *context.APIContext) {
 		apiOpts.BranchName = ctx.Repo.Repository.DefaultBranch
 	}
 
+	contentReader, err := base64Reader(apiOpts.ContentBase64)
+	if err != nil {
+		ctx.Error(http.StatusUnprocessableEntity, "Invalid base64 content", err)
+		return
+	}
+
 	opts := &files_service.ChangeRepoFilesOptions{
 		Files: []*files_service.ChangeRepoFile{
 			{
-				Operation: "create",
-				TreePath:  ctx.Params("*"),
-				Content:   apiOpts.Content,
+				Operation:     "create",
+				TreePath:      ctx.Params("*"),
+				ContentReader: contentReader,
 			},
 		},
 		Message:   apiOpts.Message,
@@ -636,14 +655,20 @@ func UpdateFile(ctx *context.APIContext) {
 		apiOpts.BranchName = ctx.Repo.Repository.DefaultBranch
 	}
 
+	contentReader, err := base64Reader(apiOpts.ContentBase64)
+	if err != nil {
+		ctx.Error(http.StatusUnprocessableEntity, "Invalid base64 content", err)
+		return
+	}
+
 	opts := &files_service.ChangeRepoFilesOptions{
 		Files: []*files_service.ChangeRepoFile{
 			{
-				Operation:    "update",
-				Content:      apiOpts.Content,
-				SHA:          apiOpts.SHA,
-				FromTreePath: apiOpts.FromPath,
-				TreePath:     ctx.Params("*"),
+				Operation:     "update",
+				ContentReader: contentReader,
+				SHA:           apiOpts.SHA,
+				FromTreePath:  apiOpts.FromPath,
+				TreePath:      ctx.Params("*"),
 			},
 		},
 		Message:   apiOpts.Message,
@@ -709,14 +734,6 @@ func createOrUpdateFiles(ctx *context.APIContext, opts *files_service.ChangeRepo
 		}
 	}
 
-	for _, file := range opts.Files {
-		content, err := base64.StdEncoding.DecodeString(file.Content)
-		if err != nil {
-			return nil, err
-		}
-		file.Content = string(content)
-	}
-
 	return files_service.ChangeRepoFiles(ctx, ctx.Repo.Repository, ctx.Doer, opts)
 }
 

routers/web/repo/editor.go

@@ -284,10 +284,10 @@ func editFilePost(ctx *context.Context, form forms.EditRepoFileForm, isNewFile b
 		Message:      message,
 		Files: []*files_service.ChangeRepoFile{
 			{
-				Operation:    operation,
-				FromTreePath: ctx.Repo.TreePath,
-				TreePath:     form.TreePath,
-				Content:      strings.ReplaceAll(form.Content, "\r", ""),
+				Operation:     operation,
+				FromTreePath:  ctx.Repo.TreePath,
+				TreePath:      form.TreePath,
+				ContentReader: strings.NewReader(strings.ReplaceAll(form.Content, "\r", "")),
 			},
 		},
 		Signoff: form.Signoff,

services/repository/files/update.go

@@ -6,6 +6,7 @@ package files
 import (
 	"context"
 	"fmt"
+	"io"
 	"path"
 	"strings"
 	"time"
@@ -35,12 +36,12 @@ type CommitDateOptions struct {
 }
 
 type ChangeRepoFile struct {
-	Operation    string
-	TreePath     string
-	FromTreePath string
-	Content      string
-	SHA          string
-	Options      *RepoFileOptions
+	Operation     string
+	TreePath      string
+	FromTreePath  string
+	ContentReader io.Reader
+	SHA           string
+	Options       *RepoFileOptions
 }
 
 // ChangeRepoFilesOptions holds the repository files update options
@@ -387,7 +388,7 @@ func CreateOrUpdateFile(ctx context.Context, t *TemporaryUploadRepository, file
 		}
 	}
 
-	treeObjectContent := file.Content
+	treeObjectContentReader := file.ContentReader
 	var lfsMetaObject *git_model.LFSMetaObject
 	if setting.LFS.StartServer && hasOldBranch {
 		// Check there is no way this can return multiple infos
@@ -402,17 +403,17 @@ func CreateOrUpdateFile(ctx context.Context, t *TemporaryUploadRepository, file
 
 		if filename2attribute2info[file.Options.treePath] != nil && filename2attribute2info[file.Options.treePath]["filter"] == "lfs" {
 			// OK so we are supposed to LFS this data!
-			pointer, err := lfs.GeneratePointer(strings.NewReader(file.Content))
+			pointer, err := lfs.GeneratePointer(treeObjectContentReader)
 			if err != nil {
 				return err
 			}
 			lfsMetaObject = &git_model.LFSMetaObject{Pointer: pointer, RepositoryID: repoID}
-			treeObjectContent = pointer.StringContent()
+			treeObjectContentReader = strings.NewReader(pointer.StringContent())
 		}
 	}
 
 	// Add the object to the database
-	objectHash, err := t.HashObject(strings.NewReader(treeObjectContent))
+	objectHash, err := t.HashObject(treeObjectContentReader)
 	if err != nil {
 		return err
 	}
@@ -439,7 +440,7 @@ func CreateOrUpdateFile(ctx context.Context, t *TemporaryUploadRepository, file
 			return err
 		}
 		if !exist {
-			if err := contentStore.Put(lfsMetaObject.Pointer, strings.NewReader(file.Content)); err != nil {
+			if err := contentStore.Put(lfsMetaObject.Pointer, file.ContentReader); err != nil {
 				if _, err2 := git_model.RemoveLFSMetaObjectByOid(ctx, repoID, lfsMetaObject.Oid); err2 != nil {
 					return fmt.Errorf("unable to remove failed inserted LFS object %s: %v (Prev Error: %w)", lfsMetaObject.Oid, err2, err)
 				}

templates/swagger/v1_json.tmpl

@@ -5,6 +5,7 @@ package integration
 
 import (
 	"net/url"
+	"strings"
 	"testing"
 	"time"
 
@@ -64,9 +65,9 @@ func TestPullRequestTargetEvent(t *testing.T) {
 		addWorkflowToBaseResp, err := files_service.ChangeRepoFiles(git.DefaultContext, baseRepo, user2, &files_service.ChangeRepoFilesOptions{
 			Files: []*files_service.ChangeRepoFile{
 				{
-					Operation: "create",
-					TreePath:  ".gitea/workflows/pr.yml",
-					Content:   "name: test\non: pull_request_target\njobs:\n  test:\n    runs-on: ubuntu-latest\n    steps:\n      - run: echo helloworld\n",
+					Operation:     "create",
+					TreePath:      ".gitea/workflows/pr.yml",
+					ContentReader: strings.NewReader("name: test\non: pull_request_target\njobs:\n  test:\n    runs-on: ubuntu-latest\n    steps:\n      - run: echo helloworld\n"),
 				},
 			},
 			Message:   "add workflow",
@@ -92,9 +93,9 @@ func TestPullRequestTargetEvent(t *testing.T) {
 		addFileToForkedResp, err := files_service.ChangeRepoFiles(git.DefaultContext, forkedRepo, user3, &files_service.ChangeRepoFilesOptions{
 			Files: []*files_service.ChangeRepoFile{
 				{
-					Operation: "create",
-					TreePath:  "file_1.txt",
-					Content:   "file1",
+					Operation:     "create",
+					TreePath:      "file_1.txt",
+					ContentReader: strings.NewReader("file1"),
 				},
 			},
 			Message:   "add file1",

tests/integration/actions_trigger_test.go

@@ -46,7 +46,7 @@ func getCreateFileOptions() api.CreateFileOptions {
 				Committer: time.Unix(978307190, 0),
 			},
 		},
-		Content: contentEncoded,
+		ContentBase64: contentEncoded,
 	}
 }
 

tests/integration/api_repo_file_create_test.go

@@ -4,6 +4,8 @@
 package integration
 
 import (
+	"strings"
+
 	repo_model "code.gitea.io/gitea/models/repo"
 	user_model "code.gitea.io/gitea/models/user"
 	"code.gitea.io/gitea/modules/git"
@@ -15,9 +17,9 @@ func createFileInBranch(user *user_model.User, repo *repo_model.Repository, tree
 	opts := &files_service.ChangeRepoFilesOptions{
 		Files: []*files_service.ChangeRepoFile{
 			{
-				Operation: "create",
-				TreePath:  treePath,
-				Content:   content,
+				Operation:     "create",
+				TreePath:      treePath,
+				ContentReader: strings.NewReader(content),
 			},
 		},
 		OldBranch: branchName,

tests/integration/api_repo_file_helpers.go

@@ -44,7 +44,7 @@ func getUpdateFileOptions() *api.UpdateFileOptions {
 			},
 			SHA: "103ff9234cefeee5ec5361d22b49fbb04d385885",
 		},
-		Content: contentEncoded,
+		ContentBase64: contentEncoded,
 	}
 }
 

tests/integration/api_repo_file_update_test.go

@@ -44,13 +44,13 @@ func getChangeFilesOptions() *api.ChangeFilesOptions {
 		},
 		Files: []*api.ChangeFileOperation{
 			{
-				Operation: "create",
-				Content:   newContentEncoded,
+				Operation:     "create",
+				ContentBase64: newContentEncoded,
 			},
 			{
-				Operation: "update",
-				Content:   updateContentEncoded,
-				SHA:       "103ff9234cefeee5ec5361d22b49fbb04d385885",
+				Operation:     "update",
+				ContentBase64: updateContentEncoded,
+				SHA:           "103ff9234cefeee5ec5361d22b49fbb04d385885",
 			},
 			{
 				Operation: "delete",