[content-service] add prestop hook to extract git status

components/content-service/pkg/git/git.go

@@ -215,6 +215,54 @@ func (c *Client) Git(ctx context.Context, subcommand string, args ...string) (er
 	return nil
 }
 
+// GitStatusFromFiles same as Status but reads git output from preexisting files that were generated by prestop hook
+func GitStatusFromFiles(ctx context.Context, loc string) (res *Status, err error) {
+	gitout, err := os.ReadFile(filepath.Join(loc, "git_status.txt"))
+	if err != nil {
+		return nil, err
+	}
+	porcelain, err := parsePorcelain(bytes.NewReader(gitout))
+	if err != nil {
+		return nil, err
+	}
+
+	unpushedCommits := make([]string, 0)
+	gitout, err = os.ReadFile(filepath.Join(loc, "git_log_1.txt"))
+	if err != nil && !strings.Contains(err.Error(), errNoCommitsYet) {
+		return nil, err
+	}
+	if gitout != nil {
+		out, err := io.ReadAll(bytes.NewReader(gitout))
+		if err != nil {
+			return nil, xerrors.Errorf("cannot determine unpushed commits: %w", err)
+		}
+		for _, l := range strings.Split(string(out), "\n") {
+			tl := strings.TrimSpace(l)
+			if tl != "" {
+				unpushedCommits = append(unpushedCommits, tl)
+			}
+		}
+	}
+	if len(unpushedCommits) == 0 {
+		unpushedCommits = nil
+	}
+
+	latestCommit := ""
+	gitout, err = os.ReadFile(filepath.Join(loc, "git_log_2.txt"))
+	if err != nil && !strings.Contains(err.Error(), errNoCommitsYet) {
+		return nil, err
+	}
+	if len(gitout) > 0 {
+		latestCommit = strings.TrimSpace(string(gitout))
+	}
+
+	return &Status{
+		porcelainStatus: *porcelain,
+		UnpushedCommits: unpushedCommits,
+		LatestCommit:    latestCommit,
+	}, nil
+}
+
 // Status runs git status
 func (c *Client) Status(ctx context.Context) (res *Status, err error) {
 	gitout, err := c.GitWithOutput(ctx, "status", "--porcelain=v2", "--branch", "-uall")

components/content-service/pkg/git/git_test.go

@@ -8,6 +8,7 @@ import (
 	"context"
 	"os"
 	"path/filepath"
+	"strings"
 	"testing"
 	"time"
 
@@ -228,6 +229,246 @@ func TestGitStatus(t *testing.T) {
 	}
 }
 
+func TestGitStatusFromFiles(t *testing.T) {
+	tests := []struct {
+		Name   string
+		Prep   func(context.Context, *Client) error
+		Result *Status
+		Error  error
+	}{
+		{
+			"no commits",
+			func(ctx context.Context, c *Client) error {
+				if err := c.Git(ctx, "init"); err != nil {
+					return err
+				}
+				return nil
+			},
+			&Status{
+				porcelainStatus: porcelainStatus{
+					BranchOID:  "(initial)",
+					BranchHead: "master",
+				},
+			},
+			nil,
+		},
+		{
+			"clean copy",
+			func(ctx context.Context, c *Client) error {
+				if err := initFromRemote(ctx, c); err != nil {
+					return err
+				}
+				return nil
+			},
+			&Status{
+				porcelainStatus: porcelainStatus{
+					BranchHead: "master",
+					BranchOID:  notEmpty,
+				},
+				LatestCommit: notEmpty,
+			},
+			nil,
+		},
+		{
+			"untracked files",
+			func(ctx context.Context, c *Client) error {
+				if err := initFromRemote(ctx, c); err != nil {
+					return err
+				}
+				if err := os.WriteFile(filepath.Join(c.Location, "another-file"), []byte{}, 0755); err != nil {
+					return err
+				}
+				return nil
+			},
+			&Status{
+				porcelainStatus: porcelainStatus{
+					BranchHead:     "master",
+					BranchOID:      notEmpty,
+					UntrackedFiles: []string{"another-file"},
+				},
+				LatestCommit: notEmpty,
+			},
+			nil,
+		},
+		{
+			"uncommitted files",
+			func(ctx context.Context, c *Client) error {
+				if err := initFromRemote(ctx, c); err != nil {
+					return err
+				}
+				if err := os.WriteFile(filepath.Join(c.Location, "first-file"), []byte("foobar"), 0755); err != nil {
+					return err
+				}
+				return nil
+			},
+			&Status{
+				porcelainStatus: porcelainStatus{
+					BranchHead:      "master",
+					BranchOID:       notEmpty,
+					UncommitedFiles: []string{"first-file"},
+				},
+				LatestCommit: notEmpty,
+			},
+			nil,
+		},
+		{
+			"unpushed commits",
+			func(ctx context.Context, c *Client) error {
+				if err := initFromRemote(ctx, c); err != nil {
+					return err
+				}
+				if err := os.WriteFile(filepath.Join(c.Location, "first-file"), []byte("foobar"), 0755); err != nil {
+					return err
+				}
+				if err := c.Git(ctx, "commit", "-a", "-m", "foo"); err != nil {
+					return err
+				}
+				return nil
+			},
+			&Status{
+				porcelainStatus: porcelainStatus{
+					BranchHead: "master",
+					BranchOID:  notEmpty,
+				},
+				UnpushedCommits: []string{notEmpty},
+				LatestCommit:    notEmpty,
+			},
+			nil,
+		},
+		{
+			"unpushed commits in new branch",
+			func(ctx context.Context, c *Client) error {
+				if err := initFromRemote(ctx, c); err != nil {
+					return err
+				}
+				if err := c.Git(ctx, "checkout", "-b", "otherbranch"); err != nil {
+					return err
+				}
+				if err := os.WriteFile(filepath.Join(c.Location, "first-file"), []byte("foobar"), 0755); err != nil {
+					return err
+				}
+				if err := c.Git(ctx, "commit", "-a", "-m", "foo"); err != nil {
+					return err
+				}
+				return nil
+			},
+			&Status{
+				porcelainStatus: porcelainStatus{
+					BranchHead: "otherbranch",
+					BranchOID:  notEmpty,
+				},
+				UnpushedCommits: []string{notEmpty},
+				LatestCommit:    notEmpty,
+			},
+			nil,
+		},
+
+		{
+			"pending in sub-dir files",
+			func(ctx context.Context, c *Client) error {
+				if err := initFromRemote(ctx, c); err != nil {
+					return err
+				}
+				if err := os.MkdirAll(filepath.Join(c.Location, "this/is/a/nested/test"), 0755); err != nil {
+					return err
+				}
+				if err := os.WriteFile(filepath.Join(c.Location, "this/is/a/nested/test/first-file"), []byte("foobar"), 0755); err != nil {
+					return err
+				}
+				return nil
+			},
+			&Status{
+				porcelainStatus: porcelainStatus{
+					BranchHead:     "master",
+					BranchOID:      notEmpty,
+					UntrackedFiles: []string{"this/is/a/nested/test/first-file"},
+				},
+				LatestCommit: notEmpty,
+			},
+			nil,
+		},
+	}
+
+	for _, test := range tests {
+		t.Run(test.Name, func(t *testing.T) {
+			ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
+			defer cancel()
+
+			client, err := newGitClient(ctx)
+			if err != nil {
+				t.Errorf("cannot prep %s: %v", test.Name, err)
+				return
+			}
+
+			err = test.Prep(ctx, client)
+			if err != nil {
+				t.Errorf("cannot prep %s: %v", test.Name, err)
+				return
+			}
+
+			gitout, err := client.GitWithOutput(ctx, "status", "--porcelain=v2", "--branch", "-uall")
+			if err != nil {
+				t.Errorf("error calling GitWithOutput: %v", err)
+				return
+			}
+			if err := os.WriteFile(filepath.Join("/tmp", "git_status.txt"), gitout, 0755); err != nil {
+				t.Errorf("error creating file: %v", err)
+				return
+			}
+
+			gitout, err = client.GitWithOutput(ctx, "log", "--pretty=%h: %s", "--branches", "--not", "--remotes")
+			if err != nil {
+				t.Errorf("error calling GitWithOutput: %v", err)
+				return
+			}
+			if err := os.WriteFile(filepath.Join("/tmp", "git_log_1.txt"), gitout, 0755); err != nil {
+				t.Errorf("error creating file: %v", err)
+				return
+			}
+
+			gitout, err = client.GitWithOutput(ctx, "log", "--pretty=%H", "-n", "1")
+			if err != nil && !strings.Contains(err.Error(), "fatal: your current branch 'master' does not have any commits yet") {
+				t.Errorf("error calling GitWithOutput: %v", err)
+				return
+			}
+			if err := os.WriteFile(filepath.Join("/tmp", "git_log_2.txt"), gitout, 0755); err != nil {
+				t.Errorf("error creating file: %v", err)
+				return
+			}
+
+			status, err := GitStatusFromFiles(ctx, "/tmp")
+			if err != test.Error {
+				t.Errorf("expected error does not match for %s: %v != %v", test.Name, err, test.Error)
+				return
+			}
+
+			if status != nil {
+				if test.Result.BranchOID == notEmpty && status.LatestCommit != "" {
+					test.Result.BranchOID = status.LatestCommit
+				}
+				if test.Result.LatestCommit == notEmpty && status.LatestCommit != "" {
+					test.Result.LatestCommit = status.LatestCommit
+				}
+				for _, c := range test.Result.UnpushedCommits {
+					if c == notEmpty {
+						if len(status.UnpushedCommits) == 0 {
+							t.Errorf("expected unpushed commits")
+						}
+
+						test.Result.UnpushedCommits = status.UnpushedCommits
+						break
+					}
+				}
+			}
+
+			if diff := cmp.Diff(test.Result, status, cmp.AllowUnexported(Status{})); diff != "" {
+				t.Errorf("unexpected status (-want +got):\n%s", diff)
+			}
+
+		})
+	}
+}
+
 func newGitClient(ctx context.Context) (*Client, error) {
 	loc, err := os.MkdirTemp("", "gittest")
 	if err != nil {

components/content-service/pkg/layer/provider.go

@@ -465,15 +465,27 @@ func contentDescriptorToLayer(cdesc []byte) (*Layer, error) {
 	)
 }
 
+var prestophookScript = `#!/bin/bash
+cd ${GITPOD_REPO_ROOT}
+git config --global --add safe.directory ${GITPOD_REPO_ROOT}
+git status --porcelain=v2 --branch -uall > /.workspace/prestophookdata/git_status.txt
+git log --pretty='%h: %s' --branches --not --remotes > /.workspace/prestophookdata/git_log_1.txt
+git log --pretty=%H -n 1 > /.workspace/prestophookdata/git_log_2.txt
+`
+
 // version of this function for persistent volume claim feature
 // we cannot use /workspace folder as when mounting /workspace folder through PVC
 // it will mask anything that was in container layer, hence we are using /.workspace instead here
 func contentDescriptorToLayerPVC(cdesc []byte) (*Layer, error) {
-	return layerFromContent(
-		fileInLayer{&tar.Header{Typeflag: tar.TypeDir, Name: "/.workspace", Uid: initializer.GitpodUID, Gid: initializer.GitpodGID, Mode: 0755}, nil},
-		fileInLayer{&tar.Header{Typeflag: tar.TypeDir, Name: "/.workspace/.gitpod", Uid: initializer.GitpodUID, Gid: initializer.GitpodGID, Mode: 0755}, nil},
-		fileInLayer{&tar.Header{Typeflag: tar.TypeReg, Name: "/.workspace/.gitpod/content.json", Uid: initializer.GitpodUID, Gid: initializer.GitpodGID, Mode: 0755, Size: int64(len(cdesc))}, cdesc},
-	)
+	layers := []fileInLayer{
+		{&tar.Header{Typeflag: tar.TypeDir, Name: "/.workspace", Uid: initializer.GitpodUID, Gid: initializer.GitpodGID, Mode: 0755}, nil},
+		{&tar.Header{Typeflag: tar.TypeDir, Name: "/.workspace/.gitpod", Uid: initializer.GitpodUID, Gid: initializer.GitpodGID, Mode: 0755}, nil},
+		{&tar.Header{Typeflag: tar.TypeReg, Name: "/.supervisor/prestophook.sh", Uid: 0, Gid: 0, Mode: 0775, Size: int64(len(prestophookScript))}, []byte(prestophookScript)},
+	}
+	if len(cdesc) > 0 {
+		layers = append(layers, fileInLayer{&tar.Header{Typeflag: tar.TypeReg, Name: "/.workspace/.gitpod/content.json", Uid: initializer.GitpodUID, Gid: initializer.GitpodGID, Mode: 0755, Size: int64(len(cdesc))}, cdesc})
+	}
+	return layerFromContent(layers...)
 }
 
 func workspaceReadyLayer(src csapi.WorkspaceInitSource) (*Layer, error) {

components/ws-daemon/pkg/content/service.go

@@ -11,6 +11,7 @@ import (
 	"fmt"
 	"math"
 	"os"
+	"os/exec"
 	"path/filepath"
 	"syscall"
 	"time"
@@ -240,6 +241,22 @@ func (s *WorkspaceService) InitWorkspace(ctx context.Context, req *api.InitWorks
 		}
 	}
 
+	if req.PersistentVolumeClaim {
+		// create a folder that is used to store data from running prestophook
+		deamonDir := fmt.Sprintf("%s-daemon", req.Id)
+		prestophookDir := filepath.Join(s.config.WorkingArea, deamonDir, "prestophookdata")
+		err = os.MkdirAll(prestophookDir, 0755)
+		if err != nil {
+			log.WithError(err).WithField("workspaceId", req.Id).Error("cannot create prestophookdata folder")
+			return nil, status.Error(codes.FailedPrecondition, fmt.Sprintf("cannot create prestophookdata: %v", err))
+		}
+		_, err = exec.CommandContext(ctx, "chown", "-R", fmt.Sprintf("%d:%d", wsinit.GitpodUID, wsinit.GitpodGID), prestophookDir).CombinedOutput()
+		if err != nil {
+			log.WithError(err).WithField("workspaceId", req.Id).Error("cannot chown prestophookdata folder")
+			return nil, status.Error(codes.FailedPrecondition, fmt.Sprintf("cannot chown prestophookdata: %v", err))
+		}
+	}
+
 	// Tell the world we're done
 	err = workspace.MarkInitDone(ctx)
 	if err != nil {

components/ws-manager/pkg/manager/create.go

@@ -543,6 +543,15 @@ func (m *Manager) createDefiniteWorkspacePod(startContext *startWorkspaceContext
 			// not needed, since it is using dedicated disk
 			pod.Spec.Containers[0].VolumeMounts[0].MountPropagation = nil
 
+			// add prestop hook to capture git status
+			pod.Spec.Containers[0].Lifecycle = &corev1.Lifecycle{
+				PreStop: &corev1.LifecycleHandler{
+					Exec: &corev1.ExecAction{
+						Command: []string{"/bin/sh", "-c", "/.supervisor/workspacekit lift /.supervisor/prestophook.sh"},
+					},
+				},
+			}
+
 			// pavel: 133332 is the Gitpod UID (33333) shifted by 99999. The shift happens inside the workspace container due to the user namespace use.
 			// We set this magical ID to make sure that gitpod user inside the workspace can write into /workspace folder mounted by PVC
 			gitpodGUID := int64(133332)