localrepo: Allow Size to pass exclude flag

There are times when we want to exclude certain refs from being included
in the size calculation for a repository. Add the ability to do so.

Changelog: added

2 files changed, 81 insertions, 8 deletions

diff --git a/internal/git/localrepo/repo.go b/internal/git/localrepo/repo.go
index 7d7f3e409..97525a1fc 100644
--- a/internal/git/localrepo/repo.go
+++ b/internal/git/localrepo/repo.go
@@ -94,18 +94,55 @@ func errorWithStderr(err error, stderr []byte) error {
 	return fmt.Errorf("%w, stderr: %q", err, stderr)
 }
 
+// repoSizeConfig can be used to pass in different options to
+// git rev-list in determining the size of a repository.
+type repoSizeConfig struct {
+	// Excludes is a list of ref glob patterns to exclude from the size
+	// calculation.
+	Excludes []string
+}
+
+// RepoSizeOption is an option which can be passed to Size
+type RepoSizeOption func(*repoSizeConfig)
+
+// WithExcludes is an option for Size that excludes certain refs from the size
+// calculation. The format must be a glob pattern.
+// see https://git-scm.com/docs/git-rev-list#Documentation/git-rev-list.txt---excludeltglob-patterngt
+func WithExcludes(excludes ...string) RepoSizeOption {
+	return func(cfg *repoSizeConfig) {
+		cfg.Excludes = excludes
+	}
+}
+
 // Size calculates the size of all reachable objects in bytes
-func (repo *Repo) Size(ctx context.Context) (int64, error) {
+func (repo *Repo) Size(ctx context.Context, opts ...RepoSizeOption) (int64, error) {
 	var stdout bytes.Buffer
+
+	var cfg repoSizeConfig
+
+	for _, opt := range opts {
+		opt(&cfg)
+	}
+
+	var options []git.Option
+	for _, exclude := range cfg.Excludes {
+		options = append(
+			options,
+			git.Flag{Name: fmt.Sprintf("--exclude=%s", exclude)},
+		)
+	}
+
+	options = append(options,
+		git.Flag{Name: "--all"},
+		git.Flag{Name: "--objects"},
+		git.Flag{Name: "--use-bitmap-index"},
+		git.Flag{Name: "--disk-usage"},
+	)
+
 	if err := repo.ExecAndWait(ctx,
 		git.SubCmd{
-			Name: "rev-list",
-			Flags: []git.Option{
-				git.Flag{Name: "--all"},
-				git.Flag{Name: "--objects"},
-				git.Flag{Name: "--use-bitmap-index"},
-				git.Flag{Name: "--disk-usage"},
-			},
+			Name:  "rev-list",
+			Flags: options,
 		},
 		git.WithStdout(&stdout),
 	); err != nil {
diff --git a/internal/git/localrepo/repo_test.go b/internal/git/localrepo/repo_test.go
index c305519c8..c9f620731 100644
--- a/internal/git/localrepo/repo_test.go
+++ b/internal/git/localrepo/repo_test.go
@@ -171,3 +171,39 @@ func TestSize(t *testing.T) {
 		})
 	}
 }
+
+func TestSize_excludes(t *testing.T) {
+	cfg := testcfg.Build(t)
+	gitCmdFactory := gittest.NewCommandFactory(t, cfg)
+	catfileCache := catfile.NewCache(cfg)
+	t.Cleanup(catfileCache.Stop)
+
+	pbRepo, repoPath := gittest.CloneRepo(t, cfg, cfg.Storages[0])
+	blob := bytes.Repeat([]byte("a"), 1000)
+	blobOID := gittest.WriteBlob(t, cfg, repoPath, blob)
+	treeOID := gittest.WriteTree(t, cfg, repoPath, []gittest.TreeEntry{
+		{
+			OID:  blobOID,
+			Mode: "100644",
+			Path: "1kbblob",
+		},
+	})
+	commitOID := gittest.WriteCommit(t, cfg, repoPath, gittest.WithTree(treeOID))
+
+	repo := New(config.NewLocator(cfg), gitCmdFactory, catfileCache, pbRepo)
+
+	ctx := testhelper.Context(t)
+	sizeBeforeKeepAround, err := repo.Size(ctx)
+	require.NoError(t, err)
+
+	gittest.WriteRef(t, cfg, repoPath, git.ReferenceName("refs/keep-around/keep1"), commitOID)
+
+	sizeWithKeepAround, err := repo.Size(ctx)
+	require.NoError(t, err)
+	assert.Less(t, sizeBeforeKeepAround, sizeWithKeepAround)
+
+	sizeWithoutKeepAround, err := repo.Size(ctx, WithExcludes("refs/keep-around/*"))
+	require.NoError(t, err)
+
+	assert.Equal(t, sizeBeforeKeepAround, sizeWithoutKeepAround)
+}