@@ -0,0 +1,193 @@ |
| 1 | +// SPDX-License-Identifier: AGPL-3.0-or-later |
| 2 | + |
| 3 | +// Package git wraps the git plumbing commands we need to build the |
| 4 | +// optional initial commit on a freshly-init'd bare repo. Plumbing-only, |
| 5 | +// no working tree: hash-object → update-index (with GIT_INDEX_FILE) → |
| 6 | +// write-tree → commit-tree → update-ref. This is what `git` itself does |
| 7 | +// internally; we drive it from outside via short-lived subprocesses so |
| 8 | +// we don't need to vendor a Go-native git implementation. |
| 9 | +// |
| 10 | +// Every shell-out is constrained to a caller-supplied gitDir we control |
| 11 | +// (`storage.RepoFS.RepoPath` produces it from a strict path whitelist). |
| 12 | +// `git -C` is used instead of changing the process working directory. |
| 13 | +package git |
| 14 | + |
| 15 | +import ( |
| 16 | + "bytes" |
| 17 | + "context" |
| 18 | + "errors" |
| 19 | + "fmt" |
| 20 | + "os" |
| 21 | + "os/exec" |
| 22 | + "strings" |
| 23 | + "time" |
| 24 | +) |
| 25 | + |
| 26 | +// FileEntry is one file destined for the initial commit. Path is the |
| 27 | +// repository-relative slash-separated path (no leading slash). Body is |
| 28 | +// the raw bytes; we hash whatever we're given without normalization. |
| 29 | +type FileEntry struct { |
| 30 | + Path string |
| 31 | + Body []byte |
| 32 | +} |
| 33 | + |
| 34 | +// InitialCommit describes the single commit produced for a freshly |
| 35 | +// initialized repo whose owner ticked "initialize with README/license/ |
| 36 | +// .gitignore" on the create form. Everything except Files defaults to |
| 37 | +// sensible values; callers must always set GitDir + Author* + Files. |
| 38 | +type InitialCommit struct { |
| 39 | + GitDir string // absolute bare-repo path (must already exist + be inited) |
| 40 | + AuthorName string |
| 41 | + AuthorEmail string |
| 42 | + Message string // defaults to "Initial commit" |
| 43 | + Branch string // defaults to "trunk" |
| 44 | + When time.Time // defaults to time.Now() |
| 45 | + Files []FileEntry |
| 46 | +} |
| 47 | + |
| 48 | +// Build runs the full plumbing sequence and returns the new commit OID. |
| 49 | +// On error the bare repo may have orphan blobs/trees written; the |
| 50 | +// caller is expected to RemoveAll the bare repo dir on failure. |
| 51 | +func (ic InitialCommit) Build(ctx context.Context) (string, error) { |
| 52 | + if ic.GitDir == "" { |
| 53 | + return "", errors.New("git plumbing: GitDir is required") |
| 54 | + } |
| 55 | + if ic.AuthorName == "" || ic.AuthorEmail == "" { |
| 56 | + return "", errors.New("git plumbing: author name/email are required") |
| 57 | + } |
| 58 | + if len(ic.Files) == 0 { |
| 59 | + return "", errors.New("git plumbing: at least one file is required") |
| 60 | + } |
| 61 | + if ic.Message == "" { |
| 62 | + ic.Message = "Initial commit" |
| 63 | + } |
| 64 | + if ic.Branch == "" { |
| 65 | + ic.Branch = "trunk" |
| 66 | + } |
| 67 | + if ic.When.IsZero() { |
| 68 | + ic.When = time.Now() |
| 69 | + } |
| 70 | + |
| 71 | + indexFile, err := os.CreateTemp("", "shithub-index-*") |
| 72 | + if err != nil { |
| 73 | + return "", fmt.Errorf("git plumbing: temp index: %w", err) |
| 74 | + } |
| 75 | + indexPath := indexFile.Name() |
| 76 | + _ = indexFile.Close() |
| 77 | + defer func() { _ = os.Remove(indexPath) }() |
| 78 | + // git refuses to use an empty file as an index — delete now and let |
| 79 | + // `update-index` recreate it on first use. |
| 80 | + _ = os.Remove(indexPath) |
| 81 | + |
| 82 | + for _, f := range ic.Files { |
| 83 | + oid, err := ic.hashObject(ctx, f.Body) |
| 84 | + if err != nil { |
| 85 | + return "", fmt.Errorf("hash-object %s: %w", f.Path, err) |
| 86 | + } |
| 87 | + if err := ic.updateIndex(ctx, indexPath, oid, f.Path); err != nil { |
| 88 | + return "", fmt.Errorf("update-index %s: %w", f.Path, err) |
| 89 | + } |
| 90 | + } |
| 91 | + |
| 92 | + tree, err := ic.writeTree(ctx, indexPath) |
| 93 | + if err != nil { |
| 94 | + return "", fmt.Errorf("write-tree: %w", err) |
| 95 | + } |
| 96 | + commit, err := ic.commitTree(ctx, tree) |
| 97 | + if err != nil { |
| 98 | + return "", fmt.Errorf("commit-tree: %w", err) |
| 99 | + } |
| 100 | + if err := ic.updateRef(ctx, commit); err != nil { |
| 101 | + return "", fmt.Errorf("update-ref: %w", err) |
| 102 | + } |
| 103 | + return commit, nil |
| 104 | +} |
| 105 | + |
| 106 | +// hashObject pipes body through `git hash-object -w --stdin` and |
| 107 | +// returns the resulting OID. |
| 108 | +func (ic InitialCommit) hashObject(ctx context.Context, body []byte) (string, error) { |
| 109 | + //nolint:gosec // G204: gitDir is constrained by storage.RepoFS path validation. |
| 110 | + cmd := exec.CommandContext(ctx, "git", "-C", ic.GitDir, "hash-object", "-w", "--stdin") |
| 111 | + cmd.Stdin = bytes.NewReader(body) |
| 112 | + out, err := cmd.Output() |
| 113 | + if err != nil { |
| 114 | + return "", wrapExecErr(err) |
| 115 | + } |
| 116 | + return strings.TrimSpace(string(out)), nil |
| 117 | +} |
| 118 | + |
| 119 | +// updateIndex stages a blob at path under the temp index pointed at by |
| 120 | +// indexPath. Mode is fixed at 100644 (regular file); the spec doesn't |
| 121 | +// emit symlinks or executables for the initial commit. |
| 122 | +func (ic InitialCommit) updateIndex(ctx context.Context, indexPath, oid, path string) error { |
| 123 | + cacheinfo := fmt.Sprintf("100644,%s,%s", oid, path) |
| 124 | + //nolint:gosec // G204: gitDir + path are validated upstream. |
| 125 | + cmd := exec.CommandContext(ctx, "git", "-C", ic.GitDir, "update-index", "--add", "--cacheinfo", cacheinfo) |
| 126 | + cmd.Env = append(os.Environ(), "GIT_INDEX_FILE="+indexPath) |
| 127 | + if out, err := cmd.CombinedOutput(); err != nil { |
| 128 | + return fmt.Errorf("%w: %s", wrapExecErr(err), strings.TrimSpace(string(out))) |
| 129 | + } |
| 130 | + return nil |
| 131 | +} |
| 132 | + |
| 133 | +// writeTree turns the staged index into a tree object and returns its OID. |
| 134 | +func (ic InitialCommit) writeTree(ctx context.Context, indexPath string) (string, error) { |
| 135 | + //nolint:gosec // G204: gitDir validated upstream. |
| 136 | + cmd := exec.CommandContext(ctx, "git", "-C", ic.GitDir, "write-tree") |
| 137 | + cmd.Env = append(os.Environ(), "GIT_INDEX_FILE="+indexPath) |
| 138 | + out, err := cmd.Output() |
| 139 | + if err != nil { |
| 140 | + return "", wrapExecErr(err) |
| 141 | + } |
| 142 | + return strings.TrimSpace(string(out)), nil |
| 143 | +} |
| 144 | + |
| 145 | +// commitTree builds a commit object pointing at tree, with no parent. |
| 146 | +// Author + committer are both populated from ic.Author*; ic.When fixes |
| 147 | +// both timestamps so the test suite gets deterministic OIDs. |
| 148 | +func (ic InitialCommit) commitTree(ctx context.Context, tree string) (string, error) { |
| 149 | + //nolint:gosec // G204: tree is git's stdout (40-char OID); gitDir validated. |
| 150 | + cmd := exec.CommandContext(ctx, "git", "-C", ic.GitDir, "commit-tree", tree, "-m", ic.Message) |
| 151 | + stamp := ic.When.Format(time.RFC3339) |
| 152 | + cmd.Env = append(os.Environ(), |
| 153 | + "GIT_AUTHOR_NAME="+ic.AuthorName, |
| 154 | + "GIT_AUTHOR_EMAIL="+ic.AuthorEmail, |
| 155 | + "GIT_AUTHOR_DATE="+stamp, |
| 156 | + "GIT_COMMITTER_NAME="+ic.AuthorName, |
| 157 | + "GIT_COMMITTER_EMAIL="+ic.AuthorEmail, |
| 158 | + "GIT_COMMITTER_DATE="+stamp, |
| 159 | + ) |
| 160 | + out, err := cmd.Output() |
| 161 | + if err != nil { |
| 162 | + return "", wrapExecErr(err) |
| 163 | + } |
| 164 | + return strings.TrimSpace(string(out)), nil |
| 165 | +} |
| 166 | + |
| 167 | +// updateRef points refs/heads/<Branch> at commit. After this call the |
| 168 | +// bare repo's HEAD (a symbolic ref to refs/heads/<Branch>) finally |
| 169 | +// resolves to a real commit. |
| 170 | +func (ic InitialCommit) updateRef(ctx context.Context, commit string) error { |
| 171 | + ref := "refs/heads/" + ic.Branch |
| 172 | + //nolint:gosec // G204: ref is constructed from a non-empty branch name we set. |
| 173 | + cmd := exec.CommandContext(ctx, "git", "-C", ic.GitDir, "update-ref", ref, commit) |
| 174 | + if out, err := cmd.CombinedOutput(); err != nil { |
| 175 | + return fmt.Errorf("%w: %s", wrapExecErr(err), strings.TrimSpace(string(out))) |
| 176 | + } |
| 177 | + return nil |
| 178 | +} |
| 179 | + |
| 180 | +// wrapExecErr unwraps an *exec.ExitError to expose stderr in the |
| 181 | +// returned message; on other errors it passes through. Useful when the |
| 182 | +// caller logs %w errors and we want the actual git stderr in the line. |
| 183 | +func wrapExecErr(err error) error { |
| 184 | + var ee *exec.ExitError |
| 185 | + if errors.As(err, &ee) { |
| 186 | + stderr := strings.TrimSpace(string(ee.Stderr)) |
| 187 | + if stderr != "" { |
| 188 | + return fmt.Errorf("%w: %s", err, stderr) |
| 189 | + } |
| 190 | + } |
| 191 | + return err |
| 192 | +} |
| 193 | + |