git.go

v1.0.0

package git

import (
	"fmt"
	"os"
	"os/exec"
	"path/filepath"
	"sort"
	"strings"
	"sync"
	"time"

	"golang.org/x/mod/semver"

	"example.com/curator/internal/metrics"
)

// Cache manages bare git clones with a mutex to serialize clone/fetch operations.
type Cache struct {
	dir string
	mu  sync.Mutex
}

func NewCache(dir string) *Cache {
	return &Cache{dir: dir}
}

// Dir returns the cache directory path.
func (c *Cache) Dir() string {
	return c.dir
}

// CloneOrFetch clones the repo if not cached, otherwise fetches new tags.
func (c *Cache) CloneOrFetch(name, repoURL string) (string, error) {
	repoPath := filepath.Join(c.dir, name)

	c.mu.Lock()
	defer c.mu.Unlock()

	if _, err := os.Stat(repoPath); err != nil {
		if err := os.MkdirAll(c.dir, 0o755); err != nil {
			return "", fmt.Errorf("create cache dir: %w", err)
		}

		cmd := exec.Command("git", "clone", "--bare", repoURL, repoPath)
		if out, err := cmd.CombinedOutput(); err != nil {
			return "", fmt.Errorf("git clone: %s: %w", out, err)
		}
		metrics.GitOperationsTotal.WithLabelValues("clone").Inc()
	} else {
		cmd := exec.Command("git", "-C", repoPath, "fetch", "--tags", "origin")
		if out, err := cmd.CombinedOutput(); err != nil {
			return "", fmt.Errorf("git fetch: %s: %w", out, err)
		}
		metrics.GitOperationsTotal.WithLabelValues("fetch").Inc()
	}

	return repoPath, nil
}

// ListTags returns all valid semver tags from the repository, sorted.
func ListTags(repoPath string) ([]string, error) {
	cmd := exec.Command("git", "-C", repoPath, "tag", "-l", "v*")
	out, err := cmd.Output()
	if err != nil {
		return nil, fmt.Errorf("git tag: %w", err)
	}

	var tags []string
	for _, line := range strings.Split(strings.TrimSpace(string(out)), "\n") {
		if line != "" && semver.IsValid(line) {
			tags = append(tags, line)
		}
	}

	semver.Sort(tags)

	return tags, nil
}

// FilterTagsByMajor returns only the tags that match the given module path's
// major version. For a module without a /vN suffix, only v0 and v1 tags are
// returned. For a module ending in /v2, only v2 tags are returned, etc.
func FilterTagsByMajor(tags []string, modulePath string) []string {
	major := MajorForModule(modulePath)

	var filtered []string
	for _, tag := range tags {
		if MatchesMajor(tag, major) {
			filtered = append(filtered, tag)
		}
	}
	return filtered
}

// MajorForModule extracts the major version that a module path expects.
// "foo/bar" → "v0v1", "foo/bar/v2" → "v2", "foo/bar/v3" → "v3".
func MajorForModule(modulePath string) string {
	// Check for /vN suffix.
	if i := strings.LastIndex(modulePath, "/v"); i != -1 {
		suffix := modulePath[i+1:] // "v2", "v3", etc.
		if semver.Major(""+suffix+".0.0") == suffix {
			return suffix
		}
	}
	return "v0v1" // sentinel: match both v0 and v1
}

// MatchesMajor reports whether a semver tag matches the expected major version.
// major is either "v0v1" (for unversioned module paths) or "vN".
func MatchesMajor(tag, major string) bool {
	tagMajor := semver.Major(tag)
	if major == "v0v1" {
		return tagMajor == "v0" || tagMajor == "v1"
	}
	return tagMajor == major
}

// LatestRelease returns the highest non-prerelease version from a sorted list
// of tags. If no release version exists, it falls back to the highest tag.
// This matches Go's @latest behavior: prefer stable releases over pre-releases.
func LatestRelease(tags []string) string {
	if len(tags) == 0 {
		return ""
	}

	// Walk backwards (tags are sorted) to find the first non-prerelease.
	for i := len(tags) - 1; i >= 0; i-- {
		if semver.Prerelease(tags[i]) == "" {
			return tags[i]
		}
	}

	// All versions are pre-releases, return the highest.
	return tags[len(tags)-1]
}

func ReadFile(repoPath, rev, filePath string) ([]byte, error) {
	cmd := exec.Command("git", "-C", repoPath, "show", rev+":"+filePath)
	out, err := cmd.Output()
	if err != nil {
		return nil, err
	}

	return out, nil
}

// CommitInfo returns the full commit hash and commit time for any git ref.
func CommitInfo(repoPath, rev string) (hash string, t time.Time, err error) {
	cmd := exec.Command("git", "-C", repoPath, "log", "-1", "--format=%H %cI", rev)
	out, err := cmd.Output()
	if err != nil {
		return "", time.Time{}, fmt.Errorf("git log %s: %w", rev, err)
	}

	parts := strings.SplitN(strings.TrimSpace(string(out)), " ", 2)
	if len(parts) != 2 {
		return "", time.Time{}, fmt.Errorf("unexpected git log output: %s", out)
	}

	t, err = time.Parse(time.RFC3339, parts[1])
	if err != nil {
		return "", time.Time{}, err
	}

	return parts[0], t, nil
}

// TreeEntry represents a file or directory in a git tree.
type TreeEntry struct {
	Name  string
	IsDir bool
}

// ListTree lists files and directories at a specific path in a revision.
func ListTree(repoPath, rev, dir string) ([]TreeEntry, error) {
	arg := rev
	if dir != "" {
		arg = rev + ":" + dir
	}

	cmd := exec.Command("git", "-C", repoPath, "ls-tree", "--name-only", arg)
	out, err := cmd.Output()
	if err != nil {
		return nil, fmt.Errorf("git ls-tree: %w", err)
	}

	// We need a second call to get types. Use ls-tree with -l or parse mode bits.
	cmd2 := exec.Command("git", "-C", repoPath, "ls-tree", arg)
	out2, err := cmd2.Output()
	if err != nil {
		return nil, fmt.Errorf("git ls-tree: %w", err)
	}

	_ = out // not used, we parse the full output instead

	var entries []TreeEntry
	for _, line := range strings.Split(strings.TrimSpace(string(out2)), "\n") {
		if line == "" {
			continue
		}
		// Format: <mode> <type> <hash>\t<name>
		parts := strings.SplitN(line, "\t", 2)
		if len(parts) != 2 {
			continue
		}
		name := parts[1]
		isDir := strings.HasPrefix(line, "040000") || strings.Contains(parts[0], "tree")
		entries = append(entries, TreeEntry{Name: name, IsDir: isDir})
	}

	return entries, nil
}

// ReadDir reads all .go files (excluding test files) from a directory at a revision.
// Returns a map of filename to contents.
func ReadDir(repoPath, rev, dir string) (map[string][]byte, error) {
	entries, err := ListTree(repoPath, rev, dir)
	if err != nil {
		return nil, err
	}

	files := make(map[string][]byte)
	for _, e := range entries {
		if e.IsDir {
			continue
		}
		if !strings.HasSuffix(e.Name, ".go") {
			continue
		}
		if strings.HasSuffix(e.Name, "_test.go") {
			continue
		}

		path := e.Name
		if dir != "" {
			path = dir + "/" + e.Name
		}

		data, err := ReadFile(repoPath, rev, path)
		if err != nil {
			return nil, fmt.Errorf("read %s: %w", path, err)
		}
		files[e.Name] = data
	}

	return files, nil
}

// ListSubDirs returns directories within a path that contain .go files (i.e., Go packages).
func ListSubDirs(repoPath, rev, dir string) ([]string, error) {
	cmd := exec.Command("git", "-C", repoPath, "ls-tree", "-r", "--name-only", rev)
	out, err := cmd.Output()
	if err != nil {
		return nil, fmt.Errorf("git ls-tree: %w", err)
	}

	dirSet := make(map[string]bool)
	prefix := ""
	if dir != "" {
		prefix = dir + "/"
	}

	for _, line := range strings.Split(strings.TrimSpace(string(out)), "\n") {
		if line == "" {
			continue
		}
		if !strings.HasSuffix(line, ".go") {
			continue
		}
		if strings.HasSuffix(line, "_test.go") {
			continue
		}
		// Skip vendor and hidden directories.
		if strings.Contains(line, "vendor/") || strings.Contains(line, "/.") || strings.HasPrefix(line, ".") {
			continue
		}
		// Skip files starting with _
		base := filepath.Base(line)
		if strings.HasPrefix(base, "_") {
			continue
		}

		d := filepath.Dir(line)
		if d == "." {
			continue // root package, not a sub-directory
		}
		if prefix != "" {
			if !strings.HasPrefix(d, prefix) {
				continue
			}
			d = strings.TrimPrefix(d, prefix)
		}
		dirSet[d] = true
	}

	var dirs []string
	for d := range dirSet {
		dirs = append(dirs, d)
	}

	// Sort for stable output.
	sort.Strings(dirs)

	return dirs, nil
}

// LatestTag returns the most recent semver tag reachable from a commit,
// or "" if none exists.
func LatestTag(repoPath, rev string) string {
	cmd := exec.Command("git", "-C", repoPath, "describe", "--tags", "--abbrev=0", "--match=v*", rev)
	out, err := cmd.Output()
	if err != nil {
		return ""
	}

	tag := strings.TrimSpace(string(out))
	if semver.IsValid(tag) {
		return tag
	}

	return ""
}


1	package git
2
3	import (
4	"fmt"
5	"os"
6	"os/exec"
7	"path/filepath"
8	"sort"
9	"strings"
10	"sync"
11	"time"
12
13	"golang.org/x/mod/semver"
14
15	"example.com/curator/internal/metrics"
16	)
17
18	// Cache manages bare git clones with a mutex to serialize clone/fetch operations.
19	type Cache struct {
20	dir string
21	mu sync.Mutex
22	}
23
24	func NewCache(dir string) *Cache {
25	return &Cache{dir: dir}
26	}
27
28	// Dir returns the cache directory path.
29	func (c *Cache) Dir() string {
30	return c.dir
31	}
32
33	// CloneOrFetch clones the repo if not cached, otherwise fetches new tags.
34	func (c *Cache) CloneOrFetch(name, repoURL string) (string, error) {
35	repoPath := filepath.Join(c.dir, name)
36
37	c.mu.Lock()
38	defer c.mu.Unlock()
39
40	if _, err := os.Stat(repoPath); err != nil {
41	if err := os.MkdirAll(c.dir, 0o755); err != nil {
42	return "", fmt.Errorf("create cache dir: %w", err)
43	}
44
45	cmd := exec.Command("git", "clone", "--bare", repoURL, repoPath)
46	if out, err := cmd.CombinedOutput(); err != nil {
47	return "", fmt.Errorf("git clone: %s: %w", out, err)
48	}
49	metrics.GitOperationsTotal.WithLabelValues("clone").Inc()
50	} else {
51	cmd := exec.Command("git", "-C", repoPath, "fetch", "--tags", "origin")
52	if out, err := cmd.CombinedOutput(); err != nil {
53	return "", fmt.Errorf("git fetch: %s: %w", out, err)
54	}
55	metrics.GitOperationsTotal.WithLabelValues("fetch").Inc()
56	}
57
58	return repoPath, nil
59	}
60
61	// ListTags returns all valid semver tags from the repository, sorted.
62	func ListTags(repoPath string) ([]string, error) {
63	cmd := exec.Command("git", "-C", repoPath, "tag", "-l", "v*")
64	out, err := cmd.Output()
65	if err != nil {
66	return nil, fmt.Errorf("git tag: %w", err)
67	}
68
69	var tags []string
70	for _, line := range strings.Split(strings.TrimSpace(string(out)), "\n") {
71	if line != "" && semver.IsValid(line) {
72	tags = append(tags, line)
73	}
74	}
75
76	semver.Sort(tags)
77
78	return tags, nil
79	}
80
81	// FilterTagsByMajor returns only the tags that match the given module path's
82	// major version. For a module without a /vN suffix, only v0 and v1 tags are
83	// returned. For a module ending in /v2, only v2 tags are returned, etc.
84	func FilterTagsByMajor(tags []string, modulePath string) []string {
85	major := MajorForModule(modulePath)
86
87	var filtered []string
88	for _, tag := range tags {
89	if MatchesMajor(tag, major) {
90	filtered = append(filtered, tag)
91	}
92	}
93	return filtered
94	}
95
96	// MajorForModule extracts the major version that a module path expects.
97	// "foo/bar" → "v0v1", "foo/bar/v2" → "v2", "foo/bar/v3" → "v3".
98	func MajorForModule(modulePath string) string {
99	// Check for /vN suffix.
100	if i := strings.LastIndex(modulePath, "/v"); i != -1 {
101	suffix := modulePath[i+1:] // "v2", "v3", etc.
102	if semver.Major(""+suffix+".0.0") == suffix {
103	return suffix
104	}
105	}
106	return "v0v1" // sentinel: match both v0 and v1
107	}
108
109	// MatchesMajor reports whether a semver tag matches the expected major version.
110	// major is either "v0v1" (for unversioned module paths) or "vN".
111	func MatchesMajor(tag, major string) bool {
112	tagMajor := semver.Major(tag)
113	if major == "v0v1" {
114	return tagMajor == "v0" \|\| tagMajor == "v1"
115	}
116	return tagMajor == major
117	}
118
119	// LatestRelease returns the highest non-prerelease version from a sorted list
120	// of tags. If no release version exists, it falls back to the highest tag.
121	// This matches Go's @latest behavior: prefer stable releases over pre-releases.
122	func LatestRelease(tags []string) string {
123	if len(tags) == 0 {
124	return ""
125	}
126
127	// Walk backwards (tags are sorted) to find the first non-prerelease.
128	for i := len(tags) - 1; i >= 0; i-- {
129	if semver.Prerelease(tags[i]) == "" {
130	return tags[i]
131	}
132	}
133
134	// All versions are pre-releases, return the highest.
135	return tags[len(tags)-1]
136	}
137
138	func ReadFile(repoPath, rev, filePath string) ([]byte, error) {
139	cmd := exec.Command("git", "-C", repoPath, "show", rev+":"+filePath)
140	out, err := cmd.Output()
141	if err != nil {
142	return nil, err
143	}
144
145	return out, nil
146	}
147
148	// CommitInfo returns the full commit hash and commit time for any git ref.
149	func CommitInfo(repoPath, rev string) (hash string, t time.Time, err error) {
150	cmd := exec.Command("git", "-C", repoPath, "log", "-1", "--format=%H %cI", rev)
151	out, err := cmd.Output()
152	if err != nil {
153	return "", time.Time{}, fmt.Errorf("git log %s: %w", rev, err)
154	}
155
156	parts := strings.SplitN(strings.TrimSpace(string(out)), " ", 2)
157	if len(parts) != 2 {
158	return "", time.Time{}, fmt.Errorf("unexpected git log output: %s", out)
159	}
160
161	t, err = time.Parse(time.RFC3339, parts[1])
162	if err != nil {
163	return "", time.Time{}, err
164	}
165
166	return parts[0], t, nil
167	}
168
169	// TreeEntry represents a file or directory in a git tree.
170	type TreeEntry struct {
171	Name string
172	IsDir bool
173	}
174
175	// ListTree lists files and directories at a specific path in a revision.
176	func ListTree(repoPath, rev, dir string) ([]TreeEntry, error) {
177	arg := rev
178	if dir != "" {
179	arg = rev + ":" + dir
180	}
181
182	cmd := exec.Command("git", "-C", repoPath, "ls-tree", "--name-only", arg)
183	out, err := cmd.Output()
184	if err != nil {
185	return nil, fmt.Errorf("git ls-tree: %w", err)
186	}
187
188	// We need a second call to get types. Use ls-tree with -l or parse mode bits.
189	cmd2 := exec.Command("git", "-C", repoPath, "ls-tree", arg)
190	out2, err := cmd2.Output()
191	if err != nil {
192	return nil, fmt.Errorf("git ls-tree: %w", err)
193	}
194
195	_ = out // not used, we parse the full output instead
196
197	var entries []TreeEntry
198	for _, line := range strings.Split(strings.TrimSpace(string(out2)), "\n") {
199	if line == "" {
200	continue
201	}
202	// Format: <mode> <type> <hash>\t<name>
203	parts := strings.SplitN(line, "\t", 2)
204	if len(parts) != 2 {
205	continue
206	}
207	name := parts[1]
208	isDir := strings.HasPrefix(line, "040000") \|\| strings.Contains(parts[0], "tree")
209	entries = append(entries, TreeEntry{Name: name, IsDir: isDir})
210	}
211
212	return entries, nil
213	}
214
215	// ReadDir reads all .go files (excluding test files) from a directory at a revision.
216	// Returns a map of filename to contents.
217	func ReadDir(repoPath, rev, dir string) (map[string][]byte, error) {
218	entries, err := ListTree(repoPath, rev, dir)
219	if err != nil {
220	return nil, err
221	}
222
223	files := make(map[string][]byte)
224	for _, e := range entries {
225	if e.IsDir {
226	continue
227	}
228	if !strings.HasSuffix(e.Name, ".go") {
229	continue
230	}
231	if strings.HasSuffix(e.Name, "_test.go") {
232	continue
233	}
234
235	path := e.Name
236	if dir != "" {
237	path = dir + "/" + e.Name
238	}
239
240	data, err := ReadFile(repoPath, rev, path)
241	if err != nil {
242	return nil, fmt.Errorf("read %s: %w", path, err)
243	}
244	files[e.Name] = data
245	}
246
247	return files, nil
248	}
249
250	// ListSubDirs returns directories within a path that contain .go files (i.e., Go packages).
251	func ListSubDirs(repoPath, rev, dir string) ([]string, error) {
252	cmd := exec.Command("git", "-C", repoPath, "ls-tree", "-r", "--name-only", rev)
253	out, err := cmd.Output()
254	if err != nil {
255	return nil, fmt.Errorf("git ls-tree: %w", err)
256	}
257
258	dirSet := make(map[string]bool)
259	prefix := ""
260	if dir != "" {
261	prefix = dir + "/"
262	}
263
264	for _, line := range strings.Split(strings.TrimSpace(string(out)), "\n") {
265	if line == "" {
266	continue
267	}
268	if !strings.HasSuffix(line, ".go") {
269	continue
270	}
271	if strings.HasSuffix(line, "_test.go") {
272	continue
273	}
274	// Skip vendor and hidden directories.
275	if strings.Contains(line, "vendor/") \|\| strings.Contains(line, "/.") \|\| strings.HasPrefix(line, ".") {
276	continue
277	}
278	// Skip files starting with _
279	base := filepath.Base(line)
280	if strings.HasPrefix(base, "_") {
281	continue
282	}
283
284	d := filepath.Dir(line)
285	if d == "." {
286	continue // root package, not a sub-directory
287	}
288	if prefix != "" {
289	if !strings.HasPrefix(d, prefix) {
290	continue
291	}
292	d = strings.TrimPrefix(d, prefix)
293	}
294	dirSet[d] = true
295	}
296
297	var dirs []string
298	for d := range dirSet {
299	dirs = append(dirs, d)
300	}
301
302	// Sort for stable output.
303	sort.Strings(dirs)
304
305	return dirs, nil
306	}
307
308	// LatestTag returns the most recent semver tag reachable from a commit,
309	// or "" if none exists.
310	func LatestTag(repoPath, rev string) string {
311	cmd := exec.Command("git", "-C", repoPath, "describe", "--tags", "--abbrev=0", "--match=v*", rev)
312	out, err := cmd.Output()
313	if err != nil {
314	return ""
315	}
316
317	tag := strings.TrimSpace(string(out))
318	if semver.IsValid(tag) {
319	return tag
320	}
321
322	return ""
323	}
324

git.go

Source Files