chore: remove releaser v2 flow and drop v1 naming (#27421)

## Summary Removes the GitHub Actions-driven releaser **v2** pipeline so the interactive release wizard is the only release path, and drops the `v1` naming now that it is the sole implementation. ## Changes - Delete `.github/workflows/tag-and-release.yaml` (the v2 workflow). - Delete `scripts/releaser/v2/`. - Move `scripts/releaser/v1/` into `scripts/releaser/` as `package main`. - Rewrite `scripts/releaser/main.go` to a single wizard command: drop the `--legacy` flag and the v2 `rc`/`branch`/`release` subcommands and hidden CI compat commands. `--dry-run` is preserved. - Update `scripts/release.sh` to run `go run ./scripts/releaser "$@"` (no `--legacy`). The legacy `release.yaml` workflow (triggered by `scripts/release.sh`) is unchanged and remains the release pipeline. ## Validation - `go build ./scripts/releaser/...` - `go test ./scripts/releaser/...` - `go vet` + `golangci-lint run ./scripts/releaser/...` - `gofmt -l` clean > [!NOTE] > The GPG signing key check removal is handled in a stacked follow-up PR based on this branch. <details> <summary>Implementation plan</summary> - v2 flow = `scripts/releaser/v2/` + `.github/workflows/tag-and-release.yaml` (uses `go run ./scripts/releaser prepare-release|generate-notes`). `v2` was imported only by `main.go`; the workflow was referenced nowhere else. - v1 flow = interactive wizard in `scripts/releaser/v1/`, reached via `--legacy`, driving `release.yaml` (triggered by `scripts/release.sh`). - No `docs/` referenced the releaser tool or these workflows. - Steps: delete the v2 workflow and package; move `v1/*` up to `scripts/releaser/` (`package main`, including test files); rewrite `main.go` to a single wizard command; update `release.sh`. </details> --- Generated by Coder Agents on behalf of @f0ssel.
2026-09-24 15:04:27 +08:00 · 2026-07-24 12:52:01 -04:00
parent 320817fa30
commit 591f357574
30 changed files with 78 additions and 4004 deletions
@@ -0,0 +1,228 @@
+package main
+
+import (
+	"regexp"
+	"sort"
+	"strconv"
+	"strings"
+)
+
+// commitEntry represents a single non-merge commit.
+type commitEntry struct {
+	SHA       string
+	FullSHA   string
+	Title     string
+	PRCount   int // 0 if no PR number found
+	Timestamp int64
+}
+
+var prNumRe = regexp.MustCompile(`\(#(\d+)\)`)
+
+// cherryPickPRRe matches cherry-pick bot titles like
+// "chore: foo bar (cherry-pick #42) (#43)".
+var cherryPickPRRe = regexp.MustCompile(`\(cherry-pick #(\d+)\)\s*\(#\d+\)$`)
+
+// commitLog returns non-merge commits in the given range, filtering
+// out left-side commits (already in the base) and deduplicating
+// cherry-picks using git's --cherry-mark.
+func commitLog(commitRange string) ([]commitEntry, error) {
+	// Use --left-right --cherry-mark to identify equivalent
+	// (cherry-picked) commits and left-side-only commits.
+	out, err := gitOutput("log", "--no-merges", "--left-right", "--cherry-mark",
+		"--pretty=format:%m %ct %h %H %s", commitRange)
+	if err != nil {
+		return nil, err
+	}
+	if out == "" {
+		return nil, nil
+	}
+
+	// Collect cherry-pick equivalent commits (marked with '=') so
+	// we can skip duplicates. We keep only the right-side version.
+	seen := make(map[string]bool)
+
+	var entries []commitEntry
+	for _, line := range strings.Split(out, "\n") {
+		line = strings.TrimSpace(line)
+		if line == "" {
+			continue
+		}
+		// Format: %m %ct %h %H %s
+		// mark timestamp shortSHA fullSHA title...
+		parts := strings.SplitN(line, " ", 5)
+		if len(parts) < 5 {
+			continue
+		}
+		mark := parts[0]
+		ts, _ := strconv.ParseInt(parts[1], 10, 64)
+		shortSHA := parts[2]
+		fullSHA := parts[3]
+		title := parts[4]
+
+		// Skip left-side commits (already in the old version).
+		if mark == "<" {
+			continue
+		}
+		// Skip cherry-pick equivalents that we've already seen
+		// (marked '=' by --cherry-mark).
+		if mark == "=" {
+			if seen[title] {
+				continue
+			}
+			seen[title] = true
+		}
+
+		// Normalize cherry-pick bot titles:
+		// "chore: foo (cherry-pick #42) (#43)" → "chore: foo (#42)"
+		if m := cherryPickPRRe.FindStringSubmatch(title); m != nil {
+			title = title[:cherryPickPRRe.FindStringIndex(title)[0]] + "(#" + m[1] + ")"
+		}
+
+		e := commitEntry{
+			SHA:       shortSHA,
+			FullSHA:   fullSHA,
+			Title:     title,
+			Timestamp: ts,
+		}
+		if m := prNumRe.FindStringSubmatch(e.Title); m != nil {
+			e.PRCount, _ = strconv.Atoi(m[1])
+		}
+		entries = append(entries, e)
+	}
+
+	// Sort by conventional commit prefix, then by timestamp
+	// (matching the bash script's sort -k3,3 -k1,1n).
+	sort.SliceStable(entries, func(i, j int) bool {
+		pi := commitSortPrefix(entries[i].Title)
+		pj := commitSortPrefix(entries[j].Title)
+		if pi != pj {
+			return pi < pj
+		}
+		return entries[i].Timestamp < entries[j].Timestamp
+	})
+
+	return entries, nil
+}
+
+// commitSortPrefix extracts the first word of a title for sorting.
+func commitSortPrefix(title string) string {
+	idx := strings.IndexAny(title, " (:")
+	if idx < 0 {
+		return title
+	}
+	return title[:idx]
+}
+
+// humanizedAreas maps conventional commit scopes to human-readable area
+// names. Order matters: more specific prefixes must come first so that
+// the first partial match wins.
+var humanizedAreas = []struct {
+	Prefix string
+	Area   string
+}{
+	{"agent/agentssh", "Agent SSH"},
+	{"coderd/database", "Database"},
+	{"enterprise/audit", "Auditing"},
+	{"enterprise/cli", "CLI"},
+	{"enterprise/coderd", "Server"},
+	{"enterprise/dbcrypt", "Database"},
+	{"enterprise/derpmesh", "Networking"},
+	{"enterprise/provisionerd", "Provisioner"},
+	{"enterprise/tailnet", "Networking"},
+	{"enterprise/wsproxy", "Workspace Proxy"},
+	{"agent", "Agent"},
+	{"cli", "CLI"},
+	{"coderd", "Server"},
+	{"codersdk", "SDK"},
+	{"docs", "Documentation"},
+	{"enterprise", "Enterprise"},
+	{"examples", "Examples"},
+	{"helm", "Helm"},
+	{"install.sh", "Installer"},
+	{"provisionersdk", "SDK"},
+	{"provisionerd", "Provisioner"},
+	{"provisioner", "Provisioner"},
+	{"pty", "CLI"},
+	{"scaletest", "Scale Testing"},
+	{"site", "Dashboard"},
+	{"support", "Support"},
+	{"tailnet", "Networking"},
+}
+
+// conventionalPrefixRe extracts prefix, scope, and rest from a
+// conventional commit title. Does NOT match breaking "!" suffix —
+// those titles are left as-is (matching bash behavior).
+var conventionalPrefixRe = regexp.MustCompile(`^([a-z]+)(\((.+)\))?:\s*(.*)$`)
+
+// humanizeTitle converts a conventional commit title to a
+// human-readable form, e.g. "feat(site): add bar" → "Dashboard: Add bar".
+func humanizeTitle(title string) string {
+	m := conventionalPrefixRe.FindStringSubmatch(title)
+	if m == nil {
+		return title
+	}
+	scope := m[3] // may be empty
+	rest := m[4]
+	if rest == "" {
+		return title
+	}
+	// Capitalize the first letter of the rest.
+	rest = strings.ToUpper(rest[:1]) + rest[1:]
+
+	if scope == "" {
+		return rest
+	}
+
+	// Look up scope in humanizedAreas (first partial match wins).
+	for _, ha := range humanizedAreas {
+		if strings.HasPrefix(scope, ha.Prefix) {
+			return ha.Area + ": " + rest
+		}
+	}
+	// Scope not found in map — return as-is.
+	return title
+}
+
+// breakingCommitRe matches conventional commit "!:" breaking changes.
+var breakingCommitRe = regexp.MustCompile(`^[a-zA-Z]+(\(.+\))?!:`)
+
+// categorizeCommit determines the release note section for a commit.
+// The priority order matches the bash script: breaking title first,
+// then labels (breaking, security, experimental), then prefix.
+func categorizeCommit(title string, labels []string) string {
+	// Check breaking title first (matches bash behavior).
+	if breakingCommitRe.MatchString(title) {
+		return "breaking"
+	}
+
+	// Label-based categorization.
+	for _, l := range labels {
+		if l == "release/breaking" {
+			return "breaking"
+		}
+		if l == "security" {
+			return "security"
+		}
+		if l == "release/experimental" {
+			return "experimental"
+		}
+	}
+
+	// Extract the conventional commit prefix (e.g. "feat", "fix(scope)").
+	prefixRe := regexp.MustCompile(`^([a-z]+)(\(.+\))?[!]?:`)
+	m := prefixRe.FindStringSubmatch(title)
+	if m == nil {
+		return "other"
+	}
+
+	validPrefixes := []string{
+		"feat", "fix", "docs", "refactor", "perf",
+		"test", "build", "ci", "chore", "revert",
+	}
+	for _, p := range validPrefixes {
+		if m[1] == p {
+			return p
+		}
+	}
+	return "other"
+}