fix: make WAL segment directory fsync failure fatal (C6)

Per design §3.2 line 248-272, segment directory fsync is a hard
requirement for durable-ready state, not best-effort. rename is atomic
in memory but not guaranteed to survive power loss without a directory
fsync. The previous code silently swallowed both os.Open(dir) and
dirFD.Sync() errors, leaving WAL writer to confirm batches as durable
when their segment might not exist after a crash.

Failure propagation:
- Initial segment creation: NewSegmentWriter fails -> NewSegmentManager
  fails -> DB.Open fails (user sees error, no data promise violated).
- Rotation during AppendBatch: NewSegmentWriter fails -> AppendBatch
  fails -> WalWriter.stopWithError(ErrCommitUnknown) -> write-stopped
  (per design line 272).

Changes:
- wal/segment_writer.go: extract dirFsync helper (Open -> f.Stat ->
  IsDir -> f.Sync, avoiding TOCTOU window), replace silent swallow with
  fatal error; on failure clean up resources (fd.Close + os.Remove) and
  surface cleanup errors via errors.Join so nothing is silently lost.
- wal/dir_fsync_test.go (new): unit test the helper with valid dir,
  non-existent dir (fails at os.Open), and not-a-dir (fails at IsDir).
- wal/segment_writer_test.go: add TestNewSegmentWriterDirFsyncFailure
  (injects failure via package-level dirFsyncFn override; documents the
  not-parallel-safe constraint), TestNewSegmentWriterNormalPathStillWorks
  (regression), and TestNewSegmentWriterRetryAfterDirFsyncFailure
  (verifies cleanup is effective for retry).
- wal/segment_manager_test.go: add TestSegmentManagerRotateFailsOnDirFsyncFailure
  (fills segment until rotation triggers, injects failure, verifies
  propagation through AppendBatch path) and TestNewSegmentManagerFailsOnDirFsyncFailure
  (covers the DB.Open failure path).

dirFsyncFn injection note: tests that override this package-level var
must not use t.Parallel(). All existing wal tests run serially within
the package; this is the lightest mechanism that doesn't require
interface indirection in production code.

Verified: each new test fails on pre-fix code (silent swallow returned
nil error) and passes after the fix. Full suite green including
go test -race ./... .

Audit context: docs/audit-3.2.md C6 (Oracle-verified bg_ef425776).

wal/segment_writer_test.go

@@ -1,9 +1,11 @@
 package wal
 
 import (
+	"errors"
 	"fmt"
 	"os"
 	"path/filepath"
+	"strings"
 	"testing"
 
 	"github.com/dailz/go-kv/config"
@@ -344,3 +346,99 @@ func TestSegmentWriterSync(t *testing.T) {
 		t.Fatalf("Close: %v", err)
 	}
 }
+
+// Regression guard for C6: NewSegmentWriter must return an error when
+// directory fsync fails, instead of silently succeeding. Per design §3.2
+// line 272, segment must NOT become active if durable-ready fails.
+//
+// dirFsyncFn is a package-level var; tests that override it must not use
+// t.Parallel(). All wal tests run serially within the package.
+func TestNewSegmentWriterDirFsyncFailure(t *testing.T) {
+	dir := t.TempDir()
+	cfg := config.Defaults()
+
+	orig := dirFsyncFn
+	dirFsyncFn = func(string) error { return errors.New("simulated dir fsync failure") }
+	t.Cleanup(func() { dirFsyncFn = orig })
+
+	sw, err := NewSegmentWriter(dir, 0, 0, &cfg)
+	if err == nil {
+		if sw != nil {
+			sw.Close()
+		}
+		t.Fatal("NewSegmentWriter: expected error on dir fsync failure, got nil")
+	}
+
+	if !strings.Contains(err.Error(), "fsync directory") {
+		t.Errorf("error should mention 'fsync directory', got: %v", err)
+	}
+
+	entries, err := os.ReadDir(dir)
+	if err != nil {
+		t.Fatalf("ReadDir: %v", err)
+	}
+	for _, e := range entries {
+		name := e.Name()
+		if strings.Contains(name, "segment-0") {
+			t.Errorf("segment file should be cleaned up, found: %s", name)
+		}
+	}
+}
+
+// Regression guard for C6: ensure normal path still works after the fix.
+func TestNewSegmentWriterNormalPathStillWorks(t *testing.T) {
+	dir := t.TempDir()
+	cfg := config.Defaults()
+
+	sw, err := NewSegmentWriter(dir, 0, 0, &cfg)
+	if err != nil {
+		t.Fatalf("NewSegmentWriter normal path: %v", err)
+	}
+	defer sw.Close()
+
+	if _, err := os.Stat(sw.SegmentPath()); err != nil {
+		t.Errorf("segment file should exist: %v", err)
+	}
+}
+
+// Regression guard for C6: after a failed NewSegmentWriter due to dir
+// fsync, retrying with fsync restored must succeed and not leak state.
+func TestNewSegmentWriterRetryAfterDirFsyncFailure(t *testing.T) {
+	dir := t.TempDir()
+	cfg := config.Defaults()
+
+	orig := dirFsyncFn
+	dirFsyncFn = func(string) error { return errors.New("simulated") }
+	_, err := NewSegmentWriter(dir, 0, 0, &cfg)
+	if err == nil {
+		t.Fatal("expected first NewSegmentWriter to fail")
+	}
+	dirFsyncFn = orig
+
+	sw, err := NewSegmentWriter(dir, 0, 0, &cfg)
+	if err != nil {
+		t.Fatalf("retry NewSegmentWriter: %v", err)
+	}
+	defer sw.Close()
+
+	entries, err := os.ReadDir(dir)
+	if err != nil {
+		t.Fatalf("ReadDir: %v", err)
+	}
+	tmpCount := 0
+	walCount := 0
+	for _, e := range entries {
+		if strings.HasSuffix(e.Name(), ".tmp") {
+			tmpCount++
+		}
+		if strings.HasSuffix(e.Name(), ".wal") {
+			walCount++
+		}
+	}
+	if tmpCount != 0 {
+		t.Errorf("leftover .tmp files: %d", tmpCount)
+	}
+	if walCount != 1 {
+		t.Errorf("expected exactly 1 .wal file (from successful retry), got %d", walCount)
+	}
+}