feat: polish & windviz & deploy

internal/datasets/lock_unix.go

@@ -0,0 +1,50 @@
+//go:build unix
+
+package datasets
+
+import (
+	"context"
+	"errors"
+	"fmt"
+	"os"
+	"syscall"
+	"time"
+)
+
+// lockPollInterval is how often a contended lock is retried. The lock is held
+// for the duration of a dataset download (minutes), so sub-second acquisition
+// latency is irrelevant.
+const lockPollInterval = 150 * time.Millisecond
+
+// flockExclusive acquires an exclusive flock on path, creating the lock file
+// if needed, and blocks until it is held or ctx is cancelled.
+//
+// It uses non-blocking LOCK_NB attempts in a poll loop rather than a blocking
+// flock in a goroutine: the file descriptor is only ever touched by this
+// goroutine, so there is no race between a pending syscall and Close on
+// cancellation.
+func flockExclusive(ctx context.Context, path string) (func(), error) {
+	f, err := os.OpenFile(path, os.O_CREATE|os.O_RDWR, 0o644)
+	if err != nil {
+		return nil, fmt.Errorf("open lock file: %w", err)
+	}
+	for {
+		err := syscall.Flock(int(f.Fd()), syscall.LOCK_EX|syscall.LOCK_NB)
+		if err == nil {
+			return func() {
+				_ = syscall.Flock(int(f.Fd()), syscall.LOCK_UN)
+				_ = f.Close()
+			}, nil
+		}
+		if !errors.Is(err, syscall.EWOULDBLOCK) {
+			f.Close()
+			return nil, fmt.Errorf("flock: %w", err)
+		}
+		select {
+		case <-ctx.Done():
+			f.Close()
+			return nil, ctx.Err()
+		case <-time.After(lockPollInterval):
+		}
+	}
+}