Use bigger offset in test to avoid random test failures.

Since Windows cannot to a SIGHUP test to check if a process is live,
the test will fail because the lock isn't detected as stale if a process with
that ID exists. Process IDs are re-used agressively on Windows, so add 500000 makes
the test extremely unlikely (if not impossible) to fail.

lock_test.go

@@ -124,7 +124,7 @@ var staleLockTests = []struct {
 	{
 		timestamp: time.Now(),
 		stale:     true,
-		pid:       os.Getpid() + 500,
+		pid:       os.Getpid() + 500000,
 	},
 }
 
@@ -158,7 +158,7 @@ func TestLockWithStaleLock(t *testing.T) {
 	id2, err := createFakeLock(repo, time.Now().Add(-time.Minute), os.Getpid())
 	OK(t, err)
 
-	id3, err := createFakeLock(repo, time.Now().Add(-time.Minute), os.Getpid()+500)
+	id3, err := createFakeLock(repo, time.Now().Add(-time.Minute), os.Getpid()+500000)
 	OK(t, err)
 
 	OK(t, restic.RemoveStaleLocks(repo))