fix issues

internal/ultrabeam/ultrabeam.go

@@ -5,6 +5,7 @@ package ultrabeam
 
 import (
 	"bufio"
+	"errors"
 	"fmt"
 	"log"
 	"net"
@@ -12,6 +13,17 @@ import (
 	"time"
 )
 
+// Connection tuning. Remote operation (the antenna controller reached over the
+// internet, not the LAN) sees real latency and jitter, so the read timeout is
+// generous and a few transient timeouts are tolerated before the link is torn
+// down — otherwise a single slow reply dropped the whole connection and the
+// client churned reconnect/disconnect.
+const (
+	ubReadTimeout    = 4 * time.Second  // was 1s — too tight for a remote link
+	ubKeepAlive      = 15 * time.Second // OS-level TCP keepalive
+	ubMaxPollTimeout = 3                // consecutive read timeouts tolerated before reconnecting
+)
+
 // Protocol constants
 const (
 	STX byte = 0xF5 // 245 decimal
@@ -110,6 +122,7 @@ func (c *Client) pollLoop() {
 	defer ticker.Stop()
 
 	pollCount := 0
+	pollFails := 0 // consecutive failed status polls (transient timeouts tolerated)
 
 	for {
 		select {
@@ -120,7 +133,8 @@ func (c *Client) pollLoop() {
 			c.connMu.Lock()
 			if c.conn == nil {
 				log.Printf("Ultrabeam: Not connected, attempting connection...")
-				conn, err := net.DialTimeout("tcp", fmt.Sprintf("%s:%d", c.host, c.port), 5*time.Second)
+				dialer := net.Dialer{Timeout: 5 * time.Second, KeepAlive: ubKeepAlive}
+				conn, err := dialer.Dial("tcp", net.JoinHostPort(c.host, fmt.Sprintf("%d", c.port)))
 				if err != nil {
 					log.Printf("Ultrabeam: Connection failed: %v", err)
 					c.connMu.Unlock()
@@ -133,6 +147,7 @@ func (c *Client) pollLoop() {
 				}
 				c.conn = conn
 				c.reader = bufio.NewReader(c.conn)
+				pollFails = 0
 				log.Printf("Ultrabeam: Connected to %s:%d", c.host, c.port)
 			}
 			c.connMu.Unlock()
@@ -140,8 +155,17 @@ func (c *Client) pollLoop() {
 			// Query status
 			status, err := c.queryStatus()
 			if err != nil {
-				log.Printf("Ultrabeam: Failed to query status: %v", err)
-				// Close connection and retry
+				// A single slow/lost reply over a remote link is normal — keep
+				// the connection (and the last status) for a few tries before
+				// tearing it down, so we don't churn reconnect/disconnect.
+				var ne net.Error
+				transient := errors.As(err, &ne) && ne.Timeout()
+				pollFails++
+				if transient && pollFails < ubMaxPollTimeout {
+					log.Printf("Ultrabeam: status timeout (%d/%d), keeping link: %v", pollFails, ubMaxPollTimeout, err)
+					continue
+				}
+				log.Printf("Ultrabeam: Failed to query status, reconnecting: %v", err)
 				c.connMu.Lock()
 				if c.conn != nil {
 					c.conn.Close()
@@ -156,6 +180,7 @@ func (c *Client) pollLoop() {
 				c.statusMu.Unlock()
 				continue
 			}
+			pollFails = 0
 
 			// Mark as connected
 			status.Connected = true
@@ -318,8 +343,8 @@ func (c *Client) sendCommand(cmd byte, data []byte) ([]byte, error) {
 		return nil, fmt.Errorf("failed to write: %w", err)
 	}
 
-	// Read reply with timeout
-	c.conn.SetReadDeadline(time.Now().Add(1 * time.Second)) // Reduced from 2s to 1s
+	// Read reply with timeout (generous — tolerates remote-link latency).
+	c.conn.SetReadDeadline(time.Now().Add(ubReadTimeout))
 
 	// Read until we get a complete packet
 	var buffer []byte