OTA deleted from memory after done

boot.py

@@ -0,0 +1,64 @@
+"""
+boot.py — runs before main.py on every ESP32 boot
+
+Connects WiFi, runs OTA update, then hands off to main.py.
+Keep this file as simple as possible — it is never OTA-updated itself
+(it lives outside the repo folder) so bugs here require USB to fix.
+"""
+import gauge
+import network
+import gc
+import utime
+import sys
+
+import ota
+
+ota.load_config()
+WIFI_SSID, WIFI_PASSWORD = ota.WIFI_SSID, ota.WIFI_PASSWORD
+
+def _connect_wifi(timeout_s=20):
+    sta = network.WLAN(network.STA_IF)
+    sta.active(True)
+    sta.config(txpower=15)
+    if sta.isconnected():
+        return True
+    sta.connect(WIFI_SSID, WIFI_PASSWORD)
+    deadline = utime.time() + timeout_s
+    while not sta.isconnected():
+        if utime.time() > deadline:
+            return False
+        utime.sleep_ms(300)
+    return True
+
+if WIFI_SSID is None:
+    print("[boot] No WiFi credentials — cannot connect, skipping OTA")
+elif _connect_wifi():
+    ip = network.WLAN(network.STA_IF).ifconfig()[0]
+    print(f"[boot] WiFi connected — {ip}")
+
+    try:
+        ota.update()
+    except Exception as e:
+        print(f"[boot] OTA error: {e} — continuing with existing files")
+        sys.print_exception(e)
+    utime.sleep_ms(5000)
+    ota._fetch_commit_sha = None
+    ota._fetch_manifest = None
+    ota._fetch_dir = None
+    ota._api_get = None
+    ota._download = None
+    ota.urequests = None
+    del ota.urequests
+    del ota
+    gc.collect()
+    del sys.modules["ota"]
+    gc.collect()
+
+else:
+    print("[boot] WiFi failed — skipping OTA, booting with existing files")
+
+# main.py runs automatically after boot.py
+
+
+
+

gauge.py

@@ -28,6 +28,16 @@ import gc
 from umqtt.robust import MQTTClient
 from machine import UART
 
+# Activate WiFi driver before any heavy heap allocation so it can claim its
+# contiguous DRAM block before the Python heap fragments the address space.
+# Only activate if not already running (e.g. boot.py may have started it).
+gc.collect()
+_early_wlan = network.WLAN(network.STA_IF)
+if not _early_wlan.active():
+    _early_wlan.active(True)
+del _early_wlan
+gc.collect()
+
 # ---------------------------------------------------------------------------
 # Logging
 # ---------------------------------------------------------------------------
@@ -524,46 +534,56 @@ _DEVICE = {
 _wifi_check_interval_ms = 30000
 _last_wifi_check = 0
 _wifi_sta = None
+_WIFI_CONNECT_ATTEMPTS = 3
+
+
+def _reset_wifi_interface():
+    global _wifi_sta
+    _wifi_sta = network.WLAN(network.STA_IF)
+    if _wifi_sta.active():
+        _wifi_sta.active(False)
+        utime.sleep_ms(200)
+    _wifi_sta.active(True)
+    utime.sleep_ms(500)
 
 
 def connect_wifi(ssid, password, timeout_s=15, force_reconnect=False):
     global _wifi_sta
     _wifi_sta = network.WLAN(network.STA_IF)
 
-    if force_reconnect:
-        try:
-            _wifi_sta.disconnect()
-        except Exception:
-            pass
-        try:
-            _wifi_sta.active(False)
-            utime.sleep_ms(250)
-        except Exception:
-            pass
-
-    _wifi_sta.active(True)
     if _wifi_sta.isconnected() and not force_reconnect:
         ip, mask, gw, dns = _wifi_sta.ifconfig()
         info("WiFi already connected")
         info(f"  IP:{ip}  mask:{mask}  gw:{gw}  dns:{dns}")
         utime.sleep_ms(250)
         return ip
-    info(f"WiFi connecting to '{ssid}' ...")
+
+    last_error = None
+    for attempt in range(_WIFI_CONNECT_ATTEMPTS):
+        info(f"WiFi connecting to '{ssid}' (attempt {attempt + 1}/{_WIFI_CONNECT_ATTEMPTS}) ...")
+        _reset_wifi_interface()
+        try:
             _wifi_sta.connect(ssid, password)
             deadline = utime.time() + timeout_s
             while not _wifi_sta.isconnected():
                 if utime.time() > deadline:
-            log_err(f"WiFi connect timeout after {timeout_s}s")
                     raise OSError("WiFi connect timeout")
-        utime.sleep_ms(200)
+                utime.sleep_ms(250)
+
             ip, mask, gw, dns = _wifi_sta.ifconfig()
             mac = ":".join(f"{b:02x}" for b in _wifi_sta.config("mac"))
             info("WiFi connected!")
             info(f"  SSID    : {ssid}")
             info(f"  MAC     : {mac}")
             info(f"  IP      : {ip}  mask:{mask}  gw:{gw}  dns:{dns}")
-    utime.sleep_ms(250)
+            utime.sleep_ms(500)
             return ip
+        except Exception as e:
+            last_error = e
+            log_err(f"WiFi connect attempt {attempt + 1} failed: {e}")
+            utime.sleep_ms(1000)
+
+    raise last_error
 
 
 def check_wifi():
@@ -581,15 +601,7 @@ def check_wifi():
 
     log_err("WiFi lost connection — attempting reconnect...")
     try:
-        _wifi_sta.active(True)
-        _wifi_sta.connect(WIFI_SSID, WIFI_PASSWORD)
-        deadline = utime.time() + 15
-        while not _wifi_sta.isconnected():
-            if utime.time() > deadline:
-                log_err("WiFi reconnect timeout")
-                return
-            utime.sleep_ms(200)
-        ip, mask, gw, dns = _wifi_sta.ifconfig()
+        ip = connect_wifi(WIFI_SSID, WIFI_PASSWORD, timeout_s=15, force_reconnect=True)
         info(f"WiFi reconnected! IP:{ip}")
     except Exception as e:
         log_err(f"WiFi reconnect failed: {e}")
@@ -874,6 +886,7 @@ def connect_mqtt():
     info(f"Connecting to MQTT broker {MQTT_BROKER}:{MQTT_PORT} ...")
     last_error = None
     for attempt in range(3):
+        gc.collect()
         try:
             if client_ref is not None:
                 try:
@@ -897,7 +910,12 @@ def connect_mqtt():
             return
         except Exception as e:
             last_error = e
-            log_err(f"MQTT connect attempt {attempt + 1} failed: {e}")
+            log_err(f"MQTT connect attempt {attempt + 1} failed: {type(e).__name__}: {e}")
+            try:
+                client.sock.close()
+            except Exception:
+                pass
+            gc.collect()
             utime.sleep_ms(1000)
 
     _mqtt_connected = False
@@ -909,7 +927,7 @@ _last_mqtt_check = 0
 _discovery_queue = []
 _discovery_idx = 0
 _last_discovery_ms = 0
-_DISCOVERY_INTERVAL_MS = 200
+_DISCOVERY_INTERVAL_MS = 350
 
 
 def check_mqtt():
@@ -953,6 +971,11 @@ def check_mqtt():
             return True
         except Exception as e2:
             log_err(f"MQTT reconnect attempt {attempt + 1} failed: {e2}")
+            try:
+                client_ref.sock.close()
+            except Exception:
+                pass
+            gc.collect()
             utime.sleep_ms(2000)
 
     log_err("MQTT reconnection failed after 3 attempts")
@@ -960,6 +983,7 @@ def check_mqtt():
 
 
 def _publish_discovery_entity(client, topic, payload, log_msg):
+    gc.collect()
     client.publish(topic, ujson.dumps(payload), retain=True)
     info(log_msg)
 
@@ -1233,6 +1257,7 @@ def service_discovery():
     if _last_discovery_ms and utime.ticks_diff(now, _last_discovery_ms) < _DISCOVERY_INTERVAL_MS:
         return
 
+    gc.collect()
     topic, payload, log_msg = _discovery_queue[_discovery_idx]
     try:
         if isinstance(payload, bytes):
@@ -1243,7 +1268,6 @@ def service_discovery():
         log_err(f"Discovery publish failed for {topic}: {e}")
     _discovery_idx += 1
     _last_discovery_ms = utime.ticks_ms()
-    if (_discovery_idx & 3) == 0:
     gc.collect()
 
 
@@ -1274,13 +1298,30 @@ def apply_motion_defaults():
 
 
 def main():
+    gc.collect()
     info("=" * 48)
     info("Gauge MQTT controller starting")
+    info(f"Heap free: {gc.mem_free()} bytes")
     info("=" * 48)
 
+    gc.collect()
     connect_wifi(WIFI_SSID, WIFI_PASSWORD, force_reconnect=True)
 
+    mqtt_attempts = 0
+    while True:
+        try:
             connect_mqtt()
+            break
+        except Exception as e:
+            mqtt_attempts += 1
+            log_err(f"MQTT connect failed: {e} (attempt {mqtt_attempts})")
+            if mqtt_attempts % 3 == 0:
+                log_err("WiFi may be stale — forcing reconnect...")
+                try:
+                    connect_wifi(WIFI_SSID, WIFI_PASSWORD, force_reconnect=True)
+                except Exception as we:
+                    log_err(f"WiFi reconnect failed: {we}")
+            utime.sleep_ms(5000)
     _subscribe_all(client_ref)
     schedule_discovery()