@alteriom/painlessmesh 1.9.13 → 1.9.15

package/CHANGELOG.md

@@ -9,15 +9,47 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### Added
 
-- TBD
+### Changed
+
+### Fixed
+
+## [1.9.15] - 2025-12-19
 
 ### Changed
 
-- TBD
+- **Release Consolidation** - Consolidated v1.9.13 and v1.9.14 fixes into unified v1.9.15 release
+  - Gateway Connection Timeout fix for slow HTTP APIs (WhatsApp/CallmeBot)
+  - ESP32-C6 AsyncClient deletion spacing fix for heap corruption prevention
+  - Both fixes fully tested and production-ready
+  - Complete backward compatibility maintained
+
+## [1.9.14] - 2025-12-19
 
 ### Fixed
 
-- TBD
+- **Gateway Connection Timeout During Long HTTP Requests** - Fixed `sendToInternet()` timeouts with slow APIs like CallmeBot WhatsApp
+  - **Issue**: Mesh connections closed after 10s while HTTP requests could take 30s, causing "Request timed out" errors even when messages were successfully delivered
+  - **Fix**: Bridge now disables connection timeout during HTTP request processing, ensuring ACK delivery regardless of request duration
+  - **Impact**: WhatsApp/CallmeBot users and other slow API integrations now receive proper acknowledgments
+  - **Compatibility**: Fully backward compatible, no API changes required
+  - See GATEWAY_CONNECTION_TIMEOUT_FIX.md for detailed analysis
+- **Hard Reset on ESP32-C6 - Insufficient AsyncClient Deletion Spacing** - Fixed ESP32-C6 heap corruption crashes caused by insufficient spacing between AsyncClient deletions during network disruptions
+  - **Root Cause**: The 250ms spacing between AsyncClient deletions was marginally insufficient for ESP32-C6 hardware, which uses AsyncTCP v3.3.0+ and has different timing characteristics due to RISC-V architecture and enhanced cleanup validation
+  - **Symptom**: Device crashes with "CORRUPT HEAP: Bad head at 0x40838a24. Expected 0xabba1234 got 0x4081fae4" even with deletions spaced 264ms apart (only 14ms above minimum), particularly during TCP retries, channel changes, and sendToInternet() operations
+  - **Hardware Specific**: ESP32-C6 requires more cleanup time due to:
+    - AsyncTCP v3.3.0+ with additional validation steps
+    - RISC-V architecture vs. Xtensa (different instruction timing)
+    - Enhanced heap allocator with more aggressive corruption detection
+    - Different WiFi stack and memory management implementation
+  - **Solution**: Increased `TCP_CLIENT_DELETION_SPACING_MS` from 250ms to 500ms
+    - Provides 2x safety margin for universal ESP32 compatibility
+    - Accommodates ESP32-C6, ESP32-S2/S3, ESP32-C3, ESP32-H2, and future variants
+    - Maintains backward compatibility with ESP32/ESP8266
+    - Minimal performance impact (250ms additional delay in worst-case scenarios)
+  - **Analysis**: Crash occurred at 264ms spacing, demonstrating 250ms was too close to minimum requirement for ESP32-C6; 500ms provides adequate safety margin (89% buffer vs. observed crash)
+  - **Testing**: All test suites pass (68+ assertions in TCP/connection tests, 1000+ total assertions)
+  - **Documentation**: Added ISSUE_HARD_RESET_ESP32C6_SPACING_FIX.md with detailed ESP32-C6 analysis, architecture comparison, and AsyncTCP v3.3.0+ timing requirements
+  - **Impact**: Eliminates critical heap corruption on ESP32-C6 and provides enhanced safety for all ESP32 variants
 
 ## [1.9.13] - 2025-12-19
 

package/README.md

@@ -4,7 +4,7 @@
 
 <div align="center">
 
-**Version 1.9.13** - Latest release with HTTP 203 retry logic and AsyncClient deletion race condition fixes
+**Version 1.9.15** - Consolidated release with ESP32-C6 spacing fix and gateway timeout improvements
 
 [![CI/CD Pipeline](https://github.com/Alteriom/painlessMesh/actions/workflows/ci.yml/badge.svg)](https://github.com/Alteriom/painlessMesh/actions/workflows/ci.yml)
 [![Documentation](https://github.com/Alteriom/painlessMesh/actions/workflows/docs.yml/badge.svg)](https://github.com/Alteriom/painlessMesh/actions/workflows/docs.yml)

package/library.json

@@ -6,7 +6,7 @@
     "type": "git",
     "url": "https://github.com/Alteriom/painlessMesh"
   },
-  "version": "1.9.13",
+  "version": "1.9.15",
   "frameworks": [
     "arduino"
   ],

package/library.properties

@@ -1,5 +1,5 @@
 name=Alteriom PainlessMesh
-version=1.9.13
+version=1.9.15
 author=Coopdis,Scotty Franzyshen,Edwin van Leeuwen,Germán Martín,Maximilian Schwarz,Doanh Doanh,Alteriom
 maintainer=Alteriom
 sentence=A painless way to setup a mesh with ESP8266 and ESP32 devices with Alteriom extensions

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@alteriom/painlessmesh",
-  "version": "1.9.13",
+  "version": "1.9.15",
   "description": "painlessMesh is a user-friendly library for creating mesh networks with ESP8266 and ESP32 devices. This Alteriom fork includes additional packages for sensor data (SensorPackage), device commands (CommandPackage), and status monitoring (StatusPackage). It handles routing and network management automatically, so you can focus on your application. The library uses JSON-based messaging and syncs time across all nodes, making it ideal for coordinated behaviour like synchronized light displays or sensor networks reporting to a central node.",
   "keywords": [
     "arduino",

package/src/AlteriomPainlessMesh.h

@@ -29,10 +29,10 @@
 /**
  * @brief AlteriomPainlessMesh library version information
  */
-#define ALTERIOM_PAINLESS_MESH_VERSION "1.9.13"
+#define ALTERIOM_PAINLESS_MESH_VERSION "1.9.15"
 #define ALTERIOM_PAINLESS_MESH_VERSION_MAJOR 1
 #define ALTERIOM_PAINLESS_MESH_VERSION_MINOR 9
-#define ALTERIOM_PAINLESS_MESH_VERSION_PATCH 13
+#define ALTERIOM_PAINLESS_MESH_VERSION_PATCH 15
 
 /**
  * @brief Library description and usage information

package/src/arduino/wifi.hpp

@@ -855,8 +855,9 @@ class Mesh : public painlessmesh::Mesh<Connection> {
             Log(CONNECTION,
                 "tcpConnect(): Starting TCP connection after stabilization\n");
             AsyncClient* pConn = new AsyncClient();
-            painlessmesh::tcp::connect<Connection,
-                                       painlessmesh::Mesh<Connection>>(
+            // Use wifi::Mesh type to enable blocklist functionality
+            // This allows tcp::connect to call blockNodeAfterTCPFailure on retry exhaustion
+            painlessmesh::tcp::connect<Connection, wifi::Mesh>(
                 (*pConn), targetIP, targetPort, (*this));
           });
     } else {
@@ -864,6 +865,32 @@ class Mesh : public painlessmesh::Mesh<Connection> {
     }
   }
 
+  /**
+   * Block a node after TCP connection failure
+   * 
+   * This prevents the node from being repeatedly selected for connection attempts
+   * when its TCP server is unresponsive. The block is temporary and will expire
+   * after the configured duration.
+   * 
+   * @param ip The IP address of the failed node
+   * @param blockDurationMs Duration to block the node in milliseconds (default: 60s)
+   */
+  void blockNodeAfterTCPFailure(IPAddress ip, uint32_t blockDurationMs = painlessmesh::tcp::TCP_FAILURE_BLOCK_DURATION_MS) {
+    using namespace logger;
+    uint32_t nodeId = painlessmesh::tcp::decodeNodeIdFromIP(ip);
+    
+    if (nodeId == 0) {
+      Log(CONNECTION, "blockNodeAfterTCPFailure(): Invalid mesh IP %s, cannot block\n",
+          ip.toString().c_str());
+      return;
+    }
+    
+    Log(CONNECTION, "blockNodeAfterTCPFailure(): Blocking node %u (IP: %s) for %u ms\n",
+        nodeId, ip.toString().c_str(), blockDurationMs);
+    
+    stationScan.blockNodeAfterTCPFailure(nodeId, blockDurationMs);
+  }
+
   bool setHostname(const char* hostname) {
 #ifdef ESP8266
     return WiFi.hostname(hostname);
@@ -1246,9 +1273,16 @@ class Mesh : public painlessmesh::Mesh<Connection> {
     Log(STARTUP,
         "initBridgeStatusBroadcast(): Setting up bridge status broadcast\n");
 
+    // CRITICAL FIX: Schedule tasks with a small delay to avoid crashes when
+    // called immediately after stop()/init cycle. The delay allows the scheduler
+    // and internal task structures to stabilize before adding new tasks.
+    // This fixes the "Load access fault" Guru Meditation error that occurred
+    // when promoting to bridge role.
+    const uint32_t INIT_DELAY_MS = 100;
+
     // Register ourselves as a bridge in the knownBridges list
     // This ensures the bridge knows about itself and reports correct status
-    this->addTask([this]() {
+    this->addTask(INIT_DELAY_MS, TASK_ONCE, [this]() {
       // Check Internet connectivity: WiFi connected AND valid IP address
       bool hasInternet = (WiFi.status() == WL_CONNECTED) &&
                          (WiFi.localIP() != IPAddress(0, 0, 0, 0));
@@ -1269,12 +1303,16 @@ class Mesh : public painlessmesh::Mesh<Connection> {
     });
 
     // Create periodic task to broadcast bridge status
-    bridgeStatusTask = this->addTask(this->bridgeStatusIntervalMs, TASK_FOREVER,
-                                     [this]() { this->sendBridgeStatus(); });
+    // Schedule with delay to avoid crashes during stop/init cycle
+    this->addTask(INIT_DELAY_MS, TASK_ONCE, [this]() {
+      bridgeStatusTask = this->addTask(this->bridgeStatusIntervalMs, TASK_FOREVER,
+                                       [this]() { this->sendBridgeStatus(); });
+    });
 
     // Send immediate broadcast so nodes can discover this bridge right away
     // This ensures bridge is discoverable before the first periodic broadcast
-    this->addTask([this]() {
+    // Use slightly larger delay to allow bridge status task to be set up first
+    this->addTask(INIT_DELAY_MS + 50, TASK_ONCE, [this]() {
       Log(STARTUP, "Sending initial bridge status broadcast\n");
       this->sendBridgeStatus();
     });
@@ -2101,13 +2139,26 @@ class Mesh : public painlessmesh::Mesh<Connection> {
 
     this->callbackList.onPackage(
         protocol::GATEWAY_DATA, [this](protocol::Variant& variant,
-                                       std::shared_ptr<Connection>, uint32_t) {
+                                       std::shared_ptr<Connection> connection, uint32_t) {
           auto pkg = variant.to<gateway::GatewayDataPackage>();
 
           Log(COMMUNICATION,
               "Gateway received Internet request: msgId=%u dest=%s\n",
               pkg.messageId, pkg.destination.c_str());
 
+          // Disable connection timeout during HTTP request processing
+          // HTTP requests can take up to 30 seconds (GATEWAY_HTTP_TIMEOUT_MS)
+          // but mesh connections timeout after 10 seconds (NODE_TIMEOUT).
+          // We disable the timeout here to prevent connection drop during
+          // long-running HTTP requests. The timeout will be automatically
+          // re-enabled when the next sync packet is received.
+          if (connection) {
+            connection->timeOutTask.disable();
+            Log(COMMUNICATION,
+                "Gateway disabled connection timeout for node %u during HTTP request\n",
+                connection->nodeId);
+          }
+
           // Check Internet connectivity
           if (WiFi.status() != WL_CONNECTED) {
             sendGatewayAck(pkg, false, 0, "Gateway not connected to Internet");

package/src/painlessMesh.h

@@ -5,7 +5,7 @@
  * @file painlessMesh.h
  * @brief Main header file for Alteriom painlessMesh library
  * 
- * @version 1.9.13
+ * @version 1.9.15
  * @date 2025-12-19
  * 
  * painlessMesh is a user-friendly library for creating mesh networks with 

package/src/painlessMeshSTA.cpp

@@ -144,16 +144,72 @@ void ICACHE_FLASH_ATTR StationScan::scanComplete() {
   });
 }
 
+void ICACHE_FLASH_ATTR StationScan::blockNodeAfterTCPFailure(uint32_t nodeId, uint32_t blockDurationMs) {
+  using namespace painlessmesh::logger;
+  uint32_t blockUntil = millis() + blockDurationMs;
+  tcpFailureBlocklist[nodeId] = blockUntil;
+  Log(CONNECTION, "blockNodeAfterTCPFailure(): Node %u blocked until %u (duration: %u ms)\n",
+      nodeId, blockUntil, blockDurationMs);
+}
+
+bool ICACHE_FLASH_ATTR StationScan::isNodeBlocked(uint32_t nodeId) const {
+  auto it = tcpFailureBlocklist.find(nodeId);
+  if (it == tcpFailureBlocklist.end()) {
+    return false;  // Not in blocklist
+  }
+  
+  uint32_t now = millis();
+  uint32_t blockUntil = it->second;
+  
+  // Handle millis() rollover using signed arithmetic
+  // If blockUntil - now is positive and < MILLIS_ROLLOVER_THRESHOLD, the block is still active
+  int32_t timeRemaining = (int32_t)(blockUntil - now);
+  return (timeRemaining > 0 && timeRemaining < MILLIS_ROLLOVER_THRESHOLD);
+}
+
+void ICACHE_FLASH_ATTR StationScan::cleanupBlocklist() {
+  using namespace painlessmesh::logger;
+  uint32_t now = millis();
+  
+  auto it = tcpFailureBlocklist.begin();
+  while (it != tcpFailureBlocklist.end()) {
+    uint32_t blockUntil = it->second;
+    int32_t timeRemaining = (int32_t)(blockUntil - now);
+    
+    // Remove expired entries (timeRemaining <= 0 or in far future due to rollover)
+    if (timeRemaining <= 0 || timeRemaining >= MILLIS_ROLLOVER_THRESHOLD) {
+      Log(CONNECTION, "cleanupBlocklist(): Removing expired entry for node %u\n", it->first);
+      it = tcpFailureBlocklist.erase(it);
+    } else {
+      ++it;
+    }
+  }
+}
+
 void ICACHE_FLASH_ATTR StationScan::filterAPs() {
+  // First, clean up expired blocklist entries
+  cleanupBlocklist();
+  
   auto ap = aps.begin();
   while (ap != aps.end()) {
     auto apNodeId = painlessmesh::tcp::encodeNodeId(ap->bssid);
+    
+    // Filter out nodes we're already connected to
     if (painlessmesh::router::findRoute<painlessmesh::Connection>(
             (*mesh), apNodeId) != NULL) {
       ap = aps.erase(ap);
-    } else {
-      ap++;
+      continue;
+    }
+    
+    // Filter out nodes that are temporarily blocked due to TCP failures
+    if (isNodeBlocked(apNodeId)) {
+      using namespace painlessmesh::logger;
+      Log(CONNECTION, "filterAPs(): Skipping blocked node %u (TCP server unresponsive)\n", apNodeId);
+      ap = aps.erase(ap);
+      continue;
     }
+    
+    ap++;
   }
 }
 

package/src/painlessMeshSTA.h

@@ -6,6 +6,7 @@
 #include "painlessmesh/mesh.hpp"
 
 #include <list>
+#include <map>
 
 typedef struct {
   uint8_t bssid[6];
@@ -13,6 +14,12 @@ typedef struct {
   int8_t rssi;
 } WiFi_AP_Record_t;
 
+// Entry for tracking nodes where TCP connection failed
+typedef struct {
+  uint32_t nodeId;
+  uint32_t blockUntil;  // millis() timestamp when this node can be retried
+} TCPFailureBlocklistEntry;
+
 class StationScan {
  public:
   Task task;  // Station scanning for connections
@@ -46,6 +53,16 @@ class StationScan {
   uint16_t getConsecutiveEmptyScans() const {
     return consecutiveEmptyScans;
   }
+  
+  // Add a node to the TCP failure blocklist
+  // This prevents repeated connection attempts to nodes where TCP server is unresponsive
+  void blockNodeAfterTCPFailure(uint32_t nodeId, uint32_t blockDurationMs = 60000);
+  
+  // Check if a node is currently blocked due to TCP failures
+  bool isNodeBlocked(uint32_t nodeId) const;
+  
+  // Clean up expired entries from the blocklist
+  void cleanupBlocklist();
 
   /// Valid APs found during the last scan
   std::list<WiFi_AP_Record_t> lastAPs;
@@ -68,6 +85,14 @@ class StationScan {
   // Track consecutive scans with no mesh nodes found (for channel re-detection)
   uint16_t consecutiveEmptyScans = 0;
   static const uint16_t EMPTY_SCAN_THRESHOLD = 6; // ~30 seconds at default SCAN_INTERVAL
+  
+  // TCP failure blocklist to prevent infinite retry loops
+  // Maps nodeId -> blockUntil timestamp (millis())
+  std::map<uint32_t, uint32_t> tcpFailureBlocklist;
+  
+  // Threshold for detecting millis() rollover in time comparisons
+  // Using 2^30 (~12 days) as reasonable limit - any time difference larger is likely rollover
+  static constexpr int32_t MILLIS_ROLLOVER_THRESHOLD = (int32_t)(1U << 30);
 
   friend painlessMesh;
 };

package/src/painlessmesh/connection.hpp

@@ -26,7 +26,9 @@ static const uint32_t TCP_CLIENT_CLEANUP_DELAY_MS = 1000; // 1000ms delay before
 // When multiple AsyncClients are deleted in rapid succession, the AsyncTCP library's
 // internal cleanup routines can interfere with each other, causing heap corruption
 // This spacing ensures each deletion completes before the next one begins
-static const uint32_t TCP_CLIENT_DELETION_SPACING_MS = 250; // 250ms spacing between deletions
+// Increased from 250ms to 500ms to support ESP32-C6 and other ESP32 variants which
+// require more time for AsyncTCP internal cleanup operations
+static const uint32_t TCP_CLIENT_DELETION_SPACING_MS = 500; // 500ms spacing between deletions
 
 // Global state to track AsyncClient deletion scheduling and execution
 // This ensures deletions are spaced out even when multiple deletion requests arrive simultaneously

package/src/painlessmesh/tcp.hpp

@@ -26,18 +26,49 @@ static const uint32_t TCP_CONNECT_STABILIZATION_DELAY_MS = 500; // Delay after I
 // Gives the TCP server more time to recover and reduces network congestion
 static const uint32_t TCP_EXHAUSTION_RECONNECT_DELAY_MS = 10000; // 10 seconds before reconnection
 
+// Duration to block a node after TCP connection retry exhaustion (60 seconds)
+// This prevents repeatedly trying to connect to nodes with unresponsive TCP servers
+static const uint32_t TCP_FAILURE_BLOCK_DURATION_MS = 60000;
+
 inline uint32_t encodeNodeId(const uint8_t *hwaddr) {
   using namespace painlessmesh::logger;
   Log(GENERAL, "encodeNodeId():\n");
   uint32_t value = 0;
 
-  value |= hwaddr[2] << 24;  // Big endian (aka "network order"):
-  value |= hwaddr[3] << 16;
-  value |= hwaddr[4] << 8;
-  value |= hwaddr[5];
+  // Extract last 4 bytes of MAC address (skip first 2 bytes)
+  // Big endian (aka "network order") encoding
+  value |= hwaddr[2] << 24;  // Byte 2 -> bits 31-24
+  value |= hwaddr[3] << 16;  // Byte 3 -> bits 23-16
+  value |= hwaddr[4] << 8;   // Byte 4 -> bits 15-8
+  value |= hwaddr[5];        // Byte 5 -> bits 7-0
   return value;
 }
 
+// Decode nodeId from mesh IP address
+// Mesh IPs follow format: 10.(nodeId >> 8).(nodeId & 0xFF).1
+inline uint32_t decodeNodeIdFromIP(IPAddress ip) {
+#if defined(PAINLESSMESH_BOOST)
+  // In test environment, IPAddress doesn't support indexing
+  // Return 0 to indicate invalid/unknown node ID
+  (void)ip;  // Suppress unused parameter warning
+  return 0;
+#else
+  // Validate mesh network IP format: 10.x.x.1
+  // First octet must be 10, last octet must be 1
+  const uint8_t MESH_IP_FIRST_OCTET = 10;
+  const uint8_t MESH_IP_LAST_OCTET = 1;
+  if (ip[0] != MESH_IP_FIRST_OCTET || ip[3] != MESH_IP_LAST_OCTET) {
+    return 0;  // Invalid mesh IP
+  }
+  
+  // Extract nodeId from second and third octets
+  // NodeId = (octet2 << 8) | octet3
+  const uint8_t BYTE_SHIFT = 8;  // Bits per byte
+  uint32_t nodeId = ((uint32_t)ip[1] << BYTE_SHIFT) | (uint32_t)ip[2];
+  return nodeId;
+#endif
+}
+
 template <class T, class M>
 void initServer(AsyncServer &server, M &mesh) {
   using namespace logger;
@@ -146,8 +177,24 @@ void connect(AsyncClient &client, IPAddress ip, uint16_t port, M &mesh,
       // All retries exhausted - schedule delayed reconnection
       // Adding a significant delay before reconnection prevents rapid reconnection loops
       // when the TCP server is persistently unavailable or overloaded
-      Log(CONNECTION, "tcp_err(): All %d retries exhausted, scheduling WiFi reconnection in %u ms\n",
-          TCP_CONNECT_MAX_RETRIES + 1, TCP_EXHAUSTION_RECONNECT_DELAY_MS);
+      Log(CONNECTION, "tcp_err(): All %d retries exhausted for IP %s\n",
+          TCP_CONNECT_MAX_RETRIES + 1, ip.toString().c_str());
+      
+      // Block this node temporarily to prevent immediate reconnection to the same unresponsive node
+      // This helps when the bridge's TCP server is down but WiFi AP is still running
+      // The blocklist is only used during AP filtering, so it won't affect existing connections
+      #if !defined(PAINLESSMESH_BOOST)
+      // Try to decode nodeId from IP and block it
+      // Only works for mesh IPs (format: 10.x.x.1)
+      uint32_t failedNodeId = decodeNodeIdFromIP(ip);
+      if (failedNodeId != 0) {
+        // Note: This requires M to be wifi::Mesh which has blockNodeAfterTCPFailure
+        mesh.blockNodeAfterTCPFailure(ip, TCP_FAILURE_BLOCK_DURATION_MS);
+      }
+      #endif
+      
+      Log(CONNECTION, "tcp_err(): Scheduling WiFi reconnection in %u ms\n",
+          TCP_EXHAUSTION_RECONNECT_DELAY_MS);
       
       // Defer deletion of the failed AsyncClient to prevent heap corruption
       // Use the centralized deletion scheduler to ensure proper spacing between deletions