From d31c6ced415b5a1390b976ac52ca8b3a9500b2c1 Mon Sep 17 00:00:00 2001
From: Sergey Polyakov <sergeuz@gmail.com>
Date: Fri, 16 Feb 2024 17:33:57 +0200
Subject: [PATCH 1/6] Remove the unused active object queue implementation

---
 system/inc/active_object.h |  36 --
 system/inc/channel.h       | 852 -------------------------------------
 2 files changed, 888 deletions(-)
 delete mode 100644 system/inc/channel.h

diff --git a/system/inc/active_object.h b/system/inc/active_object.h
index af2af4e90f..2752fca1af 100644
--- a/system/inc/active_object.h
+++ b/system/inc/active_object.h
@@ -29,7 +29,6 @@
 #include <future>
 #include <cstring>
 
-#include "channel.h"
 #include "concurrent_hal.h"
 #include "hal_platform.h"
 
@@ -344,41 +343,6 @@ class ActiveObjectBase
 
 };
 
-
-template <size_t queue_size=50>
-class ActiveObjectChannel : public ActiveObjectBase
-{
-    cpp::channel<Item, queue_size> _channel;
-
-protected:
-
-    virtual bool take(Item& item) override
-    {
-        return cpp::select().recv_only(_channel, item).try_once();
-    }
-
-    virtual bool put(const Item& item) override
-    {
-        _channel.send(item);
-        return true;
-    }
-
-
-public:
-
-    ActiveObjectChannel(ActiveObjectConfiguration& config) : ActiveObjectBase(config) {}
-
-    /**
-     * Start the asynchronous processing for this active object.
-     */
-    void start()
-    {
-        _channel = cpp::channel<Item*, queue_size>();
-        start_thread();
-    }
-
-};
-
 class ActiveObjectQueue : public ActiveObjectBase
 {
 protected:
diff --git a/system/inc/channel.h b/system/inc/channel.h
deleted file mode 100644
index 6e3146a41d..0000000000
--- a/system/inc/channel.h
+++ /dev/null
@@ -1,852 +0,0 @@
-// Copyright 2014, Alex Horn. All rights reserved.
-// Use of this source code is governed by a BSD-style
-// license that can be found in the LICENSE file.
-
-#ifndef CPP_CHANNEL_H
-#define CPP_CHANNEL_H
-
-#include <deque>
-#include <mutex>
-#include <vector>
-#include <limits>
-#include <random>
-#include <memory>
-#include <thread>
-#include <cstddef>
-#include <cassert>
-#include <functional>
-#include <type_traits>
-#include <condition_variable>
-
-namespace cpp
-{
-
-namespace internal
-{
-
-#if __cplusplus <= 201103L
-// since C++14 in std, see Herb Sutter's blog
-template<class T, class ...Args>
-std::unique_ptr<T> make_unique(Args&& ...args)
-{
-  return std::unique_ptr<T>(new T(std::forward<Args>(args)...));
-}
-#else
-  using std::make_unique;
-#endif
-
-template<class T>
-struct _is_exception_safe :
-  std::integral_constant<bool,
-    std::is_nothrow_copy_constructible<T>::value or
-    std::is_nothrow_move_constructible<T>::value>
-{};
-
-// Note that currently handshakes between send/receives inside selects
-// have higher priority compared to sends/receives outside selects.
-
-// TODO: investigate and ideally also discuss other scheduling algorithms
-template<class T, std::size_t N>
-class _channel
-{
-static_assert(N < std::numeric_limits<std::size_t>::max(),
-  "N must be strictly less than the largest possible size_t value");
-
-private:
-  std::mutex m_mutex;
-  std::condition_variable m_send_begin_cv;
-  std::condition_variable m_send_end_cv;
-  std::condition_variable m_recv_cv;
-
-  // FIFO order
-  std::deque<std::pair</* sender */ std::thread::id, T>> m_queue;
-
-  bool m_is_send_done;
-  bool m_is_try_send_done;
-  bool m_is_recv_ready;
-  bool m_is_try_send_ready;
-  bool m_is_try_recv_ready;
-
-  bool is_full() const
-  {
-    return m_queue.size() > N;
-  }
-
-  // Is nonblocking receive and nonblocking send ready for handshake?
-  bool is_try_ready() const
-  {
-    return m_is_try_recv_ready && m_is_try_send_ready;
-  }
-
-  // Block calling thread until queue becomes nonempty. While waiting
-  // (i.e. queue is empty), give try_send() a chance to succeed.
-  //
-  // \pre: calling thread owns lock
-  // \post: queue is nonempty and calling thread still owns lock
-  void _pre_blocking_recv(std::unique_lock<std::mutex>& lock)
-  {
-    m_is_recv_ready = true;
-    m_recv_cv.wait(lock, [this]{ return !m_queue.empty(); });
-
-    // TODO: support the case where both ends of a channel are inside a select
-    assert(!is_try_ready());
-  }
-
-  // Pop front of queue and unblock one _send() (if any)
-  //
-  // \pre: calling thread must own lock and queue is nonempty
-  // \post: calling thread doesn't own lock anymore, and protocol with
-  //    try_send() and try_recv() is fulfilled
-  void _post_blocking_recv(std::unique_lock<std::mutex>& lock)
-  {
-    // If queue is full, then there exists either a _send() waiting
-    // for m_send_end_cv, or try_send() has just enqueued an element.
-    //
-    // In general, the converse is false: if there exists a blocking send,
-    // then a nonblocking receive might have just dequeued an element,
-    // i.e. queue is not full.
-    assert(!is_full() || !m_is_send_done || !m_is_try_send_done);
-
-    // blocking and nonblocking send can never occur simultaneously
-    assert(m_is_try_send_done || m_is_send_done);
-
-    m_queue.pop_front();
-    assert(!is_full());
-
-    // protocol with nonblocking calls
-    m_is_try_send_done = true;
-    m_is_recv_ready = false;
-    m_is_try_recv_ready = false;
-    m_is_try_send_ready = false;
-
-    // Consider two concurrent _send() calls denoted by s and s'.
-    //
-    // Suppose s is waiting to enqueue an element (i.e. m_send_begin_cv),
-    // whereas s' is waiting for an acknowledgment (i.e. m_send_end_cv)
-    // that its previously enqueued element has been dequeued. Since s'
-    // is waiting and the flag m_is_send_done is only modified by _send(),
-    // m_is_send_done is false. Hence, we notify m_send_end_cv. This
-    // causes s' to notify s, thereby allowing s to proceed if possible.
-    //
-    // Now suppose there is no such s' (say, due to the fact that the
-    // queue never became full). Then, m_is_send_done == true. Thus,
-    // m_send_begin_cv is notified and s proceeds if possible. Note
-    // that if we hadn't notified s this way, then it could deadlock
-    // in case that it waited on m_is_try_send_done to become true.
-    if (m_is_send_done)
-    {
-      // unlock before notifying threads; otherwise, the
-      // notified thread would unnecessarily block again
-      lock.unlock();
-
-      // nonblocking, see also note below about notifications
-      m_send_begin_cv.notify_one();
-    }
-    else
-    {
-      // unlock before notifying threads; otherwise, the
-      // notified thread would unnecessarily block again
-      lock.unlock();
-
-      // we rely on the following semantics of notify_one():
-      //
-      // if a notification n is issued to s (i.e. s is chosen from among
-      // a set of threads waiting on a condition variable associated with
-      // mutex m) but another thread t locks m before s wakes up (i.e. s
-      // has not owned yet m after n had been issued), then n is retried
-      // as soon as t unlocks m. The retries repeat until n arrives at s
-      // meaning that s actually owns m and checks its wait guard.
-      m_send_end_cv.notify_one();
-    }
-  }
-
-  template<class U>
-  void _send(U&&);
-
-public:
-  // \pre: calling thread must own mutex()
-  // \post: calling thread doesn't own mutex() anymore
-  template<class U>
-  bool try_send(std::unique_lock<std::mutex>&, U&&);
-
-  // \pre: calling thread must own mutex()
-  // \post: calling thread doesn't own mutex() anymore
-  std::pair<bool, std::unique_ptr<T>> try_recv_ptr(
-    std::unique_lock<std::mutex>&);
-
-  _channel(const _channel&) = delete;
-
-  // Propagates exceptions thrown by std::condition_variable constructor
-  _channel()
-  : m_mutex(),
-    m_send_begin_cv(),
-    m_send_end_cv(),
-    m_recv_cv(),
-    m_queue(),
-    m_is_send_done(true),
-    m_is_try_send_done(true),
-    m_is_recv_ready(false),
-    m_is_try_send_ready(false),
-    m_is_try_recv_ready(false) {}
-
-  // channel lock
-  std::mutex& mutex()
-  {
-    return m_mutex;
-  }
-
-  // Propagates exceptions thrown by std::condition_variable::wait()
-  void send(const T& t)
-  {
-    _send(t);
-  }
-
-  // Propagates exceptions thrown by std::condition_variable::wait()
-  void send(T&& t)
-  {
-    _send(std::move(t));
-  }
-
-  // Propagates exceptions thrown by std::condition_variable::wait()
-  T recv();
-
-  // Propagates exceptions thrown by std::condition_variable::wait()
-  void recv(T&);
-
-  // Propagates exceptions thrown by std::condition_variable::wait()
-  std::unique_ptr<T> recv_ptr();
-};
-
-}
-
-template<class T, std::size_t N> class ichannel;
-template<class T, std::size_t N> class ochannel;
-
-/// Go-style concurrency
-
-/// Thread synchronization mechanism as in the Go language.
-/// As in Go, cpp::channel<T, N> are first-class values.
-///
-/// Unlike Go, however, cpp::channels<T, N> cannot be nil
-/// not closed. This simplifies the usage of the library.
-///
-/// The template arguments are as follows:
-///
-/// * T -- type of data to be communicated over channel
-/// * N is zero -- synchronous channel
-/// * N is positive -- asynchronous channel with queue size N
-///
-/// Note that cpp::channel<T, N>::recv() is only supported if T is
-/// exception safe. This is automatically checked at compile time.
-/// If T is not exception safe, use any of the other receive member
-/// functions.
-///
-/// \see http://golang.org/ref/spec#Channel_types
-/// \see http://golang.org/ref/spec#Send_statements
-/// \see http://golang.org/ref/spec#Receive_operator
-/// \see http://golang.org/doc/effective_go.html#chan_of_chan
-template<class T, std::size_t N = 0>
-class channel
-{
-static_assert(N < std::numeric_limits<std::size_t>::max(),
-  "N must be strictly less than the largest possible size_t value");
-
-private:
-  friend class ichannel<T, N>;
-  friend class ochannel<T, N>;
-
-  std::shared_ptr<internal::_channel<T, N>> m_channel_ptr;
-
-public:
-  channel(const channel& other) noexcept
-  : m_channel_ptr(other.m_channel_ptr) {}
-
-  // Propagates exceptions thrown by std::condition_variable constructor
-  channel()
-  : m_channel_ptr(std::make_shared<internal::_channel<T, N>>()) {}
-
-  channel& operator=(const channel& other) noexcept
-  {
-    m_channel_ptr = other.m_channel_ptr;
-    return *this;
-  }
-
-  bool operator==(const channel& other) const noexcept
-  {
-    return m_channel_ptr == other.m_channel_ptr;
-  }
-
-  bool operator!=(const channel& other) const noexcept
-  {
-    return m_channel_ptr != other.m_channel_ptr;
-  }
-
-  bool operator==(const ichannel<T, N>&) const noexcept;
-  bool operator!=(const ichannel<T, N>&) const noexcept;
-
-  bool operator==(const ochannel<T, N>&) const noexcept;
-  bool operator!=(const ochannel<T, N>&) const noexcept;
-
-  // Propagates exceptions thrown by std::condition_variable::wait()
-  void send(const T& t)
-  {
-    m_channel_ptr->send(t);
-  }
-
-  // Propagates exceptions thrown by std::condition_variable::wait()
-  void send(T&& t)
-  {
-    m_channel_ptr->send(std::move(t));
-  }
-
-  // Propagates exceptions thrown by std::condition_variable::wait()
-  T recv()
-  {
-    static_assert(internal::_is_exception_safe<T>::value,
-      "Cannot guarantee exception safety, use another recv operator");
-
-    return m_channel_ptr->recv();
-  }
-
-  // Propagates exceptions thrown by std::condition_variable::wait()
-  std::unique_ptr<T> recv_ptr()
-  {
-    return m_channel_ptr->recv_ptr();
-  }
-
-  // Propagates exceptions thrown by std::condition_variable::wait()
-  void recv(T& t)
-  {
-    m_channel_ptr->recv(t);
-  }
-};
-
-class select;
-
-/// Can only be used to receive elements of type T
-template<class T, std::size_t N = 0>
-class ichannel
-{
-private:
-  friend class select;
-  friend class channel<T, N>;
-  std::shared_ptr<internal::_channel<T, N>> m_channel_ptr;
-
-public:
-  ichannel(const channel<T, N>& other) noexcept
-  : m_channel_ptr(other.m_channel_ptr) {}
-
-  ichannel(const ichannel& other) noexcept
-  : m_channel_ptr(other.m_channel_ptr) {}
-
-  ichannel(ichannel&& other) noexcept
-  : m_channel_ptr(std::move(other.m_channel_ptr)) {}
-
-  ichannel& operator=(const ichannel& other) noexcept
-  {
-    m_channel_ptr = other.m_channel_ptr;
-    return *this;
-  }
-
-  bool operator==(const ichannel& other) const noexcept
-  {
-    return m_channel_ptr == other.m_channel_ptr;
-  }
-
-  bool operator!=(const ichannel& other) const noexcept
-  {
-    return m_channel_ptr != other.m_channel_ptr;
-  }
-
-  // Propagates exceptions thrown by std::condition_variable::wait()
-  T recv()
-  {
-    static_assert(internal::_is_exception_safe<T>::value,
-      "Cannot guarantee exception safety, use another recv operator");
-
-    return m_channel_ptr->recv();
-  }
-
-  // Propagates exceptions thrown by std::condition_variable::wait()
-  void recv(T& t)
-  {
-    m_channel_ptr->recv(t);
-  }
-
-  // Propagates exceptions thrown by std::condition_variable::wait()
-  std::unique_ptr<T> recv_ptr()
-  {
-    return m_channel_ptr->recv_ptr();
-  }
-
-};
-
-/// Can only be used to send elements of type T
-template<class T, std::size_t N = 0>
-class ochannel
-{
-private:
-  friend class select;
-  friend class channel<T, N>;
-  std::shared_ptr<internal::_channel<T, N>> m_channel_ptr;
-
-public:
-  ochannel(const channel<T, N>& other) noexcept
-  : m_channel_ptr(other.m_channel_ptr) {}
-
-  ochannel(const ochannel& other) noexcept
-  : m_channel_ptr(other.m_channel_ptr) {}
-
-  ochannel(ochannel&& other) noexcept
-  : m_channel_ptr(std::move(other.m_channel_ptr)) {}
-
-  ochannel& operator=(const ochannel& other) noexcept
-  {
-    m_channel_ptr = other.m_channel_ptr;
-    return *this;
-  }
-
-  bool operator==(const ochannel& other) const noexcept
-  {
-    return m_channel_ptr == other.m_channel_ptr;
-  }
-
-  bool operator!=(const ochannel& other) const noexcept
-  {
-    return m_channel_ptr != other.m_channel_ptr;
-  }
-
-  // Propagates exceptions thrown by std::condition_variable::wait()
-  void send(const T& t)
-  {
-    m_channel_ptr->send(t);
-  }
-
-  // Propagates exceptions thrown by std::condition_variable::wait()
-  void send(T&& t)
-  {
-    m_channel_ptr->send(std::move(t));
-  }
-};
-
-/// Go's select statement
-
-/// \see http://golang.org/ref/spec#Select_statements
-///
-/// \warning select objects must not be shared between threads
-///
-// TODO: investigate and ideally discuss pseudo-random distribution
-class select
-{
-private:
-  template<class T, std::size_t N, class NullaryFunction>
-  class try_send_nullary
-  {
-  private:
-    template<class U, class V = typename std::remove_reference<U>::type>
-    static bool _run(ochannel<V, N>& c, U&& u, NullaryFunction f)
-    {
-      internal::_channel<V, N>& _c = *c.m_channel_ptr;
-      std::unique_lock<std::mutex> lock(_c.mutex(), std::defer_lock);
-      if (lock.try_lock() && _c.try_send(lock, std::forward<U>(u)))
-      {
-        assert(!lock.owns_lock());
-        f();
-        return true;
-      }
-
-      return false;
-    }
-
-  public:
-    bool operator()(ochannel<T, N>& c, const T& t, NullaryFunction f)
-    {
-      return _run(c, t, f);
-    }
-
-    bool operator()(ochannel<T, N>& c, T&& t, NullaryFunction f)
-    {
-      return _run(c, std::move(t), f);
-    }
-  };
-
-  template<class T, std::size_t N, class NullaryFunction>
-  struct try_recv_nullary
-  {
-    bool operator()(ichannel<T, N>& c, T& t, NullaryFunction f)
-    {
-      internal::_channel<T, N>& _c = *c.m_channel_ptr;
-      std::unique_lock<std::mutex> lock(_c.mutex(), std::defer_lock);
-      if (lock.try_lock())
-      {
-        std::pair<bool, std::unique_ptr<T>> pair = _c.try_recv_ptr(lock);
-        if (pair.first)
-        {
-          assert(!lock.owns_lock());
-          t = *pair.second;
-          f();
-          return true;
-        }
-      }
-
-      return false;
-    }
-  };
-
-  template<class T, std::size_t N, class UnaryFunction>
-  struct try_recv_unary
-  {
-    bool operator()(ichannel<T, N>& c, UnaryFunction f)
-    {
-      internal::_channel<T, N>& _c = *c.m_channel_ptr;
-      std::unique_lock<std::mutex> lock(_c.mutex(), std::defer_lock);
-      if (lock.try_lock())
-      {
-        std::pair<bool, std::unique_ptr<T>> pair = _c.try_recv_ptr(lock);
-        if (pair.first)
-        {
-          assert(!lock.owns_lock());
-          f(std::move(*pair.second));
-          return true;
-        }
-      }
-
-      return false;
-    }
-  };
-
-  typedef std::function<bool()> try_function;
-  typedef std::vector<try_function> try_functions;
-  try_functions m_try_functions;
-
-  std::mt19937 random_gen;
-
-public:
-  select()
-  : m_try_functions(),
-    random_gen(std::time(nullptr)) {}
-
-  /* send cases */
-
-  template<class T, std::size_t N,
-    class U = typename std::remove_reference<T>::type>
-  select& send_only(channel<U, N> c, T&& t)
-  {
-    return send_only(ochannel<U, N>(c), std::forward<T>(t));
-  }
-
-  template<class T, std::size_t N,
-    class U = typename std::remove_reference<T>::type>
-  select& send_only(ochannel<U, N> c, T&& t)
-  {
-    return send(c, std::forward<T>(t), [](){ /* skip */ });
-  }
-
-  template<class T, std::size_t N, class NullaryFunction,
-    class U = typename std::remove_reference<T>::type>
-  select& send(channel<U, N> c, T&& t, NullaryFunction f)
-  {
-    return send(ochannel<U, N>(c), std::forward<T>(t), f);
-  }
-
-  template<class T, std::size_t N, class NullaryFunction,
-    class U = typename std::remove_reference<T>::type>
-  select& send(ochannel<U, N> c, T&& t, NullaryFunction f)
-  {
-    m_try_functions.push_back(std::bind(
-      try_send_nullary<U, N, NullaryFunction>(), c, std::forward<T>(t), f));
-    return *this;
-  }
-
-  /* receive cases */
-
-  template<class T, std::size_t N>
-  select& recv_only(channel<T, N> c, T& t)
-  {
-    return recv_only(ichannel<T, N>(c), t);
-  }
-
-  template<class T, std::size_t N>
-  select& recv_only(ichannel<T, N> c, T& t)
-  {
-    return recv(c, t, [](){ /* skip */ });
-  }
-
-  template<class T, std::size_t N, class NullaryFunction>
-  select& recv(channel<T, N> c, T& t, NullaryFunction f)
-  {
-    return recv(ichannel<T, N>(c), t, f);
-  }
-
-  template<class T, std::size_t N, class NullaryFunction>
-  select& recv(ichannel<T, N> c, T& t, NullaryFunction f)
-  {
-    m_try_functions.push_back(std::bind(
-      try_recv_nullary<T, N, NullaryFunction>(), c, std::ref(t), f));
-    return *this;
-  }
-
-  template<class T, std::size_t N, class UnaryFunction>
-  select& recv(channel<T, N> c, UnaryFunction f)
-  {
-    return recv(ichannel<T, N>(c), f);
-  }
-
-  template<class T, std::size_t N, class UnaryFunction>
-  select& recv(ichannel<T, N> c, UnaryFunction f)
-  {
-    m_try_functions.push_back(std::bind(
-      try_recv_unary<T, N, UnaryFunction>(), c, f));
-    return *this;
-  }
-
-  /// Nonblocking like Go's select statement with default case
-
-  /// Returns true if and only if exactly one case succeeded
-  bool try_once()
-  {
-    const try_functions::size_type n = m_try_functions.size(), i = random_gen();
-    for(try_functions::size_type j = 0; j < n; j++)
-    {
-      if (m_try_functions.at((i + j) % n)())
-        return true;
-    }
-    return false;
-  }
-
-  void wait()
-  {
-    const try_functions::size_type n = m_try_functions.size();
-    try_functions::size_type i = random_gen();
-    for(;;)
-    {
-      i = (i + 1) % n;
-      if (m_try_functions.at(i)())
-        break;
-    }
-  }
-
-  // Propagates any exception thrown by std::this_thread::sleep_for
-  template<class Rep, class Period>
-  void wait(const std::chrono::duration<Rep, Period>& sleep)
-  {
-    const try_functions::size_type n = m_try_functions.size();
-    try_functions::size_type i = random_gen();
-    for(;;)
-    {
-      i = (i + 1) % n;
-      if (m_try_functions.at(i)())
-        break;
-
-      std::this_thread::sleep_for(sleep);
-    }
-  }
-};
-
-template<class T, std::size_t N>
-bool channel<T, N>::operator==(const ichannel<T, N>& other) const noexcept
-{
-  return m_channel_ptr == other.m_channel_ptr;
-}
-
-template<class T, std::size_t N>
-bool channel<T, N>::operator!=(const ichannel<T, N>& other) const noexcept
-{
-  return m_channel_ptr != other.m_channel_ptr;
-}
-
-template<class T, std::size_t N>
-bool channel<T, N>::operator==(const ochannel<T, N>& other) const noexcept
-{
-  return m_channel_ptr == other.m_channel_ptr;
-}
-
-template<class T, std::size_t N>
-bool channel<T, N>::operator!=(const ochannel<T, N>& other) const noexcept
-{
-  return m_channel_ptr != other.m_channel_ptr;
-}
-
-template<class T, std::size_t N>
-template<class U>
-bool internal::_channel<T, N>::try_send(
-  std::unique_lock<std::mutex>& lock, U&& u)
-{
-  m_is_try_send_ready = true;
-
-  // TODO: support the case where both ends of a channel are inside a select
-  assert(!is_try_ready());
-
-  if ((!m_is_send_done || !m_is_try_send_done || is_full() ||
-       (0 == N - m_queue.size() && !m_is_recv_ready)))
-  {
-    // TODO: investigate potential LLVM libc++ RAII unlocking issue
-    lock.unlock();
-    return false;
-  }
-
-  assert(m_is_send_done);
-  assert(m_is_try_send_done);
-  assert(!is_full());
-
-  // if enqueue should block, there must be a receiver waiting
-  m_is_try_send_done = 0 < N - m_queue.size();
-  assert(m_is_try_send_done || m_is_recv_ready);
-
-  m_queue.emplace_back(std::this_thread::get_id(), std::forward<U>(u));
-
-  // Let v be the value enqueued by try_send(). If m_is_try_send_done
-  // is false, no other sender (whether blocking or not) can enqueue a
-  // value until a receiver has dequeued v, thereby ensuring the channel
-  // FIFO order when the queue is filled up by try_send(). Moreover, in
-  // that case, since m_is_try_send_done implies m_is_recv_ready, such a
-  // receiver is guaranteed to exist, and it will reset m_is_try_send_done
-  // to true so that other senders can make progress after v has been
-  // dequeued. And by notifying m_recv_cv, other receivers waiting for
-  // the queue to become nonempty can make progress as well.
-  lock.unlock();
-  m_recv_cv.notify_one();
-  return true;
-}
-
-template<class T, std::size_t N>
-std::pair<bool, std::unique_ptr<T>> internal::_channel<T, N>::try_recv_ptr(
-  std::unique_lock<std::mutex>& lock)
-{
-  m_is_try_recv_ready = true;
-
-  if (m_queue.empty())
-    return std::make_pair(false, std::unique_ptr<T>(nullptr));
-
-  // If queue is full, then there exists either a _send() waiting
-  // for m_send_end_cv, or try_send() has just enqueued an element.
-  //
-  // In general, the converse is false: if there exists a blocking send,
-  // then a nonblocking receive might have just dequeued an element,
-  // i.e. queue is not full.
-  assert(!is_full() || !m_is_send_done || !m_is_try_send_done);
-
-  // blocking and nonblocking send can never occur simultaneously
-  assert(m_is_try_send_done || m_is_send_done);
-
-  std::pair<std::thread::id, T> pair(std::move(m_queue.front()));
-  assert(!is_full() || std::this_thread::get_id() != pair.first);
-
-  // move/copy before pop_front() to ensure strong exception safety
-  std::unique_ptr<T> t_ptr(make_unique<T>(std::move(pair.second)));
-
-  m_queue.pop_front();
-  assert(!is_full());
-
-  // protocol with nonblocking calls
-  m_is_try_send_done = true;
-  m_is_try_recv_ready = false;
-  m_is_try_send_ready = false;
-
-  // see also explanation in _channel::_post_blocking_recv()
-  if (m_is_send_done)
-  {
-    lock.unlock();
-    m_send_begin_cv.notify_one();
-  }
-  else
-  {
-    lock.unlock();
-    m_send_end_cv.notify_one();
-  }
-
-  return std::make_pair(true, std::move(t_ptr));
-}
-
-template<class T, std::size_t N>
-template<class U>
-void internal::_channel<T, N>::_send(U&& u)
-{
-  // unlock before notifying threads; otherwise, the
-  // notified thread would unnecessarily block again
-  {
-    // wait (if necessary) until queue is no longer full and any
-    // previous _send() has successfully enqueued element
-    std::unique_lock<std::mutex> lock(m_mutex);
-    m_send_begin_cv.wait(lock, [this]{ return m_is_send_done &&
-      m_is_try_send_done && !is_full(); });
-
-    assert(m_is_send_done);
-    assert(m_is_try_send_done);
-    assert(!is_full());
-
-    // TODO: support the case where both ends of a channel are inside a select
-    assert(!is_try_ready());
-
-    m_queue.emplace_back(std::this_thread::get_id(), std::forward<U>(u));
-    m_is_send_done = false;
-  }
-
-  // nonblocking
-  m_recv_cv.notify_one();
-
-  // wait (if necessary) until u has been received by another thread
-  {
-    std::unique_lock<std::mutex> lock(m_mutex);
-
-    // It is enough to check !is_full() because m_is_send_done == false.
-    // Therefore, no other thread could have caused the queue to fill up
-    // during the brief time we didn't own the lock.
-    //
-    // Performance note: unblocks after at least N successful recv calls
-    m_send_end_cv.wait(lock, [this]{ return !is_full(); });
-    m_is_send_done = true;
-  }
-
-  // see also explanation in _channel<T, N>::recv()
-  m_send_begin_cv.notify_one();
-}
-
-template<class T, std::size_t N>
-T internal::_channel<T, N>::recv()
-{
-  static_assert(internal::_is_exception_safe<T>::value,
-    "Cannot guarantee exception safety, use another recv operator");
-
-  std::unique_lock<std::mutex> lock(m_mutex);
-  _pre_blocking_recv(lock);
-
-  std::pair<std::thread::id, T> pair(std::move(m_queue.front()));
-  assert(!is_full() || std::this_thread::get_id() != pair.first);
-
-  _post_blocking_recv(lock);
-  return std::move(pair.second);
-}
-
-template<class T, std::size_t N>
-void internal::_channel<T, N>::recv(T& t)
-{
-  std::unique_lock<std::mutex> lock(m_mutex);
-  _pre_blocking_recv(lock);
-
-  std::pair<std::thread::id, T> pair(std::move(m_queue.front()));
-  assert(!is_full() || std::this_thread::get_id() != pair.first);
-
-  // assignment before pop_front() to ensure strong exception safety
-  t = std::move(pair.second);
-  _post_blocking_recv(lock);
-}
-
-template<class T, std::size_t N>
-std::unique_ptr<T> internal::_channel<T, N>::recv_ptr()
-{
-  std::unique_lock<std::mutex> lock(m_mutex);
-  _pre_blocking_recv(lock);
-
-  std::pair<std::thread::id, T> pair(std::move(m_queue.front()));
-  assert(!is_full() || std::this_thread::get_id() != pair.first);
-
-  // move/copy before pop_front() to ensure strong exception safety
-  std::unique_ptr<T> t_ptr(make_unique<T>(std::move(pair.second)));
-  _post_blocking_recv(lock);
-  return t_ptr;
-}
-
-}
-
-#endif
\ No newline at end of file

From 4f1d836cfb9ee269c19ca0276050536b64901d54 Mon Sep 17 00:00:00 2001
From: Sergey Polyakov <sergeuz@gmail.com>
Date: Fri, 16 Feb 2024 18:32:11 +0200
Subject: [PATCH 2/6] Drop chunk notification events if the event queue is full

---
 system/inc/active_object.h     | 14 +++++++-------
 system/inc/system_event.h      |  3 ++-
 system/inc/system_threading.h  |  7 +++++++
 system/src/firmware_update.cpp |  2 +-
 system/src/system_event.cpp    | 10 +++++++---
 5 files changed, 24 insertions(+), 12 deletions(-)

diff --git a/system/inc/active_object.h b/system/inc/active_object.h
index 2752fca1af..79d95f0f87 100644
--- a/system/inc/active_object.h
+++ b/system/inc/active_object.h
@@ -283,7 +283,7 @@ class ActiveObjectBase
 
     // todo - concurrent queue should be a strategy so it's pluggable without requiring inheritance
     virtual bool take(Item& item)=0;
-    virtual bool put(Item& item)=0;
+    virtual bool put(Item& item, bool dontBlock = false)=0;
 
     /**
      * Static thread entrypoint to run this active object loop.
@@ -315,13 +315,13 @@ class ActiveObjectBase
         return started;
     }
 
-    template<typename R> void invoke_async(const std::function<R(void)>& work)
+    template<typename R> void invoke_async(const std::function<R(void)>& work, bool dontBlock = false)
     {
         auto task = new AsyncTask<R>(work);
         if (task)
         {
 			Item message = task;
-			if (!put(message))
+			if (!put(message, dontBlock))
 				delete task;
         }
 	}
@@ -354,9 +354,9 @@ class ActiveObjectQueue : public ActiveObjectBase
         return !os_queue_take(queue, &result, configuration.take_wait, nullptr);
     }
 
-    virtual bool put(Item& item)
+    virtual bool put(Item& item, bool dontBlock)
     {
-        return !os_queue_put(queue, &item, configuration.put_wait, nullptr);
+        return !os_queue_put(queue, &item, dontBlock ? 0 : configuration.put_wait, nullptr);
     }
 
     void createQueue()
@@ -422,9 +422,9 @@ class ActiveObjectThreadQueue : public ActiveObjectQueue
         return r;
     }
 
-    virtual bool put(Item& item) override
+    virtual bool put(Item& item, bool dontBlock) override
     {
-        bool r = ActiveObjectQueue::put(item);
+        bool r = ActiveObjectQueue::put(item, dontBlock);
         if (r && _thread != OS_THREAD_INVALID_HANDLE) {
             os_thread_notify(_thread, nullptr);
         }
diff --git a/system/inc/system_event.h b/system/inc/system_event.h
index a350bad032..885d9af48d 100644
--- a/system/inc/system_event.h
+++ b/system/inc/system_event.h
@@ -113,7 +113,8 @@ enum SystemEventsParam {
  * Flags altering the behavior of the `system_notify_event()` function.
  */
 enum SystemNotifyEventFlag {
-    NOTIFY_SYNCHRONOUSLY = 0x01
+    NOTIFY_SYNCHRONOUSLY = 0x01,
+    NOTIFY_IF_POSSIBLE = 0x02
 };
 
 #define SYSTEM_EVENT_CONTEXT_VERSION        (2)
diff --git a/system/inc/system_threading.h b/system/inc/system_threading.h
index c48852bcf5..d5e4a2b88b 100644
--- a/system/inc/system_threading.h
+++ b/system/inc/system_threading.h
@@ -86,6 +86,13 @@ os_mutex_recursive_t mutex_usb_serial();
         return; \
     }
 
+#define _THREAD_CONTEXT_TRY_ASYNC(thread, fn) \
+    if (thread.isStarted() && !thread.isCurrentThread()) { \
+        auto lambda = [=]() { (fn); }; \
+        thread.invoke_async(particle::FFL(lambda), true /* dontBlock */); \
+        return; \
+    }
+
 // execute synchronously on the system thread. Since the parameter lifetime is
 // assumed to be bound by the caller, the parameters don't need marshalling
 // fn: the function call to perform. This is textually substitued into a lambda, with the
diff --git a/system/src/firmware_update.cpp b/system/src/firmware_update.cpp
index 8a6be95312..7c6de77e99 100644
--- a/system/src/firmware_update.cpp
+++ b/system/src/firmware_update.cpp
@@ -259,7 +259,7 @@ int FirmwareUpdate::saveChunk(const char* chunkData, size_t chunkSize, size_t ch
     // Generate a system event
     fileDesc_.chunk_address = fileDesc_.file_address + chunkOffset;
     fileDesc_.chunk_size = chunkSize;
-    system_notify_event(firmware_update, firmware_update_progress, &fileDesc_);
+    system_notify_event(firmware_update, firmware_update_progress, &fileDesc_, nullptr /* fn */, nullptr /* fndata */, NOTIFY_IF_POSSIBLE);
     lastActiveTime_ = HAL_Timer_Get_Milli_Seconds();
     return 0;
 }
diff --git a/system/src/system_event.cpp b/system/src/system_event.cpp
index a84f5e1338..eb504f8e36 100644
--- a/system/src/system_event.cpp
+++ b/system/src/system_event.cpp
@@ -87,9 +87,13 @@ void system_notify_event_impl(system_event_t event, uint32_t data, void* pointer
     }
 }
 
-void system_notify_event_async(system_event_t event, uint32_t data, void* pointer, void (*fn)(void* data), void* fndata) {
+void system_notify_event_async(system_event_t event, uint32_t data, void* pointer, void (*fn)(void* data), void* fndata, bool dontBlock = false) {
     // run event notifications on the application thread
-    APPLICATION_THREAD_CONTEXT_ASYNC(system_notify_event_async(event, data, pointer, fn, fndata));
+    if (dontBlock) {
+        APPLICATION_THREAD_CONTEXT_ASYNC(system_notify_event_async(event, data, pointer, fn, fndata));
+    } else {
+        _THREAD_CONTEXT_TRY_ASYNC(particle::ApplicationThread, system_notify_event_async(event, data, pointer, fn, fndata)); // FIXME
+    }
 #if PLATFORM_THREADING
     std::lock_guard<RecursiveMutex> lk(sSubscriptionsMutex);
 #endif // PLATFORM_THREADING
@@ -203,7 +207,7 @@ void system_notify_event(system_event_t event, uint32_t data, void* pointer, voi
             SystemISRTaskQueue.enqueue(task);
         };
     } else {
-        system_notify_event_async(event, data, pointer, fn, fndata);
+        system_notify_event_async(event, data, pointer, fn, fndata, flags & NOTIFY_IF_POSSIBLE);
     }
 }
 

From d6fa39d4aecb4c271560515e4a581240d7f43e54 Mon Sep 17 00:00:00 2001
From: Sergey Polyakov <sergeuz@gmail.com>
Date: Mon, 19 Feb 2024 18:54:23 +0200
Subject: [PATCH 3/6] Minor fixes

---
 hal/src/newhal/static_recursive_mutex.h | 31 +++++++++++++++++++
 system/inc/active_object.h              | 17 +++++-----
 system/inc/system_event.h               |  4 +--
 system/inc/system_threading.h           |  4 ++-
 system/src/firmware_update.cpp          |  2 +-
 system/src/main.cpp                     |  2 ++
 system/src/system_event.cpp             | 41 ++++++++++++++++---------
 7 files changed, 75 insertions(+), 26 deletions(-)
 create mode 100644 hal/src/newhal/static_recursive_mutex.h

diff --git a/hal/src/newhal/static_recursive_mutex.h b/hal/src/newhal/static_recursive_mutex.h
new file mode 100644
index 0000000000..061650b234
--- /dev/null
+++ b/hal/src/newhal/static_recursive_mutex.h
@@ -0,0 +1,31 @@
+/*
+ * Copyright (c) 2024 Particle Industries, Inc.  All rights reserved.
+ *
+ * This library is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation, either
+ * version 3 of the License, or (at your option) any later version.
+ *
+ * This library is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with this library; if not, see <http://www.gnu.org/licenses/>.
+ */
+
+#pragma once
+
+class StaticRecursiveMutex {
+public:
+    StaticRecursiveMutex() = default;
+
+    bool lock(unsigned timeout = 0) {
+        return true;
+    }
+
+    bool unlock() {
+        return true;
+    }
+};
diff --git a/system/inc/active_object.h b/system/inc/active_object.h
index 79d95f0f87..35b2afb593 100644
--- a/system/inc/active_object.h
+++ b/system/inc/active_object.h
@@ -315,16 +315,19 @@ class ActiveObjectBase
         return started;
     }
 
-    template<typename R> void invoke_async(const std::function<R(void)>& work, bool dontBlock = false)
+    template<typename R> bool invoke_async(const std::function<R(void)>& work, bool dontBlock = false)
     {
         auto task = new AsyncTask<R>(work);
-        if (task)
-        {
-			Item message = task;
-			if (!put(message, dontBlock))
-				delete task;
+        if (!task) {
+            return false;
+        }
+        Item message = task;
+        if (!put(message, dontBlock)) {
+            delete task;
+            return false;
         }
-	}
+        return true;
+    }
 
     template<typename R> SystemPromise<R>* invoke_future(const std::function<R(void)>& work)
     {
diff --git a/system/inc/system_event.h b/system/inc/system_event.h
index 885d9af48d..3e6d8eab14 100644
--- a/system/inc/system_event.h
+++ b/system/inc/system_event.h
@@ -113,8 +113,8 @@ enum SystemEventsParam {
  * Flags altering the behavior of the `system_notify_event()` function.
  */
 enum SystemNotifyEventFlag {
-    NOTIFY_SYNCHRONOUSLY = 0x01,
-    NOTIFY_IF_POSSIBLE = 0x02
+    NOTIFY_SYNCHRONOUSLY = 0x01, ///< Invoke the event handlers directly in the calling thread.
+    NOTIFY_DONT_BLOCK = 0x02 ///< Ignore the event if the event queue of the application thread is full.
 };
 
 #define SYSTEM_EVENT_CONTEXT_VERSION        (2)
diff --git a/system/inc/system_threading.h b/system/inc/system_threading.h
index d5e4a2b88b..a5a912ebd0 100644
--- a/system/inc/system_threading.h
+++ b/system/inc/system_threading.h
@@ -86,7 +86,7 @@ os_mutex_recursive_t mutex_usb_serial();
         return; \
     }
 
-#define _THREAD_CONTEXT_TRY_ASYNC(thread, fn) \
+#define _THREAD_CONTEXT_ASYNC_TRY(thread, fn) \
     if (thread.isStarted() && !thread.isCurrentThread()) { \
         auto lambda = [=]() { (fn); }; \
         thread.invoke_async(particle::FFL(lambda), true /* dontBlock */); \
@@ -112,6 +112,7 @@ os_mutex_recursive_t mutex_usb_serial();
 #else // !PLATFORM_THREADING
 
 #define _THREAD_CONTEXT_ASYNC(thread, fn)
+#define _THREAD_CONTEXT_ASYNC_TRY(thread, fn)
 #define _THREAD_CONTEXT_ASYNC_RESULT(thread, fn, result)
 #define SYSTEM_THREAD_CONTEXT_SYNC(fn)
 
@@ -123,6 +124,7 @@ os_mutex_recursive_t mutex_usb_serial();
 #define SYSTEM_THREAD_CONTEXT_ASYNC(fn) _THREAD_CONTEXT_ASYNC(particle::SystemThread, fn)
 #define SYSTEM_THREAD_CONTEXT_ASYNC_RESULT(fn, result) _THREAD_CONTEXT_ASYNC_RESULT(particle::SystemThread, fn, result)
 #define APPLICATION_THREAD_CONTEXT_ASYNC(fn) _THREAD_CONTEXT_ASYNC(particle::ApplicationThread, fn)
+#define APPLICATION_THREAD_CONTEXT_ASYNC_TRY(fn) _THREAD_CONTEXT_ASYNC_TRY(particle::ApplicationThread, fn)
 #define APPLICATION_THREAD_CONTEXT_ASYNC_RESULT(fn, result) _THREAD_CONTEXT_ASYNC_RESULT(particle::ApplicationThread, fn, result)
 
 // Perform an asynchronous function call if not on the system thread,
diff --git a/system/src/firmware_update.cpp b/system/src/firmware_update.cpp
index 7c6de77e99..56dedafecc 100644
--- a/system/src/firmware_update.cpp
+++ b/system/src/firmware_update.cpp
@@ -259,7 +259,7 @@ int FirmwareUpdate::saveChunk(const char* chunkData, size_t chunkSize, size_t ch
     // Generate a system event
     fileDesc_.chunk_address = fileDesc_.file_address + chunkOffset;
     fileDesc_.chunk_size = chunkSize;
-    system_notify_event(firmware_update, firmware_update_progress, &fileDesc_, nullptr /* fn */, nullptr /* fndata */, NOTIFY_IF_POSSIBLE);
+    system_notify_event(firmware_update, firmware_update_progress, &fileDesc_, nullptr /* fn */, nullptr /* fndata */, NOTIFY_DONT_BLOCK);
     lastActiveTime_ = HAL_Timer_Get_Milli_Seconds();
     return 0;
 }
diff --git a/system/src/main.cpp b/system/src/main.cpp
index bb1f99c3d5..57f5954d14 100644
--- a/system/src/main.cpp
+++ b/system/src/main.cpp
@@ -31,6 +31,8 @@
 // STATIC_ASSERT macro clashes with the nRF SDK
 #define NO_STATIC_ASSERT
 
+#include <cmath>
+
 #include "debug.h"
 #include "system_event.h"
 #include "system_mode.h"
diff --git a/system/src/system_event.cpp b/system/src/system_event.cpp
index eb504f8e36..5dbbbae48f 100644
--- a/system/src/system_event.cpp
+++ b/system/src/system_event.cpp
@@ -17,14 +17,17 @@
  ******************************************************************************
  */
 
+#define NO_STATIC_ASSERT
+
 #include "system_event.h"
 #include "system_threading.h"
 #include "interrupts_hal.h"
 #include "system_task.h"
+#include "static_recursive_mutex.h"
+#include "scope_guard.h"
 #include <stdint.h>
 #include "spark_wiring_vector.h"
 #include "spark_wiring_interrupts.h"
-#include "spark_wiring_thread.h"
 #include <algorithm>
 
 namespace {
@@ -75,10 +78,20 @@ struct SystemEventSubscription {
 // for now a simple implementation
 spark::Vector<SystemEventSubscription> subscriptions;
 #if PLATFORM_THREADING
-RecursiveMutex sSubscriptionsMutex;
+StaticRecursiveMutex sSubscriptionsMutex;
 #endif // PLATFORM_THREADING
 
-void system_notify_event_impl(system_event_t event, uint32_t data, void* pointer, void (*fn)(void* data), void* fndata) {
+void system_notify_event_impl(system_event_t event, uint32_t data, void* pointer, void (*fn)(void* data), void* fndata, bool isIsr) {
+#if PLATFORM_THREADING
+    if (!isIsr) {
+        sSubscriptionsMutex.lock();
+    }
+    SCOPE_GUARD({
+        if (!isIsr) {
+            sSubscriptionsMutex.unlock();
+        }
+    });
+#endif // PLATFORM_THREADING
     for (SystemEventSubscription& subscription : subscriptions) {
         subscription.notify(event, data, pointer);
     }
@@ -90,14 +103,11 @@ void system_notify_event_impl(system_event_t event, uint32_t data, void* pointer
 void system_notify_event_async(system_event_t event, uint32_t data, void* pointer, void (*fn)(void* data), void* fndata, bool dontBlock = false) {
     // run event notifications on the application thread
     if (dontBlock) {
-        APPLICATION_THREAD_CONTEXT_ASYNC(system_notify_event_async(event, data, pointer, fn, fndata));
+        APPLICATION_THREAD_CONTEXT_ASYNC_TRY(system_notify_event_async(event, data, pointer, fn, fndata, dontBlock));
     } else {
-        _THREAD_CONTEXT_TRY_ASYNC(particle::ApplicationThread, system_notify_event_async(event, data, pointer, fn, fndata)); // FIXME
+        APPLICATION_THREAD_CONTEXT_ASYNC(system_notify_event_async(event, data, pointer, fn, fndata, dontBlock));
     }
-#if PLATFORM_THREADING
-    std::lock_guard<RecursiveMutex> lk(sSubscriptionsMutex);
-#endif // PLATFORM_THREADING
-    system_notify_event_impl(event, data, pointer, fn, fndata);
+    system_notify_event_impl(event, data, pointer, fn, fndata, false /* isIsr */);
 }
 
 class SystemEventTask : public ISRTaskQueue::Task {
@@ -149,7 +159,7 @@ int system_subscribe_event(system_event_t events, system_event_handler_t* handle
     // Modification of subscriptions normally happens from thread context, so for events generated outside ISR
     // context, only mutex acquisition is sufficient to keep things thread safe (see system_notify_event_async())
 #if PLATFORM_THREADING
-    std::lock_guard<RecursiveMutex> lk(sSubscriptionsMutex);
+    std::lock_guard lk(sSubscriptionsMutex);
 #endif // PLATFORM_THREADING
     int r = 0;
     ATOMIC_BLOCK() {
@@ -168,7 +178,7 @@ void system_unsubscribe_event(system_event_t events, system_event_handler_t* han
     // Modification of subscriptions normally happens from thread context, so for events generated outside ISR
     // context, only mutex acquisition is sufficient to keep things thread safe (see system_notify_event_async())
 #if PLATFORM_THREADING
-    std::lock_guard<RecursiveMutex> lk(sSubscriptionsMutex);
+    std::lock_guard lk(sSubscriptionsMutex);
 #endif // PLATFORM_THREADING
     ATOMIC_BLOCK() {
         auto it = std::remove_if(subscriptions.begin(), subscriptions.end(), [events, handler, context](const SystemEventSubscription& sub) {
@@ -199,15 +209,16 @@ void system_notify_event(system_event_t event, uint32_t data, void* pointer, voi
         unsigned flags) {
     // TODO: Add an API that would allow user applications to control which event handlers can be
     // executed synchronously, possibly in the context of an ISR
+    bool isIsr = hal_interrupt_is_isr();
     if (flags & NOTIFY_SYNCHRONOUSLY) {
-        system_notify_event_impl(event, data, pointer, fn, fndata);
-    } else if (hal_interrupt_is_isr()) {
+        system_notify_event_impl(event, data, pointer, fn, fndata, isIsr);
+    } else if (isIsr) {
         auto task = systemPoolNew<SystemEventTask>(event, data, pointer, fn, fndata);
         if (task) {
             SystemISRTaskQueue.enqueue(task);
-        };
+        }
     } else {
-        system_notify_event_async(event, data, pointer, fn, fndata, flags & NOTIFY_IF_POSSIBLE);
+        system_notify_event_async(event, data, pointer, fn, fndata, flags & NOTIFY_DONT_BLOCK);
     }
 }
 

From 570aab766e2436547f60ec9ad4adb6ea5c1bee33 Mon Sep 17 00:00:00 2001
From: Sergey Polyakov <sergeuz@gmail.com>
Date: Mon, 19 Feb 2024 20:50:28 +0200
Subject: [PATCH 4/6] Add test

---
 user/tests/wiring/no_fixture/thread.cpp | 18 ++++++++++++++++++
 1 file changed, 18 insertions(+)

diff --git a/user/tests/wiring/no_fixture/thread.cpp b/user/tests/wiring/no_fixture/thread.cpp
index f433e1c611..83ddf7ef84 100644
--- a/user/tests/wiring/no_fixture/thread.cpp
+++ b/user/tests/wiring/no_fixture/thread.cpp
@@ -248,6 +248,24 @@ test(THREAD_08_newlib_reent_impure_ptr_changes_on_context_switch)
     assertEqual((uintptr_t)testImpure, (uintptr_t)_impure_ptr);
 }
 
+test(THREAD_09_dont_block_event_queue_option)
+{
+    if (system_thread_get_state(nullptr) != spark::feature::ENABLED) {
+        skip();
+        return;
+    }
+    ActiveObjectBase* app = (ActiveObjectBase*)system_internal(0, nullptr); // Returns application thread instance
+    test_val_fn1 = 0;
+    std::function<void(void)> fn = increment;
+    for (int i = 0; i < 20; ++i) {
+        assertTrue(app->invoke_async(fn, true /* dontBlock */));
+    }
+    assertFalse(app->invoke_async(fn, true));
+    while (Particle.process()) {
+    }
+    assertEqual((int)test_val_fn1, 20);
+}
+
 // todo - test for SingleThreadedSection
 
 

From b5657393dc8f191301dc912e11935c21ac588f07 Mon Sep 17 00:00:00 2001
From: Sergey Polyakov <sergeuz@gmail.com>
Date: Thu, 22 Feb 2024 18:33:55 +0200
Subject: [PATCH 5/6] Inline methods that depend on
 PARTICLE_WIRING_PRINT_NO_FLOAT to avoid linker errors

---
 wiring/inc/spark_wiring_print.h   | 76 +++++++++++++++++++++++++++++--
 wiring/src/spark_wiring_print.cpp | 73 -----------------------------
 2 files changed, 71 insertions(+), 78 deletions(-)

diff --git a/wiring/inc/spark_wiring_print.h b/wiring/inc/spark_wiring_print.h
index adc2748f14..6eea33abc3 100644
--- a/wiring/inc/spark_wiring_print.h
+++ b/wiring/inc/spark_wiring_print.h
@@ -35,6 +35,7 @@
 
 #include "spark_wiring_printable.h"
 #include "spark_wiring_fixed_point.h"
+#include <cmath>
 #include <climits>
 #include <cstdarg>
 
@@ -65,9 +66,60 @@ class Print
 
     size_t printNumber(unsigned long, uint8_t);
     size_t printNumber(unsigned long long, uint8_t);
+
 #ifndef PARTICLE_WIRING_PRINT_NO_FLOAT
-    size_t printFloat(double, uint8_t);
+    size_t printFloat(double number, uint8_t digits) {
+        size_t n = 0;
+
+        if (std::isnan(number)) {
+            return print("nan");
+        }
+        if (std::isinf(number)) {
+            return print("inf");
+        }
+        if (number > 4294967040.0) {
+            return print ("ovf"); // constant determined empirically
+        }
+        if (number <-4294967040.0) {
+            return print ("ovf"); // constant determined empirically
+        }
+
+        // Handle negative numbers
+        if (number < 0.0) {
+            n += print('-');
+            number = -number;
+        }
+
+        // Round correctly so that print(1.999, 2) prints as "2.00"
+        double rounding = 0.5;
+        for (uint8_t i = 0; i < digits; ++i) {
+            rounding /= 10.0;
+        }
+
+        number += rounding;
+
+        // Extract the integer part of the number and print it
+        unsigned long int_part = (unsigned long)number;
+        double remainder = number - (double)int_part;
+        n += print(int_part);
+
+        // Print the decimal point, but only if there are digits beyond
+        if (digits > 0) {
+            n += print(".");
+        }
+
+        // Extract digits from the remainder one at a time
+        while (digits-- > 0) {
+            remainder *= 10.0;
+            int toPrint = int(remainder);
+            n += print(toPrint);
+            remainder -= toPrint;
+        }
+
+        return n;
+    }
 #endif // PARTICLE_WIRING_PRINT_NO_FLOAT
+
     size_t printVariant(const particle::Variant& var);
 
   protected:
@@ -92,9 +144,15 @@ class Print
     template <typename T, std::enable_if_t<!std::is_base_of<Printable, T>::value && (std::is_integral<T>::value || std::is_convertible<T, unsigned long long>::value ||
         std::is_convertible<T, long long>::value), int> = 0>
     size_t print(T, int = DEC);
+
 #ifndef PARTICLE_WIRING_PRINT_NO_FLOAT
-    size_t print(float, int = 2);
-    size_t print(double, int = 2);
+    size_t print(float n, int digits) {
+        return printFloat((double)n, digits);
+    }
+
+    size_t print(double n, int digits) {
+        return printFloat(n, digits);
+    }
 #endif // PARTICLE_WIRING_PRINT_NO_FLOAT
 
     // Prevent implicit constructors of Variant from affecting overload resolution
@@ -115,9 +173,17 @@ class Print
         n += println();
         return n;
     }
+
 #ifndef PARTICLE_WIRING_PRINT_NO_FLOAT
-    size_t println(float, int = 2);
-    size_t println(double, int = 2);
+    size_t println(float num, int digits) {
+        return println((double)num, digits);
+    }
+
+    size_t println(double num, int digits) {
+        size_t n = print(num, digits);
+        n += println();
+        return n;
+    }
 #endif // PARTICLE_WIRING_PRINT_NO_FLOAT
     
     template<typename T, typename std::enable_if_t<std::is_same_v<T, particle::Variant>, int> = 0>
diff --git a/wiring/src/spark_wiring_print.cpp b/wiring/src/spark_wiring_print.cpp
index 779db026bf..1944cffac2 100644
--- a/wiring/src/spark_wiring_print.cpp
+++ b/wiring/src/spark_wiring_print.cpp
@@ -124,18 +124,6 @@ size_t Print::print(char c)
   return write(c);
 }
 
-#ifndef PARTICLE_WIRING_PRINT_NO_FLOAT
-size_t Print::print(float n, int digits)
-{
-  return printFloat((double)n, digits);
-}
-
-size_t Print::print(double n, int digits)
-{
-  return printFloat(n, digits);
-}
-#endif // PARTICLE_WIRING_PRINT_NO_FLOAT
-
  size_t Print::print(const Printable& x)
  {
    return x.printTo(*this);
@@ -167,20 +155,6 @@ size_t Print::println(char c)
   return n;
 }
 
-#ifndef PARTICLE_WIRING_PRINT_NO_FLOAT
-size_t Print::println(float num, int digits)
-{
-    return println((double)num, digits);
-}
-
-size_t Print::println(double num, int digits)
-{
-  size_t n = print(num, digits);
-  n += println();
-  return n;
-}
-#endif // PARTICLE_WIRING_PRINT_NO_FLOAT
-
  size_t Print::println(const Printable& x)
  {
    size_t n = print(x);
@@ -234,53 +208,6 @@ size_t Print::printNumber(unsigned long n, uint8_t base) {
   return write(str);
 }
 
-#ifndef PARTICLE_WIRING_PRINT_NO_FLOAT
-size_t Print::printFloat(double number, uint8_t digits)
-{
-  size_t n = 0;
-
-  if (isnan(number)) return print("nan");
-  if (isinf(number)) return print("inf");
-  if (number > 4294967040.0) return print ("ovf");  // constant determined empirically
-  if (number <-4294967040.0) return print ("ovf");  // constant determined empirically
-
-  // Handle negative numbers
-  if (number < 0.0)
-  {
-     n += print('-');
-     number = -number;
-  }
-
-  // Round correctly so that print(1.999, 2) prints as "2.00"
-  double rounding = 0.5;
-  for (uint8_t i=0; i<digits; ++i)
-    rounding /= 10.0;
-
-  number += rounding;
-
-  // Extract the integer part of the number and print it
-  unsigned long int_part = (unsigned long)number;
-  double remainder = number - (double)int_part;
-  n += print(int_part);
-
-  // Print the decimal point, but only if there are digits beyond
-  if (digits > 0) {
-    n += print(".");
-  }
-
-  // Extract digits from the remainder one at a time
-  while (digits-- > 0)
-  {
-    remainder *= 10.0;
-    int toPrint = int(remainder);
-    n += print(toPrint);
-    remainder -= toPrint;
-  }
-
-  return n;
-}
-#endif // PARTICLE_WIRING_PRINT_NO_FLOAT
-
 size_t Print::printVariant(const Variant& var) {
     JSONStreamWriter writer(*this);
     writeVariant(var, writer);

From 28e9d840d7a86ebc1efded263bc86b025b0fc3b7 Mon Sep 17 00:00:00 2001
From: Andrey Tolstoy <avtolstoy@gmail.com>
Date: Mon, 26 Feb 2024 18:41:09 +0700
Subject: [PATCH 6/6] [wiring] Print: fix default float digits

---
 wiring/inc/spark_wiring_print.h | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)

diff --git a/wiring/inc/spark_wiring_print.h b/wiring/inc/spark_wiring_print.h
index 6eea33abc3..3af8efcf62 100644
--- a/wiring/inc/spark_wiring_print.h
+++ b/wiring/inc/spark_wiring_print.h
@@ -68,6 +68,9 @@ class Print
     size_t printNumber(unsigned long long, uint8_t);
 
 #ifndef PARTICLE_WIRING_PRINT_NO_FLOAT
+
+    static constexpr auto FLOAT_DEFAULT_FRACTIONAL_DIGITS = 2;
+
     size_t printFloat(double number, uint8_t digits) {
         size_t n = 0;
 
@@ -146,11 +149,11 @@ class Print
     size_t print(T, int = DEC);
 
 #ifndef PARTICLE_WIRING_PRINT_NO_FLOAT
-    size_t print(float n, int digits) {
+    size_t print(float n, int digits = FLOAT_DEFAULT_FRACTIONAL_DIGITS) {
         return printFloat((double)n, digits);
     }
 
-    size_t print(double n, int digits) {
+    size_t print(double n, int digits = FLOAT_DEFAULT_FRACTIONAL_DIGITS) {
         return printFloat(n, digits);
     }
 #endif // PARTICLE_WIRING_PRINT_NO_FLOAT
@@ -175,11 +178,11 @@ class Print
     }
 
 #ifndef PARTICLE_WIRING_PRINT_NO_FLOAT
-    size_t println(float num, int digits) {
+    size_t println(float num, int digits = FLOAT_DEFAULT_FRACTIONAL_DIGITS) {
         return println((double)num, digits);
     }
 
-    size_t println(double num, int digits) {
+    size_t println(double num, int digits = FLOAT_DEFAULT_FRACTIONAL_DIGITS) {
         size_t n = print(num, digits);
         n += println();
         return n;