// We need to lock the print method because in some cases we print different

// parts of messages from different threads.

print_mutex.lock();

LengthEncodedMessage length_encoded_message =

get_message_with_encoded_length(str);

char *buff = new char[length_encoded_message.length];

std::setvbuf(destinaton, buff, _IOFBF, length_encoded_message.length);

fwrite(length_encoded_message.message, length_encoded_message.length, 1,

destinaton);

fflush(destinaton);

delete[] buff;

print_mutex.unlock();

std::stringstream str;

str << num;

trace(str.str().c_str());

std::string method_id, int code) {

nlohmann::json exception;

exception[kError][kMessage] = message;

exception[kError][kCode] = code;

exception[kError][kDeviceId] = device_identifier;

exception[kDeviceId] = device_identifier;

exception[kId] = method_id;

// Decided it's a better idea to print everything to stdout

// Not a good practice, but the client process wouldn't have to monitor both

// streams fprintf(stderr, "%s", exception.dump().c_str());

print(exception);

std::string device_identifier, std::string method_id,

int code) {

std::ostringstream joined_error_message;

std::copy(messages.begin(), messages.end(),

std::ostream_iterator<std::string>(joined_error_message, "\n"));

print_error(joined_error_message.str().c_str(), device_identifier, method_id,

code);