Re: [fltk.general] Strange artefacts visible on controls - Forums

That sounds most curious - if the async call is not interacting with fltk at all, I’d expect it to be “safe”, so this is unexpected. Does it (the async method) interact with the GUI context in some way, or something?

As far as I know - it doesn't at all.

It might be useful if you could show us the code for the callback, in case something useful occurs to anyone on seeing it...

The connectToServer() callback calls gui.getNet().sendMessage(NetProtocol::MsgHello::ID), which calls sendMessage(nullptr, id), which calls getClient().write(msg.encode()), which calls session->write().

void GUIDlgNetClient::connectToServer(void)
{
srvIPAddressInput->deactivate();
srvPortInput->deactivate();
netMessages->value(u8"");
addMessage(gui.getTranslation().Translate(u8"Connecting to server..."));
Fl::check();
int ret = gui.getNet().getClient().start();
if (ret != Network::NET_OK)
{
addMessage(GUI::translateErrorCode(ret));
if (gui.getNet().getClient().isActive())
gui.getNet().getClient().stop();
}
else
{
addMessage(gui.getTranslation().Translate(u8"Connected!").c_str());

gui.getNet().sendMessage(NetProtocol::MsgHello::ID); // <<<<<< This line makes the artifacts!
}

}

bool Network::sendMessage(const unsigned char id)
{
return sendMessage(nullptr, id);
}

bool Network::sendMessage(const SessionPtr session, const unsigned char id)
{

switch (id)
{
case NetProtocol::MsgHello::ID:
{
NetProtocol::MsgHello msg;

getClient().write(msg.encode());
break;
}
case .....
case .....

.....
}

void NetClient::write(std::string msg)
{
session->addMessage(msg);
if (session->canWrite())
session->write();
}

void NetSession::write(void)
{
boost::asio::async_write(*socket.get(),
boost::asio::buffer(writeQueue.front()),
[&](boost::system::error_code error, std::size_t bytesTransferred)
{
if (!error)
{
writeQueue.pop();
if (!writeQueue.empty())
write();
}
else
{
auto err = error;
netConnector.getErrorHandler()(shared_from_this(), err);
}
}
);
}