Drogon hangs during shutdown after an exception

[arves100]

Hello, I am currently making a server with Drogon and struggled on an issue on how to close the server.
Currently, my idea is to throw an exception during the initialization of a custom plugin (like throw std::runtime_error) to exit the run() function, then I try to queue the quit operation, the function does return to 0 but drogon hangs in the deconstructor calls.

This is my main code:

int main(int argc, char** argv)
{
    try
    {
        drogon::app()
            .loadConfigFile("./config.json")
            .registerBeginningAdvice([]() {
            MigrationManager::RunMigrations(drogon::app().getDbClient());
                })
            .run()
            ;
    }
    catch (const std::exception& ex)
    {
        printf("Fatal exception during execution: %s\n", ex.what());
    }

    drogon::HttpAppFramework::instance().getLoop()->queueInLoop([]()
        {
            drogon::HttpAppFramework::instance().quit();
    });
    return 0;

This is what happens in the console during shutdown:
20250514 16:34:03.655000 UTC 31772 TRACE [~CacheMap] CacheMap destruct! - drogon/CacheMap.h:161

The code is stuck at the destructor of EventLoopThread:

    std::call_once(once_, [this]() {
        auto f = promiseForLoop_.get_future();
        promiseForRun_.set_value(1);
        // Make sure the event loop loops before returning.
        (void)f.get(); // this is where it gets stuck
    });

here I am checking that "f" is empty, not sure if that can be the issue.

Specifically the run function called here:

EventLoopThread::~EventLoopThread()
{
    run(); // this is where it hangs
    std::shared_ptr<EventLoop> loop;
    {
        std::unique_lock<std::mutex> lk(loopMutex_);
        loop = loop_;
    }

Perhaps I'm simply doing the stop part wrong?