Segfault when restarting the server

[dkalinowski]

Describe the bug There is segfault when the server is closed using ::quit() and then restarted using ::run().

GDB:

[New Thread 0x7fffda5d7640 (LWP 210860)]
[New Thread 0x7fffd9dd6640 (LWP 210861)]
[New Thread 0x7fffd95d5640 (LWP 210862)]
[New Thread 0x7fffd8dd4640 (LWP 210863)]
[Thread 0x7fffda5d7640 (LWP 210860) exited]
[Thread 0x7fffd8dd4640 (LWP 210863) exited]
[Thread 0x7fffd95d5640 (LWP 210862) exited]
[New Thread 0x7fffda5d7640 (LWP 210881)]
[Thread 0x7fffd9dd6640 (LWP 210861) exited]

Thread 1 "ovms_test" received signal SIGSEGV, Segmentation fault.
0x0000555556b23cf7 in drogon::ListenerManager::addListener(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, unsigned short, bool, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, bool, std::vector<std::pair<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > >, std::allocator<std::pair<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > > > > const&) ()
(gdb) 

To Reproduce

#include <drogon/drogon.h>

#include <thread>
#include <chrono>

...

    for (int i = 0; i < 2; i++) {
        std::thread k([] {
            std::this_thread::sleep_for(std::chrono::seconds(2));
            drogon::app().quit();
        });
        drogon::app()
                    .setThreadNum(3)
                    .setIdleConnectionTimeout(0)
                    .addListener("0.0.0.0", 11933)
                    .run();
        k.join();
    }

Expected behavior No segfault, server restarts

Desktop (please complete the following information):

• OS: ubuntu22

[dkalinowski]

Any hints? Is it not supported?

[an-tao]

This is not a common way to use Drogon. In fact, Drogon's design does not take into account the scenario of calling the run function again after it exits.

[dkalinowski]

It is valid use case for our library users to use C-API to stop and re-start the server with different parameters within the same process. I understand Drogon is not designed to support that (probably this is why HttpAppFramework instance is a singleton) - but is there any chance that Drogon team add support for that? The http server we migrate from (Tensorflow's net_http module) does support that.

Thank you