v3: Gateway reconnect panic #286
Comments
What's your code for this? The library should already be handling reconnection. FWIW, that is eventually going to be refactored to be completely thread-safe, |
|
Yes, for every "session", if This is because reconnection is not working well in my experience, if I see something like "going away" or "requested to reconnect", IIRC it hangs indefinitely, simply close it and create a new state solved it. func session (sCloser chan struct{}) (err error) {
var sessionSelfCloser chan struct{} = make(chan struct{})
logger.Logger.Infof("[MAIN] Session is starting...")
sv, err = storage.Sqlite()
if err != nil {
return errors.Wrap(err, "Failed to get storage")
}
binding.Setup(sv)
s, err := state.New("Bot " + *globalFlags.token)
if err != nil {
return errors.Wrap(err, "Failed to get new bot state")
}
s.AddIntents(gateway.IntentDirectMessages)
s.AddIntents(gateway.IntentGuildMessages)
s.AddIntents(gateway.IntentGuildMessageReactions)
pendingEmbeds = make(chan *pendingEmbed, 512)
redirectorClosed := redirectorLoop(s, sessionSelfCloser)
err = assureCommands(s)
if err != nil {
logger.Logger.Fatalf("[MAIN] %v", err)
}
addEventHandlers(s)
addInteractionHandlers(s)
s.ErrorLog = func(innerErr error) {
logger.Logger.Errorf("[MAIN] Gateway error: %v", innerErr)
err = innerErr
select {
case <-sessionSelfCloser:
default:
close(sessionSelfCloser)
}
}
s.FatalErrorCallback = func(innerErr error) {
logger.Logger.Errorf("[MAIN] Fatal gateway error: %v", err)
err = innerErr
select {
case <-sessionSelfCloser:
default:
close(sessionSelfCloser)
}
}
s.AfterClose = func(innerErr error) {
logger.Logger.Errorf("[MAIN] After gateway closed: %v", err)
err = innerErr
select {
case <-sessionSelfCloser:
default:
close(sessionSelfCloser)
}
}
if err := s.Open(context.Background()); err != nil {
logger.Logger.Errorf("[MAIN] %v", err)
select {
case <-sessionSelfCloser:
default:
close(sessionSelfCloser)
}
}
defer s.Close()
u, err := s.Me()
if err != nil {
log.Fatalln("Failed to get myself:", err)
}
logger.Logger.Infof("Session: %d", u.ID)
s.UpdateStatus(gateway.UpdateStatusData{
Since: 0,
Activities: [] discord.Activity {
{
Name: locale.ACTIVITY,
Type: discord.WatchingActivity,
},
},
Status: discord.OnlineStatus,
AFK: false,
})
logger.Logger.Infof("====== %s at your service ======", u.Username)
promptClosed := startPromptLoop(s, sessionSelfCloser)
select {
case <-sCloser:
close(sessionSelfCloser)
case <-sessionSelfCloser:
}
s.ErrorLog = nil
os.Stdin.WriteString("o")
select {
case <-promptClosed:
case <-time.NewTimer(time.Second*5).C:
}
<-redirectorClosed
err = sv.Close()
if err != nil {
return err
}
logger.Logger.Infof("[MAIN] Session is closed.")
return nil
} |
This is worthy of its own issue. I think it would be much better if you just I'll still leave this issue open and look into why it happens, but I strongly |
|
Well yeah, I was rushing to put together the bot for a private server, and disconnection casually happens ~10 times a day, so I didn't consider the whole connection issue is something not-awared-of; The session loop is also to handle errors in my own logic, it worked well, therefore I just kept it as-is so far. Actually I decided to skip the re-connection very soon so it could be that I misunderstood it's not working, so I need to confirm the issue. But not very soon I'm afraid, once I got spare time to test re-connection out and discover something, I'll post. |
And several seconds before that:
Whenever a connection issue occurred, instead of tried to figure out how to reconnect, I made it to start a new state all over again, I'm wondering this could be the issue.
The text was updated successfully, but these errors were encountered: