Closed markstory closed 10 years ago
Does anyone have confirmation that this fixes #58?
From my limited local testing with ejabberd it solved problems where hubot would not come online after ejabberd restart
.
I'm still seeing duplicated responses and log entries, even with the new changes:
hubot-dev [9:05 AM] PONG
bnied [9:05 AM] [this command causes a reset with Slack] hubot-dev show storage
bnied [9:07 AM] hubot-dev ping
hubot-dev [9:07 AM] PONG
hubot-dev [9:07 AM] PONG
hubot-dev [9:07 AM] PONG
hubot-dev [9:07 AM] PONG
hubot-dev [9:07 AM] PONG
hubot-dev [9:07 AM] PONG
I'm not sure of what the problem is then. How is the connection being reset?
I'm running a command which is causing Slack to send an offline event.
I think this resolves #58 and #63. When ejabberd - and I assume other xmpp servers - are hard reset node-xmpp gets a
ECONNREFUSED
error which prevents thereconnect
option from working. In these situations, theend
event is emitted. The adapter now handles this event and attempts to reconnect.If this reconnection process fails the bot process will be terminated. This will allow process managers like supervisord to restart the bot.