Bot is on a channel but doesn't seem to believe it

[x]

Hello,

this is one of the strangest problems I've ever had with my dear bots in these many
years of my bot usage. One bot in my botnet, more precisely, a bot with soft version
1.6.18 and ipv6-patch applied. I also use the newest Netbots script with my botnet.
Has been working great since the beginning, but started to act strange around a week ago.

Why strange? Well, it connects and joins to all channels where it is supposed
to join. But there is one channel, where it stopped to ask ops via netbots script
system (you know), So I started to inspect that what's going on and found out
that this bot doesn't see anyone on that very channel, althought it has joined there
and me + all the other bots see it. I tried to connect to this very server my strange
bot is using and used all the same details with my client and joined the chan, but
I was able to see everyone there and everything seemed normal.

I've tried to .jump my bot, it jumps and joins to this chan but doesn't see anyone,
not even itself there. When I do .say #channel something - it says "I'm not on that channel.".
I also removed the channel and added again ( -chan -> +chan) it parted
and joined back, but still continues with this strange behaviour.

So now I am asking, that does anyone have at least a some kind of idea, where
this problem lies eventually? If it's not the irc-server with desync or something,
or bot connects and acts normally on all the other channels, I really don't know
where to look at in order to find a solution. The channel name doesn't contain
any strange letters or stuff like that, the bot used to work fine on that very
channel, like on the others right now.

Ideas?

[r0t3n]

try and dump a NAMES and WHOIS and maybe a WHO line on the channel and see if it changes anything.

ie

.dump NAMES #mychannel
.dump WHOIS #mychannel
.dump WHO #mychannel

[x]

Tosser^^: When I did this thing -> .dump WHO #channelname

The bot immediately asked its OPs and works now normally.
Command .channel #channelname also shows everyone who
is there on this channel.

But it's weird though, this problem might just reoccur randomly
I guess. I tried earlier shutting the whole bot down and the bringing
it back up and online again, the problem still existed.

Is it possible that the ipv6-patch has modified something in a way
which makes bot not making itself aware of the "population" of
all the chans it sits? Or maybe something with version 1.6.18?
This is actually the second bot where I run 1.6.18 and this is
the first time this problem was found, the only difference between
these 2 bots is, that this problem bot uses also ipv6-patch and
I've no choice there I could of course downgrade to 1.6.15
or smthng, but lets see now how this continues to work.

Thanks!

PS. The problem bot still doesn't recognize for example channel
mode flags.

[r0t3n]

For the channel names, .dump MODE #channel

I have no idea if the patch changes anything, it shouldn't touch the channel module.

As the problem still exists, you could maybe add a few lines to the bottom of your eggdrops config file to execute the WHO,WHOIS,NAMES,MODE commands a few minutes after startup.

For example: (put this under your source lines in bots config file)

Code: Select all

timer 2 [list sync]

proc sync {} {
    foreach channel [channels] {
        if {![botonchan $channel] && ![channel get $channel inactive]} {
            putquick "JOIN $channel" -next
            putquick "WHO $channel"
            putquick "NAMES $channel"
            putquick "MODE $channel"
            putlog "Syncing $channel due to bot not being on channel and channel is not set inactive!"
        }
    }
}

This will execute itself 2 minutes after bot startup/restart/rehash and checks each channel, if the bot is not on the channel and the channel is not set inactive, it will execute a join, who, names and mode line for the channel.

[nml375]

May I ask which ipv6-patch you are using? There are a few out there that are seriously flawed and should not be used. One of the more common issues is breaking the global function export table, capable of causing apparently unrelated issues and crashes.

The last eggdrop with somewhat reliable ipv6 functionability is 1.6.13 as I recall - however this version has several other bugs.

With later eggdrops, it is recommended that you use some tunneling application such as 6tunnel whenever possible.

[x]

Thanks for the code Tosser^^!

nml375: Of course. I am using the patch by Free-bob, the one
which can be found on egghelp.org. There's a link on the front
page to this patch.

I also thought about tunneling app, but sadly it's not a possibility
in my case. I've one bot which has been using version 1.6.15
for a long time and successfully, nothing weird has happened
with that bot so I suppose I could start using this version.

[nml375]

Unfortunately, that patch is flawed, and will cause your bot to crash under certain conditions, and might very well be the cause for your problems aswell. The issue is somewhat documented here: http://forum.egghelp.org/viewtopic.php?t=12124

[x]

I hear you, thanks for the note! I switched to an earlier
release of Eggdrop, if everything works somewhat
well, I am happy enough

[nml375]

As I recall, there are some instructions in that thread on how to fix the known issues with that patch, although I cannot guarantee those are the only issues with that patch (which is also the reason why I have not released a fixed patch myself).