more feature requests

[dac]

1) ability for the host of a game to get the IP of anybody who joins his game.
2) ability for the host to wank by ip address
3) ability for the host to wank in game pregame lobby

[punkUser]

#2 would have to be a metaserver feature, not Myth and #3 would be problematic because once you're in the pregame lobby you are no longer connected to the metaserver (in terms of chat, etc).

#1 is potentially doable at least for the host, but without the other features, how much use would it be? It also might not work easily for the clients.

[dac]

if you join a game retain an int (userid) default to 0 if not provided, if you join a game send userid in with you to the join game params. Wank by userid command is passed to the server if userid of target is > 0. Obviously you still have connection information because if you leave game you connect to the metaserver again so dont tell me its not doable.

you could add the same chatroom abilities in pregame built into the client as you do the metaserver. if you can boot you probably have IP information and you could probably add a button to display that or soemthing. hosts have a right to know who is in their games.

if textentered is a command then
1) forward command to metaserver lobby
2) forward output back to game lobby

has to be possible especially since the metaserver has you dimmed in a game when you're in-game.

[BIG KROK V8 SS]

I disapprove of all of these and tell you that I know 100% none of these would ever work.

[punkUser]

It is probably possible to auto-kick people from your host based on a local wanklist, but that won't operate the same way as the mariusnet list (i.e. not even show games hosted by that player). Ideally we shouldn't re-implement the same feature on the client.

It may be possible to forward server commands back to mariusnet when entered in a lobby but I'm not sure. You can't just assume that it must work since it's totally possible that mariusnet does not accept chat/commands from users who are in games. It may, but it depends on how the protocol was written.

As I mentioned, yes, the host can probably display IPs. Not necessarily any of the other players though as they only have a connection to the host.

The metaserver having you "dimmed" doesn't actually mean anything about whether there's still a connection there or not. There probably is, but it could conceptually be a simple state machine too (i.e. I see a use go into a game, assume he's in that game until I am told otherwise).

[dac]

except that when you crash the account will sign itself off after a few minutes - this implies a stayalive message.

i understand the protocol you're talking about, but looking at the code it accepts a packet and pushes one back. thats all the server is doing. the .commands send the packet back to the person who hit .whatever so that you dont spam the lobby with somebody's commands (great) which means you get a packet to somewhere, from somewhere, and that means you can trigger the packet being sent whether or not you're actually connected to the metaserver.

and thats why magma would have to be involved in such a feature since the client would have to be changed, the metaserver actually would go untouched in terms of allowing metaserver commands in a game lobby.

[punkUser]

... which means you get a packet to somewhere, from somewhere, and that means you can trigger the packet being sent whether or not you're actually connected to the metaserver.

No, I'm pretty sure it's a TCP connection that this is all sent over, so you need a connection live. That said, if it's not doing any checking on the state of the person sending the message, it may well be possible to send it from within a lobby (or even within a game). I haven't looked at that code in detail.

Also if we want this to work on mariusnet we'll have to test it there. AFAIK they reverse engineered the protocol rather than implemented it from the Bungie source, so they may or may not do things the same way in places where functionality has never been used/tested in past versions of Myth.

[BIG KROK V8 SS]

wanklist is unconstitutional. ban everyone who uses it.

[punkUser]

Should we remove boot too Kirk?

[BIG KROK V8 SS]

No, because boot isn't permanent; its the perfect solution to get someone to stop misbehaving. I actually don't boot anyone, even people who boot me. By booting/wanking in return, they continue their original behaviors.

[dac]

what makes you think you can test it there? ppe doesnt care about the server, he hardly ever checks it and doesnt reply to posts or emails.

if the server is getting unknown packets it should be able to throw them out and ignore them without failing. put a response timeout on the client (which i assume is built in for these communications anyways) and blue bar accordingly.

[punkUser]

what makes you think you can test it there? ppe doesnt care about the server, he hardly ever checks it and doesnt reply to posts or emails.

Right, but testing whether the server will accept chat commands from a user who is ingame does not require the participation of mariusnet

if the server is getting unknown packets it should be able to throw them out and ignore them without failing. put a response timeout on the client (which i assume is built in for these communications anyways) and blue bar accordingly.

I think you misunderstood what I meant. I was just saying that we don't know for sure if the server accepts chat from ingame players (even if a connection is present, which I agree is likely). This behavior could also be different between mariusnet and the Bungie code. I'm just saying it needs testing - we can't conclude based on anything observed that it *must* work, since that functionality has never been used by the Myth client.

I'll also note that adding such a feature does cross a minor line in the sand that has to my knowledge not yet been crossed, namely effectively incorporating the relevant "." commands into the metaserver protocol. i.e. if we make it pass ".wank" commands up the metaserver, then that's making an implicit assumption that the metaserver understands ".wank", which to my knowledge was never a part of the original metaserver code. Of course nothing will blow up if it doesn't (although if a given metaserver doesn't recognize . commands at all for instance, it could spam them to the chat lobby), but it's now a part of the protocol for any future metaservers. That's probably fine, but that raises the awkward question of who is in charge of the metaserver protocol? All theoretical of course, but these are the sorts of issues one always has to consider when making modifications to network protocols.

[dac]

agree that it may not accept it from uesrs not "active" in lobby for some reason, but that seems unlikely and okay requires testing. fair enough.

disagree that adding the .command would do anything. go try to type .abcdefg in the lobby and you'll get a blue bar that says unrecognized command. this happened on bungie too. clearly there's a handler for unknown command types so forwarding to lobby should still be viable. you dont want to bypass that anyways becuase thats where the privileges (sp?) are checked.

[punkUser]

disagree that adding the .command would do anything. go try to type .abcdefg in the lobby and you'll get a blue bar that says unrecognized command. this happened on bungie too. clearly there's a handler for unknown command types so forwarding to lobby should still be viable. you dont want to bypass that anyways becuase thats where the privileges (sp?) are checked.

If it's already part of the protocol then that's good as it at least won't cause harm. Still, any commands that Myth forwards become "standards" for any other implementation of the metaserver as well, so is it fair for us (or mariusnet, or anyone) to decide that "wank" is the official name of the command for now and forever?

I guess an alternative would be for Myth to simply forward any . commands that it doesn't recognize as local lobby commands back up to the metaserver. Or I guess we could even have a special prefix/notation for metaserver commands (could be just "." and change the local game ones) to differentiate them from local game ones. That would be a bit more extensible in the future, although it could possibly pose problems if the metaserver has commands that only make sense in the lobby. Worth some thought.