Forged Alliance Forever

Re: some new intel on the lobby bug

2012-08-02T15:31:51+02:00

Ze_PilOt wrote:
The server send a "disconnectFromPeer" command to forget them.

Also, I've seen crashes with only people connecting, (like 3-4, no disconnection before), and all correctly (well except the last one because he produces a crashs, but on a rehost he connect correctly).

the server might forget them, FA.exe might still keep info on them (e.g. remember the result of the reverse name lookup).
if your crash involved less ppl, than it is a different bug.

Statistics: Posted by rootbeer23 — 02 Aug 2012, 15:31

Re: some new intel on the lobby bug

2012-08-02T15:28:19+02:00

The server send a "disconnectFromPeer" command to forget them.

Also, I've seen crashes with only people connecting, (like 3-4, no disconnection before), and all correctly (well except the last one because he produces a crashs, but on a rehost he connect correctly).

Statistics: Posted by Ze_PilOt — 02 Aug 2012, 15:28

Re: some new intel on the lobby bug

2012-08-02T15:22:09+02:00

Ze_PilOt wrote:
I still don't understand the value "12" in the code, relative to the connections.
The 12 players lobby is an arbitrary value Moritz used. There is also a 16 players lobby and it could have been 10.

It's look like a strange coincidence.

Also, all these limits are supposed to be taken from this file :
https://bitbucket.org/thepilot/forged-a ... byComm.lua

The mystery remains. It is not uncommon to see an STL implementation choose different algorithms depending on the size of the input, i.e. use quicksort for vectors larger than a few elements, use another sorting algorithm for smaller vectors.
In our case 12 is also not a limit of anything, the 13th player you connect to is just as good as the other 12.
For clarification: the counter in question is also not strictly the number of connected players, but the number of unique players that were ever connected to the game. FA seems to remember peers, indicated by the log message:
"Disconnected from ..., waiting for them to reconnect"

Statistics: Posted by rootbeer23 — 02 Aug 2012, 15:22

Re: some new intel on the lobby bug

2012-08-02T09:22:28+02:00

I still don't understand the value "12" in the code, relative to the connections.
The 12 players lobby is an arbitrary value Moritz used. There is also a 16 players lobby and it could have been 10.

It's look like a strange coincidence.

Also, all these limits are supposed to be taken from this file :
https://bitbucket.org/thepilot/forged-a ... byComm.lua

Statistics: Posted by Ze_PilOt — 02 Aug 2012, 09:22

Re: some new intel on the lobby bug

2012-08-02T06:54:38+02:00

Ze_PilOt wrote:
I confirm, some guys still have crashes.

Yeah, there is more than one bug in the engine. For me the lobby-full crash made up 95% of the trouble, i like big games.
If a crash does not happen to me i cannot try to fix it. And next time we might not be so lucky. On a difficulty scale of 1 to 10
this one was just a 1.2

Statistics: Posted by rootbeer23 — 02 Aug 2012, 06:54

Re: some new intel on the lobby bug

2012-08-01T21:45:38+02:00

I confirm, some guys still have crashes.

Statistics: Posted by Ze_PilOt — 01 Aug 2012, 21:45

Re: some new intel on the lobby bug

2012-08-01T19:15:03+02:00

It's live.

Statistics: Posted by Ze_PilOt — 01 Aug 2012, 19:15

Re: some new intel on the lobby bug

2012-08-01T14:23:09+02:00

Yes of course

Was just trying to not put your expectations too high, I hope it fix it, but I have the feeling that rootbeer job is not over yet

Statistics: Posted by Ze_PilOt — 01 Aug 2012, 14:23

Re: some new intel on the lobby bug

2012-08-01T14:15:38+02:00

Well if it fixes the problem then thats amazing. If it fixes anything then good. If it fixes nothing then still three cheers because he's putting the effort in and thats to be celebrated in itself- an example to us all.

Statistics: Posted by Softly — 01 Aug 2012, 14:15

Re: some new intel on the lobby bug

2012-08-01T14:13:00+02:00

If it's fix the problem, it's really great. But I have the feeling that it's not the only problem here..

Statistics: Posted by Ze_PilOt — 01 Aug 2012, 14:13

Re: some new intel on the lobby bug

2012-08-01T14:12:17+02:00

Nice work root. Three cheers I say.

Statistics: Posted by Softly — 01 Aug 2012, 14:12

Re: some new intel on the lobby bug

2012-08-01T13:26:49+02:00

Thanks. I will put that online tonight (in about 6 hours).

Statistics: Posted by Ze_PilOt — 01 Aug 2012, 13:26

Re: some new intel on the lobby bug

2012-08-01T12:52:56+02:00

http://www.filedropper.com/forgedalliance

Statistics: Posted by rootbeer23 — 01 Aug 2012, 12:52

Re: some new intel on the lobby bug

2012-08-01T12:42:08+02:00

You know what ? Send me the .exe

Statistics: Posted by Ze_PilOt — 01 Aug 2012, 12:42

Re: some new intel on the lobby bug

2012-08-01T12:36:13+02:00

Ze_PilOt wrote:
Can you give me the offset in a editor like Hex Neo (http://www.hhdsoftware.com/free-hex-editor) please ?

lolwhut ^^
the offset is 523842, that is the 523842th byte of the file.
the 16 hex characters are the immediate context of that location from 523840 - 523856

Statistics: Posted by rootbeer23 — 01 Aug 2012, 12:36