Daytona USA Progress

[ioncannon]

Spent a ton of time fixing bugs and getting player join/leave of the lobby/teams working correctly.

I've figured out the packets to start the session as well as to reconnect back to the lobby after one. Session of course doesn't start as we don't have a game server yet.





Only thing left for the lobby server now is:

* Some minor bugs when a player d/cs
* Still gotta figure out how player positions work. Currently you are always p1 regardless.
* Fix a bug with shared memory being set to all players in the team.

Then all that is left is the game server; which should be pretty isolated from the lobby.

[jial]

You are doing a great job, any effort about reviving Dreamcast services deserve our gratitude.

[Sonic1994DC]

Spent a ton of time fixing bugs and getting player join/leave of the lobby/teams working correctly.

I've figured out the packets to start the session as well as to reconnect back to the lobby after one. Session of course doesn't start as we don't have a game server yet.





Only thing left for the lobby server now is:

* Some minor bugs when a player d/cs
* Still gotta figure out how player positions work. Currently you are always p1 regardless.
* Fix a bug with shared memory being set to all players in the team.

Then all that is left is the game server; which should be pretty isolated from the lobby.

What a fantastic job you've done. How close would you say this is almost playable online again?

[ncman071]

wow totally impressed with these accomplishments! Daytona is such a great game. Hard to believe just how many games work online for the dreamcast

[ioncannon]

Not much to update in terms of features; but I did find all the socket library functions. Originally I had reverse engineered the low-level Access TCP lib that all these games used, but turns out the upper functions I had labeled were part of the old Berkley Sockets lib that is used in Windows, Unix, and Dreamcast (an implementation of it).

This let me map out the whole network stack for both the Lobby and Net portion of the game. I can see when a socket is made, bound, connected, and recv/send data, and if it is using UDP or TCP. This led me to find that there is a UDP socket being made for pinging individual clients; and completes the rest of the Player struct I was confused about.

Still looking at the NET ENGINE side of things and trying to figure out if it's Peer2Peer (clients connect into each other) or it's expecting a relay server of some kind. I do know non-host clients try to send a UDP packet into the ether on port 0x7F7F. It's just 4 bytes of random data.

[Sennar83]

Great. What is missing yet?

[rjbr]

Will we have this game online again soon?

[3style]

Great job! What is the progress approximately so far, more than 50% ?

[Sennar83]

Not much to update in terms of features; but I did find all the socket library functions. Originally I had reverse engineered the low-level Access TCP lib that all these games used, but turns out the upper functions I had labeled were part of the old Berkley Sockets lib that is used in Windows, Unix, and Dreamcast (an implementation of it).

This let me map out the whole network stack for both the Lobby and Net portion of the game. I can see when a socket is made, bound, connected, and recv/send data, and if it is using UDP or TCP. This led me to find that there is a UDP socket being made for pinging individual clients; and completes the rest of the Player struct I was confused about.

Still looking at the NET ENGINE side of things and trying to figure out if it's Peer2Peer (clients connect into each other) or it's expecting a relay server of some kind. I do know non-host clients try to send a UDP packet into the ether on port 0x7F7F. It's just 4 bytes of random data.

Hi

News?

[ncman071]

i'm sure it will be updated when more progress happens...i dont know what it takes...but i do know it takes so much effort in reversing these games/servers and on top of that...these people have their own lives/jobs/families they deal with in real life