[DiscordArchive] There is alot of containers that are local to map, e.g. what if 2 of your threads spawns a creature
[DiscordArchive] There is alot of containers that are local to map, e.g. what if 2 of your threads spawns a creature
Original source
There is alot of containers that are local to map, e.g. what if 2 of your threads spawns a creature at the same time on the map ?
rektbyfaith
04-02-2024, 03:08 PM #1Archived author: shelby • Posted: 2024-04-02T15:08:58.622000+00:00
Original source
There is alot of containers that are local to map, e.g. what if 2 of your threads spawns a creature at the same time on the map ?
Original source
There is a grid container but also a map local to the Map for the guid look up
rektbyfaith
04-02-2024, 03:09 PM #2Archived author: shelby • Posted: 2024-04-02T15:09:12.401000+00:00
Original source
There is a grid container but also a map local to the Map for the guid look up
Original source
And that just a exemple but there is alot of things like that
rektbyfaith
04-02-2024, 03:09 PM #3Archived author: shelby • Posted: 2024-04-02T15:09:40.229000+00:00
Original source
And that just a exemple but there is alot of things like that
Original source
What if there is a TP local to the map in the middle of the update that put you out of your thread, then your unit will deal with non safe unit etc
rektbyfaith
04-02-2024, 03:10 PM #4Archived author: shelby • Posted: 2024-04-02T15:10:33.960000+00:00
Original source
What if there is a TP local to the map in the middle of the update that put you out of your thread, then your unit will deal with non safe unit etc
Original source
I think its really difficult to achieve
rektbyfaith
04-02-2024, 03:10 PM #5Archived author: shelby • Posted: 2024-04-02T15:10:39.706000+00:00
Original source
I think its really difficult to achieve
Original source
vmangos still has this, reportedly working fine in production. Not a lot of code as well, continent areas are instanced, players moving across areas are switched from one instance to another after map updates
rektbyfaith
04-02-2024, 03:18 PM #6Archived author: Shudza • Posted: 2024-04-02T15:18:43.305000+00:00
Original source
vmangos still has this, reportedly working fine in production. Not a lot of code as well, continent areas are instanced, players moving across areas are switched from one instance to another after map updates
Original source
Hi all, I am learning MMO netcode architecture and am reading the TC code since yesterday. I noticed when a user moves they send a move packet with the client timestamp. That gets converted to server time on the server and then used in packets that get send back to the client. Is this correct? (I assume timestamps are also used for things like spell cast timing etc?). I am wondering what prevent a client from manipulating their timestamps? And how exactly is this timestamp then used on the client (when received from the server, let's say you get the move packet from your friend). I hope my question is a little bit clear.
rektbyfaith
04-02-2024, 03:24 PM #7Archived author: Balthazar • Posted: 2024-04-02T15:24:35.472000+00:00
Original source
Hi all, I am learning MMO netcode architecture and am reading the TC code since yesterday. I noticed when a user moves they send a move packet with the client timestamp. That gets converted to server time on the server and then used in packets that get send back to the client. Is this correct? (I assume timestamps are also used for things like spell cast timing etc?). I am wondering what prevent a client from manipulating their timestamps? And how exactly is this timestamp then used on the client (when received from the server, let's say you get the move packet from your friend). I hope my question is a little bit clear.
Original source
How would a client misuse this if
- clock is monotonic
- too big gaps in clocks are filtered?
- server can always ignore packets arriving after tick
rektbyfaith
04-02-2024, 03:45 PM #8Archived author: Kakukkokkal • Posted: 2024-04-02T15:45:05.407000+00:00
Original source
How would a client misuse this if
- clock is monotonic
- too big gaps in clocks are filtered?
- server can always ignore packets arriving after tick
Original source
I haven't checked this part of the TC code yet, but a naive implementation could be to have e.g a 20Hz server tick rate, every packet which has the timestamp between the last and the next tick time is handled (and taking care of monotonity, etc), others are rejected/ignored/whatever
rektbyfaith
04-02-2024, 03:47 PM #9Archived author: Kakukkokkal • Posted: 2024-04-02T15:47:16.529000+00:00
Original source
I haven't checked this part of the TC code yet, but a naive implementation could be to have e.g a 20Hz server tick rate, every packet which has the timestamp between the last and the next tick time is handled (and taking care of monotonity, etc), others are rejected/ignored/whatever
Original source
https://www.snapnet.dev/blog/netcode-arc...-lockstep/
[Embed: Netcode Architectures Part 1: Lockstep]
Explore what's commonly referred to as "lockstep" netcode architecture
https://www.snapnet.dev/blog/netcode-arc...-lockstep/
rektbyfaith
04-02-2024, 03:48 PM #10Archived author: Kakukkokkal • Posted: 2024-04-02T15:48:01.411000+00:00
Original source
https://www.snapnet.dev/blog/netcode-arc...-lockstep/
[Embed: Netcode Architectures Part 1: Lockstep]
Explore what's commonly referred to as "lockstep" netcode architecture
https://www.snapnet.dev/blog/netcode-arc...-lockstep/
1 Guest(s)
1 Guest(s)