com.unity.services.samples.game-lobby

文件历史

nathaniel.buck@unity3d.com 8306b622 Basic solution to that bug with the Relay disconnect not happening. This ensures that the disconnects will happen eventually, as the host stops pinging relay on disconnect and also clients will request disconnection from hosts immediately if possible. However, a few more issues to consider addressing: We could boot the host from the lobby manually if a relay disconnect is detected. A client could leave the lobby immediately if the host has left. We need to clean up properly so that the automatic disconnect can be done repeatedly.		3 年前
..
AsyncRequestRelay.cs	Merging the error code exception branch.	3 年前
AsyncRequestRelay.cs.meta	Changing the error pop-up to handle the LobbyServiceException with more player-facing error output. The same work for Relay will follow if the package is in a usable state.	3 年前
RelayAPIInterface.cs	Merging the error code exception branch.	3 年前
RelayAPIInterface.cs.meta	Latest Dev and Readme without Vivox	3 年前
RelayUtpClient.cs	Basic solution to that bug with the Relay disconnect not happening. This ensures that the disconnects will happen eventually, as the host stops pinging relay on disconnect and also clients will request disconnection from hosts immediately if possible. However, a few more issues to consider addressing: We could boot the host from the lobby manually if a relay disconnect is detected. A client could leave the lobby immediately if the host has left. We need to clean up properly so that the automatic disconnect can be done repeatedly.	3 年前
RelayUtpClient.cs.meta	Still partial progress. I have a little handoff going where the client identifies itself to the server and the server responds with an ID to use in further communication. However, I don't think that will actually work since all clients would also need those IDs? I can just send the user ID, which takes up more bandwidth but not dramatically so. Also, this pinpoints the heartbeat for keeping the Relay allocation up. I'm trying to pull out all the job system things, since that both seems like an extra layer of complexity that devs would have to learn and it also isn't, like, actually used correctly in the sample?	3 年前
RelayUtpHost.cs	Basic solution to that bug with the Relay disconnect not happening. This ensures that the disconnects will happen eventually, as the host stops pinging relay on disconnect and also clients will request disconnection from hosts immediately if possible. However, a few more issues to consider addressing: We could boot the host from the lobby manually if a relay disconnect is detected. A client could leave the lobby immediately if the host has left. We need to clean up properly so that the automatic disconnect can be done repeatedly.	3 年前
RelayUtpHost.cs.meta	Still partial progress. I have a little handoff going where the client identifies itself to the server and the server responds with an ID to use in further communication. However, I don't think that will actually work since all clients would also need those IDs? I can just send the user ID, which takes up more bandwidth but not dramatically so. Also, this pinpoints the heartbeat for keeping the Relay allocation up. I'm trying to pull out all the job system things, since that both seems like an extra layer of complexity that devs would have to learn and it also isn't, like, actually used correctly in the sample?	3 年前
RelayUtpSetup.cs	Basic solution to that bug with the Relay disconnect not happening. This ensures that the disconnects will happen eventually, as the host stops pinging relay on disconnect and also clients will request disconnection from hosts immediately if possible. However, a few more issues to consider addressing: We could boot the host from the lobby manually if a relay disconnect is detected. A client could leave the lobby immediately if the host has left. We need to clean up properly so that the automatic disconnect can be done repeatedly.	3 年前
RelayUtpSetup.cs.meta	Partial progress: Following a trio of Relay+UTP samples, I seem to have something that will bind a transport to Relay and keep the connection open (as evident from the fact that a client can join after what would be the 10s timeout on Relay and successfully get the JoinAllocation, when they wouldn't without keeping the connection open even when binding was correct). However, this current code is very slapdash and sloppy; I'm just committing now as a very rough draft.	3 年前