chore!: change package name & asmdefs#1026
Conversation
|
@MFatihMAR have we tested what this PR would do to a project using the old package and would need to update to this package ? I suspect that its going to be somewhat of a pain because of the asmdef change and potentially the script links? I assume there will be some document somewhere to help folks make that migration? |
yeah, very good point. I haven't tested or wrote a doc about how to migrate yet. this PR is still a draft. lots of changes went into develop recently already. this PR is the last gate (except docs & md files, including CHANGELOG -should we change names there retrospectively?) before we finalize package renaming. so I'd expect this PR to be updated a few more times with fixes, guides and other issues. BUT, I wanted to keep the ball rolling here as I imagine this PR to be the last one :) back to migration doc, I could work with SA folks for that. I could also summarize my own "migration experience" as a baseline. package name change would break codebases for sure because devs now required to update dependencies section of their manifest.json and package.json files. there needs to be a place for these things. |
| { | ||
| "name": "com.unity.netcode.gameobjects", | ||
| "displayName": "Netcode for GameObjects", | ||
| "description": "Netcode for GameObjects is a high-level netcode SDK that provides networking capabilities to GameObject/MonoBehaviour workflows within Unity and sits on top of underlying transport layer.", |
There was a problem hiding this comment.
|
I have plans to address namespace/asmdef migration under a separate PR (with MovedFrom attrs & MigrateFromMLAPI.md doc). Can we get this in? |
JesseOlmer
left a comment
There was a problem hiding this comment.
lgtm except for a few unintended changes w/ comments
com.unity.netcode.gameobjects/Runtime/Messaging/MessageQueue/MessageQueueHistoryFrame.cs
Show resolved
Hide resolved
|
I notice we still have a |
…nsform * develop: (21 commits) test: adding more details to multiprocess readme (#1050) refactor!: convert NetworkTransform.NetworkState to `struct` (#1061) fix: networkmanager destroy on app quit (#1011) feat: snapshot, using unreliable packets, now that the underlying foundation supports it. Only merge after #1062 (#1064) feat: snapshot. Fully integrated despawn, mtt-1092, mtt-1056 (#1062) fix: eliminate bad use-after-free(destroy) pattern (#1068) chore: cleanup meta files for empty dirs (#1067) chore: minor MLAPI to Netcode rename (#1065) feat: report network behaviour name to the profiler (#1058) fix: player movement (#1063) test: Add unit tests for NetworkTime properties (#1053) chore: remove authority & netvar perms from NetworkTransform (#1059) feat: networktransform pos/rot/sca thresholds on state sync (#1055) feat: expose network behaviour type name internally (#1057) chore: remove all the old profiling code (#1048) fix: if-guard `NetworkManager.__rpc_name_table` access (#1056) fix: Disabling fixedupdate portion of SpawnRpcDespawn test because it's failing for known reasons that will be fixed in the IMessage refactor. (#1049) feat: Implement metrics for the new network profiler (#960) chore!: change package name & asmdefs (#1026) feat: per axis networktransform state sync (+bitwise state comp) (#1042) ... # Conflicts: # com.unity.netcode.gameobjects/Prototyping/NetworkTransform.cs
* chore: change package name * update Unity Transport for Netcode for GameObjects description * update manual.md * add `MigrateFromMLAPI.md` * mark public types with appropriate `[MovedFrom(...)]` attributes * add namespace * undo MovedFrom attrs * update package description * nuke Documentation~ * fix bad merge * put Documentation~ back (due to package validation) * update CODEOWNERS
I'll hold onto this PR for a while until I make sure both Yamato and UPM wouldn't break.