mirror of
https://github.com/rojo-rbx/rojo.git
synced 2026-10-23 05:51:39 +00:00
bfb5815f5e
- Removed outdated git configuration command from TODO.txt for clarity. - Added a command to set the `core.longpaths` configuration in the `git_init_repo` function, improving support for long file paths in the repository initialization process. These changes aim to streamline the TODO list and enhance the git initialization workflow.
102 lines
No EOL
6.6 KiB
Text
102 lines
No EOL
6.6 KiB
Text
[vscode extension]
|
|
get the schemas for everything into cursor somehow
|
|
|
|
[merging]
|
|
we should make the property diffs have arrows next to eachother so you can like "undo" certain things
|
|
need proper merging support in this UI.
|
|
make sure for fully deleted/added files we can use the merge UI as expected.
|
|
|
|
[syncback]
|
|
maybe make syncback automatically create a branch and merge it in, rather than overwriting your project files?
|
|
in order to do this, we must make dedup work flawlessly.
|
|
|
|
[Move tests to dedicated location]
|
|
Right now we have a severe amount of tests in the actual implementation files. I'd like to move those to a dedicated location because it's bloating the implementation files.
|
|
|
|
[Make instance creation/syncing more consistent and levelled up with how rbxsync does it]
|
|
* Terrain syncback / sync
|
|
* Order of operations (e.g. applying the Tag after full-sync routine complete so that things listening for tag updates get updated after it has been fully constructed, including it's children)
|
|
* CSG parts - requires very special handling to reproduce
|
|
* MeshPart creation must be done via CreateMeshPartAsync https://create.roblox.com/docs/reference/engine/classes/AssetService#CreateMeshPartAsync
|
|
* Special handling props (e.g. Skybox, ColorMap/ColorMapContent stuff)
|
|
* MeshId swaps for updates, need to re-create via CreateMeshPartAsync
|
|
* Queueing up material updates to properly support MaterialService changes so they appear correctly in studio without having to reboot
|
|
* Refer to rbxsync D:\UserGenerated\rbxsync\plugin\src\Serializer.luau and D:\UserGenerated\rbxsync\plugin\src\Sync.luau for reference, this system works correctly for the above use cases.
|
|
* You can search that whole rbxsync project for details/information.
|
|
|
|
[reify]
|
|
I noticed that in the reify/encodeInstance/decodeInstance/diff etc system we are not applying tags? Should the tag application / update behavior not be in here?
|
|
Tags should be applied once ALL updates have been performed on the instance
|
|
Tags should not be applied until EVERYTHING, including refs/attributes/properties are set.
|
|
Also, I noticed that attributes were not being doing in that system either, that should be done there I think?
|
|
Also, I noticed that it can fail to create an Instance - do we properly handle creating MeshPart's using InsertService?
|
|
Do we handle rebuilding CSG parts appropriately?
|
|
Do we handle swapping MeshId properly?
|
|
Are we performing operations in the correct order?
|
|
I feel like we added a lot of sync-logic that was supposed to be in these systems, but was placed elsewhere..
|
|
Refer to rbxsync D:\UserGenerated\rbxsync\plugin\src\Serializer.luau and D:\UserGenerated\rbxsync\plugin\src\Sync.luau for reference, this system works correctly for the above use cases.
|
|
You can search that whole rbxsync project for details/information.
|
|
|
|
[--!strict the plugin and make it --!native in hotpaths]
|
|
Need to get stuff inlined...
|
|
the entire syncing system needs to have it's types corrected and use strict,native,optimize.
|
|
we need to optimize the matching system heavily, it's so slow...
|
|
|
|
[Always mode]
|
|
MEDIUM - Issue 5: Ongoing "Always" mode patches don't support per-item selections yet
|
|
The existing code at ServeSession.lua line 311-314 explicitly says:
|
|
> "Table-type responses from the selection-based UI are not yet handled here. Ongoing WebSocket patches with per-item selections would need the same logic."
|
|
So in "Always" mode, ongoing patches currently only support "Accept"/"Abort"/"Skip" -- not per-item push/pull/ignore. The git defaults would be set in the PatchTree but the user can't act on them for ongoing patches. This is a pre-existing limitation, not introduced by us, but worth noting.
|
|
|
|
[two-way sync-back non scripts with a checkbox?]
|
|
we should just two-way sync everything backwards property wise, its really about not forward-syncing properties when that scripts only mode is on..
|
|
|
|
[testing which actually runs studio, syncs, etc]
|
|
we could make an automated testing system which truly simulates opening a place file and syncing to a running server automatically
|
|
this would close the gap on testing HEAVILY
|
|
use run-in-roblox utility like we use everywhere else.
|
|
so this expands beyond just that TestEZ system, where it tests spec files. this is separate from that, but same idea, just expanded upon.
|
|
|
|
[two way sync]
|
|
Make the two-way sync system respect the syncback configuration fully
|
|
ensure that live-syncback respects the syncback configuration fully
|
|
|
|
[refs]
|
|
-- Ref properties cannot be encoded during instance addition because the
|
|
-- target instance has no server ID yet. They are handled separately by
|
|
-- encodePatchUpdate.lua via the InstanceMap.
|
|
if descriptor.dataType == "Ref" then return end
|
|
I thought we were supposed to generate a uid and then send that table to the server so it can resolve those based on like a client uid?
|
|
|
|
[cli inspection]
|
|
we should make commands for inspecting .rbxl and .rbxm files for ai agents
|
|
even inspecting ones from the website directly
|
|
could we find an endpoint for downloading the "client side" of any game on the platform for context?
|
|
|
|
[Crates]
|
|
publish as a crate https://crates.io/crates/rojo
|
|
|
|
[vscode file watcher]
|
|
Also, we need to explore how VSCode deals with this. They seem to have ZERO issues.
|
|
Maybe we can have a mode to tie into the vscode file watcher ourselves? luau-lsp does this.
|
|
|
|
[fix dedup]
|
|
dedup is broken due to relying on sort order
|
|
recreate broken state by forcing random order in all matching dedup systems of the incoming children list
|
|
if it cannot match correctly with randomization, then it is inherently broken
|
|
in all 3 matchers, simply randomize the order of the children - this will break the tests.
|
|
|
|
[syncback via extension]
|
|
can we allow the terminal to stay so i can like run some git commands myself without having to make a whole new terminal?
|
|
i also want to just cleanup terminals / use one terminal for syncback and serve...
|
|
|
|
[skill]
|
|
'/playtest-quick'
|
|
simply runs the game and checks for literal errors, typically created just after completing a code change
|
|
resolve these errors in the project, or if in plan mode create a plan to resolve the errors
|
|
this should utilize the mcp to run the game and fetch the errors, make sure the server is fully booted up (INTELLIGENT 5-10s wait) and then fix stop the game
|
|
|
|
modify /atlas-push to always trigger /playtest-quick after it finishes pushing, looping back around to fix issues if needed. make sure the errors are related to what we actually pushed, so it doesnt try to fix others work. if it's others work, ask the user if they want to fix it - thought recommend they should do a planning phase.
|
|
"/atlas-push no test" or similar it just skips the automated test
|
|
|
|
make clone support multi-placeids like we did with defend-67 |