master
6
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 EmilandClaude Sonnet 5
|
cd8f221ddd |
Fix real bugs from independent review (docs/review-handoff.md)
A second AI session read the whole codebase in parallel (read-only, no code changes) and left a handoff doc. Addressed the correctness findings: - PhysicsWorld.Sync's early return compared _bodies.Count to live.Count, not their contents — same size, different membership (destroy one tracked GameObject, gain one untracked-because-no-collider one; or any Restore where the scene has both a Rigidbody+collider object and a Rigidbody-without-one) skipped cleanup entirely, leaking the native Box3D body forever. Fixed by checking the actual stale set. Two new regression tests reproduce the review's own two scenarios via a new Lingua_GetBodyCount native export, asserting on the native table's own count rather than PhysicsWorld's C#-side bookkeeping. - TryCreateBody stored handle -1 (native shim refused: invalid world, or its 8192-slot body table full) as if it were real — every later GetBodyTransform on it silently teleported the GameObject to the origin with a degenerate rotation, no error anywhere. Now checked and warned once, same as the missing-collider case. - PluginHost.Load didn't roll back anything when Configure threw partway through: Schedule.Add/Events.Subscribe registrations it already made stayed forever, and its ALC was never unloaded — neither loaded (no _loaded entry) nor cleanly unloadable. Fixed with try/catch: best-effort Shutdown (the only thing that knows which services this plugin provided), RemoveAllFrom on both Schedule and EventBus, best-effort ALC unload, rethrow. New fixture plugin (sandbox.failing-configure, mirrors sandbox.echo's own real-load pattern) registers a system against a shared Ping component then throws, so FailedConfigureRollbackTests can assert the dangling system actually stops firing — an earlier version tried to prove this via AssemblyLoadContext.All instead, which passed even against the deliberately-reverted buggy code (the ALC turned out to get collected either way once its only references went out of scope); watching the dangling system is what actually distinguishes rolled-back from not, confirmed by deliberately reverting the fix and watching this specific test fail before restoring it. - Engine.Host's "r <id>" left a plugin unloaded on a failed reload with no honest indication of that, and retrying threw "not loaded" instead of ever reaching Load again. Added PluginHost.IsLoaded so the handler only calls Unload when there's something to unload, and the failure message now says the plugin is unloaded, not just "failed." - AssetService.ReloadWithRetry's `when (attempt < 4)` guard meant the 5th and final IOException fell out of the loop and propagated from a discarded fire-and-forget Task — no log, no event, nothing. Now logged. - EditorState.Selected kept pointing at a GameObject Restore had already destroyed after ExitPlay, so Inspector/gizmo would silently keep editing something no longer in the world. EditorPlugin.DrawUi now compares IsPlaying against its own previous frame (not just reacting to the Stop button) so this is caught whether Play was exited via the button or the "stop" stdin command — the same stdin-vs-real-control gap already hit once earlier this session — and re-resolves the selection by name. - README.md and kernel-contract.md's Event Bus/Time rows had fallen a milestone behind (still said "no asset system yet" and "no fixed-step accumulator yet" after both shipped). Full suite: 98 tests. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01N1qPfzq8TDCUMFMV3UwV5N |
||
|
EmilandClaude Sonnet 5
|
e0e6c1d503 |
M3: real 3D translate gizmo — the "полноценный 3D-пайплайн" it was chosen for
TranslateGizmo draws three axis handles at the selected GameObject's world position by projecting real 3D points through the real camera's View/Projection (GizmoMath.WorldToScreen) onto ImGui's foreground draw list. Nothing OpenGL-side draws lines yet, so this isn't 3D geometry in the GL sense — but it IS driven by the actual camera matrices, which is what the earlier scoped-2D-vs-full-3D-pipeline choice was actually about: a handle dragged in screen space has to map onto a real 3D axis, and that only means something once there's a real camera to project through. The screenshot below shows exactly that — the axes aren't screen-perpendicular, because the camera at (0,3,6) looking at the origin means they shouldn't be. Dragging a handle re-projects the mouse delta onto the axis's own screen-space direction (GizmoMath.ProjectDragOntoAxis) and writes the result back through GizmoMath.WorldToLocalPosition, which inverts the parent's WorldMatrix rather than writing LocalPosition directly — a parent with non-identity scale or rotation means "move 1 world unit" and "add 1 to LocalPosition" are different amounts, and samples/WindowDemo's ChildQuad (parented under a (2,2,1)-scaled Quad) is exactly that case. Split the actual math into GizmoMath (Engine.Editor.Contracts, no GL/ ImGui/mouse dependency) so it's unit-testable without a window or a real mouse — neither exists in a headless test run, and dragging is exactly the kind of interaction that's easy to get subtly wrong (screen-space ratio direction, perspective sign, parent-scale correctness) without something to check it against beyond eyeballing a screenshot. New Engine.Editor.Tests project, 8 tests: screen-center projection, a behind-camera point returning null, drag-ratio math on both an axis-aligned and a diagonal screen direction, and the parent-scale/ parent-translation inverse-transform cases. All 8 passed on the first run — including the non-uniform-scale case, the one most likely to be subtly wrong. Full suite: 73 tests, all green (65 previous + 8 new). Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01N1qPfzq8TDCUMFMV3UwV5N |
||
|
EmilandClaude Sonnet 5
|
d7fcc20f5e |
M3: Play/Stop — Engine.Host stops ticking Stage.Update outside Play mode
IPlayModeController (Engine.Editor.Contracts) wraps IWorld.Snapshot()/ Restore() as EnterPlay/ExitPlay — the entire mechanism is that one snapshot, per Snapshot's own doc comment. Exposed as a service, not just something engine.editor's UI calls directly, so a future non-UI driver (a test harness, a headless "play for N frames" CLI command) can drive Play mode without depending on ImGui. Engine.Host now looks up IPlayModeController once and checks IsPlaying every frame before running Stage.Update: Edit mode still renders the scene (so the view isn't frozen and the editor UI stays responsive) but never ticks it, same distinction Unity draws between its Scene and Game views. A project with no engine.editor loaded sees no behavior change — Update runs unconditionally, same as before this existed. Added "play"/"stop" as stdin commands alongside the existing "r" and "screenshot", both real controls (not just test scaffolding) for driving Play mode without a mouse — which is also how this got verified: an interactive run sent "play", screenshotted, sent "stop", and the log shows "Entered Play mode in 13.2ms", the real editor path exercising the same under-100ms budget WorldSnapshotTests already proves at the kernel level. The Lingua Editor panel's button/label flip Play/(Edit mode) to Stop/(Playing) correctly across both screenshots. Also fixes a real layout bug hit while verifying this: SetNextWindowPos with ImGuiCond.FirstUseEver only applies with no prior imgui.ini entry for that window title, and this project had already accumulated one from earlier runs (all three panels stacked exactly on top of each other) — gave each panel its own default position and .gitignore'd imgui.ini, which is per-machine session state, not source. Full suite still green: 65 tests. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01N1qPfzq8TDCUMFMV3UwV5N |
||
|
EmilandClaude Sonnet 5
|
20e0f8c53f |
M3: reflection-based Inspector — edit any component's fields live
InspectorPanel shows EditorState.Selected's Transform (Position/Scale as draggable Vector3 widgets, Rotation shown read-only as a quaternion — a real rotate control needs Euler conversion or a gizmo, out of scope for just the Inspector) and, below it, every attached Component's public fields via reflection. No per-component-type drawer code exists anywhere: int/float/bool/string/Vector3 fields get a live-editable widget, anything else falls back to a read-only ToString(), and a brand new component type in any plugin gets an Inspector for free the moment it's attached — that genericity is the entire reason to do this by reflection instead of a registry. Field edits write straight back through FieldInfo.SetValue, same gap GameObject.AddComponent's doc comment already names: components are plain fields with nothing to intercept a direct mutation, so this needs no SystemAccessScope declaration any more than a hand-written `component. Value = 5` would. EditorPlugin now defaults EditorState.Selected to the first root once, if nothing's been clicked yet — an empty Inspector on every fresh launch had nothing useful to show. Verified by screenshot: Inspector displays "Quad"'s real Position (0,0,0) and Scale (2,2,1), matching scene.json exactly, plus a QuadRenderer header correctly showing "(no fields)" since it has none yet. Full suite still green: 65 tests. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01N1qPfzq8TDCUMFMV3UwV5N |
||
|
EmilandClaude Sonnet 5
|
bccb8c3ba7 |
M3: Hierarchy panel — click a GameObject to select it
HierarchyPanel walks IWorld.Roots/GameObject.Children recursively as ImGui tree nodes, clicking one sets EditorState.Selected — the shared selection the not-yet-built Inspector panel will read from the same instance. PushID(go.GetHashCode()) scopes each node's ID by object identity rather than name, since nothing stops two sibling GameObjects sharing a Name and ImGui's default label-based IDs would otherwise merge their open/selected state. Also gives samples/WindowDemo/scene.json a child GameObject (offset, half-scale, parented under Quad) — the existing scene only had one root, nothing to show a tree with. Verified by screenshot: both quads render at their correct composed WorldMatrix (the child visibly smaller and offset, confirming parent/child composition is still correct through this change), and Hierarchy lists "Quad" as a collapsible node. Full suite still green: 65 tests. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01N1qPfzq8TDCUMFMV3UwV5N |
||
|
EmilandClaude Sonnet 5
|
4c67c2c905 |
M3: editor shell boots — ImGui overlay drawn over the live scene
Introduces engine.editor, a new plugin that renders an ImGui panel on top of whatever the scene is drawing, using Silk.NET.OpenGL.Extensions.ImGui's ImGuiController against the same GL context and window engine.render already owns. Getting the overlay to actually show up in the same frame (not delayed by one, which is what happens if UI draws after SwapBuffers) required splitting engine.render's Draw system: SwapBuffers moves out into its own Present system on a new Stage.Present, run by Engine.Host after every Stage.Render system — this plugin's Draw and, now, engine.editor's ImGui pass — has drawn into the same back buffer. Stage.Present is the second addition to the frame stage set (after FixedUpdate was scoped out for M4); docs/kernel-contract.md already committed to the set being fixed and kernel-defined, not plugin-extensible, and to only adding a stage when there's something real to run in it — SwapBuffers already needed somewhere to live once a second Render-stage system existed to race it. Also extends IEngineInput with a Native IInputContext property, the same escape hatch IEngineWindow.Native already is — ImGuiController's constructor needs the raw Silk.NET input context directly. Verified with a real windowed run: --screenshot after 5 frames shows the "Lingua" ImGui panel (FPS/frame counters) correctly composited over the still-correct 3D checkerboard quad from the M3 camera pipeline, in the same frame. Full suite still green: 65 tests. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01N1qPfzq8TDCUMFMV3UwV5N |