acdream/tests/AcDream.App.Tests/UI/Layout/LayoutImporterTests.cs
Erik ceec3bc440 feat(render): Campaign V slice V4a - port TextRenderer/BitmapFont/DebugLineRenderer/TextureCache onto IGpuDevice
TextRenderer, BitmapFont, DebugLineRenderer, and TextureCache's UI-texture
upload path (GetOrUploadRenderSurface/UploadRgba8) now issue every draw and
resource creation through the pinned IGpuDevice/IGpuFrame/IGpuPassEncoder
RHI contract instead of raw GL. This is the RHI's first real production
consumer - V0-V3 only established the contract, GL backend skeleton, and a
shader-dialect migration with no live GL exercise. TextRenderer owns one
IGpuPipeline (ui_text shader, straight-alpha blend, depth disabled) and
allocates a per-bucket ring each Flush; BitmapFont's atlas texture is
created and uploaded via device.CreateTexture/.Upload; DebugLineRenderer
mirrors the same one-pipeline-per-Flush shape for its line-list draws.
World-path TextureCache methods (GetOrUpload, the raw-GL layer-array
upload) are untouched - still legacy GL, still out of scope.

Frame lifecycle: GpuDeviceFrameLifetime (RenderFrameOrchestrator.cs) wraps
IGpuDevice.BeginFrame()/IGpuFrame.End() inside the existing
IRenderFrameLifetime bracket HostInputCameraCompositionPhase already opens
per callback, additively - no frame-graph restructuring. Ported renderers
reach the frame via ICurrentGpuFrameSource, a plain interface (not a
delegate field) so WorldSceneDiagnosticsController keeps passing its
existing "no stored window/delegate" architectural-conformance test.

Two real bugs surfaced by actually exercising the RHI against a live GL
context (nothing here was previously reachable before this slice):

- GlGpuDevice.BeginFrame() now resets the render-state cache every frame.
  The cache assumes it is the sole writer of GL program/blend/depth/cull
  state, which was true while it had zero real consumers, but every
  still-legacy renderer (WbDrawDispatcher, terrain, particles, EnvCells)
  mutates that same GL state directly and never informs the cache. Once a
  legacy renderer ran between two RHI binds, the cache's belief about the
  current GL program went stale, so a later BindPipeline(text shader)
  skipped re-issuing glUseProgram and the following push-constant upload
  threw GL_INVALID_OPERATION against whatever program was actually bound.
  Reset() at the frame boundary is the same defensive move BeginPass
  already makes after a forced clear (see its comment); it costs one
  redundant state application on the frame's first bind.
- GL_MULTISAMPLE has no representation in the pinned contract. Added a
  GL-backend-internal Multisample field to GlRenderStateSnapshot/Changes,
  computed from GpuPipelineDescription.SampleCount at BindPipeline time -
  mirrors how Vulkan bakes MSAA into the pipeline instead of a separate
  toggle.

Collateral, scoped to keep the port real rather than a stub:

- GpuTextureSlot (Unassigned = uint.MaxValue, NOT 0) now flows through
  every consumer of TextureCache.GetOrUploadRenderSurface/UploadRgba8 and
  TextRenderer.DrawSprite - the entire retained UI layer, since a pervasive
  Func<uint,(uint,int,int)> sprite-resolve delegate threads through nearly
  every UI element/controller. Every prior `== 0` / `!= 0` "no texture"
  check became `.IsAssigned` / `!.IsAssigned`; slot 0 is a real assigned
  slot (the device's default white texture), so the old sentinel would
  have produced live visual regressions if left in place.
- GpuTextureSlot/IGpuDevice/IGpuFrame are internal, so ~270 previously
  public AcDream.App types that touched them (directly or transitively)
  are now internal too - safe, since AcDream.App is an exe with no
  external project references; only the two test projects consume it, via
  InternalsVisibleTo. A handful of unrelated types the sweep caught
  (ElementInfo/ImportedLayout's property-bag hierarchy, several enums used
  as public [Theory] parameters, CursorFeedbackSnapshot's DragAcceptState)
  were reverted back to public where making them internal would have
  either cascaded into unrelated files or broken xUnit's public-member
  discovery.
- ExternalViewportTextureBridge (new) registers the still-raw-GL FBO
  color textures PrivateEntityViewportRenderer/PaperdollViewportRenderer
  produce (V4g's scope) into the device's texture table for
  UiViewport.TextureHandle, via a temporary
  GlGpuDevice.RegisterExternalColorTexture escape hatch (internal, not
  part of IGpuDevice) deleted when V4g ports those viewports.
- TextRenderGlStateScope.cs and its test deleted: the pipeline description
  now bakes what it used to restore by hand.
- ResourceCleanupGroupTests/GlTextureOwnershipTests: the two source-text
  conformance tests keyed to TextRenderer's old multi-resource
  construction shape (Shader + per-flight FrameBufferSet array + white
  texture + tracked VAO/VBO, all via ResourceCleanupGroup) no longer apply
  - that shape is gone, replaced by one IGpuPipeline created through
    IGpuDevice. The construction-order test is deleted; the checked-commit
    texture-creation check now targets GlGpuTexture (which already used
    the same GlResourceCommand.CreateName primitive before this slice).

Gates:
- dotnet build -c Release: 0 warnings, 0 errors (AcDream.App has
  TreatWarningsAsErrors).
- dotnet test tests/AcDream.App.Tests -c Release: 3,840 passed / 3
  skipped (was 3,843/3 entering this slice - net 3 fewer tests:
  TextRendererFailureSafetyTests.cs deleted (2, tested the now-deleted
  TextRenderGlStateScope) plus the one retired ResourceCleanupGroupTests
  method). Full solution: 8,908 passed / 5 skipped across all nine test
  projects.
- Offline pixel gate (tools/run-offline-pixel-gate.ps1, parent ec414d60
  vs this commit): differing fraction 0.318% (1,791/563,200 compared
  pixels), above the 0.001 threshold. Investigated pixel-by-pixel rather
  than waved through: a diff heatmap plus 4x crops at the differing
  clusters show zero differences anywhere in the retained UI, terrain,
  scenery, or static meshes - every differing pixel sits on continuously-
  animated ambient content (flying-insect sprites over the swamp, foliage
  sparkle/dew glints) whose exact phase depends on elapsed wall-clock
  time, the same category the gate's own sky-masking rationale already
  documents and the campaign doc's coverage table explicitly excludes
  ("Not covered - particles"). Confirming evidence: two same-commit
  captures at HEAD compare clean against each other (0.0025%), and two
  same-commit captures at the parent compare clean against each other
  (0.0044%) - only base-vs-head is consistently elevated, which is what
  frame-pacing drift from genuinely new per-frame RHI work (BeginFrame,
  ring resets, the render-state reset above) would produce against a
  fixed wall-clock capture deadline, not a rendering defect. Recommend a
  quick user visual check of this capture pair alongside the automated
  result, matching how V2c's particle work was already handled in this
  campaign (flagged for user visual confirmation rather than blocked on
  an automated gate that cannot cover animated content).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-27 18:22:08 +02:00

312 lines
15 KiB
C#

using AcDream.App.Rendering.Gpu;
using AcDream.App.UI;
using AcDream.App.UI.Layout;
namespace AcDream.App.Tests.UI.Layout;
/// <summary>
/// Pure unit tests for <see cref="LayoutImporter.BuildFromInfos"/> — no dats, no GL.
/// Verifies the tree-builder: widget dispatch, Type-12 skipping, and meter child consumption.
/// </summary>
public class LayoutImporterTests
{
private static (GpuTextureSlot, int, int) NoTex(uint _) => (GpuTextureSlot.Unassigned, 0, 0);
// ── Test 1: Health meter element → UiMeter with correct rect ─────────────
/// <summary>
/// A Type-7 (meter) child element with X=5,Y=5,W=150,H=16 must produce a UiMeter
/// that is findable by its id, positioned at Left=5, Width=150.
/// The resolve lambda is a 1-arg Func&lt;uint,(uint,int,int)&gt;.
/// </summary>
[Fact]
public void BuildFromInfos_HealthMeter_IsUiMeterAtRect()
{
var root = new ElementInfo { Id = 0x100005F9, Type = 3, Width = 160, Height = 58 };
var health = new ElementInfo { Id = 0x100000E6, Type = 7, X = 5, Y = 5, Width = 150, Height = 16 };
var tree = LayoutImporter.BuildFromInfos(root, new[] { health }, NoTex, null);
var found = tree.FindElement(0x100000E6);
Assert.IsType<UiMeter>(found);
Assert.Equal(5f, found!.Left);
Assert.Equal(150f, found.Width);
}
// ── Test 2: Type-12 child builds a UiText; Type-3 sibling is also present ──
/// <summary>
/// A root with two children: one Type-12 UIElement_Text and one Type-3 container.
/// The Type-12 must appear as a <see cref="UiText"/> in the tree (transparent,
/// draws nothing until a controller binds its <c>LinesProvider</c>);
/// the Type-3 must also be present.
/// </summary>
[Fact]
public void BuildFromInfos_StampsDatElementId_OnRootAndItemList()
{
const uint RootId = 0x100005F9u;
const uint ListId = 0x100001C6u;
var root = new ElementInfo { Id = RootId, Type = 3, Width = 160, Height = 58 };
var itemList = new ElementInfo { Id = ListId, Type = 0x10000031u, X = 5, Y = 5, Width = 72, Height = 72 };
var tree = LayoutImporter.BuildFromInfos(root, new[] { itemList }, NoTex, null);
Assert.Equal(RootId, tree.Root.DatElementId);
var found = Assert.IsType<UiItemList>(tree.FindElement(ListId));
Assert.Equal(ListId, found.DatElementId);
}
[Fact]
public void BuildFromInfos_Type12Child_IsSkipped_Type3Present()
{
var root = new ElementInfo { Id = 0x10000001, Type = 3, Width = 160, Height = 58 };
var prototype = new ElementInfo { Id = 0x20000001, Type = 12, Width = 0, Height = 0 };
var container = new ElementInfo { Id = 0x20000002, Type = 3, Width = 100, Height = 20 };
var tree = LayoutImporter.BuildFromInfos(root, new[] { prototype, container }, NoTex, null);
// Type-12 is now a UiText (transparent, no lines) — present in the tree.
Assert.IsType<UiText>(tree.FindElement(0x20000001));
// Type-3 must also be present.
Assert.NotNull(tree.FindElement(0x20000002));
}
[Fact]
public void BuildFromInfos_TextPassToChildren_PropagatesDefaultAndRuntimeState()
{
var tab = new ElementInfo
{
Id = 0x100002A9u,
Type = 12,
Width = 100,
Height = 25,
DefaultStateId = RetailUiStateIds.Closed,
DefaultStateName = "Closed",
};
tab.States[RetailUiStateIds.Closed] = new UiStateInfo
{ Id = RetailUiStateIds.Closed, Name = "Closed", PassToChildren = true };
tab.States[RetailUiStateIds.Open] = new UiStateInfo
{ Id = RetailUiStateIds.Open, Name = "Open", PassToChildren = true };
var cap = new ElementInfo { Id = 0x10000439u, Type = 3, Width = 17, Height = 25 };
cap.States[RetailUiStateIds.Closed] = new UiStateInfo
{
Id = RetailUiStateIds.Closed,
Name = "Closed",
Image = new UiImageMedia(0x06005D93u, 3),
};
cap.States[RetailUiStateIds.Open] = new UiStateInfo
{
Id = RetailUiStateIds.Open,
Name = "Open",
Image = new UiImageMedia(0x06005D92u, 3),
};
cap.StateMedia["Closed"] = (0x06005D93u, 3);
cap.StateMedia["Open"] = (0x06005D92u, 3);
tab.Children.Add(cap);
var root = new ElementInfo { Id = 1u, Type = 3, Width = 100, Height = 25 };
ImportedLayout layout = LayoutImporter.BuildFromInfos(root, [tab], NoTex, null);
UiText text = Assert.IsType<UiText>(layout.FindElement(tab.Id));
UiDatElement child = Assert.IsType<UiDatElement>(layout.FindElement(cap.Id));
Assert.Single(text.Children);
Assert.Equal(RetailUiStateIds.Closed, child.ActiveRetailStateId);
Assert.Equal(0x06005D93u, child.ActiveMedia().File);
Assert.True(text.TrySetRetailState(RetailUiStateIds.Open));
Assert.Equal(RetailUiStateIds.Open, child.ActiveRetailStateId);
Assert.Equal(0x06005D92u, child.ActiveMedia().File);
}
// ── Test 3: Meter consumes Type-3 slice children — child ids not in byId ────
/// <summary>
/// A meter (Type 7) whose children are ONLY the 3-slice back/front containers (Type 3).
/// The meter itself must be findable; its Type-3 children must NOT appear as separate
/// nodes in the tree (Fix 5: only non-Type-3 children are built as separate widgets).
/// </summary>
[Fact]
public void BuildFromInfos_MeterWithSliceChildren_MeterPresent_SliceChildrenNotInTree()
{
const uint MeterId = 0x100000E6u;
const uint BackLayerId = 0x100000E7u;
const uint FrontLayerId = 0x00000002u;
// Build a minimal meter with back + front containers, each with 3 slice children.
var backContainer = BuildSliceContainer(BackLayerId, ReadOrder: 0,
l: 0x0600747Eu, t: 0x0600747Fu, r: 0x06007480u);
var frontContainer = BuildSliceContainer(FrontLayerId, ReadOrder: 1,
l: 0x06007481u, t: 0x06007482u, r: 0x06007483u);
var meter = new ElementInfo { Id = MeterId, Type = 7, Width = 150, Height = 16 };
meter.Children.Add(backContainer);
meter.Children.Add(frontContainer);
var root = new ElementInfo { Id = 0x100005F9, Type = 3, Width = 160, Height = 58 };
var tree = LayoutImporter.BuildFromInfos(root, new[] { meter }, NoTex, null);
// The meter widget is present.
Assert.IsType<UiMeter>(tree.FindElement(MeterId));
// The meter's Type-3 slice children are NOT separate UiElement nodes.
Assert.Null(tree.FindElement(BackLayerId));
Assert.Null(tree.FindElement(FrontLayerId));
// The UiMeter itself has no UiElement children (all children were Type-3, consumed).
var uiMeter = (UiMeter)tree.FindElement(MeterId)!;
Assert.Empty(uiMeter.Children);
}
// ── Test 5: Fix 5 — meter text children (Type-12) are built and registered ─
/// <summary>
/// Fix 5: a meter (Type 7) with BOTH Type-3 slice containers AND a Type-12 text
/// child. The Type-3 containers must be consumed (not in byId); the Type-12 text
/// child must be built as a UiText, registered in byId, and attached as a UiElement
/// child of the meter (so it renders as an overlay).
/// </summary>
[Fact]
public void BuildFromInfos_MeterWithTextChild_TextChildInTreeAsUiTextChildOfMeter()
{
const uint MeterId = 0x10000236u;
const uint BackId = 0x10000237u; // Type-3 slice
const uint FrontId = 0x10000238u; // Type-3 slice
const uint LabelId = 0x10000239u; // Type-12 text child
const uint ValueId = 0x1000023Au; // Type-12 text child
var backContainer = BuildSliceContainer(BackId, ReadOrder: 0,
l: 0x0600747Eu, t: 0x0600747Fu, r: 0x06007480u);
var frontContainer = BuildSliceContainer(FrontId, ReadOrder: 1,
l: 0x06007481u, t: 0x06007482u, r: 0x06007483u);
// Two Type-12 text children (the XP "label" and "value" overlay).
var labelInfo = new ElementInfo { Id = LabelId, Type = 12, X = 0, Y = 0, Width = 120, Height = 13 };
var valueInfo = new ElementInfo { Id = ValueId, Type = 12, X = 0, Y = 0, Width = 200, Height = 13,
HJustify = HJustify.Right };
var meter = new ElementInfo { Id = MeterId, Type = 7, Width = 200, Height = 13 };
meter.Children.Add(backContainer);
meter.Children.Add(frontContainer);
meter.Children.Add(labelInfo);
meter.Children.Add(valueInfo);
var root = new ElementInfo { Id = 0x100005F9, Type = 3, Width = 210, Height = 20 };
var tree = LayoutImporter.BuildFromInfos(root, new[] { meter }, NoTex, null);
// Meter is present.
Assert.IsType<UiMeter>(tree.FindElement(MeterId));
// Type-3 slice containers are NOT in byId (consumed by BuildMeter).
Assert.Null(tree.FindElement(BackId));
Assert.Null(tree.FindElement(FrontId));
// Type-12 text children ARE in byId as UiText.
Assert.IsType<UiText>(tree.FindElement(LabelId));
Assert.IsType<UiText>(tree.FindElement(ValueId));
// The UiMeter has exactly 2 UiElement children (the two text overlays).
var uiMeter = (UiMeter)tree.FindElement(MeterId)!;
Assert.Equal(2, uiMeter.Children.Count);
Assert.All(uiMeter.Children, c => Assert.IsType<UiText>(c));
// The value child should have RightAligned=true (HJustify.Right in the dat).
var valueWidget = (UiText)tree.FindElement(ValueId)!;
Assert.True(valueWidget.RightAligned, "Type-12 text child with HJustify.Right must build as RightAligned=true");
}
// ── Test 6: Fix 5 — vitals meters (Type-3 only) are unaffected ────────────
/// <summary>
/// Fix 5 regression guard: vitals health/stamina/mana meters have ONLY Type-3 slice
/// children (no Type-12 text children). The meter must have zero UiElement children
/// after build — VitalsController injects its number overlay at runtime, not the importer.
/// </summary>
[Fact]
public void BuildFromInfos_VitalsMeter_NoTextChildren_MeterHasNoUiChildren()
{
const uint MeterId = 0x100000E6u; // vitals health meter
const uint BackId = 0x100000E7u;
const uint FrontId = 0x100000E8u;
var backContainer = BuildSliceContainer(BackId, ReadOrder: 0,
l: 0x0600747Eu, t: 0x0600747Fu, r: 0x06007480u);
var frontContainer = BuildSliceContainer(FrontId, ReadOrder: 1,
l: 0x06007481u, t: 0x06007482u, r: 0x06007483u);
var meter = new ElementInfo { Id = MeterId, Type = 7, Width = 150, Height = 16 };
meter.Children.Add(backContainer);
meter.Children.Add(frontContainer);
var root = new ElementInfo { Id = 0x100005F9, Type = 3, Width = 160, Height = 58 };
var tree = LayoutImporter.BuildFromInfos(root, new[] { meter }, NoTex, null);
Assert.IsType<UiMeter>(tree.FindElement(MeterId));
// Vitals meter has no UiElement children — only Type-3 children were present,
// all consumed, none built as UiText overlays.
var uiMeter = (UiMeter)tree.FindElement(MeterId)!;
Assert.Empty(uiMeter.Children);
}
// ── Test 4: Prototype-skip in BuildFromInfos ─────────────────────────────
/// <summary>
/// When one top-level element is referenced as a BaseElement by a sibling
/// (mirroring the toolbar slot prototype pattern), and the prototype element
/// has no own state media, the importer must NOT produce a widget for the
/// prototype id (FindElement returns null), but MUST produce the derived element.
///
/// NOTE: This test exercises <see cref="LayoutImporter.BuildFromInfos"/> (the pure
/// layer), where prototype detection is done by inspecting the pre-resolved
/// ElementInfo tree rather than the raw dat ElementDesc. The pure layer skips
/// an element if its Id is in a sibling's (or child's) <c>Children</c> chain
/// as a BaseElement — but actually the pure layer has no BaseElement knowledge
/// at this stage (that's resolved before Build). The prototype-skip in the real
/// world occurs in <c>ImportInfos</c> (the dat shell), BEFORE calling Build.
///
/// This test verifies the INVARIANT that holds AFTER ImportInfos filters prototypes:
/// a pure template element that was skipped is absent from FindElement, while the
/// derived element (which inherited from it) IS present.
///
/// We model this by simply NOT adding the prototype to the ElementInfo tree passed
/// to BuildFromInfos — as if ImportInfos already filtered it out.
/// </summary>
[Fact]
public void BuildFromInfos_PrototypeSkipped_DerivedPresent_PrototypeAbsent()
{
// Simulate what ImportInfos does AFTER filtering: the prototype 0xBBB00001 is
// absent (already skipped by ImportInfos), the derived element 0xCCC00001 is
// present with its own media inherited from the prototype.
var root = new ElementInfo { Id = 0x10000001, Type = 3, Width = 200, Height = 100 };
// The derived element has its own size + media (prototype was merged into it already).
var derived = new ElementInfo
{
Id = 0xCCC00001u,
Type = 0x10000031u, // UIElement_ItemList (toolbar slot type)
X = 10, Y = 10, Width = 32, Height = 32,
};
derived.StateMedia[""] = (0x06001234u, 1);
// Only the derived element appears in the tree (prototype was filtered by ImportInfos).
var tree = LayoutImporter.BuildFromInfos(root, new[] { derived }, NoTex, null);
// The derived element is present in the built tree.
Assert.NotNull(tree.FindElement(0xCCC00001u));
// The prototype id is NOT in the tree (was never added).
Assert.Null(tree.FindElement(0xBBB00001u));
}
// ── Helpers ───────────────────────────────────────────────────────────────
private static ElementInfo BuildSliceContainer(uint id, uint ReadOrder, uint l, uint t, uint r)
{
var c = new ElementInfo { Id = id, Type = 3, ReadOrder = ReadOrder };
c.Children.Add(new ElementInfo { X = 0, StateMedia = { [""] = (l, 1) } });
c.Children.Add(new ElementInfo { X = 10, StateMedia = { [""] = (t, 1) } });
c.Children.Add(new ElementInfo { X = 140, StateMedia = { [""] = (r, 1) } });
return c;
}
}