erik/acdream
1
0
Fork 0
Code Issues Pull requests Projects Releases 6 Packages Wiki Activity Actions

feat(render): Campaign V slice V4a - port TextRenderer/BitmapFont/DebugLineRenderer/TextureCache onto IGpuDevice

Browse source 
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>
This commit is contained in:
Erik 2026-07-27 18:22:08 +02:00
parent ec414d60cd
commit ceec3bc440
334 changed files with 3660 additions and 3840 deletions
Download patch file Download diff file Expand all files Collapse all files

44
src/AcDream.App/Rendering/PortalProjection.cs
View file

@ -1,11 +1,11 @@
// PortalProjection.cs
// PortalProjection.cs
//
// Phase A8.F: project a cell-local portal polygon to NDC screen space. Homogeneous frustum clip
// in CLIP SPACE (before the perspective divide): first the IN-FRONT-OF-EYE half-space (keep where
// w > MinW) so a portal straddling the camera does not invert under the divide and the divide
// stays bounded away from the w=0 eye singularity, then the 4 SIDE planes (x,y within ±w) so every
// stays bounded away from the w=0 eye singularity, then the 4 SIDE planes (x,y within ±w) so every
// surviving vertex lands on the screen [-1,1] by construction. The side-plane clip is the R1
// void-flap fix (2026-06-05) — see ProjectToNdc.
// void-flap fix (2026-06-05) — see ProjectToNdc.
//
// The clip is NEAR-INDEPENDENT on purpose. We only use the projected x/y for the visibility clip
// REGION, so a vertex in front of the eye is meaningful even if it is closer than the projection's
@ -23,7 +23,7 @@ using System.Numerics;
namespace AcDream.App.Rendering;
public static class PortalProjection
internal static class PortalProjection
{
internal ref struct ClipPolygonLease
{
@ -118,17 +118,17 @@ public static class PortalProjection
Vector4[]? second = null;
// Homogeneous frustum clip in CLIP SPACE, before the perspective divide. First the
// in-front-of-eye half-space (w > MinW) — near-INDEPENDENT, so a portal the camera is
// standing in still projects (see header); then the 4 SIDE planes (x,y within ±w). The
// in-front-of-eye half-space (w > MinW) — near-INDEPENDENT, so a portal the camera is
// standing in still projects (see header); then the 4 SIDE planes (x,y within ±w). The
// side clip is the R1 void-flap fix (2026-06-05): without it, a portal WITHIN the near
// plane projected small-w verts to wildly off-screen NDC (the probe saw (10.2,-67.4)),
// which corrupted the downstream 2D ScreenPolygonClip into an EMPTY region -> OutsideView
// empty -> terrain Skip -> the bluish doorway "void". Clipping the side planes here bounds
// every surviving vertex to the screen [-1,1] by construction, so a screen-covering doorway
// clips to the screen (non-empty) instead of collapsing. The eye plane is clipped FIRST so
// all survivors have w > 0, making the side-plane functionals (w ± x, w ± y) well defined.
// all survivors have w > 0, making the side-plane functionals (w ± x, w ± y) well defined.
// Near/far are intentionally NOT clipped (near-independence). Retail PView::GetClip
// (decomp:0x005a4320) projects + frustum-clips the portal poly likewise (research doc A §3.5).
// (decomp:0x005a4320) projects + frustum-clips the portal poly likewise (research doc A §3.5).
try
{
second = vectorPool.Rent(capacity);
@ -155,7 +155,7 @@ public static class PortalProjection
(current, output) = (output, current);
}
// Perspective divide → NDC xy. This is the only result allocation.
// Perspective divide → NDC xy. This is the only result allocation.
var ndc = new Vector2[currentCount];
for (int i = 0; i < currentCount; i++)
{
@ -174,21 +174,21 @@ public static class PortalProjection
/// <summary>Faithful homogeneous projection (retail PrimD3DRender::xformStart + the W=0 clip of
/// ACRender::polyClipFinish, decomp 424310 / 702749): transform the portal to clip space and clip
/// ONLY the eye plane (w &gt;= 0, EXACT), keeping homogeneous coords — NO perspective divide, NO
/// ONLY the eye plane (w &gt;= 0, EXACT), keeping homogeneous coords — NO perspective divide, NO
/// frustum side-plane clamp. The screen bound is applied later by <see cref="ClipToRegion"/>
/// against the view region (the root region is the full screen), exactly as retail clips the portal
/// against the accumulated portal_view rather than fixed side planes.
///
/// <para>The W=0 clip is exact on purpose (the knife-edge port, 2026-06-11; pseudocode at
/// docs/research/2026-06-11-polyclipfinish-w0-clip-pseudocode.md): boundary intersections land
/// at w == 0 — homogeneous DIRECTIONS — so a portal the eye is crossing (stair openings, decks)
/// at w == 0 — homogeneous DIRECTIONS — so a portal the eye is crossing (stair openings, decks)
/// yields the correct UNBOUNDED half-region, which the bounded view-region clip then cuts to the
/// screen. The previous EyePlaneW = 1e-4 produced finite ~1e4-NDC boundary verts whose region
/// intersections sat at the dedup/merge degeneracy threshold — the climb-strobe class. A w=0
/// intersections sat at the dedup/merge degeneracy threshold — the climb-strobe class. A w=0
/// vertex can never survive ClipToRegion into its divide (a nonzero direction fails at least one
/// edge test of any BOUNDED convex region), so no divide-by-zero path exists; the measure-zero
/// corner case is guarded in ClipToRegion. Matches polyClipFinish part 1: clip pass runs only
/// when some vertex has w &lt; 0; &lt;3 survivors → reject (empty).</para></summary>
/// when some vertex has w &lt; 0; &lt;3 survivors → reject (empty).</para></summary>
public static Vector4[] ProjectToClip(IReadOnlyList<Vector3> localPoly, Matrix4x4 cellToWorld, Matrix4x4 viewProj)
{
using ClipPolygonLease lease = ProjectToClipLease(localPoly, cellToWorld, viewProj);
@ -274,7 +274,7 @@ public static class PortalProjection
/// (CCW convex) with w-aware Sutherland-Hodgman edge tests, then divide the survivors to NDC and
/// normalize to CCW. Ports retail ACRender::polyClipFinish's view-region clip (decomp 702749): the
/// edge test multiplies through w (which is &gt; 0 after the eye-plane clip) so it never divides a
/// near-eye vertex, and the final divide runs only on survivors already bounded to the region —
/// near-eye vertex, and the final divide runs only on survivors already bounded to the region —
/// stable by construction. Returns &lt;3 verts when the portal does not intersect the region.</summary>
public static Vector2[] ClipToRegion(IReadOnlyList<Vector4> subjectClip, IReadOnlyList<Vector2> regionCcwNdc)
{
@ -339,7 +339,7 @@ public static class PortalProjection
// Homogeneous Sutherland-Hodgman: clip the (w > 0) subject against each CCW edge of the NDC
// region. f(P) below is the NDC inside test cross(edge, P_ndc - a) multiplied through P.W,
// which is > 0 after the eye-plane clip — so the sign is the NDC sign yet no near-eye vertex
// which is > 0 after the eye-plane clip — so the sign is the NDC sign yet no near-eye vertex
// is ever divided (retail polyClipFinish, decomp 702749).
int regionCount = regionCcwNdc.Count;
int capacity = checked(subjectClip.Length + regionCount);
@ -370,7 +370,7 @@ public static class PortalProjection
if (currentCount < 3)
return System.Array.Empty<Vector2>();
// Divide survivors → NDC. They are already inside the bounded region. A w=0
// Divide survivors → NDC. They are already inside the bounded region. A w=0
// measure-zero corner remains the same empty knife-edge result as the prior path.
ndcScratch = vector2Pool.Rent(currentCount);
Span<Vector2> ndc = ndcScratch.AsSpan(0, currentCount);
@ -408,8 +408,8 @@ public static class PortalProjection
// Retail copy_view's ~1-pixel vertex merge (see ClipToRegion). Collapses
// runs of consecutive near-identical vertices, including across the
// wrap-around. A polygon that collapses below 3 distinct vertices is
// degenerate (sub-pixel sliver) and returns empty — exactly retail's
// "<3 surviving verts → output count 0".
// degenerate (sub-pixel sliver) and returns empty — exactly retail's
// "<3 surviving verts → output count 0".
private const float VertexMergeEpsilonNdc = 2f / 1080f;
private static int MergeSubPixelVertices(Span<Vector2> poly)
@ -428,7 +428,7 @@ public static class PortalProjection
}
poly[kept++] = vertex;
}
// Wrap-around: last ≈ first.
// Wrap-around: last ≈ first.
while (kept >= 2)
{
Vector2 first = poly[0];
@ -442,9 +442,9 @@ public static class PortalProjection
return kept;
}
// One Sutherland-Hodgman half-plane against the directed NDC edge a→b, keeping the CCW-inside
// One Sutherland-Hodgman half-plane against the directed NDC edge a→b, keeping the CCW-inside
// (left) part of a HOMOGENEOUS polygon. Inside test for vertex P (clip space): the NDC cross
// product cross(b-a, P/P.W - a) scaled by P.W (> 0): ex·(P.Y - P.W·a.Y) - ey·(P.X - P.W·a.X) ≥ 0.
// product cross(b-a, P/P.W - a) scaled by P.W (> 0): ex·(P.Y - P.W·a.Y) - ey·(P.X - P.W·a.X) ≥ 0.
// Crossings interpolate in homogeneous coords (perspective-correct), via the shared Lerp.
private static int ClipHomogeneousEdge(
ReadOnlySpan<Vector4> polygon,
@ -489,7 +489,7 @@ public static class PortalProjection
if (area2 < 0f) System.Array.Reverse(poly);
}
// Minimum clip-space w (≈ metres in front of the eye) to keep a vertex. Excludes the eye
// Minimum clip-space w (≈ metres in front of the eye) to keep a vertex. Excludes the eye
// (w=0) singularity and the ~5 cm right at it (bounding the perspective divide), but is
// INTENTIONALLY far closer than the projection's 1.0 m near plane so a doorway the camera is
// standing in still projects and the cell behind it stays visible. See the file header.