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

fix #426: extract solid-colour (NO_POS_UVS) faces; skip untextured subsets only on building shells and cells like retail

Browse source 
The Holtburg windmill axle (GfxObj 0x010010CE, 8 polygons, all
Stippling.NoPos + SurfaceType.Base1Solid) extracted to a 0-vertex mesh.
NoPos ("NO_POS_UVS", acclient.h:7380-7388) means "this side has no
texture coordinates" — true of every solid-colour polygon, since
nothing samples them — not "there is no positive face". Extraction read
it as the latter and dropped the polygon entirely, client-wide, for
every untextured polygon on every object.

Retail's D3DPolyRender::DrawMesh (@0x0059d4a0, named-retail decomp
~line 426048) draws an untextured subset on an ordinary object exactly
like a textured one; the only retail cases that skip an untextured
subset are a building shell (RenderDeviceD3D::DrawBuilding @0x0059f2a0
sets ObjBuildingOrBuildingPart=1) or an EnvCell interior
(RenderDeviceD3D::DrawEnvCell @0x0059f170, arg4=1). The #119
investigation's "retail's skipNoTexture never draws them either"
conclusion was itself wrong as a general rule.

- MeshExtractor.PrepareGfxObjMeshData / GfxObjMesh.Build: emit the
  positive side whenever PosSurface is a valid index, regardless of
  NoPos; the existing UV-index-0 fallback already produces zero
  texcoords for a NoPos polygon with no UVs on the wire.
- RetailUntexturedSurfacePolicy.IsUntextured(SurfaceType): the one
  place that answers "is this surface textured"
  ((type & (Base1Image|Base1ClipMap)) == 0), replacing the old
  `isSolid = NoPos || Base1Solid` (which also mis-classified a NEG-side
  batch by the POS-side's NoPos flag).
- RetailUntexturedSubsetPolicy.Draws(isBuildingShell, isUntextured):
  the shared draw-time gate wired into WbDrawDispatcher.ClassifyBatches,
  .PackedOracle.ClassifyPackedBatches, and
  .DirectionalShadows.AddDirectionalShadowBatches — one predicate so the
  three walks cannot drift (Campaign VM VM6 lesson).
- CellMesh.cs / MeshExtractor.PrepareCellStructMeshData deliberately
  KEEP their NoPos-gated skip for cell-wall geometry — retail's
  DrawEnvCell really does skip untextured subsets there; register row
  AP-234 documents the NoPos-vs-Surface.Type approximation.
- PakFormat.CurrentBakeToolVersion 4->5 (LauncherInstallRecordStore in
  lockstep): a pak baked by an older tool is missing every untextured
  face. No bake was run as part of this commit.

Also fixed: WorldBuilder's own upstream ObjectMeshManager.cs has the
identical NoPos bug (ObjectMeshManager.cs:959,984) — our port had
faithfully carried it over, and our own conformance test
(Build_NoPosFlag_OnlyEmitsNegSide) asserted the bug as correct WB
conformance. Renamed/reworded to Build_NoPosFlag_EmitsBothPosAndNegSide
with a citation for why retail decomp overrides WB here.

Issue119UpNullGfxObjDumpTests re-run against the installed DAT:
#119's own two objects (0x010002B4 9/9 polys, 0x010008A8 1/1 poly) now
gate DRAWS on every polygon instead of extracting to nothing.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
This commit is contained in:
Erik 2026-08-23 11:20:24 +02:00
parent 51a5fe99ef
commit 517d17b4b3
18 changed files with 755 additions and 40 deletions
Download patch file Download diff file Expand all files Collapse all files

42
tests/AcDream.Core.Tests/Conformance/Issue119UpNullGfxObjDumpTests.cs
View file

@ -1,6 +1,7 @@
using System;
using System.Collections.Generic;
using System.Linq;
using AcDream.Core.Meshing;
using DatReaderWriter;
using DatReaderWriter.DBObjs;
using DatReaderWriter.Enums;
@ -22,6 +23,22 @@ namespace AcDream.Core.Tests.Conformance;
/// and replicates MeshExtractor.PrepareGfxObjMeshData's gates (moved from
/// ObjectMeshManager in MP1a)
/// so the zeroing gate reads directly off the output.
///
/// #426 (2026-08-23) UPDATE: the gate replica below now mirrors the FIXED
/// extraction rule — the positive side is added whenever PosSurface is a
/// valid index, regardless of Stippling.NoPos (NoPos means "no positive
/// UVs", not "no positive face"; see RetailUntexturedSurfacePolicy). Run
/// against the installed DAT, both of #119's original objects now gate
/// DRAWS on every single polygon (0x010002B4: 9/9 polys; 0x010008A8: 1/1
/// poly — both all-NoPos+Base1Solid, wouldAddPos == polygon count) — they
/// are NEITHER all-degenerate NOR all-invalid-surface-index; they were
/// dropped for being all-solid, and #426 now extracts that geometry. The
/// #119 filing's "retail never draws untextured subsets" conclusion was
/// simply wrong (see #426's ISSUES.md entry) — it never held for these two
/// objects specifically. Whether their solid faces end up VISIBLE on
/// screen (vs. skipped again at draw time by RetailUntexturedSubsetPolicy,
/// if either object turns out to be a building-shell part) is a separate,
/// unverified question this dump does not answer.
/// </summary>
[Trait("Lane", "InstalledDat")]
public sealed class Issue119UpNullGfxObjDumpTests
@ -57,7 +74,10 @@ public sealed class Issue119UpNullGfxObjDumpTests
}
// Replicate the extraction gates (PrepareGfxObjMeshData):
// pos added when !NoPos
// pos added whenever PosSurface is a valid index — #426
// (2026-08-23): NoPos means "no positive UVs", not "no positive
// face"; ordinary objects draw untextured (solid) subsets the same
// as textured ones (RetailUntexturedSurfacePolicy).
// neg added when Negative || Both || (!NoNeg && SidesType==Clockwise)
// surface index must be in [0, Surfaces.Count)
int wouldAddPos = 0, wouldAddNeg = 0, degenerate = 0;
@ -68,8 +88,7 @@ public sealed class Issue119UpNullGfxObjDumpTests
if (poly.VertexIds.Count < 3) { degenerate++; gate = "degenerate(<3 verts)"; }
else
{
bool pos = !poly.Stippling.HasFlag(StipplingType.NoPos)
&& poly.PosSurface >= 0 && poly.PosSurface < gfx.Surfaces.Count;
bool pos = poly.PosSurface >= 0 && poly.PosSurface < gfx.Surfaces.Count;
bool neg = (poly.Stippling.HasFlag(StipplingType.Negative)
|| poly.Stippling.HasFlag(StipplingType.Both)
|| (!poly.Stippling.HasFlag(StipplingType.NoNeg) && poly.SidesType == CullMode.Clockwise))
@ -96,6 +115,18 @@ public sealed class Issue119UpNullGfxObjDumpTests
/// are "regular shell polys" — render digest user axiom). A non-zero
/// "DROPPED but textured" count names the extraction as the stairs-miss
/// mechanism; zero exonerates the per-poly gates.
///
/// #426 (2026-08-23) UPDATE: post-fix, the POS side can only be dropped
/// when PosSurface itself is out of range — a textured pos-side surface
/// with a valid index is now ALWAYS emitted (that's the whole point of
/// #426). So this test's remaining bite is almost entirely the NEG-side
/// gate (unchanged by #426 — a textured NEG surface can still legitimately
/// be dropped when none of the three double-sided conditions hold). Kept
/// under its original name/assertion (zero textured drops) because that
/// invariant is still exactly what we want to hold; `SurfaceIsTextured`
/// now calls the shared RetailUntexturedSurfacePolicy predicate instead of
/// a bare Base1Solid check, so this test and the extraction it's checking
/// literally cannot drift on what "textured" means.
/// </summary>
[Theory]
[InlineData(0x010014C3u)]
@ -112,7 +143,7 @@ public sealed class Issue119UpNullGfxObjDumpTests
{
if (idx < 0 || idx >= gfx!.Surfaces.Count) return false;
if (!dats.Portal.TryGet<Surface>(gfx.Surfaces[idx], out var surf) || surf is null) return false;
return !surf.Type.HasFlag(SurfaceType.Base1Solid);
return !RetailUntexturedSurfacePolicy.IsUntextured(surf.Type);
}
int draws = 0;
@ -120,8 +151,7 @@ public sealed class Issue119UpNullGfxObjDumpTests
foreach (var (pid, poly) in gfx!.Polygons.OrderBy(kv => kv.Key))
{
if (poly.VertexIds.Count < 3) continue;
bool pos = !poly.Stippling.HasFlag(StipplingType.NoPos)
&& poly.PosSurface >= 0 && poly.PosSurface < gfx.Surfaces.Count;
bool pos = poly.PosSurface >= 0 && poly.PosSurface < gfx.Surfaces.Count;
bool neg = (poly.Stippling.HasFlag(StipplingType.Negative)
|| poly.Stippling.HasFlag(StipplingType.Both)
|| (!poly.Stippling.HasFlag(StipplingType.NoNeg) && poly.SidesType == CullMode.Clockwise))