fix(ui): Campaign LA gate round 2 — character-select scales as one authored canvas

Third iteration on the screen, completing AD-98. The previous substitution
stretched only the root BACKGROUND while the child widgets stayed at their
authored 800x600 pixel positions - and the background painting carries
visual anchors (the World/Characters captions are art), so the user gate
showed captions overlapping the listbox and every widget misaligned
against the stretched art.

Retail model (established at 71bf24fb): fixed-canvas pre-world screens
render at authored 800x600 and the whole composed frame stretches once at
presentation; the blitter has no stretch mode. Our equivalent now does the
same one stage earlier:

- UiRoot.FixedCanvasSize: while the char-select screen is active, the
  retained tree lays out in its authored canvas and Draw scopes a uniform
  scale onto TextRenderer.CanvasScale; the mouse entry points apply the
  exact inverse so MouseX/MouseY and every hit test live in canvas space.
- TextRenderer.AppendQuad is the single emission chokepoint - sprites,
  rects, AND glyphs scale together, including retail-authentic non-uniform
  aspect distortion and stretched text. World-space HUD stays native (the
  scale resets outside UiRoot.Draw).
- CharacterManagementUiController stops resizing Root to the viewport;
  activate/deactivate/dispose set and clear the host canvas.
- UiDatElement returns to retail-pure copy-or-tile; the interim
  StretchOwnBackgroundToFill flag is deleted.
- AD-98 updated to describe the completed substitution.

Tests: canvas-scale quad math, inverse input mapping (window click lands
on the canvas-space widget), degenerate-size guards, controller keeps
authored extent + sets/clears the canvas. App suite 5085/6 skips; live-DAT
char-select probes 3/3.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

docs/architecture/retail-divergence-register.md

@@ -189,7 +189,7 @@ readiness/requeue adaptation. See
 | AD-92 | **Filed 2026-08-13 at the #376/#388 review fix round (blast M6 / mechanism M4).** Two switcher adaptations with no retail counterpart: (1) the fullscreen refresh rate is the monitor's HIGHEST for the picked WxH — retail passed the device mode's own refresh as-is (`Device::ForceDisplayResolution`); (2) an invalid/unsupported fullscreen request is a logged refusal that leaves the window unchanged — retail attempted the switch and surfaced the device error. The persisted-flag divergence a refusal leaves behind is ISSUES #392. | `src/AcDream.App/Settings/DisplayModeSwitching.cs` (`TryFindRefreshRate`, the refusal paths); `src/AcDream.App/Settings/RuntimeSettingsTargets.cs` (`Apply`'s refused-mode logging) | Highest-refresh is strictly better on modern variable-refresh panels (retail predates them); refuse-and-log is #388's own no-crash requirement. | A capture comparing retail's exact chosen refresh for a mode will differ; a server/tooling flow expecting an error dialog on an invalid mode sees a console line instead. | `Device::ForceDisplayResolution @gmClient::Init 0x004047af`; docs/research/2026-08-13-376-388-{mechanism,blast}-review.md |
 | AD-94 | **Filed 2026-08-14 at the secure-trade feature.** Retail's `Event_AcceptTrade` payload (`Trade::Pack @0x005B9FF0`) appends two `PackableList<ContentProfile>` staged-item lists after the six fixed fields; acdream sends both as ZERO-COUNT lists. ACE parses and then discards the ENTIRE payload (`HandleActionAcceptTrade()` takes zero arguments — server trade state is fully self-derived; lane B §quirks), so the difference is unobservable against ACE; a byte-capture comparison against a real retail client would differ from offset 40. | `src/AcDream.Core.Net/Messages/TradeRequests.cs` (`BuildAcceptTrade`) | The `ContentProfile` pack layout was not byte-verified (ACE never reads it — no reader to check against), and guessing a wire struct violates the workflow; zero-count lists are well-formed `PackableList`s. | A future server that actually validates the accept echo would see empty item lists and could refuse or desync the accept. | `Trade::Pack @0x005B9FF0`; `GameActionAcceptTrade.cs:11-16`; `docs/research/2026-08-14-trade-laneB-wire.md` Table 1 |
 | AD-96 | **Filed 2026-08-14 at the OP8 re-gate fix round (key-name display).** Retail's `GetNameFromKey_Internal @0x00687800` falls back from the DAT string tables (key enum 4 → `0x2300000A`, meta enum 5 → `0x2300000B`) to the OS keyboard layout's own key name via DirectInput `IDirectInputDevice8::GetObjectInfo` (`tszName` — "SKIFT" on a Swedish layout). acdream reads the SAME layout-resident name data through Win32 `GetKeyNameTextW` instead (no DirectInput device exists in-process); on non-Windows hosts there is no OS lookup at all and the DIK-suffix spelling shows (un-localized English, e.g. "LSHIFT"). Mouse chords keep the pre-existing enum spelling — retail names them through the DirectInput mouse device. | `src/AcDream.App/Platform/PlatformKeyNameProvider.cs`; `src/AcDream.App/UI/Layout/RetailKeyNames.cs` (`Describe`, the mouse-device early-out) | GetKeyNameText and DirectInput's key names both come from the active keyboard-layout tables; adding a DirectInput device solely for name strings would be a heavyweight, dead-end dependency. Linux graphical work is parked at Slice L1. | A key whose GetKeyNameTextW name differs from DirectInput's `tszName` on some layout shows a slightly different caption than retail did; Linux graphical shows English DIK-suffix names where retail-on-Wine would localize; a mouse-chord caption reads as the Silk enum, not retail's device string. | `CInputManager_WIN32::GetNameFromKey_Internal @0x00687800`; `GetNameFromKey @0x00687F40`; `ControlSpecification::GetDIKName @0x0068ACB0`; `DBCache::GetDIDFromEnumStatic` category-4 probe 2026-08-14 (`KeyboardConfigLiveMountProbeTests.ProbeKeyboardFontsAndKeyNameStrings`) |
-| AD-98 | **Filed 2026-08-15 at Campaign LA gate round 2 (character-select background tiling).** The LA8 root (0x1000039A) authors LeftEdge=TopEdge=RightEdge=BottomEdge=0 ("no anchor") in the installed DAT, so retail's own `UIElement::UpdateForParentSizeChange` (0x00462640) never resizes this element — it stays a fixed 800x600 rect in retail's own widget tree. Retail's generic sprite blit, `Graphic::Draw` (0x00693b20) dispatching to `Graphic::PutImage` (0x00693a30) for an exact/undersized destination or a modulo-wrapped tile loop otherwise, has no third "stretch" mode (confirmed against `BlitMode`, acclient.h ~line 3135, and `MD_Data_Image::m_drawMode`/`DrawModeType` — both are COLOR-blend selectors, not tile-vs-stretch geometry modes). The only way retail's whole pre-world scene (background AND buttons AND listbox together) can still fill an arbitrary window resolution with no element ever resizing and a blitter that can only copy-or-tile is that these "flow" screens render into a fixed 800x600 target and the WHOLE FRAME is stretched once at presentation, outside the UI element/sprite system. acdream has no offscreen fixed-resolution UI render target / present-time scale pass; `CharacterManagementUiController`'s constructor instead resizes the MOUNTED ROOT element itself to the live viewport, and `UiDatElement.StretchOwnBackgroundToFill` makes that resized root's own background draw as one UV-0..1 quad instead of tiling. | `src/AcDream.App/UI/Layout/UiDatElement.cs` (`StretchOwnBackgroundToFill`, `OnDraw`); `src/AcDream.App/UI/Layout/CharacterManagementUiController.cs` (constructor) | Reproducing retail's real mechanism (a fixed 800x600 UI render target scaled at presentation) would touch the render/swapchain pipeline (`GameWindow`, framebuffer setup) far beyond a background-draw fix; resizing the mounted root and stretching only ITS OWN background quad reaches the identical visual result (no tiling, non-uniform fill matching every resolution) confirmed against the installed DAT's zero edge-anchors and the decompiled blitter's copy-or-tile-only behavior. | If acdream ever gains a genuine fixed-resolution UI render target + present-time scale pass, this flag becomes redundant (every root would already present pre-stretched) and should be deleted along with the per-root resize in `CharacterManagementUiController`. Until then, any OTHER screen-level root mounted the same way (a future login/disconnected/datapatch screen) needs the same flag set explicitly — it is not automatic for arbitrary `UiDatElement`s. | `Graphic::Draw` 0x00693b20; `Graphic::PutImage` 0x00693a30; `UIElement::UpdateForParentSizeChange` 0x00462640; `BlitMode` acclient.h ~3135; `UIElementManager::CreateRootElement` 0x0045d020 (`UIElement::SetIsRootElement`); `CharacterManagementLiveDatTests.RootAuthorsNoEdgeAnchors_RetailNeverResizesItSelf` |
+| AD-98 | **Filed 2026-08-15 at Campaign LA gate round 2 (character-select background tiling).** The LA8 root (0x1000039A) authors LeftEdge=TopEdge=RightEdge=BottomEdge=0 ("no anchor") in the installed DAT, so retail's own `UIElement::UpdateForParentSizeChange` (0x00462640) never resizes this element — it stays a fixed 800x600 rect in retail's own widget tree. Retail's generic sprite blit, `Graphic::Draw` (0x00693b20) dispatching to `Graphic::PutImage` (0x00693a30) for an exact/undersized destination or a modulo-wrapped tile loop otherwise, has no third "stretch" mode (confirmed against `BlitMode`, acclient.h ~line 3135, and `MD_Data_Image::m_drawMode`/`DrawModeType` — both are COLOR-blend selectors, not tile-vs-stretch geometry modes). The only way retail's whole pre-world scene (background AND buttons AND listbox together) can still fill an arbitrary window resolution with no element ever resizing and a blitter that can only copy-or-tile is that these "flow" screens render into a fixed 800x600 target and the WHOLE FRAME is stretched once at presentation, outside the UI element/sprite system. **COMPLETED 2026-08-15 (same gate round, misalignment follow-up):** the first substitution (resize the mounted root + stretch only its own background) stretched the ART but left the authored child widgets at 800x600 pixel positions — misaligned against a background whose painting CARRIES visual anchors (the World/Characters captions are art). The substitution now reproduces retail's whole-frame behavior: the root KEEPS its authored 800x600 extent, and while the screen is active `UiRoot.FixedCanvasSize` scales EVERY emitted quad (widgets, glyphs, art, dialogs) uniformly at `TextRenderer.AppendQuad`, with the exact inverse applied to mouse coordinates at the `UiRoot` entry points so hit-testing lives in canvas space. Non-uniform window/canvas stretch, retail-authentic (no letterbox). `UiDatElement` keeps retail's pure copy-or-tile blit; the interim `StretchOwnBackgroundToFill` flag is deleted. | `src/AcDream.App/UI/UiRoot.cs` (`FixedCanvasSize`, `CanvasScale`, `MapWindowToCanvas`, `Draw`); `src/AcDream.App/Rendering/TextRenderer.cs` (`CanvasScale`, `AppendQuad`); `src/AcDream.App/UI/Layout/CharacterManagementUiController.cs` (activate/deactivate/dispose set+clear the canvas) | Reproducing retail's literal mechanism (an offscreen fixed-resolution UI render target scaled at presentation) would add RHI surface area for an identical pixel result; scaling at the one quad-emission chokepoint with an inverse input mapping is the same math applied one stage earlier, and the world-space HUD stays native because the scale is scoped to `UiRoot.Draw`. | Glyphs stretch with the frame (retail-authentic blur at large windows). Any future fixed-canvas screen (login/disconnected/datapatch) sets `UiRoot.FixedCanvasSize` while active — per-screen opt-in, not automatic. If a genuine present-time frame-stretch pass ever lands, this collapses into it. | `Graphic::Draw` 0x00693b20; `Graphic::PutImage` 0x00693a30; `UIElement::UpdateForParentSizeChange` 0x00462640; `BlitMode` acclient.h ~3135; `UIElementManager::CreateRootElement` 0x0045d020; `CharacterManagementLiveDatTests.RootAuthorsNoEdgeAnchors_RetailNeverResizesItSelf`; `UiRootFixedCanvasTests`; `UiDatElementTests.CanvasScale_StretchesQuadGeometry_LeavesUvsAuthored` |
 | AD-97 | **Filed 2026-08-14 at Campaign LA slice LA7a (character-restore request tail).** Retail's `CharacterRestore` request (`0xF7D9`) is ≥16 bytes: `CPlayerSystem::RestoreCharacter @0x0055d760` is, in the PDB-paired binary, `push 0x008173B4; push 0x008173B4; push guid; call Proto_UI::SendAdminRestoreCharacter @0x00546cf0`, and the callee packs BOTH constant `PStringBase<char>*` arguments (`PStringBase::Pack @0x004fc6f0` emits ≥4 bytes even empty). Binary Ninja renders the two pushes as an uninitialized `edx` local plus `this` — a rendering artifact around constant `0x008173B4` (all 3 of its other pseudo-C appearances sit in provably-broken decompiles), but the arguments are real. acdream sends the 8-byte guid-only form. What the two constant strings contain is unresolved (a live cdb `db poi(0x008173b4)` would settle it). | `src/AcDream.Core.Net/Messages/CharacterRestore.cs` (`BuildRequestBody`) | ACE reads only `ReadUInt32()` and ignores any tail (`CharacterHandler.cs:331-385`), and holtburger ships guid-only from a real client command path against ACE successfully — the tail is unread by every server we can test against, and packing two strings whose CONTENT we cannot verify would be a guess. | A byte-capture comparison against a real retail client differs from offset 8; a future server that validates the full retail shape would reject our 8-byte request. | `CPlayerSystem::RestoreCharacter @0x0055d760` (binary bytes, not the BN rendering); `Proto_UI::SendAdminRestoreCharacter @0x00546cf0`; `PStringBase::Pack @0x004fc6f0`; ACE `CharacterHandler.cs:331-385`; holtburger `character_selection.rs:79-82`; LA7a Opus review F1 (2026-08-14) |
 | AD-93 | **Filed 2026-08-13 at social gate round 2, item 5 (the refused-drop notice port).** Two narrow gaps in the `ServerSaysAttemptFailed @0x0058EAE0` port: (1) **latched-guid preference** — retail's 0x00A0 dispatcher (`@0x0055B342`) PREFERS `prevRequestObjectID` over the wire guid when picking the item to name; acdream's `InventoryTransactionState.OnMoveFailed` instead REQUIRES the wire guid to match the latch (unobservable against ACE, which always sends the request's own guid on 0x00A0, and it protects a stale latch from mislabeling an unrelated failure — acdream has no retail-style latch timeout). (2) **unlatched request kinds** — retail latches `IR_MOVE`/`IR_WIELD` too; acdream's kind enum has no Move/Wield rows because wields ride `AutoWieldController` outside the single-request gate, so a refused wield/3D-move shows only the generic `HandleFailureEvent` leg, never "The X can't be wielded/moved". | `src/AcDream.Core/Items/InventoryTransactionState.cs` (`OnMoveFailed`); `src/AcDream.Core/Chat/InventoryFailureMessages.cs` (`Compose`'s absent Move/Wield rows); `src/AcDream.App/UI/ItemInteractionController.cs` (`OnInventoryRequestFailed`) | The match requirement is the compensating guard for the missing latch timeout; adding Wield/Move kinds means routing those sends through the single-request gate they deliberately bypass today — a behavior change beyond this gate item. | Only observable against a server that sends 0x00A0 with a guid that differs from the request's item (ACE never does), or on a refused wield/move, which shows no "can't be wielded/moved" verb line where retail would show one. | `ACCWeenieObject::ServerSaysAttemptFailed @0x0058EAE0`; the 0x00A0 dispatcher `@0x0055B342`; `ACCWeenieObject::RecordRequest @0x0058C220`; `docs/research/2026-08-13-confirm-and-weenie-error-display.md` §2 |
 

src/AcDream.App/Rendering/TextRenderer.cs

@@ -201,6 +201,23 @@ public sealed class TextRenderer : IDisposable
         });
     }
 
+    /// <summary>
+    /// Campaign LA gate round 2 (register AD-98): uniform canvas scale applied
+    /// to every emitted quad — sprites, rects, AND glyphs — at the single
+    /// emission chokepoint (<see cref="AppendQuad"/>). Retail renders its
+    /// fixed-canvas pre-world screens (char select's authored 800×600, root
+    /// 0x1000039A, zero edge anchors) at authored size and stretches the whole
+    /// composed frame once at presentation; its UI blitter has no stretch mode
+    /// at all (Graphic::Draw @0x00693b20 is copy-or-tile only). We have no
+    /// present-time frame stretch, so the equivalent lives here: while a
+    /// fixed-canvas screen is active, <see cref="UiRoot"/> sets this for the
+    /// duration of its Draw and everything scales together — including retail's
+    /// characteristic non-uniform aspect distortion and stretched glyphs.
+    /// UVs and colors are untouched. Always reset to One outside UiRoot.Draw
+    /// so the world-space HUD keeps native pixels.
+    /// </summary>
+    internal Vector2 CanvasScale = Vector2.One;
+
     /// <summary>Begin a HUD pass. Call once per frame before any Draw* calls.</summary>
     public void Begin(Vector2 screenSize)
     {
@@ -388,10 +405,19 @@ public sealed class TextRenderer : IDisposable
         return ns;
     }
 
-    private static void AppendQuad(List<float> buf,
+    private void AppendQuad(List<float> buf,
         float x, float y, float w, float h,
         float u0, float v0, float u1, float v1, Vector4 color)
     {
+        // AD-98 canvas stretch — see CanvasScale's doc comment. Applied after
+        // all canvas-space clipping, so geometry and UVs stay consistent.
+        if (CanvasScale != Vector2.One)
+        {
+            x *= CanvasScale.X;
+            y *= CanvasScale.Y;
+            w *= CanvasScale.X;
+            h *= CanvasScale.Y;
+        }
         // Two triangles (6 verts). CCW in pixel space is clockwise in NDC
         // because the vertex shader flips Y, so OpenGL's default front-face
         // is GL_CCW — we rely on cull-face being disabled during HUD pass.

src/AcDream.App/UI/Layout/CharacterManagementUiController.cs

@@ -38,6 +38,7 @@ internal sealed class CharacterManagementUiController : IDisposable
     private readonly List<UiButton> _rows = [];
     private readonly Dictionary<UiButton, uint> _rowIds = [];
 
+    private Vector2 _authoredCanvas;
     private RuntimeGenerationToken _lastGeneration;
     private long _lastRevision = long.MinValue;
     private uint _deleteDialogContext;
@@ -74,24 +75,22 @@ internal sealed class CharacterManagementUiController : IDisposable
 
         Root.Left = 0f;
         Root.Top = 0f;
-        Root.Anchors = AnchorEdges.Left | AnchorEdges.Top
-            | AnchorEdges.Right | AnchorEdges.Bottom;
-        if (host.Width > 0f)
-            Root.Width = host.Width;
-        if (host.Height > 0f)
-            Root.Height = host.Height;
         Root.ClickThrough = false;
         Root.Visible = false;
 
-        // Campaign LA gate round 2: this root is resized to the live viewport just
+        // Campaign LA gate round 2 (register AD-98): the root KEEPS its authored
-        // above, which is bigger than its authored 800x600 canvas at almost every
+        // 800×600 extent — retail never resizes it (zero edge anchors, verified
-        // real resolution. Its own DirectState background (RenderSurface 0x06007576)
+        // against the installed DAT) and its blitter has no stretch mode; the
-        // must scale to fill that resized rect, not tile — see
+        // whole composed screen stretches once at presentation. Our equivalent:
-        // UiDatElement.StretchOwnBackgroundToFill's doc comment for the retail
+        // while this screen is active, the host stretches the ENTIRE canvas —
-        // mechanism (a fixed-canvas screen stretched once at presentation) this
+        // widgets, glyphs, and the painted background (which carries the
-        // substitutes.
+        // "World"/"Characters" captions as art) — as one unit via
-        if (Root is UiDatElement rootBackground)
+        // UiRoot.FixedCanvasSize. Resizing the root here instead is exactly the
-            rootBackground.StretchOwnBackgroundToFill = true;
+        // half-substitution that misaligned the widgets against the stretched
+        // art at the 2026-08-15 user gate.
+        _authoredCanvas = new Vector2(
+            Root.Width > 0f ? Root.Width : 800f,
+            Root.Height > 0f ? Root.Height : 600f);
 
         // Create Character belongs to a future campaign. Keep retail's
         // authored control in place and visibly ghosted; do not hide it or
@@ -246,6 +245,7 @@ internal sealed class CharacterManagementUiController : IDisposable
         {
             _active = true;
             Root.Visible = true;
+            _host.FixedCanvasSize = _authoredCanvas;
             _host.BringToFront(Root);
         }
 
@@ -310,6 +310,7 @@ internal sealed class CharacterManagementUiController : IDisposable
         }
         finally
         {
+            _host.FixedCanvasSize = null;
             _enter.OnClick = null;
             _delete.OnClick = null;
             _restore.OnClick = null;
@@ -665,6 +666,7 @@ internal sealed class CharacterManagementUiController : IDisposable
         {
             _active = false;
             Root.Visible = false;
+            _host.FixedCanvasSize = null;
         }
         foreach (UiButton row in _rows)
         {

src/AcDream.App/UI/Layout/UiDatElement.cs

@@ -206,11 +206,9 @@ public class UiDatElement : UiElement, IUiDatStateful
     public uint? RuntimeImageTexture { get; set; }
 
     /// <summary>
-    /// When true, this element's OWN active-state background media draws as ONE quad
+    /// Retail background-blit ground truth (Campaign LA gate round 2, register
-    /// stretched to exactly fill <see cref="UiElement.Width"/>/<see cref="UiElement.Height"/>
+    /// AD-98). Every element draws its own media with the native-pixel TILE
-    /// (UV span 0,0 .. 1,1) instead of the native-pixel TILE formula every other
+    /// formula below — retail has no per-element stretch, and neither do we.
-    /// <see cref="UiDatElement"/> uses. Default false — every ordinary dat chrome/
-    /// container element (corners, edges, drag bars, tab backdrops) keeps tiling.
     ///
     /// <para>
     /// <b>Campaign LA gate round 2 (issue found in the live client: the LA8
@@ -247,21 +245,16 @@ public class UiDatElement : UiElement, IUiDatStateful
     /// </para>
     ///
     /// <para>
-    /// acdream has no offscreen fixed-resolution UI render target / present-time scale
+    /// acdream's equivalent of that present-time stretch is
-    /// pass — <see cref="AcDream.App.UI.Layout.CharacterManagementUiController"/> instead
+    /// <see cref="AcDream.App.UI.UiRoot.FixedCanvasSize"/>: while a fixed-canvas
-    /// resizes the MOUNTED ROOT ELEMENT itself to the live viewport (see its
+    /// screen (char select) is active, the WHOLE retained tree — this tile draw
-    /// constructor) so the screen still fills the window. This flag is the acknowledged
+    /// included — is scaled uniformly at the renderer's quad chokepoint, with the
-    /// divergence for that substitution (register row: acdream resizes the element,
+    /// inverse applied to mouse input. Elements therefore keep their authored
-    /// retail stretches the presented frame) — it makes the resized ROOT's own
+    /// canvas-space sizes here, and the tile formula stays exactly retail's:
-    /// background draw as one stretched quad so the VISUAL RESULT matches retail's
+    /// inside the authored canvas an element never exceeds its media's native
-    /// present-time stretch (no tiling) even though the MECHANISM differs. Set only on
+    /// span unless retail itself tiled it.
-    /// a screen-level mounted root, never on an ordinary descendant/chrome element —
-    /// those keep the native tile formula, which IS what retail's own blit does for
-    /// content that lives inside the (in retail) fixed 800x600 canvas.
     /// </para>
     /// </summary>
-    public bool StretchOwnBackgroundToFill { get; set; }
-
     protected override void OnDraw(UiRenderContext ctx)
     {
         if (MediaVisible && RuntimeImageTexture is uint runtimeTexture)
@@ -290,25 +283,16 @@ public class UiDatElement : UiElement, IUiDatStateful
             var (tex, tw, th) = _resolve(file);
             if (tex != 0 && tw != 0 && th != 0)
             {
-                if (StretchOwnBackgroundToFill)
+                // TILE at native size on both axes (UV-repeat; GL_REPEAT-wrapped
-                {
+                // UI texture) — retail's Graphic::Draw/Graphic::PutImage
-                    // One quad, UV 0..1 — see StretchOwnBackgroundToFill's doc comment
+                // (0x00693b20/0x00693a30) copy-or-tile blit; NOT ImgTex::TileCSI,
-                    // for the retail mechanism this substitutes (a fixed-canvas screen
+                // which is land-surface-only (corrected citation, see the class
-                    // stretched once at presentation).
+                // doc). Overlay/Alphablend use the same blit (the sprite shader
-                    ctx.DrawSprite(tex, 0, 0, Width, Height, 0, 0, 1, 1, Vector4.One);
+                // already alpha-blends). No Stretch mode exists in DrawModeType;
-                }
+                // whole-canvas stretching happens at UiRoot.FixedCanvasSize.
-                else
-                {
-                    // Normal → TILE at native size on both axes (UV-repeat; GL_REPEAT-wrapped
-                    // UI texture) — retail's Graphic::Draw/Graphic::PutImage (0x00693b20/
-                    // 0x00693a30) copy-or-tile blit; see StretchOwnBackgroundToFill's doc
-                    // comment for the corrected citation (NOT ImgTex::TileCSI, which is
-                    // land-surface-only). Overlay/Alphablend use the same blit (the sprite
-                    // shader already alpha-blends). No Stretch mode exists in DrawModeType.
                 ctx.DrawSprite(tex, 0, 0, Width, Height, 0, 0, Width / tw, Height / th, Vector4.One);
             }
         }
-        }
 
         DrawLabel(ctx);
     }

src/AcDream.App/UI/UiRoot.cs

@@ -32,6 +32,32 @@ public sealed class UiRoot : UiElement
     /// <summary>Single owner for named retained-window lifecycle and raise policy.</summary>
     public RetailWindowManager WindowManager { get; }
 
+    /// <summary>
+    /// Campaign LA gate round 2 (register AD-98): when set, the retained tree
+    /// is laid out in this fixed authored canvas (the char-select screen's
+    /// 800×600) and the whole tree — widgets, glyphs, art — is stretched to
+    /// the window as one unit, matching retail's present-time frame stretch
+    /// for fixed-canvas pre-world screens. Draw applies the scale at the
+    /// renderer's quad chokepoint; the mouse entry points apply the inverse,
+    /// so <see cref="MouseX"/>/<see cref="MouseY"/> and every hit test live
+    /// in canvas space. Null (the in-world default) is native 1:1.
+    /// </summary>
+    public Vector2? FixedCanvasSize { get; set; }
+
+    /// <summary>Window→canvas stretch factor; One when no fixed canvas is set.</summary>
+    public Vector2 CanvasScale =>
+        FixedCanvasSize is { X: > 0f, Y: > 0f } canvas && Width > 0f && Height > 0f
+            ? new Vector2(Width / canvas.X, Height / canvas.Y)
+            : Vector2.One;
+
+    private (int x, int y) MapWindowToCanvas(int x, int y)
+    {
+        Vector2 scale = CanvasScale;
+        return scale == Vector2.One
+            ? (x, y)
+            : ((int)MathF.Round(x / scale.X), (int)MathF.Round(y / scale.Y));
+    }
+
     // ── Device-level state ───────────────────────────────────────────────
     public int MouseX { get; private set; }
     public int MouseY { get; private set; }
@@ -370,6 +396,21 @@ public sealed class UiRoot : UiElement
     }
 
     public void Draw(UiRenderContext ctx)
+    {
+        // AD-98 fixed-canvas stretch: scope the renderer's canvas scale to
+        // exactly this tree's draws (world-space HUD stays native).
+        ctx.TextRenderer.CanvasScale = CanvasScale;
+        try
+        {
+            DrawCore(ctx);
+        }
+        finally
+        {
+            ctx.TextRenderer.CanvasScale = Vector2.One;
+        }
+    }
+
+    private void DrawCore(UiRenderContext ctx)
     {
         // Render children (panels) sorted by z-order — modal last so it
         // sits on top.
@@ -401,6 +442,7 @@ public sealed class UiRoot : UiElement
 
     public void OnMouseMove(int x, int y)
     {
+        (x, y) = MapWindowToCanvas(x, y);
         int dx = x - MouseX;
         int dy = y - MouseY;
         MouseX = x;
@@ -552,6 +594,7 @@ public sealed class UiRoot : UiElement
 
     public void OnMouseDown(UiMouseButton btn, int x, int y, uint flags = 0)
     {
+        (x, y) = MapWindowToCanvas(x, y);
         MouseX = x; MouseY = y;
         UpdateButtonFlag(btn, down: true);
         _pressX = x; _pressY = y;
@@ -707,6 +750,7 @@ public sealed class UiRoot : UiElement
 
     public void OnMouseUp(UiMouseButton btn, int x, int y, uint flags = 0)
     {
+        (x, y) = MapWindowToCanvas(x, y);
         MouseX = x; MouseY = y;
         UpdateButtonFlag(btn, down: false);
 

tests/AcDream.App.Tests/UI/Layout/CharacterManagementUiControllerTests.cs

@@ -9,19 +9,34 @@ namespace AcDream.App.Tests.UI.Layout;
 public sealed class CharacterManagementUiControllerTests
 {
     /// <summary>
-    /// Campaign LA gate round 2: the constructor resizes Root to the host viewport
+    /// Campaign LA gate round 2 (register AD-98): the root KEEPS its authored
-    /// (see the constructor's Root.Width/Height block) — its own background must
+    /// 800×600 extent (retail never resizes it — zero edge anchors), and while
-    /// therefore draw stretched, not tiled, or it visibly repeats at any resolution
+    /// the screen is active the HOST carries the fixed canvas so the whole tree
-    /// bigger than the authored 800x600 canvas. See
+    /// — widgets, glyphs, and the painted background whose art contains the
-    /// <see cref="AcDream.App.UI.Layout.UiDatElement.StretchOwnBackgroundToFill"/>.
+    /// World/Characters captions — stretches together. Resizing the root while
+    /// stretching only the art is exactly the misalignment the 2026-08-15 user
+    /// gate caught. Dispose must release the canvas so in-world UI returns to
+    /// native pixels.
     /// </summary>
     [Fact]
-    public void Constructor_MarksRootBackgroundToStretch_NotTile()
+    public void ActiveScreen_KeepsAuthoredRootExtent_AndSetsHostFixedCanvas()
     {
-        using var environment = new EnvironmentHarness();
+        var environment = new EnvironmentHarness();
+        try
+        {
+            UiElement root = environment.Controller.Root;
+            Assert.Equal(800f, root.Width);
+            Assert.Equal(600f, root.Height);
+            Assert.Equal(
+                new Vector2(root.Width, root.Height),
+                environment.Host.FixedCanvasSize);
+        }
+        finally
+        {
+            environment.Dispose();
+        }
 
-        var root = Assert.IsType<UiDatElement>(environment.Controller.Root);
+        Assert.Null(environment.Host.FixedCanvasSize);
-        Assert.True(root.StretchOwnBackgroundToFill);
     }
 
     [Fact]

tests/AcDream.App.Tests/UI/Layout/UiDatElementTests.cs

@@ -23,16 +23,14 @@ public class UiDatElementTests
     }
 
     /// <summary>
-    /// Campaign LA gate round 2: the char-select root's background (native 800x600,
+    /// Retail's blit is copy-or-tile only (Graphic::Draw @0x00693b20 — no
-    /// resolved from a JPEG surface) was drawn with the ordinary UiDatElement TILE
+    /// stretch mode exists), so an element grown past its media's native size
-    /// UV formula (u1 = Width/tw) after CharacterManagementUiController resized the
+    /// tiles: u1 = Width/tw. Whole-screen stretching is NOT this layer's job —
-    /// root to the live viewport — at 1920x1080 that produces u1 = 2.4, v1 = 1.8,
+    /// it happens uniformly at UiRoot.FixedCanvasSize / TextRenderer.CanvasScale
-    /// which GL_REPEAT wraps into a visibly tiled background instead of one stretched
+    /// (register AD-98), covered by the test below.
-    /// image. See <see cref="UiDatElement.StretchOwnBackgroundToFill"/>'s doc comment
-    /// for the retail mechanism this substitutes.
     /// </summary>
     [Fact]
-    public void StretchOwnBackgroundToFill_False_TilesUvPastOne_WhenRectExceedsNativeSize()
+    public void OwnBackground_TilesUvPastOne_WhenRectExceedsNativeSize()
     {
         var info = new ElementInfo { Width = 1920, Height = 1080 };
         info.StateMedia[""] = (0x06007576u, 1);
@@ -58,43 +56,14 @@ public class UiDatElementTests
     }
 
     /// <summary>
-    /// Campaign LA gate round 2 fix: with the flag set, the SAME oversized rect draws
+    /// AD-98 whole-canvas stretch: with the renderer's CanvasScale set (the
-    /// as one quad spanning UV 0..1 — a single stretched image, matching retail's
+    /// char-select 800×600 canvas on a 1920×1080 window), an element drawn at
-    /// observed (never-tiled) char-select background.
+    /// authored size emits a quad scaled by exactly (2.4, 1.8) in GEOMETRY while
+    /// its UVs stay authored (0..1 here) — one stretched image, no tiling, the
+    /// same math retail's present-time frame stretch produces.
     /// </summary>
     [Fact]
-    public void StretchOwnBackgroundToFill_True_ClampsUvToOne_WhenRectExceedsNativeSize()
+    public void CanvasScale_StretchesQuadGeometry_LeavesUvsAuthored()
-    {
-        var info = new ElementInfo { Width = 1920, Height = 1080 };
-        info.StateMedia[""] = (0x06007576u, 1);
-        var e = new UiDatElement(info, _ => (7u, 800, 600))
-        {
-            Left = 0,
-            Top = 0,
-            Width = 1920,
-            Height = 1080,
-            StretchOwnBackgroundToFill = true,
-        };
-
-        (TextRenderer renderer, UiRenderContext ctx) = BuildRenderContext();
-        e.DrawSelfAndChildren(ctx);
-
-        var (texture, verts) = Assert.Single(renderer.DebugSpriteSegmentVerts);
-        Assert.Equal(7u, texture);
-        float uMax = verts[1 * 8 + 2];
-        float vMax = verts[1 * 8 + 3];
-        Assert.Equal(1f, uMax, 3);
-        Assert.Equal(1f, vMax, 3);
-    }
-
-    /// <summary>
-    /// The flag must not change anything for an element whose rect already matches
-    /// its native texture size (every ordinary panel/window root today) — stretch
-    /// (UV 0..1) and tile (UV Width/tw) are numerically identical at that size, so
-    /// this only changes behavior for elements deliberately grown past their art.
-    /// </summary>
-    [Fact]
-    public void StretchOwnBackgroundToFill_True_MatchesTile_WhenRectEqualsNativeSize()
     {
         var info = new ElementInfo { Width = 800, Height = 600 };
         info.StateMedia[""] = (0x06007576u, 1);
@@ -104,13 +73,23 @@ public class UiDatElementTests
             Top = 0,
             Width = 800,
             Height = 600,
-            StretchOwnBackgroundToFill = true,
         };
 
         (TextRenderer renderer, UiRenderContext ctx) = BuildRenderContext();
+        renderer.CanvasScale = new Vector2(1920f / 800f, 1080f / 600f);
+        try
+        {
             e.DrawSelfAndChildren(ctx);
+        }
+        finally
+        {
+            renderer.CanvasScale = Vector2.One;
+        }
 
         var (_, verts) = Assert.Single(renderer.DebugSpriteSegmentVerts);
+        // Far corner (vertex 1): geometry scaled to the window, UVs authored.
+        Assert.Equal(1920f, verts[1 * 8 + 0], 3);
+        Assert.Equal(1080f, verts[1 * 8 + 1], 3);
         Assert.Equal(1f, verts[1 * 8 + 2], 3);
         Assert.Equal(1f, verts[1 * 8 + 3], 3);
     }

tests/AcDream.App.Tests/UI/UiRootFixedCanvasTests.cs

@@ -0,0 +1,121 @@
+using System.Numerics;
+using AcDream.App.UI;
+using AcDream.App.UI.Layout;
+using Xunit;
+
+namespace AcDream.App.Tests.UI;
+
+/// <summary>
+/// Campaign LA gate round 2 (register AD-98): retail renders fixed-canvas
+/// pre-world screens (char select, authored 800×600) at authored size and
+/// stretches the whole composed frame at presentation — its UI blitter has no
+/// stretch mode. acdream's equivalent is <see cref="UiRoot.FixedCanvasSize"/>:
+/// draw applies one uniform scale at the renderer's quad chokepoint, and the
+/// mouse entry points apply the exact inverse so hit-testing lives in canvas
+/// space. These tests pin the scale math and the inverse input mapping — the
+/// half that, if wrong, makes the user click on art and hit nothing.
+/// </summary>
+public class UiRootFixedCanvasTests
+{
+    [Fact]
+    public void CanvasScale_IsOne_WithoutFixedCanvas()
+    {
+        var root = new UiRoot { Width = 1920f, Height = 1080f };
+        Assert.Equal(Vector2.One, root.CanvasScale);
+    }
+
+    [Fact]
+    public void CanvasScale_IsWindowOverCanvas_WhenFixed()
+    {
+        var root = new UiRoot
+        {
+            Width = 1920f,
+            Height = 1080f,
+            FixedCanvasSize = new Vector2(800f, 600f),
+        };
+        Assert.Equal(new Vector2(2.4f, 1.8f), root.CanvasScale);
+    }
+
+    [Fact]
+    public void CanvasScale_IsOne_ForDegenerateCanvasOrWindow()
+    {
+        var zeroCanvas = new UiRoot
+        {
+            Width = 1920f,
+            Height = 1080f,
+            FixedCanvasSize = new Vector2(0f, 600f),
+        };
+        Assert.Equal(Vector2.One, zeroCanvas.CanvasScale);
+
+        var zeroWindow = new UiRoot
+        {
+            Width = 0f,
+            Height = 0f,
+            FixedCanvasSize = new Vector2(800f, 600f),
+        };
+        Assert.Equal(Vector2.One, zeroWindow.CanvasScale);
+    }
+
+    /// <summary>
+    /// The load-bearing inverse: a click at WINDOW coordinates must land on the
+    /// widget whose authored CANVAS rect the user visually clicked. The button
+    /// sits at canvas (300,400)+(120×40); at a 1920×1080 window over an 800×600
+    /// canvas it appears at window (720,720)-(1008,792). Clicking window
+    /// (860,750) — canvas (358,417) — must click it; clicking window (300,400)
+    /// — canvas (125,222), visually empty — must not.
+    /// </summary>
+    [Fact]
+    public void MouseInput_MapsWindowCoordsToCanvasSpace()
+    {
+        var root = new UiRoot
+        {
+            Width = 1920f,
+            Height = 1080f,
+            FixedCanvasSize = new Vector2(800f, 600f),
+        };
+        int clicks = 0;
+        var button = new UiButton(
+            new ElementInfo { Width = 120, Height = 40 },
+            _ => (0u, 0, 0))
+        {
+            Left = 300f,
+            Top = 400f,
+            Width = 120f,
+            Height = 40f,
+            OnClick = () => clicks++,
+        };
+        root.AddChild(button);
+
+        root.OnMouseDown(UiMouseButton.Left, 860, 750);
+        root.OnMouseUp(UiMouseButton.Left, 860, 750);
+        Assert.Equal(1, clicks);
+        Assert.Equal(358, root.MouseX);
+        Assert.Equal(417, root.MouseY);
+
+        root.OnMouseDown(UiMouseButton.Left, 300, 400);
+        root.OnMouseUp(UiMouseButton.Left, 300, 400);
+        Assert.Equal(1, clicks);
+    }
+
+    [Fact]
+    public void MouseInput_IsUntouched_WithoutFixedCanvas()
+    {
+        var root = new UiRoot { Width = 1920f, Height = 1080f };
+        int clicks = 0;
+        var button = new UiButton(
+            new ElementInfo { Width = 120, Height = 40 },
+            _ => (0u, 0, 0))
+        {
+            Left = 300f,
+            Top = 400f,
+            Width = 120f,
+            Height = 40f,
+            OnClick = () => clicks++,
+        };
+        root.AddChild(button);
+
+        root.OnMouseDown(UiMouseButton.Left, 360, 420);
+        root.OnMouseUp(UiMouseButton.Left, 360, 420);
+        Assert.Equal(1, clicks);
+    }
+}