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

feat(ui): persist retained window layouts

Browse source
This commit is contained in:
Erik 2026-07-10 23:17:29 +02:00
parent a8e9503d2e
commit 921c388e2c
22 changed files with 705 additions and 141 deletions
Download patch file Download diff file Expand all files Collapse all files

4
docs/architecture/retail-divergence-register.md
View file

@ -54,7 +54,7 @@ accepted-divergence entries (#96, #49, #50).
| IA-12 | UI toolkit mirrors retail behavior from research docs, not a byte-port — keystone.dll is outside decomp coverage; observed constants embedded (drag 3 px, tooltip 1000 ms) | `src/AcDream.App/UI/README.md:3` (window mgr: `UiRoot.cs` RegisterWindow/Show/Hide/Toggle/BringToFront — F12 inventory toggle IS retail-faithful) | keystone.dll has no PDB/decomp; semantics reconstructed from the six `docs/research/retail-ui/` deep-dives, keeping retail's event-type constants so panel switch-cases transplant cleanly | Edge-case input semantics the research under-specified (drag threshold, tooltip timing, focus hand-off, capture corners) differ silently with no oracle to diff against | keystone.dll Device DAT_00837ff4; docs/research/retail-ui/04-input-events.md |
| IA-13 | GameEventType registry deliberately omits event types retail ignores; unknown events fall through unhandled | `src/AcDream.Core.Net/Messages/GameEventType.cs:11` | Retail also ignores them — dropping matches retail by construction | If the "retail ignores X" judgment is wrong for any opcode (or a server mod uses one), the event is silently dropped with no diagnostic pointing at the omission | retail GameEvent dispatch (ignored-event set) |
| IA-14 | Rendering + dat-handling base is WorldBuilder's tested port, not a fresh retail-decomp port (Phase N.4/O design stance) | `docs/architecture/worldbuilder-inventory.md` (code at `src/AcDream.{Core,App}/Rendering/Wb/`) | WB visually verified on the AC world, MIT, same stack; known WB↔retail deltas resolved case-by-case — terrain split kept retail `FSplitNESW` (**#51**, pinned by `SplitFormulaDivergenceTest`), scenery drift accepted (AP-31) | A WB-upstream divergence not yet caught ships silently as "our" behavior; guard = the inventory doc's 🟢/🔴 split + per-formula divergence tests | retail decomp per algorithm; `tests/.../SplitFormulaDivergenceTest.cs` |
| IA-15 | D.2b gameplay UI is our own `UiHost`/`UiRoot` retained tree, not a byte-port of Keystone. `RetailWindowManager`/typed handles centralize registry, raise, focus/capture cleanup, lifecycle events, and controller teardown; `RetailWindowFrame` is now the single production/Studio mount contract for both imported-chrome and shared-wrapper windows. Production LayoutDesc imports include vitals `0x2100006C`, chat `0x21000006`, toolbar `0x21000016`, character `0x2100002E`, inventory `0x21000023` plus mounted `0x21000024/22/21`, and radar `0x21000074`. | `src/AcDream.App/UI/RetailWindowManager.cs`; `src/AcDream.App/UI/Layout/RetailWindowFrame.cs`; `src/AcDream.App/UI/Layout/LayoutImporter.cs`; production mounts in `GameWindow.cs` | Keystone has no matching PDB/decomp, so we preserve its observable ElementDesc/state/input behavior from DAT, named client call sites, and live evidence while using modern retained ownership. Real RenderSurfaces and imported element geometry remain the visual oracle. | Unported persistence semantics can still make windows drift across restart/resolution changes; lifecycle edge cases remain constrained by conformance tests rather than a Keystone byte-port | Production LayoutDesc objects; `docs/research/2026-06-15-layoutdesc-format.md`; Keystone behavior notes in `docs/research/retail-ui/` |
| IA-15 | D.2b gameplay UI is our own `UiHost`/`UiRoot` retained tree, not a byte-port of Keystone. `RetailWindowManager`/typed handles centralize registry, raise, focus/capture cleanup, lifecycle events, controller teardown, and schema-v2 per-character/per-resolution layout persistence; `RetailWindowFrame` is the single production/Studio mount contract for imported-chrome and shared-wrapper windows. Production LayoutDesc imports include vitals `0x2100006C`, chat `0x21000006`, toolbar `0x21000016`, character `0x2100002E`, inventory `0x21000023` plus mounted `0x21000024/22/21`, and radar `0x21000074`. | `src/AcDream.App/UI/RetailWindowManager.cs`; `src/AcDream.App/UI/RetailWindowLayoutPersistence.cs`; `src/AcDream.App/UI/Layout/RetailWindowFrame.cs`; `src/AcDream.App/UI/Layout/LayoutImporter.cs`; production mounts in `GameWindow.cs` | Keystone has no matching PDB/decomp, so we preserve its observable ElementDesc/state/input behavior from DAT, named client call sites, and live evidence while using modern retained ownership. Real RenderSurfaces and imported element geometry remain the visual oracle. | Persistence is behaviorally reconstructed from `saveui/loadui` semantics rather than Keystone internals; lifecycle edge cases remain constrained by conformance tests instead of a byte-port | Production LayoutDesc objects; `docs/research/2026-06-15-layoutdesc-format.md`; Keystone behavior notes in `docs/research/retail-ui/` |
| IA-17 | Toolbar chrome is toolkit-supplied through the central `RetailWindowFrame` mount (`UiCollapsibleFrame` 8-piece bevel) because LayoutDesc `0x21000016` carries no baked frame. It also supports a toolkit-defined collapse-to-one-row (bottom-edge resize snapping between a row-1-only and a two-row height, row-2 visibility tied to the stop) — retail's real collapse is keystone.dll (no decomp) and the DAT stacks both rows always. | `src/AcDream.App/UI/Layout/RetailWindowFrame.cs`; `src/AcDream.App/UI/UiCollapsibleFrame.cs`; toolbar policy in `GameWindow.cs`; spec: `docs/superpowers/specs/2026-06-20-d2b-toolbar-collapse-design.md` | The central mount now owns wrapper geometry/registration uniformly; border-over-content prevents the row-2 right cap from poking through | The collapse stops remain a toolkit reconstruction rather than a byte-port of Keystone behavior | gmToolbarUI WM chrome (keystone.dll, no PDB); no bevel ids in LayoutDesc 0x21000016 (toolbar dump) |
| IA-18 | Effect overlay tile (enum 0x10000005) is a `ReplaceColor` SURFACE SOURCE — pure-white pixels in the composited drag icon are replaced PER-PIXEL with the same (x,y) pixel of the effect tile (the SURFACE overload `SurfaceWindow::ReplaceColor` 0x004415b0), preserving the tile's texture/gradient; the tile itself is NOT blitted as an additional layer. This IS faithful retail behavior. **Anti-regression: do NOT re-implement this as a blit layer NOR as a flat-color replace (it is a per-pixel surface copy).** | `src/AcDream.App/UI/IconComposer.cs` (`ReplaceWhiteFromSurface`) | Faithful port of `IconData::RenderIcons` @407614 → the SURFACE overload `ReplaceColor` 0x004415b0 (`dst[x,y]=src[x,y]` where `dst==white`); confirmed via clean Ghidra decompile + named decomp + visual (the Energy Crystal's blue is a gradient, 2026-06-17). | A blit-layer or flat-color re-implementation would show the wrong effect look (no gradient) — the visual-verification regression that retired the mean-color approximation | `IconData::RenderIcons` acclient_2013_pseudo_c.txt:407524; `ReplaceColor` SURFACE overload 0x004415b0:71656; `docs/research/2026-06-17-stateful-icon-RESOLVED.md` |
@ -203,7 +203,7 @@ AP-94..AP-110 for the confirmed retail-UI completion gaps.
| AP-95 | `UiButton` dispatches Click only; it lacks retail hover, pressed, released-outside, disabled, selected/toggle, and missing-state fallback transitions | `src/AcDream.App/UI/UiButton.cs` | Controllers manually set a few semantic states; generic pointer visuals remain incomplete | Buttons look inert or overwrite custom states, and disabled/toggle behavior can fire incorrectly | `UIElement_Button::UpdateState_ @ 0x00471CF0`; mouse handlers `0x00471FF0..0x004721F0` |
| AP-96 | Layout/property import loses raw edge-mode semantics (especially 3/4) and treats default scalar values as absent, so explicit `false`/`0` cannot override inheritance | `src/AcDream.App/UI/Layout/ElementReader.cs`; `LayoutImporter.cs` | Existing shipped layouts are hand-gated; Wave 1 adds presence-aware properties and the exact solver | Non-reference resizing recenters/stretches incorrectly; derived DAT properties silently inherit the wrong base value | `UIElement::UpdateForParentSizeChange @ 0x00462640`; production LayoutDesc inheritance |
| ~~AP-97~~ | **RETIRED 2026-07-10 (Wave 1 retained-widget foundation)** — the prior generic Device-timer premise was a decomp misread. Named retail polls the hovered element's tooltip deadline, then broadcasts global UI time message `3`; it does not expose the assumed arbitrary timer queue. `UiRoot.Tick` now preserves that order and subtree removal clears input/time ownership before another pulse. | `src/AcDream.App/UI/UiRoot.cs`; `IUiGlobalTimeListener.cs` | — | — | `UIElementManager::UseTime` and hover/focus paths, pinned in `docs/research/2026-07-10-retained-widget-foundations-pseudocode.md` |
| AP-98 | **PARTIAL 2026-07-10 (Wave 2.2)** — typed handles/manager own lifecycle and every production/Studio window uses `RetailWindowFrame`. Chat maximize now ports the half-parent growth/restore algorithm, resizes the outer frame, uses DAT 100–360 height limits, and drives imported button states. Complete layout persistence is still absent. | `src/AcDream.App/UI/RetailWindowManager.cs`; `RetailWindowHandle.cs`; `Layout/RetailWindowFrame.cs`; `ChatWindowController.cs` | Lifecycle, mounts, and chat outer geometry now have one tested owner; Wave 2.3 adds persistence | Layouts still reset/drift across restart and resolution/character changes | `gmMainChatUI::HandleMaximizeButton @ 0x004CCE50`; retail saveui/loadui behavior |
| ~~AP-98~~ | **RETIRED 2026-07-10 (Wave 2.3)** — typed handles/manager own lifecycle; every production/Studio window uses `RetailWindowFrame`; schema-v2 `UiWindowLayout` persists outer bounds, visibility, toolbar collapse, and chat maximize by character/resolution with legacy-radar migration, nearest-resolution fallback, clamping, and a pre-login `default` write guard. Chat maximize uses DAT 100–360 limits and preserves its lower edge when growing upward. | `src/AcDream.App/UI/RetailWindowManager.cs`; `RetailWindowLayoutPersistence.cs`; `RetailWindowHandle.cs`; `Layout/RetailWindowFrame.cs`; `ChatWindowController.cs`; `SettingsStore.cs` | — | — | `gmMainChatUI::HandleMaximizeButton @ 0x004CCE50`; retail `saveui/loadui` behavior |
| AP-99 | Shared item activation implements only a subset of `ItemHolder::UseObject`/placement legality and confirmation policy | `src/AcDream.App/UI/ItemInteractionController.cs` | Current PvE use/equip/container flows cover the M2 loop; Wave 3 ports the complete pure decision matrix | Vendor/trade/busy/confirmation/target edge cases send the wrong action or permit an illegal use | `ItemHolder::DetermineUseResult @ 0x00588460`; `UseObject @ 0x00588A80`; `AttemptPlaceIn3D @ 0x00588600` |
| AP-100 | Cursor semantics cover only a reachable subset of retail states, and widget cursor media can outrank authoritative global target mode | `src/AcDream.App/UI/CursorFeedbackController.cs`; `src/AcDream.App/Rendering/RetailCursorManager.cs` | Target-use art works for current flows; Wave 3 ports the full table and precedence | Hovering UI can replace the use/examine/busy/combat cursor; unsupported branches fall back to OS art | `ClientUISystem::UpdateCursorState @ 0x00564630` |
| AP-101 | Toolbar quick-slot actions and most panel/Use/Examine/combat controls are not fully dispatched; selected-object mana and stack entry/slider are absent | `src/AcDream.App/UI/Layout/ToolbarController.cs`; `GameWindow.OnInputAction` | Item click/use, stance indicators, health/name, and some toggles are present | Bound retail keys/buttons do nothing and item/stack selections show incomplete state | `gmToolbarUI @ 0x004BD0C0..0x004BF380`; LayoutDesc `0x21000016` |