2de9cc1c19
Stream B.1 of the 2026-06-18 handoff — the shared widget-level drag-drop infra that both shortcut-drag (B) and the inventory window (C) sit on. Four design decisions confirmed with the user this session: 1. Payload = typed ItemDragPayload record snapshotted at drag-begin (ObjId/SourceKind/SourceSlot/SourceCell); SourceContainer derived at drop from ClientObjectTable (single source of truth). 2. Cursor ghost painted by UiRoot via a generic UiElement.GetDragGhost() hook — keeps UiRoot item-agnostic; floats above all windows. 3. Cell (UiItemSlot) is the drop-target hit unit + accept/reject overlay owner, delegating the decision + dispatch UP to its parent UiItemList's registered IItemListDragHandler (faithful to retail cell->ItemList_DragOver ->m_dragHandler; scales to the inventory N-cell grid). 4. PR ships infra + a visible toolbar STUB handler (logs, no wire) so the ghost/overlay/dispatch are confirmable this session; AddShortcut/Remove wire is Stream B.2. Retail-grounded: InqDropIconInfo flags (&0xE==0 fresh / &4 reorder, reject state 0x10000040) confirmed live at gmToolbarUI 0x004bd162; the cell begin-drag/CatchDroppedItem/RegisterItemListDragHandler chain at decomp 229344/229744/230461. Planned register rows: AP-47 (ghost reuses full icon at reduced alpha vs retail m_pDragIcon) + TS-33 (toolbar drop stub pending B.2). No new wire format in the spine itself. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> |
||
|---|---|---|
| .. | ||
| architecture | ||
| audit | ||
| plans | ||
| research | ||
| superpowers | ||
| bugs.md | ||
| ISSUES.md | ||