more work done

docs/sprite-node-class-layout.md

@@ -109,14 +109,19 @@ So the safe future modeling strategy is:
 
 ## Candidate Virtual Slot Map
 
-The currently verified slots are already good enough for a first typed vtable.
+`DispatchEvent` now gives a materially tighter slot map than the earlier placeholder `A/B/C/D` read.
 
 | Slot offset | Current best role | Evidence |
 |---|---|---|
-| `+0x14` | event handler A | `sprite_node_dispatch_event` dispatches here for one event class |
-| `+0x18` | event handler B | same dispatcher |
-| `+0x20` | event handler C | same dispatcher |
-| `+0x24` | event handler D | same dispatcher |
+| `+0x04` | event `1` handler | `DispatchEvent` routes event code `1` here directly |
+| `+0x08` | event `2` handler | same dispatcher |
+| `+0x0c` | event `4` handler | same dispatcher |
+| `+0x10` | event `8` handler | same dispatcher |
+| `+0x14` | event `0x10` handler | same dispatcher |
+| `+0x18` | event `0x20` handler | same dispatcher |
+| `+0x1c` | event `0x40` self handler | called after the dispatcher walks child nodes for the same event |
+| `+0x24` | event `0x100` handler | same dispatcher |
+| `+0x34` | child-broadcast event `0x40` slot | used on child nodes during the `0x40` walk, not on the root dispatch object itself |
 
 The seg091 default-slot helpers are also useful evidence:
 
@@ -153,11 +158,45 @@ When class authoring begins, the safest sequence for this family is:
 4. create provisional vtable with slots `+0x14`, `+0x18`, `+0x20`, `+0x24`
 5. keep recursive tree helpers outside the class until decompiler output shows they benefit from becoming methods
 
+## Live Ghidra Authoring Status
+
+Verified first live `SpriteNode` batch landed on 2026-04-08.
+
+- Created class owner `Remorse::SpriteNode` in the active `CRUSADER.EXE` database.
+- Re-anchored the strongest old `000b:` method batch into the live `1360:` segment by preserved offset delta from `000b:326e -> 1360:046e`.
+- Created minimal live datatypes `/Remorse/SpriteNodeBase` and `/Remorse/SpriteNodeVtable` from the current safest note anchors.
+- `/Remorse/SpriteNodeBase` currently names only the bounded field block that is directly supported by the live method batch:
+	- `+0x19 = child_or_next_farptr`
+	- `+0x21 = local_x_offset`
+	- `+0x23 = local_y_offset`
+	- `+0x29 = dirty_flags`
+- `/Remorse/SpriteNodeVtable` is still the earlier minimal shell in-session, but the live `DispatchEvent` read now supports a deeper slot map than the first authoring batch encoded:
+	- direct self dispatch uses `+0x04`, `+0x08`, `+0x0c`, `+0x10`, `+0x14`, `+0x18`, `+0x1c`, and `+0x24`
+	- child broadcast during event `0x40` uses child slot `+0x34`
+- Moved the first bounded method set under the class owner with short provenance comments:
+	- `1360:036a` -> `Create` (best current constructor-style anchor; still keep the higher-wrapper caveat visible)
+	- `1360:046e` -> `Destroy` (older note anchor `000b:326e`)
+	- `1360:0580` -> `IsDirty` (older note anchor `000b:3380`)
+	- `1360:05a6` -> `MarkDirty` (older note anchor `000b:33a6`)
+	- `1360:0955` -> `GetOrTraverse` (best current live anchor for older note anchor `000a:b988`)
+	- `1360:0cb2` -> `DispatchEvent` (older note anchor `000b:3ab2`)
+	- `1360:12ee` -> `UpdateAndDispatch` (older note anchor `000b:40ee`)
+- The live decompiler now makes the core family surface much easier to navigate directly in-session:
+	- `Create` now exists live as the current safest constructor-style anchor: it allocates `0x34` bytes when `this` is null, stamps the `0x501a` SpriteNode vtable, initializes the child-link/core offset fields, and links the incoming parent/child lane.
+	- Direct callers now narrow the subtype story materially: the current call set is dominated by `GumpCreate_*` and adjacent UI wrapper constructors, which supports treating `Create` as the compact shared base-node constructor used by higher-level gump/display objects rather than as a one-off derived leaf.
+	- `Destroy` restores the `0x501a` base vtable, clears the global focus pointer when `this` owns it, releases child linkage, and optionally frees self.
+	- `IsDirty` and `MarkDirty` read and mutate the `+0x29` dirty-state lane exactly as the note predicted.
+	- `GetOrTraverse` now exists live as the best current `000a:b988` re-anchor: it recursively walks the child-linked subtree, adjusts the incoming query coordinates by the local offsets, and returns either the matched child node or the default sentinel through the out pointer.
+	- `DispatchEvent` updates the global focus pointer at `0x4fd0:0x4fd2` and now ties concrete event codes to concrete slot offsets: `1/2/4/8 -> +0x04/+0x08/+0x0c/+0x10`, `0x10/0x20 -> +0x14/+0x18`, `0x40 -> child +0x34 then optional self +0x1c`, and `0x100 -> +0x24`.
+	- `UpdateAndDispatch` now clearly shows the `IsDirty` / `MarkDirty` / recompute / child-walk sequence instead of staying as an anonymous seg1360 helper.
+- The `SpriteNode` family is therefore no longer note-only. What remains open is narrower now: chiefly whether `Create` should remain the family's public constructor-style entry or later be split into a higher derived/UI wrapper and a smaller base-node constructor once more callers and subtype evidence land, plus the deeper vtable-slot and subtype-layout questions.
+
 ## Open Questions
 
+- whether the current `Create` signature itself can be cleaned up further now that its caller set points to a shared compact base-node constructor used by higher-level gump wrappers
 - exact root vtable address or addresses for the main SpriteNode family
 - whether the `+0x17e` redraw flag belongs to a derived display node rather than the compact base node
-- which event-code cases map to which slot semantically beyond the current `A/B/C/D` placeholder naming
+- whether the recovered event-code map can now be promoted from raw event-code labels to prettier semantic slot names without overfitting UI behavior too early
 - whether `sprite_tree_accumulate_pos` should become a class method, a static helper, or a separate geometry utility
 
 ## Immediate Follow-Up Value