feat(headless): Implement GhostObjectManagerDummy

[bobtista]

Summary

• Adds GhostObjectManagerDummy class that provides no-op implementations for headless mode
• Factory in W3DGameLogic returns dummy when m_headless is true
• Intentionally does NOT override crc/xfer/loadPostProcess to maintain save compatibility

Related

Split from #2139 per @xezon's suggestion to review each dummy class separately.

[greptile-apps]

Greptile Summary

This PR introduces GhostObjectManagerDummy, a no-op subclass of GhostObjectManager used in headless server mode. The factory method createGhostObjectManager is updated in both GameLogic (base) and W3DGameLogic (W3D device) to accept a bool dummy flag driven by TheGlobalData->m_headless, and all changes are applied symmetrically across both the Generals/ and GeneralsMD/ trees.

Key points:

• GhostObjectManagerDummy stubs out all rendering-related virtual methods with no-ops and returns nullptr from addGhostObject, correctly preventing any ghost-object bookkeeping in headless mode.
• crc/xfer/loadPostProcess are intentionally not overridden so that save-game serialisation uses the base-class path unchanged — this is well-documented in both the PR description and the in-code comment.
• A static_cast<GhostObjectManager*> that looks redundant is a known VC6 compiler requirement (acknowledged in a prior review thread).
• Both W3DGhostObject.cpp files also drop the W3DDisplay::m_3DScene != nullptr && guard in W3DRenderObjectSnapshot::addToScene(). Since this code is now exclusively reached in W3D (non-headless) mode — where m_3DScene should always be initialised — the change is logically safe, but it removes a defensive check that previously prevented silent failures; see the inline comment for details.

Confidence Score: 4/5

• Safe to merge; the only notable change beyond the dummy class itself is the removal of a defensive null guard that is now logically unreachable in its original context.
• The core GhostObjectManagerDummy implementation is clean and well-reasoned. The changes are mirrored correctly across both game trees. The single concern is the removal of W3DDisplay::m_3DScene != nullptr in addToScene() — in practice safe because that code path is only reached in W3D non-headless mode, but it eliminates a safety net that was present for a reason.
• Generals/Code/GameEngineDevice/Source/W3DDevice/GameLogic/W3DGhostObject.cpp and its GeneralsMD counterpart — specifically the removed null guard in W3DRenderObjectSnapshot::addToScene().

Important Files Changed

Filename	Overview
Generals/Code/GameEngine/Include/GameLogic/GhostObject.h	Adds GhostObjectManagerDummy class with clean no-op overrides for all virtual methods; intentionally skips crc/xfer/loadPostProcess for save compatibility as documented.
Generals/Code/GameEngine/Source/GameLogic/System/GameLogic.cpp	Passes TheGlobalData->m_headless to createGhostObjectManager; base factory now returns GhostObjectManagerDummy when headless. Clean and straightforward.
Generals/Code/GameEngineDevice/Include/W3DDevice/GameLogic/W3DGameLogic.h	Override updated to accept bool dummy parameter; returns GhostObjectManagerDummy for headless and W3DGhostObjectManager otherwise. The static_cast is a known VC6 requirement (acknowledged in prior review).
Generals/Code/GameEngineDevice/Source/W3DDevice/GameLogic/W3DGhostObject.cpp	Removes W3DDisplay::m_3DScene != nullptr && null guard from addToScene(). Safe in practice now that headless mode uses the dummy manager, but removes defensive protection for W3D edge cases.
GeneralsMD/Code/GameEngine/Include/GameLogic/GhostObject.h	Identical GhostObjectManagerDummy addition as in Generals counterpart; correctly mirrors the Generals changes with proper Zero Hour copyright header.
GeneralsMD/Code/GameEngineDevice/Source/W3DDevice/GameLogic/W3DGhostObject.cpp	Same null guard removal as the Generals variant — same concern applies regarding losing the defensive check for m_3DScene.
GeneralsMD/Code/GameEngine/Source/GameLogic/System/GameLogic.cpp	Mirrors the Generals change; createGhostObjectManager(TheGlobalData->m_headless) correctly wires headless detection into the factory.
GeneralsMD/Code/GameEngineDevice/Include/W3DDevice/GameLogic/W3DGameLogic.h	Identical to Generals counterpart; correctly overrides createGhostObjectManager(bool dummy) and dispatches to GhostObjectManagerDummy or W3DGhostObjectManager.
Generals/Code/GameEngine/Include/GameLogic/GameLogic.h	Virtual createGhostObjectManager() signature updated to createGhostObjectManager(bool dummy = false) — default value preserves backward compatibility for any subclasses that don't override it.
GeneralsMD/Code/GameEngine/Include/GameLogic/GameLogic.h	Mirrors Generals counterpart; createGhostObjectManager(bool dummy = false) declaration with default parameter is consistent and safe.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A["GameLogic::init()"] --> B["createGhostObjectManager(TheGlobalData->m_headless)"]
    B --> C{dummy == true?}
    C -- "headless mode" --> D["NEW GhostObjectManagerDummy\n(no-op stubs)"]
    C -- "normal mode\n(base GameLogic)" --> E["NEW GhostObjectManager"]
    C -- "normal mode\n(W3DGameLogic)" --> F["NEW W3DGhostObjectManager\n(full W3D rendering)"]
    D --> G["TheGhostObjectManager = dummy\nAll calls are no-ops"]
    E --> H["TheGhostObjectManager = base"]
    F --> I["TheGhostObjectManager = W3D\nW3DRenderObjectSnapshot::addToScene()\ncalls m_3DScene->Add_Render_Object()"]

Loading

Prompt To Fix All With AI

This is a comment left during a code review.
Path: Generals/Code/GameEngineDevice/Source/W3DDevice/GameLogic/W3DGhostObject.cpp
Line: 154-162

Comment:
**Removed null guard for `m_3DScene`**

The previous check `W3DDisplay::m_3DScene != nullptr &&` guarded against a null pointer dereference on `W3DDisplay::m_3DScene->Add_Render_Object(m_robj)`. With `GhostObjectManagerDummy` now correctly handling headless mode, this path should only be reached when `m_3DScene` is valid — but if `m_3DScene` is ever null in W3D mode (e.g., during startup or teardown edge cases), this will crash instead of failing silently.

The same change exists in `GeneralsMD/Code/GameEngineDevice/Source/W3DDevice/GameLogic/W3DGhostObject.cpp` line 160.

Consider whether the old silent-failure behaviour was masking a real bug, or whether the guard should be preserved as a `DEBUG_ASSERTCRASH` to catch regressions:</p>

```cpp
Bool W3DRenderObjectSnapshot::addToScene()
{
    DEBUG_ASSERTCRASH(W3DDisplay::m_3DScene != nullptr, ("W3DRenderObjectSnapshot::addToScene - m_3DScene is null"));
    if (!m_robj->Is_In_Scene())
    {
        W3DDisplay::m_3DScene->Add_Render_Object(m_robj);
        return true;
    }
    return false;
}
```

How can I resolve this? If you propose a fix, please make it concise.

_{Last reviewed commit: "style(headless): Exp..."}

[greptile-apps]

_{2 files reviewed, 2 comments}

_{Edit Code Review Agent Settings | Greptile}

[xezon]

has left over comments

[xezon]

OK