feat(options): Implement game options for texture filter mode, anisotropy and MSAA selection

[Mauller]

Merge By Rebase

Merge after #2629

Closes: #2474
Closes: #2375

This PR implements options for allowing the enablement of MSAA and the selection of different texture filtering modes and setting the anisotropic filtering level.

The following options are added for anti aliasing

AntiAliasing = 0, 2, 4, 8
For now the numbers relate to the MSAA sample level and Zero disables MSAA as we only support MSAA at the moment.

TextureFilter = None, Point, Bilinear, Trilinear, Anisotropic
Self explanatory, just choose one and see how the game looks

AnisotropyLevel = 2, 4, 8, 16
The anisotropy level only comes into play when Anisotropic filtering is set. 2 is also the minimum value.

---

Generals is broken untill the changes are replicated due to changes in texture filter init.

• Replicate in generals

[greptile-apps]

Greptile Summary

This PR adds user-configurable MSAA (0/2/4/8x), texture filter mode (None/Point/Bilinear/Trilinear/Anisotropic), and anisotropy level (2/4/8/16x) to the options system for both Generals and Zero Hour. The implementation introduces highestBit() to round raw sample/anisotropy values to the nearest power of two, stores settings in GlobalData, and applies them after successful render device creation in W3DDisplay. Previous review concerns (incorrect > vs <= condition, stray semicolon, copy-paste bug in string parsing, and unguarded filter init after device failure) appear addressed in the current HEAD.

Confidence Score: 4/5

Safe to merge after addressing the fallback inconsistency and the header layering concern; no crash-level regressions found in the current diff.

All previously flagged P1/P0 issues are resolved. The remaining findings are: a silent quality regression when a TextureFilter INI value is misspelled (falls back to NONE rather than BILINEAR), a rendering-header inclusion in a widely-shared core header (OptionPreferences.h), and minor comment alignment. None are blocking crashes, but the fallback inconsistency could surprise users editing their config manually.

Core/GameEngine/Include/Common/OptionPreferences.h (header layering) and Core/Libraries/Source/WWVegas/WW3D2/texturefilter.cpp (fallback return value).

Important Files Changed

Filename	Overview
Core/GameEngine/Include/Common/OptionPreferences.h	Adds AntiAliasingMode enum and three new getter declarations; now includes rendering-layer headers (ww3d.h, texturefilter.h) in a widely-included core common header, creating a layering concern.
Core/Libraries/Source/WWVegas/WW3D2/texturefilter.cpp	Adds TextureFilterModeString lookup table and getTextureFilterMode() helper (falls back to TEXTURE_FILTER_NONE on mismatch, inconsistent with the BILINEAR default used elsewhere), and updates _Init_Filters signature to accept anisotropy level.
Core/Libraries/Include/Lib/BaseType.h	Adds highestBit() template that rounds a value down to the nearest power of two; correct for both signed and unsigned inputs with appropriate static_assert guard.
Generals/Code/GameEngine/Include/Common/GlobalData.h	Replaces m_antiAliasBoxValue with m_antiAliasLevel and adds m_textureFilteringMode and m_textureAnisotropyLevel; comment alignment across the three fields is inconsistent.
Generals/Code/GameEngineDevice/Source/W3DDevice/GameClient/W3DDisplay.cpp	Sets MSAA mode from global data before render device creation, then applies texture filter and anisotropy settings only on WW3D_ERROR_OK — correctly guarded against failed device creation.
Generals/Code/Libraries/Source/WWVegas/WW3D2/ww3d.cpp	Adds AnisotropyLevel static, Set_Anisotropy_Level() using clamp+highestBit, and updates Set_Texture_Filter() to pass current AnisotropyLevel to _Init_Filters; logic and clamping are correct.

Sequence Diagram

sequenceDiagram
    participant INI as Options.ini
    participant OP as OptionPreferences
    participant GD as GlobalData
    participant W3DD as W3DDisplay::init()
    participant WW3D as WW3D
    participant TF as TextureFilterClass

    INI->>OP: load("Options.ini")
    OP->>GD: parseGameDataDefinition()

    W3DD->>WW3D: Set_MSAA_Mode(m_antiAliasLevel)
    W3DD->>WW3D: Set_Render_Device(...)
    WW3D->>TF: _Init_Filters(TextureFilter, AnisotropyLevel)

    alt renderDeviceError == WW3D_ERROR_OK
        W3DD->>WW3D: Set_Texture_Filter(m_textureFilteringMode)
        WW3D->>TF: _Init_Filters(TextureFilter, AnisotropyLevel)
        W3DD->>WW3D: Set_Anisotropy_Level(m_textureAnisotropyLevel)
        WW3D->>TF: _Set_Max_Anisotropy(AnisotropyLevel)
    end

Loading

Prompt To Fix All With AI

This is a comment left during a code review.
Path: Core/Libraries/Source/WWVegas/WW3D2/texturefilter.cpp
Line: 148

Comment:
**Invalid string falls back to `TEXTURE_FILTER_NONE`, not `TEXTURE_FILTER_BILINEAR`**

When `getTextureFilterMode()` receives an unrecognized string (e.g. a typo like `"Billinear"`), it returns `TEXTURE_FILTER_NONE`, but `OptionPreferences::getTextureFilterMode()` returns `TEXTURE_FILTER_BILINEAR` when the `TextureFilter` key is entirely absent. This means a misconfigured INI key silently degrades the user to *no* texture filtering — worse than the intended default — rather than falling back gracefully to bilinear.

Consider returning the intended default (`TEXTURE_FILTER_BILINEAR`) from the static helper when no match is found:

```cpp
	return TextureFilterClass::TEXTURE_FILTER_BILINEAR;
```

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

---

This is a comment left during a code review.
Path: Core/GameEngine/Include/Common/OptionPreferences.h
Line: 9-10

Comment:
**Rendering-layer headers pulled into a core common header**

Including `ww3d.h` and `texturefilter.h` directly in `OptionPreferences.h` creates a hard dependency from the core game engine common layer on the rendering library. Any translation unit that transitively includes `OptionPreferences.h` (which is widely used) will now also pull in all rendering headers and their transitive dependencies, which may cause compilation issues in translation units that don't have the WW3D include paths configured, and makes the core layer non-portable.

Consider forward-declaring the return types (or using `Int`/`UnsignedInt` as the public API and casting internally inside the `.cpp` file where the rendering headers are already included), keeping `OptionPreferences.h` free of graphics-layer includes.

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

---

This is a comment left during a code review.
Path: Generals/Code/GameEngine/Include/Common/GlobalData.h
Line: 400-402

Comment:
**Comment columns not aligned across the three new fields**

The `///<` comment delimiters on the three new fields are not column-aligned: `m_antiAliasLevel` has 10 spaces before the comment, `m_textureFilteringMode` has 7, and `m_textureAnisotropyLevel` has 5. Per the codebase's data-block alignment convention these should be aligned to the same column. The same pattern applies to the parallel block in `GeneralsMD/Code/GameEngine/Include/Common/GlobalData.h`.

```cpp
	UnsignedInt m_antiAliasLevel;             ///< value of selected antialias level in the game options
	UnsignedInt m_textureFilteringMode;        ///< value related to TextureFilterClass::TextureFilterModeEnum
	UnsignedInt m_textureAnisotropyLevel;      ///< value related to TextureFilterClass::AnisotropicFilterMode
```

**Rule Used:** Align columns in data-oriented blocks (like lookup... ([source](https://app.greptile.com/review/custom-context?memory=e61e83cf-a14f-4759-bed8-3bab008b25b5))

**Learned From**
[TheSuperHackers/GeneralsGameCode#2067](https://github.com/TheSuperHackers/GeneralsGameCode/pull/2067#discussion_r2704192059)

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

_{Reviews (15): Last reviewed commit: "feat(options): Implement game options fo..." | Re-trigger Greptile}

[Mauller]

Generals is currently broken till i replicate across due to changes in initaialising the texture filtering

[xezon]

This implementation can be polished more.

[Mauller]

Updated based on feedback

[xezon]

There is a lot of casting from enum type to int and back to enum type. Why is that like so? Can we not stick to enum types for all these integers? And if not, then what is the point of the enum types if we need to use integers?

[xezon]

Can we make m_antiAliasLevel that enum type?

[Mauller]

Tweaked based on recent Feedback

[xezon]

This review appears to be outdated.

[Mauller]

Tweaked based on majority of feedback

[Mauller]

Updated based on feedback, just need to fix merge conflicts

[Mauller]

Rebased with main, but main changes were based on last two review comments in

TextureFilterClass::getTextureFilterMode(const char* str)
and
WW3D::Set_Anisotropy_level(int level)

[Mauller]

Had to add an extra fix commit to get the builds working in release again, no idea why it would now alert about the unused variable, unless the error was originaly being supressed.

[Mauller]

Updated, if good to go i will replicate across to generals.

I also re-ordered the commits so the one fixing the unused variable is first.

[xezon]

Can the first commit be a separate pull then? It looks to be not related to game options.

[Mauller]

Can the first commit be a separate pull then? It looks to be not related to game options.

It's not related but was required to allow the PR to build in release. I seperated it out into #2629 now. This PR will start failing to build in zero hour release again once i update it.

[Mauller]

Removed build fix commit, PR will currently fail to build zero hour release builds untill the build fix is merged to main and this PR is rebased.

[greptile-apps]

Texture filter init called inside retry loop when device may not be ready

WW3D::Set_Texture_Filter (and Set_Anisotropy_Level) are called unconditionally after Set_Render_Device, even if it returned an error. Set_Texture_Filter internally calls TextureFilterClass::_Init_Filters, which immediately dereferences DX8Wrapper::Get_Current_Caps()->Get_DX8_Caps(). If device creation failed on the first retry attempt, the caps pointer may be null or stale, leading to a crash or undefined behaviour before the second retry has a chance to run.

Guard these calls so they only execute on a successful device creation:

			TheWritableGlobalData->m_antiAliasLevel = (UnsignedInt)WW3D::Get_MSAA_Mode();
			if (renderDeviceError == WW3D_ERROR_OK)
			{
				WW3D::Set_Texture_Filter(TheWritableGlobalData->m_textureFilteringMode);
				TheWritableGlobalData->m_textureFilteringMode = WW3D::Get_Texture_Filter();
				WW3D::Set_Anisotropy_Level(TheWritableGlobalData->m_textureAnisotropyLevel);
				TheWritableGlobalData->m_textureAnisotropyLevel = WW3D::Get_Anisotropy_Level();
			}

Prompt To Fix With AI

This is a comment left during a code review.
Path: GeneralsMD/Code/GameEngineDevice/Source/W3DDevice/GameClient/W3DDisplay.cpp
Line: 780-786

Comment:
**Texture filter init called inside retry loop when device may not be ready**

`WW3D::Set_Texture_Filter` (and `Set_Anisotropy_Level`) are called unconditionally after `Set_Render_Device`, even if it returned an error. `Set_Texture_Filter` internally calls `TextureFilterClass::_Init_Filters`, which immediately dereferences `DX8Wrapper::Get_Current_Caps()->Get_DX8_Caps()`. If device creation failed on the first retry attempt, the caps pointer may be null or stale, leading to a crash or undefined behaviour before the second retry has a chance to run.

Guard these calls so they only execute on a successful device creation:
```cpp
			TheWritableGlobalData->m_antiAliasLevel = (UnsignedInt)WW3D::Get_MSAA_Mode();
			if (renderDeviceError == WW3D_ERROR_OK)
			{
				WW3D::Set_Texture_Filter(TheWritableGlobalData->m_textureFilteringMode);
				TheWritableGlobalData->m_textureFilteringMode = WW3D::Get_Texture_Filter();
				WW3D::Set_Anisotropy_Level(TheWritableGlobalData->m_textureAnisotropyLevel);
				TheWritableGlobalData->m_textureAnisotropyLevel = WW3D::Get_Anisotropy_Level();
			}
```

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

[Mauller]

Rebased on main and fixed initialisation issue brought up by greptile

If everything is good i will replicate to generals.

[Mauller]

Replicated to generals

[xezon]

Greptile shows 2 P2 findings.

[Mauller]

Greptile shows 2 P2 findings.

The (val>=0) one is a guard for future use that replicates other settings blocks that already existed in the game.
Mostly for when sliders or menus are giong to be used to set the options etc.

The Set_Texture_Filter one is not something i touched in this PR. But i can still tweak it anyway.

[Mauller]

tweaked to use a clamp instead of the two if statements, this should be good to go now.