style: Convert leading spaces to tabs in Core and Dependencies by bobtista · Pull Request #2561 · TheSuperHackers/GeneralsGameCode

bobtista · 2026-04-09T19:08:30Z

Summary

Convert leading spaces/mixed whitespace to tabs, including access specifier indentation
All changes are whitespace-only (git diff -w is empty)
Uses tree-sitter-cpp for accurate C++ parsing with macro preprocessing

Access specifiers (public:/private:/protected:) are placed at the class
brace level, matching the existing codebase convention.

Script

The formatting script is included in the PR for reference but is not intended
to be merged — it shows how the changes were generated.

See scripts/cpp/convert_leading_spaces_to_tabs.py for details.

Part 1 of 4 — Core/ and Dependencies/ (499 files).
See also: #2562, #2563, #2564.

greptile-apps · 2026-04-09T19:12:48Z

Greptile Summary

This PR performs a whitespace-only conversion of leading spaces to tabs across 452 C++ files in Core/ and Dependencies/, using a tree-sitter-based Python script for accurate AST-guided reformatting. git diff -w is confirmed empty, meaning no logic was changed.

The main() function in the committed script still lists Dependencies/Utility in top_dirs, but the HEAD commit (a9cbcc22) explicitly reverted those 7 files because the script "cannot reliably handle" them. Re-running the script as shipped would re-apply incorrect indentation to those excluded headers.

Confidence Score: 4/5

Safe to merge for the C++ files; the script has one correctness issue that should be fixed before re-use.

All 452 C++ file changes are confirmed whitespace-only (git diff -w is empty). The single P1 finding is in the committed script: top_dirs still targets Dependencies/Utility even though the HEAD commit explicitly reverted those files as incompatible — re-running the script would re-corrupt them.

scripts/cpp/convert_leading_spaces_to_tabs.py — top_dirs needs Dependencies/Utility removed

Important Files Changed

Filename	Overview
scripts/cpp/convert_leading_spaces_to_tabs.py	New script documenting how whitespace was converted; `top_dirs` still includes `Dependencies/Utility` even though the HEAD commit reverted those files as incompatible with the script
Core/GameEngine/Source/GameNetwork/GameInfo.cpp	Whitespace-only: leading spaces converted to tabs; no logic changes
Core/GameEngineDevice/Source/W3DDevice/GameClient/W3DView.cpp	Whitespace-only: leading spaces converted to tabs, including previously mixed tab+space lines; no logic changes
Core/GameEngine/Source/Common/System/GameMemory.cpp	Whitespace-only: corrects mixed space+tab indentation to pure tabs; no logic changes
Core/Libraries/Source/WWVegas/WWLib/cpudetect.cpp	Whitespace-only: GNU `__asm__` call-style lines correctly converted; inline asm brace blocks preserved as-is
Core/GameEngine/Include/Common/INI.h	Whitespace-only: private member declarations and one public method corrected from spaces to tabs

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A[Read C++ file\ncp1252 encoding] --> B[Preprocess: expand\nconfusing macros]
    B --> C[tree-sitter parse\nC++ AST]
    C --> D{Too many\nparse errors?}
    D -->|Yes| E[Skip file]
    D -->|No| F[Process each line]
    F --> G{In block comment\nor ASM block?}
    G -->|Yes| H[Preserve as-is]
    G -->|No| I{Preprocessor\ndirective?}
    I -->|Yes| J[Preserve, track pp_depth]
    I -->|No| K{Already tabs /\ncontinuation line?}
    K -->|Yes| H
    K -->|No| L[Query AST node\nat line start col]
    L --> M[get_indent_depth:\ncount scope ancestors]
    M --> N{depth==0 and\nspaces>=4 or pp_depth>0?}
    N -->|Yes - sanity skip| H
    N -->|No| O[Replace leading\nwhitespace with N tabs]
    O --> P[Write converted\nfile back]

Prompt To Fix All With AI

This is a comment left during a code review.
Path: scripts/cpp/convert_leading_spaces_to_tabs.py
Line: 443-454

Comment:
**`top_dirs` still includes `Dependencies/Utility`**

The HEAD commit (`a9cbcc22`) explicitly reverted the seven `Dependencies/Utility/Utility/*.h` files because the script "cannot reliably handle the interaction between preprocessor nesting and C++ nesting in these files." However, `top_dirs` still lists that directory, so re-running the script as committed would re-process — and likely corrupt — those same files.

```suggestion
        top_dirs = [
            os.path.join(root_dir, 'Core'),
            os.path.join(root_dir, 'Generals'),
            os.path.join(root_dir, 'GeneralsMD'),
        ]
```

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

_{Reviews (5): Last reviewed commit: "style: Exclude Dependencies/Utility comp..." | Re-trigger Greptile}

xezon

By what rules are these whitespaces edited? How can it be reviewed?

xezon · 2026-04-10T19:55:15Z

+			//using the IMC_SETCOMPOSITIONWINDOW message.
+			//
+			//I'm not sure what to do here.
+			m_result = 1;


This does not look right.

Still look wrong

bobtista · 2026-04-11T18:00:10Z

By what rules are these whitespaces edited? How can it be reviewed?

Leading whitespace is regenerated from the C++ AST, not counted from spaces. For more details on how it's done, check scripts/cpp/convert_leading_spaces_to_tabs.py it will:

Parse the file with tree-sitter-cpp
For any line whose leading whitespace contains a space, locate the AST node at the first non-whitespace column.
Walk up the AST counting "scope-creating" ancestors: compound_statement, field_declaration_list, declaration_list (namespace body), enumerator_list, initializer_list, case_statement. Add one level for the braceless body/consequence/alternative of if/while/for/do/else. A compound_statement directly under a case_statement does not add a level (the case already provides it).
Replace the leading whitespace with exactly that many tab characters.

It skips a bunch of things eg block comments, and things that are addressed in later PRs to make review easier. To review it, I'd make sure you like the approach in the script, then spot check the results.

xezon

There are still some indentation errors.

xezon · 2026-04-12T08:05:19Z


 	const Coord3D *pos = obj->getPosition();
-  Real angle = obj->getOrientation();
+		Real angle = obj->getOrientation();


1 indent too much

xezon · 2026-04-12T08:05:43Z

- 			Int cy = REAL_TO_INT_FLOOR((y + 0.5f)/PATHFIND_CELL_SIZE_F);
- 			if (cx >= 0 && cy >= 0 && cx < m_extent.hi.x && cy < m_extent.hi.y)
- 			{
+		for (Int iy = 0; iy < numStepsY; ++iy, tl_x += ydx, tl_y += ydy)


1 indent too much

Many times

xezon · 2026-04-12T08:10:30Z

+		}
+	}
+
+	return ;


Side note: I thought we had all the superfluous returns already removed.

looks like the space between return and semicolon was never considered.

This return is at the end of the function and can be removed. We already had such a change previously, but perhaps it missed W3DView.

I'm aware of that. I was merely making a note that the space between return and the semicolon, which the original script/regexp may not have considered.

The original script is only focused on leading spaces. Removing the space between the return and semicolon would be a separate PR, as would removing the return

Ah I see it was started but never finished: #1811

I can pick that up.

xezon · 2026-04-12T08:11:59Z

@@ -525,8 +525,8 @@ BoneMgrDialogClass::Remove_Object_From_Bone
 {
 	// Loop through all the children of this bone
 	for (HTREEITEM hchild_item = m_BoneTree.GetChildItem (bone_item);
-		  (hchild_item != nullptr);
-		  hchild_item = m_BoneTree.GetNextSiblingItem (hchild_item)) {
+	(hchild_item != nullptr);


Can we put one more indent after for, while, if, else if?

Skyaero42

Sampled a few files. Comments are recurring throughout the PR.

Skyaero42 · 2026-04-12T12:50:22Z

 	Bool m_use3DSoundRangeVolumeFade; // TheSuperHackers @feature Enables 3D sound range volume fade as originally intended
 	Real m_3DSoundRangeVolumeFadeExponent; // TheSuperHackers @feature Sets 3D sound range volume fade exponent for non-linear fade.
-	                                       // The higher the exponent, the sharper the decline at the max range.
+	// The higher the exponent, the sharper the decline at the max range.


Incorrect indentation, now it looks like this comment belongs to the next line, rather than the previous

Skyaero42 · 2026-04-12T12:51:45Z

+			mov esi,[buf]
+			mov ecx,[len]
+			dec ecx
      mov edi,[crcPtr]


Incorrect indentation. Multiple times. This function is now a mess

Skyaero42 · 2026-04-12T12:52:33Z

-																																			  *  memory pointed at by buffer. Returns the number of bytes read.
-																																			  *  Returns -1 if an error occurred.
-																																			  */
+	*  memory pointed at by buffer. Returns the number of bytes read.


Documentation block comment should be placed before function - multiple times in this is file

Skyaero42 · 2026-04-12T12:53:12Z

 	Int winActivate();  ///< pop window to top of list and activate
 	Int winBringToTop();  ///< bring this window to the top of the win list
 	Int winEnable( Bool enable );  /**< enable/disable a window, a disbled
-																 window can be seen but accepts no input */


Place documentation comment before function

Skyaero42 · 2026-04-12T12:57:05Z

 	public:

- 		virtual	VideoStreamInterface* next() override;				///< Returns next open stream
+	virtual	VideoStreamInterface* next() override;				///< Returns next open stream


incorrect indentation

DevGeniusCode

I manually reviewed some of the files (started from the bottom of the list and worked my way up). I left comments on the main indentation issues I noticed.

DevGeniusCode · 2026-04-12T17:58:06Z

    #if __has_builtin(__builtin_return_address)
    static inline uintptr_t _ReturnAddress()
    {
-        return reinterpret_cast<uintptr_t>(__builtin_return_address(0));
+	return reinterpret_cast<uintptr_t>(__builtin_return_address(0));
    }
    #else
    #error "No implementation for _ReturnAddress"


The script seems to have flattened the indentation within this preprocessor block. The function body should be indented further than the function signature.

DevGeniusCode · 2026-04-12T17:59:52Z

+		}
+
+		return wsOut;
 }

 inline char* WINAPI ConvertBSTRToString(BSTR pSrc)
 {
-    DWORD cb, cwch;
-    char *szOut = nullptr;
-
-    if (!pSrc)
-        return nullptr;
-
-    // Retrieve the size of the BSTR with the null terminator
-    cwch = SysStringLen(pSrc) + 1;
-
-    // Compute the needed size with the null terminator
-    cb = WideCharToMultiByte(CP_ACP, 0, pSrc, cwch, nullptr, 0, nullptr, nullptr);
-    if (cb == 0)
-    {
-        cwch = GetLastError();
-        _com_issue_error(!IS_ERROR(cwch) ? HRESULT_FROM_WIN32(cwch) : cwch);
-        return nullptr;
-    }
-
-    // Allocate the string
-    szOut = (char*)::operator new(cb * sizeof(char));
-    if (!szOut)
-    {
-        _com_issue_error(HRESULT_FROM_WIN32(ERROR_OUTOFMEMORY));
-        return nullptr;
-    }
-
-    // Convert the string and null-terminate
-    szOut[cb - 1] = '\0';
-    if (WideCharToMultiByte(CP_ACP, 0, pSrc, cwch, szOut, cb, nullptr, nullptr) == 0)
-    {
-        // We failed, clean everything up
-        cwch = GetLastError();
-
-        ::operator delete(szOut);
-        szOut = nullptr;
-
-        _com_issue_error(!IS_ERROR(cwch) ? HRESULT_FROM_WIN32(cwch) : cwch);
-    }
-
-    return szOut;
+		DWORD cb, cwch;
+		char *szOut = nullptr;
+
+		if (!pSrc)
+			return nullptr;


The script seems to be adding unnecessary padding here. The function body is misaligned, and the indentation level is no longer consistent with the rest of the code.

DevGeniusCode · 2026-04-12T18:02:21Z

 inline void Matrix3::Get_Z_Vector(Vector3 * set) const
 {
-   set->Set(Row[0][2], Row[1][2], Row[2][2]);
+	set->Set(Row[0][2], Row[1][2], Row[2][2]);
 }


While the script works correctly in simple cases like this one, it remains inconsistent overall, leading to the broken indentation seen in more complex blocks (like preprocessor directives).

DevGeniusCode · 2026-04-12T18:50:43Z

            (m_stringBackgroundBMP.CompareNoCase (pszBackgroundBMP) != 0))
-        {
+		{
 				// Create a new instance of the BMP object to use


DevGeniusCode · 2026-04-12T18:50:50Z

+		if (m_pCBackgroundBMP)
+		{
+			// Remove the background BMP from the scene
+			// and release its pointer
 				m_pCBackgroundBMP->Remove ();


DevGeniusCode · 2026-04-12T18:50:56Z

-    {
+	ASSERT (m_pC2DScene);
+	if (m_pC2DScene)
+	{
 		  // Remove the old background BMP if there was one.


DevGeniusCode · 2026-04-12T18:51:22Z

+				if ((use_global_reset_flag && m_bAutoCameraReset) ||
 					     ((use_global_reset_flag == false) && allow_reset) ||
 						  m_bOneTimeReset) {
 						pCGraphicView->Reset_Camera_To_Display_Object (*m_pCRenderObj);
 						m_bOneTimeReset = false;
 					 }


DevGeniusCode · 2026-04-12T18:52:05Z

-            // Update the animation frame
+		if (m_pCRenderObj)
+		{
+			// Update the animation frame
 			  for (int i = 0; i < m_pCAnimCombo->Get_Num_Anims(); i++)


greptile-apps · 2026-04-12T20:05:07Z

+        top_dirs = [
+            os.path.join(root_dir, 'Core'),
+            os.path.join(root_dir, 'Generals'),
+            os.path.join(root_dir, 'GeneralsMD'),
+            os.path.join(root_dir, 'Dependencies', 'Utility'),
+        ]
+        file_list = []
+        for ext in ['*.cpp', '*.h', '*.inl']:
+            for top_dir in top_dirs:
+                file_list.extend(
+                    glob.glob(os.path.join(top_dir, '**', ext), recursive=True)
+                )


top_dirs still includes Dependencies/Utility

The HEAD commit (a9cbcc22) explicitly reverted the seven Dependencies/Utility/Utility/*.h files because the script "cannot reliably handle the interaction between preprocessor nesting and C++ nesting in these files." However, top_dirs still lists that directory, so re-running the script as committed would re-process — and likely corrupt — those same files.

Suggested change

top_dirs = [

os.path.join(root_dir, 'Core'),

os.path.join(root_dir, 'Generals'),

os.path.join(root_dir, 'GeneralsMD'),

os.path.join(root_dir, 'Dependencies', 'Utility'),

]

file_list = []

for ext in ['*.cpp', '*.h', '*.inl']:

for top_dir in top_dirs:

file_list.extend(

glob.glob(os.path.join(top_dir, '**', ext), recursive=True)

)

top_dirs = [

os.path.join(root_dir, 'Core'),

os.path.join(root_dir, 'Generals'),

os.path.join(root_dir, 'GeneralsMD'),

]

Prompt To Fix With AI

This is a comment left during a code review. Path: scripts/cpp/convert_leading_spaces_to_tabs.py Line: 443-454 Comment: **`top_dirs` still includes `Dependencies/Utility`** The HEAD commit (`a9cbcc22`) explicitly reverted the seven `Dependencies/Utility/Utility/*.h` files because the script "cannot reliably handle the interaction between preprocessor nesting and C++ nesting in these files." However, `top_dirs` still lists that directory, so re-running the script as committed would re-process — and likely corrupt — those same files. ```suggestion top_dirs = [ os.path.join(root_dir, 'Core'), os.path.join(root_dir, 'Generals'), os.path.join(root_dir, 'GeneralsMD'), ] ``` How can I resolve this? If you propose a fix, please make it concise.

DevGeniusCode · 2026-04-14T12:22:06Z

   #ifdef _UNIX
-      pos=fseek(Handle, pos, dir);
+	pos=fseek(Handle, pos, dir);
   #else


Space instead of tab, and missing tab

DevGeniusCode · 2026-04-14T12:22:34Z

@@ -735,7 +735,7 @@ int RawFileClass::Write(void const * buffer, int size)
 		opened = true;
 	}

-   int writeok=TRUE;
+	int writeok=TRUE;
   #ifdef _UNIX


same before

DevGeniusCode · 2026-04-14T12:25:00Z

 (
 	0.0,	1.0,	0.0,	0.0,
  -1.0,	0.0,	0.0,	0.0,
- 	0.0,	0.0,	1.0,	0.0
+	0.0,	0.0,	1.0,	0.0
 );


space instated of tab in lines started with -1

DevGeniusCode · 2026-04-14T12:26:41Z


-  if (c->NumTimeCodes > 1) {
-  	if (c->NumTimeCodes > 2)  {
+	if (c->NumTimeCodes > 1) {
+	if (c->NumTimeCodes > 2)  {

-  		float32 *pVecSrc, *pVecDst, *pVecOrg;
-      uint32	*pTcSrc,  *pTcDst,  *pTcOrg;
+		float32 *pVecSrc, *pVecDst, *pVecOrg;
+			uint32	*pTcSrc,  *pTcDst,  *pTcOrg;

 		for(uint32 try_idx = 0; try_idx < (c->NumTimeCodes - 2); try_idx++)  {



not look good

DevGeniusCode · 2026-04-14T12:27:26Z

@@ -1335,7 +1335,7 @@ uint32 VectorChannelClass::find_least_useful_packetQ(W3dTimeCodedAnimChannelStru
 				continue;	// can't get rid of this guy
 	      }
 	      if (*pTcDst & W3D_TIMECODED_BINARY_MOVEMENT_FLAG)  {
-	      	continue;	// can't get rid of this guy either
+		continue;	// can't get rid of this guy either
 	      }



not look good, and spaces instead of tabs

DevGeniusCode · 2026-04-14T12:39:17Z

- 		unsigned long P_Encrypt[ROUNDS+2];
- 		unsigned long P_Decrypt[ROUNDS+2];
+		unsigned long P_Encrypt[ROUNDS+2];
+		unsigned long P_Decrypt[ROUNDS+2];


double tab?

DevGeniusCode · 2026-04-14T12:40:07Z

+	xlEdgeBottom = 9,
+	xlEdgeLeft = 7,
+	xlEdgeRight = 10,
+	xlEdgeTop = 8
 } XlBordersIndex;

 enum {


this enum have double tab unlike the other enum

DevGeniusCode · 2026-04-14T12:41:25Z

@@ -124,7 +124,7 @@ class StandardFileClass
 		int 	Query_Size			( void );
 		bool 	Query_Open			( void );
 		char *	Query_Name_String	( void );
-     	int		End_Of_File			( void );
+	int		End_Of_File			( void );


wording missing tab?

DevGeniusCode · 2026-04-14T12:42:53Z

@@ -984,26 +984,26 @@ void StandardFileClass::Reset( void )
 int StandardFileClass::End_Of_File	( void )
 {
 	#if( SUPPORT_HANDLES )
-   	return( TRUE );
+	return( TRUE );


missing tab

DevGeniusCode · 2026-04-14T12:44:39Z


-    // Return a pointer to the prototype
+	// Return a pointer to the prototype
 	 return pprototype;


greptile-apps · 2026-04-15T16:40:07Z

Too many files changed for review. (538 files found, 500 file limit)

… to tabs

bobtista · 2026-04-15T17:06:52Z

How do we feel about using this tool + script to make some improvements and merge things - then to use clang-format for the files where it would end up being a game of whackamole with regex fixes? I think these manual tweaks are error prone, and we have tons of random tabs/spaces that are a bit of a mess eg matrix3d.cpp that it would probably be better to use a proper formatter on vs tweaking scripts and doing things by hand

This was referenced Apr 9, 2026

style: Convert continuation lines to tabs in GeneralsMD #2572

Open

style: Convert continuation lines to tabs in Generals #2571

Open

xezon reviewed Apr 10, 2026

View reviewed changes

xezon reviewed Apr 12, 2026

View reviewed changes

Skyaero42 reviewed Apr 12, 2026

View reviewed changes

DevGeniusCode self-requested a review April 12, 2026 17:29

DevGeniusCode requested changes Apr 12, 2026

View reviewed changes

greptile-apps bot reviewed Apr 12, 2026

View reviewed changes

DevGeniusCode requested changes Apr 14, 2026

View reviewed changes

bobtista added 2 commits April 15, 2026 12:45

feat(scripts): Add tree-sitter based script to convert leading spaces…

8f17eb8

… to tabs

style: Convert leading spaces to tabs in Core and Dependencies

2718314

bobtista force-pushed the bobtista/feat/access-spec-indent-core branch from 5e923ff to 2718314 Compare April 15, 2026 16:48

Conversation

bobtista commented Apr 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Script

Uh oh!

greptile-apps bot commented Apr 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Greptile Summary

Confidence Score: 4/5

Important Files Changed

Flowchart

Uh oh!

xezon left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

bobtista commented Apr 11, 2026

Uh oh!

xezon left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Skyaero42 left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

DevGeniusCode left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

bobtista commented Apr 9, 2026 •

edited

Loading

greptile-apps bot commented Apr 9, 2026 •

edited

Loading

greptile-apps bot commented Apr 15, 2026 •

edited

Loading