Initial commit for macOS support

.github/workflows/main.yml

@@ -20,7 +20,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        name: [windows-latest, ubuntu-24.04]
+        name: [windows-latest, ubuntu-24.04, macos-latest]
         include:
           - name: windows-latest
             os: windows-latest
@@ -34,6 +34,13 @@ jobs:
             physx: linux
             physxkey: linux
             gpulang: linux
+            anyfx: anyfxcompiler-linux
+          - name: macos-latest
+            os: macos-latest
+            config: darwin-ninja-debug
+            physx: linux
+            physxkey: linux
+            anyfx: anyfxcompiler-macos
 
       # Steps represent a sequence of tasks that will be executed as part of the job
     steps:

code/foundation/CMakeLists.txt

@@ -2,10 +2,14 @@
 # Foundation
 #-------------------------------------------------------------------------------
 
-if(FIPS_LINUX OR FIPS_MACOS)
+if(FIPS_LINUX OR FIPS_OSX)
     FIND_PACKAGE(CURL)
 endif()
 
+if(FIPS_OSX)
+find_library(COCOA_LIBRARY Cocoa)
+endif()
+
 if (EXISTS "${NROOT}/syswork/export.zip")
     create_resource("${NROOT}/syswork/" "${NROOT}/syswork/export.zip" system_resources)
     set_target_properties(system_resources PROPERTIES FOLDER Resources)
@@ -18,6 +22,9 @@ nebula_begin_module(foundation)
     if(FIPS_LINUX)
         fips_libs(uuid stdc++fs)
     endif()
+    if (FIPS_OSX)
+        fips_libs(${COCOA_LIBRARY})
+    endif()
     target_include_directories(foundation PUBLIC ${CMAKE_CURRENT_SOURCE_DIR})
     fips_deps(nebula-libs zlib tinyxml nflatbuffer)
     if(TARGET system_resources)
@@ -525,12 +532,12 @@ nebula_begin_module(foundation)
             jobs2.cc
             jobs2.h
         )
-        fips_dir(fibers)
-        fips_files(
-            fiber.h
-            fibers.cc
-            fibers.h
-        )
+        #fips_dir(fibers)
+        #fips_files(
+        #    fiber.h
+        #    fibers.cc
+        #    fibers.h
+        #)
         fips_dir(framesync)
         fips_files(
             framesynctimer.cc
@@ -611,10 +618,10 @@ nebula_begin_module(foundation)
                 win32/win32systeminfo.cc
                 win32/win32systeminfo.h
             )
-        fips_dir(fibers GROUP "fibers/win32")
-            fips_files(
-                win32/win32fiber.cc
-            )
+        #fips_dir(fibers GROUP "fibers/win32")
+        #    fips_files(
+        #        win32/win32fiber.cc
+        #    )
         fips_dir(.)
     endif()
     if (FIPS_POSIX)
@@ -623,7 +630,7 @@ nebula_begin_module(foundation)
             core/posix/posixsingleton.h
             core/posix/posixsysfunc.cc
             core/posix/posixsysfunc.h
-            fibers/posix/posixfiber.cc
+            #fibers/posix/posixfiber.cc
             memory/posix/posixmemory.h
             memory/posix/posixheap.cc
             memory/posix/posixheap.h
@@ -688,6 +695,18 @@ nebula_begin_module(foundation)
         )
         fips_dir(threading/gcc GROUP "threading/gcc")
         fips_files(gccinterlocked.cc)
+    elseif(FIPS_MACOS)
+        fips_dir(threading/posix GROUP "threading/posix")
+            fips_files(
+                posixbarrier.h
+                posixcriticalsection.h
+                posixevent.h
+
+                posixthread.cc
+                posixthread.h
+            )
+        fips_dir(threading/gcc GROUP "threading/gcc")
+        fips_files(gccinterlocked.cc)
     else()
         fips_dir(threading/posix GROUP "threading/posix")
         fips_files(

code/foundation/core/posix/posixsysfunc.cc

@@ -43,15 +43,11 @@ SysFunc::Setup()
     if (!SetupCalled)
     {
          SetupCalled = true;
-        #if !__MAYA__
         //Threading::Thread::SetMyThreadName("MainThread");
-        #endif
         Memory::SetupHeaps();
         Memory::Heap::Setup();
         Blob::Setup();
-        #if !__MAYA__
         Net::Socket::InitNetwork();
-        #endif   
 
         globalStringAtomTable = new Util::GlobalStringAtomTable;
         #if NEBULA_ENABLE_THREADLOCAL_STRINGATOM_TABLES

code/foundation/core/posix/precompiled.h

@@ -3,7 +3,7 @@
 #define CORE_POSIX_PRECOMPILED_H
 //------------------------------------------------------------------------------
 /**
-    @file core/win32/precompiled.h
+    @file core/posix/precompiled.h
 
     Contains precompiled headers on the Posix platform.
 
@@ -28,12 +28,16 @@
 #include <limits.h>
 
 // sse intrinsics
+#ifdef __linux__
 #if __SSE2__ || __SSE4_1__ || __AVX2__
 #include <emmintrin.h>
 #include <immintrin.h>
 #elif __aarc64__
 #include <arm_neon.h>
 #endif
+#elif __APPLE__
+#include <sse2neon.h>
+#endif
 
 //------------------------------------------------------------------------------
 #endif

code/foundation/core/singleton.h

@@ -11,9 +11,7 @@
 */
 #if __WIN32__
 #include "core/win32/win32singleton.h"
-#elif __OSX__
-#include "core/osx/osxsingleton.h"
-#elif __linux__
+#elif __linux__ || __APPLE__
 #include "core/posix/posixsingleton.h"
 #else
 #error "IMPLEMENT ME!"

code/foundation/core/types.h

@@ -34,6 +34,7 @@ typedef unsigned int   uint;
 typedef unsigned short ushort;
 typedef unsigned char  uchar;
 typedef unsigned char  ubyte;
+typedef signed char byte;
 
 typedef uintptr_t uintptr;
 typedef ptrdiff_t ptrdiff;
@@ -109,7 +110,7 @@ OnlyBits(const FLAGS flags, const BITS bits)
 #define BITS_TO_BYTES(x) (((x)+7)>>3)
 #define BYTES_TO_BITS(x) ((x)<<3)
 
-#if (__OSX__ || __linux__)
+#if (__APPLE__ || __linux__)
 inline ushort                _byteswap_ushort(ushort x)              { return ((x>>8) | (x<<8)); }
 inline ulong                 _byteswap_ulong(ulong x)                { return ((x&0xff000000)>>24) | ((x&0x00ff0000)>>8) | ((x&0x00000ff00)<<8) | ((x&0x000000ff)<<24); }
 inline unsigned long long    _byteswap_uint64(unsigned long long x)  { return ((((unsigned long long)_byteswap_ulong((ulong)(x & 0xffffffff))) << 32) | ((unsigned long long)_byteswap_ulong((ulong)(x >> 32)))); }
@@ -121,7 +122,7 @@ typedef unsigned char byte;
 #if __WIN32__
 #define n_stricmp stricmp
 #define n_snprintf StringCchPrintf
-#elif (__OSX__ || __APPLE__ || __linux__ )
+#elif __APPLE__ || __linux__
 #define n_stricmp strcasecmp
 #define n_snprintf sprintf
 #else
@@ -132,7 +133,7 @@ typedef unsigned char byte;
 #define ThreadLocal __declspec(thread)
 #elif __linux__
 #define ThreadLocal __thread
-#elif (__OSX__ || __APPLE__)
+#elif __APPLE__
 #define ThreadLocal thread_local
 #else
 #error "Unsupported platform!"

code/foundation/io/archfs/archive.cc

@@ -8,7 +8,7 @@
 
 namespace IO
 {
-#if __WIN32__ || __linux__
+#if __WIN32__ || __linux__ || __APPLE__
 __ImplementClass(IO::Archive, 'ARCV', IO::ZipArchive);
 #else
 #error "IO::Archive not implemented on this platform!"

code/foundation/io/archfs/archive.h

@@ -9,7 +9,7 @@
     (C) 2009 Radon Labs GmbH
     (C) 2013-2020 Individual contributors, see AUTHORS file
 */    
-#if __WIN32__ || __linux__
+#if __WIN32__ || __linux__ || __APPLE__
 #include "io/zipfs/ziparchive.h"
 namespace IO
 {

code/foundation/io/archfs/archivefilesystem.cc

@@ -8,7 +8,7 @@
 
 namespace IO
 {
-#if __WIN32__ || __linux__
+#if __WIN32__ || __linux__ || __APPLE__
 __ImplementClass(IO::ArchiveFileSystem, 'ARFS', IO::ZipFileSystem);
 __ImplementInterfaceSingleton(IO::ArchiveFileSystem);
 #else

code/foundation/io/archfs/archivefilesystem.h

@@ -9,7 +9,7 @@
     (C) 2009 Radon Labs GmbH
     (C) 2013-2020 Individual contributors, see AUTHORS file
 */
-#if __WIN32__ || __linux__
+#if __WIN32__ || __linux__ || __APPLE__
 #include "io/zipfs/zipfilesystem.h"
 namespace IO
 {

code/foundation/io/console.cc

@@ -8,9 +8,7 @@
 #include "threading/thread.h"
 #if __WIN32__
 #include "io/win32/win32consolehandler.h"
-#elif __OSX__
-#include "io/osx/osxconsolehandler.h"
-#elif __linux__
+#elif __linux__ || __APPLE__
 #include "io/posix/posixconsolehandler.h"
 #endif
 
@@ -59,9 +57,7 @@ Console::Open()
     // create default console handlers
     #if __WIN32__
     Ptr<ConsoleHandler> consoleHandler = Win32::Win32ConsoleHandler::Create();
-    #elif __OSX__
-    Ptr<ConsoleHandler> consoleHandler = OSX::OSXConsoleHandler::Create();
-    #elif __linux__
+    #elif __linux__ || __APPLE__
     Ptr<ConsoleHandler> consoleHandler = Posix::PosixConsoleHandler::Create();
     #endif
     this->AttachHandler(consoleHandler);    

code/foundation/io/filetime.h

@@ -18,13 +18,7 @@ namespace IO
 {
 typedef Win32::Win32FileTime FileTime;
 }
-#elif __OSX__
-#include "io/osx/osxfiletime.h"
-namespace IO
-{
-typedef OSX::OSXFileTime FileTime;
-}
-#elif __linux__
+#elif __linux__ || __APPLE__
 #include "io/posix/posixfiletime.h"
 namespace IO
 {

code/foundation/io/filewatcher.h

@@ -22,7 +22,7 @@
 #include "util/bitfield.h"
 #if __WIN32__
 #include "io/win32/win32filewatcher.h"
-#elif __linux__
+#elif __linux__ || __APPLE__
 #include "io/posix/linuxfilewatcher.h"
 #else
 #error "not implemented"

code/foundation/io/gamecontentserver.cc

@@ -8,7 +8,7 @@
 
 namespace IO
 {
-#if (__WIN32__ || __OSX__ || __linux__)
+#if (__WIN32__ || __APPLE__ || __linux__)
 __ImplementClass(IO::GameContentServer, 'IGCS', Base::GameContentServerBase);
 #else
 #error "IO::GameContentServer not implemented on this platform!"

code/foundation/io/gamecontentserver.h

@@ -12,7 +12,7 @@
     (C) 2013-2020 Individual contributors, see AUTHORS file
 */
 #include "core/config.h"
-#if (__WIN32__ || __OSX__ || __linux__)
+#if (__WIN32__ || __APPLE__ || __linux__)
 #include "io/base/gamecontentserverbase.h"
 namespace IO
 {

code/foundation/math/point.h

@@ -231,7 +231,7 @@ point::storeu(scalar* ptr) const
 __forceinline void
 point::store3(scalar* ptr) const
 {
-    __m128 v = _mm_permute_ps(this->vec, _MM_SHUFFLE(2, 2, 2, 2));
+    __m128 v = _mm_shuffle_ps(this->vec, this->vec, _MM_SHUFFLE(2, 2, 2, 2));
     _mm_storel_epi64(reinterpret_cast<__m128i*>(ptr), _mm_castps_si128(this->vec));
     _mm_store_ss(&ptr[2], v);
 }
@@ -243,8 +243,8 @@ point::store3(scalar* ptr) const
 __forceinline void
 point::storeu3(scalar* ptr) const
 {
-    __m128 t1 = _mm_permute_ps(this->vec, _MM_SHUFFLE(1, 1, 1, 1));
-    __m128 t2 = _mm_permute_ps(this->vec, _MM_SHUFFLE(2, 2, 2, 2));
+    __m128 t1 = _mm_shuffle_ps(this->vec, this->vec, _MM_SHUFFLE(1, 1, 1, 1));
+    __m128 t2 = _mm_shuffle_ps(this->vec, this->vec, _MM_SHUFFLE(2, 2, 2, 2));
     _mm_store_ss(&ptr[0], this->vec);
     _mm_store_ss(&ptr[1], t1);
     _mm_store_ss(&ptr[2], t2);

code/foundation/math/sse.h

@@ -9,7 +9,14 @@
     (C) 2018 Individual contributors, see AUTHORS file
 */
 //------------------------------------------------------------------------------
-
+#ifdef __WIN32__ || __linux__
+#include <xmmintrin.h>
+#include <emmintrin.h>
+#include <smmintrin.h>
+#include <immintrin.h>
+#elif __APPLE__
+#include <sse2neon.h>
+#endif
 
 namespace Math
 {

code/foundation/math/vec4.h

@@ -211,7 +211,7 @@ vec4::storeu(scalar* ptr) const
 __forceinline void 
 vec4::store3(scalar* ptr) const
 {
-    __m128 vv = _mm_permute_ps(this->vec, _MM_SHUFFLE(2, 2, 2, 2));
+    __m128 vv = _mm_shuffle_ps(this->vec, this->vec, _MM_SHUFFLE(2, 2, 2, 2));
     _mm_storel_epi64(reinterpret_cast<__m128i*>(ptr), _mm_castps_si128(this->vec));
     _mm_store_ss(&ptr[2], vv);
 }
@@ -222,8 +222,8 @@ vec4::store3(scalar* ptr) const
 __forceinline void 
 vec4::storeu3(scalar* ptr) const
 {
-    __m128 t1 = _mm_permute_ps(this->vec, _MM_SHUFFLE(1, 1, 1, 1));
-    __m128 t2 = _mm_permute_ps(this->vec, _MM_SHUFFLE(2, 2, 2, 2));
+    __m128 t1 = _mm_shuffle_ps(this->vec, this->vec, _MM_SHUFFLE(1, 1, 1, 1));
+    __m128 t2 = _mm_shuffle_ps(this->vec, this->vec, _MM_SHUFFLE(2, 2, 2, 2));
     _mm_store_ss(&ptr[0], this->vec);
     _mm_store_ss(&ptr[1], t1);
     _mm_store_ss(&ptr[2], t2);
@@ -1073,8 +1073,13 @@ permute(const vec4& v0, const vec4& v1, unsigned int i0, unsigned int i1, unsign
     __m128i vSelect = _mm_cmpgt_epi32(vControl, three);
     vControl = _mm_and_si128(vControl, three);
 
+#if N_USE_NEON
+    __m128 shuffled1 = vbslq_f32(vreinterpretq_u32_f32(vControl), v0.vec, v0.vec);
+    __m128 shuffled2 = vbslq_f32(vreinterpretq_u32_f32(vControl), v1.vec, v1.vec);
+#elif N_USE_AVX
     __m128 shuffled1 = _mm_permutevar_ps(v0.vec, vControl);
     __m128 shuffled2 = _mm_permutevar_ps(v1.vec, vControl);
+#endif
 
     __m128 masked1 = _mm_andnot_ps(_mm_castsi128_ps(vSelect), shuffled1);
     __m128 masked2 = _mm_and_ps(_mm_castsi128_ps(vSelect), shuffled2);

code/foundation/math/vector.h

@@ -179,7 +179,7 @@ vector::loadu(const scalar* ptr)
 __forceinline void
 vector::store(scalar* ptr) const
 {
-    __m128 v = _mm_permute_ps(this->vec, _MM_SHUFFLE(2, 2, 2, 2));
+    __m128 v = _mm_shuffle_ps(this->vec, this->vec, _MM_SHUFFLE(2, 2, 2, 2));
     _mm_storel_epi64(reinterpret_cast<__m128i*>(ptr), _mm_castps_si128(this->vec));
     _mm_store_ss(&ptr[2], v);
 }
@@ -191,8 +191,8 @@ vector::store(scalar* ptr) const
 __forceinline void
 vector::storeu(scalar* ptr) const
 {
-    __m128 t1 = _mm_permute_ps(this->vec, _MM_SHUFFLE(1, 1, 1, 1));
-    __m128 t2 = _mm_permute_ps(this->vec, _MM_SHUFFLE(2, 2, 2, 2));
+    __m128 t1 = _mm_shuffle_ps(this->vec, this->vec, _MM_SHUFFLE(1, 1, 1, 1));
+    __m128 t2 = _mm_shuffle_ps(this->vec, this->vec, _MM_SHUFFLE(2, 2, 2, 2));
     _mm_store_ss(&ptr[0], this->vec);
     _mm_store_ss(&ptr[1], t1);
     _mm_store_ss(&ptr[2], t2);