ref_dx2 (WIP)

ref/dx2/dllmain.cpp

@@ -0,0 +1,19 @@
+// dllmain.cpp : Defines the entry point for the DLL application.
+#include "windows.h"
+
+BOOL APIENTRY DllMain( HMODULE hModule,
+                       DWORD  ul_reason_for_call,
+                       LPVOID lpReserved
+                     )
+{
+    switch (ul_reason_for_call)
+    {
+    case DLL_PROCESS_ATTACH:
+    case DLL_THREAD_ATTACH:
+    case DLL_THREAD_DETACH:
+    case DLL_PROCESS_DETACH:
+        break;
+    }
+    return TRUE;
+}
+

ref/dx2/r_image.c

@@ -0,0 +1,722 @@
+/*
+r_context.c -- dx2 renderer image
+Copyright (C) 2025 lewa_j
+
+This program is free software: you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation, either version 3 of the License, or
+(at your option) any later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+*/
+
+#include "r_local.h"
+#include "com_strings.h"
+#include "xash3d_mathlib.h"
+#include "crtlib.h"
+#include "crclib.h"
+
+#define Assert(x) if(!( x )) gEngfuncs.Host_Error( "assert failed at %s:%i\n", __FILE__, __LINE__ )
+
+
+static dx_texture_t		dx_textures[MAX_TEXTURES];
+static dx_texture_t *dx_texturesHashTable[TEXTURES_HASH_SIZE];
+static uint		dx_numTextures;
+
+static rgbdata_t *GL_FakeImage(int width, int height, int depth, int flags)
+{
+	static byte	data2D[1024]; // 16x16x4
+	static rgbdata_t	r_image;
+
+	// also use this for bad textures, but without alpha
+	r_image.width = Q_max(1, width);
+	r_image.height = Q_max(1, height);
+	r_image.depth = Q_max(1, depth);
+	r_image.flags = flags;
+	r_image.type = PF_BGRA_32;
+	r_image.size = r_image.width * r_image.height * r_image.depth * 4;
+	r_image.buffer = (r_image.size > sizeof(data2D)) ? NULL : data2D;
+	r_image.palette = NULL;
+	r_image.numMips = 1;
+	r_image.encode = 0;
+
+	if (FBitSet(r_image.flags, IMAGE_CUBEMAP))
+		r_image.size *= 6;
+	memset(data2D, 0xFF, sizeof(data2D));
+
+	return &r_image;
+}
+
+static byte    dottexture[8][8] =
+{
+	  {0,1,1,0,0,0,0,0},
+	  {1,1,1,1,0,0,0,0},
+	  {1,1,1,1,0,0,0,0},
+	  {0,1,1,0,0,0,0,0},
+	  {0,0,0,0,0,0,0,0},
+	  {0,0,0,0,0,0,0,0},
+	  {0,0,0,0,0,0,0,0},
+	  {0,0,0,0,0,0,0,0},
+};
+
+static void GL_CreateInternalTextures(void)
+{
+	int	dx2, dy, d;
+	int	x, y;
+	rgbdata_t *pic;
+
+	// emo-texture from quake1
+	pic = GL_FakeImage(16, 16, 1, IMAGE_HAS_COLOR);
+
+	for (y = 0; y < 16; y++)
+	{
+		for (x = 0; x < 16; x++)
+		{
+			if ((y < 8) ^ (x < 8))
+				((uint *)pic->buffer)[y * 16 + x] = 0xFFFF00FF;
+			else ((uint *)pic->buffer)[y * 16 + x] = 0xFF000000;
+		}
+	}
+
+	tr.defaultTexture = GL_LoadTextureFromBuffer(REF_DEFAULT_TEXTURE, pic, TF_COLORMAP, false);
+
+	// particle texture from quake1
+	pic = GL_FakeImage(8, 8, 1, IMAGE_HAS_COLOR | IMAGE_HAS_ALPHA);
+
+	for (x = 0; x < 8; x++)
+	{
+		for (y = 0; y < 8; y++)
+		{
+			if (dottexture[x][y])
+				pic->buffer[(y * 8 + x) * 4 + 3] = 255;
+			else pic->buffer[(y * 8 + x) * 4 + 3] = 0;
+		}
+	}
+
+	tr.particleTexture = GL_LoadTextureFromBuffer(REF_PARTICLE_TEXTURE, pic, TF_CLAMP, false);
+
+	// white texture
+	pic = GL_FakeImage(4, 4, 1, IMAGE_HAS_COLOR);
+	for (x = 0; x < 16; x++)
+		((uint *)pic->buffer)[x] = 0xFFFFFFFF;
+	tr.whiteTexture = GL_LoadTextureFromBuffer(REF_WHITE_TEXTURE, pic, TF_COLORMAP, false);
+
+	// gray texture
+	pic = GL_FakeImage(4, 4, 1, IMAGE_HAS_COLOR);
+	for (x = 0; x < 16; x++)
+		((uint *)pic->buffer)[x] = 0xFF7F7F7F;
+	tr.grayTexture = GL_LoadTextureFromBuffer(REF_GRAY_TEXTURE, pic, TF_COLORMAP, false);
+
+	// black texture
+	pic = GL_FakeImage(4, 4, 1, IMAGE_HAS_COLOR);
+	for (x = 0; x < 16; x++)
+		((uint *)pic->buffer)[x] = 0xFF000000;
+	tr.blackTexture = GL_LoadTextureFromBuffer(REF_BLACK_TEXTURE, pic, TF_COLORMAP, false);
+
+	// cinematic dummy
+	pic = GL_FakeImage(640, 100, 1, IMAGE_HAS_COLOR);
+	tr.cinTexture = GL_LoadTextureFromBuffer("*cintexture", pic, TF_NOMIPMAP | TF_CLAMP, false);
+}
+
+void R_InitImages(void)
+{
+	memset(dx_textures, 0, sizeof(dx_textures));
+	memset(dx_texturesHashTable, 0, sizeof(dx_texturesHashTable));
+	dx_numTextures = 0;
+
+	// create unused 0-entry
+	Q_strncpy(dx_textures->name, "*unused*", sizeof(dx_textures->name));
+	dx_textures->hashValue = COM_HashKey(dx_textures->name, TEXTURES_HASH_SIZE);
+	dx_textures->nextHash = dx_texturesHashTable[dx_textures->hashValue];
+	dx_texturesHashTable[dx_textures->hashValue] = dx_textures;
+	dx_numTextures = 1;
+
+	GL_CreateInternalTextures();
+}
+
+dx_texture_t *R_GetTexture(unsigned int texnum)
+{
+	Assert(texnum < MAX_TEXTURES);
+	return &dx_textures[texnum];
+}
+
+void R_ShowTextures( void )
+{
+
+}
+
+const byte *R_GetTextureOriginalBuffer(unsigned int idx)
+{
+	return NULL;
+}
+
+static qboolean GL_CheckTexName(const char *name)
+{
+	int len;
+
+	if( COM_StringEmptyOrNULL( name ))
+		return false;
+
+	len = Q_strlen(name);
+
+	// because multi-layered textures can exceed name string
+	if (len >= sizeof(dx_textures->name))
+	{
+		gEngfuncs.Con_Printf(S_ERROR "LoadTexture: too long name %s (%d)\n", name, len);
+		return false;
+	}
+
+	return true;
+}
+
+static dx_texture_t *GL_TextureForName(const char *name)
+{
+	dx_texture_t *tex;
+	uint		hash;
+
+	// find the texture in array
+	hash = COM_HashKey(name, TEXTURES_HASH_SIZE);
+
+	for (tex = dx_texturesHashTable[hash]; tex != NULL; tex = tex->nextHash)
+	{
+		if (!Q_stricmp(tex->name, name))
+			return tex;
+	}
+
+	return NULL;
+}
+
+static dx_texture_t *GL_AllocTexture(const char *name, texFlags_t flags)
+{
+	dx_texture_t *tex = NULL;
+
+	// find a free texture_t slot
+	uint i;
+
+	//skip 0
+	for (i = 1; i < MAX_TEXTURES; i++)
+	{
+		if (dx_textures[i].used)
+			continue;
+
+		tex = &dx_textures[i];
+		break;
+	}
+
+	if (tex == NULL)
+	{
+		gEngfuncs.Host_Error("%s: MAX_TEXTURES limit exceeds\n", __func__);
+		return NULL;
+	}
+
+	// copy initial params
+	Q_strncpy(tex->name, name, sizeof(tex->name));
+	tex->used = 1;
+	tex->flags = flags;
+
+	// increase counter
+	dx_numTextures = Q_max((tex - dx_textures) + 1, dx_numTextures);
+
+	// add to hash table
+	tex->hashValue = COM_HashKey(name, TEXTURES_HASH_SIZE);
+	tex->nextHash = dx_texturesHashTable[tex->hashValue];
+	dx_texturesHashTable[tex->hashValue] = tex;
+
+	return tex;
+}
+
+static void GL_ProcessImage(dx_texture_t *tex, rgbdata_t *pic)
+{
+	uint	img_flags = 0;
+
+	// force upload texture as RGB or RGBA (detail textures requires this)
+	if (tex->flags & TF_FORCE_COLOR) pic->flags |= IMAGE_HAS_COLOR;
+	if (pic->flags & IMAGE_HAS_ALPHA) tex->flags |= TF_HAS_ALPHA;
+
+	if (ImageCompressed(pic->type))
+	{
+		if (!pic->numMips)
+			tex->flags |= TF_NOMIPMAP; // disable mipmapping by user request
+
+		// clear all the unsupported flags
+		tex->flags &= ~TF_KEEP_SOURCE;
+	}
+	else
+	{
+		// copy flag about luma pixels
+		if (pic->flags & IMAGE_HAS_LUMA)
+			tex->flags |= TF_HAS_LUMA;
+
+		if (pic->flags & IMAGE_QUAKEPAL)
+			tex->flags |= TF_QUAKEPAL;
+
+		// create luma texture from quake texture
+		if (tex->flags & TF_MAKELUMA)
+		{
+			img_flags |= IMAGE_MAKE_LUMA;
+			tex->flags &= ~TF_MAKELUMA;
+		}
+
+		if (!FBitSet(tex->flags, TF_IMG_UPLOADED) && FBitSet(tex->flags, TF_KEEP_SOURCE))
+			tex->original = gEngfuncs.FS_CopyImage(pic); // because current pic will be expanded to rgba
+
+		// we need to expand image into RGBA buffer
+		if (pic->type == PF_INDEXED_24 || pic->type == PF_INDEXED_32)
+			img_flags |= IMAGE_FORCE_RGBA;
+
+		// processing image before uploading (force to rgba, make luma etc)
+		if (pic->buffer) gEngfuncs.Image_Process(&pic, 0, 0, img_flags, 0);
+
+		if (FBitSet(tex->flags, TF_LUMINANCE))
+			ClearBits(pic->flags, IMAGE_HAS_COLOR);
+	}
+}
+
+static int GL_CalcMipmapCount(dx_texture_t *tex, qboolean haveBuffer)
+{
+	int	width, height;
+	int	mipcount;
+
+	Assert(tex != NULL);
+
+	//TEMP
+	return 1;
+
+	if (!haveBuffer)
+		return 1;
+
+	// generate mip-levels by user request
+	if (FBitSet(tex->flags, TF_NOMIPMAP))
+		return 1;
+
+	// mip-maps can't exceeds 16
+	for (mipcount = 0; mipcount < 16; mipcount++)
+	{
+		width = Q_max(1, (tex->width >> mipcount));
+		height = Q_max(1, (tex->height >> mipcount));
+		if (width == 1 && height == 1)
+			break;
+	}
+
+	return mipcount + 1;
+}
+
+static byte *GL_ResampleTexture(const byte *source, int inWidth, int inHeight, int outWidth, int outHeight, qboolean isNormalMap)
+{
+	uint		frac, fracStep;
+	uint *in = (uint *)source;
+	uint		p1[0x1000], p2[0x1000];
+	byte *pix1, *pix2, *pix3, *pix4;
+	uint *out, *inRow1, *inRow2;
+	static byte *scaledImage = NULL;	// pointer to a scaled image
+	vec3_t		normal;
+	int		i, x, y;
+
+	if (!source) return NULL;
+
+	scaledImage = Mem_Realloc(r_temppool, scaledImage, outWidth * outHeight * 4);
+	fracStep = inWidth * 0x10000 / outWidth;
+	out = (uint *)scaledImage;
+
+	frac = fracStep >> 2;
+	for (i = 0; i < outWidth; i++)
+	{
+		p1[i] = 4 * (frac >> 16);
+		frac += fracStep;
+	}
+
+	frac = (fracStep >> 2) * 3;
+	for (i = 0; i < outWidth; i++)
+	{
+		p2[i] = 4 * (frac >> 16);
+		frac += fracStep;
+	}
+
+	if (isNormalMap)
+	{
+		for (y = 0; y < outHeight; y++, out += outWidth)
+		{
+			inRow1 = in + inWidth * (int)(((float)y + 0.25f) * inHeight / outHeight);
+			inRow2 = in + inWidth * (int)(((float)y + 0.75f) * inHeight / outHeight);
+
+			for (x = 0; x < outWidth; x++)
+			{
+				pix1 = (byte *)inRow1 + p1[x];
+				pix2 = (byte *)inRow1 + p2[x];
+				pix3 = (byte *)inRow2 + p1[x];
+				pix4 = (byte *)inRow2 + p2[x];
+
+				normal[0] = MAKE_SIGNED(pix1[0]) + MAKE_SIGNED(pix2[0]) + MAKE_SIGNED(pix3[0]) + MAKE_SIGNED(pix4[0]);
+				normal[1] = MAKE_SIGNED(pix1[1]) + MAKE_SIGNED(pix2[1]) + MAKE_SIGNED(pix3[1]) + MAKE_SIGNED(pix4[1]);
+				normal[2] = MAKE_SIGNED(pix1[2]) + MAKE_SIGNED(pix2[2]) + MAKE_SIGNED(pix3[2]) + MAKE_SIGNED(pix4[2]);
+
+				if (!VectorNormalizeLength(normal))
+					VectorSet(normal, 0.5f, 0.5f, 1.0f);
+
+				((byte *)(out + x))[0] = 128 + (byte)(127.0f * normal[0]);
+				((byte *)(out + x))[1] = 128 + (byte)(127.0f * normal[1]);
+				((byte *)(out + x))[2] = 128 + (byte)(127.0f * normal[2]);
+				((byte *)(out + x))[3] = 255;
+			}
+		}
+	}
+	else
+	{
+		for (y = 0; y < outHeight; y++, out += outWidth)
+		{
+			inRow1 = in + inWidth * (int)(((float)y + 0.25f) * inHeight / outHeight);
+			inRow2 = in + inWidth * (int)(((float)y + 0.75f) * inHeight / outHeight);
+
+			for (x = 0; x < outWidth; x++)
+			{
+				pix1 = (byte *)inRow1 + p1[x];
+				pix2 = (byte *)inRow1 + p2[x];
+				pix3 = (byte *)inRow2 + p1[x];
+				pix4 = (byte *)inRow2 + p2[x];
+
+				((byte *)(out + x))[0] = (pix1[0] + pix2[0] + pix3[0] + pix4[0]) >> 2;
+				((byte *)(out + x))[1] = (pix1[1] + pix2[1] + pix3[1] + pix4[1]) >> 2;
+				((byte *)(out + x))[2] = (pix1[2] + pix2[2] + pix3[2] + pix4[2]) >> 2;
+				((byte *)(out + x))[3] = (pix1[3] + pix2[3] + pix3[3] + pix4[3]) >> 2;
+			}
+		}
+	}
+
+	return scaledImage;
+}
+
+static void GL_TextureImageRAW(dx_texture_t *tex, int width, int height, int type, const char *data)
+{
+	if (!data)
+		return;
+
+	DDSURFACEDESC ddsd = {};
+	ZeroMemory(&ddsd, sizeof(ddsd));
+	ddsd.dwSize = sizeof(ddsd);
+	ddsd.dwFlags = DDSD_LPSURFACE;
+	DXCheck(IDirectDrawSurface_Lock(tex->dds, NULL, &ddsd, DDLOCK_SURFACEMEMORYPTR, NULL));
+
+	int pitch = ddsd.lPitch;
+	if (!pitch)
+		pitch = ddsd.dwWidth * 4;
+	char *dst = (char *)ddsd.lpSurface;
+
+	for (int j = 0; j < (int)ddsd.dwHeight; j++)
+	{
+		//RGBA to BGRA
+		if (type == PF_RGBA_32)
+		{
+			for (int i = 0; i < ddsd.dwWidth; i++)
+			{
+				dst[i * 4 + 0] = data[i * 4 + 2];
+				dst[i * 4 + 1] = data[i * 4 + 1];
+				dst[i * 4 + 2] = data[i * 4 + 0];
+				dst[i * 4 + 3] = data[i * 4 + 3];
+			}
+		}
+		else if (type == PF_BGRA_32)
+		{
+			memcpy(dst, data, ddsd.dwWidth * 4);
+		}
+
+		dst += pitch;
+		data += ddsd.dwWidth * 4;
+	}
+
+	DXCheck(IDirectDrawSurface_Unlock(tex->dds, ddsd.lpSurface));
+}
+
+static qboolean GL_UploadTexture(dx_texture_t *tex, rgbdata_t *pic)
+{
+	byte *buf, *data;
+	size_t		texsize, size;
+	uint		width, height;
+	uint		i, j, numSides;
+	qboolean		normalMap;
+	const byte *bufend;
+
+	// dedicated server
+	if (!dxc.window)
+		return true;
+
+	Assert(pic != NULL);
+	Assert(tex != NULL);
+
+	qboolean supported = true;
+
+	if (FBitSet(pic->flags, IMAGE_CUBEMAP))
+		supported = false;
+	else if (FBitSet(pic->flags, IMAGE_MULTILAYER) && pic->depth >= 1)
+		supported = false;
+	else if (pic->width > 1 && pic->height > 1 && pic->depth > 1)
+		supported = false;
+	else if (FBitSet(tex->flags, TF_RECTANGLE))
+		supported = false;
+	else if (FBitSet(tex->flags, TF_MULTISAMPLE))
+		supported = false;
+
+	// make sure what target is correct
+	if (!supported)
+	{
+		gEngfuncs.Con_DPrintf(S_ERROR "%s: %s is not supported\n", __func__, tex->name);
+		return false;
+	}
+
+	//if (pic->type == PF_BC6H_SIGNED || pic->type == PF_BC6H_UNSIGNED || pic->type == PF_BC7_UNORM || pic->type == PF_BC7_SRGB)
+	if (ImageCompressed(pic->type))
+	{
+		gEngfuncs.Con_DPrintf(S_ERROR "%s: compressed textures are not supported\n", __func__);
+		return false;
+	}
+
+	// store original sizes
+	tex->srcWidth = pic->width;
+	tex->srcHeight = pic->height;
+
+	// set the texture dimensions
+	{
+		int	step = 0;// (int)gl_round_down.value;
+		int	scaled_width, scaled_height;
+
+		for (scaled_width = 1; scaled_width < pic->width; scaled_width <<= 1);
+
+		if (step > 0 && pic->width < scaled_width && (step == 1 || (scaled_width - pic->width) > (scaled_width >> step)))
+			scaled_width >>= 1;
+
+		for (scaled_height = 1; scaled_height < pic->height; scaled_height <<= 1);
+
+		if (step > 0 && pic->height < scaled_height && (step == 1 || (scaled_height - pic->height) > (scaled_height >> step)))
+			scaled_height >>= 1;
+
+		tex->width = Q_max(1, scaled_width);
+		tex->height = Q_max(1, scaled_height);
+	}
+
+	DDSURFACEDESC ddsd = {};
+	ZeroMemory(&ddsd, sizeof(ddsd));
+	ddsd.dwSize = sizeof(ddsd);
+	ddsd.dwFlags = DDSD_CAPS | DDSD_WIDTH | DDSD_HEIGHT;
+	ddsd.ddsCaps.dwCaps = DDSCAPS_TEXTURE;
+	//if (pic->numMips > 1)
+	//	ddsd.ddsCaps.dwCaps |= DDSCAPS_MIPMAP | DDSCAPS_COMPLEX;
+	ddsd.dwWidth = tex->width;
+	ddsd.dwHeight = tex->height;
+
+#if 0
+	//RGBA8
+	ddsd.dwFlags |= DDSD_PIXELFORMAT;
+	ddsd.ddpfPixelFormat.dwSize = sizeof(ddsd.ddpfPixelFormat);
+	ddsd.ddpfPixelFormat.dwFlags = DDPF_RGB | DDPF_ALPHAPIXELS;
+	ddsd.ddpfPixelFormat.dwRGBBitCount = 32;
+	ddsd.ddpfPixelFormat.dwRBitMask = 0xFF;
+	ddsd.ddpfPixelFormat.dwGBitMask = 0xFF00;
+	ddsd.ddpfPixelFormat.dwBBitMask = 0xFF0000;
+	ddsd.ddpfPixelFormat.dwRGBAlphaBitMask = 0xFF000000;
+#endif
+
+	DXCheck(IDirectDraw_CreateSurface(dxc.pdd, &ddsd, &tex->dds, NULL));
+
+	tex->fogParams[0] = pic->fogParams[0];
+	tex->fogParams[1] = pic->fogParams[1];
+	tex->fogParams[2] = pic->fogParams[2];
+	tex->fogParams[3] = pic->fogParams[3];
+
+	buf = pic->buffer;
+	bufend = pic->buffer + pic->size; // total image size include all the layers, cube sides, mipmaps
+	normalMap = FBitSet(tex->flags, TF_NORMALMAP) ? true : false;
+
+	if (Q_max(1, pic->numMips) > 1)	// not-compressed DDS
+	{
+		//TEMP
+		j = 0;
+		//for (j = 0; j < Q_max(1, pic->numMips); j++)
+		{
+			// track the buffer bounds
+			if (buf != NULL && buf >= bufend)
+				gEngfuncs.Host_Error("%s: %s image buffer overflow\n", __func__, tex->name);
+
+			width = Q_max(1, (tex->width >> j));
+			height = Q_max(1, (tex->height >> j));
+			size = gEngfuncs.Image_CalcImageSize(pic->type, width, height, 1);
+			GL_TextureImageRAW(tex, width, height, pic->type, buf);
+			buf += size; // move pointer
+			tex->numMips++;
+		}
+	}
+	else // RGBA32
+	{
+		// track the buffer bounds
+		if (buf != NULL && buf >= bufend)
+			gEngfuncs.Host_Error("%s: %s image buffer overflow\n", __func__, tex->name);
+
+		int mipCount = GL_CalcMipmapCount(tex, (buf != NULL));
+
+		// NOTE: only single uncompressed textures can be resamples, no mips, no layers, no sides
+		if (((pic->width != tex->width) || (pic->height != tex->height)))
+			data = GL_ResampleTexture(buf, pic->width, pic->height, tex->width, tex->height, normalMap);
+		else data = buf;
+
+		//if (!ImageCompressed(pic->type) && !FBitSet(tex->flags, TF_NOMIPMAP) && FBitSet(pic->flags, IMAGE_ONEBIT_ALPHA))
+		//	data = GL_ApplyFilter(data, tex->width, tex->height);
+
+		// mips will be auto-generated if desired
+		for (j = 0; j < mipCount; j++)
+		{
+			width = Q_max(1, (tex->width >> j));
+			height = Q_max(1, (tex->height >> j));
+			size = gEngfuncs.Image_CalcImageSize(pic->type, width, height, 1);
+			GL_TextureImageRAW(tex, width, height, pic->type, data);
+			//if (mipCount > 1)
+			//	GL_BuildMipMap(data, width, height, tex->depth, tex->flags);
+			tex->numMips++;
+		}
+	}
+
+	SetBits(tex->flags, TF_IMG_UPLOADED); // done
+
+	return true;
+}
+
+int GL_LoadTextureFromBuffer(const char *name, rgbdata_t *pic, texFlags_t flags, qboolean update)
+{
+	dx_texture_t *tex;
+
+	if (!GL_CheckTexName(name))
+		return 0;
+
+	// see if already loaded
+	if ((tex = GL_TextureForName(name)) && !update)
+		return (tex - dx_textures);
+
+	// couldn't loading image
+	if (!pic) return 0;
+
+	if (update)
+	{
+		if (tex == NULL)
+			gEngfuncs.Host_Error("%s: couldn't find texture %s for update\n", __func__, name);
+		SetBits(tex->flags, flags);
+	}
+	else
+	{
+		// allocate the new one
+		tex = GL_AllocTexture(name, flags);
+	}
+
+	GL_ProcessImage(tex, pic);
+
+	if (!GL_UploadTexture(tex, pic))
+	{
+		memset(tex, 0, sizeof(dx_texture_t));
+		return 0;
+	}
+
+	return (tex - dx_textures);
+}
+
+void GL_ProcessTexture(int texnum, float gamma, int topColor, int bottomColor)
+{
+	;
+}
+
+void GetDetailScaleForTexture(int texture, float *xScale, float *yScale)
+{
+	*xScale = *yScale = 1.0f;
+}
+
+void GetExtraParmsForTexture(int texture, byte *red, byte *green, byte *blue, byte *alpha)
+{
+	*red = *green = *blue = *alpha = 0;
+}
+
+int GL_FindTexture(const char *name)
+{
+	gEngfuncs.Con_Printf("GL_FindTexture(%s)\n", name);
+	return 0;
+}
+
+const char *GL_TextureName(unsigned int texnum)
+{
+	return NULL;
+}
+
+const byte *GL_TextureData(unsigned int texnum)
+{
+	return NULL;
+}
+
+int GL_LoadTexture( const char *name, const byte *buf, size_t size, int flags )
+{
+	dx_texture_t *tex;
+	rgbdata_t *pic;
+	uint		picFlags = 0;
+
+	if (!GL_CheckTexName(name))
+		return 0;
+
+	// see if already loaded
+	if ((tex = GL_TextureForName(name)))
+		return (tex - dx_textures);
+
+	gEngfuncs.Con_Printf("GL_LoadTexture(%s, %p, %zu, %X)\n", name, buf, size, flags);
+
+	if (FBitSet(flags, TF_NOFLIP_TGA))
+		SetBits(picFlags, IL_DONTFLIP_TGA);
+
+	if (FBitSet(flags, TF_KEEP_SOURCE) && !FBitSet(flags, TF_EXPAND_SOURCE))
+		SetBits(picFlags, IL_KEEP_8BIT);
+
+	// set some image flags
+	gEngfuncs.Image_SetForceFlags(picFlags);
+
+	pic = gEngfuncs.FS_LoadImage(name, buf, size);
+	if (!pic) return 0; // couldn't loading image
+
+	// allocate the new one
+	tex = GL_AllocTexture(name, flags);
+	GL_ProcessImage(tex, pic);
+
+	if (!GL_UploadTexture(tex, pic))
+	{
+		memset(tex, 0, sizeof(dx_texture_t));
+		gEngfuncs.FS_FreeImage(pic); // release source texture
+		return 0;
+	}
+
+	//GL_ApplyTextureParams(tex); // update texture filter, wrap etc
+	gEngfuncs.FS_FreeImage(pic); // release source texture
+
+	// NOTE: always return texnum as index in array or engine will stop work !!!
+	return tex - dx_textures;
+}
+
+int GL_CreateTexture(const char *name, int width, int height, const void *buffer, texFlags_t flags)
+{
+	return 0;
+}
+
+int GL_LoadTextureArray(const char **names, int flags)
+{
+	return 0;
+}
+
+int GL_CreateTextureArray(const char *name, int width, int height, int depth, const void *buffer, texFlags_t flags)
+{
+	return 0;
+}
+
+void GL_FreeTexture(unsigned int texnum)
+{
+	if (texnum <= 0 || texnum >= MAX_TEXTURES)
+		return;
+
+	gEngfuncs.Con_Printf("GL_FreeTexture(%d) %s\n", texnum, dx_textures[texnum].name);
+}
+
+void R_OverrideTextureSourceSize(unsigned int textnum, unsigned int srcWidth, unsigned int srcHeight)
+{
+
+}

ref/dx2/r_local.h

@@ -0,0 +1,220 @@
+#pragma once
+
+#include "const.h"
+#include "cvardef.h"
+#include "ref_api.h"
+#include "render_api.h"
+#include "com_image.h"
+
+#define DIRECTDRAW_VERSION 0x0200
+#define DIRECT3D_VERSION 0x0500//D3DCOLORMODEL bug
+#include <ddraw.h>
+#include <d3d.h>
+
+extern ref_api_t      gEngfuncs;
+
+extern poolhandle_t r_temppool;
+
+
+#define MAX_TEXTURES            8192
+#define TEXTURES_HASH_SIZE	(MAX_TEXTURES >> 2)
+
+// refparams
+#define RP_NONE		0
+#define RP_ENVVIEW		BIT( 0 )	// used for cubemapshot
+#define RP_OLDVIEWLEAF	BIT( 1 )
+#define RP_CLIPPLANE	BIT( 2 )
+
+#define MAX_DRAW_STACK	2		// normal view and menu view
+
+typedef struct dx_texture_s
+{
+	char		name[256];	// game path, including extension (can be store image programs)
+	word		srcWidth;		// keep unscaled sizes
+	word		srcHeight;
+	word		width;		// upload width\height
+	word		height;
+
+	byte		numMips;		// mipmap count
+
+	int used;
+	texFlags_t	flags;
+	rgba_t		fogParams;	// some water textures
+	// contain info about underwater fog
+	rgbdata_t *original;	// keep original image
+#if 0
+	word		depth;		// texture depth or count of layers for 2D_ARRAY
+
+	GLuint		target;		// glTarget
+	GLuint		texnum;		// gl texture binding
+	GLint		format;		// uploaded format
+
+	// debug info
+	size_t		size;		// upload size for debug targets
+
+	// detail textures stuff
+	float		xscale;
+	float		yscale;
+
+#endif
+	IDirectDrawSurface *dds;
+	uint		hashValue;
+	struct dx_texture_s *nextHash;
+} dx_texture_t;
+
+typedef struct
+{
+	int		params;		// rendering parameters
+
+	qboolean		drawWorld;	// ignore world for drawing PlayerModel
+	qboolean		isSkyVisible;	// sky is visible
+	qboolean		onlyClientDraw;	// disabled by client request
+	qboolean		drawOrtho;	// draw world as orthogonal projection
+
+	float		fov_x, fov_y;	// current view fov
+
+	cl_entity_t *currententity;
+	model_t *currentmodel;
+	cl_entity_t *currentbeam;	// same as above but for beams
+
+	int		viewport[4];
+	//gl_frustum_t	frustum;
+
+	mleaf_t *viewleaf;
+	mleaf_t *oldviewleaf;
+	vec3_t		pvsorigin;
+	vec3_t		vieworg;		// locked vieworigin
+	vec3_t		viewangles;
+	vec3_t		vforward;
+	vec3_t		vright;
+	vec3_t		vup;
+
+	vec3_t		cullorigin;
+	vec3_t		cull_vforward;
+	vec3_t		cull_vright;
+	vec3_t		cull_vup;
+
+	float		farClip;
+
+	qboolean		fogCustom;
+	qboolean		fogEnabled;
+	qboolean		fogSkybox;
+	vec4_t		fogColor;
+	float		fogDensity;
+	float		fogStart;
+	float		fogEnd;
+	int		cached_contents;	// in water
+	int		cached_waterlevel;	// was in water
+
+	float		skyMins[2][SKYBOX_MAX_SIDES];
+	float		skyMaxs[2][SKYBOX_MAX_SIDES];
+
+	matrix4x4		objectMatrix;		// currententity matrix
+	matrix4x4		worldviewMatrix;		// modelview for world
+	matrix4x4		modelviewMatrix;		// worldviewMatrix * objectMatrix
+
+	matrix4x4		projectionMatrix;
+	matrix4x4		worldviewProjectionMatrix;	// worldviewMatrix * projectionMatrix
+	byte		visbytes[( MAX_MAP_LEAFS + 7 ) / 8];// actual PVS for current frame
+
+	float		viewplanedist;
+	mplane_t		clipPlane;
+} ref_instance_t;
+
+typedef struct
+{
+	cl_entity_t *solid_entities[MAX_VISIBLE_PACKET];	// opaque moving or alpha brushes
+	cl_entity_t *trans_entities[MAX_VISIBLE_PACKET];	// translucent brushes
+	cl_entity_t *beam_entities[MAX_VISIBLE_PACKET];
+	uint		num_solid_entities;
+	uint		num_trans_entities;
+	uint		num_beam_entities;
+} draw_list_t;
+
+typedef struct
+{
+	int		defaultTexture;   	// use for bad textures
+	int		particleTexture;
+	int		whiteTexture;
+	int		grayTexture;
+	int		blackTexture;
+	int		cinTexture;      	// cinematic texture
+	// entity lists
+	draw_list_t	draw_stack[MAX_DRAW_STACK];
+	int		draw_stack_pos;
+	draw_list_t *draw_list;
+
+	int		realframecount;	// not including viewpasses
+	int		framecount;
+
+	qboolean		fCustomRendering;
+	qboolean		fResetVis;
+
+	double		frametime;	// special frametime for multipass rendering (will set to 0 on a nextview)
+
+	// get from engine
+	cl_entity_t *entities;
+	uint max_entities;
+} gl_globals_t;
+
+extern gl_globals_t	tr;
+
+struct dx_context_s
+{
+	HWND window;
+	int windowWidth;
+	int windowHeight;
+	IDirectDraw *pdd;
+	IDirectDrawSurface *pddsPrimary;
+	IDirectDrawSurface *pddsBack;
+	IDirectDrawClipper *pddClipper;
+	IDirectDrawClipper *pddBackClipper;
+
+	int renderMode;
+};
+typedef struct dx_context_s dx_context_t;
+
+extern dx_context_t dxc;
+
+const char *dxResultToStr( HRESULT r );
+
+#define DXCheck(CALL) {\
+HRESULT r = CALL;\
+if (r != DD_OK)\
+	gEngfuncs.Con_Printf( S_ERROR "%s:%d: DirectX error %X(%d 0x%X %d) %s at \"" #CALL "\"\n", __FUNCTION__, __LINE__, r, ( r >> 31 ) & 1, ( r >> 16 ) & 0x7FFF, r & 0xFFFF, dxResultToStr( r ) ); \
+}
+
+void GL_SetRenderMode( int mode );
+
+//r_image.c
+void R_InitImages( void );
+dx_texture_t *R_GetTexture( unsigned int texnum );
+
+void R_ShowTextures( void );
+const byte *R_GetTextureOriginalBuffer( unsigned int idx );
+int GL_LoadTextureFromBuffer( const char *name, rgbdata_t *pic, texFlags_t flags, qboolean update );
+void GL_ProcessTexture( int texnum, float gamma, int topColor, int bottomColor );
+void GetDetailScaleForTexture( int texture, float *xScale, float *yScale );
+void GetExtraParmsForTexture( int texture, byte *red, byte *green, byte *blue, byte *alpha );
+int GL_FindTexture( const char *name );
+const char *GL_TextureName( unsigned int texnum );
+const byte *GL_TextureData( unsigned int texnum );
+int GL_LoadTexture( const char *name, const byte *buf, size_t size, int flags );
+int GL_CreateTexture( const char *name, int width, int height, const void *buffer, texFlags_t flags );
+int GL_LoadTextureArray( const char **names, int flags );
+int GL_CreateTextureArray( const char *name, int width, int height, int depth, const void *buffer, texFlags_t flags );
+void GL_FreeTexture( unsigned int texnum );
+void R_OverrideTextureSourceSize( unsigned int textnum, unsigned int srcWidth, unsigned int srcHeight );
+
+//
+// engine shared convars
+//
+DECLARE_ENGINE_SHARED_CVAR_LIST( )
+
+#define Mem_Malloc( pool, size ) _Mem_Alloc( pool, size, false, __FILE__, __LINE__ )
+#define Mem_Calloc( pool, size ) _Mem_Alloc( pool, size, true, __FILE__, __LINE__ )
+#define Mem_Realloc( pool, ptr, size ) gEngfuncs._Mem_Realloc( pool, ptr, size, true, __FILE__, __LINE__ )
+#define Mem_Free( mem ) _Mem_Free( mem, __FILE__, __LINE__ )
+#define Mem_AllocPool( name ) gEngfuncs._Mem_AllocPool( name, __FILE__, __LINE__ )
+#define Mem_FreePool( pool ) gEngfuncs._Mem_FreePool( pool, __FILE__, __LINE__ )
+#define Mem_EmptyPool( pool ) gEngfuncs._Mem_EmptyPool( pool, __FILE__, __LINE__ )

ref/dx2/wscript

@@ -0,0 +1,18 @@
+#! /usr/bin/env python
+# encoding: utf-8
+# mittorn, 2018
+
+def options(opt):
+	pass
+
+def configure(conf):
+	pass
+
+def build(bld):
+	bld.shlib(
+		source   = ['r_context.c', 'r_image.c'],
+		target   = 'ref_dx2',
+		defines  = 'REF_DLL',
+		use      = 'engine_includes sdk_includes werror public USER32 DDRAW',
+		install_path = bld.env.LIBDIR
+	)