cache vid.height

saves us some pointless work going through all planes if we dont need to draw the sky

src/am_map.c

@@ -785,7 +785,7 @@ static void AM_drawPixel(INT32 xx, INT32 yy, INT32 cc)
 {
 	if (xx < 0 || yy < 0 || xx >= f_w || yy >= f_h)
 		return; // off the screen
-	am_buf[(yy*f_w) + xx] = cc;
+	am_buf[(xx*f_h) + yy] = cc;
 }
 
 //

src/android/i_system.c

@@ -110,7 +110,7 @@ void I_Error(const char *error, ...)
   char logbuf[8192];
 
   va_start(argptr, error);
-  vsprintf(logbuf, error, argptr);
+  vsnprintf(logbuf, 8192, error, argptr);
   va_end(argptr);
 
   LOGE(logbuf);
@@ -166,7 +166,7 @@ void I_OutputMsg(const char *fmt, ...)
   char logbuf[8192];
 
   va_start(argptr, fmt);
-  vsprintf(logbuf, fmt, argptr);
+  vsnprintf(logbuf, 8192, fmt, argptr);
   va_end(argptr);
 
   LOGD(logbuf);

src/command.c

@@ -1650,13 +1650,13 @@ static void CV_SetCVar(consvar_t *var, const char *value, boolean stealth)
 		UINT8 *p = buf;
 		if (!(server || (IsPlayerAdmin(consoleplayer))))
 		{
-			CONS_Printf(M_GetText("Only the server or admin can change: %s %s\n"), var->name, var->string);
+			CONS_Alert(CONS_NOTICE, "Only the server or admin can change: %s %s\n", var->name, var->string);
 			return;
 		}
 
 		if (var == &cv_kartencore && !M_SecretUnlocked(SECRET_ENCORE))
 		{
-			CONS_Printf(M_GetText("You haven't unlocked Encore Mode yet!\n"));
+			CONS_Alert(CONS_NOTICE, "You haven't unlocked Encore Mode yet!\n");
 			return;
 		}
 
@@ -1664,7 +1664,7 @@ static void CV_SetCVar(consvar_t *var, const char *value, boolean stealth)
 		{
 			if (!stricmp(value, "Hard") || atoi(value) == 2)
 			{
-				CONS_Printf(M_GetText("You haven't unlocked this yet!\n"));
+				CONS_Alert(CONS_NOTICE, "You haven't unlocked this yet!\n");
 				return;
 			}
 		}
@@ -1681,14 +1681,13 @@ static void CV_SetCVar(consvar_t *var, const char *value, boolean stealth)
 		else
 			Setvalue(var, value, stealth);
 	}
+	else if ((var->flags & CV_NOTINNET) && netgame)
+	{
+		CONS_Alert(CONS_NOTICE, "This variable can't be changed while in netgame: %s %s\n", var->name, var->string);
+		return;
+	}
 	else
-		if ((var->flags & CV_NOTINNET) && netgame)
-		{
-			CONS_Printf(M_GetText("This variable can't be changed while in netgame: %s %s\n"), var->name, var->string);
-			return;
-		}
-		else
-			Setvalue(var, value, stealth);
+		Setvalue(var, value, stealth);
 }
 
 /** Sets a value to a variable without calling its callback function.
@@ -1771,7 +1770,7 @@ void CV_AddValue(consvar_t *var, INT32 increment)
 				newvalue = var->value - 1;
 				do
 				{
-					if(increment > 0) // Going up!
+					if (increment > 0) // Going up!
 					{
 						if (++newvalue == NUMMAPS)
 							newvalue = -1;
@@ -1785,7 +1784,7 @@ void CV_AddValue(consvar_t *var, INT32 increment)
 					if (newvalue == oldvalue)
 						break; // don't loop forever if there's none of a certain gametype
 
-					if(!mapheaderinfo[newvalue])
+					if (newvalue >= 0 && !mapheaderinfo[newvalue])
 						continue; // Don't allocate the header.  That just makes memory usage skyrocket.
 
 				} while (!M_CanShowLevelInList(newvalue, gt));

src/console.c

@@ -1282,7 +1282,7 @@ void CONS_Alert(alerttype_t level, const char *fmt, ...)
 		txt = malloc(8192);
 
 	va_start(argptr, fmt);
-	vsprintf(txt, fmt, argptr);
+	vsnprintf(txt, 8192, fmt, argptr);
 	va_end(argptr);
 
 	switch (level)
@@ -1318,7 +1318,7 @@ void CONS_Debug(INT32 debugflags, const char *fmt, ...)
 		txt = malloc(8192);
 
 	va_start(argptr, fmt);
-	vsprintf(txt, fmt, argptr);
+	vsnprintf(txt, 8192, fmt, argptr);
 	va_end(argptr);
 
 	// Again I am lazy, oh well

src/d_clisrv.c

@@ -4161,7 +4161,10 @@ static void Joinable_OnChange(void)
 	UINT8 maxplayer;
 
 	if (!server)
+	{
+		CONS_Alert(CONS_NOTICE, ("Only the server can change: %s %s\n"), cv_allownewplayer.name, cv_allownewplayer.string);
 		return;
+	}
 
 #ifdef SATURNJOIN
 	// disabling joins should also disable saturn joins unless its called with CV_StealthSet and vice versa to make life a bit easier

src/d_main.c

@@ -609,7 +609,7 @@ static boolean D_Display(void)
 		{
 			if (rendermode == render_soft)
 			{
-				VID_BlitLinearScreen(vid.screens[0], vid.screens[1], vid.width, vid.height, vid.width, vid.width);
+				VID_BlitLinearScreen(vid.screens[0], vid.screens[1], vid.width, vid.height, vid.height, vid.height);
 			}
 
 			lastdraw = false;
@@ -774,6 +774,7 @@ void D_SRB2Loop(void)
 		precise_t capbudget;
 		precise_t enterprecise = I_GetPreciseTime();
 
+		memset(&g_dc, 0, sizeof(g_dc));
 		Z_Frame_Reset();
 
 		// Casting the return value of a function is bad practice (apparently)
@@ -889,12 +890,6 @@ void D_SRB2Loop(void)
 				// always update console and hud
 				// otherwise it may take minutes to open it
 				CON_Drawer();
-
-				if (gamestate == GS_LEVEL)
-				{
-					ST_Drawer();
-					HU_Drawer();
-				}
 			}
 		}
 
@@ -1026,7 +1021,7 @@ void D_ClearState(void)
 	S_StopSounds();
 	S_ResetKeepAndSpecialMus(); // just in case
 
-	P_FreeLevelState();
+	//P_FreeLevelState();
 
 	G_SetGamestate(GS_NULL);
 	wipegamestate = GS_NULL;

src/dedicated/i_system.c

@@ -1468,7 +1468,7 @@ void I_Error(const char *error, ...)
 		if (errorcount > 20)
 		{
 			va_start(argptr, error);
-			vsprintf(buffer, error, argptr);
+			vsnprintf(buffer, 8192, error, argptr);
 			va_end(argptr);
 
 			I_OutputMsg("SRB2Kart %s Recursive Error", buffer);
@@ -1482,7 +1482,7 @@ void I_Error(const char *error, ...)
 
 	// Display error message in the console before we start shutting it down
 	va_start(argptr, error);
-	vsprintf(buffer, error, argptr);
+	vsnprintf(buffer, 8192, error, argptr);
 	va_end(argptr);
 	I_OutputMsg("\nI_Error(): %s\n", buffer);
 

src/f_wipe.c

@@ -249,16 +249,16 @@ static void F_DoWipe(fademask_t *fademask)
 			draw_linestart = scrypos[masky];
 			draw_lineend   = scrypos[masky + 1];
 
-			relativepos = (draw_linestart * vid.width) + draw_rowstart;
-			draw_linestogo = draw_lineend - draw_linestart;
+			relativepos = (draw_rowstart * vid.height) + draw_linestart;
+			draw_linestogo = draw_rowend - draw_rowstart;
 
 			if (*mask == 0)
 			{
 				// shortcut - memcpy source to work
 				while (draw_linestogo--)
 				{
-					M_Memcpy(w_base+relativepos, s_base+relativepos, draw_rowend-draw_rowstart);
-					relativepos += vid.width;
+					M_Memcpy(w_base+relativepos, s_base+relativepos, draw_lineend-draw_linestart);
+					relativepos += vid.height;
 				}
 			}
 			else if (*mask == 10)
@@ -266,8 +266,8 @@ static void F_DoWipe(fademask_t *fademask)
 				// shortcut - memcpy target to work
 				while (draw_linestogo--)
 				{
-					M_Memcpy(w_base+relativepos, e_base+relativepos, draw_rowend-draw_rowstart);
-					relativepos += vid.width;
+					M_Memcpy(w_base+relativepos, e_base+relativepos, draw_lineend-draw_linestart);
+					relativepos += vid.height;
 				}
 			}
 			else
@@ -281,12 +281,12 @@ static void F_DoWipe(fademask_t *fademask)
 					w = w_base + relativepos;
 					s = s_base + relativepos;
 					e = e_base + relativepos;
-					draw_rowstogo = draw_rowend - draw_rowstart;
+					draw_rowstogo = draw_lineend - draw_linestart;
 
 					while (draw_rowstogo--)
 						*w++ = transtbl[ ( *e++ << 8 ) + *s++ ];
 
-					relativepos += vid.width;
+					relativepos += vid.height;
 				}
 				// END DRAWING LOOP
 			}
@@ -314,7 +314,7 @@ void F_WipeStartScreen(void)
 	}
 #endif
 	wipe_scr_start = vid.screens[3];
-	I_ReadScreen(wipe_scr_start, 1);
+	M_Memcpy(wipe_scr_start, vid.screens[0], vid.width*vid.height);
 #endif
 }
 
@@ -331,7 +331,7 @@ void F_WipeEndScreen(void)
 	}
 #endif
 	wipe_scr_end = vid.screens[4];
-	I_ReadScreen(wipe_scr_end, 1);
+	M_Memcpy(wipe_scr_end, vid.screens[0], vid.width*vid.height);
 	V_DrawBlock(0, 0, 0, vid.width, vid.height, wipe_scr_start);
 #endif
 }

src/hu_stuff.c

@@ -2540,17 +2540,91 @@ Ping_conversion (UINT32 lag)
 // HU_drawPing
 //
 
+static void HU_drawOldPing(INT32 x, INT32 y, UINT32 lag, INT32 flags, boolean gentleman)
+{
+	const INT32 measureid = cv_pingmeasurement.value ? 1 : 0;
+
+	if (vid.width >= 640)	// how sad, we're using a shit resolution.
+	{
+		if (measureid == 1)
+		{
+			V_DrawRightAlignedSmallString(x+12, y+13, V_ALLOWLOWERCASE|flags, va("%dms", Ping_conversion(lag)));
+		}
+		else if (measureid == 0)
+		{
+			V_DrawRightAlignedSmallString(x+12, y+13, flags, va("d%d", Ping_conversion(lag)));
+		}
+	}
+
+	if (cv_pingicon.value)
+	{
+		UINT8 numbars = 0; // how many ping bars do we draw?
+		UINT8 barcolor = 31; // color we use for the bars (green, yellow, red or black)
+		SINT8 i = 0;
+		SINT8 yoffset = 6;
+
+		switch (lag)
+		{
+			case 0 ... 1:
+				numbars = 3;
+				barcolor = 215; // Blue
+				break;
+			case 2 ... 3:
+				numbars = 3;
+				barcolor = 184; // Green
+				break;
+			case 4 ... 6:
+				numbars = 2;    // Apparently ternaries w/ multiple statements don't look good in C so I decided against it.
+				barcolor = 103; // Yellow
+				break;
+			case 7 ... 9:
+				numbars = 1;
+				barcolor = 155; // Red
+				break;
+			default:            // Brazil
+				numbars = 0;
+				barcolor = 31;  // black
+				break;
+		}
+
+		if (gentleman)
+		{
+			barcolor = 194; // make it purplish
+			// bars get indirectly set earlier
+		}
+
+		for (i = 0; (i < 3); i++) // Draw the ping bar
+		{
+			V_DrawFill(x+2 *(i-1)+7, y+8+yoffset-4, 2, 8-yoffset, 31|flags);
+			if (i < numbars)
+				V_DrawFill(x+2 *(i-1)+7, y+8+yoffset-3, 1, 8-yoffset-1, barcolor|flags);
+
+			yoffset -= 2;
+		}
+	}
+}
+
+static void HU_drawKartPing(INT32 x, INT32 y, UINT32 lag, INT32 flags, boolean gentleman)
+{
+	const INT32 measureid = cv_pingmeasurement.value ? 1 : 0;
+
+	if (measureid == 1)
+		V_DrawScaledPatch(x+11 - pingmeasure[measureid]->width, y+9, flags, pingmeasure[measureid]);
+
+	if (cv_pingicon.value)
+	{
+		INT32 gfxnum = Ping_gfx_num(lag); // gfx to draw
+		V_DrawScaledPatch(x+2, y, flags, pinggfx[gfxnum]);
+	}
+
+	x = V_DrawPingNum(x + (measureid == 1 ? 11 - pingmeasure[measureid]->width : 10), y+9, flags, Ping_conversion(lag), Ping_gfx_colormap(lag, gentleman));
+
+	if (measureid == 0)
+		V_DrawScaledPatch(x+1 - pingmeasure[measureid]->width, y+9, flags, pingmeasure[measureid]);
+}
+
 void HU_drawPlayerPing(INT32 x, INT32 y, INT32 pnum, INT32 flags)
 {
-	INT32 measureid = cv_pingmeasurement.value ? 1 : 0;
-	INT32 gfxnum; // gfx to draw
-
-	//SRB2/Kart v1.0 style
-	UINT8 numbars = 0; // how many ping bars do we draw?
-	UINT8 barcolor = 31; // color we use for the bars (green, yellow, red or black)
-	SINT8 i = 0;
-	SINT8 yoffset = 6;
-
 	UINT32 lag = playerpingtable[pnum];
 	const boolean gentleman = (cv_mindelay.value && (lag < (tic_t)simulated_lag));
 
@@ -2561,74 +2635,11 @@ void HU_drawPlayerPing(INT32 x, INT32 y, INT32 pnum, INT32 flags)
 
 	if (cv_pingstyle.value == 0) // kart
 	{
-		gfxnum = Ping_gfx_num(lag);
-
-		if (measureid == 1)
-			V_DrawScaledPatch(x+11 - pingmeasure[measureid]->width, y+9, flags, pingmeasure[measureid]);
-
-		if (cv_pingicon.value)
-			V_DrawScaledPatch(x+2, y, flags, pinggfx[gfxnum]);
-
-		x = V_DrawPingNum(x + (measureid == 1 ? 11 - pingmeasure[measureid]->width : 10), y+9, flags, Ping_conversion(lag), Ping_gfx_colormap(lag, gentleman));
-
-		if (measureid == 0)
-			V_DrawScaledPatch(x+1 - pingmeasure[measureid]->width, y+9, flags, pingmeasure[measureid]);
+		HU_drawKartPing(x, y, lag, flags, gentleman);
 	}
 	else if (cv_pingstyle.value == 1) // old style ping
 	{
-		if (vid.width >= 640)	// how sad, we're using a shit resolution.
-		{
-			if (measureid == 1)
-			{
-				V_DrawRightAlignedSmallString(x+12, y+13, V_ALLOWLOWERCASE|flags, va("%dms", Ping_conversion(lag)));
-			}
-			else if (measureid == 0)
-			{
-				V_DrawRightAlignedSmallString(x+12, y+13, flags, va("d%d", Ping_conversion(lag)));
-			}
-		}
-
-		if (cv_pingicon.value)
-		{
-			switch (lag)
-			{
-				case 0 ... 1:
-					numbars = 3;
-					barcolor = 215; // Blue
-					break;
-				case 2 ... 3:
-					numbars = 3;
-					barcolor = 184; // Green
-					break;
-				case 4 ... 6:
-					numbars = 2;    // Apparently ternaries w/ multiple statements don't look good in C so I decided against it.
-					barcolor = 103; // Yellow
-					break;
-				case 7 ... 9:
-					numbars = 1;
-					barcolor = 155; // Red
-					break;
-				default:            // Brazil
-					numbars = 0;
-					barcolor = 31;  // black
-					break;
-			}
-
-			if (gentleman)
-			{
-				barcolor = 194; // make it purplish
-				// bars get indirectly set earlier
-			}
-
-			for (i = 0; (i < 3); i++) // Draw the ping bar
-			{
-				V_DrawFill(x+2 *(i-1)+7, y+8+yoffset-4, 2, 8-yoffset, 31|flags);
-				if (i < numbars)
-					V_DrawFill(x+2 *(i-1)+7, y+8+yoffset-3, 1, 8-yoffset-1, barcolor|flags);
-
-				yoffset -= 2;
-			}
-		}
+		HU_drawOldPing(x, y, lag, flags, gentleman);
 	}
 }
 

src/k_stats.c

@@ -1,3 +1,14 @@
+// SONIC ROBO BLAST 2 KART SATURN
+//-----------------------------------------------------------------------------
+// Copyright (C) 2025 by Indev.
+//
+// This program is free software distributed under the
+// terms of the GNU General Public License, version 2.
+// See the 'LICENSE' file for more details.
+//-----------------------------------------------------------------------------
+/// \file  k_stats.c
+/// \brief Game statistics handling
+
 #include "k_stats.h"
 #include "doomstat.h"
 #include "g_game.h"

src/k_stats.h

@@ -1,3 +1,14 @@
+// SONIC ROBO BLAST 2 KART SATURN
+//-----------------------------------------------------------------------------
+// Copyright (C) 2025 by Indev.
+//
+// This program is free software distributed under the
+// terms of the GNU General Public License, version 2.
+// See the 'LICENSE' file for more details.
+//-----------------------------------------------------------------------------
+/// \file  k_stats.h
+/// \brief Game statistics handling
+
 #ifndef __K_STATS__
 #define __K_STATS__
 

src/lua_udatalib.c

@@ -1,3 +1,14 @@
+// SONIC ROBO BLAST 2 KART SATURN
+//-----------------------------------------------------------------------------
+// Copyright (C) 2025 by Indev.
+//
+// This program is free software distributed under the
+// terms of the GNU General Public License, version 2.
+// See the 'LICENSE' file for more details.
+//-----------------------------------------------------------------------------
+/// \file  lua_udatalib.c
+/// \brief Implementation of lua userdata object metatables based on getter/setter functions
+
 #include "doomtype.h"
 #include "m_fixed.h"
 #include "lua_udatalib.h"

src/lua_udatalib.h

@@ -1,3 +1,14 @@
+// SONIC ROBO BLAST 2 KART SATURN
+//-----------------------------------------------------------------------------
+// Copyright (C) 2025 by Indev.
+//
+// This program is free software distributed under the
+// terms of the GNU General Public License, version 2.
+// See the 'LICENSE' file for more details.
+//-----------------------------------------------------------------------------
+/// \file  lua_udatalib.h
+/// \brief Implementation of lua userdata object metatables based on getter/setter functions
+
 #ifndef __LUA_UDATALIB_H__
 #define __LUA_UDATALIB_H__
 
@@ -100,4 +111,4 @@ void udatalib_addfields(lua_State *L, int mt, const udata_field_t fields[]);
     return 0; \
 }
 
-#endif
+#endif

src/m_curl.c

@@ -1,3 +1,14 @@
+// SONIC ROBO BLAST 2 KART SATURN
+//-----------------------------------------------------------------------------
+// Copyright (C) 2025 by Indev.
+//
+// This program is free software distributed under the
+// terms of the GNU General Public License, version 2.
+// See the 'LICENSE' file for more details.
+//-----------------------------------------------------------------------------
+/// \file  m_curl.c
+/// \brief Common curl-related functions shared across http-mserv and file download
+
 #include "m_curl.h"
 #include "doomtype.h"
 #include "i_system.h" // I_OutputMsg

src/m_curl.h

@@ -1,3 +1,14 @@
+// SONIC ROBO BLAST 2 KART SATURN
+//-----------------------------------------------------------------------------
+// Copyright (C) 2025 by Indev.
+//
+// This program is free software distributed under the
+// terms of the GNU General Public License, version 2.
+// See the 'LICENSE' file for more details.
+//-----------------------------------------------------------------------------
+/// \file  m_curl.h
+/// \brief Common curl-related functions shared across http-mserv and file download
+
 #ifndef __M_CURL__
 #define __M_CURL__
 

src/m_emotes.cpp

@@ -1,3 +1,14 @@
+// SONIC ROBO BLAST 2 KART SATURN
+//-----------------------------------------------------------------------------
+// Copyright (C) 2025 by Indev.
+//
+// This program is free software distributed under the
+// terms of the GNU General Public License, version 2.
+// See the 'LICENSE' file for more details.
+//-----------------------------------------------------------------------------
+/// \file  m_emotes.cpp
+/// \brief Chat emotes handling
+
 #include <map>
 #include <string>
 #include <sstream>

src/m_emotes.h

@@ -1,3 +1,14 @@
+// SONIC ROBO BLAST 2 KART SATURN
+//-----------------------------------------------------------------------------
+// Copyright (C) 2025 by Indev.
+//
+// This program is free software distributed under the
+// terms of the GNU General Public License, version 2.
+// See the 'LICENSE' file for more details.
+//-----------------------------------------------------------------------------
+/// \file  m_emotes.h
+/// \brief Chat emotes handling
+
 #ifndef __M_EMOTES__
 #define __M_EMOTES__
 

src/m_menu.c

@@ -2713,6 +2713,10 @@ static void M_PrepareLevelSelect(void)
 //
 boolean M_CanShowLevelInList(INT32 mapnum, INT32 gt)
 {
+	// invalid mapnum
+	if (mapnum < -1)
+		return false;
+
 	// Random map!
 	if (mapnum == -1)
 		return (gamestate != GS_TIMEATTACK && !modeattacking);
@@ -7220,7 +7224,7 @@ static void M_DrawLevelSelectOnly(boolean leftfade, boolean rightfade)
 			if (i == oldval)
 				return;
 
-			if(!mapheaderinfo[i])
+			if (i >= 0 && !mapheaderinfo[i])
 				continue; // Don't allocate the header.  That just makes memory usage skyrocket.
 
 		} while (!M_CanShowLevelInList(i, cv_newgametype.value));
@@ -7258,7 +7262,7 @@ static void M_DrawLevelSelectOnly(boolean leftfade, boolean rightfade)
 			if (i == oldval)
 				return;
 
-			if(!mapheaderinfo[i])
+			if (i >= 0 && !mapheaderinfo[i])
 				continue; // Don't allocate the header.  That just makes memory usage skyrocket.
 
 		} while (!M_CanShowLevelInList(i, cv_newgametype.value));

src/m_textinput.c

@@ -1,3 +1,14 @@
+// SONIC ROBO BLAST 2 KART SATURN
+//-----------------------------------------------------------------------------
+// Copyright (C) 2025 by Indev.
+//
+// This program is free software distributed under the
+// terms of the GNU General Public License, version 2.
+// See the 'LICENSE' file for more details.
+//-----------------------------------------------------------------------------
+/// \file  m_textinput.c
+/// \brief User text input implementation
+
 #include "m_textinput.h"
 #include "m_menu.h" // MAXSTRINGLENGTH
 #include "v_video.h"

src/m_textinput.h

@@ -1,3 +1,14 @@
+// SONIC ROBO BLAST 2 KART SATURN
+//-----------------------------------------------------------------------------
+// Copyright (C) 2025 by Indev.
+//
+// This program is free software distributed under the
+// terms of the GNU General Public License, version 2.
+// See the 'LICENSE' file for more details.
+//-----------------------------------------------------------------------------
+/// \file  m_textinput.h
+/// \brief User text input implementation
+
 #ifndef __M_TEXTINPUT__
 #define __M_TEXTINPUT__
 

src/r_data.c

@@ -94,7 +94,7 @@ static UINT32 **texturecolumnofs; // column offset lookup table for each texture
 UINT8 **texturecache; // graphics data for each generated full-size texture
 
 // texture width is a power of 2, so it can easily repeat along sidedefs using a simple mask
-static INT32 *texturewidthmask;
+static INT32 *texturewidth;
 
 fixed_t *textureheight; // needed for texture pegging
 
@@ -241,11 +241,12 @@ static INT32 tidcachelen = 0;
 // R_DrawColumnInCache
 // Clip and draw a column from a patch into a cached post.
 //
-static inline void R_DrawColumnInCache(column_t *patch, UINT8 *cache, INT32 originy, INT32 cacheheight)
+static inline void R_DrawColumnInCache(column_t *patch, UINT8 *cache, texpatch_t *originPatch, INT32 cacheheight)
 {
 	INT32 count, position;
 	UINT8 *source;
 	INT32 topdelta, prevdelta = -1;
+	INT32 originy = originPatch->originy;
 
 	while (patch->topdelta != 0xff)
 	{
@@ -260,6 +261,7 @@ static inline void R_DrawColumnInCache(column_t *patch, UINT8 *cache, INT32 orig
 		if (position < 0)
 		{
 			count += position;
+			source -= position; // start further down the column
 			position = 0;
 		}
 
@@ -273,6 +275,50 @@ static inline void R_DrawColumnInCache(column_t *patch, UINT8 *cache, INT32 orig
 	}
 }
 
+static UINT8 *R_AllocateTextureBlock(size_t blocksize, UINT8 **user)
+{
+	texturememory += blocksize;
+
+	return Z_Malloc(blocksize, PU_LEVEL, user);
+}
+
+static UINT8 *R_AllocateDummyTextureBlock(size_t width, UINT8 **user)
+{
+	// Allocate dummy data. Keep 4-bytes aligned.
+	// Column offsets will be initialized to 0, which points to the 0xff byte (empty column flag).
+	size_t blocksize = 4 + (width * 4);
+	UINT8 *block = R_AllocateTextureBlock(blocksize, user);
+
+	memset(block, 0, blocksize);
+	block[0] = 0xff;
+
+	return block;
+}
+
+static boolean R_CheckTextureLumpLength(texture_t *texture, size_t patch)
+{
+	UINT16 wadnum = texture->patches[patch].wad;
+	UINT16 lumpnum = texture->patches[patch].lump;
+	size_t lumplength = W_LumpLengthPwad(wadnum, lumpnum);
+
+	// The header does not exist
+	if (lumplength < offsetof(softwarepatch_t, columnofs))
+	{
+		CONS_Alert(
+			CONS_ERROR,
+			 "%.8s: texture lump data is too small. Expected %s bytes, got %s. (%s)\n",
+				   texture->name,
+			 sizeu1(offsetof(softwarepatch_t, columnofs)),
+				   sizeu2(lumplength),
+				   wadfiles[wadnum]->lumpinfo[lumpnum].fullname
+		);
+
+		return false;
+	}
+
+	return true;
+}
+
 //
 // R_GenerateTexture
 //
@@ -297,6 +343,10 @@ UINT8 *R_GenerateTexture(size_t texnum)
 	column_t *patchcol;
 	UINT8 *colofs;
 
+	UINT16 wadnum;
+	lumpnum_t lumpnum;
+	size_t lumplength;
+
 	I_Assert(texnum <= (size_t)numtextures);
 	texture = textures[texnum];
 	I_Assert(texture != NULL);
@@ -311,45 +361,57 @@ UINT8 *R_GenerateTexture(size_t texnum)
 	{
 		boolean holey = false;
 		patch = texture->patches;
-		pdata = W_CacheLumpNumPwad(patch->wad, patch->lump, PU_LEVEL);
+
+		wadnum = patch->wad;
+		lumpnum = patch->lump;
+		lumplength = W_LumpLengthPwad(wadnum, lumpnum);
+
+		// The header does not exist
+		if (R_CheckTextureLumpLength(texture, 0) == false)
+		{
+			block = R_AllocateDummyTextureBlock(texture->width, &texturecache[texnum]);
+			texturecolumnofs[texnum] = (UINT32*)&block[4];
+			textures[texnum]->holes = true;
+			return block;
+		}
+
+		pdata = (UINT8*)W_CacheLumpNumPwad(wadnum, lumpnum, PU_LEVEL);
 		realpatch = (softwarepatch_t *)pdata;
 
 		// Check the patch for holes.
 		if (texture->width > SHORT(realpatch->width) || texture->height > SHORT(realpatch->height))
 			holey = true;
-		else
+
+		colofs = (UINT8 *)realpatch->columnofs;
+
+		for (x = 0; x < texture->width && !holey; x++)
 		{
-			colofs = (UINT8 *)realpatch->columnofs;
+			column_t *col = (column_t *)((UINT8 *)realpatch + LONG(*(UINT32 *)&colofs[x<<2]));
 
-			for (x = 0; x < texture->width; x++)
+			INT32 topdelta, prevdelta = -1, y = 0;
+
+			while (col->topdelta != 0xff)
 			{
-				column_t *col = (column_t *)((UINT8 *)realpatch + LONG(*(UINT32 *)&colofs[x<<2]));
-
-				INT32 topdelta, prevdelta = -1, y = 0;
-
-				while (col->topdelta != 0xff)
-				{
-					topdelta = col->topdelta;
-					if (topdelta <= prevdelta)
-						topdelta += prevdelta;
-					prevdelta = topdelta;
-					if (topdelta > y)
-						break;
-					y = topdelta + col->length + 1;
-					col = (column_t *)((UINT8 *)col + col->length + 4);
-				}
-
-				if (y < texture->height)
-					holey = true; // this texture is HOLEy! D:
+				topdelta = col->topdelta;
+				if (topdelta <= prevdelta)
+					topdelta += prevdelta;
+				prevdelta = topdelta;
+				if (topdelta > y)
+					break;
+				y = topdelta + col->length + 1;
+				col = (column_t *)((UINT8 *)col + col->length + 4);
 			}
+
+			if (y < texture->height)
+				holey = true; // this texture is HOLEy! D:
 		}
 
 		// If the patch uses transparency, we have to save it this way.
 		if (holey)
 		{
 			texture->holes = true;
-			blocksize = W_LumpLengthPwad(patch->wad, patch->lump);
-			block = Z_Calloc(blocksize, PU_LEVEL, &texturecache[texnum]);
+			blocksize = lumplength;
+			block = Z_Calloc(blocksize, PU_LEVEL, &texturecache[texnum]); // will change tag at end of this function
 			M_Memcpy(block, realpatch, blocksize);
 			texturememory += blocksize;
 
@@ -365,7 +427,6 @@ UINT8 *R_GenerateTexture(size_t texnum)
 
 			goto done;
 		}
-
 		// Otherwise, do multipatch format.
 	}
 
@@ -384,10 +445,18 @@ UINT8 *R_GenerateTexture(size_t texnum)
 	// texture data after the lookup table
 	blocktex = block + (texture->width*4);
 
+	for (x = 0; x < texture->width; ++x)
+	{
+		// generate column ofset lookup
+		*(UINT32 *)&colofs[x<<2] = LONG((x * texture->height) + (texture->width*4));
+	}
+
 	// Composite the columns together.
 	for (i = 0, patch = texture->patches; i < texture->patchcount; i++, patch++)
 	{
-		pdata = W_CacheLumpNumPwad(patch->wad, patch->lump, PU_LEVEL);
+		wadnum = patch->wad;
+		lumpnum = patch->lump;
+		pdata = (UINT8*)W_CacheLumpNumPwad(wadnum, lumpnum, PU_LEVEL);
 		realpatch = (softwarepatch_t *)pdata;
 
 		x1 = patch->originx;
@@ -418,9 +487,7 @@ UINT8 *R_GenerateTexture(size_t texnum)
 		{
 			patchcol = (column_t *)((UINT8 *)realpatch + LONG(realpatch->columnofs[x-x1]));
 
-			// generate column ofset lookup
-			*(UINT32 *)&colofs[x<<2] = LONG((x * texture->height) + (texture->width*4));
-			R_DrawColumnInCache(patchcol, block + LONG(*(UINT32 *)&colofs[x<<2]), patch->originy, texture->height);
+			R_DrawColumnInCache(patchcol, block + LONG(*(UINT32 *)&colofs[x<<2]), patch, texture->height);
 		}
 	}
 
@@ -439,6 +506,7 @@ INT32 R_GetTextureNum(INT32 texnum)
 {
 	if (texnum < 0 || texnum >= numtextures)
 		return 0;
+
 	return texturetranslation[texnum];
 }
 
@@ -456,17 +524,12 @@ void R_CheckTextureCache(INT32 tex)
 
 static inline INT32 wrap_column(fixed_t tex, INT32 col)
 {
-	if (texturewidthmask[tex])
-		col &= texturewidthmask[tex];  // set by load textures
-	else
-	{
-		const INT16 texwidth = textures[tex]->width;
+	INT32 width = texturewidth[tex];
 
-		// Odd width texture, cannot just mask.
-		// Sometime gets colnum = -1 or = width, even without tiling.
-		// Test LostCiv, Map 20, crates.
-		col = ((col % texwidth) + texwidth) % texwidth;
-	}
+	if (width & (width - 1))
+		col = (UINT32)col % width;
+	else
+		col &= (width - 1);
 
 	return col;
 }
@@ -476,13 +539,6 @@ static inline INT32 wrap_column(fixed_t tex, INT32 col)
 //
 UINT8 *R_GetColumn(fixed_t tex, INT32 col)
 {
-	if (!texturecache[tex])
-	{
-		// This must be here because cache can be freed by other operations.
-		// To prevent must lock individual texture cache on every draw.
-		R_GenerateTexture(tex);
-	}
-
 	return texturecache[tex] + LONG(texturecolumnofs[tex][wrap_column(tex, col)]);
 }
 
@@ -516,11 +572,10 @@ static INT32
 Rloadtextures (INT32 i, INT32 w)
 {
 	UINT16 j, numlumps = 0;
-	INT32 k;
 	UINT16 texstart, texend, texturesLumpPos;
-	softwarepatch_t *patchlump;
 	texpatch_t *patch;
 	texture_t *texture;
+	softwarepatch_t patchlump;
 
 	// Get the lump numbers for the markers in the WAD, if they exist.
 	if (W_FileHasFolders(wadfiles[w]))
@@ -560,7 +615,7 @@ Rloadtextures (INT32 i, INT32 w)
 				continue; // If it is then SKIP IT
 		}
 
-		patchlump = (softwarepatch_t *)W_CacheLumpNumPwad(wadnum, lumpnum, PU_STATIC);
+		W_ReadLumpHeaderPwad(wadnum, lumpnum, &patchlump, PNG_HEADER_SIZE, 0);
 
 		//CONS_Printf("\n\"%s\" is a single patch, dimensions %d x %d",W_CheckNameForNumPwad((wadnum, lumpnum), patchlump->width, patchlump->height);
 		texture = textures[i] = Z_Calloc(sizeof(texture_t) + sizeof(texpatch_t), PU_STATIC, NULL);
@@ -568,8 +623,10 @@ Rloadtextures (INT32 i, INT32 w)
 		// Set texture properties.
 		M_Memcpy(texture->name, W_CheckNameForNumPwad(wadnum, lumpnum), sizeof(texture->name));
 		texture->hash = quickncasehash(texture->name, 8);
-		texture->width = SHORT(patchlump->width);
-		texture->height = SHORT(patchlump->height);
+
+		texture->width = SHORT(patchlump.width);
+		texture->height = SHORT(patchlump.height);
+
 		texture->type = TEXTURETYPE_SINGLEPATCH;
 		texture->patchcount = 1;
 		texture->holes = false;
@@ -581,27 +638,7 @@ Rloadtextures (INT32 i, INT32 w)
 		patch->wad = wadnum;
 		patch->lump = texstart + j;
 
-		Z_Free(patchlump);
-
-		// determine width power of 2
-#if 1
-        // [WDJ] only need to determine if exact power of 2.
-        k = 1;
-        while (k < texture->width)
-            k<<=1;
-#else
-		// Largest power of 2 that fits within width.
-		k = 1;
-		while (k << 1 <= texture->width)
-			k <<= 1;
-#endif
-		if (k != texture->width)
-		{
-			// Odd width
-			k = 1;  // make texturewidthmask = 0
-		}
-
-		texturewidthmask[i] = k - 1;
+		texturewidth[i] = texture->width;
 		textureheight[i] = texture->height << FRACBITS;
 		i++;
 	}
@@ -709,7 +746,7 @@ static void R_AllocateTextures(INT32 add)
 	// Allocate texture referencing cache.
 	recallocuser(&texturecache, oldsize, newsize);
 	// Allocate texture width table.
-	recallocuser(&texturewidthmask, oldsize, newsize);
+	recallocuser(&texturewidth, oldsize, newsize);
 	// Allocate texture height table.
 	recallocuser(&textureheight, oldsize, newsize);
 	// Create translation table for global animation.
@@ -1181,7 +1218,7 @@ void R_ParseTEXTURESLump(UINT16 wadNum, UINT16 lumpNum, INT32 *texindex)
 			newTexture = R_ParseTexture(true);
 			// Store the new texture
 			textures[*texindex] = newTexture;
-			texturewidthmask[*texindex] = newTexture->width - 1;
+			texturewidth[*texindex] = newTexture->width;
 			textureheight[*texindex] = newTexture->height << FRACBITS;
 			// Increment i back in R_LoadTextures()
 			(*texindex)++;

src/r_defs.h

@@ -615,12 +615,12 @@ typedef struct drawseg_s
 	// Pointers to lists for sprite clipping, all three adjusted so [x1] is first value.
 	INT16 *sprtopclip;
 	INT16 *sprbottomclip;
-	INT16 *maskedtexturecol;
+	fixed_t *maskedtexturecol;
 
 	struct visplane_s *ffloorplanes[MAXFFLOORS];
 	INT32 numffloorplanes;
 	struct ffloor_s *thicksides[MAXFFLOORS];
-	INT16 *thicksidecol;
+	fixed_t *thicksidecol;
 	INT32 numthicksides;
 	fixed_t *frontscale;
 

src/r_draw.h

@@ -30,7 +30,7 @@ extern INT32 linesize;
 
 FUNCINLINE static ATTRINLINE UINT8 *R_Address(INT32 px, INT32 py)
 {
-	return renderscreen + (py + viewwindowy) * linesize + (viewwindowx + px);
+	return renderscreen + (py + viewwindowy) + (viewwindowx + px) * vid.height;
 }
 
 

src/r_draw_column.cpp

@@ -31,7 +31,8 @@ enum DrawColumnType
 };
 
 template<DrawColumnType Type>
-static constexpr UINT8 R_GetColumnTranslated(drawcolumndata_t* dc, UINT8 col, const UINT8 * restrict colormap)
+FUNCINLINE static ATTRINLINE constexpr UINT8
+R_GetColumnTranslated(drawcolumndata_t* dc, UINT8 col, const UINT8 * restrict colormap)
 {
 	if constexpr (Type & DrawColumnType::DC_COLORMAP)
 	{
@@ -42,7 +43,8 @@ static constexpr UINT8 R_GetColumnTranslated(drawcolumndata_t* dc, UINT8 col, co
 }
 
 template<DrawColumnType Type>
-static constexpr UINT8 R_GetColumnTranslucent(drawcolumndata_t* dc, UINT8 * restrict dest, UINT8 col, const UINT8 * restrict colormap)
+FUNCINLINE static ATTRINLINE constexpr UINT8
+R_GetColumnTranslucent(drawcolumndata_t* dc, UINT8 * restrict dest, UINT8 col, const UINT8 * restrict colormap)
 {
 	col = R_GetColumnTranslated<Type>(dc, col, colormap);
 
@@ -57,7 +59,8 @@ static constexpr UINT8 R_GetColumnTranslucent(drawcolumndata_t* dc, UINT8 * rest
 }
 
 template<DrawColumnType Type>
-static constexpr UINT8 R_DrawColumnPixel(drawcolumndata_t* dc, UINT8 * restrict dest, UINT32 bit, const UINT8 * restrict source, const UINT8 * restrict colormap)
+FUNCINLINE static ATTRINLINE constexpr UINT8
+R_DrawColumnPixel(drawcolumndata_t* dc, UINT8 * restrict dest, UINT32 bit, const UINT8 * restrict source, const UINT8 * restrict colormap)
 {
 	UINT8 col = source[bit];
 
@@ -187,7 +190,6 @@ static void R_DrawColumnTemplate(drawcolumndata_t *dc)
 		const intptr_t heightmask = dc->sourcelength-1;
 		constexpr INT32 npow2min = -1;
 		const INT32 npow2max = dc->sourcelength;
-		const INT32 stride = vid.width;
 
 		const UINT8 * restrict source = dc->source;
 		const lighttable_t * restrict colormap = dc->colormap;
@@ -208,12 +210,12 @@ static void R_DrawColumnTemplate(drawcolumndata_t *dc)
 					while(count--)
 					{
 						*dest = R_DrawColumnPixel<Type>(dc, dest, (frac>>FRACBITS) & heightmask, source, colormap);
-						dest += stride;
+						dest++;
 						frac += fracstep;
 					}
 				}
 				break;
-			case -1:
+			case npow2min:
 				{
 					if (frac < 0)
 						// adjust in case we underread
@@ -223,7 +225,7 @@ static void R_DrawColumnTemplate(drawcolumndata_t *dc)
 					while (--count >= 0)
 					{
 						*dest = R_DrawColumnPixel<Type>(dc, dest, frac>>FRACBITS, source, colormap);
-						dest += stride;
+						dest++;
 						frac += fracstep;
 					}
 				}
@@ -235,11 +237,11 @@ static void R_DrawColumnTemplate(drawcolumndata_t *dc)
 						while ((count -= 2) >= 0) // texture height is a power of 2 -- killough
 						{
 							*dest = R_DrawColumnPixel<Type>(dc, dest, (frac>>FRACBITS) & heightmask, source, colormap);
-							dest += stride;
+							dest++;
 							frac += fracstep;
 
 							*dest = R_DrawColumnPixel<Type>(dc, dest, (frac>>FRACBITS) & heightmask, source, colormap);
-							dest += stride;
+							dest++;
 							frac += fracstep;
 						}
 
@@ -277,7 +279,7 @@ static void R_DrawColumnTemplate(drawcolumndata_t *dc)
 							// and a few bytes after as well
 							*dest = R_DrawColumnPixel<Type>(dc, dest, CLAMP((frac >> FRACBITS), npow2min, npow2max), source, colormap);
 
-							dest += stride;
+							dest++;
 
 #if __SIZEOF_POINTER__ < 8  // 64-bit systems have large enough numbers for this to be a non-issue
 							// Avoid overflow.
@@ -343,7 +345,6 @@ void R_DrawFogColumn(drawcolumndata_t* dc)
 	// Framebuffer destination address.
 	dest = R_Address(dc->x, dc->yl);
 
-	const INT32 stride = vid.width;
 	const lighttable_t * restrict colormap = dc->colormap;
 
 	// Determine scaling, which is the only mapping to be done.
@@ -351,6 +352,6 @@ void R_DrawFogColumn(drawcolumndata_t* dc)
 	{
 		// Simple. Apply the colormap to what's already on the screen.
 		*dest = colormap[*dest];
-		dest += stride;
+		dest++;
 	} while (count--);
 }

src/r_draw_span.cpp

@@ -25,6 +25,9 @@
 #define local_for_thread static
 #endif
 
+#include <vector>
+#include <algorithm>
+
 // ==========================================================================
 // SPANS
 // ==========================================================================
@@ -47,7 +50,8 @@ enum DrawSpanType
 };
 
 template<DrawSpanType Type>
-static constexpr UINT8 R_GetSpanTranslated(drawspandata_t* ds, UINT8 col)
+FUNCINLINE static ATTRINLINE constexpr UINT8
+R_GetSpanTranslated(drawspandata_t* ds, UINT8 col)
 {
 	if constexpr (Type & DrawSpanType::DS_COLORMAP)
 	{
@@ -60,7 +64,8 @@ static constexpr UINT8 R_GetSpanTranslated(drawspandata_t* ds, UINT8 col)
 }
 
 template<DrawSpanType Type>
-static constexpr UINT8 R_GetSpanTranslucent(drawspandata_t* ds, UINT8 *dsrc, const UINT8 *colormap, UINT8 col)
+FUNCINLINE static ATTRINLINE constexpr UINT8
+R_GetSpanTranslucent(drawspandata_t* ds, UINT8 *dsrc, const UINT8 *colormap, UINT8 col)
 {
 	col = colormap[R_GetSpanTranslated<Type>(ds, col)];
 
@@ -75,7 +80,8 @@ static constexpr UINT8 R_GetSpanTranslucent(drawspandata_t* ds, UINT8 *dsrc, con
 }
 
 template<DrawSpanType Type>
-static constexpr UINT8 R_DrawSpanPixel(drawspandata_t* ds, UINT8 *dsrc, const UINT8 *colormap, UINT32 bit, const UINT8 *source)
+FUNCINLINE static ATTRINLINE constexpr UINT8
+R_DrawSpanPixel(drawspandata_t* ds, UINT8 *dsrc, const UINT8 *colormap, UINT32 bit, const UINT8 *source)
 {
 	UINT8 col = source[bit];
 
@@ -104,15 +110,17 @@ static void R_DrawSpanTemplate(drawspandata_t* ds)
 	UINT8 * restrict dest = R_Address(ds->x1, ds->y);
 	UINT8 * restrict dsrc;
 
-	const UINT8 * restrict deststop = vid.screens[0] + vid.width * vid.height;
+	const UINT8 * restrict deststop = vid.screens[0] + vid.height *vid.width;
 
-	if (dest+8 > deststop)
+	if (dest > deststop)
 	{
 		return;
 	}
 
 	intptr_t count = (ds->x2 - ds->x1 + 1);
-	size_t i;
+	//size_t i;
+
+	const INT32 stride = vid.height;
 
 	xposition = ds->xfrac; yposition = ds->yfrac;
 	xstep = ds->xstep; ystep = ds->ystep;
@@ -137,14 +145,14 @@ static void R_DrawSpanTemplate(drawspandata_t* ds)
 
 	if constexpr (Type & DS_RIPPLE)
 	{
-		dsrc = vid.screens[1] + (ds->y + ds->bgofs) * vid.width + ds->x1;
+		dsrc = vid.screens[1] + (ds->y + ds->bgofs) + ds->x1 * vid.height;
 	}
 	else
 	{
 		dsrc = dest;
 	}
 
-	while (count >= 8)
+	/*while (count >= 8)
 	{
 		// SoM: Why didn't I see this earlier? the spot variable is a waste now because we don't
 		// have the uber complicated math to calculate it now, so that was a memory write we didn't
@@ -164,7 +172,7 @@ static void R_DrawSpanTemplate(drawspandata_t* ds)
 		dsrc += 8;
 
 		count -= 8;
-	}
+	}*/
 
 	while (count-- && dest <= deststop)
 	{
@@ -172,8 +180,8 @@ static void R_DrawSpanTemplate(drawspandata_t* ds)
 
 		*dest = R_DrawSpanPixel<Type>(ds, dsrc, colormap, bit, source);
 
-		dest++;
-		dsrc++;
+		dest += stride;
+		dsrc += stride;
 
 		xposition += xstep;
 		yposition += ystep;
@@ -182,31 +190,39 @@ static void R_DrawSpanTemplate(drawspandata_t* ds)
 
 // R_CalcTiltedLighting
 // Exactly what it says on the tin. I wish I wasn't too lazy to explain things properly.
-static void R_CalcTiltedLighting(INT32 *lightbuffer, INT32 x1, INT32 x2, fixed_t start, fixed_t end)
+static void R_CalcTiltedLighting(std::vector<INT32>& lightbuffer, INT32 x1, INT32 x2, fixed_t start, fixed_t end)
 {
 	// ZDoom uses a different lighting setup to us, and I couldn't figure out how to adapt their version
 	// of this function. Here's my own.
-	INT32 i, left = x1, right = x2;
+	INT32 i;
 	const fixed_t step = (end-start)/(x2 - x1 + 1);
 
 	// I wanna do some optimizing by checking for out-of-range segments on either side to fill in all at once,
 	// but I'm too bad at coding to not crash the game trying to do that. I guess this is fast enough for now...
 
-	for (i = left; i <= right; i++)
+	for (i = x1; i <= x2; i++)
 	{
-		lightbuffer[i] = (start += step) >> FRACBITS;
-
-		if (lightbuffer[i] < 0)
-		{
-			lightbuffer[i] = 0;
-		}
-		else if (lightbuffer[i] >= MAXLIGHTSCALE)
-		{
-			lightbuffer[i] = MAXLIGHTSCALE-1;
-		}
+		fixed_t light = start >> FRACBITS;
+		lightbuffer[i] = CLAMP(light, 0, MAXLIGHTSCALE - 1);
+		start += step;
 	}
 }
 
+static void R_GetTiltedLighting(std::vector<INT32>& tiltlighting, const drawspandata_t* ds, const float iz, const int width, const INT32 stride)
+{
+	float planelightfloat = PLANELIGHTFLOAT;
+	const fixed_t lightstart = FloatToFixed(iz * planelightfloat);
+	const fixed_t lightend   = FloatToFixed((iz + ds->szp.x * width) * planelightfloat);
+
+	if (tiltlighting.size() != (size_t)viewwidth)
+	{
+		tiltlighting.resize(viewwidth);
+	}
+
+	R_CalcTiltedLighting(tiltlighting, ds->x1, ds->x2, lightstart, lightend);
+	//CONS_Printf("tilted lighting %f to %f (foc %f)\n", FixedToFloat(lightstart), FixedToFloat(lightend), focallengthf);
+}
+
 template<DrawSpanType Type>
 static void R_DrawTiltedSpanTemplate(drawspandata_t* ds)
 {
@@ -223,35 +239,18 @@ static void R_DrawTiltedSpanTemplate(drawspandata_t* ds)
 	float endz, endu, endv;
 	UINT32 stepu, stepv;
 	UINT32 bit;
-	local_for_thread INT32 *tiltlighting = NULL;
-	local_for_thread INT32 oldviewwidth = 0;
 
-	// dont realloc every frame pls thx
-	if (tiltlighting == NULL || oldviewwidth != viewwidth)
-	{
-		tiltlighting = static_cast<INT32*>(realloc(tiltlighting, sizeof(*tiltlighting) * viewwidth));
-		oldviewwidth = viewwidth;
-	}
-
-	INT32 x1 = ds->x1;
+	//INT32 x1 = ds->x1;
 	const INT32 nflatxshift = ds->nflatxshift;
 	const INT32 nflatyshift = ds->nflatyshift;
 	const INT32 nflatmask = ds->nflatmask;
-	const INT32 stride = vid.width;
+	const INT32 stride = vid.height;
+	local_for_thread std::vector<INT32> tiltlighting;
 
 	iz = ds->szp.z + ds->szp.y*(centery-ds->y) + ds->szp.x*(ds->x1-centerx);
 
 	// Lighting is simple. It's just linear interpolation from start to end
-	{
-		float planelightfloat = PLANELIGHTFLOAT;
-		float lightstart, lightend;
-
-		lightend = (iz + ds->szp.x*width) * planelightfloat;
-		lightstart = iz * planelightfloat;
-
-		R_CalcTiltedLighting(tiltlighting, ds->x1, ds->x2, FLOAT_TO_FIXED(lightstart), FLOAT_TO_FIXED(lightend));
-		//CONS_Printf("tilted lighting %f to %f (foc %f)\n", lightstart, lightend, focallengthf);
-	}
+	R_GetTiltedLighting(tiltlighting, ds, iz, width, stride);
 
 	uz = ds->sup.z + ds->sup.y*(centery-ds->y) + ds->sup.x*(ds->x1-centerx);
 	vz = ds->svp.z + ds->svp.y*(centery-ds->y) + ds->svp.x*(ds->x1-centerx);
@@ -261,7 +260,7 @@ static void R_DrawTiltedSpanTemplate(drawspandata_t* ds)
 
 	if constexpr (Type & DS_RIPPLE)
 	{
-		dsrc = vid.screens[1] + (ds->y + ds->bgofs) * stride + ds->x1;
+		dsrc = vid.screens[1] + (ds->y + ds->bgofs) + ds->x1 * stride;
 	}
 	else
 	{
@@ -313,18 +312,31 @@ static void R_DrawTiltedSpanTemplate(drawspandata_t* ds)
 		u = (INT64)(startu);
 		v = (INT64)(startv);
 
-		x1 = ds->x1;
+		//x1 = ds->x1;
 
-		for (i = 0; i < SPANSIZE; i++)
+		// not sure which one does a better job
+
+		for (i = SPANSIZE-1; i >= 0; i--)
+		{
+			bit = ((v >> nflatyshift) & nflatmask) | (u >> nflatxshift);
+			colormap = ds->planezlight[tiltlighting[ds->x1++]] + (ds->colormap - colormaps);
+			*dest = R_DrawSpanPixel<Type>(ds, dsrc, colormap, bit, source);
+			dest += stride;
+			dsrc += stride;
+			u += stepu;
+			v += stepv;
+		}
+
+		/*for (i = 0; i < SPANSIZE; i++)
 		{
 			bit = (((v + stepv * i) >> nflatyshift) & nflatmask) | ((u + stepu * i) >> nflatxshift);
 			colormap = ds->planezlight[tiltlighting[x1 + i]] + (ds->colormap - colormaps);
-			dest[i] = R_DrawSpanPixel<Type>(ds, &dsrc[i], colormap, bit, source);
+			dest[i * stride] = R_DrawSpanPixel<Type>(ds, &dsrc[i * stride], colormap, bit, source);
 		}
 
 		ds->x1 += SPANSIZE;
-		dest += SPANSIZE;
-		dsrc += SPANSIZE;
+		dest += SPANSIZE * stride;
+		dsrc += SPANSIZE * stride;*/
 		startu = endu;
 		startv = endv;
 		width -= SPANSIZE;
@@ -362,9 +374,9 @@ static void R_DrawTiltedSpanTemplate(drawspandata_t* ds)
 				bit = ((v >> ds->nflatyshift) & ds->nflatmask) | (u >> ds->nflatxshift);
 				colormap = ds->planezlight[tiltlighting[ds->x1]] + (ds->colormap - colormaps);
 				*dest = R_DrawSpanPixel<Type>(ds, dsrc, colormap, bit, source);
-				dest++;
+				dest += stride;
+				dsrc += stride;
 				ds->x1++;
-				dsrc++;
 				u += stepu;
 				v += stepv;
 			}
@@ -400,7 +412,9 @@ void R_DrawFogSpan(drawspandata_t* ds)
 
 	intptr_t count = ds->x2 - ds->x1 + 1;
 
-	while (count >= 4)
+	const INT32 stride = vid.height;
+
+	/*while (count >= 4)
 	{
 		dest[0] = colormap[dest[0]];
 		dest[1] = colormap[dest[1]];
@@ -409,12 +423,12 @@ void R_DrawFogSpan(drawspandata_t* ds)
 
 		dest += 4;
 		count -= 4;
-	}
+	}*/
 
 	while (count--)
 	{
 		*dest = colormap[*dest];
-		dest++;
+		dest += stride;
 	}
 }
 
@@ -423,37 +437,19 @@ void R_DrawFogSpan_Tilted(drawspandata_t* ds)
 	int width = ds->x2 - ds->x1;
 	float iz = ds->szp.z + ds->szp.y*(centery-ds->y) + ds->szp.x*(ds->x1-centerx);
 	UINT8 * restrict dest;
-
-	local_for_thread INT32 *tiltlighting = NULL;
-	local_for_thread INT32 oldviewwidth = 0;
-
-	// dont realloc every frame pls thx
-	if (tiltlighting == NULL || oldviewwidth != viewwidth)
-	{
-		tiltlighting = static_cast<INT32*>(realloc(tiltlighting, sizeof(*tiltlighting) * viewwidth));
-		oldviewwidth = viewwidth;
-	}
+	local_for_thread std::vector<INT32> tiltlighting;
 
 	dest = R_Address(ds->x1, ds->y);
 	const INT32 stride = vid.width;
 
 	// Lighting is simple. It's just linear interpolation from start to end
-	{
-		float planelightfloat = PLANELIGHTFLOAT;
-		float lightstart, lightend;
-
-		lightend = (iz + ds->szp.x*width) * planelightfloat;
-		lightstart = iz * planelightfloat;
-
-		R_CalcTiltedLighting(tiltlighting, ds->x1, ds->x2, FLOAT_TO_FIXED(lightstart), FLOAT_TO_FIXED(lightend));
-		//CONS_Printf("tilted lighting %f to %f (foc %f)\n", lightstart, lightend, focallengthf);
-	}
+	R_GetTiltedLighting(tiltlighting, ds, iz, width, stride);
 
 	do
 	{
 		UINT8 *colormap = ds->planezlight[tiltlighting[ds->x1++]] + (ds->colormap - colormaps);
 		*dest = colormap[*dest];
-		dest++;
+		dest += stride;
 	}
 	while (--width >= 0);
 }

src/r_main.cpp

@@ -1400,6 +1400,7 @@ void R_RenderPlayerView(player_t *player)
 
 		R_ClearClipSegs();
 		R_ClearDrawSegs();
+		R_ClearSegTables();
 		R_ClearPlanes();
 		R_ClearSprites();
 
@@ -1434,6 +1435,7 @@ void R_RenderPlayerView(player_t *player)
 		R_ClearClipSegs();
 	}
 	R_ClearDrawSegs();
+	R_ClearSegTables();
 	R_ClearSprites();
 
 	// check for new console commands.

src/r_patch.h

@@ -36,6 +36,8 @@ typedef struct
 	boolean available;
 } spriteinfo_t;
 
+#define PNG_HEADER_SIZE (8)
+
 // Patch functions
 patch_t *Patch_Create(softwarepatch_t *source, size_t srcsize, void *dest);
 void Patch_Free(patch_t *patch);

src/r_plane.cpp

@@ -60,10 +60,6 @@ INT32 numffloors;
 #define visplane_hash(picnum,lightlevel,height) \
   ((unsigned)((picnum)*3+(lightlevel)+(height)*7) & VISPLANEHASHMASK)
 
-//SoM: 3/23/2000: Use boom opening limit removal
-size_t maxopenings;
-INT16 *openings, *lastopening; /// \todo free leak
-
 //
 // Clip values are the solid pixel bounding the range.
 //  floorclip starts out SCREENHEIGHT
@@ -339,13 +335,13 @@ void R_ClearPlanes(void)
 	R_ClearFFloorClips();
 
 	for (i = 0; i < MAXVISPLANES; i++)
+	{
 		for (*freehead = visplanes[i], visplanes[i] = NULL;
 			freehead && *freehead ;)
 		{
 			freehead = &(*freehead)->next;
 		}
-
-	lastopening = openings;
+	}
 
 	// left to right mapping
 	angle = (viewangle-ANGLE_90)>>ANGLETOFINESHIFT;
@@ -634,8 +630,9 @@ static void R_MakeSpans(void (*mapfunc)(drawspandata_t* ds, void(*spanfunc)(draw
 	// coarse enough to not be too slow due to task scheduling overhead.
 	// To safely do this, we need to copy part of spanstart to a local.
 	// This is essentially loop unrolling across threads.
-	constexpr const int kSpanTaskGranularity = 8;
+	constexpr const int kSpanTaskGranularity = 16;
 	drawspandata_t ds_copy = *ds;
+
 	while (t1 < t2 && t1 <= b1)
 	{
 		INT32 spanstartcopy[kSpanTaskGranularity] = {0};
@@ -753,6 +750,13 @@ void R_DrawSkyPlanes(void)
 	visplane_t *pl;
 	INT32 i;
 
+	// If we're not supposed to draw the sky (e.g. for skyboxes), don't do anything!
+	// This probably utterly ruins sky rendering for FOFs and polyobjects, unfortunately
+	if (!newview->sky)
+	{
+		return;
+	}
+
 #ifdef HAVE_THREADS
 	srb2::ThreadPool::Sema tp_sema;
 	srb2::g_main_threadpool->begin_sema();
@@ -781,14 +785,8 @@ static void R_DrawSkyPlane(visplane_t *pl, void(*colfunc2)(drawcolumndata_t*), b
 	if (!(pl->minx <= pl->maxx))
 		return;
 
-	// If we're not supposed to draw the sky (e.g. for skyboxes), don't do anything!
-	// This probably utterly ruins sky rendering for FOFs and polyobjects, unfortunately
-	if (!newview->sky)
-	{
-		return;
-	}
-
 	drawcolumndata_t dc = {};
+	const INT32 texture = texturetranslation[skytexture];
 
 	// Reset column drawer function (note: couldn't we just call walldrawerfunc directly?)
 	// (that is, unless we'll need to switch drawers in future for some reason)
@@ -807,24 +805,21 @@ static void R_DrawSkyPlane(visplane_t *pl, void(*colfunc2)(drawcolumndata_t*), b
 		dc.colormap += COLORMAP_REMAPOFFSET;
 
 	dc.texturemid = skytexturemid;
-	dc.texheight = textureheight[skytexture] >>FRACBITS;
+	dc.texheight = textureheight[texture] >>FRACBITS;
 	dc.sourcelength = dc.texheight;
 
 	x = pl->minx;
 
 	// Precache the texture so we don't corrupt the zoned heap off-main thread
-	if (!texturecache[texturetranslation[skytexture]])
-	{
-		R_GenerateTexture(texturetranslation[skytexture]);
-	}
+	R_CheckTextureCache(texture);
 
 #ifdef HAVE_THREADS
+	// Tune concurrency granularity here to maximize throughput
+	// The cheaper colfunc is, the more coarse the task should be
+	constexpr const int kSkyPlaneMacroColumns = 8;
+
 	while (x <= pl->maxx)
 	{
-		// Tune concurrency granularity here to maximize throughput
-		// The cheaper colfunc is, the more coarse the task should be
-		constexpr const int kSkyPlaneMacroColumns = 8;
-
 		auto thunk = [=]() mutable -> void {
 			for (int i = 0; i < kSkyPlaneMacroColumns && i + x <= pl->maxx; i++)
 			{
@@ -841,9 +836,7 @@ static void R_DrawSkyPlane(visplane_t *pl, void(*colfunc2)(drawcolumndata_t*), b
 
 				dc.iscale = FixedMul(skyscale, FINECOSINE(xtoviewangle[x + i]>>ANGLETOFINESHIFT));
 				dc.x = x + i;
-				dc.source =
-					R_GetColumn(texturetranslation[skytexture],
-						-angle); // get negative of angle for each column to display sky correct way round! --Monster Iestyn 27/01/18
+				dc.source = R_GetColumn(texture, -angle); // get negative of angle for each column to display sky correct way round! --Monster Iestyn 27/01/18
 
 				colfunc2(&dc);
 			}
@@ -1164,13 +1157,14 @@ void R_DrawSinglePlane(drawspandata_t* ds, visplane_t *pl, boolean allow_paralle
 						}
 					}
 
-					offset = (scry*vid.width) + scrx;
+					//offset = (scry*vid.width) + scrx;
+					offset = (scrx*vid.height) + scry;
 
 					// No idea if this works
 					VID_BlitLinearScreen(renderscreen + offset,
-										 vid.screens[1] + (top*vid.width), // intentionally not +offset
+										 vid.screens[1] + top, // intentionally not +offset
 										 viewwidth, bottom-top,
-										 vid.width, vid.width);
+										 vid.height, vid.height);
 				}
 			}
 		}

src/r_plane.h

@@ -67,8 +67,6 @@ extern visplane_t *floorplane;
 extern visplane_t *ceilingplane;
 
 // Visplane related.
-extern INT16 *lastopening, *openings;
-extern size_t maxopenings;
 
 extern INT16 *floorclip, *ceilingclip;
 extern fixed_t *frontscale, *yslopetab;

src/r_segs.cpp

@@ -78,9 +78,21 @@ static fixed_t topfrac, topstep;
 static fixed_t bottomfrac, bottomstep;
 
 static lighttable_t **walllights;
-static INT16 *maskedtexturecol;
+static fixed_t *maskedtexturecol;
 static fixed_t *maskedtextureheight = NULL;
 
+static size_t numopenings;
+static INT16 *openings, *lastopening;
+
+static size_t texturecolumntablesize;
+static fixed_t *texturecolumntable, *curtexturecolumntable;
+
+void R_ClearSegTables(void)
+{
+	lastopening = openings;
+	curtexturecolumntable = texturecolumntable;
+}
+
 // ==========================================================================
 // R_RenderMaskedSegRange
 // ==========================================================================
@@ -140,7 +152,18 @@ static void R_Render2sidedMultiPatchColumn(drawcolumndata_t* dc, column_t *colum
 
 transnum_t R_GetLinedefTransTable(fixed_t alpha)
 {
-	return static_cast<transnum_t>((20*(FRACUNIT - alpha - 1) + FRACUNIT) >> (FRACBITS+1));
+	if (alpha >= FRACUNIT)
+	{
+		return static_cast<transnum_t>(0);
+	}
+	else if (alpha <= 0)
+	{
+		return NUMTRANSMAPS;
+	}
+	else
+	{
+		return static_cast<transnum_t>((20*(FRACUNIT - alpha - 1) + FRACUNIT) >> (FRACBITS+1));
+	}
 }
 
 static inline boolean R_OverflowTest(drawcolumndata_t* dc)
@@ -173,7 +196,7 @@ static void R_RenderMaskedSegLoop(drawcolumndata_t* dc, drawseg_t *drawseg, INT3
 
 	ldef = curline->linedef;
 
-	range = std::max(drawseg->x2-drawseg->x1, 1);
+	range = std::max<INT32>(drawseg->x2-drawseg->x1, 1);
 
 	// Setup lighting based on the presence/lack-of 3D floors.
 	dc->numlights = 0;
@@ -306,141 +329,137 @@ static void R_RenderMaskedSegLoop(drawcolumndata_t* dc, drawseg_t *drawseg, INT3
 			else
 				dc->texturemid -= (textureheight[texnum])*times;
 
-			const INT16 colnum = maskedtexturecol[dc->x];
+			// Check for overflows first
+			if (R_OverflowTest(dc))
+			{
+				// Eh, no, go away, don't waste our time
+				if (dc->numlights)
+				{
+					for (i = 0; i < dc->numlights; i++)
+					{
+						rlight = &dc->lightlist[i];
+						rlight->height += rlight->heightstep;
+					}
+				}
+
+				spryscale += rw_scalestep;
+				continue;
+			}
 
 			// calculate lighting
-			if (colnum != INT16_MAX)
+			if (dc->numlights)
 			{
-				// Check for overflows first
-				if (R_OverflowTest(dc))
+				lighttable_t **xwalllights;
+
+				sprbotscreen = INT32_MAX;
+				sprtopscreen = windowtop = (centeryfrac - FixedMul(dc->texturemid, spryscale));
+
+				realbot = windowbottom = FixedMul(textureheight[texnum], spryscale) + sprtopscreen;
+				dc->iscale = 0xffffffffu / (unsigned)spryscale;
+
+				// draw the texture
+				col = (column_t *)((UINT8 *)R_GetColumn(texnum, (maskedtexturecol[dc->x] >> FRACBITS)) - 3);
+
+				auto set_light_vars = [&](INT32 k)
 				{
-					// Eh, no, go away, don't waste our time
-					if (dc->numlights)
+					rlight = &dc->lightlist[k];
+
+					lightnum = rlight->lightnum;
+
+					if (lightnum < 0)
+						xwalllights = scalelight[0];
+					else if (lightnum >= LIGHTLEVELS)
+						xwalllights = scalelight[LIGHTLEVELS-1];
+					else
+						xwalllights = scalelight[lightnum];
+
+					pindex = FixedMul(spryscale, LIGHTRESOLUTIONFIX)>>LIGHTSCALESHIFT;
+
+					if (pindex >= MAXLIGHTSCALE)
+						pindex = MAXLIGHTSCALE - 1;
+
+					if (rlight->extra_colormap)
+						rlight->rcolormap = rlight->extra_colormap->colormap + (xwalllights[pindex] - colormaps);
+					else
+						rlight->rcolormap = xwalllights[pindex];
+				};
+
+				auto set_colormap_below_light = [&]
+				{
+					dc->colormap = rlight->rcolormap;
+					if (encoremap && !(ldef->flags & ML_TFERLINE))
+						dc->colormap += COLORMAP_REMAPOFFSET;
+				};
+
+				// Use the base sector's light level above the first FOF.
+				// You can imagine it as the sky casting its light on top of the highest FOF.
+				set_light_vars(0);
+				set_colormap_below_light();
+
+				for (i = 0; i < dc->numlights; i++)
+				{
+					if ((dc->lightlist[i].flags & FF_NOSHADE))
+						continue;
+
+					set_light_vars(i);
+
+					height = rlight->height;
+					rlight->height += rlight->heightstep;
+
+					if (height <= windowtop)
 					{
-						for (i = 0; i < dc->numlights; i++)
+						set_colormap_below_light();
+						continue;
+					}
+
+					windowbottom = height;
+					if (windowbottom >= realbot)
+					{
+						windowbottom = realbot;
+						colfunc_2s(dc, col);
+						for (i++; i < dc->numlights; i++)
 						{
 							rlight = &dc->lightlist[i];
 							rlight->height += rlight->heightstep;
 						}
+
+						continue;
 					}
 
-					spryscale += rw_scalestep;
-					continue;
-				}
+					colfunc_2s(dc, col);
 
-				if (dc->numlights)
-				{
-					lighttable_t **xwalllights;
-
-					sprbotscreen = INT32_MAX;
-					sprtopscreen = windowtop = (centeryfrac - FixedMul(dc->texturemid, spryscale));
-
-					realbot = windowbottom = FixedMul(textureheight[texnum], spryscale) + sprtopscreen;
-					dc->iscale = 0xffffffffu / (unsigned)spryscale;
-
-					// draw the texture
-					col = (column_t *)((UINT8 *)R_GetColumn(texnum, colnum) - 3);
-
-					auto set_light_vars = [&](INT32 k)
-					{
-						rlight = &dc->lightlist[k];
-
-						lightnum = rlight->lightnum;
-
-						if (lightnum < 0)
-							xwalllights = scalelight[0];
-						else if (lightnum >= LIGHTLEVELS)
-							xwalllights = scalelight[LIGHTLEVELS-1];
-						else
-							xwalllights = scalelight[lightnum];
-
-						pindex = FixedMul(spryscale, LIGHTRESOLUTIONFIX)>>LIGHTSCALESHIFT;
-
-						if (pindex >= MAXLIGHTSCALE)
-							pindex = MAXLIGHTSCALE - 1;
-
-						if (rlight->extra_colormap)
-							rlight->rcolormap = rlight->extra_colormap->colormap + (xwalllights[pindex] - colormaps);
-						else
-							rlight->rcolormap = xwalllights[pindex];
-					};
-
-					auto set_colormap_below_light = [&]
-					{
-						dc->colormap = rlight->rcolormap;
-						if (encoremap && !(ldef->flags & ML_TFERLINE))
-							dc->colormap += COLORMAP_REMAPOFFSET;
-					};
-
-					// Use the base sector's light level above the first FOF.
-					// You can imagine it as the sky casting its light on top of the highest FOF.
-					set_light_vars(0);
+					windowtop = windowbottom + 1;
 					set_colormap_below_light();
-
-					for (i = 0; i < dc->numlights; i++)
-					{
-						if ((dc->lightlist[i].flags & FF_NOSHADE))
-							continue;
-
-						set_light_vars(i);
-
-						height = rlight->height;
-						rlight->height += rlight->heightstep;
-
-						if (height <= windowtop)
-						{
-							set_colormap_below_light();
-							continue;
-						}
-
-						windowbottom = height;
-						if (windowbottom >= realbot)
-						{
-							windowbottom = realbot;
-							colfunc_2s(dc, col);
-							for (i++; i < dc->numlights; i++)
-							{
-								rlight = &dc->lightlist[i];
-								rlight->height += rlight->heightstep;
-							}
-
-							continue;
-						}
-						colfunc_2s(dc, col);
-
-						windowtop = windowbottom + 1;
-						set_colormap_below_light();
-					}
-					windowbottom = realbot;
-					if (windowtop < windowbottom)
-						colfunc_2s(dc, col);
-
-					spryscale += rw_scalestep;
-					continue;
 				}
+				windowbottom = realbot;
+				if (windowtop < windowbottom)
+					colfunc_2s(dc, col);
 
-				// calculate lighting
-				pindex = FixedMul(spryscale, LIGHTRESOLUTIONFIX)>>LIGHTSCALESHIFT;
-
-				if (pindex >= MAXLIGHTSCALE)
-					pindex = MAXLIGHTSCALE - 1;
-
-				dc->colormap = walllights[pindex];
-				if (encoremap && !(ldef->flags & ML_TFERLINE))
-					dc->colormap += COLORMAP_REMAPOFFSET;
-
-				if (frontsector->extra_colormap)
-					dc->colormap = frontsector->extra_colormap->colormap + (dc->colormap - colormaps);
-
-				sprtopscreen = centeryfrac - FixedMul(dc->texturemid, spryscale);
-				dc->iscale = 0xffffffffu / (unsigned)spryscale;
-
-				// draw the texture
-				col = (column_t *)((UINT8 *)R_GetColumn(texnum, colnum) - 3);
-
-				colfunc_2s(dc, col);
+				spryscale += rw_scalestep;
+				continue;
 			}
 
+			// calculate lighting
+			pindex = FixedMul(spryscale, LIGHTRESOLUTIONFIX)>>LIGHTSCALESHIFT;
+
+			if (pindex >= MAXLIGHTSCALE)
+				pindex = MAXLIGHTSCALE - 1;
+
+			dc->colormap = walllights[pindex];
+			if (encoremap && !(ldef->flags & ML_TFERLINE))
+				dc->colormap += COLORMAP_REMAPOFFSET;
+
+			if (frontsector->extra_colormap)
+				dc->colormap = frontsector->extra_colormap->colormap + (dc->colormap - colormaps);
+
+			sprtopscreen = centeryfrac - FixedMul(dc->texturemid, spryscale);
+			dc->iscale = 0xffffffffu / (unsigned)spryscale;
+
+			// draw the texture
+			col = (column_t *)((UINT8 *)R_GetColumn(texnum, (maskedtexturecol[dc->x] >> FRACBITS)) - 3);
+
+			colfunc_2s(dc, col);
+
 			spryscale += rw_scalestep;
 		}
 	}
@@ -632,6 +651,7 @@ void R_RenderThickSideRange(drawseg_t *drawseg, INT32 x1, INT32 x2, ffloor_t *pf
 
 	void (*colfunc_2s) (drawcolumndata_t* dc, column_t *);
 
+
 	// Calculate light table.
 	// Use different light tables
 	//   for horizontal / vertical / diagonal. Diagonal?
@@ -674,7 +694,7 @@ void R_RenderThickSideRange(drawseg_t *drawseg, INT32 x1, INT32 x2, ffloor_t *pf
 		fog = true;
 	}
 
-	range = std::max(drawseg->x2-drawseg->x1, 1);
+	range = std::max<INT32>(drawseg->x2-drawseg->x1, 1);
 	//SoM: Moved these up here so they are available for my lightlist calculations
 	rw_scalestep = drawseg->scalestep;
 	spryscale = drawseg->scale1 + (x1 - drawseg->x1)*rw_scalestep;
@@ -899,14 +919,11 @@ void R_RenderThickSideRange(drawseg_t *drawseg, INT32 x1, INT32 x2, ffloor_t *pf
 	// draw the columns
 	for (dc->x = x1; dc->x <= x2; dc->x++)
 	{
-		if (maskedtexturecol[dc->x] == INT16_MAX)
-			continue;
-
 		// skew FOF walls
 		if (ffloortextureslide)
 		{
 			if (oldx != -1)
-				dc->texturemid += FixedMul(ffloortextureslide, (maskedtexturecol[oldx]-maskedtexturecol[dc->x])<<FRACBITS);
+				dc->texturemid += FixedMul(ffloortextureslide, (maskedtexturecol[oldx]-maskedtexturecol[dc->x]));
 			oldx = dc->x;
 		}
 
@@ -936,9 +953,6 @@ void R_RenderThickSideRange(drawseg_t *drawseg, INT32 x1, INT32 x2, ffloor_t *pf
 			continue;
 		}
 
-		// Get data for the column
-		col = (column_t *)((UINT8 *)R_GetColumn(texnum, maskedtexturecol[dc->x]) - 3);
-
 		if (!fog)
 		{
 			dc->iscale = 0xffffffffu / (unsigned)spryscale;
@@ -949,6 +963,9 @@ void R_RenderThickSideRange(drawseg_t *drawseg, INT32 x1, INT32 x2, ffloor_t *pf
 				R_SetColumnFunc(BASEDRAWFUNC);
 		}
 
+		// Get data for the column
+		col = (column_t *)((UINT8 *)R_GetColumn(texnum, (maskedtexturecol[dc->x] >> FRACBITS)) - 3);
+
 		// SoM: New code does not rely on R_DrawColumnShadowed which
 		// will (hopefully) put less strain on the stack.
 		if (dc->numlights)
@@ -1159,13 +1176,14 @@ static inline void R_ExpandPlaneY(visplane_t *pl, INT32 x, INT16 top, INT16 bott
 
 static void R_DrawWallColumn(drawcolumndata_t* dc, INT32 yl, INT32 yh, fixed_t mid, fixed_t texturecolumn, INT32 texture, boolean remap)
 {
+	const INT32 itexturecolumn = texturecolumn >> FRACBITS;
 	dc->yl = yl;
 	dc->yh = yh;
 	dc->texturemid = mid;
-	dc->source = R_GetColumn(texture, texturecolumn);
+	dc->source = R_GetColumn(texture, itexturecolumn);
 	dc->texheight = textureheight[texture] >> FRACBITS;
 	dc->sourcelength = dc->texheight;
-	//R_SetColumnFunc(colfunctype);
+	R_SetColumnFunc(colfunctype);
 	coldrawfunc_t* colfunccopy = colfunc;
 	drawcolumndata_t dc_copy = *dc;
 	if (remap)
@@ -1191,13 +1209,21 @@ static void R_RenderSegLoop(drawcolumndata_t* dc)
 	INT32     bottom;
 	INT32     i;
 
+	if (midtexture)
+		R_CheckTextureCache(midtexture);
+	if (toptexture)
+		R_CheckTextureCache(toptexture);
+	if (bottomtexture)
+		R_CheckTextureCache(bottomtexture);
+
 	// Loop over width
 	for (; rw_x < rw_stopx; rw_x++)
 	{
 		// mark floor / ceiling areas
 		yl = (topfrac+HEIGHTUNIT-1)>>HEIGHTBITS;
+		yh = bottomfrac>>HEIGHTBITS;
 
-		// no space above wall?
+		// Mark ceiling
 		top = ceilingclip[rw_x]+1;
 
 		// no space above wall?
@@ -1214,10 +1240,10 @@ static void R_RenderSegLoop(drawcolumndata_t* dc)
 			}
 		}
 
-		yh = bottomfrac>>HEIGHTBITS;
-
+		// Mark floor
 		bottom = floorclip[rw_x]-1;
 
+		// no space below floor?
 		if (yh > bottom)
 			yh = bottom;
 
@@ -1346,7 +1372,7 @@ static void R_RenderSegLoop(drawcolumndata_t* dc)
 		// calculate texture offset
 		angle = (rw_centerangle + xtoviewangle[rw_x])>>ANGLETOFINESHIFT;
 		// Mask 4095 to guarantee this index is within bounds
-		texturecolumn = rw_offset-FixedMul(FINETANGENT(angle & 4095),rw_distance);
+		texturecolumn = rw_offset-FixedMul(FINETANGENT(angle & 4095), rw_distance);
 
 		if (oldtexturecolumn != -1)
 		{
@@ -1357,8 +1383,6 @@ static void R_RenderSegLoop(drawcolumndata_t* dc)
 		}
 		oldtexturecolumn = texturecolumn;
 
-		texturecolumn >>= FRACBITS;
-
 		// texturecolumn and lighting are independent of wall tiers
 		if (segtextured)
 		{
@@ -1515,7 +1539,7 @@ static void R_RenderSegLoop(drawcolumndata_t* dc)
 		if (maskedtexture || numthicksides)
 		{
 			// save texturecol for backdrawing of masked mid texture
-			maskedtexturecol[rw_x] = (INT16)texturecolumn;
+			maskedtexturecol[rw_x] = texturecolumn;
 
 			if (maskedtextureheight != NULL)
 			{
@@ -1561,7 +1585,7 @@ static void R_RenderSegLoop(drawcolumndata_t* dc)
 
 static void R_MarkSegBounds(void)
 {
-	INT32     yl, yh;
+	INT32 yl, yh;
 	INT32 top, bottom;
 	INT16 topclip, bottomclip;
 
@@ -1570,8 +1594,9 @@ static void R_MarkSegBounds(void)
 	{
 		// mark floor / ceiling areas
 		yl = (topfrac+HEIGHTUNIT-1)>>HEIGHTBITS;
+		yh = bottomfrac>>HEIGHTBITS;
 
-		// no space above wall?
+		// Mark ceiling
 		top = ceilingclip[rw_x]+1;
 
 		// no space above wall?
@@ -1588,10 +1613,10 @@ static void R_MarkSegBounds(void)
 			}
 		}
 
-		yh = bottomfrac>>HEIGHTBITS;
-
+		// Mark floor
 		bottom = floorclip[rw_x]-1;
 
+		// no space below floor?
 		if (yh > bottom)
 			yh = bottom;
 
@@ -1616,12 +1641,6 @@ static void R_MarkSegBounds(void)
 		if (markfloor) // no bottom wall
 			floorclip[rw_x] = bottomclip;
 
-		if (floorclip[rw_x] <= ceilingclip[rw_x] + 1)
-		{
-			solidcol[rw_x] = 1;
-			didsolidcol = true;
-		}
-
 		rw_scale += rw_scalestep;
 		topfrac += topstep;
 		bottomfrac += bottomstep;
@@ -1663,15 +1682,87 @@ void R_AllocSegMemory(void)
 	fixed_t *frontscale_p = frontscaletable;
 	fixed_t *maskedheight_p = maskedheighttable;
 
-	for (drawseg_t *ds = drawsegs; ds < lastseg; ds++)
+	for (drawseg_t *drawseg = drawsegs; drawseg < lastseg; drawseg++)
 	{
-		ds->frontscale = frontscale_p;
-		ds->maskedtextureheight = maskedheight_p;
+		drawseg->frontscale = frontscale_p;
+		drawseg->maskedtextureheight = maskedheight_p;
 		frontscale_p += viewwidth;
 		maskedheight_p += viewwidth;
 	}
 }
 
+
+//SoM: Code to remove limits on openings.
+static void R_AllocClippingTables(size_t range)
+{
+	size_t pos = lastopening - openings;
+	size_t need = range * 2; // for both sprtopclip and sprbottomclip
+
+	if (pos + need < numopenings)
+		return;
+
+	INT16 *oldopenings = openings;
+	INT16 *oldlast = lastopening;
+
+	if (numopenings == 0)
+		numopenings = 16384;
+
+	numopenings += need;
+	openings = static_cast<INT16*>(Z_Realloc(openings, numopenings * sizeof (*openings), PU_STATIC, NULL));
+	lastopening = openings + pos;
+
+	if (oldopenings == NULL)
+		return;
+
+	// borrowed fix from *cough* zdoom *cough*
+	// [RH] We also need to adjust the openings pointers that
+	//    were already stored in drawsegs.
+	for (drawseg_t *drawseg = drawsegs; drawseg < ds_p; drawseg++)
+	{
+		// Check if it's in range of the openings
+#define CHECK(which) \
+		if (which + drawseg->x1 >= oldopenings && which + drawseg->x1 <= oldlast) \
+			which = (which - oldopenings) + openings
+		CHECK(drawseg->sprtopclip);
+		CHECK(drawseg->sprbottomclip);
+#undef CHECK
+	}
+}
+
+static void R_AllocTextureColumnTables(size_t range)
+{
+	size_t pos = curtexturecolumntable - texturecolumntable;
+
+	// For both tables, we reserve exactly an amount of memory that's equivalent to
+	// how many columns the seg will take on the entire screen (think about it)
+	if (pos + range < texturecolumntablesize)
+		return;
+
+	fixed_t *oldtable = texturecolumntable;
+	fixed_t *oldlast = curtexturecolumntable;
+
+	if (texturecolumntablesize == 0)
+		texturecolumntablesize = 16384;
+
+	texturecolumntablesize += range;
+	texturecolumntable = static_cast<fixed_t*>(Z_Realloc(texturecolumntable, texturecolumntablesize * sizeof (*texturecolumntable), PU_STATIC, NULL));
+	curtexturecolumntable = texturecolumntable + pos;
+
+	if (oldtable == NULL)
+		return;
+
+	for (drawseg_t *drawseg = drawsegs; drawseg < ds_p; drawseg++)
+	{
+		// Check if it's in range of the tables
+#define CHECK(which) \
+		if (which + drawseg->x1 >= oldtable && which + drawseg->x1 <= oldlast) \
+			which = (which - oldtable) + texturecolumntable
+		CHECK(drawseg->maskedtexturecol);
+		CHECK(drawseg->thicksidecol);
+#undef CHECK
+	}
+}
+
 //
 // R_StoreWallRange
 // A wall segment will be drawn
@@ -1684,12 +1775,13 @@ void R_StoreWallRange(INT32 start, INT32 stop)
 	angle_t       distangle, offsetangle;
 	boolean longboi;
 	INT32           lightnum;
-	INT32           i, p;
+	INT32           i = 0, p;
 	lightlist_t   *light;
 	r_lightlist_t *rlight;
 	INT32 range;
 	vertex_t segleft, segright;
 	fixed_t ceilingfrontslide, floorfrontslide, ceilingbackslide, floorbackslide;
+	const INT32 twosidedmidtexture = R_GetTextureNum(curline->sidedef->midtexture);
 	drawcolumndata_t dc = {};
 
 	maskedtextureheight = NULL;
@@ -1745,41 +1837,6 @@ void R_StoreWallRange(INT32 start, INT32 stop)
 	ds_p->curline = curline;
 	rw_stopx = stop+1;
 
-	//SoM: Code to remove limits on openings.
-	{
-		size_t pos = lastopening - openings;
-		size_t need = (rw_stopx - start)*4 + pos;
-
-		if (need > maxopenings)
-		{
-			drawseg_t *ds;  //needed for fix from *cough* zdoom *cough*
-			INT16 *oldopenings = openings;
-			INT16 *oldlast = lastopening;
-
-			do
-				maxopenings = maxopenings ? maxopenings*2 : 16384;
-			while (need > maxopenings);
-			openings = static_cast<INT16*>(Z_Realloc(openings, maxopenings * sizeof (*openings), PU_STATIC, NULL));
-			lastopening = openings + pos;
-
-			if (oldopenings == NULL)
-				return;
-
-			// borrowed fix from *cough* zdoom *cough*
-			// [RH] We also need to adjust the openings pointers that
-			//    were already stored in drawsegs.
-			for (ds = drawsegs; ds < ds_p; ds++)
-			{
-#define ADJUST(p) if (ds->p + ds->x1 >= oldopenings && ds->p + ds->x1 <= oldlast) ds->p = (ds->p - oldopenings) + openings;
-				ADJUST(maskedtexturecol);
-				ADJUST(sprtopclip);
-				ADJUST(sprbottomclip);
-				ADJUST(thicksidecol);
-#undef ADJUST
-			}
-		}
-	}  // end of code to remove limits on openings
-
 	// calculate scale at both ends and step
 	ds_p->scale1 = rw_scale = R_ScaleFromGlobalAngle(viewangle + xtoviewangle[start]);
 
@@ -1874,8 +1931,7 @@ void R_StoreWallRange(INT32 start, INT32 stop)
 
 	numbackffloors = 0;
 
-	for (i = 0; i < MAXFFLOORS; i++)
-		ds_p->thicksides[i] = NULL;
+	memset(ds_p->thicksides, 0, MAXFFLOORS * sizeof(*ds_p->thicksides));
 
 	if (numffloors)
 	{
@@ -2160,6 +2216,8 @@ void R_StoreWallRange(INT32 start, INT32 stop)
 		rw_toptexturemid += sidedef->rowoffset;
 		rw_bottomtexturemid += sidedef->rowoffset;
 
+		R_AllocTextureColumnTables(rw_stopx - start);
+
 		// allocate space for masked texture tables
 		if (frontsector && backsector && (frontsector->tag != backsector->tag) && (backsector->ffloors || frontsector->ffloors))
 		{
@@ -2173,8 +2231,8 @@ void R_StoreWallRange(INT32 start, INT32 stop)
 
 			maskedtexture = true;
 
-			ds_p->thicksidecol = maskedtexturecol = lastopening - rw_x;
-			lastopening += rw_stopx - rw_x;
+			ds_p->thicksidecol = maskedtexturecol = curtexturecolumntable - rw_x;
+			curtexturecolumntable += rw_stopx - rw_x;
 
 			lowcut       = std::max(worldbottom, worldlow) + viewz;
 			highcut      = std::min(worldtop, worldhigh) + viewz;
@@ -2353,13 +2411,14 @@ void R_StoreWallRange(INT32 start, INT32 stop)
 			ds_p->numthicksides = numthicksides = i;
 		}
 
-		if (sidedef->midtexture > 0 && sidedef->midtexture < numtextures)
+		if (twosidedmidtexture)
 		{
 			// masked midtexture
+
 			if (!ds_p->thicksidecol)
 			{
-				ds_p->maskedtexturecol = maskedtexturecol = lastopening - rw_x;
-				lastopening += rw_stopx - rw_x;
+				ds_p->maskedtexturecol = maskedtexturecol = curtexturecolumntable - rw_x;
+				curtexturecolumntable += rw_stopx - rw_x;
 			}
 			else
 				ds_p->maskedtexturecol = ds_p->thicksidecol;
@@ -2939,18 +2998,23 @@ void R_StoreWallRange(INT32 start, INT32 stop)
 	}
 
 	// save sprite clipping info
-	if (((ds_p->silhouette & SIL_TOP) || maskedtexture) && !ds_p->sprtopclip)
+	if (maskedtexture || (ds_p->silhouette & (SIL_TOP | SIL_BOTTOM)))
 	{
-		M_Memcpy(lastopening, ceilingclip+start, 2*(rw_stopx - start));
-		ds_p->sprtopclip = lastopening - start;
-		lastopening += rw_stopx - start;
-	}
+		R_AllocClippingTables(rw_stopx - start);
 
-	if (((ds_p->silhouette & SIL_BOTTOM) || maskedtexture) && !ds_p->sprbottomclip)
-	{
-		M_Memcpy(lastopening, floorclip + start, 2*(rw_stopx-start));
-		ds_p->sprbottomclip = lastopening - start;
-		lastopening += rw_stopx - start;
+		if (((ds_p->silhouette & SIL_TOP) || maskedtexture) && !ds_p->sprtopclip)
+		{
+			M_Memcpy(lastopening, ceilingclip + start, 2*(rw_stopx - start));
+			ds_p->sprtopclip = lastopening - start;
+			lastopening += rw_stopx - start;
+		}
+
+		if (((ds_p->silhouette & SIL_BOTTOM) || maskedtexture) && !ds_p->sprbottomclip)
+		{
+			M_Memcpy(lastopening, floorclip + start, 2*(rw_stopx - start));
+			ds_p->sprbottomclip = lastopening - start;
+			lastopening += rw_stopx - start;
+		}
 	}
 
 	if (maskedtexture)
@@ -2958,13 +3022,13 @@ void R_StoreWallRange(INT32 start, INT32 stop)
 		if (!(ds_p->silhouette & SIL_TOP))
 		{
 			ds_p->silhouette |= SIL_TOP;
-			ds_p->tsilheight = (sidedef->midtexture > 0 && sidedef->midtexture < numtextures) ? INT32_MIN: INT32_MAX;
+			ds_p->tsilheight = (twosidedmidtexture) ? INT32_MIN : INT32_MAX;
 		}
 
 		if (!(ds_p->silhouette & SIL_BOTTOM))
 		{
 			ds_p->silhouette |= SIL_BOTTOM;
-			ds_p->bsilheight = (sidedef->midtexture > 0 && sidedef->midtexture < numtextures) ? INT32_MAX: INT32_MIN;
+			ds_p->bsilheight = (twosidedmidtexture) ? INT32_MAX : INT32_MIN;
 		}
 	}
 

src/r_segs.h

@@ -25,6 +25,7 @@ extern "C" {
 void R_RenderMaskedSegRange(drawseg_t *ds, INT32 x1, INT32 x2);
 void R_RenderThickSideRange(drawseg_t *ds, INT32 x1, INT32 x2, ffloor_t *pffloor);
 void R_StoreWallRange(INT32 start, INT32 stop);
+void R_ClearSegTables(void);
 
 void R_AllocSegMemory(void);
 

src/r_things.cpp

@@ -94,13 +94,6 @@ static drawseg_xrange_item_t *drawsegs_xrange;
 static size_t drawsegs_xrange_size = 0;
 static INT32 drawsegs_xrange_count = 0;
 
-INT32 R_ThingLightLevel(mobj_t* thing)
-{
-	INT32 lightlevel = thing->lightlevel;
-
-	return lightlevel;
-}
-
 //
 // Sprite rotation 0 is facing the viewer,
 //  rotation 1 is one angle turn CLOCKWISE around the axis.
@@ -312,7 +305,7 @@ boolean R_AddSingleSpriteDef(const char *sprname, spritedef_t *spritedef, UINT16
 
 		// store sprite info in lookup tables
 		//FIXME : numspritelumps do not duplicate sprite replacements
-		W_ReadLumpHeaderPwad(wadnum, l, &patch, (sizeof(INT16) *4), 0);
+		W_ReadLumpHeaderPwad(wadnum, l, &patch, PNG_HEADER_SIZE, 0);
 		spritecachedinfo[numspritelumps].width = (INT32)(SHORT(patch.width))<<FRACBITS;
 		spritecachedinfo[numspritelumps].offset = (INT32)(SHORT(patch.leftoffset))<<FRACBITS;
 		spritecachedinfo[numspritelumps].topoffset = (INT32)(SHORT(patch.topoffset))<<FRACBITS;
@@ -2249,11 +2242,13 @@ static void R_CreateDrawNodes(maskcount_t* mask, drawnode_t* head, boolean temps
 			}
 			ds->curline->polyseg->visplane = NULL;
 		}
+
 		if (ds->maskedtexturecol)
 		{
 			entry = R_CreateDrawNode(head);
 			entry->seg = ds;
 		}
+
 		if (ds->numffloorplanes)
 		{
 			for (i = 0; i < ds->numffloorplanes; i++)
@@ -2824,6 +2819,7 @@ void R_ClipSprites(drawseg_t* dsstart, portal_t* portal)
 
 	if (drawsegs_xrange_size < maxdrawsegs)
 	{
+		// haleyjd: fix reallocation to track 2x size
 		drawsegs_xrange_size = 2 * maxdrawsegs;
 
 		for (i = 0; i < DS_RANGES_COUNT; i++)
@@ -2863,6 +2859,9 @@ void R_ClipSprites(drawseg_t* dsstart, portal_t* portal)
 		}
 	}
 
+	// haleyjd: terminate with a nullptr user for faster loop - adds ~3 FPS
+	drawsegs_xranges[0].items[drawsegs_xranges[0].count].user = nullptr;
+
 	for (; clippedvissprites < visspritecount; clippedvissprites++)
 	{
 		vissprite_t *spr = R_GetVisSprite(clippedvissprites);
@@ -3028,7 +3027,7 @@ static void R_DrawMaskedList(drawnode_t* head)
 		else if (r2->seg && r2->seg->maskedtexturecol != NULL)
 		{
 			next = r2->prev;
-			R_RenderMaskedSegRange( r2->seg, r2->seg->x1, r2->seg->x2);
+			R_RenderMaskedSegRange(r2->seg, r2->seg->x1, r2->seg->x2);
 			r2->seg->maskedtexturecol = NULL;
 			R_DoneWithNode(r2);
 			r2 = next;

src/r_things.h

@@ -23,10 +23,10 @@ extern "C" {
 
 // number of sprite lumps for spritewidth,offset,topoffset lookup tables
 // Fab: this is a hack : should allocate the lookup tables per sprite
-#if defined(__x86_64__) || defined(__amd64__) || defined(__aarch64__) || defined(__arm64__) // only for 64bit (idk how else to proper check lmao)
-#define MAXVISSPRITES 4096
-#else
+#if __SIZEOF_POINTER__ < 8
 #define MAXVISSPRITES 2048 // added 2-2-98 was 128
+#else
+#define MAXVISSPRITES 4096 // only for 64bit
 #endif
 
 #define VISSPRITECHUNKBITS 6	// 2^6 = 64 sprites per chunk
@@ -63,7 +63,10 @@ extern fixed_t windowtop;
 extern fixed_t windowbottom;
 extern INT32 lengthcol;
 
-INT32 R_ThingLightLevel(mobj_t *thing);
+FUNCINLINE static ATTRINLINE INT32 R_ThingLightLevel(mobj_t* thing)
+{
+	return (INT32)thing->lightlevel;
+}
 fixed_t R_GetSpriteDirectionalLighting(angle_t angle);
 
 fixed_t R_GetShadowZ(mobj_t *thing, pslope_t **shadowslope);
@@ -83,7 +86,7 @@ typedef struct
 {
 	size_t drawsegs[2];
 	size_t vissprites[2];
-	fixed_t viewx, viewy, viewz;			/**< View z stored at the time of the BSP traversal for the view/portal. Masked sorting/drawing needs it. */
+	fixed_t viewx, viewy, viewz; /**< View z stored at the time of the BSP traversal for the view/portal. Masked sorting/drawing needs it. */
 	sector_t* viewsector;
 } maskcount_t;
 

src/screen.c

@@ -400,74 +400,97 @@ void SCR_CalculateFPS(void)
 #endif
 }
 
+static void SCR_DrawOldTicRate(UINT32 cap, UINT32 benchmark, double fps, INT32 fpsflags)
+{
+	const char *fps_string;
+	INT32 ticcntcolor = 0;
+
+	if (fps > (benchmark - 5))
+		ticcntcolor = V_GREENMAP;
+	else if (fps < 20)
+		ticcntcolor = V_REDMAP;
+
+	if (cap != 0)
+		fps_string = va("%d/%d\x82", (INT32)fps, cap);
+	else
+		fps_string = va("%d\x82", (INT32)fps);
+
+	// draw "FPS"
+	if (cv_ticrate.value == 3)
+		V_DrawRightAlignedString(319, 181, V_YELLOWMAP|fpsflags, "FPS");
+
+	V_DrawRightAlignedString(319, 190, ticcntcolor|fpsflags, fps_string);
+}
+
+static void SCR_DrawKartTicRate(UINT32 cap, UINT32 benchmark, double fps, INT32 fpsflags)
+{
+	UINT8 *ticcntcolor = NULL;
+	INT32 x = 318;
+
+	// draw "FPS"
+	if (cv_ticrate.value == 1)
+	{
+		ticcntcolor = R_GetTranslationColormap(TC_RAINBOW, SKINCOLOR_YELLOW, GTC_CACHE);
+		V_DrawFixedPatch(306<<FRACBITS, 183<<FRACBITS, FRACUNIT, fpsflags, framecounter, ticcntcolor);
+	}
+
+	if (fps > (benchmark - 5))
+		ticcntcolor = R_GetTranslationColormap(TC_RAINBOW, SKINCOLOR_MINT, GTC_CACHE);
+	else if (fps < 20)
+		ticcntcolor = R_GetTranslationColormap(TC_RAINBOW, SKINCOLOR_RASPBERRY, GTC_CACHE);
+	else
+		ticcntcolor = NULL;
+
+	if (cap != 0)
+	{
+		UINT32 digits = 1;
+		UINT32 c2 = cap;
+
+		while (c2 > 0)
+		{
+			c2 = c2 / 10;
+			digits++;
+		}
+
+		// draw total frame:
+		V_DrawPingNum(x, 190, fpsflags, cap, ticcntcolor);
+
+		x -= digits * 4;
+
+		// draw "/"
+		V_DrawFixedPatch(x<<FRACBITS, 190<<FRACBITS, FRACUNIT, fpsflags, frameslash, ticcntcolor);
+	}
+
+	// draw our actual framerate
+	V_DrawPingNum(x, 190, fpsflags, fps, ticcntcolor);
+}
+
 void SCR_DisplayTicRate(void)
 {
-	UINT32 cap = R_GetFramerateCap();
-	UINT32 benchmark = (cap == 0) ? I_GetRefreshRate() : cap;
-	double fps = round(averageFPS);
-	INT32 fpsflags = V_LocalTransFlag()|V_SNAPTOBOTTOM|V_SNAPTORIGHT;
+	UINT32 cap, benchmark;
+	double fps;
+	INT32 fpsflags;
 
 	if (gamestate == GS_NULL)
 		return;
 
-	// new kart counter
-	if (cv_ticrate.value == 1 || cv_ticrate.value == 2)
+	cap = R_GetFramerateCap();
+	benchmark = (cap == 0) ? I_GetRefreshRate() : cap;
+	fps = round(averageFPS);
+	fpsflags = V_LocalTransFlag()|V_SNAPTOBOTTOM|V_SNAPTORIGHT;
+
+	switch (cv_ticrate.value)
 	{
-		const UINT8 *ticcntcolor = NULL;
-		INT32 x = 318;
-
-		// draw "FPS"
-		if (cv_ticrate.value == 1)
-			V_DrawFixedPatch(306<<FRACBITS, 183<<FRACBITS, FRACUNIT, fpsflags, framecounter, R_GetTranslationColormap(TC_RAINBOW, SKINCOLOR_YELLOW, GTC_CACHE));
-
-		if (fps > (benchmark - 5))
-			ticcntcolor = R_GetTranslationColormap(TC_RAINBOW, SKINCOLOR_MINT, GTC_CACHE);
-		else if (fps < 20)
-			ticcntcolor = R_GetTranslationColormap(TC_RAINBOW, SKINCOLOR_RASPBERRY, GTC_CACHE);
-
-		if (cap != 0)
-		{
-			UINT32 digits = 1;
-			UINT32 c2 = cap;
-
-			while (c2 > 0)
-			{
-				c2 = c2 / 10;
-				digits++;
-			}
-
-			// draw total frame:
-			V_DrawPingNum(x, 190, fpsflags, cap, ticcntcolor);
-
-			x -= digits * 4;
-
-			// draw "/"
-			V_DrawFixedPatch(x<<FRACBITS, 190<<FRACBITS, FRACUNIT, fpsflags, frameslash, ticcntcolor);
-		}
-
-		// draw our actual framerate
-		V_DrawPingNum(x, 190, fpsflags, fps, ticcntcolor);
-	}
-	else if (cv_ticrate.value == 3 || cv_ticrate.value == 4) // kart v1.0/srb2 counter
-	{
-		const char *fps_string;
-		INT32 ticcntcolor2 = 0;
-
-		if (fps > (benchmark - 5))
-			ticcntcolor2 = V_GREENMAP;
-		else if (fps < 20)
-			ticcntcolor2 = V_REDMAP;
-
-		if (cap != 0)
-			fps_string = va("%d/%d\x82", (INT32)fps, cap);
-		else
-			fps_string = va("%d\x82", (INT32)fps);
-
-		// draw "FPS"
-		if (cv_ticrate.value == 3)
-			V_DrawRightAlignedString(319, 181, V_YELLOWMAP|fpsflags, "FPS");
-
-		V_DrawRightAlignedString(319, 190, ticcntcolor2|fpsflags, fps_string);
+		case 1: // new kart counter
+		case 2:
+			SCR_DrawKartTicRate(cap, benchmark, fps, fpsflags);
+			break;
+		case 3: // kart v1.0/srb2 counter
+		case 4:
+			SCR_DrawOldTicRate(cap, benchmark, fps, fpsflags);
+			break;
+		default:
+			break;
 	}
 }
 
@@ -482,7 +505,6 @@ void SCR_DisplayLocalPing(void)
 	if (cv_showping.value == 1 || (cv_showping.value == 2 && ping > servermaxping)) // only show 2 (warning) if our ping is at a bad level
 	{
 		INT32 dispy = (cv_ticrate.value == 1) ? 165 : ((cv_ticrate.value == 2 || cv_ticrate.value == 4) ? 172 : ((cv_ticrate.value == 3) ? 163 : 181)); // absolute buttpain
-
 		HU_drawPlayerPing(308, dispy, consoleplayer, pingflags); // consoleplayer's ping is everyone's ping in a splitnetgame :P
 	}
 }

src/sdl/i_system.cpp

@@ -2021,7 +2021,7 @@ FUNCIERROR void ATTRNORETURN I_Error(const char *error, ...)
 		if (errorcount > 20)
 		{
 			va_start(argptr, error);
-			vsprintf(buffer, error, argptr);
+			vsnprintf(buffer, 8192, error, argptr);
 			va_end(argptr);
 			// Implement message box with SDL_ShowSimpleMessageBox,
 			// which should fail gracefully if it can't put a message box up
@@ -2039,7 +2039,7 @@ FUNCIERROR void ATTRNORETURN I_Error(const char *error, ...)
 
 	// Display error message in the console before we start shutting it down
 	va_start(argptr, error);
-	vsprintf(buffer, error, argptr);
+	vsnprintf(buffer, 8192, error, argptr);
 	va_end(argptr);
 	I_OutputMsg("\nI_Error(): %s\n", buffer);
 

src/sdl/i_video.c

@@ -1304,7 +1304,7 @@ static void SDLSetMode(INT32 width, INT32 height, SDL_bool fullscreen)
 			SDL_DestroyTexture(texture);
 		}
 
-		texture = SDL_CreateTexture(renderer, sw_texture_format, SDL_TEXTUREACCESS_STREAMING, width, height);
+		texture = SDL_CreateTexture(renderer, sw_texture_format, SDL_TEXTUREACCESS_STREAMING, height, width);
 	}
 }
 
@@ -1339,6 +1339,7 @@ void I_UpdateNoBlit(void)
 // I_FinishUpdate
 //
 static SDL_Rect src_rect = { 0, 0, 0, 0 };
+static SDL_Rect dst_rect = { 0, 0, 0, 0 };
 
 void I_FinishUpdate(void)
 {
@@ -1368,13 +1369,13 @@ void I_FinishUpdate(void)
 		void *pixels;
 		int pitch;
 		SDL_LockTexture(texture, NULL, &pixels, &pitch);
-		int step = pitch / 4 - vid.width;
+		int step = pitch / 4 - vid.height;
 		UINT32 *restrict dst = pixels;
 		UINT8 *restrict src = vid.screens[0];
 		UINT32 *restrict palette = localPalette;
-		for (INT32 y = 0; y < vid.height; y++)
+		for (INT32 y = 0; y < vid.width; y++)
 		{
-			UINT8 *restrict end = src + vid.width;
+			UINT8 *restrict end = src + vid.height;
 			do *dst++ = palette[*src++];
 			while (src < end);
 			dst += step;
@@ -1382,7 +1383,8 @@ void I_FinishUpdate(void)
 		SDL_UnlockTexture(texture);
 
 		SDL_RenderClear(renderer);
-		SDL_RenderCopy(renderer, texture, &src_rect, NULL);
+		//SDL_RenderCopy(renderer, texture, &src_rect, NULL);
+		SDL_RenderCopyEx(renderer, texture, NULL, &dst_rect, 90.0, NULL, SDL_FLIP_VERTICAL);
 		SDL_RenderPresent(renderer);
 	}
 #ifdef HWRENDER
@@ -1413,18 +1415,17 @@ void I_ReadScreen(UINT8 * restrict scr, INT32 scale)
 {
 	if (rendermode != render_soft)
 		I_Error("I_ReadScreen: called while in non-software mode");
-	else if (scale == 1)
-		VID_BlitLinearScreen(vid.screens[0], scr, vid.width, vid.height, vid.width, vid.width);
 	else
 	{
 		UINT8 * restrict source = vid.screens[0];
+
 		uintptr_t w = vid.width/scale*scale, h = vid.height/scale*scale;
 
 		// size_t saves a lea + movsxd over INT32. mind your types!
 		// uintptr_t is even better since it's guaranteed to be the size of a pointer
 		for (uintptr_t y = 0; y < h; y += scale)
 			for (uintptr_t x = 0; x < w; x += scale)
-				*scr++ = source[y*vid.width + x];
+				*scr++ = source[x*vid.height + y];
 	}
 }
 
@@ -1562,8 +1563,15 @@ INT32 VID_SetMode(INT32 modeNum)
 
 	SDLSetMode(vid.width, vid.height, USE_FULLSCREEN);
 
-	src_rect.w = vid.width;
-	src_rect.h = vid.height;
+	src_rect.w = vid.height;
+	src_rect.h = vid.width;
+
+
+	dst_rect.x = (vid.width - vid.height) / 2;
+	dst_rect.y = (vid.height - vid.width) / 2;
+	dst_rect.w = vid.height;
+	dst_rect.h = vid.width;
+
 
 	refresh_rate = VID_GetRefreshRate();
 

src/v_video.c

@@ -614,9 +614,9 @@ static void CV_constextsize_OnChange(void)
 // --------------------------------------------------------------------------
 void VID_BlitLinearScreen(const UINT8 *restrict srcptr, UINT8 *restrict destptr, INT32 width, INT32 height, size_t srcrowbytes, size_t destrowbytes)
 {
-	if (srcrowbytes == destrowbytes && srcrowbytes == (size_t)width)
+	if (srcrowbytes == destrowbytes && srcrowbytes == (size_t)height)
 	{
-		size_t i = srcrowbytes * height;
+		size_t i = srcrowbytes * width;
 #if defined(__SSE__)
 		while (i >= 16)
 		{
@@ -631,9 +631,9 @@ void VID_BlitLinearScreen(const UINT8 *restrict srcptr, UINT8 *restrict destptr,
 	}
 	else
 	{
-		while (height--)
+		while (width--)
 		{
-			M_Memcpy(destptr, srcptr, width);
+			M_Memcpy(destptr, srcptr, height);
 
 			destptr += destrowbytes;
 			srcptr += srcrowbytes;
@@ -663,7 +663,7 @@ void V_DrawStretchyFixedPatch(fixed_t x, fixed_t y, fixed_t pscale, fixed_t vsca
 	fixed_t col, ofs, colfrac, rowfrac, fdup, vdup;
 	INT32 dup;
 	const column_t *column;
-	UINT8 *desttop, *dest, *deststart, *destend;
+	UINT8 *desttop, *dest/*, *deststart, *destend*/;
 	const UINT8 *source, *deststop;
 	fixed_t pwidth; // patch width
 	fixed_t offx = 0; // x offset
@@ -791,13 +791,10 @@ void V_DrawStretchyFixedPatch(fixed_t x, fixed_t y, fixed_t pscale, fixed_t vsca
 	if (!desttop)
 		return;
 
-	deststop = desttop + vid.width * vid.height;
-
 	if (scrn & V_NOSCALESTART)
 	{
 		x >>= FRACBITS;
 		y >>= FRACBITS;
-		desttop += (y*vid.width) + x;
 	}
 	else
 	{
@@ -843,9 +840,8 @@ void V_DrawStretchyFixedPatch(fixed_t x, fixed_t y, fixed_t pscale, fixed_t vsca
 					y += (vid.height - (BASEVIDHEIGHT * dup)) / 2;
 			}
 		}
-
-		desttop += (y*vid.width) + x;
 	}
+	desttop += (x*vid.height) + y;
 
 	if (pscale != FRACUNIT) // scale width properly
 	{
@@ -857,29 +853,29 @@ void V_DrawStretchyFixedPatch(fixed_t x, fixed_t y, fixed_t pscale, fixed_t vsca
 	else
 		pwidth = patch->width * dup;
 
-	deststart = desttop;
-	destend = desttop + pwidth;
+	//deststart = desttop;
+	//destend = desttop + pwidth*vid.height;
 
-	const INT32 stride = vid.width;
-
-	for (col = 0; (col>>FRACBITS) < patch->width; col += colfrac, ++offx, desttop++)
+	for (col = 0; (col>>FRACBITS) < patch->width; col += colfrac, ++offx, desttop += vid.height)
 	{
 		INT32 topdelta, prevdelta = -1;
 		if (scrn & V_FLIP) // offx is measured from right edge instead of left
 		{
 			if (x+pwidth-offx < 0) // don't draw off the left of the screen (WRAP PREVENTION)
 				break;
-			if (x+pwidth-offx >= stride) // don't draw off the right of the screen (WRAP PREVENTION)
+			if (x+pwidth-offx >= vid.width) // don't draw off the right of the screen (WRAP PREVENTION)
 				continue;
 		}
 		else
 		{
 			if (x+offx < 0) // don't draw off the left of the screen (WRAP PREVENTION)
 				continue;
-			if (x+offx >= stride) // don't draw off the right of the screen (WRAP PREVENTION)
+			if (x+offx >= vid.width) // don't draw off the right of the screen (WRAP PREVENTION)
 				break;
 		}
-		column = (const column_t *)((const UINT8 *)(patch->columns) + (patch->columnofs[col>>FRACBITS]));
+
+		INT32 realcol = (scrn & V_FLIP) ? ((patch->width-(col>>FRACBITS)-1)<<FRACBITS) : col;
+		column = (const column_t *)((const UINT8 *)(patch->columns) + (patch->columnofs[realcol>>FRACBITS]));
 
 		switch (patchdrawtype)
 		{
@@ -892,15 +888,16 @@ void V_DrawStretchyFixedPatch(fixed_t x, fixed_t y, fixed_t pscale, fixed_t vsca
 					prevdelta = topdelta;
 					source = (const UINT8 *)(column) + 3;
 					dest = desttop;
-					if (scrn & V_FLIP)
-						dest = deststart + (destend - desttop);
-					dest += FixedInt(FixedMul(topdelta<<FRACBITS,vdup))*stride;
+					dest += FixedInt(FixedMul(topdelta<<FRACBITS,vdup));
 
-					for (ofs = 0; dest < deststop && (ofs>>FRACBITS) < column->length; ofs += rowfrac)
+					UINT8 *col_start = vid.screens[0] + ((x + offx)*vid.height);
+					deststop = col_start + vid.height;
+
+					for (ofs = 0; dest < deststop && (size_t)(ofs>>FRACBITS) < column->length; ofs += rowfrac)
 					{
-						if (dest >= vid.screens[scrn&V_PARAMMASK]) // don't draw off the top of the screen (CRASH PREVENTION)
+						if (dest >= col_start) // don't draw off the top of the screen (CRASH PREVENTION)
 							*dest = source[ofs>>FRACBITS];
-						dest += stride;
+						dest++;
 					}
 					column = (const column_t *)((const UINT8 *)column + column->length + 4);
 				}
@@ -915,15 +912,16 @@ void V_DrawStretchyFixedPatch(fixed_t x, fixed_t y, fixed_t pscale, fixed_t vsca
 					prevdelta = topdelta;
 					source = (const UINT8 *)(column) + 3;
 					dest = desttop;
-					if (scrn & V_FLIP)
-						dest = deststart + (destend - desttop);
-					dest += FixedInt(FixedMul(topdelta<<FRACBITS,vdup))*stride;
+					dest += FixedInt(FixedMul(topdelta<<FRACBITS,vdup));
 
-					for (ofs = 0; dest < deststop && (ofs>>FRACBITS) < column->length; ofs += rowfrac)
+					UINT8 *col_start = vid.screens[0] + ((x + offx)*vid.height);
+					deststop = col_start + vid.height;
+
+					for (ofs = 0; dest < deststop && (size_t)(ofs>>FRACBITS) < column->length; ofs += rowfrac)
 					{
-						if (dest >= vid.screens[scrn&V_PARAMMASK]) // don't draw off the top of the screen (CRASH PREVENTION)
+						if (dest >= col_start) // don't draw off the top of the screen (CRASH PREVENTION)
 							*dest = *(v_colormap + source[ofs>>FRACBITS]);
-						dest += stride;
+						dest++;
 					}
 					column = (const column_t *)((const UINT8 *)column + column->length + 4);
 				}
@@ -938,15 +936,16 @@ void V_DrawStretchyFixedPatch(fixed_t x, fixed_t y, fixed_t pscale, fixed_t vsca
 					prevdelta = topdelta;
 					source = (const UINT8 *)(column) + 3;
 					dest = desttop;
-					if (scrn & V_FLIP)
-						dest = deststart + (destend - desttop);
-					dest += FixedInt(FixedMul(topdelta<<FRACBITS,vdup))*stride;
+					dest += FixedInt(FixedMul(topdelta<<FRACBITS,vdup));
 
-					for (ofs = 0; dest < deststop && (ofs>>FRACBITS) < column->length; ofs += rowfrac)
+					UINT8 *col_start = vid.screens[0] + ((x + offx)*vid.height);
+					deststop = col_start + vid.height;
+
+					for (ofs = 0; dest < deststop && (size_t)(ofs>>FRACBITS) < column->length; ofs += rowfrac)
 					{
-						if (dest >= vid.screens[scrn&V_PARAMMASK]) // don't draw off the top of the screen (CRASH PREVENTION)
+						if (dest >= col_start) // don't draw off the top of the screen (CRASH PREVENTION)
 							*dest = *(v_translevel + ((source[ofs>>FRACBITS]<<8)&0xff00) + (*dest&0xff));
-						dest += stride;
+						dest++;
 					}
 					column = (const column_t *)((const UINT8 *)column + column->length + 4);
 				}
@@ -961,15 +960,16 @@ void V_DrawStretchyFixedPatch(fixed_t x, fixed_t y, fixed_t pscale, fixed_t vsca
 					prevdelta = topdelta;
 					source = (const UINT8 *)(column) + 3;
 					dest = desttop;
-					if (scrn & V_FLIP)
-						dest = deststart + (destend - desttop);
-					dest += FixedInt(FixedMul(topdelta<<FRACBITS,vdup))*stride;
+					dest += FixedInt(FixedMul(topdelta<<FRACBITS,vdup));
 
-					for (ofs = 0; dest < deststop && (ofs>>FRACBITS) < column->length; ofs += rowfrac)
+					UINT8 *col_start = vid.screens[0] + ((x + offx)*vid.height);
+					deststop = col_start + vid.height;
+
+					for (ofs = 0; dest < deststop && (size_t)(ofs>>FRACBITS) < column->length; ofs += rowfrac)
 					{
-						if (dest >= vid.screens[scrn&V_PARAMMASK]) // don't draw off the top of the screen (CRASH PREVENTION)
+						if (dest >= col_start) // don't draw off the top of the screen (CRASH PREVENTION)
 							*dest = *(v_translevel + (((*(v_colormap + source[ofs>>FRACBITS]))<<8)&0xff00) + (*dest&0xff));
-						dest += stride;
+						dest++;
 					}
 					column = (const column_t *)((const UINT8 *)column + column->length + 4);
 				}
@@ -1039,13 +1039,10 @@ void V_DrawCroppedPatch(fixed_t x, fixed_t y, fixed_t pscale, INT32 scrn, patch_
 	if (!desttop)
 		return;
 
-	deststop = desttop + vid.width * vid.height;
-
 	if (scrn & V_NOSCALESTART)
 	{
 		x >>= FRACBITS;
 		y >>= FRACBITS;
-		desttop += (y*vid.width) + x;
 	}
 	else
 	{
@@ -1079,20 +1076,18 @@ void V_DrawCroppedPatch(fixed_t x, fixed_t y, fixed_t pscale, INT32 scrn, patch_
 					y += (vid.height - (BASEVIDHEIGHT * vid.dup)) / 2;
 			}
 		}
-
-		desttop += (y*vid.width) + x;
 	}
 
-	const INT32 stride = vid.width;
+	desttop += (y*vid.width) + x;
 
-	for (col = sx<<FRACBITS; (col>>FRACBITS) < patch->width && ((col>>FRACBITS) - sx) < w; col += colfrac, ++x, desttop++)
+	for (col = sx<<FRACBITS; (col>>FRACBITS) < patch->width && ((col>>FRACBITS) - sx) < w; col += colfrac, ++x, desttop += vid.height)
 	{
 		INT32 topdelta, prevdelta = -1;
 
 		if (x < 0) // don't draw off the left of the screen (WRAP PREVENTION)
 			continue;
 
-		if (x >= stride) // don't draw off the right of the screen (WRAP PREVENTION)
+		if (x >= vid.width) // don't draw off the right of the screen (WRAP PREVENTION)
 			break;
 
 		column = (const column_t *)((const UINT8 *)(patch->columns) + (patch->columnofs[col>>FRACBITS]));
@@ -1110,17 +1105,20 @@ void V_DrawCroppedPatch(fixed_t x, fixed_t y, fixed_t pscale, INT32 scrn, patch_
 					dest = desttop;
 					if (topdelta-sy > 0)
 					{
-						dest += FixedInt(FixedMul((topdelta-sy)<<FRACBITS,fdup))*stride;
+						dest += FixedInt(FixedMul((topdelta-sy)<<FRACBITS,fdup));
 						ofs = 0;
 					}
 					else
 						ofs = (sy-topdelta)<<FRACBITS;
 
-					for (; dest < deststop && (ofs>>FRACBITS) < column->length && (((ofs>>FRACBITS) - sy) + topdelta) < h; ofs += rowfrac)
+					UINT8 *col_start = vid.screens[0] + (x*vid.height);
+					deststop = col_start + vid.height;
+
+					for (; dest < deststop && (size_t)(ofs>>FRACBITS) < column->length && (((ofs>>FRACBITS) - sy) + topdelta) < h; ofs += rowfrac)
 					{
-						if (dest >= vid.screens[scrn&V_PARAMMASK]) // don't draw off the top of the screen (CRASH PREVENTION)
+						if (dest >= col_start) // don't draw off the top of the screen (CRASH PREVENTION)
 							*dest = source[ofs>>FRACBITS];
-						dest += stride;
+						dest++;
 					}
 					column = (const column_t *)((const UINT8 *)column + column->length + 4);
 				}
@@ -1137,17 +1135,20 @@ void V_DrawCroppedPatch(fixed_t x, fixed_t y, fixed_t pscale, INT32 scrn, patch_
 					dest = desttop;
 					if (topdelta-sy > 0)
 					{
-						dest += FixedInt(FixedMul((topdelta-sy)<<FRACBITS,fdup))*stride;
+						dest += FixedInt(FixedMul((topdelta-sy)<<FRACBITS,fdup));
 						ofs = 0;
 					}
 					else
 						ofs = (sy-topdelta)<<FRACBITS;
 
-					for (; dest < deststop && (ofs>>FRACBITS) < column->length && (((ofs>>FRACBITS) - sy) + topdelta) < h; ofs += rowfrac)
+					UINT8 *col_start = vid.screens[0] + (x*vid.height);
+					deststop = col_start + vid.height;
+
+					for (; dest < deststop && (size_t)(ofs>>FRACBITS) < column->length && (((ofs>>FRACBITS) - sy) + topdelta) < h; ofs += rowfrac)
 					{
-						if (dest >= vid.screens[scrn&V_PARAMMASK]) // don't draw off the top of the screen (CRASH PREVENTION)
+						if (dest >= col_start) // don't draw off the top of the screen (CRASH PREVENTION)
 							*dest = *(v_translevel + ((source[ofs>>FRACBITS]<<8)&0xff00) + (*dest&0xff));
-						dest += stride;
+						dest++;
 					}
 					column = (const column_t *)((const UINT8 *)column + column->length + 4);
 				}
@@ -1189,8 +1190,8 @@ void V_DrawBlock(INT32 x, INT32 y, INT32 scrn, INT32 width, INT32 height, const
 		I_Error("Bad V_DrawBlock");
 #endif
 
-	dest = vid.screens[scrn] + y*vid.width + x;
-	deststop = vid.screens[scrn] + vid.width * vid.height;
+	dest = vid.screens[scrn] + x*vid.height + y;
+	deststop = vid.screens[scrn] + vid.height * vid.width;
 
 	while (height--)
 	{
@@ -1209,7 +1210,7 @@ void V_DrawBlock(INT32 x, INT32 y, INT32 scrn, INT32 width, INT32 height, const
 // alug: now with translucency support X)
 void V_DrawFill(INT32 x, INT32 y, INT32 w, INT32 h, INT32 c)
 {
-	UINT8 *dest;
+	UINT8 *dest, *desttop;
 	const UINT8 *deststop;
 	UINT32 alphalevel = ((c & V_ALPHAMASK) >> V_ALPHASHIFT);
 
@@ -1285,8 +1286,8 @@ void V_DrawFill(INT32 x, INT32 y, INT32 w, INT32 h, INT32 c)
 	if (y + h > vid.height)
 		h = vid.height-y;
 
-	dest = vid.screens[0] + y*vid.width + x;
-	deststop = vid.screens[0] + vid.width * vid.height;
+	dest = desttop = vid.screens[0] + x*vid.height + y;
+	deststop = vid.screens[0] + vid.height * vid.width;
 
 	if (alphalevel)
 	{
@@ -1312,16 +1313,31 @@ void V_DrawFill(INT32 x, INT32 y, INT32 w, INT32 h, INT32 c)
 	if (alphalevel)
 	{
 		const UINT8 *fadetable = ((UINT8 *)transtables + ((alphalevel-1)<<FF_TRANSSHIFT) + (c*256));
-		for (;(--h >= 0) && dest < deststop; dest += vid.width)
+		for (;(--h >= 0) && desttop < deststop; desttop++)
 		{
-			for (x = 0; x < w; x++)
-				dest[x] = fadetable[dest[x]];
+			dest = desttop;
+
+			UINT8 *row_end = dest + (w * vid.height);
+			while (dest < row_end)
+			{
+				*dest = fadetable[*dest];
+				dest += vid.height;
+			}
 		}
 	}
 	else
 	{
-		for (;(--h >= 0) && dest < deststop; dest += vid.width)
-			memset(dest, c, w);
+		for (;(--h >= 0) && desttop < deststop; desttop++)
+		{
+			dest = desttop;
+
+			UINT8 *row_end = dest + (w * vid.height);
+			while (dest < row_end)
+			{
+				*dest = c;
+				dest += vid.height;
+			}
+		}
 	}
 }
 
@@ -1367,8 +1383,8 @@ static UINT32 V_GetHWConsBackColor(void)
 
 void V_DrawFillConsoleMap(INT32 x, INT32 y, INT32 w, INT32 h, INT32 c)
 {
-	UINT8 *dest;
-	INT32 u, v;
+	UINT8 *dest, *desttop;
+	INT32 v;
 	UINT32 alphalevel = 0;
 
 	if (rendermode == render_none)
@@ -1435,7 +1451,7 @@ void V_DrawFillConsoleMap(INT32 x, INT32 y, INT32 w, INT32 h, INT32 c)
 	if (y + h > vid.height)
 		h = vid.height-y;
 
-	dest = vid.screens[0] + y*vid.width + x;
+	dest = desttop = vid.screens[0] + x*vid.height + y;
 
 	if ((alphalevel = ((c & V_ALPHAMASK) >> V_ALPHASHIFT)))
 	{
@@ -1460,11 +1476,14 @@ void V_DrawFillConsoleMap(INT32 x, INT32 y, INT32 w, INT32 h, INT32 c)
 
 	if (!alphalevel)
 	{
-		for (v = 0; v < h; v++, dest += vid.width)
+		for (v = 0; v < h; v++, desttop++)
 		{
-			for (u = 0; u < w; u++)
+			dest = desttop;
+			UINT8 *row_end = dest + w*vid.height;
+			while (dest < row_end)
 			{
-				dest[u] = consolebgmap[dest[u]];
+				*dest = consolebgmap[*dest];
+				dest += vid.height;
 			}
 		}
 	}
@@ -1475,11 +1494,14 @@ void V_DrawFillConsoleMap(INT32 x, INT32 y, INT32 w, INT32 h, INT32 c)
 		w = clip(w,vid.width);
 		h = clip(h,vid.height);
 #undef clip
-		for (v = 0; v < h; v++, dest += vid.width)
+		for (v = 0; v < h; v++, desttop++)
 		{
-			for (u = 0; u < w; u++)
+			dest = desttop;
+			UINT8 *row_end = dest + w*vid.height;
+			while (dest < row_end)
 			{
-				dest[u] = fadetable[consolebgmap[dest[u]]];
+				*dest = fadetable[consolebgmap[*dest]];
+				dest += vid.height;
 			}
 		}
 	}
@@ -1494,7 +1516,8 @@ void V_DrawFillConsoleMap(INT32 x, INT32 y, INT32 w, INT32 h, INT32 c)
 //
 void V_DrawDiag(INT32 x, INT32 y, INT32 wh, INT32 c)
 {
-	UINT8 *dest;
+	UINT8 *dest, *desttop;
+
 	const UINT8 *deststop;
 	INT32 w, h, wait = 0;
 
@@ -1569,19 +1592,28 @@ void V_DrawDiag(INT32 x, INT32 y, INT32 wh, INT32 c)
 	if (h > w)
 		h = w;
 
-	dest = vid.screens[0] + y*vid.width + x;
-	deststop = vid.screens[0] + vid.width * vid.height;
+	dest = desttop = vid.screens[0] + x*vid.height + y;
+	deststop = vid.screens[0] + vid.height * vid.width;
 
 	c &= 255;
 
-	for (;(--h >= 0) && dest < deststop; dest += vid.width)
+	for (;(--h >= 0) && desttop < deststop; desttop++)
 	{
-		memset(dest, c, w);
+		dest = desttop;
+
+		UINT8 *row_end = dest + (w * vid.height);
+		while (dest < row_end)
+		{
+			*dest = c;
+			dest += vid.height;
+		}
+
 		if (wait)
 			wait--;
 		else
 			w--;
 	}
+
 }
 
 //
@@ -1592,7 +1624,7 @@ void V_DrawFlatFill(INT32 x, INT32 y, INT32 w, INT32 h, lumpnum_t flatnum)
 	INT32 u, v;
 	fixed_t dx, dy, xfrac, yfrac;
 	const UINT8 *src, *deststop;
-	UINT8 *flat, *dest;
+	UINT8 *flat, *dest, *desttop;
 	size_t size, lflatsize, flatshift;
 
 #ifdef HWRENDER
@@ -1639,7 +1671,7 @@ void V_DrawFlatFill(INT32 x, INT32 y, INT32 w, INT32 h, lumpnum_t flatnum)
 
 	flat = W_CacheLumpNum(flatnum, PU_CACHE);
 
-	dest = vid.screens[0] + y*vid.dup*vid.width + x*vid.dup;
+	dest = desttop = vid.screens[0] + y*vid.dup*vid.width + x*vid.dup;
 	deststop = vid.screens[0] + vid.width * vid.height;
 
 	// from V_DrawScaledPatch
@@ -1663,15 +1695,18 @@ void V_DrawFlatFill(INT32 x, INT32 y, INT32 w, INT32 h, lumpnum_t flatnum)
 	dy = FixedDiv(FRACUNIT, vid.dup<<(FRACBITS-2));
 
 	yfrac = 0;
-	for (v = 0; v < h; v++, dest += vid.width)
+	for (v = 0; v < h; v++, desttop++)
 	{
 		xfrac = 0;
 		src = flat + (((yfrac>>FRACBITS) & (lflatsize - 1)) << flatshift);
+		dest = desttop;
+
 		for (u = 0; u < w; u++)
 		{
-			if (&dest[u] > deststop)
+			if (dest > deststop)
 				return;
-			dest[u] = src[(xfrac>>FRACBITS)&(lflatsize-1)];
+			*dest = src[(xfrac>>FRACBITS)&(lflatsize-1)];
+			dest += vid.height;
 			xfrac += dx;
 		}
 		yfrac += dy;
@@ -1743,7 +1778,7 @@ void V_DrawVhsEffect(boolean rewind)
 
 	UINT8 *buf = vid.screens[0], *tmp = vid.screens[4];
 	UINT16 y;
-	UINT32 x, pos = 0;
+	INT32 x;
 
 	UINT8 *normalmapstart = ((UINT8 *)transtables + (8<<FF_TRANSSHIFT|(19<<8)));
 #ifdef HQ_VHS
@@ -1777,32 +1812,40 @@ void V_DrawVhsEffect(boolean rewind)
 	}
 #endif
 
-	for (y = 0; y < vid.height; y+=2)
+	// Since we go through columns now instead of rows, we need to keep same random offset values for each row
+	// Don't need a random value for each row, just a bunch of values to cycle through will do
+#define ROWOFFSLEN 128
+	SINT8 rowoffs[ROWOFFSLEN];
+	for (y = 0; y < ROWOFFSLEN; y++)
+		rowoffs[y] = M_RandomKey(vid.dup<<1);
+
+	for (x = 0; x < vid.width; ++x)
 	{
-		thismapstart = normalmapstart;
-		offs = 0;
-
-		if (y >= uby && y < uby+barsize)
+		for (y = 0; y < vid.height; y++)
 		{
-			thismapstart -= (2<<FF_TRANSSHIFT) - (5<<8);
-			offs += updistort * 2.0f * min(y-uby, uby+barsize-y) / barsize;
-		}
+			offs = rowoffs[y/2 % ROWOFFSLEN];
 
-		if (y >= dby && y < dby+barsize)
-		{
-			thismapstart -= (2<<FF_TRANSSHIFT) - (5<<8);
-			offs -= downdistort * 2.0f * min(y-dby, dby+barsize-y) / barsize;
-		}
+			thismapstart = normalmapstart;
 
-		offs += M_RandomKey(vid.dup<<1);
+			if (y >= uby && y < uby+barsize)
+			{
+				thismapstart -= (2<<FF_TRANSSHIFT) - (5<<8);
+				offs += updistort * 2.0f * min(y-uby, uby+barsize-y) / barsize;
+			}
 
-		// lazy way to avoid crashes
-		if ((y == 0 && offs < 0) || (y >= vid.height-2 && offs > 0))
-			offs = 0;
+			if (y >= dby && y < dby+barsize)
+			{
+				thismapstart -= (2<<FF_TRANSSHIFT) - (5<<8);
+				offs -= downdistort * 2.0f * min(y-dby, dby+barsize-y) / barsize;
+			}
 
-		for (x = min(pos+(size_t)vid.width*2, (size_t)vid.width*vid.height); pos < x; pos++)
-		{
-			tmp[pos] = thismapstart[buf[pos+offs]];
+			INT32 pos = x*vid.height + y;
+			INT32 offspos = pos + offs*vid.height;
+
+			if (offspos < 0 || offspos > vid.width*vid.height)
+				offspos = pos;
+
+			tmp[pos] = thismapstart[buf[offspos]];
 #ifdef HQ_VHS
 			tmp[pos] = tmapstart[buf[pos]<<8 | tmp[pos]];
 #endif
@@ -1810,6 +1853,7 @@ void V_DrawVhsEffect(boolean rewind)
 	}
 
 	memcpy(buf, tmp, vid.width*vid.height);
+#undef ROWOFFSLEN
 }
 
 //
@@ -1837,7 +1881,7 @@ void V_DrawFadeScreen(UINT16 color, UINT8 strength)
 		: ((color & 0xFF00) // Color is not palette index?
 		? ((UINT8 *)colormaps + strength*256) // Do COLORMAP fade.
 		: ((UINT8 *)transtables + ((9-strength)<<FF_TRANSSHIFT) + color*256)); // Else, do TRANSMAP** fade.
-	const UINT8 *deststop = vid.screens[0] + vid.width * vid.height;
+	const UINT8 *deststop = vid.screens[0] + vid.height * vid.width;
 	UINT8 *buf = vid.screens[0];
 
 	// heavily simplified -- we don't need to know x or y
@@ -1849,7 +1893,7 @@ void V_DrawFadeScreen(UINT16 color, UINT8 strength)
 // Simple translucency with one color, over a set number of lines starting from the top.
 void V_DrawFadeConsBack(INT32 plines)
 {
-	UINT8 *deststop, *buf;
+	UINT8 *deststop, *dest, *desttop;
 
 #ifdef HWRENDER // not win32 only 19990829 by Kin
 	if (rendermode == render_opengl)
@@ -1862,9 +1906,18 @@ void V_DrawFadeConsBack(INT32 plines)
 
 	// heavily simplified -- we don't need to know x or y position,
 	// just the stop position
-	deststop = vid.screens[0] + vid.width * min(plines, vid.height);
-	for (buf = vid.screens[0]; buf < deststop; ++buf)
-		*buf = consolebgmap[*buf];
+	dest = desttop = vid.screens[0];
+	deststop = vid.screens[0] + min(plines, vid.height);
+	for (; desttop < deststop; desttop++)
+	{
+		dest = desttop;
+		UINT8 *row_end = dest + (vid.width * vid.height);
+		while (dest < row_end)
+		{
+			*dest = consolebgmap[*dest];
+			dest += vid.height;
+		}
+	}
 }
 
 // Gets string colormap, used for 0x80 color codes
@@ -3599,7 +3652,7 @@ void V_DoPostProcessor(INT32 view, INT32 param)
 
 	if (thiscam->postimg & POSTIMG_WATER)
 	{
-		INT32 y;
+		INT32 y, x;
 		// Set disStart to a range from 0 to FINEANGLE, incrementing by 128 per tic
 		angle_t disStart = (((leveltime-1)*128) + (R_GetTimeFrac(RTF_LEVEL) / (FRACUNIT/128))) & FINEMASK;
 		INT32 newpix;
@@ -3609,7 +3662,7 @@ void V_DoPostProcessor(INT32 view, INT32 param)
 		for (y = yoffset; y < yoffset+viewheight; y++)
 		{
 			sine = (FINESINE(disStart)*5)>>FRACBITS;
-			newpix = abs(sine);
+			/*newpix = abs(sine);
 
 			if (sine < 0)
 			{
@@ -3632,7 +3685,7 @@ void V_DoPostProcessor(INT32 view, INT32 param)
 					tmpscr[(y*vid.width)+xoffset+viewwidth-newpix] = srcscr[(y*vid.width)+xoffset+(viewwidth-1)];
 					newpix--;
 				}
-			}
+			}*/
 
 			/*
 			 Unoptimized version
@@ -3648,6 +3701,18 @@ void V_DoPostProcessor(INT32 view, INT32 param)
 			 	tmpscr[y*vid.width + x] = srcscr[y*vid.width+newpix]; // *(transme + (srcscr[y*vid.width+x]<<8) + srcscr[y*vid.width+newpix]);
 			 }*/
 
+			for (x = 0; x < vid.width; x++)
+			{
+				newpix = (x + sine);
+
+				if (newpix < 0)
+					newpix = 0;
+				else if (newpix >= vid.width)
+					newpix = vid.width-1;
+
+				tmpscr[y + x*vid.height] = srcscr[y+newpix*vid.height]; // *(transme + (srcscr[y*vid.width+x]<<8) + srcscr[y*vid.width+newpix]);
+			}
+
 			disStart += 22;//the offset into the displacement map, increment each game loop
 			disStart &= FINEMASK; //clip it to FINEMASK
 		}
@@ -3658,7 +3723,7 @@ void V_DoPostProcessor(INT32 view, INT32 param)
 	}
 	else if (thiscam->postimg & POSTIMG_HEAT) // Heat wave
 	{
-		INT32 y;
+		INT32 y, x;
 
 		// Make sure table is built
 		if (heatshifter == NULL || lastheight != viewheight)
@@ -3683,11 +3748,12 @@ void V_DoPostProcessor(INT32 view, INT32 param)
 			if (heatshifter[heatindex[view]++])
 			{
 				// Shift this row of pixels to the right by 2
-				tmpscr[(y*vid.width)+xoffset] = srcscr[(y*vid.width)+xoffset];
-				M_Memcpy(&tmpscr[(y*vid.width)+xoffset], &srcscr[(y*vid.width)+xoffset+vid.dup], viewwidth-vid.dup);
+				tmpscr[y] = srcscr[y];
+				for (x = vid.dup; x < vid.width; x++)
+					tmpscr[y+x*vid.height] = srcscr[y+(x-vid.dup)*vid.height];
 			}
-			else
-				M_Memcpy(&tmpscr[(y*vid.width)+xoffset], &srcscr[(y*vid.width)+xoffset], viewwidth);
+			else for (x = 0; x < vid.width; x++)
+				tmpscr[y+x*vid.height] = srcscr[y+x*vid.height];
 
 			heatindex[view] %= viewheight;
 		}
@@ -3719,10 +3785,11 @@ void V_DoPostProcessor(INT32 view, INT32 param)
 
 	if ((thiscam->postimg & POSTIMG_FLIP) && !(thiscam->postimg & POSTIMG_MIRROR)) // Flip the screen upside-down
 	{
-		INT32 y, y2;
+		INT32 x, y, y2;
 
-		for (y = yoffset, y2 = yoffset+viewheight - 1; y < yoffset+viewheight; y++, y2--)
-			M_Memcpy(&tmpscr[(y2*vid.width)+xoffset], &srcscr[(y*vid.width)+xoffset], viewwidth);
+		for (x = xoffset; x < xoffset+viewwidth; x++)
+			for (y = yoffset, y2 = yoffset+(viewheight-1); y < yoffset+viewheight; y++, y2--)
+				tmpscr[x*vid.height + y2] = srcscr[x*vid.height + y];
 
 		UINT8 *tmp = tmpscr;
 		tmpscr = srcscr;
@@ -3730,11 +3797,10 @@ void V_DoPostProcessor(INT32 view, INT32 param)
 	}
 	else if ((thiscam->postimg & POSTIMG_MIRROR) && !(thiscam->postimg & POSTIMG_FLIP)) // Flip the screen on the x axis
 	{
-		INT32 y, x, x2;
+		INT32 x, x2;
 
-		for (y = yoffset; y < yoffset+viewheight; y++)
-			for (x = xoffset, x2 = xoffset+(viewwidth-1); x < xoffset+viewwidth; x++, x2--)
-				tmpscr[y*vid.width + x2] = srcscr[y*vid.width + x];
+		for (x = xoffset, x2 = xoffset+viewwidth - 1; x < xoffset+viewwidth; x++, x2--)
+			M_Memcpy(&tmpscr[(x2*vid.height)+yoffset], &srcscr[(x*vid.height)+yoffset], viewheight);
 
 		UINT8 *tmp = tmpscr;
 		tmpscr = srcscr;
@@ -3754,7 +3820,7 @@ void V_DoPostProcessor(INT32 view, INT32 param)
 	}
 
 	VID_BlitLinearScreen(srcscr+vid.width*yoffset+xoffset, tmpscr+vid.width*yoffset+xoffset,
-						 viewwidth, viewheight, vid.width, vid.width);
+						 viewwidth, viewheight, vid.height, vid.height);
 #endif
 }
 
@@ -3810,7 +3876,7 @@ UINT8 GetColorLUTDirect(colorlookup_t *lut, UINT8 r, UINT8 g, UINT8 b)
 void V_Init(void)
 {
 	INT32 i;
-	INT32 screensize = vid.width * vid.height;
+	INT32 screensize = vid.height * vid.width;
 
 	for (i = 0; i < NUMSCREENS; i++)
 	{

src/w_wad.c

@@ -45,6 +45,7 @@
 #include "w_wad.h"
 #include "z_zone.h"
 #include "fastcmp.h"
+#include "console.h"
 
 #include "g_game.h" // G_LoadGameData
 #include "d_main.h"
@@ -760,11 +761,7 @@ static UINT16 W_InitFileError(const char *filename, boolean exitworthy)
 {
 	if (exitworthy)
 	{
-#ifdef _DEBUG
 		CONS_Error(va("%s was not found or not valid.\nCheck the log for more details.\n", filename));
-#else
-		I_Error("%s was not found or not valid.\nCheck the log for more details.\n", filename);
-#endif
 	}
 	else
 		CONS_Printf(M_GetText("Errors occurred while loading %s; not added.\n"), filename);

src/win32/win_dbg.c

@@ -118,7 +118,7 @@ static VOID FPrintf(HANDLE fileHandle, LPCSTR lpFmt, ...)
 	DWORD   bytesWritten;
 
 	va_start(arglist, lpFmt);
-	vsprintf(str, lpFmt, arglist);
+	vsnprintf(str, 1999, lpFmt, arglist);
 	va_end(arglist);
 
 	WriteFile(fileHandle, str, (DWORD)strlen(str), &bytesWritten, NULL);

src/y_inter.c

@@ -450,7 +450,7 @@ void Y_IntermissionDrawer(void)
 	else
 	{
 		if (rendermode == render_soft)
-			VID_BlitLinearScreen(vid.screens[1], vid.screens[0], vid.width, vid.height, vid.width, vid.width);
+			VID_BlitLinearScreen(vid.screens[1], vid.screens[0], vid.width, vid.height, vid.height, vid.height);
 #ifdef HWRENDER
 		else if (rendermode == render_opengl)
 			HWR_DrawIntermissionBG();