stjr/SRB2
0
0
Fork 0
mirror of https://git.do.srb2.org/STJr/SRB2.git synced 2024-12-13 06:10:50 +00:00
Code Issues Projects Releases Packages Wiki Activity
SRB2/src/hardware/hw_main.c

6769 lines
212 KiB
C
Raw Normal View History

SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// Emacs style mode select -*- C++ -*-
//-----------------------------------------------------------------------------
//
// Copyright (C) 1998-2000 by DooM Legacy Team.
//
// This program is free software; you can redistribute it and/or
// modify it under the terms of the GNU General Public License
// as published by the Free Software Foundation; either version 2
// of the License, or (at your option) any later version.
//
// This program is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU General Public License for more details.
//-----------------------------------------------------------------------------
/// \file
/// \brief hardware renderer, using the standard HardWareRender driver DLL for SRB2
#include <math.h>
#include "../doomstat.h"
#ifdef HWRENDER
#include "hw_glob.h"
#include "hw_light.h"
#include "hw_drv.h"
#include "../i_video.h" // for rendermode == render_glide
#include "../v_video.h"
#include "../p_local.h"
#include "../p_setup.h"
#include "../r_local.h"
#include "../r_bsp.h"
#include "../d_clisrv.h"
#include "../w_wad.h"
#include "../z_zone.h"
#include "../r_splats.h"
#include "../g_game.h"
#include "../st_stuff.h"
#include "../i_system.h"
#include "../m_cheat.h"
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
#ifdef ESLOPE
#include "../p_slopes.h"
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
#include "hw_md2.h"
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#ifdef NEWCLIP
#include "hw_clip.h"
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
#define R_FAKEFLOORS
YUP, I just fixed OGL Precipitation with this branch too. What the fuck? Squashing ancient bugs, woo.
2016-10-27 22:41:24 +00:00
#define HWPRECIP
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
#define SORTING
//#define POLYSKY
// ==========================================================================
// the hardware driver object
// ==========================================================================
struct hwdriver_s hwdriver;
// ==========================================================================
// PROTOS
// ==========================================================================
static void HWR_AddSprites(sector_t *sec);
static void HWR_ProjectSprite(mobj_t *thing);
#ifdef HWPRECIP
static void HWR_ProjectPrecipitationSprite(precipmobj_t *thing);
#endif
#ifdef SORTING
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
void HWR_AddTransparentFloor(lumpnum_t lumpnum, extrasubsector_t *xsub, boolean isceiling, fixed_t fixedheight,
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
INT32 lightlevel, INT32 alpha, sector_t *FOFSector, FBITFIELD blend, boolean fogplane, extracolormap_t *planecolormap);
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
void HWR_AddTransparentPolyobjectFloor(lumpnum_t lumpnum, polyobj_t *polysector, boolean isceiling, fixed_t fixedheight,
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
INT32 lightlevel, INT32 alpha, sector_t *FOFSector, FBITFIELD blend, extracolormap_t *planecolormap);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
#else
static void HWR_Add3DWater(lumpnum_t lumpnum, extrasubsector_t *xsub, fixed_t fixedheight,
INT32 lightlevel, INT32 alpha, sector_t *FOFSector);
static void HWR_Render3DWater(void);
static void HWR_RenderTransparentWalls(void);
#endif
static void HWR_FoggingOn(void);
static UINT32 atohex(const char *s);
static void CV_filtermode_ONChange(void);
static void CV_anisotropic_ONChange(void);
static void CV_FogDensity_ONChange(void);
static void CV_grFov_OnChange(void);
// ==========================================================================
// 3D ENGINE COMMANDS & CONSOLE VARS
// ==========================================================================
static CV_PossibleValue_t grfov_cons_t[] = {{0, "MIN"}, {179*FRACUNIT, "MAX"}, {0, NULL}};
static CV_PossibleValue_t grfiltermode_cons_t[]= {{HWD_SET_TEXTUREFILTER_POINTSAMPLED, "Nearest"},
{HWD_SET_TEXTUREFILTER_BILINEAR, "Bilinear"}, {HWD_SET_TEXTUREFILTER_TRILINEAR, "Trilinear"},
{HWD_SET_TEXTUREFILTER_MIXED1, "Linear_Nearest"},
{HWD_SET_TEXTUREFILTER_MIXED2, "Nearest_Linear"},
{HWD_SET_TEXTUREFILTER_MIXED3, "Nearest_Mipmap"},
{0, NULL}};
CV_PossibleValue_t granisotropicmode_cons_t[] = {{1, "MIN"}, {16, "MAX"}, {0, NULL}};
boolean drawsky = true;
// needs fix: walls are incorrectly clipped one column less
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#ifndef NEWCLIP
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
static consvar_t cv_grclipwalls = {"gr_clipwalls", "Off", 0, CV_OnOff, NULL, 0, NULL, NULL, 0, 0, NULL};
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
//development variables for diverse uses
static consvar_t cv_gralpha = {"gr_alpha", "160", 0, CV_Unsigned, NULL, 0, NULL, NULL, 0, 0, NULL};
static consvar_t cv_grbeta = {"gr_beta", "0", 0, CV_Unsigned, NULL, 0, NULL, NULL, 0, 0, NULL};
static float HWRWipeCounter = 1.0f;
consvar_t cv_grrounddown = {"gr_rounddown", "Off", 0, CV_OnOff, NULL, 0, NULL, NULL, 0, 0, NULL};
consvar_t cv_grfov = {"gr_fov", "90", CV_FLOAT|CV_CALL, grfov_cons_t, CV_grFov_OnChange, 0, NULL, NULL, 0, 0, NULL};
consvar_t cv_grfogdensity = {"gr_fogdensity", "150", CV_CALL|CV_NOINIT, CV_Unsigned,
CV_FogDensity_ONChange, 0, NULL, NULL, 0, 0, NULL};
// Unfortunately, this can no longer be saved..
consvar_t cv_grfiltermode = {"gr_filtermode", "Nearest", CV_CALL, grfiltermode_cons_t,
CV_filtermode_ONChange, 0, NULL, NULL, 0, 0, NULL};
consvar_t cv_granisotropicmode = {"gr_anisotropicmode", "1", CV_CALL, granisotropicmode_cons_t,
CV_anisotropic_ONChange, 0, NULL, NULL, 0, 0, NULL};
//static consvar_t cv_grzbuffer = {"gr_zbuffer", "On", 0, CV_OnOff};
Uncomment HWR_CorrectSWTricks but set gr_correcttricks to be off by default I wasn't aware of the cvar, this should do for now since I don't believe any maps use these software tricks, probably an old leftover from DOOM.
2018-03-30 22:12:44 +00:00
consvar_t cv_grcorrecttricks = {"gr_correcttricks", "Off", 0, CV_OnOff, NULL, 0, NULL, NULL, 0, 0, NULL};
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
consvar_t cv_grsolvetjoin = {"gr_solvetjoin", "On", 0, CV_OnOff, NULL, 0, NULL, NULL, 0, 0, NULL};
static void CV_FogDensity_ONChange(void)
{
HWD.pfnSetSpecialState(HWD_SET_FOG_DENSITY, cv_grfogdensity.value);
}
static void CV_filtermode_ONChange(void)
{
HWD.pfnSetSpecialState(HWD_SET_TEXTUREFILTERMODE, cv_grfiltermode.value);
}
static void CV_anisotropic_ONChange(void)
{
HWD.pfnSetSpecialState(HWD_SET_TEXTUREANISOTROPICMODE, cv_granisotropicmode.value);
}
/*
* lookuptable for lightvalues
* calculated as follow:
* floatlight = (1.0-exp((light^3)*gamma)) / (1.0-exp(1.0*gamma));
* gamma=-0,2;-2,0;-4,0;-6,0;-8,0
* light = 0,0 .. 1,0
*/
static const float lighttable[5][256] =
{
{
0.00000f,0.00000f,0.00000f,0.00000f,0.00000f,0.00001f,0.00001f,0.00002f,0.00003f,0.00004f,
0.00006f,0.00008f,0.00010f,0.00013f,0.00017f,0.00020f,0.00025f,0.00030f,0.00035f,0.00041f,
0.00048f,0.00056f,0.00064f,0.00073f,0.00083f,0.00094f,0.00106f,0.00119f,0.00132f,0.00147f,
0.00163f,0.00180f,0.00198f,0.00217f,0.00237f,0.00259f,0.00281f,0.00305f,0.00331f,0.00358f,
0.00386f,0.00416f,0.00447f,0.00479f,0.00514f,0.00550f,0.00587f,0.00626f,0.00667f,0.00710f,
0.00754f,0.00800f,0.00848f,0.00898f,0.00950f,0.01003f,0.01059f,0.01117f,0.01177f,0.01239f,
0.01303f,0.01369f,0.01437f,0.01508f,0.01581f,0.01656f,0.01734f,0.01814f,0.01896f,0.01981f,
0.02069f,0.02159f,0.02251f,0.02346f,0.02444f,0.02544f,0.02647f,0.02753f,0.02862f,0.02973f,
0.03088f,0.03205f,0.03325f,0.03448f,0.03575f,0.03704f,0.03836f,0.03971f,0.04110f,0.04252f,
0.04396f,0.04545f,0.04696f,0.04851f,0.05009f,0.05171f,0.05336f,0.05504f,0.05676f,0.05852f,
0.06031f,0.06214f,0.06400f,0.06590f,0.06784f,0.06981f,0.07183f,0.07388f,0.07597f,0.07810f,
0.08027f,0.08248f,0.08473f,0.08702f,0.08935f,0.09172f,0.09414f,0.09659f,0.09909f,0.10163f,
0.10421f,0.10684f,0.10951f,0.11223f,0.11499f,0.11779f,0.12064f,0.12354f,0.12648f,0.12946f,
0.13250f,0.13558f,0.13871f,0.14188f,0.14511f,0.14838f,0.15170f,0.15507f,0.15850f,0.16197f,
0.16549f,0.16906f,0.17268f,0.17635f,0.18008f,0.18386f,0.18769f,0.19157f,0.19551f,0.19950f,
0.20354f,0.20764f,0.21179f,0.21600f,0.22026f,0.22458f,0.22896f,0.23339f,0.23788f,0.24242f,
0.24702f,0.25168f,0.25640f,0.26118f,0.26602f,0.27091f,0.27587f,0.28089f,0.28596f,0.29110f,
0.29630f,0.30156f,0.30688f,0.31226f,0.31771f,0.32322f,0.32879f,0.33443f,0.34013f,0.34589f,
0.35172f,0.35761f,0.36357f,0.36960f,0.37569f,0.38185f,0.38808f,0.39437f,0.40073f,0.40716f,
0.41366f,0.42022f,0.42686f,0.43356f,0.44034f,0.44718f,0.45410f,0.46108f,0.46814f,0.47527f,
0.48247f,0.48974f,0.49709f,0.50451f,0.51200f,0.51957f,0.52721f,0.53492f,0.54271f,0.55058f,
0.55852f,0.56654f,0.57463f,0.58280f,0.59105f,0.59937f,0.60777f,0.61625f,0.62481f,0.63345f,
0.64217f,0.65096f,0.65984f,0.66880f,0.67783f,0.68695f,0.69615f,0.70544f,0.71480f,0.72425f,
0.73378f,0.74339f,0.75308f,0.76286f,0.77273f,0.78268f,0.79271f,0.80283f,0.81304f,0.82333f,
0.83371f,0.84417f,0.85472f,0.86536f,0.87609f,0.88691f,0.89781f,0.90880f,0.91989f,0.93106f,
0.94232f,0.95368f,0.96512f,0.97665f,0.98828f,1.00000f
},
{
0.00000f,0.00000f,0.00000f,0.00000f,0.00001f,0.00002f,0.00003f,0.00005f,0.00007f,0.00010f,
0.00014f,0.00019f,0.00024f,0.00031f,0.00038f,0.00047f,0.00057f,0.00069f,0.00081f,0.00096f,
0.00112f,0.00129f,0.00148f,0.00170f,0.00193f,0.00218f,0.00245f,0.00274f,0.00306f,0.00340f,
0.00376f,0.00415f,0.00456f,0.00500f,0.00547f,0.00597f,0.00649f,0.00704f,0.00763f,0.00825f,
0.00889f,0.00957f,0.01029f,0.01104f,0.01182f,0.01264f,0.01350f,0.01439f,0.01532f,0.01630f,
0.01731f,0.01836f,0.01945f,0.02058f,0.02176f,0.02298f,0.02424f,0.02555f,0.02690f,0.02830f,
0.02974f,0.03123f,0.03277f,0.03436f,0.03600f,0.03768f,0.03942f,0.04120f,0.04304f,0.04493f,
0.04687f,0.04886f,0.05091f,0.05301f,0.05517f,0.05738f,0.05964f,0.06196f,0.06434f,0.06677f,
0.06926f,0.07181f,0.07441f,0.07707f,0.07979f,0.08257f,0.08541f,0.08831f,0.09126f,0.09428f,
0.09735f,0.10048f,0.10368f,0.10693f,0.11025f,0.11362f,0.11706f,0.12056f,0.12411f,0.12773f,
0.13141f,0.13515f,0.13895f,0.14281f,0.14673f,0.15072f,0.15476f,0.15886f,0.16303f,0.16725f,
0.17153f,0.17587f,0.18028f,0.18474f,0.18926f,0.19383f,0.19847f,0.20316f,0.20791f,0.21272f,
0.21759f,0.22251f,0.22748f,0.23251f,0.23760f,0.24274f,0.24793f,0.25318f,0.25848f,0.26383f,
0.26923f,0.27468f,0.28018f,0.28573f,0.29133f,0.29697f,0.30266f,0.30840f,0.31418f,0.32001f,
0.32588f,0.33179f,0.33774f,0.34374f,0.34977f,0.35585f,0.36196f,0.36810f,0.37428f,0.38050f,
0.38675f,0.39304f,0.39935f,0.40570f,0.41207f,0.41847f,0.42490f,0.43136f,0.43784f,0.44434f,
0.45087f,0.45741f,0.46398f,0.47057f,0.47717f,0.48379f,0.49042f,0.49707f,0.50373f,0.51041f,
0.51709f,0.52378f,0.53048f,0.53718f,0.54389f,0.55061f,0.55732f,0.56404f,0.57075f,0.57747f,
0.58418f,0.59089f,0.59759f,0.60429f,0.61097f,0.61765f,0.62432f,0.63098f,0.63762f,0.64425f,
0.65086f,0.65746f,0.66404f,0.67060f,0.67714f,0.68365f,0.69015f,0.69662f,0.70307f,0.70948f,
0.71588f,0.72224f,0.72857f,0.73488f,0.74115f,0.74739f,0.75359f,0.75976f,0.76589f,0.77199f,
0.77805f,0.78407f,0.79005f,0.79599f,0.80189f,0.80774f,0.81355f,0.81932f,0.82504f,0.83072f,
0.83635f,0.84194f,0.84747f,0.85296f,0.85840f,0.86378f,0.86912f,0.87441f,0.87964f,0.88482f,
0.88995f,0.89503f,0.90005f,0.90502f,0.90993f,0.91479f,0.91959f,0.92434f,0.92903f,0.93366f,
0.93824f,0.94276f,0.94723f,0.95163f,0.95598f,0.96027f,0.96451f,0.96868f,0.97280f,0.97686f,
0.98086f,0.98481f,0.98869f,0.99252f,0.99629f,1.00000f
},
{
0.00000f,0.00000f,0.00000f,0.00001f,0.00002f,0.00003f,0.00005f,0.00008f,0.00013f,0.00018f,
0.00025f,0.00033f,0.00042f,0.00054f,0.00067f,0.00083f,0.00101f,0.00121f,0.00143f,0.00168f,
0.00196f,0.00227f,0.00261f,0.00299f,0.00339f,0.00383f,0.00431f,0.00483f,0.00538f,0.00598f,
0.00661f,0.00729f,0.00802f,0.00879f,0.00961f,0.01048f,0.01140f,0.01237f,0.01340f,0.01447f,
0.01561f,0.01680f,0.01804f,0.01935f,0.02072f,0.02215f,0.02364f,0.02520f,0.02682f,0.02850f,
0.03026f,0.03208f,0.03397f,0.03594f,0.03797f,0.04007f,0.04225f,0.04451f,0.04684f,0.04924f,
0.05172f,0.05428f,0.05691f,0.05963f,0.06242f,0.06530f,0.06825f,0.07129f,0.07441f,0.07761f,
0.08089f,0.08426f,0.08771f,0.09125f,0.09487f,0.09857f,0.10236f,0.10623f,0.11019f,0.11423f,
0.11836f,0.12257f,0.12687f,0.13125f,0.13571f,0.14027f,0.14490f,0.14962f,0.15442f,0.15931f,
0.16427f,0.16932f,0.17445f,0.17966f,0.18496f,0.19033f,0.19578f,0.20130f,0.20691f,0.21259f,
0.21834f,0.22417f,0.23007f,0.23605f,0.24209f,0.24820f,0.25438f,0.26063f,0.26694f,0.27332f,
0.27976f,0.28626f,0.29282f,0.29944f,0.30611f,0.31284f,0.31962f,0.32646f,0.33334f,0.34027f,
0.34724f,0.35426f,0.36132f,0.36842f,0.37556f,0.38273f,0.38994f,0.39718f,0.40445f,0.41174f,
0.41907f,0.42641f,0.43378f,0.44116f,0.44856f,0.45598f,0.46340f,0.47084f,0.47828f,0.48573f,
0.49319f,0.50064f,0.50809f,0.51554f,0.52298f,0.53042f,0.53784f,0.54525f,0.55265f,0.56002f,
0.56738f,0.57472f,0.58203f,0.58932f,0.59658f,0.60381f,0.61101f,0.61817f,0.62529f,0.63238f,
0.63943f,0.64643f,0.65339f,0.66031f,0.66717f,0.67399f,0.68075f,0.68746f,0.69412f,0.70072f,
0.70726f,0.71375f,0.72017f,0.72653f,0.73282f,0.73905f,0.74522f,0.75131f,0.75734f,0.76330f,
0.76918f,0.77500f,0.78074f,0.78640f,0.79199f,0.79751f,0.80295f,0.80831f,0.81359f,0.81880f,
0.82393f,0.82898f,0.83394f,0.83883f,0.84364f,0.84836f,0.85301f,0.85758f,0.86206f,0.86646f,
0.87078f,0.87502f,0.87918f,0.88326f,0.88726f,0.89118f,0.89501f,0.89877f,0.90245f,0.90605f,
0.90957f,0.91301f,0.91638f,0.91966f,0.92288f,0.92601f,0.92908f,0.93206f,0.93498f,0.93782f,
0.94059f,0.94329f,0.94592f,0.94848f,0.95097f,0.95339f,0.95575f,0.95804f,0.96027f,0.96244f,
0.96454f,0.96658f,0.96856f,0.97049f,0.97235f,0.97416f,0.97591f,0.97760f,0.97924f,0.98083f,
0.98237f,0.98386f,0.98530f,0.98669f,0.98803f,0.98933f,0.99058f,0.99179f,0.99295f,0.99408f,
0.99516f,0.99620f,0.99721f,0.99817f,0.99910f,1.00000f
},
{
0.00000f,0.00000f,0.00000f,0.00001f,0.00002f,0.00005f,0.00008f,0.00012f,0.00019f,0.00026f,
0.00036f,0.00048f,0.00063f,0.00080f,0.00099f,0.00122f,0.00148f,0.00178f,0.00211f,0.00249f,
0.00290f,0.00335f,0.00386f,0.00440f,0.00500f,0.00565f,0.00636f,0.00711f,0.00793f,0.00881f,
0.00975f,0.01075f,0.01182f,0.01295f,0.01416f,0.01543f,0.01678f,0.01821f,0.01971f,0.02129f,
0.02295f,0.02469f,0.02652f,0.02843f,0.03043f,0.03252f,0.03469f,0.03696f,0.03933f,0.04178f,
0.04433f,0.04698f,0.04973f,0.05258f,0.05552f,0.05857f,0.06172f,0.06498f,0.06834f,0.07180f,
0.07537f,0.07905f,0.08283f,0.08672f,0.09072f,0.09483f,0.09905f,0.10337f,0.10781f,0.11236f,
0.11701f,0.12178f,0.12665f,0.13163f,0.13673f,0.14193f,0.14724f,0.15265f,0.15817f,0.16380f,
0.16954f,0.17538f,0.18132f,0.18737f,0.19351f,0.19976f,0.20610f,0.21255f,0.21908f,0.22572f,
0.23244f,0.23926f,0.24616f,0.25316f,0.26023f,0.26739f,0.27464f,0.28196f,0.28935f,0.29683f,
0.30437f,0.31198f,0.31966f,0.32740f,0.33521f,0.34307f,0.35099f,0.35896f,0.36699f,0.37506f,
0.38317f,0.39133f,0.39952f,0.40775f,0.41601f,0.42429f,0.43261f,0.44094f,0.44929f,0.45766f,
0.46604f,0.47443f,0.48283f,0.49122f,0.49962f,0.50801f,0.51639f,0.52476f,0.53312f,0.54146f,
0.54978f,0.55807f,0.56633f,0.57457f,0.58277f,0.59093f,0.59905f,0.60713f,0.61516f,0.62314f,
0.63107f,0.63895f,0.64676f,0.65452f,0.66221f,0.66984f,0.67739f,0.68488f,0.69229f,0.69963f,
0.70689f,0.71407f,0.72117f,0.72818f,0.73511f,0.74195f,0.74870f,0.75536f,0.76192f,0.76839f,
0.77477f,0.78105f,0.78723f,0.79331f,0.79930f,0.80518f,0.81096f,0.81664f,0.82221f,0.82768f,
0.83305f,0.83832f,0.84347f,0.84853f,0.85348f,0.85832f,0.86306f,0.86770f,0.87223f,0.87666f,
0.88098f,0.88521f,0.88933f,0.89334f,0.89726f,0.90108f,0.90480f,0.90842f,0.91194f,0.91537f,
0.91870f,0.92193f,0.92508f,0.92813f,0.93109f,0.93396f,0.93675f,0.93945f,0.94206f,0.94459f,
0.94704f,0.94941f,0.95169f,0.95391f,0.95604f,0.95810f,0.96009f,0.96201f,0.96386f,0.96564f,
0.96735f,0.96900f,0.97059f,0.97212f,0.97358f,0.97499f,0.97634f,0.97764f,0.97888f,0.98007f,
0.98122f,0.98231f,0.98336f,0.98436f,0.98531f,0.98623f,0.98710f,0.98793f,0.98873f,0.98949f,
0.99021f,0.99090f,0.99155f,0.99218f,0.99277f,0.99333f,0.99387f,0.99437f,0.99486f,0.99531f,
0.99575f,0.99616f,0.99654f,0.99691f,0.99726f,0.99759f,0.99790f,0.99819f,0.99847f,0.99873f,
0.99897f,0.99920f,0.99942f,0.99963f,0.99982f,1.00000f
},
{
0.00000f,0.00000f,0.00000f,0.00001f,0.00003f,0.00006f,0.00010f,0.00017f,0.00025f,0.00035f,
0.00048f,0.00064f,0.00083f,0.00106f,0.00132f,0.00163f,0.00197f,0.00237f,0.00281f,0.00330f,
0.00385f,0.00446f,0.00513f,0.00585f,0.00665f,0.00751f,0.00845f,0.00945f,0.01054f,0.01170f,
0.01295f,0.01428f,0.01569f,0.01719f,0.01879f,0.02048f,0.02227f,0.02415f,0.02614f,0.02822f,
0.03042f,0.03272f,0.03513f,0.03765f,0.04028f,0.04303f,0.04589f,0.04887f,0.05198f,0.05520f,
0.05855f,0.06202f,0.06561f,0.06933f,0.07318f,0.07716f,0.08127f,0.08550f,0.08987f,0.09437f,
0.09900f,0.10376f,0.10866f,0.11369f,0.11884f,0.12414f,0.12956f,0.13512f,0.14080f,0.14662f,
0.15257f,0.15865f,0.16485f,0.17118f,0.17764f,0.18423f,0.19093f,0.19776f,0.20471f,0.21177f,
0.21895f,0.22625f,0.23365f,0.24117f,0.24879f,0.25652f,0.26435f,0.27228f,0.28030f,0.28842f,
0.29662f,0.30492f,0.31329f,0.32175f,0.33028f,0.33889f,0.34756f,0.35630f,0.36510f,0.37396f,
0.38287f,0.39183f,0.40084f,0.40989f,0.41897f,0.42809f,0.43723f,0.44640f,0.45559f,0.46479f,
0.47401f,0.48323f,0.49245f,0.50167f,0.51088f,0.52008f,0.52927f,0.53843f,0.54757f,0.55668f,
0.56575f,0.57479f,0.58379f,0.59274f,0.60164f,0.61048f,0.61927f,0.62799f,0.63665f,0.64524f,
0.65376f,0.66220f,0.67056f,0.67883f,0.68702f,0.69511f,0.70312f,0.71103f,0.71884f,0.72655f,
0.73415f,0.74165f,0.74904f,0.75632f,0.76348f,0.77053f,0.77747f,0.78428f,0.79098f,0.79756f,
0.80401f,0.81035f,0.81655f,0.82264f,0.82859f,0.83443f,0.84013f,0.84571f,0.85117f,0.85649f,
0.86169f,0.86677f,0.87172f,0.87654f,0.88124f,0.88581f,0.89026f,0.89459f,0.89880f,0.90289f,
0.90686f,0.91071f,0.91445f,0.91807f,0.92157f,0.92497f,0.92826f,0.93143f,0.93450f,0.93747f,
0.94034f,0.94310f,0.94577f,0.94833f,0.95081f,0.95319f,0.95548f,0.95768f,0.95980f,0.96183f,
0.96378f,0.96565f,0.96744f,0.96916f,0.97081f,0.97238f,0.97388f,0.97532f,0.97669f,0.97801f,
0.97926f,0.98045f,0.98158f,0.98266f,0.98369f,0.98467f,0.98560f,0.98648f,0.98732f,0.98811f,
0.98886f,0.98958f,0.99025f,0.99089f,0.99149f,0.99206f,0.99260f,0.99311f,0.99359f,0.99404f,
0.99446f,0.99486f,0.99523f,0.99559f,0.99592f,0.99623f,0.99652f,0.99679f,0.99705f,0.99729f,
0.99751f,0.99772f,0.99792f,0.99810f,0.99827f,0.99843f,0.99857f,0.99871f,0.99884f,0.99896f,
0.99907f,0.99917f,0.99926f,0.99935f,0.99943f,0.99951f,0.99958f,0.99964f,0.99970f,0.99975f,
0.99980f,0.99985f,0.99989f,0.99993f,0.99997f,1.00000f
}
};
#define gld_CalcLightLevel(lightlevel) (lighttable[1][max(min((lightlevel),255),0)])
// ==========================================================================
// VIEW GLOBALS
// ==========================================================================
// Fineangles in the SCREENWIDTH wide window.
#define FIELDOFVIEW ANGLE_90
#define ABS(x) ((x) < 0 ? -(x) : (x))
static angle_t gr_clipangle;
// The viewangletox[viewangle + FINEANGLES/4] lookup
// maps the visible view angles to screen X coordinates,
// flattening the arc to a flat projection plane.
// There will be many angles mapped to the same X.
static INT32 gr_viewangletox[FINEANGLES/2];
// The xtoviewangleangle[] table maps a screen pixel
// to the lowest viewangle that maps back to x ranges
// from clipangle to -clipangle.
static angle_t gr_xtoviewangle[MAXVIDWIDTH+1];
// ==========================================================================
// GLOBALS
// ==========================================================================
// uncomment to remove the plane rendering
#define DOPLANES
//#define DOWALLS
// test of drawing sky by polygons like in software with visplane, unfortunately
// this doesn't work since we must have z for pixel and z for texture (not like now with z = oow)
//#define POLYSKY
// test change fov when looking up/down but bsp projection messup :(
//#define NOCRAPPYMLOOK
// base values set at SetViewSize
static float gr_basecentery;
float gr_baseviewwindowy, gr_basewindowcentery;
float gr_viewwidth, gr_viewheight; // viewport clipping boundaries (screen coords)
float gr_viewwindowx;
static float gr_centerx, gr_centery;
static float gr_viewwindowy; // top left corner of view window
static float gr_windowcenterx; // center of view window, for projection
static float gr_windowcentery;
static float gr_pspritexscale, gr_pspriteyscale;
static seg_t *gr_curline;
static side_t *gr_sidedef;
static line_t *gr_linedef;
static sector_t *gr_frontsector;
static sector_t *gr_backsector;
// --------------------------------------------------------------------------
// STUFF FOR THE PROJECTION CODE
// --------------------------------------------------------------------------
FTransform atransform;
// duplicates of the main code, set after R_SetupFrame() passed them into sharedstruct,
// copied here for local use
static fixed_t dup_viewx, dup_viewy, dup_viewz;
static angle_t dup_viewangle;
static float gr_viewx, gr_viewy, gr_viewz;
static float gr_viewsin, gr_viewcos;
// Maybe not necessary with the new T&L code (needs to be checked!)
static float gr_viewludsin, gr_viewludcos; // look up down kik test
static float gr_fovlud;
// ==========================================================================
// LIGHT stuffs
// ==========================================================================
static UINT8 lightleveltonumlut[256];
// added to SRB2's sector lightlevel to make things a bit brighter (sprites/walls/planes)
FUNCMATH UINT8 LightLevelToLum(INT32 l)
{
return (UINT8)(255*gld_CalcLightLevel(l));
}
static inline void InitLumLut(void)
{
INT32 i, k = 0;
for (i = 0; i < 256; i++)
{
if (i > 128)
k += 2;
else
k = 1;
lightleveltonumlut[i] = (UINT8)(k);
}
}
//#define FOGFACTOR 300 //was 600 >> Covered by cv_grfogdensity
#define NORMALFOG 0x00000000
#define FADEFOG 0x19000000
#define CALCFOGDENSITY(x) ((float)((5220.0f*(1.0f/((x)/41.0f+1.0f)))-(5220.0f*(1.0f/(255.0f/41.0f+1.0f))))) // Approximate fog calculation based off of software walls
#define CALCFOGDENSITYFLOOR(x) ((float)((40227.0f*(1.0f/((x)/11.0f+1.0f)))-(40227.0f*(1.0f/(255.0f/11.0f+1.0f))))) // Approximate fog calculation based off of software floors
#define CALCLIGHT(x,y) ((float)(x)*((y)/255.0f))
UINT32 HWR_Lighting(INT32 light, UINT32 color, UINT32 fadecolor, boolean fogblockpoly, boolean plane)
{
RGBA_t realcolor, fogcolor, surfcolor;
INT32 alpha, fogalpha;
(void)fogblockpoly;
Negative Light Light no longer goes wrong when lightlevel of a sector is negative.
2014-03-24 17:11:45 +00:00
// Don't go out of bounds
if (light < 0)
light = 0;
else if (light > 255)
light = 255;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
realcolor.rgba = color;
fogcolor.rgba = fadecolor;
alpha = (realcolor.s.alpha*255)/25;
fogalpha = (fogcolor.s.alpha*255)/25;
if (cv_grfog.value && cv_grsoftwarefog.value) // Only do this when fog is on, software fog mode is on, and the poly is not from a fog block
{
// Modulate the colors by alpha.
realcolor.s.red = (UINT8)(CALCLIGHT(alpha,realcolor.s.red));
realcolor.s.green = (UINT8)(CALCLIGHT(alpha,realcolor.s.green));
realcolor.s.blue = (UINT8)(CALCLIGHT(alpha,realcolor.s.blue));
// Set the surface colors and further modulate the colors by light.
surfcolor.s.red = (UINT8)(CALCLIGHT((0xFF-alpha),255)+CALCLIGHT(realcolor.s.red,255));
surfcolor.s.green = (UINT8)(CALCLIGHT((0xFF-alpha),255)+CALCLIGHT(realcolor.s.green,255));
surfcolor.s.blue = (UINT8)(CALCLIGHT((0xFF-alpha),255)+CALCLIGHT(realcolor.s.blue,255));
surfcolor.s.alpha = 0xFF;
// Modulate the colors by alpha.
fogcolor.s.red = (UINT8)(CALCLIGHT(fogalpha,fogcolor.s.red));
fogcolor.s.green = (UINT8)(CALCLIGHT(fogalpha,fogcolor.s.green));
fogcolor.s.blue = (UINT8)(CALCLIGHT(fogalpha,fogcolor.s.blue));
}
else
{
// Modulate the colors by alpha.
realcolor.s.red = (UINT8)(CALCLIGHT(alpha,realcolor.s.red));
realcolor.s.green = (UINT8)(CALCLIGHT(alpha,realcolor.s.green));
realcolor.s.blue = (UINT8)(CALCLIGHT(alpha,realcolor.s.blue));
// Set the surface colors and further modulate the colors by light.
surfcolor.s.red = (UINT8)(CALCLIGHT((0xFF-alpha),light)+CALCLIGHT(realcolor.s.red,light));
surfcolor.s.green = (UINT8)(CALCLIGHT((0xFF-alpha),light)+CALCLIGHT(realcolor.s.green,light));
surfcolor.s.blue = (UINT8)(CALCLIGHT((0xFF-alpha),light)+CALCLIGHT(realcolor.s.blue,light));
// Modulate the colors by alpha.
fogcolor.s.red = (UINT8)(CALCLIGHT(fogalpha,fogcolor.s.red));
fogcolor.s.green = (UINT8)(CALCLIGHT(fogalpha,fogcolor.s.green));
fogcolor.s.blue = (UINT8)(CALCLIGHT(fogalpha,fogcolor.s.blue));
// Set the surface colors and further modulate the colors by light.
surfcolor.s.red = surfcolor.s.red+((UINT8)(CALCLIGHT((0xFF-fogalpha),(255-light))+CALCLIGHT(fogcolor.s.red,(255-light))));
surfcolor.s.green = surfcolor.s.green+((UINT8)(CALCLIGHT((0xFF-fogalpha),(255-light))+CALCLIGHT(fogcolor.s.green,(255-light))));
surfcolor.s.blue = surfcolor.s.blue+((UINT8)(CALCLIGHT((0xFF-fogalpha),(255-light))+CALCLIGHT(fogcolor.s.blue,(255-light))));
surfcolor.s.alpha = 0xFF;
}
if(cv_grfog.value)
{
if (cv_grsoftwarefog.value)
{
fogcolor.s.red = (UINT8)((CALCLIGHT(fogcolor.s.red,(255-light)))+(CALCLIGHT(realcolor.s.red,light)));
fogcolor.s.green = (UINT8)((CALCLIGHT(fogcolor.s.green,(255-light)))+(CALCLIGHT(realcolor.s.green,light)));
fogcolor.s.blue = (UINT8)((CALCLIGHT(fogcolor.s.blue,(255-light)))+(CALCLIGHT(realcolor.s.blue,light)));
// Set the fog options.
if (cv_grsoftwarefog.value == 1 && plane) // With floors, software draws them way darker for their distance
HWD.pfnSetSpecialState(HWD_SET_FOG_DENSITY, (INT32)(CALCFOGDENSITYFLOOR(light)));
else // everything else is drawn like walls
HWD.pfnSetSpecialState(HWD_SET_FOG_DENSITY, (INT32)(CALCFOGDENSITY(light)));
}
else
{
fogcolor.s.red = (UINT8)((CALCLIGHT(fogcolor.s.red,(255-light)))+(CALCLIGHT(realcolor.s.red,light)));
fogcolor.s.green = (UINT8)((CALCLIGHT(fogcolor.s.green,(255-light)))+(CALCLIGHT(realcolor.s.green,light)));
fogcolor.s.blue = (UINT8)((CALCLIGHT(fogcolor.s.blue,(255-light)))+(CALCLIGHT(realcolor.s.blue,light)));
fogalpha = (UINT8)((CALCLIGHT(fogalpha,(255-light)))+(CALCLIGHT(alpha,light)));
// Set the fog options.
light = (UINT8)(CALCLIGHT(light,(255-fogalpha)));
HWD.pfnSetSpecialState(HWD_SET_FOG_DENSITY, (INT32)(cv_grfogdensity.value-(cv_grfogdensity.value*(float)light/255.0f)));
}
HWD.pfnSetSpecialState(HWD_SET_FOG_COLOR, (fogcolor.s.red*0x10000)+(fogcolor.s.green*0x100)+fogcolor.s.blue);
HWD.pfnSetSpecialState(HWD_SET_FOG_MODE, 1);
}
return surfcolor.rgba;
}
Better fog block colouring They still aren't perfect, but now they are at least not quite so obviously just translucent polygons over the level. A mixture between partially modulating the background colours and adding the fog colour. Notably white fog blocks look like they're brightening what's behind them. Additive was also setting noalphatest before, can probably decide that depending on what it needs anyway. I don't think it's currently used anyway.
2018-03-15 23:58:37 +00:00
static UINT8 HWR_FogBlockAlpha(INT32 light, UINT32 color) // Let's see if this can work
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
Better fog block colouring They still aren't perfect, but now they are at least not quite so obviously just translucent polygons over the level. A mixture between partially modulating the background colours and adding the fog colour. Notably white fog blocks look like they're brightening what's behind them. Additive was also setting noalphatest before, can probably decide that depending on what it needs anyway. I don't think it's currently used anyway.
2018-03-15 23:58:37 +00:00
RGBA_t realcolor, surfcolor;
INT32 alpha;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
Negative Light Light no longer goes wrong when lightlevel of a sector is negative.
2014-03-24 17:11:45 +00:00
// Don't go out of bounds
if (light < 0)
light = 0;
else if (light > 255)
light = 255;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
realcolor.rgba = color;
alpha = (realcolor.s.alpha*255)/25;
Better fog block colouring They still aren't perfect, but now they are at least not quite so obviously just translucent polygons over the level. A mixture between partially modulating the background colours and adding the fog colour. Notably white fog blocks look like they're brightening what's behind them. Additive was also setting noalphatest before, can probably decide that depending on what it needs anyway. I don't think it's currently used anyway.
2018-03-15 23:58:37 +00:00
// at 255 brightness, alpha is between 0 and 127, at 0 brightness alpha will always be 255
surfcolor.s.alpha = (alpha*light)/(2*256)+255-light;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
return surfcolor.s.alpha;
}
// ==========================================================================
// FLOOR/CEILING GENERATION FROM SUBSECTORS
// ==========================================================================
#ifdef DOPLANES
// -----------------+
// HWR_RenderPlane : Render a floor or ceiling convex polygon
// -----------------+
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
static void HWR_RenderPlane(sector_t *sector, extrasubsector_t *xsub, boolean isceiling, fixed_t fixedheight,
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
FBITFIELD PolyFlags, INT32 lightlevel, lumpnum_t lumpnum, sector_t *FOFsector, UINT8 alpha, boolean fogplane, extracolormap_t *planecolormap)
{
polyvertex_t * pv;
float height; //constant y for all points on the convex flat polygon
FOutVector *v3d;
INT32 nrPlaneVerts; //verts original define of convex flat polygon
INT32 i;
float flatxref,flatyref;
float fflatsize;
INT32 flatflag;
size_t len;
float scrollx = 0.0f, scrolly = 0.0f;
angle_t angle = 0;
FSurfaceInfo Surf;
fixed_t tempxsow, tempytow;
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
#ifdef ESLOPE
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
pslope_t *slope = NULL;
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
static FOutVector *planeVerts = NULL;
static UINT16 numAllocedPlaneVerts = 0;
Merge branch 'next' into gl-slopes # Conflicts: # src/hardware/hw_main.c
2016-04-09 21:12:29 +00:00
(void)sector; ///@TODO remove shitty unused variable
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// no convex poly were generated for this subsector
if (!xsub->planepoly)
return;
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
#ifdef ESLOPE
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
// Get the slope pointer to simplify future code
if (FOFsector)
{
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
if (FOFsector->f_slope && !isceiling)
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
slope = FOFsector->f_slope;
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
else if (FOFsector->c_slope && isceiling)
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
slope = FOFsector->c_slope;
}
else
{
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
if (gr_frontsector->f_slope && !isceiling)
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
slope = gr_frontsector->f_slope;
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
else if (gr_frontsector->c_slope && isceiling)
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
slope = gr_frontsector->c_slope;
}
// Set fixedheight to the slope's height from our viewpoint, if we have a slope
if (slope)
fixedheight = P_GetZAt(slope, viewx, viewy);
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
height = FIXED_TO_FLOAT(fixedheight);
pv = xsub->planepoly->pts;
nrPlaneVerts = xsub->planepoly->numpts;
if (nrPlaneVerts < 3) //not even a triangle ?
return;
if (nrPlaneVerts > UINT16_MAX) // FIXME: exceeds plVerts size
{
CONS_Debug(DBG_RENDER, "polygon size of %d exceeds max value of %d vertices\n", nrPlaneVerts, UINT16_MAX);
return;
}
// Allocate plane-vertex buffer if we need to
if (!planeVerts || nrPlaneVerts > numAllocedPlaneVerts)
{
numAllocedPlaneVerts = (UINT16)nrPlaneVerts;
Z_Free(planeVerts);
Z_Malloc(numAllocedPlaneVerts * sizeof (FOutVector), PU_LEVEL, &planeVerts);
}
len = W_LumpLength(lumpnum);
switch (len)
{
case 4194304: // 2048x2048 lump
fflatsize = 2048.0f;
flatflag = 2047;
break;
case 1048576: // 1024x1024 lump
fflatsize = 1024.0f;
flatflag = 1023;
break;
case 262144:// 512x512 lump
fflatsize = 512.0f;
flatflag = 511;
break;
case 65536: // 256x256 lump
fflatsize = 256.0f;
flatflag = 255;
break;
case 16384: // 128x128 lump
fflatsize = 128.0f;
flatflag = 127;
break;
case 1024: // 32x32 lump
fflatsize = 32.0f;
flatflag = 31;
break;
default: // 64x64 lump
fflatsize = 64.0f;
flatflag = 63;
break;
}
// reference point for flat texture coord for each vertex around the polygon
flatxref = (float)(((fixed_t)pv->x & (~flatflag)) / fflatsize);
flatyref = (float)(((fixed_t)pv->y & (~flatflag)) / fflatsize);
// transform
v3d = planeVerts;
if (FOFsector != NULL)
{
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
if (!isceiling) // it's a floor
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
scrollx = FIXED_TO_FLOAT(FOFsector->floor_xoffs)/fflatsize;
scrolly = FIXED_TO_FLOAT(FOFsector->floor_yoffs)/fflatsize;
angle = FOFsector->floorpic_angle>>ANGLETOFINESHIFT;
}
else // it's a ceiling
{
scrollx = FIXED_TO_FLOAT(FOFsector->ceiling_xoffs)/fflatsize;
scrolly = FIXED_TO_FLOAT(FOFsector->ceiling_yoffs)/fflatsize;
angle = FOFsector->ceilingpic_angle>>ANGLETOFINESHIFT;
}
}
else if (gr_frontsector)
{
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
if (!isceiling) // it's a floor
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
scrollx = FIXED_TO_FLOAT(gr_frontsector->floor_xoffs)/fflatsize;
scrolly = FIXED_TO_FLOAT(gr_frontsector->floor_yoffs)/fflatsize;
angle = gr_frontsector->floorpic_angle>>ANGLETOFINESHIFT;
}
else // it's a ceiling
{
scrollx = FIXED_TO_FLOAT(gr_frontsector->ceiling_xoffs)/fflatsize;
scrolly = FIXED_TO_FLOAT(gr_frontsector->ceiling_yoffs)/fflatsize;
angle = gr_frontsector->ceilingpic_angle>>ANGLETOFINESHIFT;
}
}
if (angle) // Only needs to be done if there's an altered angle
{
// This needs to be done so that it scrolls in a different direction after rotation like software
tempxsow = FLOAT_TO_FIXED(scrollx);
tempytow = FLOAT_TO_FIXED(scrolly);
scrollx = (FIXED_TO_FLOAT(FixedMul(tempxsow, FINECOSINE(angle)) - FixedMul(tempytow, FINESINE(angle))));
scrolly = (FIXED_TO_FLOAT(FixedMul(tempxsow, FINESINE(angle)) + FixedMul(tempytow, FINECOSINE(angle))));
// This needs to be done so everything aligns after rotation
// It would be done so that rotation is done, THEN the translation, but I couldn't get it to rotate AND scroll like software does
tempxsow = FLOAT_TO_FIXED(flatxref);
tempytow = FLOAT_TO_FIXED(flatyref);
flatxref = (FIXED_TO_FLOAT(FixedMul(tempxsow, FINECOSINE(angle)) - FixedMul(tempytow, FINESINE(angle))));
flatyref = (FIXED_TO_FLOAT(FixedMul(tempxsow, FINESINE(angle)) + FixedMul(tempytow, FINECOSINE(angle))));
}
for (i = 0; i < nrPlaneVerts; i++,v3d++,pv++)
{
// Hurdler: add scrolling texture on floor/ceiling
v3d->sow = (float)((pv->x / fflatsize) - flatxref + scrollx);
v3d->tow = (float)(flatyref - (pv->y / fflatsize) + scrolly);
//v3d->sow = (float)(pv->x / fflatsize);
//v3d->tow = (float)(pv->y / fflatsize);
// Need to rotate before translate
if (angle) // Only needs to be done if there's an altered angle
{
tempxsow = FLOAT_TO_FIXED(v3d->sow);
tempytow = FLOAT_TO_FIXED(v3d->tow);
v3d->sow = (FIXED_TO_FLOAT(FixedMul(tempxsow, FINECOSINE(angle)) - FixedMul(tempytow, FINESINE(angle))));
v3d->tow = (FIXED_TO_FLOAT(-FixedMul(tempxsow, FINESINE(angle)) - FixedMul(tempytow, FINECOSINE(angle))));
}
//v3d->sow = (float)(v3d->sow - flatxref + scrollx);
//v3d->tow = (float)(flatyref - v3d->tow + scrolly);
v3d->x = pv->x;
v3d->y = height;
v3d->z = pv->y;
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
#ifdef ESLOPE
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
if (slope)
{
fixedheight = P_GetZAt(slope, FLOAT_TO_FIXED(pv->x), FLOAT_TO_FIXED(pv->y));
v3d->y = FIXED_TO_FLOAT(fixedheight);
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
#endif
}
// only useful for flat coloured triangles
//Surf.FlatColor = 0xff804020;
// use different light tables
// for horizontal / vertical / diagonal
// note: try to get the same visual feel as the original
Surf.FlatColor.s.red = Surf.FlatColor.s.green =
Surf.FlatColor.s.blue = LightLevelToLum(lightlevel); // Don't take from the frontsector, or the game will crash
#if 0 // no colormap test
// colormap test
if (gr_frontsector)
{
sector_t *psector = gr_frontsector;
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
#ifdef ESLOPE
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
if (slope)
fixedheight = P_GetZAt(slope, psector->soundorg.x, psector->soundorg.y);
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (psector->ffloors)
{
ffloor_t *caster = psector->lightlist[R_GetPlaneLight(psector, fixedheight, false)].caster;
psector = caster ? &sectors[caster->secnum] : psector;
if (caster)
{
lightlevel = psector->lightlevel;
Surf.FlatColor.s.red = Surf.FlatColor.s.green = Surf.FlatColor.s.blue = LightLevelToLum(lightlevel);
}
}
if (psector->extra_colormap)
LightLevelToLum gone LightLevelToLum unused by everything now, it seemed to be giving wrong values for lightlevels.
2014-03-24 16:11:25 +00:00
Surf.FlatColor.rgba = HWR_Lighting(lightlevel,psector->extra_colormap->rgba,psector->extra_colormap->fadergba, false, true);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
else
LightLevelToLum gone LightLevelToLum unused by everything now, it seemed to be giving wrong values for lightlevels.
2014-03-24 16:11:25 +00:00
Surf.FlatColor.rgba = HWR_Lighting(lightlevel,NORMALFOG,FADEFOG, false, true);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
else
LightLevelToLum gone LightLevelToLum unused by everything now, it seemed to be giving wrong values for lightlevels.
2014-03-24 16:11:25 +00:00
Surf.FlatColor.rgba = HWR_Lighting(lightlevel,NORMALFOG,FADEFOG, false, true);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
#endif // NOPE
if (planecolormap)
{
if (fogplane)
LightLevelToLum gone LightLevelToLum unused by everything now, it seemed to be giving wrong values for lightlevels.
2014-03-24 16:11:25 +00:00
Surf.FlatColor.rgba = HWR_Lighting(lightlevel, planecolormap->rgba, planecolormap->fadergba, true, false);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
else
LightLevelToLum gone LightLevelToLum unused by everything now, it seemed to be giving wrong values for lightlevels.
2014-03-24 16:11:25 +00:00
Surf.FlatColor.rgba = HWR_Lighting(lightlevel, planecolormap->rgba, planecolormap->fadergba, false, true);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
else
{
if (fogplane)
LightLevelToLum gone LightLevelToLum unused by everything now, it seemed to be giving wrong values for lightlevels.
2014-03-24 16:11:25 +00:00
Surf.FlatColor.rgba = HWR_Lighting(lightlevel, NORMALFOG, FADEFOG, true, false);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
else
LightLevelToLum gone LightLevelToLum unused by everything now, it seemed to be giving wrong values for lightlevels.
2014-03-24 16:11:25 +00:00
Surf.FlatColor.rgba = HWR_Lighting(lightlevel, NORMALFOG, FADEFOG, false, true);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
Better fog block colouring They still aren't perfect, but now they are at least not quite so obviously just translucent polygons over the level. A mixture between partially modulating the background colours and adding the fog colour. Notably white fog blocks look like they're brightening what's behind them. Additive was also setting noalphatest before, can probably decide that depending on what it needs anyway. I don't think it's currently used anyway.
2018-03-15 23:58:37 +00:00
if (PolyFlags & (PF_Translucent|PF_Fog))
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
Surf.FlatColor.s.alpha = (UINT8)alpha;
Translucent floors shouldn't write into the depth buffer
2018-03-15 23:59:01 +00:00
PolyFlags |= PF_Modulated|PF_Clip;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
else
PolyFlags |= PF_Masked|PF_Modulated|PF_Clip;
HWD.pfnDrawPolygon(&Surf, planeVerts, nrPlaneVerts, PolyFlags);
#ifdef ALAM_LIGHTING
// add here code for dynamic lighting on planes
HWR_PlaneLighting(planeVerts, nrPlaneVerts);
#endif
}
#ifdef POLYSKY
// this don't draw anything it only update the z-buffer so there isn't problem with
// wall/things upper that sky (map12)
static void HWR_RenderSkyPlane(extrasubsector_t *xsub, fixed_t fixedheight)
{
polyvertex_t *pv;
float height; //constant y for all points on the convex flat polygon
FOutVector *v3d;
INT32 nrPlaneVerts; //verts original define of convex flat polygon
INT32 i;
// no convex poly were generated for this subsector
if (!xsub->planepoly)
return;
height = FIXED_TO_FLOAT(fixedheight);
pv = xsub->planepoly->pts;
nrPlaneVerts = xsub->planepoly->numpts;
if (nrPlaneVerts < 3) // not even a triangle?
return;
if (nrPlaneVerts > MAXPLANEVERTICES) // FIXME: exceeds plVerts size
{
CONS_Debug(DBG_RENDER, "polygon size of %d exceeds max value of %d vertices\n", nrPlaneVerts, MAXPLANEVERTICES);
return;
}
// transform
v3d = planeVerts;
for (i = 0; i < nrPlaneVerts; i++,v3d++,pv++)
{
v3d->sow = 0.0f;
v3d->tow = 0.0f;
v3d->x = pv->x;
v3d->y = height;
v3d->z = pv->y;
}
HWD.pfnDrawPolygon(NULL, planeVerts, nrPlaneVerts,
PF_Clip|PF_Invisible|PF_NoTexture|PF_Occlude);
}
#endif //polysky
#endif //doplanes
/*
wallVerts order is :
3--2
| /|
|/ |
0--1
*/
#ifdef WALLSPLATS
static void HWR_DrawSegsSplats(FSurfaceInfo * pSurf)
{
FOutVector trVerts[4], *wv;
wallVert3D wallVerts[4];
wallVert3D *pwallVerts;
wallsplat_t *splat;
GLPatch_t *gpatch;
fixed_t i;
FSurfaceInfo pSurf2;
// seg bbox
fixed_t segbbox[4];
M_ClearBox(segbbox);
M_AddToBox(segbbox,
Make new pv1/pv2 seg pointers, so AdjustSeg doesn't modify the v1/v2 pointers directly anymore Yes I know they're void * in r_defs.h's seg_t definition, it's quicker than trying to figure out if including hardware/hw_glob.h is a good idea or not
2016-11-07 21:55:56 +00:00
FLOAT_TO_FIXED(((polyvertex_t *)gr_curline->pv1)->x),
FLOAT_TO_FIXED(((polyvertex_t *)gr_curline->pv1)->y));
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
M_AddToBox(segbbox,
Make new pv1/pv2 seg pointers, so AdjustSeg doesn't modify the v1/v2 pointers directly anymore Yes I know they're void * in r_defs.h's seg_t definition, it's quicker than trying to figure out if including hardware/hw_glob.h is a good idea or not
2016-11-07 21:55:56 +00:00
FLOAT_TO_FIXED(((polyvertex_t *)gr_curline->pv2)->x),
FLOAT_TO_FIXED(((polyvertex_t *)gr_curline->pv2)->y));
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
splat = (wallsplat_t *)gr_curline->linedef->splats;
for (; splat; splat = splat->next)
{
//BP: don't draw splat extern to this seg
// this is quick fix best is explain in logboris.txt at 12-4-2000
if (!M_PointInBox(segbbox,splat->v1.x,splat->v1.y) && !M_PointInBox(segbbox,splat->v2.x,splat->v2.y))
continue;
gpatch = W_CachePatchNum(splat->patch, PU_CACHE);
HWR_GetPatch(gpatch);
wallVerts[0].x = wallVerts[3].x = FIXED_TO_FLOAT(splat->v1.x);
wallVerts[0].z = wallVerts[3].z = FIXED_TO_FLOAT(splat->v1.y);
wallVerts[2].x = wallVerts[1].x = FIXED_TO_FLOAT(splat->v2.x);
wallVerts[2].z = wallVerts[1].z = FIXED_TO_FLOAT(splat->v2.y);
i = splat->top;
if (splat->yoffset)
i += *splat->yoffset;
wallVerts[2].y = wallVerts[3].y = FIXED_TO_FLOAT(i)+(gpatch->height>>1);
wallVerts[0].y = wallVerts[1].y = FIXED_TO_FLOAT(i)-(gpatch->height>>1);
wallVerts[3].s = wallVerts[3].t = wallVerts[2].s = wallVerts[0].t = 0.0f;
wallVerts[1].s = wallVerts[1].t = wallVerts[2].t = wallVerts[0].s = 1.0f;
// transform
wv = trVerts;
pwallVerts = wallVerts;
for (i = 0; i < 4; i++,wv++,pwallVerts++)
{
wv->x = pwallVerts->x;
wv->z = pwallVerts->z;
wv->y = pwallVerts->y;
// Kalaron: TOW and SOW needed to be switched
wv->sow = pwallVerts->t;
wv->tow = pwallVerts->s;
}
M_Memcpy(&pSurf2,pSurf,sizeof (FSurfaceInfo));
switch (splat->flags & SPLATDRAWMODE_MASK)
{
case SPLATDRAWMODE_OPAQUE :
pSurf2.FlatColor.s.alpha = 0xff;
i = PF_Translucent;
break;
case SPLATDRAWMODE_TRANS :
pSurf2.FlatColor.s.alpha = 128;
i = PF_Translucent;
break;
case SPLATDRAWMODE_SHADE :
pSurf2.FlatColor.s.alpha = 0xff;
i = PF_Substractive;
break;
}
HWD.pfnDrawPolygon(&pSurf2, trVerts, 4, i|PF_Modulated|PF_Clip|PF_Decal);
}
}
#endif
// ==========================================================================
// WALL GENERATION FROM SUBSECTOR SEGS
// ==========================================================================
FBITFIELD HWR_TranstableToAlpha(INT32 transtablenum, FSurfaceInfo *pSurf)
{
switch (transtablenum)
{
case tr_trans10 : pSurf->FlatColor.s.alpha = 0xe6;return PF_Translucent;
case tr_trans20 : pSurf->FlatColor.s.alpha = 0xcc;return PF_Translucent;
case tr_trans30 : pSurf->FlatColor.s.alpha = 0xb3;return PF_Translucent;
case tr_trans40 : pSurf->FlatColor.s.alpha = 0x99;return PF_Translucent;
case tr_trans50 : pSurf->FlatColor.s.alpha = 0x80;return PF_Translucent;
case tr_trans60 : pSurf->FlatColor.s.alpha = 0x66;return PF_Translucent;
case tr_trans70 : pSurf->FlatColor.s.alpha = 0x4c;return PF_Translucent;
case tr_trans80 : pSurf->FlatColor.s.alpha = 0x33;return PF_Translucent;
case tr_trans90 : pSurf->FlatColor.s.alpha = 0x19;return PF_Translucent;
}
return PF_Translucent;
}
// v1,v2 : the start & end vertices along the original wall segment, that may have been
// clipped so that only a visible portion of the wall seg is drawn.
// floorheight, ceilingheight : depend on wall upper/lower/middle, comes from the sectors.
static void HWR_AddTransparentWall(wallVert3D *wallVerts, FSurfaceInfo * pSurf, INT32 texnum, FBITFIELD blend, boolean fogwall, INT32 lightlevel, extracolormap_t *wallcolormap);
// -----------------+
// HWR_ProjectWall :
// -----------------+
/*
wallVerts order is :
3--2
| /|
|/ |
0--1
*/
static void HWR_ProjectWall(wallVert3D * wallVerts,
FSurfaceInfo * pSurf,
FBITFIELD blendmode, INT32 lightlevel, extracolormap_t *wallcolormap)
{
FOutVector trVerts[4];
FOutVector *wv;
// transform
wv = trVerts;
// it sounds really stupid to do this conversion with the new T&L code
// we should directly put the right information in the right structure
// wallVerts3D seems ok, doesn't need FOutVector
// also remove the light copy
// More messy to unwrap, but it's also quicker, uses less memory.
wv->sow = wallVerts->s;
wv->tow = wallVerts->t;
wv->x = wallVerts->x;
wv->y = wallVerts->y;
wv->z = wallVerts->z;
wv++; wallVerts++;
wv->sow = wallVerts->s;
wv->tow = wallVerts->t;
wv->x = wallVerts->x;
wv->y = wallVerts->y;
wv->z = wallVerts->z;
wv++; wallVerts++;
wv->sow = wallVerts->s;
wv->tow = wallVerts->t;
wv->x = wallVerts->x;
wv->y = wallVerts->y;
wv->z = wallVerts->z;
wv++; wallVerts++;
wv->sow = wallVerts->s;
wv->tow = wallVerts->t;
wv->x = wallVerts->x;
wv->y = wallVerts->y;
wv->z = wallVerts->z;
if (wallcolormap)
{
LightLevelToLum gone LightLevelToLum unused by everything now, it seemed to be giving wrong values for lightlevels.
2014-03-24 16:11:25 +00:00
pSurf->FlatColor.rgba = HWR_Lighting(lightlevel, wallcolormap->rgba, wallcolormap->fadergba, false, false);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
else
{
LightLevelToLum gone LightLevelToLum unused by everything now, it seemed to be giving wrong values for lightlevels.
2014-03-24 16:11:25 +00:00
pSurf->FlatColor.rgba = HWR_Lighting(lightlevel, NORMALFOG, FADEFOG, false, false);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
HWD.pfnDrawPolygon(pSurf, trVerts, 4, blendmode|PF_Modulated|PF_Occlude|PF_Clip);
#ifdef WALLSPLATS
if (gr_curline->linedef->splats && cv_splats.value)
HWR_DrawSegsSplats(pSurf);
#endif
#ifdef ALAM_LIGHTING
//Hurdler: TDOD: do static lighting using gr_curline->lm
HWR_WallLighting(trVerts);
//Hurdler: for better dynamic light in dark area, we should draw the light first
// and then the wall all that with the right blending func
//HWD.pfnDrawPolygon(pSurf, trVerts, 4, PF_Additive|PF_Modulated|PF_Occlude|PF_Clip);
#endif
}
// ==========================================================================
// BSP, CULL, ETC..
// ==========================================================================
// return the frac from the interception of the clipping line
// (in fact a clipping plane that has a constant, so can clip with simple 2d)
// with the wall segment
//
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#ifndef NEWCLIP
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
static float HWR_ClipViewSegment(INT32 x, polyvertex_t *v1, polyvertex_t *v2)
{
float num, den;
float v1x, v1y, v1dx, v1dy, v2dx, v2dy;
angle_t pclipangle = gr_xtoviewangle[x];
// a segment of a polygon
v1x = v1->x;
v1y = v1->y;
v1dx = (v2->x - v1->x);
v1dy = (v2->y - v1->y);
// the clipping line
pclipangle = pclipangle + dup_viewangle; //back to normal angle (non-relative)
v2dx = FIXED_TO_FLOAT(FINECOSINE(pclipangle>>ANGLETOFINESHIFT));
v2dy = FIXED_TO_FLOAT(FINESINE(pclipangle>>ANGLETOFINESHIFT));
den = v2dy*v1dx - v2dx*v1dy;
if (den == 0)
return -1; // parallel
// calc the frac along the polygon segment,
//num = (v2x - v1x)*v2dy + (v1y - v2y)*v2dx;
//num = -v1x * v2dy + v1y * v2dx;
num = (gr_viewx - v1x)*v2dy + (v1y - gr_viewy)*v2dx;
return num / den;
}
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
//
// HWR_SplitWall
//
Clean up warnings
2018-11-14 15:52:16 +00:00
static void HWR_SplitWall(sector_t *sector, wallVert3D *wallVerts, INT32 texnum, FSurfaceInfo* Surf, INT32 cutflag, ffloor_t *pfloor)
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
/* SoM: split up and light walls according to the
lightlist. This may also include leaving out parts
of the wall that can't be seen */
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
float realtop, realbot, top, bot;
float pegt, pegb, pegmul;
float height = 0.0f, bheight = 0.0f;
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#ifdef ESLOPE
float endrealtop, endrealbot, endtop, endbot;
float endpegt, endpegb, endpegmul;
float endheight = 0.0f, endbheight = 0.0f;
fixed_t v1x = FLOAT_TO_FIXED(wallVerts[0].x);
wallVerts[].y is actually the map z coord, so use wallVerts[].z instead (which is actually the map y coord) Don't worry I'm not going mad, this is actually how it's supposed to be
2017-06-03 13:46:03 +00:00
fixed_t v1y = FLOAT_TO_FIXED(wallVerts[0].z); // not a typo
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
fixed_t v2x = FLOAT_TO_FIXED(wallVerts[1].x);
wallVerts[].y is actually the map z coord, so use wallVerts[].z instead (which is actually the map y coord) Don't worry I'm not going mad, this is actually how it's supposed to be
2017-06-03 13:46:03 +00:00
fixed_t v2y = FLOAT_TO_FIXED(wallVerts[1].z); // not a typo
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
// compiler complains when P_GetZAt is used in FLOAT_TO_FIXED directly
// use this as a temp var to store P_GetZAt's return value each time
fixed_t temp;
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
INT32 solid, i;
lightlist_t * list = sector->lightlist;
const UINT8 alpha = Surf->FlatColor.s.alpha;
Start with lightnum on sector lightlevel
2018-03-30 17:13:52 +00:00
FUINT lightnum = sector->lightlevel;
Some fixes and updates for HWR_SplitWall Solid walls *can* be cut Fix issues with water and fog FOFs not cutting each other out correctly Fix Fog colourmap and lighting setting that is done here. Remove HWR_SplitFog There is currently a bug with FF_DOUBLESHADOW (that also exists in software) but has a larger impact here. When 2 FF_DOUBLESHADOW lights are directly stacked on each other the bottom one has its height set incorrectly. This causes all the Fog in the timed gravity flipping section of ERZ2 to be drawn and it looks really bad.
2018-03-20 14:20:02 +00:00
extracolormap_t *colormap = NULL;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
realtop = top = wallVerts[3].y;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
realbot = bot = wallVerts[0].y;
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
pegt = wallVerts[3].t;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
pegb = wallVerts[0].t;
pegmul = (pegb - pegt) / (top - bot);
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#ifdef ESLOPE
endrealtop = endtop = wallVerts[2].y;
endrealbot = endbot = wallVerts[1].y;
endpegt = wallVerts[2].t;
endpegb = wallVerts[1].t;
endpegmul = (endpegb - endpegt) / (endtop - endbot);
#endif
Some fixes and updates for HWR_SplitWall Solid walls *can* be cut Fix issues with water and fog FOFs not cutting each other out correctly Fix Fog colourmap and lighting setting that is done here. Remove HWR_SplitFog There is currently a bug with FF_DOUBLESHADOW (that also exists in software) but has a larger impact here. When 2 FF_DOUBLESHADOW lights are directly stacked on each other the bottom one has its height set incorrectly. This causes all the Fog in the timed gravity flipping section of ERZ2 to be drawn and it looks really bad.
2018-03-20 14:20:02 +00:00
for (i = 0; i < sector->numlights; i++)
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#ifdef ESLOPE
if (endtop < endrealbot)
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (top < realbot)
return;
Some fixes and updates for HWR_SplitWall Solid walls *can* be cut Fix issues with water and fog FOFs not cutting each other out correctly Fix Fog colourmap and lighting setting that is done here. Remove HWR_SplitFog There is currently a bug with FF_DOUBLESHADOW (that also exists in software) but has a larger impact here. When 2 FF_DOUBLESHADOW lights are directly stacked on each other the bottom one has its height set incorrectly. This causes all the Fog in the timed gravity flipping section of ERZ2 to be drawn and it looks really bad.
2018-03-20 14:20:02 +00:00
// There's a compiler warning here if this comment isn't here because of indentation
if (!(list[i].flags & FF_NOSHADE))
{
if (pfloor && (pfloor->flags & FF_FOG))
{
lightnum = pfloor->master->frontsector->lightlevel;
colormap = pfloor->master->frontsector->extra_colormap;
}
else
{
lightnum = *list[i].lightlevel;
colormap = list[i].extra_colormap;
}
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
solid = false;
Some fixes and updates for HWR_SplitWall Solid walls *can* be cut Fix issues with water and fog FOFs not cutting each other out correctly Fix Fog colourmap and lighting setting that is done here. Remove HWR_SplitFog There is currently a bug with FF_DOUBLESHADOW (that also exists in software) but has a larger impact here. When 2 FF_DOUBLESHADOW lights are directly stacked on each other the bottom one has its height set incorrectly. This causes all the Fog in the timed gravity flipping section of ERZ2 to be drawn and it looks really bad.
2018-03-20 14:20:02 +00:00
if ((sector->lightlist[i].flags & FF_CUTSOLIDS) && !(cutflag & FF_EXTRA))
solid = true;
else if ((sector->lightlist[i].flags & FF_CUTEXTRA) && (cutflag & FF_EXTRA))
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
Some fixes and updates for HWR_SplitWall Solid walls *can* be cut Fix issues with water and fog FOFs not cutting each other out correctly Fix Fog colourmap and lighting setting that is done here. Remove HWR_SplitFog There is currently a bug with FF_DOUBLESHADOW (that also exists in software) but has a larger impact here. When 2 FF_DOUBLESHADOW lights are directly stacked on each other the bottom one has its height set incorrectly. This causes all the Fog in the timed gravity flipping section of ERZ2 to be drawn and it looks really bad.
2018-03-20 14:20:02 +00:00
if (sector->lightlist[i].flags & FF_EXTRA)
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
Some fixes and updates for HWR_SplitWall Solid walls *can* be cut Fix issues with water and fog FOFs not cutting each other out correctly Fix Fog colourmap and lighting setting that is done here. Remove HWR_SplitFog There is currently a bug with FF_DOUBLESHADOW (that also exists in software) but has a larger impact here. When 2 FF_DOUBLESHADOW lights are directly stacked on each other the bottom one has its height set incorrectly. This causes all the Fog in the timed gravity flipping section of ERZ2 to be drawn and it looks really bad.
2018-03-20 14:20:02 +00:00
if ((sector->lightlist[i].flags & (FF_FOG|FF_SWIMMABLE)) == (cutflag & (FF_FOG|FF_SWIMMABLE))) // Only merge with your own types
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
solid = true;
}
else
Some fixes and updates for HWR_SplitWall Solid walls *can* be cut Fix issues with water and fog FOFs not cutting each other out correctly Fix Fog colourmap and lighting setting that is done here. Remove HWR_SplitFog There is currently a bug with FF_DOUBLESHADOW (that also exists in software) but has a larger impact here. When 2 FF_DOUBLESHADOW lights are directly stacked on each other the bottom one has its height set incorrectly. This causes all the Fog in the timed gravity flipping section of ERZ2 to be drawn and it looks really bad.
2018-03-20 14:20:02 +00:00
solid = true;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
else
solid = false;
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#ifdef ESLOPE
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
if (list[i].slope)
{
temp = P_GetZAt(list[i].slope, v1x, v1y);
height = FIXED_TO_FLOAT(temp);
temp = P_GetZAt(list[i].slope, v2x, v2y);
endheight = FIXED_TO_FLOAT(temp);
}
else
height = endheight = FIXED_TO_FLOAT(list[i].height);
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
if (solid)
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
{
if (*list[i].caster->b_slope)
{
temp = P_GetZAt(*list[i].caster->b_slope, v1x, v1y);
bheight = FIXED_TO_FLOAT(temp);
temp = P_GetZAt(*list[i].caster->b_slope, v2x, v2y);
endbheight = FIXED_TO_FLOAT(temp);
}
else
bheight = endbheight = FIXED_TO_FLOAT(*list[i].caster->bottomheight);
}
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#else
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
height = FIXED_TO_FLOAT(list[i].height);
if (solid)
bheight = FIXED_TO_FLOAT(*list[i].caster->bottomheight);
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#ifdef ESLOPE
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
if (endheight >= endtop)
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (height >= top)
{
if (solid && top > bheight)
top = bheight;
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#ifdef ESLOPE
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
if (solid && endtop > endbheight)
endtop = endbheight;
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
Some fixes and updates for HWR_SplitWall Solid walls *can* be cut Fix issues with water and fog FOFs not cutting each other out correctly Fix Fog colourmap and lighting setting that is done here. Remove HWR_SplitFog There is currently a bug with FF_DOUBLESHADOW (that also exists in software) but has a larger impact here. When 2 FF_DOUBLESHADOW lights are directly stacked on each other the bottom one has its height set incorrectly. This causes all the Fog in the timed gravity flipping section of ERZ2 to be drawn and it looks really bad.
2018-03-20 14:20:02 +00:00
#ifdef ESLOPE
if (i + 1 < sector->numlights)
{
if (list[i+1].slope)
{
temp = P_GetZAt(list[i+1].slope, v1x, v1y);
bheight = FIXED_TO_FLOAT(temp);
temp = P_GetZAt(list[i+1].slope, v2x, v2y);
endbheight = FIXED_TO_FLOAT(temp);
}
else
bheight = endbheight = FIXED_TO_FLOAT(list[i+1].height);
}
else
{
bheight = realbot;
endbheight = endrealbot;
}
#else
if (i + 1 < sector->numlights)
{
bheight = FIXED_TO_FLOAT(list[i+1].height);
}
else
{
bheight = realbot;
}
#endif
OpenGL Sprite Splitting
2018-03-22 00:52:14 +00:00
#ifdef ESLOPE
Some fixes and updates for HWR_SplitWall Solid walls *can* be cut Fix issues with water and fog FOFs not cutting each other out correctly Fix Fog colourmap and lighting setting that is done here. Remove HWR_SplitFog There is currently a bug with FF_DOUBLESHADOW (that also exists in software) but has a larger impact here. When 2 FF_DOUBLESHADOW lights are directly stacked on each other the bottom one has its height set incorrectly. This causes all the Fog in the timed gravity flipping section of ERZ2 to be drawn and it looks really bad.
2018-03-20 14:20:02 +00:00
if (endbheight >= endtop)
OpenGL Sprite Splitting
2018-03-22 00:52:14 +00:00
#endif
Some fixes and updates for HWR_SplitWall Solid walls *can* be cut Fix issues with water and fog FOFs not cutting each other out correctly Fix Fog colourmap and lighting setting that is done here. Remove HWR_SplitFog There is currently a bug with FF_DOUBLESHADOW (that also exists in software) but has a larger impact here. When 2 FF_DOUBLESHADOW lights are directly stacked on each other the bottom one has its height set incorrectly. This causes all the Fog in the timed gravity flipping section of ERZ2 to be drawn and it looks really bad.
2018-03-20 14:20:02 +00:00
if (bheight >= top)
continue;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
//Found a break;
Some fixes and updates for HWR_SplitWall Solid walls *can* be cut Fix issues with water and fog FOFs not cutting each other out correctly Fix Fog colourmap and lighting setting that is done here. Remove HWR_SplitFog There is currently a bug with FF_DOUBLESHADOW (that also exists in software) but has a larger impact here. When 2 FF_DOUBLESHADOW lights are directly stacked on each other the bottom one has its height set incorrectly. This causes all the Fog in the timed gravity flipping section of ERZ2 to be drawn and it looks really bad.
2018-03-20 14:20:02 +00:00
bot = bheight;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (bot < realbot)
bot = realbot;
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#ifdef ESLOPE
Some fixes and updates for HWR_SplitWall Solid walls *can* be cut Fix issues with water and fog FOFs not cutting each other out correctly Fix Fog colourmap and lighting setting that is done here. Remove HWR_SplitFog There is currently a bug with FF_DOUBLESHADOW (that also exists in software) but has a larger impact here. When 2 FF_DOUBLESHADOW lights are directly stacked on each other the bottom one has its height set incorrectly. This causes all the Fog in the timed gravity flipping section of ERZ2 to be drawn and it looks really bad.
2018-03-20 14:20:02 +00:00
endbot = endbheight;
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
if (endbot < endrealbot)
endbot = endrealbot;
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
Surf->FlatColor.s.alpha = alpha;
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#ifdef ESLOPE
wallVerts[3].t = pegt + ((realtop - top) * pegmul);
wallVerts[2].t = endpegt + ((endrealtop - endtop) * endpegmul);
wallVerts[0].t = pegt + ((realtop - bot) * pegmul);
wallVerts[1].t = endpegt + ((endrealtop - endbot) * endpegmul);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
// set top/bottom coords
wallVerts[3].y = top;
wallVerts[2].y = endtop;
wallVerts[0].y = bot;
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
wallVerts[1].y = endbot;
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#else
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
wallVerts[3].t = wallVerts[2].t = pegt + ((realtop - top) * pegmul);
wallVerts[0].t = wallVerts[1].t = pegt + ((realtop - bot) * pegmul);
// set top/bottom coords
wallVerts[2].y = wallVerts[3].y = top;
wallVerts[0].y = wallVerts[1].y = bot;
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
Some fixes and updates for HWR_SplitWall Solid walls *can* be cut Fix issues with water and fog FOFs not cutting each other out correctly Fix Fog colourmap and lighting setting that is done here. Remove HWR_SplitFog There is currently a bug with FF_DOUBLESHADOW (that also exists in software) but has a larger impact here. When 2 FF_DOUBLESHADOW lights are directly stacked on each other the bottom one has its height set incorrectly. This causes all the Fog in the timed gravity flipping section of ERZ2 to be drawn and it looks really bad.
2018-03-20 14:20:02 +00:00
if (cutflag & FF_FOG)
HWR_AddTransparentWall(wallVerts, Surf, texnum, PF_Fog|PF_NoTexture, true, lightnum, colormap);
else if (cutflag & FF_TRANSLUCENT)
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
HWR_AddTransparentWall(wallVerts, Surf, texnum, PF_Translucent, false, lightnum, colormap);
else
HWR_ProjectWall(wallVerts, Surf, PF_Masked, lightnum, colormap);
Some fixes and updates for HWR_SplitWall Solid walls *can* be cut Fix issues with water and fog FOFs not cutting each other out correctly Fix Fog colourmap and lighting setting that is done here. Remove HWR_SplitFog There is currently a bug with FF_DOUBLESHADOW (that also exists in software) but has a larger impact here. When 2 FF_DOUBLESHADOW lights are directly stacked on each other the bottom one has its height set incorrectly. This causes all the Fog in the timed gravity flipping section of ERZ2 to be drawn and it looks really bad.
2018-03-20 14:20:02 +00:00
top = bot;
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#ifdef ESLOPE
Some fixes and updates for HWR_SplitWall Solid walls *can* be cut Fix issues with water and fog FOFs not cutting each other out correctly Fix Fog colourmap and lighting setting that is done here. Remove HWR_SplitFog There is currently a bug with FF_DOUBLESHADOW (that also exists in software) but has a larger impact here. When 2 FF_DOUBLESHADOW lights are directly stacked on each other the bottom one has its height set incorrectly. This causes all the Fog in the timed gravity flipping section of ERZ2 to be drawn and it looks really bad.
2018-03-20 14:20:02 +00:00
endtop = endbot;
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
bot = realbot;
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#ifdef ESLOPE
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
endbot = endrealbot;
if (endtop <= endrealbot)
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (top <= realbot)
return;
Some fixes and updates for HWR_SplitWall Solid walls *can* be cut Fix issues with water and fog FOFs not cutting each other out correctly Fix Fog colourmap and lighting setting that is done here. Remove HWR_SplitFog There is currently a bug with FF_DOUBLESHADOW (that also exists in software) but has a larger impact here. When 2 FF_DOUBLESHADOW lights are directly stacked on each other the bottom one has its height set incorrectly. This causes all the Fog in the timed gravity flipping section of ERZ2 to be drawn and it looks really bad.
2018-03-20 14:20:02 +00:00
Surf->FlatColor.s.alpha = alpha;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#ifdef ESLOPE
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
wallVerts[3].t = pegt + ((realtop - top) * pegmul);
wallVerts[2].t = endpegt + ((endrealtop - endtop) * endpegmul);
wallVerts[0].t = pegt + ((realtop - bot) * pegmul);
wallVerts[1].t = endpegt + ((endrealtop - endbot) * endpegmul);
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
// set top/bottom coords
wallVerts[3].y = top;
wallVerts[2].y = endtop;
wallVerts[0].y = bot;
wallVerts[1].y = endbot;
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#else
wallVerts[3].t = wallVerts[2].t = pegt + ((realtop - top) * pegmul);
wallVerts[0].t = wallVerts[1].t = pegt + ((realtop - bot) * pegmul);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
// set top/bottom coords
wallVerts[2].y = wallVerts[3].y = top;
wallVerts[0].y = wallVerts[1].y = bot;
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
Some fixes and updates for HWR_SplitWall Solid walls *can* be cut Fix issues with water and fog FOFs not cutting each other out correctly Fix Fog colourmap and lighting setting that is done here. Remove HWR_SplitFog There is currently a bug with FF_DOUBLESHADOW (that also exists in software) but has a larger impact here. When 2 FF_DOUBLESHADOW lights are directly stacked on each other the bottom one has its height set incorrectly. This causes all the Fog in the timed gravity flipping section of ERZ2 to be drawn and it looks really bad.
2018-03-20 14:20:02 +00:00
if (cutflag & FF_FOG)
HWR_AddTransparentWall(wallVerts, Surf, texnum, PF_Fog|PF_NoTexture, true, lightnum, colormap);
else if (cutflag & FF_TRANSLUCENT)
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
HWR_AddTransparentWall(wallVerts, Surf, texnum, PF_Translucent, false, lightnum, colormap);
else
HWR_ProjectWall(wallVerts, Surf, PF_Masked, lightnum, colormap);
}
OpenGL: Sky culling Gets rid of stuff being drawn where the sky should be being drawn.
2014-09-08 22:29:05 +00:00
// HWR_DrawSkyWalls
// Draw walls into the depth buffer so that anything behind is culled properly
static void HWR_DrawSkyWall(wallVert3D *wallVerts, FSurfaceInfo *Surf, fixed_t bottom, fixed_t top)
{
HWD.pfnSetTexture(NULL);
// no texture
wallVerts[3].t = wallVerts[2].t = 0;
wallVerts[0].t = wallVerts[1].t = 0;
wallVerts[0].s = wallVerts[3].s = 0;
wallVerts[2].s = wallVerts[1].s = 0;
// set top/bottom coords
wallVerts[2].y = wallVerts[3].y = FIXED_TO_FLOAT(top); // No real way to find the correct height of this
wallVerts[0].y = wallVerts[1].y = FIXED_TO_FLOAT(bottom); // worldlow/bottom because it needs to cover up the lower thok barrier wall
HWR_ProjectWall(wallVerts, Surf, PF_Invisible|PF_Clip|PF_NoTexture, 255, NULL);
// PF_Invisible so it's not drawn into the colour buffer
// PF_NoTexture for no texture
// PF_Occlude is set in HWR_ProjectWall to draw into the depth buffer
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
//
// HWR_StoreWallRange
// A portion or all of a wall segment will be drawn, from startfrac to endfrac,
// where 0 is the start of the segment, 1 the end of the segment
// Anything between means the wall segment has been clipped with solidsegs,
// reducing wall overdraw to a minimum
//
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#ifdef NEWCLIP
static void HWR_ProcessSeg(void) // Sort of like GLWall::Process in GZDoom
#else
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
static void HWR_StoreWallRange(double startfrac, double endfrac)
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
wallVert3D wallVerts[4];
v2d_t vs, ve; // start, end vertices of 2d line (view from above)
fixed_t worldtop, worldbottom;
fixed_t worldhigh = 0, worldlow = 0;
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
#ifdef ESLOPE
fixed_t worldtopslope, worldbottomslope;
fixed_t worldhighslope = 0, worldlowslope = 0;
fixed_t v1x, v1y, v2x, v2y;
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
GLTexture_t *grTex = NULL;
float cliplow = 0.0f, cliphigh = 0.0f;
INT32 gr_midtexture;
fixed_t h, l; // 3D sides and 2s middle textures
Fix FOF slope rendering in ogl should work as well as software if not better now
2016-01-25 05:57:53 +00:00
#ifdef ESLOPE
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
fixed_t hS, lS;
Fix FOF slope rendering in ogl should work as well as software if not better now
2016-01-25 05:57:53 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
FUINT lightnum = 0; // shut up compiler
extracolormap_t *colormap;
FSurfaceInfo Surf;
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#ifndef NEWCLIP
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (startfrac > endfrac)
return;
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
gr_sidedef = gr_curline->sidedef;
gr_linedef = gr_curline->linedef;
Make new pv1/pv2 seg pointers, so AdjustSeg doesn't modify the v1/v2 pointers directly anymore Yes I know they're void * in r_defs.h's seg_t definition, it's quicker than trying to figure out if including hardware/hw_glob.h is a good idea or not
2016-11-07 21:55:56 +00:00
vs.x = ((polyvertex_t *)gr_curline->pv1)->x;
vs.y = ((polyvertex_t *)gr_curline->pv1)->y;
ve.x = ((polyvertex_t *)gr_curline->pv2)->x;
ve.y = ((polyvertex_t *)gr_curline->pv2)->y;
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
#ifdef ESLOPE
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
v1x = FLOAT_TO_FIXED(vs.x);
v1y = FLOAT_TO_FIXED(vs.y);
v2x = FLOAT_TO_FIXED(ve.x);
v2y = FLOAT_TO_FIXED(ve.y);
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
#endif
#ifdef ESLOPE
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
Remove pointless drawtextured variable and redundant fake planes checks (R_FakeFlat would already have made gr_frontsector/backsector something else if they were)
2017-01-12 21:43:37 +00:00
#define SLOPEPARAMS(slope, end1, end2, normalheight) \
if (slope) { \
end1 = P_GetZAt(slope, v1x, v1y); \
end2 = P_GetZAt(slope, v2x, v2y); \
} else \
end1 = end2 = normalheight;
SLOPEPARAMS(gr_frontsector->c_slope, worldtop, worldtopslope, gr_frontsector->ceilingheight)
SLOPEPARAMS(gr_frontsector->f_slope, worldbottom, worldbottomslope, gr_frontsector->floorheight)
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
#else
Remove pointless drawtextured variable and redundant fake planes checks (R_FakeFlat would already have made gr_frontsector/backsector something else if they were)
2017-01-12 21:43:37 +00:00
worldtop = gr_frontsector->ceilingheight;
worldbottom = gr_frontsector->floorheight;
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// remember vertices ordering
// 3--2
// | /|
// |/ |
// 0--1
// make a wall polygon (with 2 triangles), using the floor/ceiling heights,
// and the 2d map coords of start/end vertices
wallVerts[0].x = wallVerts[3].x = vs.x;
wallVerts[0].z = wallVerts[3].z = vs.y;
wallVerts[2].x = wallVerts[1].x = ve.x;
wallVerts[2].z = wallVerts[1].z = ve.y;
wallVerts[0].w = wallVerts[1].w = wallVerts[2].w = wallVerts[3].w = 1.0f;
{
// x offset the texture
fixed_t texturehpeg = gr_sidedef->textureoffset + gr_curline->offset;
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#ifndef NEWCLIP
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// clip texture s start/end coords with solidsegs
if (startfrac > 0.0f && startfrac < 1.0f)
cliplow = (float)(texturehpeg + (gr_curline->flength*FRACUNIT) * startfrac);
else
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
cliplow = (float)texturehpeg;
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#ifndef NEWCLIP
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (endfrac > 0.0f && endfrac < 1.0f)
cliphigh = (float)(texturehpeg + (gr_curline->flength*FRACUNIT) * endfrac);
else
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
cliphigh = (float)(texturehpeg + (gr_curline->flength*FRACUNIT));
}
lightnum = gr_frontsector->lightlevel;
colormap = gr_frontsector->extra_colormap;
if (gr_frontsector)
{
Surf.FlatColor.s.alpha = 255;
}
if (gr_backsector)
{
Created R_GetTextureNum to make sure top/bottom/midtexture texture ids are always valid in rendering code for both software and OpenGL (and also for the Solid Midtexture effect physics code)
2016-12-16 21:38:53 +00:00
INT32 gr_toptexture, gr_bottomtexture;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// two sided line
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
Remove pointless drawtextured variable and redundant fake planes checks (R_FakeFlat would already have made gr_frontsector/backsector something else if they were)
2017-01-12 21:43:37 +00:00
#ifdef ESLOPE
SLOPEPARAMS(gr_backsector->c_slope, worldhigh, worldhighslope, gr_backsector->ceilingheight)
SLOPEPARAMS(gr_backsector->f_slope, worldlow, worldlowslope, gr_backsector->floorheight)
#undef SLOPEPARAMS
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
#else
Remove pointless drawtextured variable and redundant fake planes checks (R_FakeFlat would already have made gr_frontsector/backsector something else if they were)
2017-01-12 21:43:37 +00:00
worldhigh = gr_backsector->ceilingheight;
worldlow = gr_backsector->floorheight;
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// hack to allow height changes in outdoor areas
// This is what gets rid of the upper textures if there should be sky
if (gr_frontsector->ceilingpic == skyflatnum &&
gr_backsector->ceilingpic == skyflatnum)
{
worldtop = worldhigh;
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
#ifdef ESLOPE
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
worldtopslope = worldhighslope;
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
Created R_GetTextureNum to make sure top/bottom/midtexture texture ids are always valid in rendering code for both software and OpenGL (and also for the Solid Midtexture effect physics code)
2016-12-16 21:38:53 +00:00
gr_toptexture = R_GetTextureNum(gr_sidedef->toptexture);
gr_bottomtexture = R_GetTextureNum(gr_sidedef->bottomtexture);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// check TOP TEXTURE
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
if ((
#ifdef ESLOPE
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
worldhighslope < worldtopslope ||
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
#endif
worldhigh < worldtop
Created R_GetTextureNum to make sure top/bottom/midtexture texture ids are always valid in rendering code for both software and OpenGL (and also for the Solid Midtexture effect physics code)
2016-12-16 21:38:53 +00:00
) && gr_toptexture)
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
{
fixed_t texturevpegtop; // top
Created R_GetTextureNum to make sure top/bottom/midtexture texture ids are always valid in rendering code for both software and OpenGL (and also for the Solid Midtexture effect physics code)
2016-12-16 21:38:53 +00:00
grTex = HWR_GetTexture(gr_toptexture);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// PEGGING
if (gr_linedef->flags & ML_DONTPEGTOP)
texturevpegtop = 0;
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
#ifdef ESLOPE
else if (gr_linedef->flags & ML_EFFECT1)
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
texturevpegtop = worldhigh + textureheight[gr_sidedef->toptexture] - worldtop;
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
else
texturevpegtop = gr_backsector->ceilingheight + textureheight[gr_sidedef->toptexture] - gr_frontsector->ceilingheight;
#else
else
texturevpegtop = worldhigh + textureheight[gr_sidedef->toptexture] - worldtop;
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
texturevpegtop += gr_sidedef->rowoffset;
OpenGL wall Y offset Fixed wall textures breaking when Y offset of the wall was too high.
2014-03-28 23:28:00 +00:00
// This is so that it doesn't overflow and screw up the wall, it doesn't need to go higher than the texture's height anyway
Created R_GetTextureNum to make sure top/bottom/midtexture texture ids are always valid in rendering code for both software and OpenGL (and also for the Solid Midtexture effect physics code)
2016-12-16 21:38:53 +00:00
texturevpegtop %= SHORT(textures[gr_toptexture]->height)<<FRACBITS;
OpenGL wall Y offset Fixed wall textures breaking when Y offset of the wall was too high.
2014-03-28 23:28:00 +00:00
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
wallVerts[3].t = wallVerts[2].t = texturevpegtop * grTex->scaleY;
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
wallVerts[0].t = wallVerts[1].t = (texturevpegtop + gr_frontsector->ceilingheight - gr_backsector->ceilingheight) * grTex->scaleY;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
wallVerts[0].s = wallVerts[3].s = cliplow * grTex->scaleX;
wallVerts[2].s = wallVerts[1].s = cliphigh * grTex->scaleX;
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
#ifdef ESLOPE
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
// Adjust t value for sloped walls
if (!(gr_linedef->flags & ML_EFFECT1))
{
// Unskewed
wallVerts[3].t -= (worldtop - gr_frontsector->ceilingheight) * grTex->scaleY;
wallVerts[2].t -= (worldtopslope - gr_frontsector->ceilingheight) * grTex->scaleY;
wallVerts[0].t -= (worldhigh - gr_backsector->ceilingheight) * grTex->scaleY;
wallVerts[1].t -= (worldhighslope - gr_backsector->ceilingheight) * grTex->scaleY;
}
else if (gr_linedef->flags & ML_DONTPEGTOP)
{
// Skewed by top
wallVerts[0].t = (texturevpegtop + worldtop - worldhigh) * grTex->scaleY;
wallVerts[1].t = (texturevpegtop + worldtopslope - worldhighslope) * grTex->scaleY;
}
else
{
// Skewed by bottom
OpenGL: Fix upper texture Effect 1 only skewing
2016-06-01 18:22:54 +00:00
wallVerts[0].t = wallVerts[1].t = (texturevpegtop + worldtop - worldhigh) * grTex->scaleY;
wallVerts[3].t = wallVerts[0].t - (worldtop - worldhigh) * grTex->scaleY;
wallVerts[2].t = wallVerts[1].t - (worldtopslope - worldhighslope) * grTex->scaleY;
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
}
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
// set top/bottom coords
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
#ifdef ESLOPE
wallVerts[3].y = FIXED_TO_FLOAT(worldtop);
wallVerts[0].y = FIXED_TO_FLOAT(worldhigh);
wallVerts[2].y = FIXED_TO_FLOAT(worldtopslope);
wallVerts[1].y = FIXED_TO_FLOAT(worldhighslope);
#else
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
wallVerts[2].y = wallVerts[3].y = FIXED_TO_FLOAT(worldtop);
wallVerts[0].y = wallVerts[1].y = FIXED_TO_FLOAT(worldhigh);
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (gr_frontsector->numlights)
Some fixes and updates for HWR_SplitWall Solid walls *can* be cut Fix issues with water and fog FOFs not cutting each other out correctly Fix Fog colourmap and lighting setting that is done here. Remove HWR_SplitFog There is currently a bug with FF_DOUBLESHADOW (that also exists in software) but has a larger impact here. When 2 FF_DOUBLESHADOW lights are directly stacked on each other the bottom one has its height set incorrectly. This causes all the Fog in the timed gravity flipping section of ERZ2 to be drawn and it looks really bad.
2018-03-20 14:20:02 +00:00
HWR_SplitWall(gr_frontsector, wallVerts, gr_toptexture, &Surf, FF_CUTLEVEL, NULL);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
else if (grTex->mipmap.flags & TF_TRANSPARENT)
Created R_GetTextureNum to make sure top/bottom/midtexture texture ids are always valid in rendering code for both software and OpenGL (and also for the Solid Midtexture effect physics code)
2016-12-16 21:38:53 +00:00
HWR_AddTransparentWall(wallVerts, &Surf, gr_toptexture, PF_Environment, false, lightnum, colormap);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
else
HWR_ProjectWall(wallVerts, &Surf, PF_Masked, lightnum, colormap);
}
// check BOTTOM TEXTURE
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
if ((
#ifdef ESLOPE
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
worldlowslope > worldbottomslope ||
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
#endif
Created R_GetTextureNum to make sure top/bottom/midtexture texture ids are always valid in rendering code for both software and OpenGL (and also for the Solid Midtexture effect physics code)
2016-12-16 21:38:53 +00:00
worldlow > worldbottom) && gr_bottomtexture) //only if VISIBLE!!!
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
{
fixed_t texturevpegbottom = 0; // bottom
Created R_GetTextureNum to make sure top/bottom/midtexture texture ids are always valid in rendering code for both software and OpenGL (and also for the Solid Midtexture effect physics code)
2016-12-16 21:38:53 +00:00
grTex = HWR_GetTexture(gr_bottomtexture);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// PEGGING
GL slope walls and fixed plane culling
2016-01-24 08:41:30 +00:00
#ifdef ESLOPE
if (!(gr_linedef->flags & ML_DONTPEGBOTTOM))
texturevpegbottom = 0;
else if (gr_linedef->flags & ML_EFFECT1)
OpenGL: Fix lower unpegged texture offset, fix lower unpegged + effect 1 so the texture actually skews
2016-06-01 17:51:38 +00:00
texturevpegbottom = worldbottom - worldlow;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
else
OpenGL: Fix lower unpegged texture offset, fix lower unpegged + effect 1 so the texture actually skews
2016-06-01 17:51:38 +00:00
texturevpegbottom = gr_frontsector->floorheight - gr_backsector->floorheight;
GL slope walls and fixed plane culling
2016-01-24 08:41:30 +00:00
#else
if (gr_linedef->flags & ML_DONTPEGBOTTOM)
OpenGL: Fix lower unpegged texture offset, fix lower unpegged + effect 1 so the texture actually skews
2016-06-01 17:51:38 +00:00
texturevpegbottom = worldbottom - worldlow;
GL slope walls and fixed plane culling
2016-01-24 08:41:30 +00:00
else
texturevpegbottom = 0;
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
texturevpegbottom += gr_sidedef->rowoffset;
OpenGL wall Y offset Fixed wall textures breaking when Y offset of the wall was too high.
2014-03-28 23:28:00 +00:00
// This is so that it doesn't overflow and screw up the wall, it doesn't need to go higher than the texture's height anyway
Created R_GetTextureNum to make sure top/bottom/midtexture texture ids are always valid in rendering code for both software and OpenGL (and also for the Solid Midtexture effect physics code)
2016-12-16 21:38:53 +00:00
texturevpegbottom %= SHORT(textures[gr_bottomtexture]->height)<<FRACBITS;
OpenGL wall Y offset Fixed wall textures breaking when Y offset of the wall was too high.
2014-03-28 23:28:00 +00:00
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
wallVerts[3].t = wallVerts[2].t = texturevpegbottom * grTex->scaleY;
GL slope walls and fixed plane culling
2016-01-24 08:41:30 +00:00
wallVerts[0].t = wallVerts[1].t = (texturevpegbottom + gr_backsector->floorheight - gr_frontsector->floorheight) * grTex->scaleY;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
wallVerts[0].s = wallVerts[3].s = cliplow * grTex->scaleX;
wallVerts[2].s = wallVerts[1].s = cliphigh * grTex->scaleX;
GL slope walls and fixed plane culling
2016-01-24 08:41:30 +00:00
#ifdef ESLOPE
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
// Adjust t value for sloped walls
if (!(gr_linedef->flags & ML_EFFECT1))
{
// Unskewed
wallVerts[0].t -= (worldbottom - gr_frontsector->floorheight) * grTex->scaleY;
wallVerts[1].t -= (worldbottomslope - gr_frontsector->floorheight) * grTex->scaleY;
wallVerts[3].t -= (worldlow - gr_backsector->floorheight) * grTex->scaleY;
wallVerts[2].t -= (worldlowslope - gr_backsector->floorheight) * grTex->scaleY;
}
else if (gr_linedef->flags & ML_DONTPEGBOTTOM)
{
// Skewed by bottom
OpenGL: Fix lower unpegged texture offset, fix lower unpegged + effect 1 so the texture actually skews
2016-06-01 17:51:38 +00:00
wallVerts[0].t = wallVerts[1].t = (texturevpegbottom + worldlow - worldbottom) * grTex->scaleY;
//wallVerts[3].t = wallVerts[0].t - (worldlow - worldbottom) * grTex->scaleY; // no need, [3] is already this
wallVerts[2].t = wallVerts[1].t - (worldlowslope - worldbottomslope) * grTex->scaleY;
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
}
else
{
// Skewed by top
wallVerts[0].t = (texturevpegbottom + worldlow - worldbottom) * grTex->scaleY;
wallVerts[1].t = (texturevpegbottom + worldlowslope - worldbottomslope) * grTex->scaleY;
}
GL slope walls and fixed plane culling
2016-01-24 08:41:30 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
// set top/bottom coords
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
#ifdef ESLOPE
wallVerts[3].y = FIXED_TO_FLOAT(worldlow);
wallVerts[0].y = FIXED_TO_FLOAT(worldbottom);
wallVerts[2].y = FIXED_TO_FLOAT(worldlowslope);
wallVerts[1].y = FIXED_TO_FLOAT(worldbottomslope);
#else
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
wallVerts[2].y = wallVerts[3].y = FIXED_TO_FLOAT(worldlow);
wallVerts[0].y = wallVerts[1].y = FIXED_TO_FLOAT(worldbottom);
Begin work on OGL slope support unfinished
2016-01-17 04:10:38 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (gr_frontsector->numlights)
Some fixes and updates for HWR_SplitWall Solid walls *can* be cut Fix issues with water and fog FOFs not cutting each other out correctly Fix Fog colourmap and lighting setting that is done here. Remove HWR_SplitFog There is currently a bug with FF_DOUBLESHADOW (that also exists in software) but has a larger impact here. When 2 FF_DOUBLESHADOW lights are directly stacked on each other the bottom one has its height set incorrectly. This causes all the Fog in the timed gravity flipping section of ERZ2 to be drawn and it looks really bad.
2018-03-20 14:20:02 +00:00
HWR_SplitWall(gr_frontsector, wallVerts, gr_bottomtexture, &Surf, FF_CUTLEVEL, NULL);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
else if (grTex->mipmap.flags & TF_TRANSPARENT)
Created R_GetTextureNum to make sure top/bottom/midtexture texture ids are always valid in rendering code for both software and OpenGL (and also for the Solid Midtexture effect physics code)
2016-12-16 21:38:53 +00:00
HWR_AddTransparentWall(wallVerts, &Surf, gr_bottomtexture, PF_Environment, false, lightnum, colormap);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
else
HWR_ProjectWall(wallVerts, &Surf, PF_Masked, lightnum, colormap);
}
Created R_GetTextureNum to make sure top/bottom/midtexture texture ids are always valid in rendering code for both software and OpenGL (and also for the Solid Midtexture effect physics code)
2016-12-16 21:38:53 +00:00
gr_midtexture = R_GetTextureNum(gr_sidedef->midtexture);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (gr_midtexture)
{
FBITFIELD blendmode;
sector_t *front, *back;
fixed_t popentop, popenbottom, polytop, polybottom, lowcut, highcut;
fixed_t texturevpeg = 0;
INT32 repeats;
if (gr_linedef->frontsector->heightsec != -1)
front = &sectors[gr_linedef->frontsector->heightsec];
else
front = gr_linedef->frontsector;
if (gr_linedef->backsector->heightsec != -1)
back = &sectors[gr_linedef->backsector->heightsec];
else
back = gr_linedef->backsector;
if (gr_sidedef->repeatcnt)
repeats = 1 + gr_sidedef->repeatcnt;
else if (gr_linedef->flags & ML_EFFECT5)
{
fixed_t high, low;
if (front->ceilingheight > back->ceilingheight)
high = back->ceilingheight;
else
high = front->ceilingheight;
if (front->floorheight > back->floorheight)
low = front->floorheight;
else
low = back->floorheight;
repeats = (high - low)/textureheight[gr_sidedef->midtexture];
if ((high-low)%textureheight[gr_sidedef->midtexture])
repeats++; // tile an extra time to fill the gap -- Monster Iestyn
}
else
repeats = 1;
// SoM: a little note: This code re-arranging will
// fix the bug in Nimrod map02. popentop and popenbottom
// record the limits the texture can be displayed in.
// polytop and polybottom, are the ideal (i.e. unclipped)
// heights of the polygon, and h & l, are the final (clipped)
// poly coords.
#ifdef POLYOBJECTS
// NOTE: With polyobjects, whenever you need to check the properties of the polyobject sector it belongs to,
// you must use the linedef's backsector to be correct
// From CB
if (gr_curline->polyseg)
{
popentop = back->ceilingheight;
popenbottom = back->floorheight;
}
else
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#endif
{
#ifdef ESLOPE
popentop = min(worldtop, worldhigh);
popenbottom = max(worldbottom, worldlow);
#else
popentop = min(front->ceilingheight, back->ceilingheight);
popenbottom = max(front->floorheight, back->floorheight);
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#ifdef ESLOPE
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
if (gr_linedef->flags & ML_EFFECT2)
{
if (!!(gr_linedef->flags & ML_DONTPEGBOTTOM) ^ !!(gr_linedef->flags & ML_EFFECT3))
{
polybottom = max(front->floorheight, back->floorheight) + gr_sidedef->rowoffset;
polytop = polybottom + textureheight[gr_midtexture]*repeats;
}
else
{
polytop = min(front->ceilingheight, back->ceilingheight) + gr_sidedef->rowoffset;
polybottom = polytop - textureheight[gr_midtexture]*repeats;
}
}
else if (!!(gr_linedef->flags & ML_DONTPEGBOTTOM) ^ !!(gr_linedef->flags & ML_EFFECT3))
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#else
if (gr_linedef->flags & ML_DONTPEGBOTTOM)
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
polybottom = popenbottom + gr_sidedef->rowoffset;
polytop = polybottom + textureheight[gr_midtexture]*repeats;
}
else
{
polytop = popentop + gr_sidedef->rowoffset;
polybottom = polytop - textureheight[gr_midtexture]*repeats;
}
// CB
#ifdef POLYOBJECTS
// NOTE: With polyobjects, whenever you need to check the properties of the polyobject sector it belongs to,
// you must use the linedef's backsector to be correct
if (gr_curline->polyseg)
{
lowcut = polybottom;
highcut = polytop;
}
#endif
else
{
// The cut-off values of a linedef can always be constant, since every line has an absoulute front and or back sector
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
lowcut = popenbottom;
highcut = popentop;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
h = min(highcut, polytop);
l = max(polybottom, lowcut);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
// PEGGING
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#ifdef ESLOPE
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
if (!!(gr_linedef->flags & ML_DONTPEGBOTTOM) ^ !!(gr_linedef->flags & ML_EFFECT3))
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#else
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (gr_linedef->flags & ML_DONTPEGBOTTOM)
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
texturevpeg = textureheight[gr_sidedef->midtexture]*repeats - h + polybottom;
else
texturevpeg = polytop - h;
grTex = HWR_GetTexture(gr_midtexture);
wallVerts[3].t = wallVerts[2].t = texturevpeg * grTex->scaleY;
wallVerts[0].t = wallVerts[1].t = (h - l + texturevpeg) * grTex->scaleY;
wallVerts[0].s = wallVerts[3].s = cliplow * grTex->scaleX;
wallVerts[2].s = wallVerts[1].s = cliphigh * grTex->scaleX;
}
// set top/bottom coords
// Take the texture peg into account, rather than changing the offsets past
// where the polygon might not be.
wallVerts[2].y = wallVerts[3].y = FIXED_TO_FLOAT(h);
wallVerts[0].y = wallVerts[1].y = FIXED_TO_FLOAT(l);
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#ifdef ESLOPE
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
// Correct to account for slopes
{
fixed_t midtextureslant;
if (gr_linedef->flags & ML_EFFECT2)
midtextureslant = 0;
else if (!!(gr_linedef->flags & ML_DONTPEGBOTTOM) ^ !!(gr_linedef->flags & ML_EFFECT3))
midtextureslant = worldlow < worldbottom
? worldbottomslope-worldbottom
: worldlowslope-worldlow;
else
midtextureslant = worldtop < worldhigh
? worldtopslope-worldtop
: worldhighslope-worldhigh;
polytop += midtextureslant;
polybottom += midtextureslant;
highcut += worldtop < worldhigh
? worldtopslope-worldtop
: worldhighslope-worldhigh;
lowcut += worldlow < worldbottom
? worldbottomslope-worldbottom
: worldlowslope-worldlow;
// Texture stuff
h = min(highcut, polytop);
l = max(polybottom, lowcut);
{
// PEGGING
if (!!(gr_linedef->flags & ML_DONTPEGBOTTOM) ^ !!(gr_linedef->flags & ML_EFFECT3))
texturevpeg = textureheight[gr_sidedef->midtexture]*repeats - h + polybottom;
else
texturevpeg = polytop - h;
wallVerts[2].t = texturevpeg * grTex->scaleY;
wallVerts[1].t = (h - l + texturevpeg) * grTex->scaleY;
}
wallVerts[2].y = FIXED_TO_FLOAT(h);
wallVerts[1].y = FIXED_TO_FLOAT(l);
}
Midtextures, lights, and culling fixes for ogl slopes There's a weird issue with lights that's hard to diagnose but otherwise this is ready to go I think
2016-01-27 06:00:15 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// set alpha for transparent walls (new boom and legacy linedef types)
// ooops ! this do not work at all because render order we should render it in backtofront order
switch (gr_linedef->special)
{
case 900:
blendmode = HWR_TranstableToAlpha(tr_trans10, &Surf);
break;
case 901:
blendmode = HWR_TranstableToAlpha(tr_trans20, &Surf);
break;
case 902:
blendmode = HWR_TranstableToAlpha(tr_trans30, &Surf);
break;
case 903:
blendmode = HWR_TranstableToAlpha(tr_trans40, &Surf);
break;
case 904:
blendmode = HWR_TranstableToAlpha(tr_trans50, &Surf);
break;
case 905:
blendmode = HWR_TranstableToAlpha(tr_trans60, &Surf);
break;
case 906:
blendmode = HWR_TranstableToAlpha(tr_trans70, &Surf);
break;
case 907:
blendmode = HWR_TranstableToAlpha(tr_trans80, &Surf);
break;
case 908:
blendmode = HWR_TranstableToAlpha(tr_trans90, &Surf);
break;
// Translucent
case 102:
case 121:
case 123:
case 124:
case 125:
case 141:
case 142:
case 144:
case 145:
case 174:
case 175:
case 192:
case 195:
case 221:
case 253:
case 256:
blendmode = PF_Translucent;
break;
default:
blendmode = PF_Masked;
break;
}
Polyobject top/bottom texture + translucency polyobjects do have top/bottom textures, silly. And translucency.
2014-08-30 13:10:55 +00:00
#ifdef POLYOBJECTS
if (gr_curline->polyseg && gr_curline->polyseg->translucency > 0)
{
if (gr_curline->polyseg->translucency >= NUMTRANSMAPS) // wall not drawn
Polyobject translucency quick fix. I realise what I did before may cause FOF's in the same sector to not be drawn, if they were before.
2014-08-30 13:20:07 +00:00
{
Surf.FlatColor.s.alpha = 0x00; // This shouldn't draw anything regardless of blendmode
blendmode = PF_Masked;
}
else added less redundancy.
2014-08-30 13:36:18 +00:00
else
blendmode = HWR_TranstableToAlpha(gr_curline->polyseg->translucency, &Surf);
Polyobject top/bottom texture + translucency polyobjects do have top/bottom textures, silly. And translucency.
2014-08-30 13:10:55 +00:00
}
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (gr_frontsector->numlights)
Opaque translucent midtextures Fixed translucent midtextures being opaque when there were multiple lights in a sector.
2014-03-20 20:34:56 +00:00
{
if (!(blendmode & PF_Masked))
Some fixes and updates for HWR_SplitWall Solid walls *can* be cut Fix issues with water and fog FOFs not cutting each other out correctly Fix Fog colourmap and lighting setting that is done here. Remove HWR_SplitFog There is currently a bug with FF_DOUBLESHADOW (that also exists in software) but has a larger impact here. When 2 FF_DOUBLESHADOW lights are directly stacked on each other the bottom one has its height set incorrectly. This causes all the Fog in the timed gravity flipping section of ERZ2 to be drawn and it looks really bad.
2018-03-20 14:20:02 +00:00
HWR_SplitWall(gr_frontsector, wallVerts, gr_midtexture, &Surf, FF_TRANSLUCENT, NULL);
Opaque translucent midtextures Fixed translucent midtextures being opaque when there were multiple lights in a sector.
2014-03-20 20:34:56 +00:00
else
Various Transparent Texture Fixes Draw Textures and Flats that have holes in them like a solid polygon so they use the depth buffer and don't need to be sorted Disable all linear filtering on textures and flats that have holes in them, the linear filtering introduces translucency into the textures where the edges are. Leaving them with either a black border, or causing pixels behind the slightly translucent areas to not be drawn. Doesn't apply to sprites and the HUD as they are always already sorted properly. Make the Alpha Testing more strict on non-translucent blend modes. This makes it so any transparency below 0.5 is discarded instead. Would make anything that is blended and has holes in it look slightly better, only the HUD and MD2s where the texture has holes are effected currently. Set TF_TRANSPARENT on flat texture flags when there are holes in the texture. Minor fix to make sure MD2s always set the right blend mode
2017-09-08 23:41:11 +00:00
{
Some fixes and updates for HWR_SplitWall Solid walls *can* be cut Fix issues with water and fog FOFs not cutting each other out correctly Fix Fog colourmap and lighting setting that is done here. Remove HWR_SplitFog There is currently a bug with FF_DOUBLESHADOW (that also exists in software) but has a larger impact here. When 2 FF_DOUBLESHADOW lights are directly stacked on each other the bottom one has its height set incorrectly. This causes all the Fog in the timed gravity flipping section of ERZ2 to be drawn and it looks really bad.
2018-03-20 14:20:02 +00:00
HWR_SplitWall(gr_frontsector, wallVerts, gr_midtexture, &Surf, FF_CUTLEVEL, NULL);
Various Transparent Texture Fixes Draw Textures and Flats that have holes in them like a solid polygon so they use the depth buffer and don't need to be sorted Disable all linear filtering on textures and flats that have holes in them, the linear filtering introduces translucency into the textures where the edges are. Leaving them with either a black border, or causing pixels behind the slightly translucent areas to not be drawn. Doesn't apply to sprites and the HUD as they are always already sorted properly. Make the Alpha Testing more strict on non-translucent blend modes. This makes it so any transparency below 0.5 is discarded instead. Would make anything that is blended and has holes in it look slightly better, only the HUD and MD2s where the texture has holes are effected currently. Set TF_TRANSPARENT on flat texture flags when there are holes in the texture. Minor fix to make sure MD2s always set the right blend mode
2017-09-08 23:41:11 +00:00
}
Opaque translucent midtextures Fixed translucent midtextures being opaque when there were multiple lights in a sector.
2014-03-20 20:34:56 +00:00
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
else if (!(blendmode & PF_Masked))
HWR_AddTransparentWall(wallVerts, &Surf, gr_midtexture, blendmode, false, lightnum, colormap);
else
HWR_ProjectWall(wallVerts, &Surf, blendmode, lightnum, colormap);
// If there is a colormap change, remove it.
/* if (!(Surf.FlatColor.s.red + Surf.FlatColor.s.green + Surf.FlatColor.s.blue == Surf.FlatColor.s.red/3)
{
Surf.FlatColor.s.red = Surf.FlatColor.s.green = Surf.FlatColor.s.blue = 0xff;
Surf.FlatColor.rgba = 0xffffffff;
}*/
}
OpenGL: Sky culling Gets rid of stuff being drawn where the sky should be being drawn.
2014-09-08 22:29:05 +00:00
// Isn't this just the most lovely mess
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
if (!gr_curline->polyseg) // Don't do it for polyobjects
OpenGL: Sky culling Gets rid of stuff being drawn where the sky should be being drawn.
2014-09-08 22:29:05 +00:00
{
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
if (gr_frontsector->ceilingpic == skyflatnum || gr_backsector->ceilingpic == skyflatnum)
{
fixed_t depthwallheight;
Quick sky fix Avoid Z-Fighting on the depth wall with the regular walls to stop the possibility of regular walls being made partially invisible.
2014-09-09 00:59:19 +00:00
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
if (!gr_sidedef->toptexture || (gr_frontsector->ceilingpic == skyflatnum && gr_backsector->ceilingpic == skyflatnum)) // when both sectors are sky, the top texture isn't drawn
depthwallheight = gr_frontsector->ceilingheight < gr_backsector->ceilingheight ? gr_frontsector->ceilingheight : gr_backsector->ceilingheight;
else
depthwallheight = gr_frontsector->ceilingheight > gr_backsector->ceilingheight ? gr_frontsector->ceilingheight : gr_backsector->ceilingheight;
Quick sky fix Avoid Z-Fighting on the depth wall with the regular walls to stop the possibility of regular walls being made partially invisible.
2014-09-09 00:59:19 +00:00
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
if (gr_frontsector->ceilingheight-gr_frontsector->floorheight <= 0) // current sector is a thok barrier
Quick sky fix Avoid Z-Fighting on the depth wall with the regular walls to stop the possibility of regular walls being made partially invisible.
2014-09-09 00:59:19 +00:00
{
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
if (gr_backsector->ceilingheight-gr_backsector->floorheight <= 0) // behind sector is also a thok barrier
{
if (!gr_sidedef->bottomtexture) // Only extend further down if there's no texture
HWR_DrawSkyWall(wallVerts, &Surf, worldbottom < worldlow ? worldbottom : worldlow, INT32_MAX);
else
HWR_DrawSkyWall(wallVerts, &Surf, worldbottom > worldlow ? worldbottom : worldlow, INT32_MAX);
}
// behind sector is not a thok barrier
else if (gr_backsector->ceilingheight <= gr_frontsector->ceilingheight) // behind sector ceiling is lower or equal to current sector
HWR_DrawSkyWall(wallVerts, &Surf, depthwallheight, INT32_MAX);
// gr_front/backsector heights need to be used here because of the worldtop being set to worldhigh earlier on
}
else if (gr_backsector->ceilingheight-gr_backsector->floorheight <= 0) // behind sector is a thok barrier, current sector is not
{
if (gr_backsector->ceilingheight >= gr_frontsector->ceilingheight // thok barrier ceiling height is equal to or greater than current sector ceiling height
|| gr_backsector->floorheight <= gr_frontsector->floorheight // thok barrier ceiling height is equal to or less than current sector floor height
|| gr_backsector->ceilingpic != skyflatnum) // thok barrier is not a sky
HWR_DrawSkyWall(wallVerts, &Surf, depthwallheight, INT32_MAX);
}
else // neither sectors are thok barriers
{
if ((gr_backsector->ceilingheight < gr_frontsector->ceilingheight && !gr_sidedef->toptexture) // no top texture and sector behind is lower
|| gr_backsector->ceilingpic != skyflatnum) // behind sector is not a sky
HWR_DrawSkyWall(wallVerts, &Surf, depthwallheight, INT32_MAX);
Quick sky fix Avoid Z-Fighting on the depth wall with the regular walls to stop the possibility of regular walls being made partially invisible.
2014-09-09 00:59:19 +00:00
}
OpenGL: Sky culling Gets rid of stuff being drawn where the sky should be being drawn.
2014-09-08 22:29:05 +00:00
}
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
// And now for sky floors!
if (gr_frontsector->floorpic == skyflatnum || gr_backsector->floorpic == skyflatnum)
OpenGL: Sky culling Gets rid of stuff being drawn where the sky should be being drawn.
2014-09-08 22:29:05 +00:00
{
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
fixed_t depthwallheight;
Quick sky fix Avoid Z-Fighting on the depth wall with the regular walls to stop the possibility of regular walls being made partially invisible.
2014-09-09 00:59:19 +00:00
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
if (!gr_sidedef->bottomtexture)
depthwallheight = worldbottom > worldlow ? worldbottom : worldlow;
else
depthwallheight = worldbottom < worldlow ? worldbottom : worldlow;
Quick sky fix Avoid Z-Fighting on the depth wall with the regular walls to stop the possibility of regular walls being made partially invisible.
2014-09-09 00:59:19 +00:00
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
if (gr_frontsector->ceilingheight-gr_frontsector->floorheight <= 0) // current sector is a thok barrier
Quick sky fix Avoid Z-Fighting on the depth wall with the regular walls to stop the possibility of regular walls being made partially invisible.
2014-09-09 00:59:19 +00:00
{
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
if (gr_backsector->ceilingheight-gr_backsector->floorheight <= 0) // behind sector is also a thok barrier
{
if (!gr_sidedef->toptexture) // Only extend up if there's no texture
HWR_DrawSkyWall(wallVerts, &Surf, INT32_MIN, worldtop > worldhigh ? worldtop : worldhigh);
else
HWR_DrawSkyWall(wallVerts, &Surf, INT32_MIN, worldtop < worldhigh ? worldtop : worldhigh);
}
// behind sector is not a thok barrier
else if (gr_backsector->floorheight >= gr_frontsector->floorheight) // behind sector floor is greater or equal to current sector
HWR_DrawSkyWall(wallVerts, &Surf, INT32_MIN, depthwallheight);
}
else if (gr_backsector->ceilingheight-gr_backsector->floorheight <= 0) // behind sector is a thok barrier, current sector is not
{
if (gr_backsector->floorheight <= gr_frontsector->floorheight // thok barrier floor height is equal to or less than current sector floor height
|| gr_backsector->ceilingheight >= gr_frontsector->ceilingheight // thok barrier floor height is equal to or greater than current sector ceiling height
|| gr_backsector->floorpic != skyflatnum) // thok barrier is not a sky
HWR_DrawSkyWall(wallVerts, &Surf, INT32_MIN, depthwallheight);
}
else // neither sectors are thok barriers
{
if ((gr_backsector->floorheight > gr_frontsector->floorheight && !gr_sidedef->bottomtexture) // no bottom texture and sector behind is higher
|| gr_backsector->floorpic != skyflatnum) // behind sector is not a sky
HWR_DrawSkyWall(wallVerts, &Surf, INT32_MIN, depthwallheight);
Quick sky fix Avoid Z-Fighting on the depth wall with the regular walls to stop the possibility of regular walls being made partially invisible.
2014-09-09 00:59:19 +00:00
}
OpenGL: Sky culling Gets rid of stuff being drawn where the sky should be being drawn.
2014-09-08 22:29:05 +00:00
}
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
else
{
// Single sided line... Deal only with the middletexture (if one exists)
Created R_GetTextureNum to make sure top/bottom/midtexture texture ids are always valid in rendering code for both software and OpenGL (and also for the Solid Midtexture effect physics code)
2016-12-16 21:38:53 +00:00
gr_midtexture = R_GetTextureNum(gr_sidedef->midtexture);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (gr_midtexture)
{
{
fixed_t texturevpeg;
// PEGGING
GL slope walls and fixed plane culling
2016-01-24 08:41:30 +00:00
#ifdef ESLOPE
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
if ((gr_linedef->flags & (ML_DONTPEGBOTTOM|ML_EFFECT2)) == (ML_DONTPEGBOTTOM|ML_EFFECT2))
GL slope walls and fixed plane culling
2016-01-24 08:41:30 +00:00
texturevpeg = gr_frontsector->floorheight + textureheight[gr_sidedef->midtexture] - gr_frontsector->ceilingheight + gr_sidedef->rowoffset;
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
else
GL slope walls and fixed plane culling
2016-01-24 08:41:30 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (gr_linedef->flags & ML_DONTPEGBOTTOM)
texturevpeg = worldbottom + textureheight[gr_sidedef->midtexture] - worldtop + gr_sidedef->rowoffset;
else
// top of texture at top
texturevpeg = gr_sidedef->rowoffset;
grTex = HWR_GetTexture(gr_midtexture);
wallVerts[3].t = wallVerts[2].t = texturevpeg * grTex->scaleY;
GL slope walls and fixed plane culling
2016-01-24 08:41:30 +00:00
wallVerts[0].t = wallVerts[1].t = (texturevpeg + gr_frontsector->ceilingheight - gr_frontsector->floorheight) * grTex->scaleY;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
wallVerts[0].s = wallVerts[3].s = cliplow * grTex->scaleX;
wallVerts[2].s = wallVerts[1].s = cliphigh * grTex->scaleX;
GL slope walls and fixed plane culling
2016-01-24 08:41:30 +00:00
#ifdef ESLOPE
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
// Texture correction for slopes
if (gr_linedef->flags & ML_EFFECT2) {
wallVerts[3].t += (gr_frontsector->ceilingheight - worldtop) * grTex->scaleY;
wallVerts[2].t += (gr_frontsector->ceilingheight - worldtopslope) * grTex->scaleY;
wallVerts[0].t += (gr_frontsector->floorheight - worldbottom) * grTex->scaleY;
wallVerts[1].t += (gr_frontsector->floorheight - worldbottomslope) * grTex->scaleY;
} else if (gr_linedef->flags & ML_DONTPEGBOTTOM) {
wallVerts[3].t = wallVerts[0].t + (worldbottom-worldtop) * grTex->scaleY;
wallVerts[2].t = wallVerts[1].t + (worldbottomslope-worldtopslope) * grTex->scaleY;
} else {
wallVerts[0].t = wallVerts[3].t - (worldbottom-worldtop) * grTex->scaleY;
wallVerts[1].t = wallVerts[2].t - (worldbottomslope-worldtopslope) * grTex->scaleY;
}
GL slope walls and fixed plane culling
2016-01-24 08:41:30 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
GL slope walls and fixed plane culling
2016-01-24 08:41:30 +00:00
#ifdef ESLOPE
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
//Set textures properly on single sided walls that are sloped
GL slope walls and fixed plane culling
2016-01-24 08:41:30 +00:00
wallVerts[3].y = FIXED_TO_FLOAT(worldtop);
wallVerts[0].y = FIXED_TO_FLOAT(worldbottom);
wallVerts[2].y = FIXED_TO_FLOAT(worldtopslope);
wallVerts[1].y = FIXED_TO_FLOAT(worldbottomslope);
#else
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// set top/bottom coords
wallVerts[2].y = wallVerts[3].y = FIXED_TO_FLOAT(worldtop);
wallVerts[0].y = wallVerts[1].y = FIXED_TO_FLOAT(worldbottom);
GL slope walls and fixed plane culling
2016-01-24 08:41:30 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// I don't think that solid walls can use translucent linedef types...
if (gr_frontsector->numlights)
Some fixes and updates for HWR_SplitWall Solid walls *can* be cut Fix issues with water and fog FOFs not cutting each other out correctly Fix Fog colourmap and lighting setting that is done here. Remove HWR_SplitFog There is currently a bug with FF_DOUBLESHADOW (that also exists in software) but has a larger impact here. When 2 FF_DOUBLESHADOW lights are directly stacked on each other the bottom one has its height set incorrectly. This causes all the Fog in the timed gravity flipping section of ERZ2 to be drawn and it looks really bad.
2018-03-20 14:20:02 +00:00
HWR_SplitWall(gr_frontsector, wallVerts, gr_midtexture, &Surf, FF_CUTLEVEL, NULL);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
else
{
if (grTex->mipmap.flags & TF_TRANSPARENT)
HWR_AddTransparentWall(wallVerts, &Surf, gr_midtexture, PF_Environment, false, lightnum, colormap);
else
HWR_ProjectWall(wallVerts, &Surf, PF_Masked, lightnum, colormap);
}
}
OpenGL: Sky culling Gets rid of stuff being drawn where the sky should be being drawn.
2014-09-08 22:29:05 +00:00
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
if (!gr_curline->polyseg)
{
if (gr_frontsector->ceilingpic == skyflatnum) // It's a single-sided line with sky for its sector
HWR_DrawSkyWall(wallVerts, &Surf, worldtop, INT32_MAX);
if (gr_frontsector->floorpic == skyflatnum)
HWR_DrawSkyWall(wallVerts, &Surf, INT32_MIN, worldbottom);
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
//Hurdler: 3d-floors test
#ifdef R_FAKEFLOORS
if (gr_frontsector && gr_backsector && gr_frontsector->tag != gr_backsector->tag && (gr_backsector->ffloors || gr_frontsector->ffloors))
{
ffloor_t * rover;
fixed_t highcut = 0, lowcut = 0;
OpenGL Multi-Property FOF's Multi-Property FOF's have their multiple side properties display in OpenGL.
2014-04-17 12:02:34 +00:00
INT32 texnum;
line_t * newline = NULL; // Multi-Property FOF
GL slope walls and fixed plane culling
2016-01-24 08:41:30 +00:00
///TODO add slope support (fixing cutoffs, proper wall clipping) - maybe just disable highcut/lowcut if either sector or FOF has a slope
/// to allow fun plane intersecting in OGL? But then people would abuse that and make software look bad. :C
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
highcut = gr_frontsector->ceilingheight < gr_backsector->ceilingheight ? gr_frontsector->ceilingheight : gr_backsector->ceilingheight;
lowcut = gr_frontsector->floorheight > gr_backsector->floorheight ? gr_frontsector->floorheight : gr_backsector->floorheight;
if (gr_backsector->ffloors)
{
for (rover = gr_backsector->ffloors; rover; rover = rover->next)
{
if (!(rover->flags & FF_EXISTS) || !(rover->flags & FF_RENDERSIDES) || (rover->flags & FF_INVERTSIDES))
continue;
if (*rover->topheight < lowcut || *rover->bottomheight > highcut)
continue;
Created R_GetTextureNum to make sure top/bottom/midtexture texture ids are always valid in rendering code for both software and OpenGL (and also for the Solid Midtexture effect physics code)
2016-12-16 21:38:53 +00:00
texnum = R_GetTextureNum(sides[rover->master->sidenum[0]].midtexture);
OpenGL Multi-Property FOF's Multi-Property FOF's have their multiple side properties display in OpenGL.
2014-04-17 12:02:34 +00:00
if (rover->master->flags & ML_TFERLINE)
{
size_t linenum = gr_curline->linedef-gr_backsector->lines[0];
newline = rover->master->frontsector->lines[0] + linenum;
Created R_GetTextureNum to make sure top/bottom/midtexture texture ids are always valid in rendering code for both software and OpenGL (and also for the Solid Midtexture effect physics code)
2016-12-16 21:38:53 +00:00
texnum = R_GetTextureNum(sides[newline->sidenum[0]].midtexture);
OpenGL Multi-Property FOF's Multi-Property FOF's have their multiple side properties display in OpenGL.
2014-04-17 12:02:34 +00:00
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
Fix FOF slope rendering in ogl should work as well as software if not better now
2016-01-25 05:57:53 +00:00
#ifdef ESLOPE
h = *rover->t_slope ? P_GetZAt(*rover->t_slope, v1x, v1y) : *rover->topheight;
hS = *rover->t_slope ? P_GetZAt(*rover->t_slope, v2x, v2y) : *rover->topheight;
l = *rover->b_slope ? P_GetZAt(*rover->b_slope, v1x, v1y) : *rover->bottomheight;
lS = *rover->b_slope ? P_GetZAt(*rover->b_slope, v2x, v2y) : *rover->bottomheight;
if (!(*rover->t_slope) && !gr_frontsector->c_slope && !gr_backsector->c_slope && h > highcut)
h = hS = highcut;
if (!(*rover->b_slope) && !gr_frontsector->f_slope && !gr_backsector->f_slope && l < lowcut)
l = lS = lowcut;
//Hurdler: HW code starts here
//FIXME: check if peging is correct
// set top/bottom coords
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
wallVerts[3].y = FIXED_TO_FLOAT(h);
wallVerts[2].y = FIXED_TO_FLOAT(hS);
wallVerts[0].y = FIXED_TO_FLOAT(l);
wallVerts[1].y = FIXED_TO_FLOAT(lS);
Fix FOF slope rendering in ogl should work as well as software if not better now
2016-01-25 05:57:53 +00:00
#else
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
h = *rover->topheight;
l = *rover->bottomheight;
if (h > highcut)
h = highcut;
if (l < lowcut)
l = lowcut;
//Hurdler: HW code starts here
//FIXME: check if peging is correct
// set top/bottom coords
wallVerts[2].y = wallVerts[3].y = FIXED_TO_FLOAT(h);
wallVerts[0].y = wallVerts[1].y = FIXED_TO_FLOAT(l);
Fix FOF slope rendering in ogl should work as well as software if not better now
2016-01-25 05:57:53 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (rover->flags & FF_FOG)
{
wallVerts[3].t = wallVerts[2].t = 0;
wallVerts[0].t = wallVerts[1].t = 0;
wallVerts[0].s = wallVerts[3].s = 0;
wallVerts[2].s = wallVerts[1].s = 0;
}
Remove pointless drawtextured variable and redundant fake planes checks (R_FakeFlat would already have made gr_frontsector/backsector something else if they were)
2017-01-12 21:43:37 +00:00
else
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
Fix OpenGL completely missing the ability to alter FOF wall pegging by lower unpegged flag. Stupid OpenGL. Sorry in advance Lat'!
2018-06-26 20:41:05 +00:00
fixed_t texturevpeg;
Redo OpenGL skewing support, so that it now itself supports the lower unpegged effect for FOFs in OpenGL!
2018-09-28 16:02:02 +00:00
boolean attachtobottom = false;
#ifdef ESLOPE
boolean slopeskew = false; // skew FOF walls with slopes?
#endif
OpenGL Multi-Property FOF's Multi-Property FOF's have their multiple side properties display in OpenGL.
2014-04-17 12:02:34 +00:00
Fix OpenGL completely missing the ability to alter FOF wall pegging by lower unpegged flag. Stupid OpenGL. Sorry in advance Lat'!
2018-06-26 20:41:05 +00:00
// Wow, how was this missing from OpenGL for so long?
// ...Oh well, anyway, Lower Unpegged now changes pegging of FOFs like in software
// -- Monster Iestyn 26/06/18
OpenGL Multi-Property FOF's Multi-Property FOF's have their multiple side properties display in OpenGL.
2014-04-17 12:02:34 +00:00
if (newline)
{
Fix OpenGL completely missing the ability to alter FOF wall pegging by lower unpegged flag. Stupid OpenGL. Sorry in advance Lat'!
2018-06-26 20:41:05 +00:00
texturevpeg = sides[newline->sidenum[0]].rowoffset;
Redo OpenGL skewing support, so that it now itself supports the lower unpegged effect for FOFs in OpenGL!
2018-09-28 16:02:02 +00:00
attachtobottom = !!(newline->flags & ML_DONTPEGBOTTOM);
#ifdef ESLOPE
slopeskew = !!(newline->flags & ML_DONTPEGTOP);
#endif
OpenGL Multi-Property FOF's Multi-Property FOF's have their multiple side properties display in OpenGL.
2014-04-17 12:02:34 +00:00
}
else
{
Fix OpenGL completely missing the ability to alter FOF wall pegging by lower unpegged flag. Stupid OpenGL. Sorry in advance Lat'!
2018-06-26 20:41:05 +00:00
texturevpeg = sides[rover->master->sidenum[0]].rowoffset;
Redo OpenGL skewing support, so that it now itself supports the lower unpegged effect for FOFs in OpenGL!
2018-09-28 16:02:02 +00:00
attachtobottom = !!(gr_linedef->flags & ML_DONTPEGBOTTOM);
#ifdef ESLOPE
slopeskew = !!(rover->master->flags & ML_DONTPEGTOP);
#endif
OpenGL Multi-Property FOF's Multi-Property FOF's have their multiple side properties display in OpenGL.
2014-04-17 12:02:34 +00:00
}
Fix OpenGL completely missing the ability to alter FOF wall pegging by lower unpegged flag. Stupid OpenGL. Sorry in advance Lat'!
2018-06-26 20:41:05 +00:00
grTex = HWR_GetTexture(texnum);
#ifdef ESLOPE
Redo OpenGL skewing support, so that it now itself supports the lower unpegged effect for FOFs in OpenGL!
2018-09-28 16:02:02 +00:00
if (!slopeskew) // no skewing
{
if (attachtobottom)
texturevpeg -= *rover->topheight - *rover->bottomheight;
wallVerts[3].t = (*rover->topheight - h + texturevpeg) * grTex->scaleY;
wallVerts[2].t = (*rover->topheight - hS + texturevpeg) * grTex->scaleY;
wallVerts[0].t = (*rover->topheight - l + texturevpeg) * grTex->scaleY;
wallVerts[1].t = (*rover->topheight - lS + texturevpeg) * grTex->scaleY;
}
else
{
if (!attachtobottom) // skew by top
{
wallVerts[3].t = wallVerts[2].t = texturevpeg * grTex->scaleY;
wallVerts[0].t = (h - l + texturevpeg) * grTex->scaleY;
wallVerts[1].t = (hS - lS + texturevpeg) * grTex->scaleY;
}
else // skew by bottom
{
wallVerts[0].t = wallVerts[1].t = texturevpeg * grTex->scaleY;
wallVerts[3].t = wallVerts[0].t - (h - l) * grTex->scaleY;
wallVerts[2].t = wallVerts[1].t - (hS - lS) * grTex->scaleY;
}
}
Fix OpenGL completely missing the ability to alter FOF wall pegging by lower unpegged flag. Stupid OpenGL. Sorry in advance Lat'!
2018-06-26 20:41:05 +00:00
#else
Redo OpenGL skewing support, so that it now itself supports the lower unpegged effect for FOFs in OpenGL!
2018-09-28 16:02:02 +00:00
if (attachtobottom)
texturevpeg -= *rover->topheight - *rover->bottomheight;
Fix OpenGL completely missing the ability to alter FOF wall pegging by lower unpegged flag. Stupid OpenGL. Sorry in advance Lat'!
2018-06-26 20:41:05 +00:00
wallVerts[3].t = wallVerts[2].t = (*rover->topheight - h + texturevpeg) * grTex->scaleY;
wallVerts[0].t = wallVerts[1].t = (*rover->topheight - l + texturevpeg) * grTex->scaleY;
Fix FOF slope rendering in ogl should work as well as software if not better now
2016-01-25 05:57:53 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
wallVerts[0].s = wallVerts[3].s = cliplow * grTex->scaleX;
wallVerts[2].s = wallVerts[1].s = cliphigh * grTex->scaleX;
}
if (rover->flags & FF_FOG)
{
FBITFIELD blendmode;
Better fog block colouring They still aren't perfect, but now they are at least not quite so obviously just translucent polygons over the level. A mixture between partially modulating the background colours and adding the fog colour. Notably white fog blocks look like they're brightening what's behind them. Additive was also setting noalphatest before, can probably decide that depending on what it needs anyway. I don't think it's currently used anyway.
2018-03-15 23:58:37 +00:00
blendmode = PF_Fog|PF_NoTexture;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
lightnum = rover->master->frontsector->lightlevel;
colormap = rover->master->frontsector->extra_colormap;
if (rover->master->frontsector->extra_colormap)
{
Better fog block colouring They still aren't perfect, but now they are at least not quite so obviously just translucent polygons over the level. A mixture between partially modulating the background colours and adding the fog colour. Notably white fog blocks look like they're brightening what's behind them. Additive was also setting noalphatest before, can probably decide that depending on what it needs anyway. I don't think it's currently used anyway.
2018-03-15 23:58:37 +00:00
Surf.FlatColor.s.alpha = HWR_FogBlockAlpha(rover->master->frontsector->lightlevel,rover->master->frontsector->extra_colormap->rgba);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
else
{
Better fog block colouring They still aren't perfect, but now they are at least not quite so obviously just translucent polygons over the level. A mixture between partially modulating the background colours and adding the fog colour. Notably white fog blocks look like they're brightening what's behind them. Additive was also setting noalphatest before, can probably decide that depending on what it needs anyway. I don't think it's currently used anyway.
2018-03-15 23:58:37 +00:00
Surf.FlatColor.s.alpha = HWR_FogBlockAlpha(rover->master->frontsector->lightlevel,NORMALFOG);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
if (gr_frontsector->numlights)
Some fixes and updates for HWR_SplitWall Solid walls *can* be cut Fix issues with water and fog FOFs not cutting each other out correctly Fix Fog colourmap and lighting setting that is done here. Remove HWR_SplitFog There is currently a bug with FF_DOUBLESHADOW (that also exists in software) but has a larger impact here. When 2 FF_DOUBLESHADOW lights are directly stacked on each other the bottom one has its height set incorrectly. This causes all the Fog in the timed gravity flipping section of ERZ2 to be drawn and it looks really bad.
2018-03-20 14:20:02 +00:00
HWR_SplitWall(gr_frontsector, wallVerts, 0, &Surf, rover->flags, rover);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
else
HWR_AddTransparentWall(wallVerts, &Surf, 0, blendmode, true, lightnum, colormap);
}
else
{
FBITFIELD blendmode = PF_Masked;
Various Transparent Texture Fixes Draw Textures and Flats that have holes in them like a solid polygon so they use the depth buffer and don't need to be sorted Disable all linear filtering on textures and flats that have holes in them, the linear filtering introduces translucency into the textures where the edges are. Leaving them with either a black border, or causing pixels behind the slightly translucent areas to not be drawn. Doesn't apply to sprites and the HUD as they are always already sorted properly. Make the Alpha Testing more strict on non-translucent blend modes. This makes it so any transparency below 0.5 is discarded instead. Would make anything that is blended and has holes in it look slightly better, only the HUD and MD2s where the texture has holes are effected currently. Set TF_TRANSPARENT on flat texture flags when there are holes in the texture. Minor fix to make sure MD2s always set the right blend mode
2017-09-08 23:41:11 +00:00
if (rover->flags & FF_TRANSLUCENT && rover->alpha < 256)
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
blendmode = PF_Translucent;
Surf.FlatColor.s.alpha = (UINT8)rover->alpha-1 > 255 ? 255 : rover->alpha-1;
}
if (gr_frontsector->numlights)
Some fixes and updates for HWR_SplitWall Solid walls *can* be cut Fix issues with water and fog FOFs not cutting each other out correctly Fix Fog colourmap and lighting setting that is done here. Remove HWR_SplitFog There is currently a bug with FF_DOUBLESHADOW (that also exists in software) but has a larger impact here. When 2 FF_DOUBLESHADOW lights are directly stacked on each other the bottom one has its height set incorrectly. This causes all the Fog in the timed gravity flipping section of ERZ2 to be drawn and it looks really bad.
2018-03-20 14:20:02 +00:00
HWR_SplitWall(gr_frontsector, wallVerts, texnum, &Surf, rover->flags, rover);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
else
{
if (blendmode != PF_Masked)
OpenGL Multi-Property FOF's Multi-Property FOF's have their multiple side properties display in OpenGL.
2014-04-17 12:02:34 +00:00
HWR_AddTransparentWall(wallVerts, &Surf, texnum, blendmode, false, lightnum, colormap);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
else
HWR_ProjectWall(wallVerts, &Surf, PF_Masked, lightnum, colormap);
}
}
}
}
if (gr_frontsector->ffloors) // Putting this seperate should allow 2 FOF sectors to be connected without too many errors? I think?
{
for (rover = gr_frontsector->ffloors; rover; rover = rover->next)
{
if (!(rover->flags & FF_EXISTS) || !(rover->flags & FF_RENDERSIDES) || !(rover->flags & FF_ALLSIDES))
continue;
if (*rover->topheight < lowcut || *rover->bottomheight > highcut)
continue;
Created R_GetTextureNum to make sure top/bottom/midtexture texture ids are always valid in rendering code for both software and OpenGL (and also for the Solid Midtexture effect physics code)
2016-12-16 21:38:53 +00:00
texnum = R_GetTextureNum(sides[rover->master->sidenum[0]].midtexture);
OpenGL Multi-Property FOF's Multi-Property FOF's have their multiple side properties display in OpenGL.
2014-04-17 12:02:34 +00:00
if (rover->master->flags & ML_TFERLINE)
{
size_t linenum = gr_curline->linedef-gr_backsector->lines[0];
newline = rover->master->frontsector->lines[0] + linenum;
Created R_GetTextureNum to make sure top/bottom/midtexture texture ids are always valid in rendering code for both software and OpenGL (and also for the Solid Midtexture effect physics code)
2016-12-16 21:38:53 +00:00
texnum = R_GetTextureNum(sides[newline->sidenum[0]].midtexture);
OpenGL Multi-Property FOF's Multi-Property FOF's have their multiple side properties display in OpenGL.
2014-04-17 12:02:34 +00:00
}
Fix FOF slope rendering in ogl should work as well as software if not better now
2016-01-25 05:57:53 +00:00
#ifdef ESLOPE //backsides
h = *rover->t_slope ? P_GetZAt(*rover->t_slope, v1x, v1y) : *rover->topheight;
hS = *rover->t_slope ? P_GetZAt(*rover->t_slope, v2x, v2y) : *rover->topheight;
l = *rover->b_slope ? P_GetZAt(*rover->b_slope, v1x, v1y) : *rover->bottomheight;
lS = *rover->b_slope ? P_GetZAt(*rover->b_slope, v2x, v2y) : *rover->bottomheight;
if (!(*rover->t_slope) && !gr_frontsector->c_slope && !gr_backsector->c_slope && h > highcut)
h = hS = highcut;
if (!(*rover->b_slope) && !gr_frontsector->f_slope && !gr_backsector->f_slope && l < lowcut)
l = lS = lowcut;
//Hurdler: HW code starts here
//FIXME: check if peging is correct
// set top/bottom coords
OpenGL Multi-Property FOF's Multi-Property FOF's have their multiple side properties display in OpenGL.
2014-04-17 12:02:34 +00:00
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
wallVerts[3].y = FIXED_TO_FLOAT(h);
wallVerts[2].y = FIXED_TO_FLOAT(hS);
wallVerts[0].y = FIXED_TO_FLOAT(l);
wallVerts[1].y = FIXED_TO_FLOAT(lS);
Fix FOF slope rendering in ogl should work as well as software if not better now
2016-01-25 05:57:53 +00:00
#else
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
h = *rover->topheight;
l = *rover->bottomheight;
if (h > highcut)
h = highcut;
if (l < lowcut)
l = lowcut;
//Hurdler: HW code starts here
//FIXME: check if peging is correct
// set top/bottom coords
wallVerts[2].y = wallVerts[3].y = FIXED_TO_FLOAT(h);
wallVerts[0].y = wallVerts[1].y = FIXED_TO_FLOAT(l);
Fix FOF slope rendering in ogl should work as well as software if not better now
2016-01-25 05:57:53 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (rover->flags & FF_FOG)
{
wallVerts[3].t = wallVerts[2].t = 0;
wallVerts[0].t = wallVerts[1].t = 0;
wallVerts[0].s = wallVerts[3].s = 0;
wallVerts[2].s = wallVerts[1].s = 0;
}
Remove pointless drawtextured variable and redundant fake planes checks (R_FakeFlat would already have made gr_frontsector/backsector something else if they were)
2017-01-12 21:43:37 +00:00
else
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
OpenGL Multi-Property FOF's Multi-Property FOF's have their multiple side properties display in OpenGL.
2014-04-17 12:02:34 +00:00
grTex = HWR_GetTexture(texnum);
if (newline)
{
wallVerts[3].t = wallVerts[2].t = (*rover->topheight - h + sides[newline->sidenum[0]].rowoffset) * grTex->scaleY;
wallVerts[0].t = wallVerts[1].t = (h - l + (*rover->topheight - h + sides[newline->sidenum[0]].rowoffset)) * grTex->scaleY;
}
else
{
wallVerts[3].t = wallVerts[2].t = (*rover->topheight - h + sides[rover->master->sidenum[0]].rowoffset) * grTex->scaleY;
wallVerts[0].t = wallVerts[1].t = (h - l + (*rover->topheight - h + sides[rover->master->sidenum[0]].rowoffset)) * grTex->scaleY;
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
wallVerts[0].s = wallVerts[3].s = cliplow * grTex->scaleX;
wallVerts[2].s = wallVerts[1].s = cliphigh * grTex->scaleX;
}
if (rover->flags & FF_FOG)
{
FBITFIELD blendmode;
Better fog block colouring They still aren't perfect, but now they are at least not quite so obviously just translucent polygons over the level. A mixture between partially modulating the background colours and adding the fog colour. Notably white fog blocks look like they're brightening what's behind them. Additive was also setting noalphatest before, can probably decide that depending on what it needs anyway. I don't think it's currently used anyway.
2018-03-15 23:58:37 +00:00
blendmode = PF_Fog|PF_NoTexture;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
lightnum = rover->master->frontsector->lightlevel;
colormap = rover->master->frontsector->extra_colormap;
if (rover->master->frontsector->extra_colormap)
{
Better fog block colouring They still aren't perfect, but now they are at least not quite so obviously just translucent polygons over the level. A mixture between partially modulating the background colours and adding the fog colour. Notably white fog blocks look like they're brightening what's behind them. Additive was also setting noalphatest before, can probably decide that depending on what it needs anyway. I don't think it's currently used anyway.
2018-03-15 23:58:37 +00:00
Surf.FlatColor.s.alpha = HWR_FogBlockAlpha(rover->master->frontsector->lightlevel,rover->master->frontsector->extra_colormap->rgba);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
else
{
Better fog block colouring They still aren't perfect, but now they are at least not quite so obviously just translucent polygons over the level. A mixture between partially modulating the background colours and adding the fog colour. Notably white fog blocks look like they're brightening what's behind them. Additive was also setting noalphatest before, can probably decide that depending on what it needs anyway. I don't think it's currently used anyway.
2018-03-15 23:58:37 +00:00
Surf.FlatColor.s.alpha = HWR_FogBlockAlpha(rover->master->frontsector->lightlevel,NORMALFOG);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
if (gr_backsector->numlights)
Some fixes and updates for HWR_SplitWall Solid walls *can* be cut Fix issues with water and fog FOFs not cutting each other out correctly Fix Fog colourmap and lighting setting that is done here. Remove HWR_SplitFog There is currently a bug with FF_DOUBLESHADOW (that also exists in software) but has a larger impact here. When 2 FF_DOUBLESHADOW lights are directly stacked on each other the bottom one has its height set incorrectly. This causes all the Fog in the timed gravity flipping section of ERZ2 to be drawn and it looks really bad.
2018-03-20 14:20:02 +00:00
HWR_SplitWall(gr_backsector, wallVerts, 0, &Surf, rover->flags, rover);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
else
HWR_AddTransparentWall(wallVerts, &Surf, 0, blendmode, true, lightnum, colormap);
}
else
{
FBITFIELD blendmode = PF_Masked;
Various Transparent Texture Fixes Draw Textures and Flats that have holes in them like a solid polygon so they use the depth buffer and don't need to be sorted Disable all linear filtering on textures and flats that have holes in them, the linear filtering introduces translucency into the textures where the edges are. Leaving them with either a black border, or causing pixels behind the slightly translucent areas to not be drawn. Doesn't apply to sprites and the HUD as they are always already sorted properly. Make the Alpha Testing more strict on non-translucent blend modes. This makes it so any transparency below 0.5 is discarded instead. Would make anything that is blended and has holes in it look slightly better, only the HUD and MD2s where the texture has holes are effected currently. Set TF_TRANSPARENT on flat texture flags when there are holes in the texture. Minor fix to make sure MD2s always set the right blend mode
2017-09-08 23:41:11 +00:00
if (rover->flags & FF_TRANSLUCENT && rover->alpha < 256)
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
blendmode = PF_Translucent;
Surf.FlatColor.s.alpha = (UINT8)rover->alpha-1 > 255 ? 255 : rover->alpha-1;
}
if (gr_backsector->numlights)
Some fixes and updates for HWR_SplitWall Solid walls *can* be cut Fix issues with water and fog FOFs not cutting each other out correctly Fix Fog colourmap and lighting setting that is done here. Remove HWR_SplitFog There is currently a bug with FF_DOUBLESHADOW (that also exists in software) but has a larger impact here. When 2 FF_DOUBLESHADOW lights are directly stacked on each other the bottom one has its height set incorrectly. This causes all the Fog in the timed gravity flipping section of ERZ2 to be drawn and it looks really bad.
2018-03-20 14:20:02 +00:00
HWR_SplitWall(gr_backsector, wallVerts, texnum, &Surf, rover->flags, rover);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
else
{
if (blendmode != PF_Masked)
OpenGL Multi-Property FOF's Multi-Property FOF's have their multiple side properties display in OpenGL.
2014-04-17 12:02:34 +00:00
HWR_AddTransparentWall(wallVerts, &Surf, texnum, blendmode, false, lightnum, colormap);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
else
HWR_ProjectWall(wallVerts, &Surf, PF_Masked, lightnum, colormap);
}
}
}
}
}
#endif
//Hurdler: end of 3d-floors test
}
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
// From PrBoom:
//
// e6y: Check whether the player can look beyond this line
//
#ifdef NEWCLIP
Made some efforts to improve efficiency of new code, hard to tell if I've made it better or worse though honestly R_IsEmptyLine is now a thing too btw
2017-01-10 20:07:02 +00:00
boolean checkforemptylines = true;
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
// Don't modify anything here, just check
// Kalaron: Modified for sloped linedefs
static boolean CheckClip(seg_t * seg, sector_t * afrontsector, sector_t * abacksector)
{
fixed_t frontf1,frontf2, frontc1, frontc2; // front floor/ceiling ends
fixed_t backf1, backf2, backc1, backc2; // back floor ceiling ends
// GZDoom method of sloped line clipping
#ifdef ESLOPE
if (afrontsector->f_slope || afrontsector->c_slope || abacksector->f_slope || abacksector->c_slope)
{
fixed_t v1x, v1y, v2x, v2y; // the seg's vertexes as fixed_t
v1x = FLOAT_TO_FIXED(((polyvertex_t *)gr_curline->pv1)->x);
v1y = FLOAT_TO_FIXED(((polyvertex_t *)gr_curline->pv1)->y);
v2x = FLOAT_TO_FIXED(((polyvertex_t *)gr_curline->pv2)->x);
v2y = FLOAT_TO_FIXED(((polyvertex_t *)gr_curline->pv2)->y);
#define SLOPEPARAMS(slope, end1, end2, normalheight) \
if (slope) { \
end1 = P_GetZAt(slope, v1x, v1y); \
end2 = P_GetZAt(slope, v2x, v2y); \
} else \
end1 = end2 = normalheight;
SLOPEPARAMS(afrontsector->f_slope, frontf1, frontf2, afrontsector->floorheight)
SLOPEPARAMS(afrontsector->c_slope, frontc1, frontc2, afrontsector->ceilingheight)
SLOPEPARAMS( abacksector->f_slope, backf1, backf2, abacksector->floorheight)
SLOPEPARAMS( abacksector->c_slope, backc1, backc2, abacksector->ceilingheight)
#undef SLOPEPARAMS
}
else
#endif
{
frontf1 = frontf2 = afrontsector->floorheight;
frontc1 = frontc2 = afrontsector->ceilingheight;
backf1 = backf2 = abacksector->floorheight;
backc1 = backc2 = abacksector->ceilingheight;
}
// now check for closed sectors!
if (backc1 <= frontf1 && backc2 <= frontf2)
{
Made some efforts to improve efficiency of new code, hard to tell if I've made it better or worse though honestly R_IsEmptyLine is now a thing too btw
2017-01-10 20:07:02 +00:00
checkforemptylines = false;
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
if (!seg->sidedef->toptexture)
return false;
if (abacksector->ceilingpic == skyflatnum && afrontsector->ceilingpic == skyflatnum)
return false;
return true;
}
if (backf1 >= frontc1 && backf2 >= frontc2)
{
Made some efforts to improve efficiency of new code, hard to tell if I've made it better or worse though honestly R_IsEmptyLine is now a thing too btw
2017-01-10 20:07:02 +00:00
checkforemptylines = false;
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
if (!seg->sidedef->bottomtexture)
return false;
// properly render skies (consider door "open" if both floors are sky):
if (abacksector->ceilingpic == skyflatnum && afrontsector->ceilingpic == skyflatnum)
return false;
return true;
}
if (backc1 <= backf1 && backc2 <= backf2)
{
Made some efforts to improve efficiency of new code, hard to tell if I've made it better or worse though honestly R_IsEmptyLine is now a thing too btw
2017-01-10 20:07:02 +00:00
checkforemptylines = false;
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
// preserve a kind of transparent door/lift special effect:
if (backc1 < frontc1 || backc2 < frontc2)
{
if (!seg->sidedef->toptexture)
return false;
}
if (backf1 > frontf1 || backf2 > frontf2)
{
if (!seg->sidedef->bottomtexture)
return false;
}
if (abacksector->ceilingpic == skyflatnum && afrontsector->ceilingpic == skyflatnum)
return false;
if (abacksector->floorpic == skyflatnum && afrontsector->floorpic == skyflatnum)
return false;
return true;
}
Made some efforts to improve efficiency of new code, hard to tell if I've made it better or worse though honestly R_IsEmptyLine is now a thing too btw
2017-01-10 20:07:02 +00:00
if (backc1 != frontc1 || backc2 != frontc2
|| backf1 != frontf1 || backf2 != frontf2)
{
checkforemptylines = false;
return false;
}
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
return false;
}
#else
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
//Hurdler: just like in r_bsp.c
#if 1
#define MAXSEGS MAXVIDWIDTH/2+1
#else
//Alam_GBC: Or not (may cause overflow)
#define MAXSEGS 128
#endif
// hw_newend is one past the last valid seg
static cliprange_t * hw_newend;
static cliprange_t gr_solidsegs[MAXSEGS];
static void printsolidsegs(void)
{
cliprange_t * start;
if (!hw_newend || cv_grbeta.value != 2)
return;
for (start = gr_solidsegs;start != hw_newend;start++)
{
CONS_Debug(DBG_RENDER, "%d-%d|",start->first,start->last);
}
CONS_Debug(DBG_RENDER, "\n\n");
}
//
//
//
static void HWR_ClipSolidWallSegment(INT32 first, INT32 last)
{
cliprange_t *next, *start;
float lowfrac, highfrac;
boolean poorhack = false;
// Find the first range that touches the range
// (adjacent pixels are touching).
start = gr_solidsegs;
while (start->last < first-1)
start++;
if (first < start->first)
{
if (last < start->first-1)
{
// Post is entirely visible (above start),
// so insert a new clippost.
HWR_StoreWallRange(first, last);
next = hw_newend;
hw_newend++;
while (next != start)
{
*next = *(next-1);
next--;
}
next->first = first;
next->last = last;
printsolidsegs();
return;
}
// There is a fragment above *start.
if (!cv_grclipwalls.value)
{
if (!poorhack) HWR_StoreWallRange(first, last);
poorhack = true;
}
else
{
Make new pv1/pv2 seg pointers, so AdjustSeg doesn't modify the v1/v2 pointers directly anymore Yes I know they're void * in r_defs.h's seg_t definition, it's quicker than trying to figure out if including hardware/hw_glob.h is a good idea or not
2016-11-07 21:55:56 +00:00
highfrac = HWR_ClipViewSegment(start->first+1, (polyvertex_t *)gr_curline->pv1, (polyvertex_t *)gr_curline->pv2);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
HWR_StoreWallRange(0, highfrac);
}
// Now adjust the clip size.
start->first = first;
}
// Bottom contained in start?
if (last <= start->last)
{
printsolidsegs();
return;
}
next = start;
while (last >= (next+1)->first-1)
{
// There is a fragment between two posts.
if (!cv_grclipwalls.value)
{
if (!poorhack) HWR_StoreWallRange(first,last);
poorhack = true;
}
else
{
Make new pv1/pv2 seg pointers, so AdjustSeg doesn't modify the v1/v2 pointers directly anymore Yes I know they're void * in r_defs.h's seg_t definition, it's quicker than trying to figure out if including hardware/hw_glob.h is a good idea or not
2016-11-07 21:55:56 +00:00
lowfrac = HWR_ClipViewSegment(next->last-1, (polyvertex_t *)gr_curline->pv1, (polyvertex_t *)gr_curline->pv2);
highfrac = HWR_ClipViewSegment((next+1)->first+1, (polyvertex_t *)gr_curline->pv1, (polyvertex_t *)gr_curline->pv2);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
HWR_StoreWallRange(lowfrac, highfrac);
}
next++;
if (last <= next->last)
{
// Bottom is contained in next.
// Adjust the clip size.
start->last = next->last;
goto crunch;
}
}
if (first == next->first+1) // 1 line texture
{
if (!cv_grclipwalls.value)
{
if (!poorhack) HWR_StoreWallRange(first,last);
poorhack = true;
}
else
HWR_StoreWallRange(0, 1);
}
else
{
// There is a fragment after *next.
if (!cv_grclipwalls.value)
{
if (!poorhack) HWR_StoreWallRange(first,last);
poorhack = true;
}
else
{
Make new pv1/pv2 seg pointers, so AdjustSeg doesn't modify the v1/v2 pointers directly anymore Yes I know they're void * in r_defs.h's seg_t definition, it's quicker than trying to figure out if including hardware/hw_glob.h is a good idea or not
2016-11-07 21:55:56 +00:00
lowfrac = HWR_ClipViewSegment(next->last-1, (polyvertex_t *)gr_curline->pv1, (polyvertex_t *)gr_curline->pv2);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
HWR_StoreWallRange(lowfrac, 1);
}
}
// Adjust the clip size.
start->last = last;
// Remove start+1 to next from the clip list,
// because start now covers their area.
crunch:
if (next == start)
{
printsolidsegs();
// Post just extended past the bottom of one post.
return;
}
while (next++ != hw_newend)
{
// Remove a post.
*++start = *next;
}
hw_newend = start;
printsolidsegs();
}
//
// handle LineDefs with upper and lower texture (windows)
//
static void HWR_ClipPassWallSegment(INT32 first, INT32 last)
{
cliprange_t *start;
float lowfrac, highfrac;
//to allow noclipwalls but still solidseg reject of non-visible walls
boolean poorhack = false;
// Find the first range that touches the range
// (adjacent pixels are touching).
start = gr_solidsegs;
while (start->last < first - 1)
start++;
if (first < start->first)
{
if (last < start->first-1)
{
// Post is entirely visible (above start).
HWR_StoreWallRange(0, 1);
return;
}
// There is a fragment above *start.
if (!cv_grclipwalls.value)
{ //20/08/99: Changed by Hurdler (taken from faB's code)
if (!poorhack) HWR_StoreWallRange(0, 1);
poorhack = true;
}
else
{
highfrac = HWR_ClipViewSegment(min(start->first + 1,
Make new pv1/pv2 seg pointers, so AdjustSeg doesn't modify the v1/v2 pointers directly anymore Yes I know they're void * in r_defs.h's seg_t definition, it's quicker than trying to figure out if including hardware/hw_glob.h is a good idea or not
2016-11-07 21:55:56 +00:00
start->last), (polyvertex_t *)gr_curline->pv1,
(polyvertex_t *)gr_curline->pv2);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
HWR_StoreWallRange(0, highfrac);
}
}
// Bottom contained in start?
if (last <= start->last)
return;
while (last >= (start+1)->first-1)
{
// There is a fragment between two posts.
if (!cv_grclipwalls.value)
{
if (!poorhack) HWR_StoreWallRange(0, 1);
poorhack = true;
}
else
{
Make new pv1/pv2 seg pointers, so AdjustSeg doesn't modify the v1/v2 pointers directly anymore Yes I know they're void * in r_defs.h's seg_t definition, it's quicker than trying to figure out if including hardware/hw_glob.h is a good idea or not
2016-11-07 21:55:56 +00:00
lowfrac = HWR_ClipViewSegment(max(start->last-1,start->first), (polyvertex_t *)gr_curline->pv1, (polyvertex_t *)gr_curline->pv2);
highfrac = HWR_ClipViewSegment(min((start+1)->first+1,(start+1)->last), (polyvertex_t *)gr_curline->pv1, (polyvertex_t *)gr_curline->pv2);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
HWR_StoreWallRange(lowfrac, highfrac);
}
start++;
if (last <= start->last)
return;
}
if (first == start->first+1) // 1 line texture
{
if (!cv_grclipwalls.value)
{
if (!poorhack) HWR_StoreWallRange(0, 1);
poorhack = true;
}
else
HWR_StoreWallRange(0, 1);
}
else
{
// There is a fragment after *next.
if (!cv_grclipwalls.value)
{
if (!poorhack) HWR_StoreWallRange(0,1);
poorhack = true;
}
else
{
lowfrac = HWR_ClipViewSegment(max(start->last - 1,
Make new pv1/pv2 seg pointers, so AdjustSeg doesn't modify the v1/v2 pointers directly anymore Yes I know they're void * in r_defs.h's seg_t definition, it's quicker than trying to figure out if including hardware/hw_glob.h is a good idea or not
2016-11-07 21:55:56 +00:00
start->first), (polyvertex_t *)gr_curline->pv1,
(polyvertex_t *)gr_curline->pv2);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
HWR_StoreWallRange(lowfrac, 1);
}
}
}
// --------------------------------------------------------------------------
// HWR_ClipToSolidSegs check if it is hide by wall (solidsegs)
// --------------------------------------------------------------------------
static boolean HWR_ClipToSolidSegs(INT32 first, INT32 last)
{
cliprange_t * start;
// Find the first range that touches the range
// (adjacent pixels are touching).
start = gr_solidsegs;
while (start->last < first-1)
start++;
if (first < start->first)
return true;
// Bottom contained in start?
if (last <= start->last)
return false;
return true;
}
//
// HWR_ClearClipSegs
//
static void HWR_ClearClipSegs(void)
{
gr_solidsegs[0].first = -0x7fffffff;
gr_solidsegs[0].last = -1;
gr_solidsegs[1].first = vid.width; //viewwidth;
gr_solidsegs[1].last = 0x7fffffff;
hw_newend = gr_solidsegs+2;
}
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#endif // NEWCLIP
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// -----------------+
// HWR_AddLine : Clips the given segment and adds any visible pieces to the line list.
// Notes : gr_cursectorlight is set to the current subsector -> sector -> light value
// : (it may be mixed with the wall's own flat colour in the future ...)
// -----------------+
static void HWR_AddLine(seg_t * line)
{
angle_t angle1, angle2;
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#ifndef NEWCLIP
INT32 x1, x2;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
angle_t span, tspan;
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// SoM: Backsector needs to be run through R_FakeFlat
static tempsec for R_FakeFlat I don't fully understand this, but it's what software does and it fixes the issue of the lighting in DSZ3. Also don't need the extra call to R_Prep3DFloors.
2018-03-29 22:28:54 +00:00
static sector_t tempsec;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
Added missing checks in HWR_AddLine from the software version, move v** vars to the top since R_PointToAngle calls use the same values anyway
2016-12-21 22:31:09 +00:00
fixed_t v1x, v1y, v2x, v2y; // the seg's vertexes as fixed_t
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#ifdef POLYOBJECTS
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (line->polyseg && !(line->polyseg->flags & POF_RENDERSIDES))
return;
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
gr_curline = line;
Added missing checks in HWR_AddLine from the software version, move v** vars to the top since R_PointToAngle calls use the same values anyway
2016-12-21 22:31:09 +00:00
v1x = FLOAT_TO_FIXED(((polyvertex_t *)gr_curline->pv1)->x);
v1y = FLOAT_TO_FIXED(((polyvertex_t *)gr_curline->pv1)->y);
v2x = FLOAT_TO_FIXED(((polyvertex_t *)gr_curline->pv2)->x);
v2y = FLOAT_TO_FIXED(((polyvertex_t *)gr_curline->pv2)->y);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// OPTIMIZE: quickly reject orthogonal back sides.
Added missing checks in HWR_AddLine from the software version, move v** vars to the top since R_PointToAngle calls use the same values anyway
2016-12-21 22:31:09 +00:00
angle1 = R_PointToAngle(v1x, v1y);
angle2 = R_PointToAngle(v2x, v2y);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#ifdef NEWCLIP
// PrBoom: Back side, i.e. backface culling - read: endAngle >= startAngle!
if (angle2 - angle1 < ANGLE_180)
return;
// PrBoom: use REAL clipping math YAYYYYYYY!!!
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
if (!gld_clipper_SafeCheckRange(angle2, angle1))
{
return;
}
Made some efforts to improve efficiency of new code, hard to tell if I've made it better or worse though honestly R_IsEmptyLine is now a thing too btw
2017-01-10 20:07:02 +00:00
checkforemptylines = true;
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#else
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// Clip to view edges.
span = angle1 - angle2;
// backface culling : span is < ANGLE_180 if ang1 > ang2 : the seg is facing
if (span >= ANGLE_180)
return;
// Global angle needed by segcalc.
//rw_angle1 = angle1;
angle1 -= dup_viewangle;
angle2 -= dup_viewangle;
tspan = angle1 + gr_clipangle;
if (tspan > 2*gr_clipangle)
{
tspan -= 2*gr_clipangle;
// Totally off the left edge?
if (tspan >= span)
return;
angle1 = gr_clipangle;
}
tspan = gr_clipangle - angle2;
if (tspan > 2*gr_clipangle)
{
tspan -= 2*gr_clipangle;
// Totally off the left edge?
if (tspan >= span)
return;
angle2 = (angle_t)-(signed)gr_clipangle;
}
#if 0
{
float fx1,fx2,fy1,fy2;
//BP: test with a better projection than viewangletox[R_PointToAngle(angle)]
// do not enable this at release 4 mul and 2 div
Make new pv1/pv2 seg pointers, so AdjustSeg doesn't modify the v1/v2 pointers directly anymore Yes I know they're void * in r_defs.h's seg_t definition, it's quicker than trying to figure out if including hardware/hw_glob.h is a good idea or not
2016-11-07 21:55:56 +00:00
fx1 = ((polyvertex_t *)(line->pv1))->x-gr_viewx;
fy1 = ((polyvertex_t *)(line->pv1))->y-gr_viewy;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
fy2 = (fx1 * gr_viewcos + fy1 * gr_viewsin);
if (fy2 < 0)
// the point is back
fx1 = 0;
else
fx1 = gr_windowcenterx + (fx1 * gr_viewsin - fy1 * gr_viewcos) * gr_centerx / fy2;
Make new pv1/pv2 seg pointers, so AdjustSeg doesn't modify the v1/v2 pointers directly anymore Yes I know they're void * in r_defs.h's seg_t definition, it's quicker than trying to figure out if including hardware/hw_glob.h is a good idea or not
2016-11-07 21:55:56 +00:00
fx2 = ((polyvertex_t *)(line->pv2))->x-gr_viewx;
fy2 = ((polyvertex_t *)(line->pv2))->y-gr_viewy;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
fy1 = (fx2 * gr_viewcos + fy2 * gr_viewsin);
if (fy1 < 0)
// the point is back
fx2 = vid.width;
else
fx2 = gr_windowcenterx + (fx2 * gr_viewsin - fy2 * gr_viewcos) * gr_centerx / fy1;
x1 = fx1+0.5f;
x2 = fx2+0.5f;
}
#else
// The seg is in the view range,
// but not necessarily visible.
angle1 = (angle1+ANGLE_90)>>ANGLETOFINESHIFT;
angle2 = (angle2+ANGLE_90)>>ANGLETOFINESHIFT;
x1 = gr_viewangletox[angle1];
x2 = gr_viewangletox[angle2];
#endif
// Does not cross a pixel?
// if (x1 == x2)
/* {
// BP: HERE IS THE MAIN PROBLEM !
//CONS_Debug(DBG_RENDER, "tineline\n");
return;
}
*/
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
gr_backsector = line->backsector;
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#ifdef NEWCLIP
if (!line->backsector)
{
gld_clipper_SafeAddClipRange(angle2, angle1);
}
else
{
gr_backsector = R_FakeFlat(gr_backsector, &tempsec, NULL, NULL, true);
if (CheckClip(line, gr_frontsector, gr_backsector))
{
gld_clipper_SafeAddClipRange(angle2, angle1);
Made some efforts to improve efficiency of new code, hard to tell if I've made it better or worse though honestly R_IsEmptyLine is now a thing too btw
2017-01-10 20:07:02 +00:00
checkforemptylines = false;
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
}
Made some efforts to improve efficiency of new code, hard to tell if I've made it better or worse though honestly R_IsEmptyLine is now a thing too btw
2017-01-10 20:07:02 +00:00
// Reject empty lines used for triggers and special events.
// Identical floor and ceiling on both sides,
// identical light levels on both sides,
// and no middle texture.
if (checkforemptylines && R_IsEmptyLine(line, gr_frontsector, gr_backsector))
return;
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
}
HWR_ProcessSeg(); // Doesn't need arguments because they're defined globally :D
return;
#else
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// Single sided line?
if (!gr_backsector)
goto clipsolid;
gr_backsector = R_FakeFlat(gr_backsector, &tempsec, NULL, NULL, true);
OpenGL: closed door/window detection code now accounts for slopes, just like in software
2016-06-04 17:31:21 +00:00
#ifdef ESLOPE
if (gr_frontsector->f_slope || gr_frontsector->c_slope || gr_backsector->f_slope || gr_backsector->c_slope)
{
fixed_t frontf1,frontf2, frontc1, frontc2; // front floor/ceiling ends
fixed_t backf1, backf2, backc1, backc2; // back floor ceiling ends
#define SLOPEPARAMS(slope, end1, end2, normalheight) \
if (slope) { \
end1 = P_GetZAt(slope, v1x, v1y); \
end2 = P_GetZAt(slope, v2x, v2y); \
} else \
end1 = end2 = normalheight;
SLOPEPARAMS(gr_frontsector->f_slope, frontf1, frontf2, gr_frontsector->floorheight)
SLOPEPARAMS(gr_frontsector->c_slope, frontc1, frontc2, gr_frontsector->ceilingheight)
SLOPEPARAMS( gr_backsector->f_slope, backf1, backf2, gr_backsector->floorheight)
SLOPEPARAMS( gr_backsector->c_slope, backc1, backc2, gr_backsector->ceilingheight)
#undef SLOPEPARAMS
// Closed door.
if ((backc1 <= frontf1 && backc2 <= frontf2)
|| (backf1 >= frontc1 && backf2 >= frontc2))
{
goto clipsolid;
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
Added missing checks in HWR_AddLine from the software version, move v** vars to the top since R_PointToAngle calls use the same values anyway
2016-12-21 22:31:09 +00:00
// Check for automap fix.
if (backc1 <= backf1 && backc2 <= backf2
&& ((backc1 >= frontc1 && backc2 >= frontc2) || gr_curline->sidedef->toptexture)
&& ((backf1 <= frontf1 && backf2 >= frontf2) || gr_curline->sidedef->bottomtexture)
&& (gr_backsector->ceilingpic != skyflatnum || gr_frontsector->ceilingpic != skyflatnum))
goto clipsolid;
OpenGL: closed door/window detection code now accounts for slopes, just like in software
2016-06-04 17:31:21 +00:00
// Window.
if (backc1 != frontc1 || backc2 != frontc2
|| backf1 != frontf1 || backf2 != frontf2)
{
goto clippass;
}
}
else
#endif
{
// Closed door.
if (gr_backsector->ceilingheight <= gr_frontsector->floorheight ||
gr_backsector->floorheight >= gr_frontsector->ceilingheight)
goto clipsolid;
Added missing checks in HWR_AddLine from the software version, move v** vars to the top since R_PointToAngle calls use the same values anyway
2016-12-21 22:31:09 +00:00
// Check for automap fix.
if (gr_backsector->ceilingheight <= gr_backsector->floorheight
&& ((gr_backsector->ceilingheight >= gr_frontsector->ceilingheight) || gr_curline->sidedef->toptexture)
&& ((gr_backsector->floorheight <= gr_backsector->floorheight) || gr_curline->sidedef->bottomtexture)
&& (gr_backsector->ceilingpic != skyflatnum || gr_frontsector->ceilingpic != skyflatnum))
goto clipsolid;
OpenGL: closed door/window detection code now accounts for slopes, just like in software
2016-06-04 17:31:21 +00:00
// Window.
if (gr_backsector->ceilingheight != gr_frontsector->ceilingheight ||
gr_backsector->floorheight != gr_frontsector->floorheight)
goto clippass;
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// Reject empty lines used for triggers and special events.
// Identical floor and ceiling on both sides,
// identical light levels on both sides,
// and no middle texture.
Whoops, this broke my non-NEWCLIP test build
2017-05-31 16:17:08 +00:00
if (R_IsEmptyLine(gr_curline, gr_frontsector, gr_backsector))
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
return;
clippass:
if (x1 == x2)
{ x2++;x1 -= 2; }
HWR_ClipPassWallSegment(x1, x2-1);
return;
clipsolid:
if (x1 == x2)
goto clippass;
HWR_ClipSolidWallSegment(x1, x2-1);
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
// HWR_CheckBBox
// Checks BSP node/subtree bounding box.
// Returns true
// if some part of the bbox might be visible.
//
// modified to use local variables
static boolean HWR_CheckBBox(fixed_t *bspcoord)
{
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
INT32 boxpos;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
fixed_t px1, py1, px2, py2;
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
angle_t angle1, angle2;
#ifndef NEWCLIP
INT32 sx1, sx2;
angle_t span, tspan;
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// Find the corners of the box
// that define the edges from current viewpoint.
if (dup_viewx <= bspcoord[BOXLEFT])
boxpos = 0;
else if (dup_viewx < bspcoord[BOXRIGHT])
boxpos = 1;
else
boxpos = 2;
if (dup_viewy >= bspcoord[BOXTOP])
boxpos |= 0;
else if (dup_viewy > bspcoord[BOXBOTTOM])
boxpos |= 1<<2;
else
boxpos |= 2<<2;
if (boxpos == 5)
return true;
px1 = bspcoord[checkcoord[boxpos][0]];
py1 = bspcoord[checkcoord[boxpos][1]];
px2 = bspcoord[checkcoord[boxpos][2]];
py2 = bspcoord[checkcoord[boxpos][3]];
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#ifdef NEWCLIP
angle1 = R_PointToAngle(px1, py1);
angle2 = R_PointToAngle(px2, py2);
return gld_clipper_SafeCheckRange(angle2, angle1);
#else
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// check clip list for an open space
Very large map rendering issue fixed Move old fix for too large maps having rendering issues from R_CheckBBox to OpenGL's HWR_CheckBBox From what I know, this effects at least Aerial Garden and Seraphic Skylands
2018-05-26 12:13:37 +00:00
angle1 = R_PointToAngle2(dup_viewx>>1, dup_viewy>>1, px1>>1, py1>>1) - dup_viewangle;
angle2 = R_PointToAngle2(dup_viewx>>1, dup_viewy>>1, px2>>1, py2>>1) - dup_viewangle;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
span = angle1 - angle2;
// Sitting on a line?
if (span >= ANGLE_180)
return true;
tspan = angle1 + gr_clipangle;
if (tspan > 2*gr_clipangle)
{
tspan -= 2*gr_clipangle;
// Totally off the left edge?
if (tspan >= span)
return false;
angle1 = gr_clipangle;
}
tspan = gr_clipangle - angle2;
if (tspan > 2*gr_clipangle)
{
tspan -= 2*gr_clipangle;
// Totally off the left edge?
if (tspan >= span)
return false;
angle2 = (angle_t)-(signed)gr_clipangle;
}
// Find the first clippost
// that touches the source post
// (adjacent pixels are touching).
angle1 = (angle1+ANGLE_90)>>ANGLETOFINESHIFT;
angle2 = (angle2+ANGLE_90)>>ANGLETOFINESHIFT;
sx1 = gr_viewangletox[angle1];
sx2 = gr_viewangletox[angle2];
// Does not cross a pixel.
if (sx1 == sx2)
return false;
return HWR_ClipToSolidSegs(sx1, sx2 - 1);
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
#ifdef POLYOBJECTS
//
// HWR_AddPolyObjectSegs
//
// haleyjd 02/19/06
// Adds all segs in all polyobjects in the given subsector.
// Modified for hardware rendering.
//
static inline void HWR_AddPolyObjectSegs(void)
{
size_t i, j;
seg_t *gr_fakeline = Z_Calloc(sizeof(seg_t), PU_STATIC, NULL);
polyvertex_t *pv1 = Z_Calloc(sizeof(polyvertex_t), PU_STATIC, NULL);
polyvertex_t *pv2 = Z_Calloc(sizeof(polyvertex_t), PU_STATIC, NULL);
// Sort through all the polyobjects
for (i = 0; i < numpolys; ++i)
{
// Render the polyobject's lines
for (j = 0; j < po_ptrs[i]->segCount; ++j)
{
// Copy the info of a polyobject's seg, then convert it to OpenGL floating point
M_Memcpy(gr_fakeline, po_ptrs[i]->segs[j], sizeof(seg_t));
// Now convert the line to float and add it to be rendered
pv1->x = FIXED_TO_FLOAT(gr_fakeline->v1->x);
pv1->y = FIXED_TO_FLOAT(gr_fakeline->v1->y);
pv2->x = FIXED_TO_FLOAT(gr_fakeline->v2->x);
pv2->y = FIXED_TO_FLOAT(gr_fakeline->v2->y);
Make new pv1/pv2 seg pointers, so AdjustSeg doesn't modify the v1/v2 pointers directly anymore Yes I know they're void * in r_defs.h's seg_t definition, it's quicker than trying to figure out if including hardware/hw_glob.h is a good idea or not
2016-11-07 21:55:56 +00:00
gr_fakeline->pv1 = pv1;
gr_fakeline->pv2 = pv2;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
HWR_AddLine(gr_fakeline);
}
}
// Free temporary data no longer needed
Z_Free(pv2);
Z_Free(pv1);
Z_Free(gr_fakeline);
}
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
#ifdef POLYOBJECTS_PLANES
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
static void HWR_RenderPolyObjectPlane(polyobj_t *polysector, boolean isceiling, fixed_t fixedheight,
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
FBITFIELD blendmode, UINT8 lightlevel, lumpnum_t lumpnum, sector_t *FOFsector,
UINT8 alpha, extracolormap_t *planecolormap)
{
float height; //constant y for all points on the convex flat polygon
FOutVector *v3d;
INT32 i;
float flatxref,flatyref;
float fflatsize;
INT32 flatflag;
size_t len;
float scrollx = 0.0f, scrolly = 0.0f;
angle_t angle = 0;
FSurfaceInfo Surf;
fixed_t tempxsow, tempytow;
size_t nrPlaneVerts;
static FOutVector *planeVerts = NULL;
static UINT16 numAllocedPlaneVerts = 0;
nrPlaneVerts = polysector->numVertices;
height = FIXED_TO_FLOAT(fixedheight);
if (nrPlaneVerts < 3) //not even a triangle ?
return;
if (nrPlaneVerts > UINT16_MAX) // FIXME: exceeds plVerts size
{
oh, nPlaneVerts is a size_t?, well, then we need to use sizeu#()
2015-01-10 22:27:24 +00:00
CONS_Debug(DBG_RENDER, "polygon size of %s exceeds max value of %d vertices\n", sizeu1(nrPlaneVerts), UINT16_MAX);
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
return;
}
// Allocate plane-vertex buffer if we need to
if (!planeVerts || nrPlaneVerts > numAllocedPlaneVerts)
{
numAllocedPlaneVerts = (UINT16)nrPlaneVerts;
Z_Free(planeVerts);
Z_Malloc(numAllocedPlaneVerts * sizeof (FOutVector), PU_LEVEL, &planeVerts);
}
len = W_LumpLength(lumpnum);
switch (len)
{
case 4194304: // 2048x2048 lump
fflatsize = 2048.0f;
flatflag = 2047;
break;
case 1048576: // 1024x1024 lump
fflatsize = 1024.0f;
flatflag = 1023;
break;
case 262144:// 512x512 lump
fflatsize = 512.0f;
flatflag = 511;
break;
case 65536: // 256x256 lump
fflatsize = 256.0f;
flatflag = 255;
break;
case 16384: // 128x128 lump
fflatsize = 128.0f;
flatflag = 127;
break;
case 1024: // 32x32 lump
fflatsize = 32.0f;
flatflag = 31;
break;
default: // 64x64 lump
fflatsize = 64.0f;
flatflag = 63;
break;
}
// reference point for flat texture coord for each vertex around the polygon
flatxref = (float)(((fixed_t)FIXED_TO_FLOAT(polysector->origVerts[0].x) & (~flatflag)) / fflatsize);
flatyref = (float)(((fixed_t)FIXED_TO_FLOAT(polysector->origVerts[0].y) & (~flatflag)) / fflatsize);
// transform
v3d = planeVerts;
if (FOFsector != NULL)
{
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
if (!isceiling) // it's a floor
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
{
scrollx = FIXED_TO_FLOAT(FOFsector->floor_xoffs)/fflatsize;
scrolly = FIXED_TO_FLOAT(FOFsector->floor_yoffs)/fflatsize;
angle = FOFsector->floorpic_angle>>ANGLETOFINESHIFT;
}
else // it's a ceiling
{
scrollx = FIXED_TO_FLOAT(FOFsector->ceiling_xoffs)/fflatsize;
scrolly = FIXED_TO_FLOAT(FOFsector->ceiling_yoffs)/fflatsize;
angle = FOFsector->ceilingpic_angle>>ANGLETOFINESHIFT;
}
}
else if (gr_frontsector)
{
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
if (!isceiling) // it's a floor
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
{
scrollx = FIXED_TO_FLOAT(gr_frontsector->floor_xoffs)/fflatsize;
scrolly = FIXED_TO_FLOAT(gr_frontsector->floor_yoffs)/fflatsize;
angle = gr_frontsector->floorpic_angle>>ANGLETOFINESHIFT;
}
else // it's a ceiling
{
scrollx = FIXED_TO_FLOAT(gr_frontsector->ceiling_xoffs)/fflatsize;
scrolly = FIXED_TO_FLOAT(gr_frontsector->ceiling_yoffs)/fflatsize;
angle = gr_frontsector->ceilingpic_angle>>ANGLETOFINESHIFT;
}
}
if (angle) // Only needs to be done if there's an altered angle
{
// This needs to be done so that it scrolls in a different direction after rotation like software
tempxsow = FLOAT_TO_FIXED(scrollx);
tempytow = FLOAT_TO_FIXED(scrolly);
scrollx = (FIXED_TO_FLOAT(FixedMul(tempxsow, FINECOSINE(angle)) - FixedMul(tempytow, FINESINE(angle))));
scrolly = (FIXED_TO_FLOAT(FixedMul(tempxsow, FINESINE(angle)) + FixedMul(tempytow, FINECOSINE(angle))));
// This needs to be done so everything aligns after rotation
// It would be done so that rotation is done, THEN the translation, but I couldn't get it to rotate AND scroll like software does
tempxsow = FLOAT_TO_FIXED(flatxref);
tempytow = FLOAT_TO_FIXED(flatyref);
flatxref = (FIXED_TO_FLOAT(FixedMul(tempxsow, FINECOSINE(angle)) - FixedMul(tempytow, FINESINE(angle))));
flatyref = (FIXED_TO_FLOAT(FixedMul(tempxsow, FINESINE(angle)) + FixedMul(tempytow, FINECOSINE(angle))));
}
Fix hardware / r_opengl warnings.
2014-11-02 04:15:05 +00:00
for (i = 0; i < (INT32)nrPlaneVerts; i++,v3d++)
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
{
// Hurdler: add scrolling texture on floor/ceiling
v3d->sow = (float)((FIXED_TO_FLOAT(polysector->origVerts[i].x) / fflatsize) - flatxref + scrollx); // Go from the polysector's original vertex locations
v3d->tow = (float)(flatyref - (FIXED_TO_FLOAT(polysector->origVerts[i].y) / fflatsize) + scrolly); // Means the flat is offset based on the original vertex locations
// Need to rotate before translate
if (angle) // Only needs to be done if there's an altered angle
{
tempxsow = FLOAT_TO_FIXED(v3d->sow);
tempytow = FLOAT_TO_FIXED(v3d->tow);
v3d->sow = (FIXED_TO_FLOAT(FixedMul(tempxsow, FINECOSINE(angle)) - FixedMul(tempytow, FINESINE(angle))));
v3d->tow = (FIXED_TO_FLOAT(-FixedMul(tempxsow, FINESINE(angle)) - FixedMul(tempytow, FINECOSINE(angle))));
}
Fix the polyobject crash
2014-11-14 22:19:44 +00:00
v3d->x = FIXED_TO_FLOAT(polysector->vertices[i]->x);
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
v3d->y = height;
Fix the polyobject crash
2014-11-14 22:19:44 +00:00
v3d->z = FIXED_TO_FLOAT(polysector->vertices[i]->y);
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
}
if (planecolormap)
Surf.FlatColor.rgba = HWR_Lighting(lightlevel, planecolormap->rgba, planecolormap->fadergba, false, true);
else
Surf.FlatColor.rgba = HWR_Lighting(lightlevel, NORMALFOG, FADEFOG, false, true);
if (blendmode & PF_Translucent)
{
Surf.FlatColor.s.alpha = (UINT8)alpha;
blendmode |= PF_Modulated|PF_Occlude|PF_Clip;
}
else
blendmode |= PF_Masked|PF_Modulated|PF_Clip;
HWD.pfnDrawPolygon(&Surf, planeVerts, nrPlaneVerts, blendmode);
}
static void HWR_AddPolyObjectPlanes(void)
{
size_t i;
whiteline checkup
2014-11-02 03:52:17 +00:00
sector_t *polyobjsector;
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
// Polyobject Planes need their own function for drawing because they don't have extrasubsectors by themselves
// It should be okay because polyobjects should always be convex anyway
for (i = 0; i < numpolys; i++)
{
polyobjsector = po_ptrs[i]->lines[0]->backsector; // the in-level polyobject sector
if (!(po_ptrs[i]->flags & POF_RENDERPLANES)) // Only render planes when you should
continue;
if (po_ptrs[i]->translucency >= NUMTRANSMAPS)
continue;
if (polyobjsector->floorheight <= gr_frontsector->ceilingheight
&& polyobjsector->floorheight >= gr_frontsector->floorheight
&& (viewz < polyobjsector->floorheight))
{
if (po_ptrs[i]->translucency > 0)
{
FSurfaceInfo Surf;
FBITFIELD blendmode = HWR_TranstableToAlpha(po_ptrs[i]->translucency, &Surf);
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
HWR_AddTransparentPolyobjectFloor(levelflats[polyobjsector->floorpic].lumpnum, po_ptrs[i], false, polyobjsector->floorheight,
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
polyobjsector->lightlevel, Surf.FlatColor.s.alpha, polyobjsector, blendmode, NULL);
}
else
{
HWR_GetFlat(levelflats[polyobjsector->floorpic].lumpnum);
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
HWR_RenderPolyObjectPlane(po_ptrs[i], false, polyobjsector->floorheight, PF_Occlude,
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
polyobjsector->lightlevel, levelflats[polyobjsector->floorpic].lumpnum,
polyobjsector, 255, NULL);
}
}
if (polyobjsector->ceilingheight >= gr_frontsector->floorheight
&& polyobjsector->ceilingheight <= gr_frontsector->ceilingheight
&& (viewz > polyobjsector->ceilingheight))
{
if (po_ptrs[i]->translucency > 0)
{
FSurfaceInfo Surf;
hardware: let not break MSVC support
2016-05-27 05:28:21 +00:00
FBITFIELD blendmode;
hardware: start the surf as clean
2016-05-27 05:19:16 +00:00
memset(&Surf, 0x00, sizeof(Surf));
hardware: let not break MSVC support
2016-05-27 05:28:21 +00:00
blendmode = HWR_TranstableToAlpha(po_ptrs[i]->translucency, &Surf);
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
HWR_AddTransparentPolyobjectFloor(levelflats[polyobjsector->ceilingpic].lumpnum, po_ptrs[i], true, polyobjsector->ceilingheight,
hardware: let not break MSVC support
2016-05-27 05:28:21 +00:00
polyobjsector->lightlevel, Surf.FlatColor.s.alpha, polyobjsector, blendmode, NULL);
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
}
else
{
HWR_GetFlat(levelflats[polyobjsector->ceilingpic].lumpnum);
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
HWR_RenderPolyObjectPlane(po_ptrs[i], true, polyobjsector->ceilingheight, PF_Occlude,
hardware: let not break MSVC support
2016-05-27 05:28:21 +00:00
polyobjsector->lightlevel, levelflats[polyobjsector->floorpic].lumpnum,
polyobjsector, 255, NULL);
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
}
}
}
}
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
#endif
// -----------------+
// HWR_Subsector : Determine floor/ceiling planes.
// : Add sprites of things in sector.
// : Draw one or more line segments.
// Notes : Sets gr_cursectorlight to the light of the parent sector, to modulate wall textures
// -----------------+
static void HWR_Subsector(size_t num)
{
INT16 count;
seg_t *line;
subsector_t *sub;
static tempsec for R_FakeFlat I don't fully understand this, but it's what software does and it fixes the issue of the lighting in DSZ3. Also don't need the extra call to R_Prep3DFloors.
2018-03-29 22:28:54 +00:00
static sector_t tempsec; //SoM: 4/7/2000
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
INT32 floorlightlevel;
INT32 ceilinglightlevel;
INT32 locFloorHeight, locCeilingHeight;
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
INT32 cullFloorHeight, cullCeilingHeight;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
INT32 light = 0;
extracolormap_t *floorcolormap;
extracolormap_t *ceilingcolormap;
#ifdef PARANOIA //no risk while developing, enough debugging nights!
if (num >= addsubsector)
I_Error("HWR_Subsector: ss %s with numss = %s, addss = %s\n",
sizeu1(num), sizeu2(numsubsectors), sizeu3(addsubsector));
/*if (num >= numsubsectors)
I_Error("HWR_Subsector: ss %i with numss = %i",
num,
numsubsectors);*/
#endif
if (num < numsubsectors)
{
// subsector
sub = &subsectors[num];
// sector
gr_frontsector = sub->sector;
// how many linedefs
count = sub->numlines;
// first line seg
line = &segs[sub->firstline];
}
else
{
// there are no segs but only planes
sub = &subsectors[0];
gr_frontsector = sub->sector;
count = 0;
line = NULL;
}
//SoM: 4/7/2000: Test to make Boom water work in Hardware mode.
gr_frontsector = R_FakeFlat(gr_frontsector, &tempsec, &floorlightlevel,
&ceilinglightlevel, false);
//FIXME: Use floorlightlevel and ceilinglightlevel insted of lightlevel.
floorcolormap = ceilingcolormap = gr_frontsector->extra_colormap;
// ------------------------------------------------------------------------
// sector lighting, DISABLED because it's done in HWR_StoreWallRange
// ------------------------------------------------------------------------
/// \todo store a RGBA instead of just intensity, allow coloured sector lighting
//light = (FUBYTE)(sub->sector->lightlevel & 0xFF) / 255.0f;
//gr_cursectorlight.red = light;
//gr_cursectorlight.green = light;
//gr_cursectorlight.blue = light;
//gr_cursectorlight.alpha = light;
// ----- for special tricks with HW renderer -----
if (gr_frontsector->pseudoSector)
{
GL slope walls and fixed plane culling
2016-01-24 08:41:30 +00:00
cullFloorHeight = locFloorHeight = gr_frontsector->virtualFloorheight;
cullCeilingHeight = locCeilingHeight = gr_frontsector->virtualCeilingheight;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
else if (gr_frontsector->virtualFloor)
{
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
///@TODO Is this whole virtualFloor mess even useful? I don't think it even triggers ever.
GL slope walls and fixed plane culling
2016-01-24 08:41:30 +00:00
cullFloorHeight = locFloorHeight = gr_frontsector->virtualFloorheight;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (gr_frontsector->virtualCeiling)
GL slope walls and fixed plane culling
2016-01-24 08:41:30 +00:00
cullCeilingHeight = locCeilingHeight = gr_frontsector->virtualCeilingheight;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
else
GL slope walls and fixed plane culling
2016-01-24 08:41:30 +00:00
cullCeilingHeight = locCeilingHeight = gr_frontsector->ceilingheight;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
else if (gr_frontsector->virtualCeiling)
{
GL slope walls and fixed plane culling
2016-01-24 08:41:30 +00:00
cullCeilingHeight = locCeilingHeight = gr_frontsector->virtualCeilingheight;
cullFloorHeight = locFloorHeight = gr_frontsector->floorheight;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
else
{
GL slope walls and fixed plane culling
2016-01-24 08:41:30 +00:00
cullFloorHeight = locFloorHeight = gr_frontsector->floorheight;
cullCeilingHeight = locCeilingHeight = gr_frontsector->ceilingheight;
#ifdef ESLOPE
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
if (gr_frontsector->f_slope)
{
cullFloorHeight = P_GetZAt(gr_frontsector->f_slope, viewx, viewy);
locFloorHeight = P_GetZAt(gr_frontsector->f_slope, gr_frontsector->soundorg.x, gr_frontsector->soundorg.y);
}
GL slope walls and fixed plane culling
2016-01-24 08:41:30 +00:00
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
if (gr_frontsector->c_slope)
{
cullCeilingHeight = P_GetZAt(gr_frontsector->c_slope, viewx, viewy);
locCeilingHeight = P_GetZAt(gr_frontsector->c_slope, gr_frontsector->soundorg.x, gr_frontsector->soundorg.y);
}
GL slope walls and fixed plane culling
2016-01-24 08:41:30 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
// ----- end special tricks -----
if (gr_frontsector->ffloors)
{
if (gr_frontsector->moved)
{
gr_frontsector->numlights = sub->sector->numlights = 0;
R_Prep3DFloors(gr_frontsector);
sub->sector->lightlist = gr_frontsector->lightlist;
sub->sector->numlights = gr_frontsector->numlights;
sub->sector->moved = gr_frontsector->moved = false;
}
light = R_GetPlaneLight(gr_frontsector, locFloorHeight, false);
if (gr_frontsector->floorlightsec == -1)
floorlightlevel = *gr_frontsector->lightlist[light].lightlevel;
floorcolormap = gr_frontsector->lightlist[light].extra_colormap;
light = R_GetPlaneLight(gr_frontsector, locCeilingHeight, false);
if (gr_frontsector->ceilinglightsec == -1)
ceilinglightlevel = *gr_frontsector->lightlist[light].lightlevel;
ceilingcolormap = gr_frontsector->lightlist[light].extra_colormap;
}
sub->sector->extra_colormap = gr_frontsector->extra_colormap;
// render floor ?
#ifdef DOPLANES
// yeah, easy backface cull! :)
GL slope walls and fixed plane culling
2016-01-24 08:41:30 +00:00
if (cullFloorHeight < dup_viewz)
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
if (gr_frontsector->floorpic != skyflatnum)
{
if (sub->validcount != validcount)
{
HWR_GetFlat(levelflats[gr_frontsector->floorpic].lumpnum);
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
HWR_RenderPlane(gr_frontsector, &extrasubsectors[num], false,
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
// Hack to make things continue to work around slopes.
locFloorHeight == cullFloorHeight ? locFloorHeight : gr_frontsector->floorheight,
// We now return you to your regularly scheduled rendering.
PF_Occlude, floorlightlevel, levelflats[gr_frontsector->floorpic].lumpnum, NULL, 255, false, floorcolormap);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
}
else
{
#ifdef POLYSKY
HWR_RenderSkyPlane(&extrasubsectors[num], locFloorHeight);
#endif
}
}
GL slope walls and fixed plane culling
2016-01-24 08:41:30 +00:00
if (cullCeilingHeight > dup_viewz)
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
if (gr_frontsector->ceilingpic != skyflatnum)
{
if (sub->validcount != validcount)
{
HWR_GetFlat(levelflats[gr_frontsector->ceilingpic].lumpnum);
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
HWR_RenderPlane(NULL, &extrasubsectors[num], true,
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
// Hack to make things continue to work around slopes.
locCeilingHeight == cullCeilingHeight ? locCeilingHeight : gr_frontsector->ceilingheight,
// We now return you to your regularly scheduled rendering.
PF_Occlude, ceilinglightlevel, levelflats[gr_frontsector->ceilingpic].lumpnum,NULL, 255, false, ceilingcolormap);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
}
else
{
#ifdef POLYSKY
HWR_RenderSkyPlane(&extrasubsectors[num], locCeilingHeight);
#endif
}
}
#ifndef POLYSKY
// Moved here because before, when above the ceiling and the floor does not have the sky flat, it doesn't draw the sky
if (gr_frontsector->ceilingpic == skyflatnum || gr_frontsector->floorpic == skyflatnum)
drawsky = true;
#endif
#ifdef R_FAKEFLOORS
if (gr_frontsector->ffloors)
{
/// \todo fix light, xoffs, yoffs, extracolormap ?
ffloor_t * rover;
for (rover = gr_frontsector->ffloors;
rover; rover = rover->next)
{
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
fixed_t cullHeight, centerHeight;
Fix FOF slope rendering in ogl should work as well as software if not better now
2016-01-25 05:57:53 +00:00
// bottom plane
#ifdef ESLOPE
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
if (*rover->b_slope)
{
cullHeight = P_GetZAt(*rover->b_slope, viewx, viewy);
centerHeight = P_GetZAt(*rover->b_slope, gr_frontsector->soundorg.x, gr_frontsector->soundorg.y);
}
else
Fix FOF slope rendering in ogl should work as well as software if not better now
2016-01-25 05:57:53 +00:00
#endif
cullHeight = centerHeight = *rover->bottomheight;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (!(rover->flags & FF_EXISTS) || !(rover->flags & FF_RENDERPLANES))
continue;
if (sub->validcount == validcount)
continue;
Fix FOF slope rendering in ogl should work as well as software if not better now
2016-01-25 05:57:53 +00:00
if (centerHeight <= locCeilingHeight &&
centerHeight >= locFloorHeight &&
((dup_viewz < cullHeight && !(rover->flags & FF_INVERTPLANES)) ||
(dup_viewz > cullHeight && (rover->flags & FF_BOTHPLANES || rover->flags & FF_INVERTPLANES))))
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
if (rover->flags & FF_FOG)
{
UINT8 alpha;
Fix FOF slope rendering in ogl should work as well as software if not better now
2016-01-25 05:57:53 +00:00
light = R_GetPlaneLight(gr_frontsector, centerHeight, dup_viewz < cullHeight ? true : false);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (rover->master->frontsector->extra_colormap)
Better fog block colouring They still aren't perfect, but now they are at least not quite so obviously just translucent polygons over the level. A mixture between partially modulating the background colours and adding the fog colour. Notably white fog blocks look like they're brightening what's behind them. Additive was also setting noalphatest before, can probably decide that depending on what it needs anyway. I don't think it's currently used anyway.
2018-03-15 23:58:37 +00:00
alpha = HWR_FogBlockAlpha(*gr_frontsector->lightlist[light].lightlevel, rover->master->frontsector->extra_colormap->rgba);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
else
Better fog block colouring They still aren't perfect, but now they are at least not quite so obviously just translucent polygons over the level. A mixture between partially modulating the background colours and adding the fog colour. Notably white fog blocks look like they're brightening what's behind them. Additive was also setting noalphatest before, can probably decide that depending on what it needs anyway. I don't think it's currently used anyway.
2018-03-15 23:58:37 +00:00
alpha = HWR_FogBlockAlpha(*gr_frontsector->lightlist[light].lightlevel, NORMALFOG);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
HWR_AddTransparentFloor(0,
&extrasubsectors[num],
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
false,
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
*rover->bottomheight,
*gr_frontsector->lightlist[light].lightlevel,
Better fog block colouring They still aren't perfect, but now they are at least not quite so obviously just translucent polygons over the level. A mixture between partially modulating the background colours and adding the fog colour. Notably white fog blocks look like they're brightening what's behind them. Additive was also setting noalphatest before, can probably decide that depending on what it needs anyway. I don't think it's currently used anyway.
2018-03-15 23:58:37 +00:00
alpha, rover->master->frontsector, PF_Fog|PF_NoTexture,
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
true, rover->master->frontsector->extra_colormap);
}
Various Transparent Texture Fixes Draw Textures and Flats that have holes in them like a solid polygon so they use the depth buffer and don't need to be sorted Disable all linear filtering on textures and flats that have holes in them, the linear filtering introduces translucency into the textures where the edges are. Leaving them with either a black border, or causing pixels behind the slightly translucent areas to not be drawn. Doesn't apply to sprites and the HUD as they are always already sorted properly. Make the Alpha Testing more strict on non-translucent blend modes. This makes it so any transparency below 0.5 is discarded instead. Would make anything that is blended and has holes in it look slightly better, only the HUD and MD2s where the texture has holes are effected currently. Set TF_TRANSPARENT on flat texture flags when there are holes in the texture. Minor fix to make sure MD2s always set the right blend mode
2017-09-08 23:41:11 +00:00
else if (rover->flags & FF_TRANSLUCENT && rover->alpha < 256) // SoM: Flags are more efficient
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
Fix FOF slope rendering in ogl should work as well as software if not better now
2016-01-25 05:57:53 +00:00
light = R_GetPlaneLight(gr_frontsector, centerHeight, dup_viewz < cullHeight ? true : false);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
#ifndef SORTING
HWR_Add3DWater(levelflats[*rover->bottompic].lumpnum,
&extrasubsectors[num],
*rover->bottomheight,
*gr_frontsector->lightlist[light].lightlevel,
rover->alpha-1, rover->master->frontsector);
#else
HWR_AddTransparentFloor(levelflats[*rover->bottompic].lumpnum,
&extrasubsectors[num],
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
false,
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
*rover->bottomheight,
*gr_frontsector->lightlist[light].lightlevel,
rover->alpha-1 > 255 ? 255 : rover->alpha-1, rover->master->frontsector, PF_Translucent,
false, gr_frontsector->lightlist[light].extra_colormap);
#endif
}
else
{
HWR_GetFlat(levelflats[*rover->bottompic].lumpnum);
Fix FOF slope rendering in ogl should work as well as software if not better now
2016-01-25 05:57:53 +00:00
light = R_GetPlaneLight(gr_frontsector, centerHeight, dup_viewz < cullHeight ? true : false);
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
HWR_RenderPlane(NULL, &extrasubsectors[num], false, *rover->bottomheight, PF_Occlude, *gr_frontsector->lightlist[light].lightlevel, levelflats[*rover->bottompic].lumpnum,
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
rover->master->frontsector, 255, false, gr_frontsector->lightlist[light].extra_colormap);
}
}
Fix FOF slope rendering in ogl should work as well as software if not better now
2016-01-25 05:57:53 +00:00
// top plane
#ifdef ESLOPE
Fixed compiler errors as well as a ton of tab spaces (apparently it's a common problem in this file anyway but whatever)
2016-01-28 13:56:23 +00:00
if (*rover->t_slope)
{
cullHeight = P_GetZAt(*rover->t_slope, viewx, viewy);
centerHeight = P_GetZAt(*rover->t_slope, gr_frontsector->soundorg.x, gr_frontsector->soundorg.y);
}
else
Fix FOF slope rendering in ogl should work as well as software if not better now
2016-01-25 05:57:53 +00:00
#endif
cullHeight = centerHeight = *rover->topheight;
if (centerHeight >= locFloorHeight &&
centerHeight <= locCeilingHeight &&
((dup_viewz > cullHeight && !(rover->flags & FF_INVERTPLANES)) ||
(dup_viewz < cullHeight && (rover->flags & FF_BOTHPLANES || rover->flags & FF_INVERTPLANES))))
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
if (rover->flags & FF_FOG)
{
UINT8 alpha;
Fix FOF slope rendering in ogl should work as well as software if not better now
2016-01-25 05:57:53 +00:00
light = R_GetPlaneLight(gr_frontsector, centerHeight, dup_viewz < cullHeight ? true : false);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (rover->master->frontsector->extra_colormap)
Better fog block colouring They still aren't perfect, but now they are at least not quite so obviously just translucent polygons over the level. A mixture between partially modulating the background colours and adding the fog colour. Notably white fog blocks look like they're brightening what's behind them. Additive was also setting noalphatest before, can probably decide that depending on what it needs anyway. I don't think it's currently used anyway.
2018-03-15 23:58:37 +00:00
alpha = HWR_FogBlockAlpha(*gr_frontsector->lightlist[light].lightlevel, rover->master->frontsector->extra_colormap->rgba);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
else
Better fog block colouring They still aren't perfect, but now they are at least not quite so obviously just translucent polygons over the level. A mixture between partially modulating the background colours and adding the fog colour. Notably white fog blocks look like they're brightening what's behind them. Additive was also setting noalphatest before, can probably decide that depending on what it needs anyway. I don't think it's currently used anyway.
2018-03-15 23:58:37 +00:00
alpha = HWR_FogBlockAlpha(*gr_frontsector->lightlist[light].lightlevel, NORMALFOG);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
HWR_AddTransparentFloor(0,
&extrasubsectors[num],
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
true,
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
*rover->topheight,
*gr_frontsector->lightlist[light].lightlevel,
Better fog block colouring They still aren't perfect, but now they are at least not quite so obviously just translucent polygons over the level. A mixture between partially modulating the background colours and adding the fog colour. Notably white fog blocks look like they're brightening what's behind them. Additive was also setting noalphatest before, can probably decide that depending on what it needs anyway. I don't think it's currently used anyway.
2018-03-15 23:58:37 +00:00
alpha, rover->master->frontsector, PF_Fog|PF_NoTexture,
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
true, rover->master->frontsector->extra_colormap);
}
Various Transparent Texture Fixes Draw Textures and Flats that have holes in them like a solid polygon so they use the depth buffer and don't need to be sorted Disable all linear filtering on textures and flats that have holes in them, the linear filtering introduces translucency into the textures where the edges are. Leaving them with either a black border, or causing pixels behind the slightly translucent areas to not be drawn. Doesn't apply to sprites and the HUD as they are always already sorted properly. Make the Alpha Testing more strict on non-translucent blend modes. This makes it so any transparency below 0.5 is discarded instead. Would make anything that is blended and has holes in it look slightly better, only the HUD and MD2s where the texture has holes are effected currently. Set TF_TRANSPARENT on flat texture flags when there are holes in the texture. Minor fix to make sure MD2s always set the right blend mode
2017-09-08 23:41:11 +00:00
else if (rover->flags & FF_TRANSLUCENT && rover->alpha < 256)
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
Fix FOF slope rendering in ogl should work as well as software if not better now
2016-01-25 05:57:53 +00:00
light = R_GetPlaneLight(gr_frontsector, centerHeight, dup_viewz < cullHeight ? true : false);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
#ifndef SORTING
HWR_Add3DWater(levelflats[*rover->toppic].lumpnum,
&extrasubsectors[num],
*rover->topheight,
*gr_frontsector->lightlist[light].lightlevel,
rover->alpha-1, rover->master->frontsector);
#else
HWR_AddTransparentFloor(levelflats[*rover->toppic].lumpnum,
&extrasubsectors[num],
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
true,
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
*rover->topheight,
*gr_frontsector->lightlist[light].lightlevel,
rover->alpha-1 > 255 ? 255 : rover->alpha-1, rover->master->frontsector, PF_Translucent,
false, gr_frontsector->lightlist[light].extra_colormap);
#endif
}
else
{
HWR_GetFlat(levelflats[*rover->toppic].lumpnum);
Fix FOF slope rendering in ogl should work as well as software if not better now
2016-01-25 05:57:53 +00:00
light = R_GetPlaneLight(gr_frontsector, centerHeight, dup_viewz < cullHeight ? true : false);
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
HWR_RenderPlane(NULL, &extrasubsectors[num], true, *rover->topheight, PF_Occlude, *gr_frontsector->lightlist[light].lightlevel, levelflats[*rover->toppic].lumpnum,
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
rover->master->frontsector, 255, false, gr_frontsector->lightlist[light].extra_colormap);
}
}
}
}
#endif
#endif //doplanes
#ifdef POLYOBJECTS
// Draw all the polyobjects in this subsector
if (sub->polyList)
{
polyobj_t *po = sub->polyList;
numpolys = 0;
// Count all the polyobjects, reset the list, and recount them
while (po)
{
++numpolys;
po = (polyobj_t *)(po->link.next);
}
// Sort polyobjects
R_SortPolyObjects(sub);
// Draw polyobject lines.
HWR_AddPolyObjectSegs();
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
#ifdef POLYOBJECTS_PLANES
if (sub->validcount != validcount) // This validcount situation seems to let us know that the floors have already been drawn.
{
// Draw polyobject planes
HWR_AddPolyObjectPlanes();
}
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
#endif
// Hurder ici se passe les choses INT32<33>essantes!
// on vient de tracer le sol et le plafond
// on trace <20>pr<70>ent d'abord les sprites et ensuite les murs
// hurdler: faux: on ajoute seulement les sprites, le murs sont trac<61> d'abord
if (line)
{
// draw sprites first, coz they are clipped to the solidsegs of
// subsectors more 'in front'
HWR_AddSprites(gr_frontsector);
//Hurdler: at this point validcount must be the same, but is not because
// gr_frontsector doesn't point anymore to sub->sector due to
// the call gr_frontsector = R_FakeFlat(...)
// if it's not done, the sprite is drawn more than once,
// what looks really bad with translucency or dynamic light,
// without talking about the overdraw of course.
sub->sector->validcount = validcount;/// \todo fix that in a better way
while (count--)
{
Do not draw segs that belong to polyobjects, if you're drawing subsectors adjacent to them Polyobject segs should ONLY be drawn if the polyobject itself is in the polylist of a subsector being rendered. That way you won't sometimes see polyobject walls through level boundaries, if you happen to be close enough to their pre-spawn locations outside the level (or in them, if you decided to go on a noclip journey).
2017-06-03 16:47:46 +00:00
#ifdef POLYOBJECTS
if (!line->polyseg) // ignore segs that belong to polyobjects
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
HWR_AddLine(line);
line++;
}
}
sub->validcount = validcount;
}
//
// Renders all subsectors below a given node,
// traversing subtree recursively.
// Just call with BSP root.
#ifdef coolhack
//t;b;l;r
static fixed_t hackbbox[4];
//BOXTOP,
//BOXBOTTOM,
//BOXLEFT,
//BOXRIGHT
static boolean HWR_CheckHackBBox(fixed_t *bb)
{
if (bb[BOXTOP] < hackbbox[BOXBOTTOM]) //y up
return false;
if (bb[BOXBOTTOM] > hackbbox[BOXTOP])
return false;
if (bb[BOXLEFT] > hackbbox[BOXRIGHT])
return false;
if (bb[BOXRIGHT] < hackbbox[BOXLEFT])
return false;
return true;
}
#endif
// BP: big hack for a test in lighning ref : 1249753487AB
fixed_t *hwbbox;
static void HWR_RenderBSPNode(INT32 bspnum)
{
/*//GZDoom code
if(bspnum == -1)
{
HWR_Subsector(subsectors);
return;
}
while(!((size_t)bspnum&(~NF_SUBSECTOR))) // Keep going until found a subsector
{
node_t *bsp = &nodes[bspnum];
// Decide which side the view point is on
INT32 side = R_PointOnSide(dup_viewx, dup_viewy, bsp);
// Recursively divide front space (toward the viewer)
HWR_RenderBSPNode(bsp->children[side]);
// Possibly divide back space (away from viewer)
side ^= 1;
if (!HWR_CheckBBox(bsp->bbox[side]))
return;
bspnum = bsp->children[side];
}
HWR_Subsector(bspnum-1);
*/
node_t *bsp = &nodes[bspnum];
// Decide which side the view point is on
INT32 side;
// Found a subsector?
if (bspnum & NF_SUBSECTOR)
{
if (bspnum == -1)
{
//*(gr_drawsubsector_p++) = 0;
HWR_Subsector(0);
}
else
{
//*(gr_drawsubsector_p++) = bspnum&(~NF_SUBSECTOR);
HWR_Subsector(bspnum&(~NF_SUBSECTOR));
}
return;
}
// Decide which side the view point is on.
side = R_PointOnSide(dup_viewx, dup_viewy, bsp);
// BP: big hack for a test in lighning ref : 1249753487AB
hwbbox = bsp->bbox[side];
// Recursively divide front space.
HWR_RenderBSPNode(bsp->children[side]);
// Possibly divide back space.
if (HWR_CheckBBox(bsp->bbox[side^1]))
{
// BP: big hack for a test in lighning ref : 1249753487AB
hwbbox = bsp->bbox[side^1];
HWR_RenderBSPNode(bsp->children[side^1]);
}
}
/*
//
// Clear 'stack' of subsectors to draw
//
static void HWR_ClearDrawSubsectors(void)
{
gr_drawsubsector_p = gr_drawsubsectors;
}
//
// Draw subsectors pushed on the drawsubsectors 'stack', back to front
//
static void HWR_RenderSubsectors(void)
{
while (gr_drawsubsector_p > gr_drawsubsectors)
{
HWR_RenderBSPNode(
lastsubsec->nextsubsec = bspnum & (~NF_SUBSECTOR);
}
}
*/
// ==========================================================================
// FROM R_MAIN.C
// ==========================================================================
//BP : exactely the same as R_InitTextureMapping
void HWR_InitTextureMapping(void)
{
angle_t i;
INT32 x;
INT32 t;
fixed_t focallength;
fixed_t grcenterx;
fixed_t grcenterxfrac;
INT32 grviewwidth;
#define clipanglefov (FIELDOFVIEW>>ANGLETOFINESHIFT)
grviewwidth = vid.width;
grcenterx = grviewwidth/2;
grcenterxfrac = grcenterx<<FRACBITS;
// Use tangent table to generate viewangletox:
// viewangletox will give the next greatest x
// after the view angle.
//
// Calc focallength
// so FIELDOFVIEW angles covers SCREENWIDTH.
focallength = FixedDiv(grcenterxfrac,
FINETANGENT(FINEANGLES/4+clipanglefov/2));
for (i = 0; i < FINEANGLES/2; i++)
{
if (FINETANGENT(i) > FRACUNIT*2)
t = -1;
else if (FINETANGENT(i) < -FRACUNIT*2)
t = grviewwidth+1;
else
{
t = FixedMul(FINETANGENT(i), focallength);
t = (grcenterxfrac - t+FRACUNIT-1)>>FRACBITS;
if (t < -1)
t = -1;
else if (t > grviewwidth+1)
t = grviewwidth+1;
}
gr_viewangletox[i] = t;
}
// Scan viewangletox[] to generate xtoviewangle[]:
// xtoviewangle will give the smallest view angle
// that maps to x.
for (x = 0; x <= grviewwidth; x++)
{
i = 0;
while (gr_viewangletox[i]>x)
i++;
gr_xtoviewangle[x] = (i<<ANGLETOFINESHIFT) - ANGLE_90;
}
// Take out the fencepost cases from viewangletox.
for (i = 0; i < FINEANGLES/2; i++)
{
if (gr_viewangletox[i] == -1)
gr_viewangletox[i] = 0;
else if (gr_viewangletox[i] == grviewwidth+1)
gr_viewangletox[i] = grviewwidth;
}
gr_clipangle = gr_xtoviewangle[0];
}
// ==========================================================================
// gr_things.c
// ==========================================================================
// sprites are drawn after all wall and planes are rendered, so that
// sprite translucency effects apply on the rendered view (instead of the background sky!!)
static UINT32 gr_visspritecount;
static gr_vissprite_t *gr_visspritechunks[MAXVISSPRITES >> VISSPRITECHUNKBITS] = {NULL};
// --------------------------------------------------------------------------
// HWR_ClearSprites
// Called at frame start.
// --------------------------------------------------------------------------
static void HWR_ClearSprites(void)
{
gr_visspritecount = 0;
}
// --------------------------------------------------------------------------
// HWR_NewVisSprite
// --------------------------------------------------------------------------
static gr_vissprite_t gr_overflowsprite;
static gr_vissprite_t *HWR_GetVisSprite(UINT32 num)
{
UINT32 chunk = num >> VISSPRITECHUNKBITS;
// Allocate chunk if necessary
if (!gr_visspritechunks[chunk])
Z_Malloc(sizeof(gr_vissprite_t) * VISSPRITESPERCHUNK, PU_LEVEL, &gr_visspritechunks[chunk]);
return gr_visspritechunks[chunk] + (num & VISSPRITEINDEXMASK);
}
static gr_vissprite_t *HWR_NewVisSprite(void)
{
if (gr_visspritecount == MAXVISSPRITES)
return &gr_overflowsprite;
return HWR_GetVisSprite(gr_visspritecount++);
}
// Finds a floor through which light does not pass.
static fixed_t HWR_OpaqueFloorAtPos(fixed_t x, fixed_t y, fixed_t z, fixed_t height)
{
const sector_t *sec = R_PointInSubsector(x, y)->sector;
fixed_t floorz = sec->floorheight;
if (sec->ffloors)
{
ffloor_t *rover;
fixed_t delta1, delta2;
const fixed_t thingtop = z + height;
for (rover = sec->ffloors; rover; rover = rover->next)
{
if (!(rover->flags & FF_EXISTS)
|| !(rover->flags & FF_RENDERPLANES)
|| rover->flags & FF_TRANSLUCENT
|| rover->flags & FF_FOG
|| rover->flags & FF_INVERTPLANES)
continue;
delta1 = z - (*rover->bottomheight + ((*rover->topheight - *rover->bottomheight)/2));
delta2 = thingtop - (*rover->bottomheight + ((*rover->topheight - *rover->bottomheight)/2));
if (*rover->topheight > floorz && abs(delta1) < abs(delta2))
floorz = *rover->topheight;
}
}
return floorz;
}
SRB2 2.1.12 release
2014-11-12 00:55:07 +00:00
//
// HWR_DoCulling
// Hardware version of R_DoCulling
// (see r_main.c)
static boolean HWR_DoCulling(line_t *cullheight, line_t *viewcullheight, float vz, float bottomh, float toph)
{
float cullplane;
if (!cullheight)
return false;
cullplane = FIXED_TO_FLOAT(cullheight->frontsector->floorheight);
if (cullheight->flags & ML_NOCLIMB) // Group culling
{
if (!viewcullheight)
return false;
// Make sure this is part of the same group
if (viewcullheight->frontsector == cullheight->frontsector)
{
// OK, we can cull
if (vz > cullplane && toph < cullplane) // Cull if below plane
return true;
if (bottomh > cullplane && vz <= cullplane) // Cull if above plane
return true;
}
}
else // Quick culling
{
if (vz > cullplane && toph < cullplane) // Cull if below plane
return true;
if (bottomh > cullplane && vz <= cullplane) // Cull if above plane
return true;
}
return false;
}
*Fixed OpenGL's handling of cv_translucency effects (should not remove shadows if off, but perhaps make them use default alpha?), someone was a bit hasty! *De-stupified the MD2 status checks regarding drawing of sprites - if they're already checked before calling HWR_DrawSprite, there's no point doing them WITHIN the function as well *Split off sprite shadow code into HWR_DrawSpriteShadow for convenience git-svn-id: https://code.orospakr.ca/svn/srb2/trunk@8983 6de4a73c-47e2-0310-b8c1-93d6ecd3f8cd
2015-01-02 15:14:22 +00:00
static void HWR_DrawSpriteShadow(gr_vissprite_t *spr, GLPatch_t *gpatch, float this_scale)
{
FOutVector swallVerts[4];
FSurfaceInfo sSurf;
fixed_t floorheight, mobjfloor;
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
float offset = 0;
*Fixed OpenGL's handling of cv_translucency effects (should not remove shadows if off, but perhaps make them use default alpha?), someone was a bit hasty! *De-stupified the MD2 status checks regarding drawing of sprites - if they're already checked before calling HWR_DrawSprite, there's no point doing them WITHIN the function as well *Split off sprite shadow code into HWR_DrawSpriteShadow for convenience git-svn-id: https://code.orospakr.ca/svn/srb2/trunk@8983 6de4a73c-47e2-0310-b8c1-93d6ecd3f8cd
2015-01-02 15:14:22 +00:00
mobjfloor = HWR_OpaqueFloorAtPos(
spr->mobj->x, spr->mobj->y,
spr->mobj->z, spr->mobj->height);
if (cv_shadowoffs.value)
{
angle_t shadowdir;
// Set direction
Fix HWR_ProjectSprite to check properly whether the displayed player's mobj or its subsector exists, to avoid a crash when checking for fake planes. (also use viewplayer since its available to use, silly hardware code) Also tweaked a weird splitscreen check in HWR_DrawSpriteShadow; still investigating whether stplyr is ever not player 2 when it's player 2's view, but this looks better for now
2018-06-23 17:47:32 +00:00
if (splitscreen && stplyr == &players[secondarydisplayplayer])
*Fixed OpenGL's handling of cv_translucency effects (should not remove shadows if off, but perhaps make them use default alpha?), someone was a bit hasty! *De-stupified the MD2 status checks regarding drawing of sprites - if they're already checked before calling HWR_DrawSprite, there's no point doing them WITHIN the function as well *Split off sprite shadow code into HWR_DrawSpriteShadow for convenience git-svn-id: https://code.orospakr.ca/svn/srb2/trunk@8983 6de4a73c-47e2-0310-b8c1-93d6ecd3f8cd
2015-01-02 15:14:22 +00:00
shadowdir = localangle2 + FixedAngle(cv_cam2_rotate.value);
else
shadowdir = localangle + FixedAngle(cv_cam_rotate.value);
// Find floorheight
floorheight = HWR_OpaqueFloorAtPos(
spr->mobj->x + P_ReturnThrustX(spr->mobj, shadowdir, spr->mobj->z - mobjfloor),
spr->mobj->y + P_ReturnThrustY(spr->mobj, shadowdir, spr->mobj->z - mobjfloor),
spr->mobj->z, spr->mobj->height);
// The shadow is falling ABOVE it's mobj?
// Don't draw it, then!
if (spr->mobj->z < floorheight)
return;
else
{
fixed_t floorz;
floorz = HWR_OpaqueFloorAtPos(
spr->mobj->x + P_ReturnThrustX(spr->mobj, shadowdir, spr->mobj->z - floorheight),
spr->mobj->y + P_ReturnThrustY(spr->mobj, shadowdir, spr->mobj->z - floorheight),
spr->mobj->z, spr->mobj->height);
// The shadow would be falling on a wall? Don't draw it, then.
// Would draw midair otherwise.
if (floorz < floorheight)
return;
}
floorheight = FixedInt(spr->mobj->z - floorheight);
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
offset = floorheight;
*Fixed OpenGL's handling of cv_translucency effects (should not remove shadows if off, but perhaps make them use default alpha?), someone was a bit hasty! *De-stupified the MD2 status checks regarding drawing of sprites - if they're already checked before calling HWR_DrawSprite, there's no point doing them WITHIN the function as well *Split off sprite shadow code into HWR_DrawSpriteShadow for convenience git-svn-id: https://code.orospakr.ca/svn/srb2/trunk@8983 6de4a73c-47e2-0310-b8c1-93d6ecd3f8cd
2015-01-02 15:14:22 +00:00
}
else
floorheight = FixedInt(spr->mobj->z - mobjfloor);
// create the sprite billboard
//
// 3--2
// | /|
// |/ |
// 0--1
// x1/x2 were already scaled in HWR_ProjectSprite
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
// First match the normal sprite
*Fixed OpenGL's handling of cv_translucency effects (should not remove shadows if off, but perhaps make them use default alpha?), someone was a bit hasty! *De-stupified the MD2 status checks regarding drawing of sprites - if they're already checked before calling HWR_DrawSprite, there's no point doing them WITHIN the function as well *Split off sprite shadow code into HWR_DrawSpriteShadow for convenience git-svn-id: https://code.orospakr.ca/svn/srb2/trunk@8983 6de4a73c-47e2-0310-b8c1-93d6ecd3f8cd
2015-01-02 15:14:22 +00:00
swallVerts[0].x = swallVerts[3].x = spr->x1;
swallVerts[2].x = swallVerts[1].x = spr->x2;
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
swallVerts[0].z = swallVerts[3].z = spr->z1;
swallVerts[2].z = swallVerts[1].z = spr->z2;
*Fixed OpenGL's handling of cv_translucency effects (should not remove shadows if off, but perhaps make them use default alpha?), someone was a bit hasty! *De-stupified the MD2 status checks regarding drawing of sprites - if they're already checked before calling HWR_DrawSprite, there's no point doing them WITHIN the function as well *Split off sprite shadow code into HWR_DrawSpriteShadow for convenience git-svn-id: https://code.orospakr.ca/svn/srb2/trunk@8983 6de4a73c-47e2-0310-b8c1-93d6ecd3f8cd
2015-01-02 15:14:22 +00:00
if (spr->mobj && this_scale != 1.0f)
{
// Always a pixel above the floor, perfectly flat.
swallVerts[0].y = swallVerts[1].y = swallVerts[2].y = swallVerts[3].y = spr->ty - gpatch->topoffset * this_scale - (floorheight+3);
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
// Now transform the TOP vertices along the floor in the direction of the camera
swallVerts[3].x = spr->x1 + ((gpatch->height * this_scale) + offset) * gr_viewcos;
swallVerts[2].x = spr->x2 + ((gpatch->height * this_scale) + offset) * gr_viewcos;
swallVerts[3].z = spr->z1 + ((gpatch->height * this_scale) + offset) * gr_viewsin;
swallVerts[2].z = spr->z2 + ((gpatch->height * this_scale) + offset) * gr_viewsin;
*Fixed OpenGL's handling of cv_translucency effects (should not remove shadows if off, but perhaps make them use default alpha?), someone was a bit hasty! *De-stupified the MD2 status checks regarding drawing of sprites - if they're already checked before calling HWR_DrawSprite, there's no point doing them WITHIN the function as well *Split off sprite shadow code into HWR_DrawSpriteShadow for convenience git-svn-id: https://code.orospakr.ca/svn/srb2/trunk@8983 6de4a73c-47e2-0310-b8c1-93d6ecd3f8cd
2015-01-02 15:14:22 +00:00
}
else
{
// Always a pixel above the floor, perfectly flat.
swallVerts[0].y = swallVerts[1].y = swallVerts[2].y = swallVerts[3].y = spr->ty - gpatch->topoffset - (floorheight+3);
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
// Now transform the TOP vertices along the floor in the direction of the camera
swallVerts[3].x = spr->x1 + (gpatch->height + offset) * gr_viewcos;
swallVerts[2].x = spr->x2 + (gpatch->height + offset) * gr_viewcos;
swallVerts[3].z = spr->z1 + (gpatch->height + offset) * gr_viewsin;
swallVerts[2].z = spr->z2 + (gpatch->height + offset) * gr_viewsin;
*Fixed OpenGL's handling of cv_translucency effects (should not remove shadows if off, but perhaps make them use default alpha?), someone was a bit hasty! *De-stupified the MD2 status checks regarding drawing of sprites - if they're already checked before calling HWR_DrawSprite, there's no point doing them WITHIN the function as well *Split off sprite shadow code into HWR_DrawSpriteShadow for convenience git-svn-id: https://code.orospakr.ca/svn/srb2/trunk@8983 6de4a73c-47e2-0310-b8c1-93d6ecd3f8cd
2015-01-02 15:14:22 +00:00
}
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
// We also need to move the bottom ones away when shadowoffs is on
if (cv_shadowoffs.value)
*Fixed OpenGL's handling of cv_translucency effects (should not remove shadows if off, but perhaps make them use default alpha?), someone was a bit hasty! *De-stupified the MD2 status checks regarding drawing of sprites - if they're already checked before calling HWR_DrawSprite, there's no point doing them WITHIN the function as well *Split off sprite shadow code into HWR_DrawSpriteShadow for convenience git-svn-id: https://code.orospakr.ca/svn/srb2/trunk@8983 6de4a73c-47e2-0310-b8c1-93d6ecd3f8cd
2015-01-02 15:14:22 +00:00
{
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
swallVerts[0].x = spr->x1 + offset * gr_viewcos;
swallVerts[1].x = spr->x2 + offset * gr_viewcos;
swallVerts[0].z = spr->z1 + offset * gr_viewsin;
swallVerts[1].z = spr->z2 + offset * gr_viewsin;
*Fixed OpenGL's handling of cv_translucency effects (should not remove shadows if off, but perhaps make them use default alpha?), someone was a bit hasty! *De-stupified the MD2 status checks regarding drawing of sprites - if they're already checked before calling HWR_DrawSprite, there's no point doing them WITHIN the function as well *Split off sprite shadow code into HWR_DrawSpriteShadow for convenience git-svn-id: https://code.orospakr.ca/svn/srb2/trunk@8983 6de4a73c-47e2-0310-b8c1-93d6ecd3f8cd
2015-01-02 15:14:22 +00:00
}
if (spr->flip)
{
swallVerts[0].sow = swallVerts[3].sow = gpatch->max_s;
swallVerts[2].sow = swallVerts[1].sow = 0;
}
else
{
swallVerts[0].sow = swallVerts[3].sow = 0;
swallVerts[2].sow = swallVerts[1].sow = gpatch->max_s;
}
// flip the texture coords (look familiar?)
if (spr->vflip)
{
swallVerts[3].tow = swallVerts[2].tow = gpatch->max_t;
swallVerts[0].tow = swallVerts[1].tow = 0;
}
else
{
swallVerts[3].tow = swallVerts[2].tow = 0;
swallVerts[0].tow = swallVerts[1].tow = gpatch->max_t;
}
sSurf.FlatColor.s.red = 0x00;
sSurf.FlatColor.s.blue = 0x00;
sSurf.FlatColor.s.green = 0x00;
/*if (spr->mobj->frame & FF_TRANSMASK || spr->mobj->flags2 & MF2_SHADOW)
{
sector_t *sector = spr->mobj->subsector->sector;
Some minor cleanup for OpenGL sprite/MD2 code
2015-01-28 15:16:50 +00:00
UINT8 lightlevel = 255;
*Fixed OpenGL's handling of cv_translucency effects (should not remove shadows if off, but perhaps make them use default alpha?), someone was a bit hasty! *De-stupified the MD2 status checks regarding drawing of sprites - if they're already checked before calling HWR_DrawSprite, there's no point doing them WITHIN the function as well *Split off sprite shadow code into HWR_DrawSpriteShadow for convenience git-svn-id: https://code.orospakr.ca/svn/srb2/trunk@8983 6de4a73c-47e2-0310-b8c1-93d6ecd3f8cd
2015-01-02 15:14:22 +00:00
extracolormap_t *colormap = sector->extra_colormap;
if (sector->numlights)
{
INT32 light = R_GetPlaneLight(sector, spr->mobj->floorz, false);
if (!(spr->mobj->frame & FF_FULLBRIGHT))
lightlevel = *sector->lightlist[light].lightlevel;
if (sector->lightlist[light].extra_colormap)
colormap = sector->lightlist[light].extra_colormap;
}
else
{
lightlevel = sector->lightlevel;
if (sector->extra_colormap)
colormap = sector->extra_colormap;
}
if (colormap)
sSurf.FlatColor.rgba = HWR_Lighting(lightlevel/2, colormap->rgba, colormap->fadergba, false, true);
else
sSurf.FlatColor.rgba = HWR_Lighting(lightlevel/2, NORMALFOG, FADEFOG, false, true);
}*/
// shadow is always half as translucent as the sprite itself
if (!cv_translucency.value) // use default translucency (main sprite won't have any translucency)
sSurf.FlatColor.s.alpha = 0x80; // default
else if (spr->mobj->flags2 & MF2_SHADOW)
sSurf.FlatColor.s.alpha = 0x20;
else if (spr->mobj->frame & FF_TRANSMASK)
{
HWR_TranstableToAlpha((spr->mobj->frame & FF_TRANSMASK)>>FF_TRANSSHIFT, &sSurf);
sSurf.FlatColor.s.alpha /= 2; //cut alpha in half!
}
else
sSurf.FlatColor.s.alpha = 0x80; // default
if (sSurf.FlatColor.s.alpha > floorheight/4)
{
sSurf.FlatColor.s.alpha = (UINT8)(sSurf.FlatColor.s.alpha - floorheight/4);
HWD.pfnDrawPolygon(&sSurf, swallVerts, 4, PF_Translucent|PF_Modulated|PF_Clip);
}
}
OpenGL Sprite Splitting
2018-03-22 00:52:14 +00:00
static void HWR_SplitSprite(gr_vissprite_t *spr)
{
float this_scale = 1.0f;
FOutVector wallVerts[4];
GLPatch_t *gpatch;
FSurfaceInfo Surf;
const boolean hires = (spr->mobj && spr->mobj->skin && ((skin_t *)spr->mobj->skin)->flags & SF_HIRES);
extracolormap_t *colormap;
FUINT lightlevel;
FBITFIELD blend = 0;
UINT8 alpha;
INT32 i;
float realtop, realbot, top, bot;
float towtop, towbot, towmult;
float bheight;
const sector_t *sector = spr->mobj->subsector->sector;
const lightlist_t *list = sector->lightlist;
#ifdef ESLOPE
float endrealtop, endrealbot, endtop, endbot;
float endbheight;
fixed_t temp;
fixed_t v1x, v1y, v2x, v2y;
#endif
this_scale = FIXED_TO_FLOAT(spr->mobj->scale);
if (hires)
this_scale = this_scale * FIXED_TO_FLOAT(((skin_t *)spr->mobj->skin)->highresscale);
gpatch = W_CachePatchNum(spr->patchlumpnum, PU_CACHE);
// cache the patch in the graphics card memory
//12/12/99: Hurdler: same comment as above (for md2)
//Hurdler: 25/04/2000: now support colormap in hardware mode
HWR_GetMappedPatch(gpatch, spr->colormap);
// Draw shadow BEFORE sprite
if (cv_shadow.value // Shadows enabled
&& (spr->mobj->flags & (MF_SCENERY|MF_SPAWNCEILING|MF_NOGRAVITY)) != (MF_SCENERY|MF_SPAWNCEILING|MF_NOGRAVITY) // Ceiling scenery have no shadow.
&& !(spr->mobj->flags2 & MF2_DEBRIS) // Debris have no corona or shadow.
#ifdef ALAM_LIGHTING
&& !(t_lspr[spr->mobj->sprite]->type // Things with dynamic lights have no shadow.
&& (!spr->mobj->player || spr->mobj->player->powers[pw_super])) // Except for non-super players.
#endif
&& (spr->mobj->z >= spr->mobj->floorz)) // Without this, your shadow shows on the floor, even after you die and fall through the ground.
{
////////////////////
// SHADOW SPRITE! //
////////////////////
HWR_DrawSpriteShadow(spr, gpatch, this_scale);
}
wallVerts[0].x = wallVerts[3].x = spr->x1;
wallVerts[2].x = wallVerts[1].x = spr->x2;
wallVerts[0].z = wallVerts[3].z = spr->z1;
wallVerts[1].z = wallVerts[2].z = spr->z2;
wallVerts[2].y = wallVerts[3].y = spr->ty;
if (spr->mobj && this_scale != 1.0f)
wallVerts[0].y = wallVerts[1].y = spr->ty - gpatch->height * this_scale;
else
wallVerts[0].y = wallVerts[1].y = spr->ty - gpatch->height;
v1x = FLOAT_TO_FIXED(spr->x1);
v1y = FLOAT_TO_FIXED(spr->z1);
v2x = FLOAT_TO_FIXED(spr->x2);
v2y = FLOAT_TO_FIXED(spr->z2);
if (spr->flip)
{
wallVerts[0].sow = wallVerts[3].sow = gpatch->max_s;
wallVerts[2].sow = wallVerts[1].sow = 0;
}else{
wallVerts[0].sow = wallVerts[3].sow = 0;
wallVerts[2].sow = wallVerts[1].sow = gpatch->max_s;
}
// flip the texture coords (look familiar?)
if (spr->vflip)
{
wallVerts[3].tow = wallVerts[2].tow = gpatch->max_t;
wallVerts[0].tow = wallVerts[1].tow = 0;
}else{
wallVerts[3].tow = wallVerts[2].tow = 0;
wallVerts[0].tow = wallVerts[1].tow = gpatch->max_t;
}
dispoffset OGL fix [by Azeonus]
2018-12-18 21:37:01 +00:00
// if it has a dispoffset, push it a little towards the camera
if (spr->dispoffset) {
float co = -gr_viewcos*(0.05*spr->dispoffset);
float si = -gr_viewsin*(0.05*spr->dispoffset);
wallVerts[0].z = wallVerts[3].z = wallVerts[0].z+si;
wallVerts[1].z = wallVerts[2].z = wallVerts[1].z+si;
wallVerts[0].x = wallVerts[3].x = wallVerts[0].x+co;
wallVerts[1].x = wallVerts[2].x = wallVerts[1].x+co;
}
OpenGL Sprite Splitting
2018-03-22 00:52:14 +00:00
realtop = top = wallVerts[3].y;
realbot = bot = wallVerts[0].y;
towtop = wallVerts[3].tow;
towbot = wallVerts[0].tow;
towmult = (towbot - towtop) / (top - bot);
#ifdef ESLOPE
endrealtop = endtop = wallVerts[2].y;
endrealbot = endbot = wallVerts[1].y;
#endif
if (!cv_translucency.value) // translucency disabled
{
Surf.FlatColor.s.alpha = 0xFF;
blend = PF_Translucent|PF_Occlude;
}
else if (spr->mobj->flags2 & MF2_SHADOW)
{
Surf.FlatColor.s.alpha = 0x40;
blend = PF_Translucent;
}
else if (spr->mobj->frame & FF_TRANSMASK)
blend = HWR_TranstableToAlpha((spr->mobj->frame & FF_TRANSMASK)>>FF_TRANSSHIFT, &Surf);
else
{
// BP: i agree that is little better in environement but it don't
// work properly under glide nor with fogcolor to ffffff :(
// Hurdler: PF_Environement would be cool, but we need to fix
// the issue with the fog before
Surf.FlatColor.s.alpha = 0xFF;
blend = PF_Translucent|PF_Occlude;
}
alpha = Surf.FlatColor.s.alpha;
// Start with the lightlevel and colormap from the top of the sprite
lightlevel = *list[sector->numlights - 1].lightlevel;
colormap = list[sector->numlights - 1].extra_colormap;
i = 0;
temp = FLOAT_TO_FIXED(realtop);
Fix FF_FULLBRIGHT not working in sectors with multiple light levels in OpenGL
2018-05-16 20:04:57 +00:00
if (spr->mobj->frame & FF_FULLBRIGHT)
lightlevel = 255;
OpenGL Sprite Splitting
2018-03-22 00:52:14 +00:00
#ifdef ESLOPE
for (i = 1; i < sector->numlights; i++)
{
fixed_t h = sector->lightlist[i].slope ? P_GetZAt(sector->lightlist[i].slope, spr->mobj->x, spr->mobj->y)
: sector->lightlist[i].height;
if (h <= temp)
{
Fix FF_FULLBRIGHT not working in sectors with multiple light levels in OpenGL
2018-05-16 20:04:57 +00:00
if (!(spr->mobj->frame & FF_FULLBRIGHT))
lightlevel = *list[i-1].lightlevel;
OpenGL Sprite Splitting
2018-03-22 00:52:14 +00:00
colormap = list[i-1].extra_colormap;
break;
}
}
#else
i = R_GetPlaneLight(sector, temp, false);
Fix FF_FULLBRIGHT not working in sectors with multiple light levels in OpenGL
2018-05-16 20:04:57 +00:00
if (!(spr->mobj->frame & FF_FULLBRIGHT))
lightlevel = *list[i].lightlevel;
OpenGL Sprite Splitting
2018-03-22 00:52:14 +00:00
colormap = list[i].extra_colormap;
#endif
for (i = 0; i < sector->numlights; i++)
{
#ifdef ESLOPE
if (endtop < endrealbot)
#endif
if (top < realbot)
return;
// even if we aren't changing colormap or lightlevel, we still need to continue drawing down the sprite
if (!(list[i].flags & FF_NOSHADE) && (list[i].flags & FF_CUTSPRITES))
{
Fix FF_FULLBRIGHT not working in sectors with multiple light levels in OpenGL
2018-05-16 20:04:57 +00:00
if (!(spr->mobj->frame & FF_FULLBRIGHT))
lightlevel = *list[i].lightlevel;
OpenGL Sprite Splitting
2018-03-22 00:52:14 +00:00
colormap = list[i].extra_colormap;
}
#ifdef ESLOPE
if (i + 1 < sector->numlights)
{
if (list[i+1].slope)
{
temp = P_GetZAt(list[i+1].slope, v1x, v1y);
bheight = FIXED_TO_FLOAT(temp);
temp = P_GetZAt(list[i+1].slope, v2x, v2y);
endbheight = FIXED_TO_FLOAT(temp);
}
else
bheight = endbheight = FIXED_TO_FLOAT(list[i+1].height);
}
else
{
bheight = realbot;
endbheight = endrealbot;
}
#else
if (i + 1 < sector->numlights)
{
bheight = FIXED_TO_FLOAT(list[i+1].height);
}
else
{
bheight = realbot;
}
#endif
#ifdef ESLOPE
if (endbheight >= endtop)
#endif
if (bheight >= top)
continue;
bot = bheight;
if (bot < realbot)
bot = realbot;
#ifdef ESLOPE
endbot = endbheight;
if (endbot < endrealbot)
endbot = endrealbot;
#endif
#ifdef ESLOPE
wallVerts[3].tow = towtop + ((realtop - top) * towmult);
wallVerts[2].tow = towtop + ((endrealtop - endtop) * towmult);
wallVerts[0].tow = towtop + ((realtop - bot) * towmult);
wallVerts[1].tow = towtop + ((endrealtop - endbot) * towmult);
wallVerts[3].y = top;
wallVerts[2].y = endtop;
wallVerts[0].y = bot;
wallVerts[1].y = endbot;
#else
wallVerts[3].tow = wallVerts[2].tow = towtop + ((realtop - top) * towmult);
wallVerts[0].tow = wallVerts[1].tow = towtop + ((realtop - bot) * towmult);
wallVerts[2].y = wallVerts[3].y = top;
wallVerts[0].y = wallVerts[1].y = bot;
#endif
if (colormap)
Surf.FlatColor.rgba = HWR_Lighting(lightlevel, colormap->rgba, colormap->fadergba, false, false);
else
Surf.FlatColor.rgba = HWR_Lighting(lightlevel, NORMALFOG, FADEFOG, false, false);
Surf.FlatColor.s.alpha = alpha;
HWD.pfnDrawPolygon(&Surf, wallVerts, 4, blend|PF_Modulated|PF_Clip);
top = bot;
#ifdef ESLOPE
endtop = endbot;
#endif
}
bot = realbot;
#ifdef ESLOPE
endbot = endrealbot;
if (endtop <= endrealbot)
#endif
if (top <= realbot)
return;
// If we're ever down here, somehow the above loop hasn't draw all the light levels of sprite
#ifdef ESLOPE
wallVerts[3].tow = towtop + ((realtop - top) * towmult);
wallVerts[2].tow = towtop + ((endrealtop - endtop) * towmult);
wallVerts[0].tow = towtop + ((realtop - bot) * towmult);
wallVerts[1].tow = towtop + ((endrealtop - endbot) * towmult);
wallVerts[3].y = top;
wallVerts[2].y = endtop;
wallVerts[0].y = bot;
wallVerts[1].y = endbot;
#else
wallVerts[3].tow = wallVerts[2].tow = towtop + ((realtop - top) * towmult);
wallVerts[0].tow = wallVerts[1].tow = towtop + ((realtop - bot) * towmult);
wallVerts[2].y = wallVerts[3].y = top;
wallVerts[0].y = wallVerts[1].y = bot;
#endif
if (colormap)
Surf.FlatColor.rgba = HWR_Lighting(lightlevel, colormap->rgba, colormap->fadergba, false, false);
else
Surf.FlatColor.rgba = HWR_Lighting(lightlevel, NORMALFOG, FADEFOG, false, false);
Surf.FlatColor.s.alpha = alpha;
HWD.pfnDrawPolygon(&Surf, wallVerts, 4, blend|PF_Modulated|PF_Clip);
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// -----------------+
// HWR_DrawSprite : Draw flat sprites
// : (monsters, bonuses, weapons, lights, ...)
// Returns :
// -----------------+
static void HWR_DrawSprite(gr_vissprite_t *spr)
{
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
float this_scale = 1.0f;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
FOutVector wallVerts[4];
GLPatch_t *gpatch; // sprite patch converted to hardware
FSurfaceInfo Surf;
const boolean hires = (spr->mobj && spr->mobj->skin && ((skin_t *)spr->mobj->skin)->flags & SF_HIRES);
if (spr->mobj)
this_scale = FIXED_TO_FLOAT(spr->mobj->scale);
if (hires)
this_scale = this_scale * FIXED_TO_FLOAT(((skin_t *)spr->mobj->skin)->highresscale);
if (!spr->mobj)
return;
if (!spr->mobj->subsector)
return;
OpenGL Sprite Splitting
2018-03-22 00:52:14 +00:00
if (spr->mobj->subsector->sector->numlights)
{
HWR_SplitSprite(spr);
return;
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// cache sprite graphics
//12/12/99: Hurdler:
// OK, I don't change anything for MD2 support because I want to be
// sure to do it the right way. So actually, we keep normal sprite
// in memory and we add the md2 model if it exists for that sprite
gpatch = W_CachePatchNum(spr->patchlumpnum, PU_CACHE);
#ifdef ALAM_LIGHTING
if (!(spr->mobj->flags2 & MF2_DEBRIS) && (spr->mobj->sprite != SPR_PLAY ||
(spr->mobj->player && spr->mobj->player->powers[pw_super])))
HWR_DL_AddLight(spr, gpatch);
#endif
// create the sprite billboard
//
// 3--2
// | /|
// |/ |
// 0--1
// these were already scaled in HWR_ProjectSprite
wallVerts[0].x = wallVerts[3].x = spr->x1;
wallVerts[2].x = wallVerts[1].x = spr->x2;
wallVerts[2].y = wallVerts[3].y = spr->ty;
if (spr->mobj && this_scale != 1.0f)
wallVerts[0].y = wallVerts[1].y = spr->ty - gpatch->height * this_scale;
else
wallVerts[0].y = wallVerts[1].y = spr->ty - gpatch->height;
// make a wall polygon (with 2 triangles), using the floor/ceiling heights,
// and the 2d map coords of start/end vertices
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
wallVerts[0].z = wallVerts[3].z = spr->z1;
wallVerts[1].z = wallVerts[2].z = spr->z2;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (spr->flip)
{
wallVerts[0].sow = wallVerts[3].sow = gpatch->max_s;
wallVerts[2].sow = wallVerts[1].sow = 0;
}else{
wallVerts[0].sow = wallVerts[3].sow = 0;
wallVerts[2].sow = wallVerts[1].sow = gpatch->max_s;
}
// flip the texture coords (look familiar?)
if (spr->vflip)
{
wallVerts[3].tow = wallVerts[2].tow = gpatch->max_t;
wallVerts[0].tow = wallVerts[1].tow = 0;
}else{
wallVerts[3].tow = wallVerts[2].tow = 0;
wallVerts[0].tow = wallVerts[1].tow = gpatch->max_t;
}
// cache the patch in the graphics card memory
//12/12/99: Hurdler: same comment as above (for md2)
//Hurdler: 25/04/2000: now support colormap in hardware mode
HWR_GetMappedPatch(gpatch, spr->colormap);
// Draw shadow BEFORE sprite
if (cv_shadow.value // Shadows enabled
*Fixed OpenGL's handling of cv_translucency effects (should not remove shadows if off, but perhaps make them use default alpha?), someone was a bit hasty! *De-stupified the MD2 status checks regarding drawing of sprites - if they're already checked before calling HWR_DrawSprite, there's no point doing them WITHIN the function as well *Split off sprite shadow code into HWR_DrawSpriteShadow for convenience git-svn-id: https://code.orospakr.ca/svn/srb2/trunk@8983 6de4a73c-47e2-0310-b8c1-93d6ecd3f8cd
2015-01-02 15:14:22 +00:00
&& (spr->mobj->flags & (MF_SCENERY|MF_SPAWNCEILING|MF_NOGRAVITY)) != (MF_SCENERY|MF_SPAWNCEILING|MF_NOGRAVITY) // Ceiling scenery have no shadow.
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
&& !(spr->mobj->flags2 & MF2_DEBRIS) // Debris have no corona or shadow.
#ifdef ALAM_LIGHTING
&& !(t_lspr[spr->mobj->sprite]->type // Things with dynamic lights have no shadow.
&& (!spr->mobj->player || spr->mobj->player->powers[pw_super])) // Except for non-super players.
#endif
&& (spr->mobj->z >= spr->mobj->floorz)) // Without this, your shadow shows on the floor, even after you die and fall through the ground.
{
////////////////////
// SHADOW SPRITE! //
////////////////////
*Fixed OpenGL's handling of cv_translucency effects (should not remove shadows if off, but perhaps make them use default alpha?), someone was a bit hasty! *De-stupified the MD2 status checks regarding drawing of sprites - if they're already checked before calling HWR_DrawSprite, there's no point doing them WITHIN the function as well *Split off sprite shadow code into HWR_DrawSpriteShadow for convenience git-svn-id: https://code.orospakr.ca/svn/srb2/trunk@8983 6de4a73c-47e2-0310-b8c1-93d6ecd3f8cd
2015-01-02 15:14:22 +00:00
HWR_DrawSpriteShadow(spr, gpatch, this_scale);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
dispoffset OGL fix [by Azeonus]
2018-12-18 21:37:01 +00:00
// if it has a dispoffset, push it a little towards the camera
if (spr->dispoffset) {
float co = -gr_viewcos*(0.05*spr->dispoffset);
float si = -gr_viewsin*(0.05*spr->dispoffset);
wallVerts[0].z = wallVerts[3].z = wallVerts[0].z+si;
wallVerts[1].z = wallVerts[2].z = wallVerts[1].z+si;
wallVerts[0].x = wallVerts[3].x = wallVerts[0].x+co;
wallVerts[1].x = wallVerts[2].x = wallVerts[1].x+co;
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// This needs to be AFTER the shadows so that the regular sprites aren't drawn completely black.
// sprite lighting by modulating the RGB components
/// \todo coloured
// colormap test
{
sector_t *sector = spr->mobj->subsector->sector;
Some minor cleanup for OpenGL sprite/MD2 code
2015-01-28 15:16:50 +00:00
UINT8 lightlevel = 255;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
extracolormap_t *colormap = sector->extra_colormap;
OpenGL Sprite Splitting
2018-03-22 00:52:14 +00:00
if (!(spr->mobj->frame & FF_FULLBRIGHT))
lightlevel = sector->lightlevel;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (colormap)
Surf.FlatColor.rgba = HWR_Lighting(lightlevel, colormap->rgba, colormap->fadergba, false, false);
else
Surf.FlatColor.rgba = HWR_Lighting(lightlevel, NORMALFOG, FADEFOG, false, false);
}
{
FBITFIELD blend = 0;
*Fixed OpenGL's handling of cv_translucency effects (should not remove shadows if off, but perhaps make them use default alpha?), someone was a bit hasty! *De-stupified the MD2 status checks regarding drawing of sprites - if they're already checked before calling HWR_DrawSprite, there's no point doing them WITHIN the function as well *Split off sprite shadow code into HWR_DrawSpriteShadow for convenience git-svn-id: https://code.orospakr.ca/svn/srb2/trunk@8983 6de4a73c-47e2-0310-b8c1-93d6ecd3f8cd
2015-01-02 15:14:22 +00:00
if (!cv_translucency.value) // translucency disabled
{
Surf.FlatColor.s.alpha = 0xFF;
blend = PF_Translucent|PF_Occlude;
}
else if (spr->mobj->flags2 & MF2_SHADOW)
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
Surf.FlatColor.s.alpha = 0x40;
blend = PF_Translucent;
}
else if (spr->mobj->frame & FF_TRANSMASK)
blend = HWR_TranstableToAlpha((spr->mobj->frame & FF_TRANSMASK)>>FF_TRANSSHIFT, &Surf);
else
{
// BP: i agree that is little better in environement but it don't
// work properly under glide nor with fogcolor to ffffff :(
// Hurdler: PF_Environement would be cool, but we need to fix
// the issue with the fog before
Surf.FlatColor.s.alpha = 0xFF;
blend = PF_Translucent|PF_Occlude;
}
HWD.pfnDrawPolygon(&Surf, wallVerts, 4, blend|PF_Modulated|PF_Clip);
}
}
#ifdef HWPRECIP
// Sprite drawer for precipitation
static inline void HWR_DrawPrecipitationSprite(gr_vissprite_t *spr)
{
FBITFIELD blend = 0;
FOutVector wallVerts[4];
GLPatch_t *gpatch; // sprite patch converted to hardware
FSurfaceInfo Surf;
if (!spr->mobj)
return;
if (!spr->mobj->subsector)
return;
// cache sprite graphics
gpatch = W_CachePatchNum(spr->patchlumpnum, PU_CACHE);
// create the sprite billboard
//
// 3--2
// | /|
// |/ |
// 0--1
wallVerts[0].x = wallVerts[3].x = spr->x1;
wallVerts[2].x = wallVerts[1].x = spr->x2;
wallVerts[2].y = wallVerts[3].y = spr->ty;
wallVerts[0].y = wallVerts[1].y = spr->ty - gpatch->height;
// make a wall polygon (with 2 triangles), using the floor/ceiling heights,
// and the 2d map coords of start/end vertices
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
wallVerts[0].z = wallVerts[3].z = spr->z1;
wallVerts[1].z = wallVerts[2].z = spr->z2;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
wallVerts[0].sow = wallVerts[3].sow = 0;
wallVerts[2].sow = wallVerts[1].sow = gpatch->max_s;
wallVerts[3].tow = wallVerts[2].tow = 0;
wallVerts[0].tow = wallVerts[1].tow = gpatch->max_t;
// cache the patch in the graphics card memory
//12/12/99: Hurdler: same comment as above (for md2)
//Hurdler: 25/04/2000: now support colormap in hardware mode
HWR_GetMappedPatch(gpatch, spr->colormap);
YUP, I just fixed OGL Precipitation with this branch too. What the fuck? Squashing ancient bugs, woo.
2016-10-27 22:41:24 +00:00
// colormap test
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
YUP, I just fixed OGL Precipitation with this branch too. What the fuck? Squashing ancient bugs, woo.
2016-10-27 22:41:24 +00:00
sector_t *sector = spr->mobj->subsector->sector;
UINT8 lightlevel = 255;
extracolormap_t *colormap = sector->extra_colormap;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
YUP, I just fixed OGL Precipitation with this branch too. What the fuck? Squashing ancient bugs, woo.
2016-10-27 22:41:24 +00:00
if (sector->numlights)
{
INT32 light;
light = R_GetPlaneLight(sector, spr->mobj->z + spr->mobj->height, false); // Always use the light at the top instead of whatever I was doing before
if (!(spr->mobj->frame & FF_FULLBRIGHT))
lightlevel = *sector->lightlist[light].lightlevel;
if (sector->lightlist[light].extra_colormap)
colormap = sector->lightlist[light].extra_colormap;
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
else
YUP, I just fixed OGL Precipitation with this branch too. What the fuck? Squashing ancient bugs, woo.
2016-10-27 22:41:24 +00:00
{
if (!(spr->mobj->frame & FF_FULLBRIGHT))
lightlevel = sector->lightlevel;
if (sector->extra_colormap)
colormap = sector->extra_colormap;
}
if (colormap)
Surf.FlatColor.rgba = HWR_Lighting(lightlevel, colormap->rgba, colormap->fadergba, false, false);
else
Surf.FlatColor.rgba = HWR_Lighting(lightlevel, NORMALFOG, FADEFOG, false, false);
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (spr->mobj->flags2 & MF2_SHADOW)
{
Surf.FlatColor.s.alpha = 0x40;
blend = PF_Translucent;
}
else if (spr->mobj->frame & FF_TRANSMASK)
blend = HWR_TranstableToAlpha((spr->mobj->frame & FF_TRANSMASK)>>FF_TRANSSHIFT, &Surf);
else
{
// BP: i agree that is little better in environement but it don't
// work properly under glide nor with fogcolor to ffffff :(
// Hurdler: PF_Environement would be cool, but we need to fix
// the issue with the fog before
Surf.FlatColor.s.alpha = 0xFF;
blend = PF_Translucent|PF_Occlude;
}
HWD.pfnDrawPolygon(&Surf, wallVerts, 4, blend|PF_Modulated|PF_Clip);
}
#endif
// --------------------------------------------------------------------------
// Sort vissprites by distance
// --------------------------------------------------------------------------
static gr_vissprite_t gr_vsprsortedhead;
static void HWR_SortVisSprites(void)
{
UINT32 i;
gr_vissprite_t *ds, *dsprev, *dsnext, *dsfirst;
gr_vissprite_t *best = NULL;
gr_vissprite_t unsorted;
hardware: init undef vars in HWR_SortVisSprites
2016-12-23 00:42:00 +00:00
float bestdist = 0.0f;
INT32 bestdispoffset = 0;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (!gr_visspritecount)
return;
dsfirst = HWR_GetVisSprite(0);
for (i = 0, dsnext = dsfirst, ds = NULL; i < gr_visspritecount; i++)
{
dsprev = ds;
ds = dsnext;
if (i < gr_visspritecount - 1) dsnext = HWR_GetVisSprite(i + 1);
ds->next = dsnext;
ds->prev = dsprev;
}
// Fix first and last. ds still points to the last one after the loop
dsfirst->prev = &unsorted;
unsorted.next = dsfirst;
NULL checks
2016-06-13 14:07:10 +00:00
if (ds)
ds->next = &unsorted;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
unsorted.prev = ds;
// pull the vissprites out by scale
gr_vsprsortedhead.next = gr_vsprsortedhead.prev = &gr_vsprsortedhead;
for (i = 0; i < gr_visspritecount; i++)
{
Also fixed it for MD2s
2016-02-22 05:02:09 +00:00
best = NULL;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
for (ds = unsorted.next; ds != &unsorted; ds = ds->next)
{
Also fixed it for MD2s
2016-02-22 05:02:09 +00:00
if (!best || ds->tz > bestdist)
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
bestdist = ds->tz;
dispoffset now works in OpenGL
2015-11-23 17:01:10 +00:00
bestdispoffset = ds->dispoffset;
best = ds;
}
// order visprites of same scale by dispoffset, smallest first
else if (ds->tz == bestdist && ds->dispoffset < bestdispoffset)
{
bestdispoffset = ds->dispoffset;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
best = ds;
}
}
NULL checks
2016-06-13 14:07:10 +00:00
if (best)
{
best->next->prev = best->prev;
best->prev->next = best->next;
best->next = &gr_vsprsortedhead;
best->prev = gr_vsprsortedhead.prev;
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
gr_vsprsortedhead.prev->next = best;
gr_vsprsortedhead.prev = best;
}
Small hacky fix for MD2s and sprites until sorting for walls, floors, and sprites is done Sorts all translucent sprites and MD2s so they're drawn after all the opaque ones. Fixes most of the observable issues between translucent MD2s and opaque sprites/MD2s.
2018-03-09 09:58:10 +00:00
// Sryder: Oh boy, while it's nice having ALL the sprites sorted properly, it fails when we bring MD2's into the
// mix and they want to be translucent. So let's place all the translucent sprites and MD2's AFTER
// everything else, but still ordered of course, the depth buffer can handle the opaque ones plenty fine.
// We just need to move all translucent ones to the end in order
// TODO: Fully sort all sprites and MD2s with walls and floors, this part will be unnecessary after that
best = gr_vsprsortedhead.next;
for (i = 0; i < gr_visspritecount; i++)
{
if ((best->mobj->flags2 & MF2_SHADOW) || (best->mobj->frame & FF_TRANSMASK))
{
if (best == gr_vsprsortedhead.next)
{
gr_vsprsortedhead.next = best->next;
}
best->prev->next = best->next;
best->next->prev = best->prev;
best->prev = gr_vsprsortedhead.prev;
gr_vsprsortedhead.prev->next = best;
gr_vsprsortedhead.prev = best;
ds = best;
best = best->next;
ds->next = &gr_vsprsortedhead;
}
else
best = best->next;
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
// A drawnode is something that points to a 3D floor, 3D side, or masked
// middle texture. This is used for sorting with sprites.
typedef struct
{
wallVert3D wallVerts[4];
FSurfaceInfo Surf;
INT32 texnum;
FBITFIELD blend;
INT32 drawcount;
#ifndef SORTING
fixed_t fixedheight;
#endif
boolean fogwall;
INT32 lightlevel;
extracolormap_t *wallcolormap; // Doing the lighting in HWR_RenderWall now for correct fog after sorting
} wallinfo_t;
static wallinfo_t *wallinfo = NULL;
static size_t numwalls = 0; // a list of transparent walls to be drawn
static void HWR_RenderWall(wallVert3D *wallVerts, FSurfaceInfo *pSurf, FBITFIELD blend, boolean fogwall, INT32 lightlevel, extracolormap_t *wallcolormap);
#define MAX_TRANSPARENTWALL 256
typedef struct
{
extrasubsector_t *xsub;
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
boolean isceiling;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
fixed_t fixedheight;
INT32 lightlevel;
lumpnum_t lumpnum;
INT32 alpha;
sector_t *FOFSector;
FBITFIELD blend;
boolean fogplane;
extracolormap_t *planecolormap;
INT32 drawcount;
} planeinfo_t;
static size_t numplanes = 0; // a list of transparent floors to be drawn
static planeinfo_t *planeinfo = NULL;
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
typedef struct
{
polyobj_t *polysector;
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
boolean isceiling;
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
fixed_t fixedheight;
INT32 lightlevel;
lumpnum_t lumpnum;
INT32 alpha;
sector_t *FOFSector;
FBITFIELD blend;
extracolormap_t *planecolormap;
INT32 drawcount;
} polyplaneinfo_t;
static size_t numpolyplanes = 0; // a list of transparent poyobject floors to be drawn
static polyplaneinfo_t *polyplaneinfo = NULL;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
#ifndef SORTING
size_t numfloors = 0;
#else
//static floorinfo_t *floorinfo = NULL;
//static size_t numfloors = 0;
//Hurdler: 3D water sutffs
typedef struct gr_drawnode_s
{
planeinfo_t *plane;
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
polyplaneinfo_t *polyplane;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
wallinfo_t *wall;
gr_vissprite_t *sprite;
// struct gr_drawnode_s *next;
// struct gr_drawnode_s *prev;
} gr_drawnode_t;
static INT32 drawcount = 0;
#define MAX_TRANSPARENTFLOOR 512
// This will likely turn into a copy of HWR_Add3DWater and replace it.
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
void HWR_AddTransparentFloor(lumpnum_t lumpnum, extrasubsector_t *xsub, boolean isceiling,
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
fixed_t fixedheight, INT32 lightlevel, INT32 alpha, sector_t *FOFSector, FBITFIELD blend, boolean fogplane, extracolormap_t *planecolormap)
{
static size_t allocedplanes = 0;
// Force realloc if buffer has been freed
if (!planeinfo)
allocedplanes = 0;
if (allocedplanes < numplanes + 1)
{
allocedplanes += MAX_TRANSPARENTFLOOR;
Z_Realloc(planeinfo, allocedplanes * sizeof (*planeinfo), PU_LEVEL, &planeinfo);
}
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
planeinfo[numplanes].isceiling = isceiling;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
planeinfo[numplanes].fixedheight = fixedheight;
planeinfo[numplanes].lightlevel = lightlevel;
planeinfo[numplanes].lumpnum = lumpnum;
planeinfo[numplanes].xsub = xsub;
planeinfo[numplanes].alpha = alpha;
planeinfo[numplanes].FOFSector = FOFSector;
planeinfo[numplanes].blend = blend;
planeinfo[numplanes].fogplane = fogplane;
planeinfo[numplanes].planecolormap = planecolormap;
planeinfo[numplanes].drawcount = drawcount++;
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
numplanes++;
}
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
// Adding this for now until I can create extrasubsector info for polyobjects
// When that happens it'll just be done through HWR_AddTransparentFloor and HWR_RenderPlane
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
void HWR_AddTransparentPolyobjectFloor(lumpnum_t lumpnum, polyobj_t *polysector, boolean isceiling,
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
fixed_t fixedheight, INT32 lightlevel, INT32 alpha, sector_t *FOFSector, FBITFIELD blend, extracolormap_t *planecolormap)
{
static size_t allocedpolyplanes = 0;
// Force realloc if buffer has been freed
if (!polyplaneinfo)
allocedpolyplanes = 0;
if (allocedpolyplanes < numpolyplanes + 1)
{
allocedpolyplanes += MAX_TRANSPARENTFLOOR;
Z_Realloc(polyplaneinfo, allocedpolyplanes * sizeof (*polyplaneinfo), PU_LEVEL, &polyplaneinfo);
}
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
polyplaneinfo[numpolyplanes].isceiling = isceiling;
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
polyplaneinfo[numpolyplanes].fixedheight = fixedheight;
polyplaneinfo[numpolyplanes].lightlevel = lightlevel;
polyplaneinfo[numpolyplanes].lumpnum = lumpnum;
polyplaneinfo[numpolyplanes].polysector = polysector;
polyplaneinfo[numpolyplanes].alpha = alpha;
polyplaneinfo[numpolyplanes].FOFSector = FOFSector;
polyplaneinfo[numpolyplanes].blend = blend;
polyplaneinfo[numpolyplanes].planecolormap = planecolormap;
polyplaneinfo[numpolyplanes].drawcount = drawcount++;
numpolyplanes++;
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
//
// HWR_CreateDrawNodes
// Creates and sorts a list of drawnodes for the scene being rendered.
static void HWR_CreateDrawNodes(void)
{
UINT32 i = 0, p = 0, prev = 0, loop;
const fixed_t pviewz = dup_viewz;
// Dump EVERYTHING into a huge drawnode list. Then we'll sort it!
// Could this be optimized into _AddTransparentWall/_AddTransparentPlane?
// Hell yes! But sort algorithm must be modified to use a linked list.
gr_drawnode_t *sortnode = Z_Calloc((sizeof(planeinfo_t)*numplanes)
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
+ (sizeof(polyplaneinfo_t)*numpolyplanes)
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
+ (sizeof(wallinfo_t)*numwalls)
,PU_STATIC, NULL);
// todo:
// However, in reality we shouldn't be re-copying and shifting all this information
// that is already lying around. This should all be in some sort of linked list or lists.
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
size_t *sortindex = Z_Calloc(sizeof(size_t) * (numplanes + numpolyplanes + numwalls), PU_STATIC, NULL);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// If true, swap the draw order.
boolean shift = false;
for (i = 0; i < numplanes; i++, p++)
{
sortnode[p].plane = &planeinfo[i];
sortindex[p] = p;
}
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
for (i = 0; i < numpolyplanes; i++, p++)
{
sortnode[p].polyplane = &polyplaneinfo[i];
sortindex[p] = p;
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
for (i = 0; i < numwalls; i++, p++)
{
sortnode[p].wall = &wallinfo[i];
sortindex[p] = p;
}
// p is the number of stuff to sort
// Add the 3D floors, thicksides, and masked textures...
// Instead of going through drawsegs, we need to iterate
// through the lists of masked textures and
// translucent ffloors being drawn.
// This is a bubble sort! Wahoo!
// Stuff is sorted:
// sortnode[sortindex[0]] = farthest away
// sortnode[sortindex[p-1]] = closest
// "i" should be closer. "prev" should be further.
// The lower drawcount is, the further it is from the screen.
for (loop = 0; loop < p; loop++)
{
for (i = 1; i < p; i++)
{
prev = i-1;
if (sortnode[sortindex[i]].plane)
{
// What are we comparing it with?
if (sortnode[sortindex[prev]].plane)
{
// Plane (i) is further away than plane (prev)
if (ABS(sortnode[sortindex[i]].plane->fixedheight - pviewz) > ABS(sortnode[sortindex[prev]].plane->fixedheight - pviewz))
shift = true;
}
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
if (sortnode[sortindex[prev]].polyplane)
{
// Plane (i) is further away than polyplane (prev)
if (ABS(sortnode[sortindex[i]].plane->fixedheight - pviewz) > ABS(sortnode[sortindex[prev]].polyplane->fixedheight - pviewz))
shift = true;
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
else if (sortnode[sortindex[prev]].wall)
{
// Plane (i) is further than wall (prev)
if (sortnode[sortindex[i]].plane->drawcount > sortnode[sortindex[prev]].wall->drawcount)
shift = true;
}
}
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
else if (sortnode[sortindex[i]].polyplane)
{
// What are we comparing it with?
if (sortnode[sortindex[prev]].plane)
{
// Plane (i) is further away than plane (prev)
if (ABS(sortnode[sortindex[i]].polyplane->fixedheight - pviewz) > ABS(sortnode[sortindex[prev]].plane->fixedheight - pviewz))
shift = true;
}
if (sortnode[sortindex[prev]].polyplane)
{
// Plane (i) is further away than polyplane (prev)
if (ABS(sortnode[sortindex[i]].polyplane->fixedheight - pviewz) > ABS(sortnode[sortindex[prev]].polyplane->fixedheight - pviewz))
shift = true;
}
else if (sortnode[sortindex[prev]].wall)
{
// Plane (i) is further than wall (prev)
if (sortnode[sortindex[i]].polyplane->drawcount > sortnode[sortindex[prev]].wall->drawcount)
shift = true;
}
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
else if (sortnode[sortindex[i]].wall)
{
// What are we comparing it with?
if (sortnode[sortindex[prev]].plane)
{
// Wall (i) is further than plane(prev)
if (sortnode[sortindex[i]].wall->drawcount > sortnode[sortindex[prev]].plane->drawcount)
shift = true;
}
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
if (sortnode[sortindex[prev]].polyplane)
{
// Wall (i) is further than polyplane(prev)
if (sortnode[sortindex[i]].wall->drawcount > sortnode[sortindex[prev]].polyplane->drawcount)
shift = true;
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
else if (sortnode[sortindex[prev]].wall)
{
// Wall (i) is further than wall (prev)
if (sortnode[sortindex[i]].wall->drawcount > sortnode[sortindex[prev]].wall->drawcount)
shift = true;
}
}
if (shift)
{
size_t temp;
temp = sortindex[prev];
sortindex[prev] = sortindex[i];
sortindex[i] = temp;
shift = false;
}
} //i++
} // loop++
HWD.pfnSetTransform(&atransform);
// Okay! Let's draw it all! Woo!
for (i = 0; i < p; i++)
{
if (sortnode[sortindex[i]].plane)
{
// We aren't traversing the BSP tree, so make gr_frontsector null to avoid crashes.
gr_frontsector = NULL;
if (!(sortnode[sortindex[i]].plane->blend & PF_NoTexture))
HWR_GetFlat(sortnode[sortindex[i]].plane->lumpnum);
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
HWR_RenderPlane(NULL, sortnode[sortindex[i]].plane->xsub, sortnode[sortindex[i]].plane->isceiling, sortnode[sortindex[i]].plane->fixedheight, sortnode[sortindex[i]].plane->blend, sortnode[sortindex[i]].plane->lightlevel,
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
sortnode[sortindex[i]].plane->lumpnum, sortnode[sortindex[i]].plane->FOFSector, sortnode[sortindex[i]].plane->alpha, sortnode[sortindex[i]].plane->fogplane, sortnode[sortindex[i]].plane->planecolormap);
}
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
else if (sortnode[sortindex[i]].polyplane)
{
// We aren't traversing the BSP tree, so make gr_frontsector null to avoid crashes.
gr_frontsector = NULL;
if (!(sortnode[sortindex[i]].polyplane->blend & PF_NoTexture))
HWR_GetFlat(sortnode[sortindex[i]].polyplane->lumpnum);
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
HWR_RenderPolyObjectPlane(sortnode[sortindex[i]].polyplane->polysector, sortnode[sortindex[i]].polyplane->isceiling, sortnode[sortindex[i]].polyplane->fixedheight, sortnode[sortindex[i]].polyplane->blend, sortnode[sortindex[i]].polyplane->lightlevel,
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
sortnode[sortindex[i]].polyplane->lumpnum, sortnode[sortindex[i]].polyplane->FOFSector, sortnode[sortindex[i]].polyplane->alpha, sortnode[sortindex[i]].polyplane->planecolormap);
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
else if (sortnode[sortindex[i]].wall)
{
if (!(sortnode[sortindex[i]].wall->blend & PF_NoTexture))
HWR_GetTexture(sortnode[sortindex[i]].wall->texnum);
HWR_RenderWall(sortnode[sortindex[i]].wall->wallVerts, &sortnode[sortindex[i]].wall->Surf, sortnode[sortindex[i]].wall->blend, sortnode[sortindex[i]].wall->fogwall,
sortnode[sortindex[i]].wall->lightlevel, sortnode[sortindex[i]].wall->wallcolormap);
}
}
numwalls = 0;
numplanes = 0;
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
numpolyplanes = 0;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// No mem leaks, please.
Z_Free(sortnode);
Z_Free(sortindex);
}
#endif
// --------------------------------------------------------------------------
// Draw all vissprites
// --------------------------------------------------------------------------
#ifdef SORTING
Go through and draw MD2s and sprites at the same time so they are sorted from each other (mostly) correctly
2017-09-07 23:56:58 +00:00
// added the stransform so they can be switched as drawing happenes so MD2s and sprites are sorted correctly with each other
Warnings must die
2018-03-30 17:53:23 +00:00
static void HWR_DrawSprites(void)
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
if (gr_visspritecount > 0)
{
gr_vissprite_t *spr;
// draw all vissprites back to front
for (spr = gr_vsprsortedhead.next;
spr != &gr_vsprsortedhead;
spr = spr->next)
{
#ifdef HWPRECIP
if (spr->precip)
HWR_DrawPrecipitationSprite(spr);
else
#endif
MD2's replacing skinned mobj fix MD2's that replace mobjs that have a skin, but the skin doesn't have an MD2 now display.
2014-03-19 23:40:02 +00:00
if (spr->mobj && spr->mobj->skin && spr->mobj->sprite == SPR_PLAY)
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
*Fixed OpenGL's handling of cv_translucency effects (should not remove shadows if off, but perhaps make them use default alpha?), someone was a bit hasty! *De-stupified the MD2 status checks regarding drawing of sprites - if they're already checked before calling HWR_DrawSprite, there's no point doing them WITHIN the function as well *Split off sprite shadow code into HWR_DrawSpriteShadow for convenience git-svn-id: https://code.orospakr.ca/svn/srb2/trunk@8983 6de4a73c-47e2-0310-b8c1-93d6ecd3f8cd
2015-01-02 15:14:22 +00:00
if (!cv_grmd2.value || md2_playermodels[(skin_t*)spr->mobj->skin-skins].notfound || md2_playermodels[(skin_t*)spr->mobj->skin-skins].scale < 0.0f)
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
HWR_DrawSprite(spr);
Go through and draw MD2s and sprites at the same time so they are sorted from each other (mostly) correctly
2017-09-07 23:56:58 +00:00
else
HWR_DrawMD2(spr);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
Go through and draw MD2s and sprites at the same time so they are sorted from each other (mostly) correctly
2017-09-07 23:56:58 +00:00
else
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
Go through and draw MD2s and sprites at the same time so they are sorted from each other (mostly) correctly
2017-09-07 23:56:58 +00:00
if (!cv_grmd2.value || md2_models[spr->mobj->sprite].notfound || md2_models[spr->mobj->sprite].scale < 0.0f)
HWR_DrawSprite(spr);
else
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
HWR_DrawMD2(spr);
}
}
}
}
Go through and draw MD2s and sprites at the same time so they are sorted from each other (mostly) correctly
2017-09-07 23:56:58 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// --------------------------------------------------------------------------
// HWR_AddSprites
// During BSP traversal, this adds sprites by sector.
// --------------------------------------------------------------------------
static UINT8 sectorlight;
static void HWR_AddSprites(sector_t *sec)
{
mobj_t *thing;
#ifdef HWPRECIP
precipmobj_t *precipthing;
#endif
fixed_t approx_dist, limit_dist;
// BSP is traversed by subsector.
// A sector might have been split into several
// subsectors during BSP building.
// Thus we check whether its already added.
if (sec->validcount == validcount)
return;
// Well, now it will be done.
sec->validcount = validcount;
// sprite lighting
LightLevelToLum gone LightLevelToLum unused by everything now, it seemed to be giving wrong values for lightlevels.
2014-03-24 16:11:25 +00:00
sectorlight = sec->lightlevel & 0xff;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// Handle all things in sector.
// If a limit exists, handle things a tiny bit different.
if ((limit_dist = (fixed_t)((maptol & TOL_NIGHTS) ? cv_drawdist_nights.value : cv_drawdist.value) << FRACBITS))
{
for (thing = sec->thinglist; thing; thing = thing->snext)
{
if (thing->sprite == SPR_NULL || thing->flags2 & MF2_DONTDRAW)
continue;
Base draw distances on viewx/viewy coordinates, NOT the player object's coordinates (this can cause problems with things like skyboxes for instance). Splitscreen's player 2 should not affect what sprites player 1 can see, and vice versa! Especially not for precipitation, that just looks ridiculous. git-svn-id: https://code.orospakr.ca/svn/srb2/trunk@9041 6de4a73c-47e2-0310-b8c1-93d6ecd3f8cd
2015-02-15 17:15:55 +00:00
approx_dist = P_AproxDistance(viewx-thing->x, viewy-thing->y);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
Weather is already run client-side. What if we ran it render-side, for major performance gains? This commit will answer all your questions - and more!
2018-10-07 14:00:58 +00:00
if (approx_dist > limit_dist)
continue;
HWR_ProjectSprite(thing);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
}
else
{
// Draw everything in sector, no checks
for (thing = sec->thinglist; thing; thing = thing->snext)
if (!(thing->sprite == SPR_NULL || thing->flags2 & MF2_DONTDRAW))
HWR_ProjectSprite(thing);
}
#ifdef HWPRECIP
// Someone seriously wants infinite draw distance for precipitation?
if ((limit_dist = (fixed_t)cv_drawdist_precip.value << FRACBITS))
{
for (precipthing = sec->preciplist; precipthing; precipthing = precipthing->snext)
{
Base draw distances on viewx/viewy coordinates, NOT the player object's coordinates (this can cause problems with things like skyboxes for instance). Splitscreen's player 2 should not affect what sprites player 1 can see, and vice versa! Especially not for precipitation, that just looks ridiculous. git-svn-id: https://code.orospakr.ca/svn/srb2/trunk@9041 6de4a73c-47e2-0310-b8c1-93d6ecd3f8cd
2015-02-15 17:15:55 +00:00
if (precipthing->precipflags & PCF_INVISIBLE)
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
continue;
Base draw distances on viewx/viewy coordinates, NOT the player object's coordinates (this can cause problems with things like skyboxes for instance). Splitscreen's player 2 should not affect what sprites player 1 can see, and vice versa! Especially not for precipitation, that just looks ridiculous. git-svn-id: https://code.orospakr.ca/svn/srb2/trunk@9041 6de4a73c-47e2-0310-b8c1-93d6ecd3f8cd
2015-02-15 17:15:55 +00:00
approx_dist = P_AproxDistance(viewx-precipthing->x, viewy-precipthing->y);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
Weather is already run client-side. What if we ran it render-side, for major performance gains? This commit will answer all your questions - and more!
2018-10-07 14:00:58 +00:00
if (approx_dist > limit_dist)
continue;
HWR_ProjectPrecipitationSprite(precipthing);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
}
else
{
// Draw everything in sector, no checks
for (precipthing = sec->preciplist; precipthing; precipthing = precipthing->snext)
Base draw distances on viewx/viewy coordinates, NOT the player object's coordinates (this can cause problems with things like skyboxes for instance). Splitscreen's player 2 should not affect what sprites player 1 can see, and vice versa! Especially not for precipitation, that just looks ridiculous. git-svn-id: https://code.orospakr.ca/svn/srb2/trunk@9041 6de4a73c-47e2-0310-b8c1-93d6ecd3f8cd
2015-02-15 17:15:55 +00:00
if (!(precipthing->precipflags & PCF_INVISIBLE))
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
HWR_ProjectPrecipitationSprite(precipthing);
}
#endif
}
// --------------------------------------------------------------------------
// HWR_ProjectSprite
// Generates a vissprite for a thing if it might be visible.
// --------------------------------------------------------------------------
// BP why not use xtoviexangle/viewangletox like in bsp ?....
static void HWR_ProjectSprite(mobj_t *thing)
{
gr_vissprite_t *vis;
float tr_x, tr_y;
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
float tz;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
float x1, x2;
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
float z1, z2;
float rightsin, rightcos;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
float this_scale;
float gz, gzt;
spritedef_t *sprdef;
spriteframe_t *sprframe;
size_t lumpoff;
unsigned rot;
UINT8 flip;
angle_t ang;
INT32 heightsec, phs;
if (!thing)
return;
else
this_scale = FIXED_TO_FLOAT(thing->scale);
// transform the origin point
tr_x = FIXED_TO_FLOAT(thing->x) - gr_viewx;
tr_y = FIXED_TO_FLOAT(thing->y) - gr_viewy;
// rotation around vertical axis
tz = (tr_x * gr_viewcos) + (tr_y * gr_viewsin);
// thing is behind view plane?
Fix vissprite-related crashing in OGL
2016-02-22 04:32:38 +00:00
if (tz < ZCLIP_PLANE && (!cv_grmd2.value || md2_models[thing->sprite].notfound == true)) //Yellow: Only MD2's dont disappear
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
return;
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
// The above can stay as it works for cutting sprites that are too close
tr_x = FIXED_TO_FLOAT(thing->x);
tr_y = FIXED_TO_FLOAT(thing->y);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// decide which patch to use for sprite relative to player
#ifdef RANGECHECK
if ((unsigned)thing->sprite >= numsprites)
I_Error("HWR_ProjectSprite: invalid sprite number %i ", thing->sprite);
#endif
rot = thing->frame&FF_FRAMEMASK;
//Fab : 02-08-98: 'skin' override spritedef currently used for skin
if (thing->skin && thing->sprite == SPR_PLAY)
sprdef = &((skin_t *)thing->skin)->spritedef;
else
sprdef = &sprites[thing->sprite];
if (rot >= sprdef->numframes)
{
Base draw distances on viewx/viewy coordinates, NOT the player object's coordinates (this can cause problems with things like skyboxes for instance). Splitscreen's player 2 should not affect what sprites player 1 can see, and vice versa! Especially not for precipitation, that just looks ridiculous. git-svn-id: https://code.orospakr.ca/svn/srb2/trunk@9041 6de4a73c-47e2-0310-b8c1-93d6ecd3f8cd
2015-02-15 17:15:55 +00:00
CONS_Alert(CONS_ERROR, M_GetText("HWR_ProjectSprite: invalid sprite frame %s/%s for %s\n"),
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
sizeu1(rot), sizeu2(sprdef->numframes), sprnames[thing->sprite]);
thing->sprite = states[S_UNKNOWN].sprite;
thing->frame = states[S_UNKNOWN].frame;
I was helping yellow debug a crash and we found out this thing never actually fixed the crash like it was supposed to so it's fixed now -3- git-svn-id: https://code.orospakr.ca/svn/srb2/trunk@9044 6de4a73c-47e2-0310-b8c1-93d6ecd3f8cd
2015-02-18 23:42:01 +00:00
sprdef = &sprites[thing->sprite];
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
rot = thing->frame&FF_FRAMEMASK;
thing->state->sprite = thing->sprite;
thing->state->frame = thing->frame;
}
sprframe = &sprdef->spriteframes[rot];
#ifdef PARANOIA
if (!sprframe)
I_Error("sprframes NULL for sprite %d\n", thing->sprite);
#endif
if (sprframe->rotate)
{
// choose a different rotation based on player view
ang = R_PointToAngle(thing->x, thing->y); // uses viewx,viewy
rot = (ang-thing->angle+ANGLE_202h)>>29;
//Fab: lumpid is the index for spritewidth,spriteoffset... tables
lumpoff = sprframe->lumpid[rot];
flip = sprframe->flip & (1<<rot);
}
else
{
// use single rotation for all views
rot = 0; //Fab: for vis->patch below
lumpoff = sprframe->lumpid[0]; //Fab: see note above
flip = sprframe->flip; // Will only be 0x00 or 0xFF
}
if (thing->skin && ((skin_t *)thing->skin)->flags & SF_HIRES)
this_scale = this_scale * FIXED_TO_FLOAT(((skin_t *)thing->skin)->highresscale);
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
rightsin = FIXED_TO_FLOAT(FINESINE((viewangle + ANGLE_90)>>ANGLETOFINESHIFT));
rightcos = FIXED_TO_FLOAT(FINECOSINE((viewangle + ANGLE_90)>>ANGLETOFINESHIFT));
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (flip)
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
{
x1 = (FIXED_TO_FLOAT(spritecachedinfo[lumpoff].width - spritecachedinfo[lumpoff].offset) * this_scale);
x2 = (FIXED_TO_FLOAT(spritecachedinfo[lumpoff].offset) * this_scale);
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
else
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
{
x1 = (FIXED_TO_FLOAT(spritecachedinfo[lumpoff].offset) * this_scale);
x2 = (FIXED_TO_FLOAT(spritecachedinfo[lumpoff].width - spritecachedinfo[lumpoff].offset) * this_scale);
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
z1 = tr_y + x1 * rightsin;
z2 = tr_y - x2 * rightsin;
x1 = tr_x + x1 * rightcos;
x2 = tr_x - x2 * rightcos;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (thing->eflags & MFE_VERTICALFLIP)
{
gz = FIXED_TO_FLOAT(thing->z+thing->height) - FIXED_TO_FLOAT(spritecachedinfo[lumpoff].topoffset) * this_scale;
gzt = gz + FIXED_TO_FLOAT(spritecachedinfo[lumpoff].height) * this_scale;
}
else
{
gzt = FIXED_TO_FLOAT(thing->z) + FIXED_TO_FLOAT(spritecachedinfo[lumpoff].topoffset) * this_scale;
gz = gzt - FIXED_TO_FLOAT(spritecachedinfo[lumpoff].height) * this_scale;
}
if (thing->subsector->sector->cullheight)
{
SRB2 2.1.12 release
2014-11-12 00:55:07 +00:00
if (HWR_DoCulling(thing->subsector->sector->cullheight, viewsector->cullheight, gr_viewz, gz, gzt))
return;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
heightsec = thing->subsector->sector->heightsec;
Fix HWR_ProjectSprite to check properly whether the displayed player's mobj or its subsector exists, to avoid a crash when checking for fake planes. (also use viewplayer since its available to use, silly hardware code) Also tweaked a weird splitscreen check in HWR_DrawSpriteShadow; still investigating whether stplyr is ever not player 2 when it's player 2's view, but this looks better for now
2018-06-23 17:47:32 +00:00
if (viewplayer->mo && viewplayer->mo->subsector)
phs = viewplayer->mo->subsector->sector->heightsec;
else
phs = -1;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (heightsec != -1 && phs != -1) // only clip things which are in special sectors
{
if (gr_viewz < FIXED_TO_FLOAT(sectors[phs].floorheight) ?
FIXED_TO_FLOAT(thing->z) >= FIXED_TO_FLOAT(sectors[heightsec].floorheight) :
gzt < FIXED_TO_FLOAT(sectors[heightsec].floorheight))
return;
if (gr_viewz > FIXED_TO_FLOAT(sectors[phs].ceilingheight) ?
gzt < FIXED_TO_FLOAT(sectors[heightsec].ceilingheight) && gr_viewz >= FIXED_TO_FLOAT(sectors[heightsec].ceilingheight) :
FIXED_TO_FLOAT(thing->z) >= FIXED_TO_FLOAT(sectors[heightsec].ceilingheight))
return;
}
// store information in a vissprite
vis = HWR_NewVisSprite();
vis->x1 = x1;
vis->x2 = x2;
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
vis->z1 = z1;
vis->z2 = z2;
vis->tz = tz; // Keep tz for the simple sprite sorting that happens
dispoffset now works in OpenGL
2015-11-23 17:01:10 +00:00
vis->dispoffset = thing->info->dispoffset; // Monster Iestyn: 23/11/15: HARDWARE SUPPORT AT LAST
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
vis->patchlumpnum = sprframe->lumppat[rot];
vis->flip = flip;
vis->mobj = thing;
//Hurdler: 25/04/2000: now support colormap in hardware mode
if ((vis->mobj->flags & MF_BOSS) && (vis->mobj->flags2 & MF2_FRET) && (leveltime & 1)) // Bosses "flash"
{
if (vis->mobj->type == MT_CYBRAKDEMON)
vis->colormap = R_GetTranslationColormap(TC_ALLWHITE, 0, GTC_CACHE);
else if (vis->mobj->type == MT_METALSONIC_BATTLE)
vis->colormap = R_GetTranslationColormap(TC_METALSONIC, 0, GTC_CACHE);
else
vis->colormap = R_GetTranslationColormap(TC_BOSS, 0, GTC_CACHE);
}
else if (thing->color)
{
// New colormap stuff for skins Tails 06-07-2002
if (thing->skin && thing->sprite == SPR_PLAY) // This thing is a player!
{
size_t skinnum = (skin_t*)thing->skin-skins;
vis->colormap = R_GetTranslationColormap((INT32)skinnum, thing->color, GTC_CACHE);
}
else
vis->colormap = R_GetTranslationColormap(TC_DEFAULT, vis->mobj->color ? vis->mobj->color : SKINCOLOR_CYAN, GTC_CACHE);
}
else
vis->colormap = colormaps;
// set top/bottom coords
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
vis->ty = gzt;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
//CONS_Debug(DBG_RENDER, "------------------\nH: sprite : %d\nH: frame : %x\nH: type : %d\nH: sname : %s\n\n",
// thing->sprite, thing->frame, thing->type, sprnames[thing->sprite]);
if (thing->eflags & MFE_VERTICALFLIP)
vis->vflip = true;
else
vis->vflip = false;
vis->precip = false;
}
#ifdef HWPRECIP
// Precipitation projector for hardware mode
static void HWR_ProjectPrecipitationSprite(precipmobj_t *thing)
{
gr_vissprite_t *vis;
float tr_x, tr_y;
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
float tz;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
float x1, x2;
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
float z1, z2;
float rightsin, rightcos;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
spritedef_t *sprdef;
spriteframe_t *sprframe;
size_t lumpoff;
unsigned rot = 0;
UINT8 flip;
// transform the origin point
tr_x = FIXED_TO_FLOAT(thing->x) - gr_viewx;
tr_y = FIXED_TO_FLOAT(thing->y) - gr_viewy;
// rotation around vertical axis
tz = (tr_x * gr_viewcos) + (tr_y * gr_viewsin);
// thing is behind view plane?
if (tz < ZCLIP_PLANE)
return;
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
tr_x = FIXED_TO_FLOAT(thing->x);
tr_y = FIXED_TO_FLOAT(thing->y);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// decide which patch to use for sprite relative to player
if ((unsigned)thing->sprite >= numsprites)
#ifdef RANGECHECK
I_Error("HWR_ProjectPrecipitationSprite: invalid sprite number %i ",
thing->sprite);
#else
return;
#endif
sprdef = &sprites[thing->sprite];
if ((size_t)(thing->frame&FF_FRAMEMASK) >= sprdef->numframes)
#ifdef RANGECHECK
I_Error("HWR_ProjectPrecipitationSprite: invalid sprite frame %i : %i for %s",
thing->sprite, thing->frame, sprnames[thing->sprite]);
#else
return;
#endif
sprframe = &sprdef->spriteframes[ thing->frame & FF_FRAMEMASK];
// use single rotation for all views
lumpoff = sprframe->lumpid[0];
flip = sprframe->flip; // Will only be 0x00 or 0xFF
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
rightsin = FIXED_TO_FLOAT(FINESINE((viewangle + ANGLE_90)>>ANGLETOFINESHIFT));
rightcos = FIXED_TO_FLOAT(FINECOSINE((viewangle + ANGLE_90)>>ANGLETOFINESHIFT));
if (flip)
{
x1 = FIXED_TO_FLOAT(spritecachedinfo[lumpoff].width - spritecachedinfo[lumpoff].offset);
x2 = FIXED_TO_FLOAT(spritecachedinfo[lumpoff].offset);
}
else
{
x1 = FIXED_TO_FLOAT(spritecachedinfo[lumpoff].offset);
x2 = FIXED_TO_FLOAT(spritecachedinfo[lumpoff].width - spritecachedinfo[lumpoff].offset);
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
z1 = tr_y + x1 * rightsin;
z2 = tr_y - x2 * rightsin;
x1 = tr_x + x1 * rightcos;
x2 = tr_x - x2 * rightcos;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
Weather is already run client-side. What if we ran it render-side, for major performance gains? This commit will answer all your questions - and more!
2018-10-07 14:00:58 +00:00
// okay, we can't return now... this is a hack, but weather isn't networked, so it should be ok
if (!(thing->precipflags & PCF_THUNK))
{
if (thing->precipflags & PCF_RAIN)
P_RainThinker(thing);
else
P_SnowThinker(thing);
thing->precipflags |= PCF_THUNK;
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
//
// store information in a vissprite
//
vis = HWR_NewVisSprite();
vis->x1 = x1;
YUP, I just fixed OGL Precipitation with this branch too. What the fuck? Squashing ancient bugs, woo.
2016-10-27 22:41:24 +00:00
vis->x2 = x2;
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
vis->z1 = z1;
vis->z2 = z2;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
vis->tz = tz;
dispoffset now works in OpenGL
2015-11-23 17:01:10 +00:00
vis->dispoffset = 0; // Monster Iestyn: 23/11/15: HARDWARE SUPPORT AT LAST
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
vis->patchlumpnum = sprframe->lumppat[rot];
vis->flip = flip;
vis->mobj = (mobj_t *)thing;
vis->colormap = colormaps;
// set top/bottom coords
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
vis->ty = FIXED_TO_FLOAT(thing->z + spritecachedinfo[lumpoff].topoffset);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
vis->precip = true;
}
#endif
// ==========================================================================
//
// ==========================================================================
Render the skies correctly on splitscreen in OpenGL mode
2018-11-10 19:19:41 +00:00
static void HWR_DrawSkyBackground(void)
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
FOutVector v[4];
angle_t angle;
OpenGL: Fix sky drawing sky positioning and scaling more correct compared to software.
2014-09-07 23:55:32 +00:00
float dimensionmultiply;
SRB2 2.1.12 release
2014-11-12 00:55:07 +00:00
float aspectratio;
float angleturn;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
HWR_GetTexture(skytexture);
Render the skies correctly on splitscreen in OpenGL mode
2018-11-10 19:19:41 +00:00
aspectratio = (float)vid.width/(float)vid.height;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
//Hurdler: the sky is the only texture who need 4.0f instead of 1.0
// because it's called just after clearing the screen
// and thus, the near clipping plane is set to 3.99
Decrease far clipping plane The Far clipping plane did not need to be nearly as high as it was, the new value is 32768, which I suspect is about how far software can render before it completely falls apart. It is desirable to increase the near clipping plane to between 6-10, but it can introduce more issues with close geometry not being drawn when the player or camera is scaled or viewheight is set to MIN in first person view. It would also stop sprites from being drawn ever so slightly too early, but this isn't too much of an issue and isn't too noticeable with those values. Might look into scaling near clipping plane in accordance to camera scale in the future. The reason for wanting to increase the near clipping plane is because the small value can cause very noticeable Z-fighting where there shouldn't be on older GPU's, usually Intel ones, that don't support 24-bits for the depth buffer.
2018-03-06 03:48:15 +00:00
// Sryder: Just use the near clipping plane value then
Render the skies correctly on splitscreen in OpenGL mode
2018-11-10 19:19:41 +00:00
// 3--2
// | /|
// |/ |
// 0--1
Decrease far clipping plane The Far clipping plane did not need to be nearly as high as it was, the new value is 32768, which I suspect is about how far software can render before it completely falls apart. It is desirable to increase the near clipping plane to between 6-10, but it can introduce more issues with close geometry not being drawn when the player or camera is scaled or viewheight is set to MIN in first person view. It would also stop sprites from being drawn ever so slightly too early, but this isn't too much of an issue and isn't too noticeable with those values. Might look into scaling near clipping plane in accordance to camera scale in the future. The reason for wanting to increase the near clipping plane is because the small value can cause very noticeable Z-fighting where there shouldn't be on older GPU's, usually Intel ones, that don't support 24-bits for the depth buffer.
2018-03-06 03:48:15 +00:00
v[0].x = v[3].x = -ZCLIP_PLANE-1;
v[1].x = v[2].x = ZCLIP_PLANE+1;
v[0].y = v[1].y = -ZCLIP_PLANE-1;
v[2].y = v[3].y = ZCLIP_PLANE+1;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
Decrease far clipping plane The Far clipping plane did not need to be nearly as high as it was, the new value is 32768, which I suspect is about how far software can render before it completely falls apart. It is desirable to increase the near clipping plane to between 6-10, but it can introduce more issues with close geometry not being drawn when the player or camera is scaled or viewheight is set to MIN in first person view. It would also stop sprites from being drawn ever so slightly too early, but this isn't too much of an issue and isn't too noticeable with those values. Might look into scaling near clipping plane in accordance to camera scale in the future. The reason for wanting to increase the near clipping plane is because the small value can cause very noticeable Z-fighting where there shouldn't be on older GPU's, usually Intel ones, that don't support 24-bits for the depth buffer.
2018-03-06 03:48:15 +00:00
v[0].z = v[1].z = v[2].z = v[3].z = ZCLIP_PLANE+1;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
OpenGL Skybox OpenGL skyboxes. Slight change to sky vertical offsetting.
2014-03-20 20:13:15 +00:00
// X
OpenGL: Fix sky drawing sky positioning and scaling more correct compared to software.
2014-09-07 23:55:32 +00:00
// NOTE: This doesn't work right with texture widths greater than 1024
// software doesn't draw any further than 1024 for skies anyway, but this doesn't overlap properly
// The only time this will probably be an issue is when a sky wider than 1024 is used as a sky AND a regular wall texture
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
OpenGL: Fix sky drawing sky positioning and scaling more correct compared to software.
2014-09-07 23:55:32 +00:00
angle = (dup_viewangle + gr_xtoviewangle[0]);
dimensionmultiply = ((float)textures[skytexture]->width/256.0f);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
OpenGL: Fix sky drawing sky positioning and scaling more correct compared to software.
2014-09-07 23:55:32 +00:00
v[0].sow = v[3].sow = ((float) angle / ((ANGLE_90-1)*dimensionmultiply));
v[2].sow = v[1].sow = (-1.0f/dimensionmultiply)+((float) angle / ((ANGLE_90-1)*dimensionmultiply));
OpenGL Skybox OpenGL skyboxes. Slight change to sky vertical offsetting.
2014-03-20 20:13:15 +00:00
// Y
OpenGL: Fix sky drawing sky positioning and scaling more correct compared to software.
2014-09-07 23:55:32 +00:00
angle = aimingangle;
dimensionmultiply = ((float)textures[skytexture]->height/(128.0f*aspectratio));
Render the skies correctly on splitscreen in OpenGL mode
2018-11-10 19:19:41 +00:00
if (splitscreen)
{
dimensionmultiply *= 2;
angle *= 2;
}
OpenGL: Fix sky drawing sky positioning and scaling more correct compared to software.
2014-09-07 23:55:32 +00:00
// Middle of the sky should always be at angle 0
// need to keep correct aspect ratio with X
if (atransform.flip)
{
// During vertical flip the sky should be flipped and it's y movement should also be flipped obviously
v[3].tow = v[2].tow = -(0.5f-(0.5f/dimensionmultiply));
v[0].tow = v[1].tow = (-1.0f/dimensionmultiply)-(0.5f-(0.5f/dimensionmultiply));
}
OpenGL Skybox OpenGL skyboxes. Slight change to sky vertical offsetting.
2014-03-20 20:13:15 +00:00
else
OpenGL: Fix sky drawing sky positioning and scaling more correct compared to software.
2014-09-07 23:55:32 +00:00
{
v[3].tow = v[2].tow = (-1.0f/dimensionmultiply)-(0.5f-(0.5f/dimensionmultiply));
v[0].tow = v[1].tow = -(0.5f-(0.5f/dimensionmultiply));
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
Render the skies correctly on splitscreen in OpenGL mode
2018-11-10 19:19:41 +00:00
angleturn = (((float)ANGLE_45-1.0f)*aspectratio)*dimensionmultiply;
OpenGL: Fix sky drawing sky positioning and scaling more correct compared to software.
2014-09-07 23:55:32 +00:00
if (angle > ANGLE_180) // Do this because we don't want the sky to suddenly teleport when crossing over 0 to 360 and vice versa
{
angle = InvAngle(angle);
v[3].tow = v[2].tow += ((float) angle / angleturn);
v[0].tow = v[1].tow += ((float) angle / angleturn);
}
else
{
v[3].tow = v[2].tow -= ((float) angle / angleturn);
v[0].tow = v[1].tow -= ((float) angle / angleturn);
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
HWD.pfnDrawPolygon(NULL, v, 4, 0);
}
// -----------------+
// HWR_ClearView : clear the viewwindow, with maximum z value
// -----------------+
static inline void HWR_ClearView(void)
{
// 3--2
// | /|
// |/ |
// 0--1
/// \bug faB - enable depth mask, disable color mask
HWD.pfnGClipRect((INT32)gr_viewwindowx,
(INT32)gr_viewwindowy,
(INT32)(gr_viewwindowx + gr_viewwidth),
(INT32)(gr_viewwindowy + gr_viewheight),
Decrease far clipping plane The Far clipping plane did not need to be nearly as high as it was, the new value is 32768, which I suspect is about how far software can render before it completely falls apart. It is desirable to increase the near clipping plane to between 6-10, but it can introduce more issues with close geometry not being drawn when the player or camera is scaled or viewheight is set to MIN in first person view. It would also stop sprites from being drawn ever so slightly too early, but this isn't too much of an issue and isn't too noticeable with those values. Might look into scaling near clipping plane in accordance to camera scale in the future. The reason for wanting to increase the near clipping plane is because the small value can cause very noticeable Z-fighting where there shouldn't be on older GPU's, usually Intel ones, that don't support 24-bits for the depth buffer.
2018-03-06 03:48:15 +00:00
ZCLIP_PLANE);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
HWD.pfnClearBuffer(false, true, 0);
//disable clip window - set to full size
// rem by Hurdler
// HWD.pfnGClipRect(0, 0, vid.width, vid.height);
}
// -----------------+
// HWR_SetViewSize : set projection and scaling values
// -----------------+
void HWR_SetViewSize(void)
{
// setup view size
gr_viewwidth = (float)vid.width;
gr_viewheight = (float)vid.height;
if (splitscreen)
gr_viewheight /= 2;
gr_centerx = gr_viewwidth / 2;
gr_basecentery = gr_viewheight / 2; //note: this is (gr_centerx * gr_viewheight / gr_viewwidth)
gr_viewwindowx = (vid.width - gr_viewwidth) / 2;
gr_windowcenterx = (float)(vid.width / 2);
if (gr_viewwidth == vid.width)
{
gr_baseviewwindowy = 0;
gr_basewindowcentery = gr_viewheight / 2; // window top left corner at 0,0
}
else
{
gr_baseviewwindowy = (vid.height-gr_viewheight) / 2;
gr_basewindowcentery = (float)(vid.height / 2);
}
gr_pspritexscale = gr_viewwidth / BASEVIDWIDTH;
gr_pspriteyscale = ((vid.height*gr_pspritexscale*BASEVIDWIDTH)/BASEVIDHEIGHT)/vid.width;
Optimise the screen texture setup for SDL2, Post-processor, and wipes. Only use glCopyTexImage2D when first creating the screen texture, use glCopyTexSubImage2D anytime after that as it does not define a new texture each time. Flushing of the screen textures has been implemented for when the screen size changes (so that the screen textures don't stay at a wrong size) and the game is closed, I believe they would leave a memory leak before.
2018-03-07 05:19:06 +00:00
HWD.pfnFlushScreenTextures();
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
OpenGL Skybox OpenGL skyboxes. Slight change to sky vertical offsetting.
2014-03-20 20:13:15 +00:00
// ==========================================================================
// Same as rendering the player view, but from the skybox object
// ==========================================================================
void HWR_RenderSkyboxView(INT32 viewnumber, player_t *player)
{
const float fpov = FIXED_TO_FLOAT(cv_grfov.value+player->fovadd);
SRB2 2.1.11 release
2014-08-27 03:56:30 +00:00
postimg_t *type;
if (splitscreen && player == &players[secondarydisplayplayer])
type = &postimgtype2;
else
type = &postimgtype;
OpenGL Skybox OpenGL skyboxes. Slight change to sky vertical offsetting.
2014-03-20 20:13:15 +00:00
{
// do we really need to save player (is it not the same)?
player_t *saved_player = stplyr;
stplyr = player;
ST_doPaletteStuff();
stplyr = saved_player;
#ifdef ALAM_LIGHTING
HWR_SetLights(viewnumber);
#endif
}
// note: sets viewangle, viewx, viewy, viewz
R_SkyboxFrame(player);
// copy view cam position for local use
dup_viewx = viewx;
dup_viewy = viewy;
dup_viewz = viewz;
dup_viewangle = viewangle;
// set window position
gr_centery = gr_basecentery;
gr_viewwindowy = gr_baseviewwindowy;
gr_windowcentery = gr_basewindowcentery;
if (splitscreen && viewnumber == 1)
{
gr_viewwindowy += (vid.height/2);
gr_windowcentery += (vid.height/2);
}
// check for new console commands.
NetUpdate();
gr_viewx = FIXED_TO_FLOAT(dup_viewx);
gr_viewy = FIXED_TO_FLOAT(dup_viewy);
gr_viewz = FIXED_TO_FLOAT(dup_viewz);
gr_viewsin = FIXED_TO_FLOAT(viewsin);
gr_viewcos = FIXED_TO_FLOAT(viewcos);
gr_viewludsin = FIXED_TO_FLOAT(FINECOSINE(aimingangle>>ANGLETOFINESHIFT));
gr_viewludcos = FIXED_TO_FLOAT(-FINESINE(aimingangle>>ANGLETOFINESHIFT));
//04/01/2000: Hurdler: added for T&L
// It should replace all other gr_viewxxx when finished
atransform.anglex = (float)(aimingangle>>ANGLETOFINESHIFT)*(360.0f/(float)FINEANGLES);
atransform.angley = (float)(viewangle>>ANGLETOFINESHIFT)*(360.0f/(float)FINEANGLES);
OpenGL Screen Flip Screen flipping in OpenGL is no longer done by inverting the screen polygons.
2014-04-07 16:08:04 +00:00
SRB2 2.1.11 release
2014-08-27 03:56:30 +00:00
if (*type == postimg_flip)
OpenGL Screen Flip Screen flipping in OpenGL is no longer done by inverting the screen polygons.
2014-04-07 16:08:04 +00:00
atransform.flip = true;
else
atransform.flip = false;
OpenGL Skybox OpenGL skyboxes. Slight change to sky vertical offsetting.
2014-03-20 20:13:15 +00:00
atransform.x = gr_viewx; // FIXED_TO_FLOAT(viewx)
atransform.y = gr_viewy; // FIXED_TO_FLOAT(viewy)
atransform.z = gr_viewz; // FIXED_TO_FLOAT(viewz)
atransform.scalex = 1;
Stop squashing the screen vertically in non-green resolutions
2018-03-22 01:10:53 +00:00
atransform.scaley = (float)vid.width/vid.height;
OpenGL Skybox OpenGL skyboxes. Slight change to sky vertical offsetting.
2014-03-20 20:13:15 +00:00
atransform.scalez = 1;
atransform.fovxangle = fpov; // Tails
atransform.fovyangle = fpov; // Tails
atransform.splitscreen = splitscreen;
gr_fovlud = (float)(1.0l/tan((double)(fpov*M_PIl/360l)));
//------------------------------------------------------------------------
HWR_ClearView();
if (0)
{ // I don't think this is ever used.
if (cv_grfog.value)
HWR_FoggingOn(); // First of all, turn it on, set the default user settings too
else
HWD.pfnSetSpecialState(HWD_SET_FOG_MODE, 0); // Turn it off
}
if (drawsky)
Render the skies correctly on splitscreen in OpenGL mode
2018-11-10 19:19:41 +00:00
HWR_DrawSkyBackground();
OpenGL Skybox OpenGL skyboxes. Slight change to sky vertical offsetting.
2014-03-20 20:13:15 +00:00
//Hurdler: it doesn't work in splitscreen mode
drawsky = splitscreen;
HWR_ClearSprites();
#ifdef SORTING
drawcount = 0;
#endif
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#ifdef NEWCLIP
if (rendermode == render_opengl)
{
angle_t a1 = gld_FrustumAngle();
gld_clipper_Clear();
gld_clipper_SafeAddClipRange(viewangle + a1, viewangle - a1);
#ifdef HAVE_SPHEREFRUSTRUM
gld_FrustrumSetup();
#endif
}
#else
OpenGL Skybox OpenGL skyboxes. Slight change to sky vertical offsetting.
2014-03-20 20:13:15 +00:00
HWR_ClearClipSegs();
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#endif
OpenGL Skybox OpenGL skyboxes. Slight change to sky vertical offsetting.
2014-03-20 20:13:15 +00:00
//04/01/2000: Hurdler: added for T&L
// Actually it only works on Walls and Planes
HWD.pfnSetTransform(&atransform);
validcount++;
HWR_RenderBSPNode((INT32)numnodes-1);
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#ifndef NEWCLIP
OpenGL Skybox OpenGL skyboxes. Slight change to sky vertical offsetting.
2014-03-20 20:13:15 +00:00
// Make a viewangle int so we can render things based on mouselook
if (player == &players[consoleplayer])
viewangle = localaiming;
else if (splitscreen && player == &players[secondarydisplayplayer])
viewangle = localaiming2;
// Handle stuff when you are looking farther up or down.
if ((aimingangle || cv_grfov.value+player->fovadd > 90*FRACUNIT))
{
dup_viewangle += ANGLE_90;
HWR_ClearClipSegs();
HWR_RenderBSPNode((INT32)numnodes-1); //left
dup_viewangle += ANGLE_90;
if (((INT32)aimingangle > ANGLE_45 || (INT32)aimingangle<-ANGLE_45))
{
HWR_ClearClipSegs();
HWR_RenderBSPNode((INT32)numnodes-1); //back
}
dup_viewangle += ANGLE_90;
HWR_ClearClipSegs();
HWR_RenderBSPNode((INT32)numnodes-1); //right
dup_viewangle += ANGLE_90;
}
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#endif
OpenGL Skybox OpenGL skyboxes. Slight change to sky vertical offsetting.
2014-03-20 20:13:15 +00:00
// Check for new console commands.
NetUpdate();
#ifdef ALAM_LIGHTING
//14/11/99: Hurdler: moved here because it doesn't work with
// subsector, see other comments;
HWR_ResetLights();
#endif
// Draw MD2 and sprites
#ifdef SORTING
HWR_SortVisSprites();
#endif
#ifdef SORTING
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
HWR_DrawSprites();
OpenGL Skybox OpenGL skyboxes. Slight change to sky vertical offsetting.
2014-03-20 20:13:15 +00:00
#endif
#ifdef NEWCORONAS
//Hurdler: they must be drawn before translucent planes, what about gl fog?
HWR_DrawCoronas();
#endif
#ifdef SORTING
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
if (numplanes || numpolyplanes || numwalls) //Hurdler: render 3D water and transparent walls after everything
OpenGL Skybox OpenGL skyboxes. Slight change to sky vertical offsetting.
2014-03-20 20:13:15 +00:00
{
HWR_CreateDrawNodes();
}
#else
if (numfloors || numwalls)
{
HWD.pfnSetTransform(&atransform);
if (numfloors)
HWR_Render3DWater();
if (numwalls)
HWR_RenderTransparentWalls();
}
#endif
HWD.pfnSetTransform(NULL);
// put it off for menus etc
if (cv_grfog.value)
HWD.pfnSetSpecialState(HWD_SET_FOG_MODE, 0);
// Check for new console commands.
NetUpdate();
// added by Hurdler for correct splitscreen
// moved here by hurdler so it works with the new near clipping plane
HWD.pfnGClipRect(0, 0, vid.width, vid.height, NZCLIP_PLANE);
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// ==========================================================================
//
// ==========================================================================
void HWR_RenderPlayerView(INT32 viewnumber, player_t *player)
{
const float fpov = FIXED_TO_FLOAT(cv_grfov.value+player->fovadd);
SRB2 2.1.11 release
2014-08-27 03:56:30 +00:00
postimg_t *type;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
OpenGL Skybox OpenGL skyboxes. Slight change to sky vertical offsetting.
2014-03-20 20:13:15 +00:00
const boolean skybox = (skyboxmo[0] && cv_skybox.value); // True if there's a skybox object and skyboxes are on
OpenGL remove HOMs Removes HOMs from OpenGL. This also seems to have made skyboxes not draw ontop of everything else on Intel GPU's.
2014-03-22 17:17:40 +00:00
FRGBAFloat ClearColor;
SRB2 2.1.11 release
2014-08-27 03:56:30 +00:00
if (splitscreen && player == &players[secondarydisplayplayer])
type = &postimgtype2;
else
type = &postimgtype;
OpenGL remove HOMs Removes HOMs from OpenGL. This also seems to have made skyboxes not draw ontop of everything else on Intel GPU's.
2014-03-22 17:17:40 +00:00
ClearColor.red = 0.0f;
ClearColor.green = 0.0f;
ClearColor.blue = 0.0f;
ClearColor.alpha = 1.0f;
OpenGL Splitscreen Fix First screen in split-screen is now not completely black.
2014-03-30 23:07:55 +00:00
if (viewnumber == 0) // Only do it if it's the first screen being rendered
HWD.pfnClearBuffer(true, false, &ClearColor); // Clear the Color Buffer, stops HOMs. Also seems to fix the skybox issue on Intel GPUs.
OpenGL remove HOMs Removes HOMs from OpenGL. This also seems to have made skyboxes not draw ontop of everything else on Intel GPU's.
2014-03-22 17:17:40 +00:00
OpenGL Skybox OpenGL skyboxes. Slight change to sky vertical offsetting.
2014-03-20 20:13:15 +00:00
if (skybox && drawsky) // If there's a skybox and we should be drawing the sky, draw the skybox
HWR_RenderSkyboxView(viewnumber, player); // This is drawn before everything else so it is placed behind
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
// do we really need to save player (is it not the same)?
player_t *saved_player = stplyr;
stplyr = player;
ST_doPaletteStuff();
stplyr = saved_player;
#ifdef ALAM_LIGHTING
HWR_SetLights(viewnumber);
#endif
}
// note: sets viewangle, viewx, viewy, viewz
OpenGL Skybox OpenGL skyboxes. Slight change to sky vertical offsetting.
2014-03-20 20:13:15 +00:00
R_SetupFrame(player, false); // This can stay false because it is only used to set viewsky in r_main.c, which isn't used here
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// copy view cam position for local use
dup_viewx = viewx;
dup_viewy = viewy;
dup_viewz = viewz;
dup_viewangle = viewangle;
// set window position
gr_centery = gr_basecentery;
gr_viewwindowy = gr_baseviewwindowy;
gr_windowcentery = gr_basewindowcentery;
if (splitscreen && viewnumber == 1)
{
gr_viewwindowy += (vid.height/2);
gr_windowcentery += (vid.height/2);
}
// check for new console commands.
NetUpdate();
gr_viewx = FIXED_TO_FLOAT(dup_viewx);
gr_viewy = FIXED_TO_FLOAT(dup_viewy);
gr_viewz = FIXED_TO_FLOAT(dup_viewz);
gr_viewsin = FIXED_TO_FLOAT(viewsin);
gr_viewcos = FIXED_TO_FLOAT(viewcos);
gr_viewludsin = FIXED_TO_FLOAT(FINECOSINE(aimingangle>>ANGLETOFINESHIFT));
gr_viewludcos = FIXED_TO_FLOAT(-FINESINE(aimingangle>>ANGLETOFINESHIFT));
//04/01/2000: Hurdler: added for T&L
// It should replace all other gr_viewxxx when finished
atransform.anglex = (float)(aimingangle>>ANGLETOFINESHIFT)*(360.0f/(float)FINEANGLES);
atransform.angley = (float)(viewangle>>ANGLETOFINESHIFT)*(360.0f/(float)FINEANGLES);
OpenGL Screen Flip Screen flipping in OpenGL is no longer done by inverting the screen polygons.
2014-04-07 16:08:04 +00:00
SRB2 2.1.11 release
2014-08-27 03:56:30 +00:00
if (*type == postimg_flip)
OpenGL Screen Flip Screen flipping in OpenGL is no longer done by inverting the screen polygons.
2014-04-07 16:08:04 +00:00
atransform.flip = true;
else
atransform.flip = false;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
atransform.x = gr_viewx; // FIXED_TO_FLOAT(viewx)
atransform.y = gr_viewy; // FIXED_TO_FLOAT(viewy)
atransform.z = gr_viewz; // FIXED_TO_FLOAT(viewz)
atransform.scalex = 1;
Stop squashing the screen vertically in non-green resolutions
2018-03-22 01:10:53 +00:00
atransform.scaley = (float)vid.width/vid.height;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
atransform.scalez = 1;
atransform.fovxangle = fpov; // Tails
atransform.fovyangle = fpov; // Tails
atransform.splitscreen = splitscreen;
gr_fovlud = (float)(1.0l/tan((double)(fpov*M_PIl/360l)));
//------------------------------------------------------------------------
OpenGL remove HOMs Removes HOMs from OpenGL. This also seems to have made skyboxes not draw ontop of everything else on Intel GPU's.
2014-03-22 17:17:40 +00:00
HWR_ClearView(); // Clears the depth buffer and resets the view I believe
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
if (0)
{ // I don't think this is ever used.
if (cv_grfog.value)
HWR_FoggingOn(); // First of all, turn it on, set the default user settings too
else
HWD.pfnSetSpecialState(HWD_SET_FOG_MODE, 0); // Turn it off
}
OpenGL Skybox OpenGL skyboxes. Slight change to sky vertical offsetting.
2014-03-20 20:13:15 +00:00
if (!skybox && drawsky) // Don't draw the regular sky if there's a skybox
Render the skies correctly on splitscreen in OpenGL mode
2018-11-10 19:19:41 +00:00
HWR_DrawSkyBackground();
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
//Hurdler: it doesn't work in splitscreen mode
drawsky = splitscreen;
HWR_ClearSprites();
#ifdef SORTING
drawcount = 0;
#endif
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#ifdef NEWCLIP
if (rendermode == render_opengl)
{
angle_t a1 = gld_FrustumAngle();
gld_clipper_Clear();
gld_clipper_SafeAddClipRange(viewangle + a1, viewangle - a1);
#ifdef HAVE_SPHEREFRUSTRUM
gld_FrustrumSetup();
#endif
}
#else
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
HWR_ClearClipSegs();
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
//04/01/2000: Hurdler: added for T&L
// Actually it only works on Walls and Planes
HWD.pfnSetTransform(&atransform);
validcount++;
HWR_RenderBSPNode((INT32)numnodes-1);
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#ifndef NEWCLIP
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// Make a viewangle int so we can render things based on mouselook
if (player == &players[consoleplayer])
viewangle = localaiming;
else if (splitscreen && player == &players[secondarydisplayplayer])
viewangle = localaiming2;
// Handle stuff when you are looking farther up or down.
if ((aimingangle || cv_grfov.value+player->fovadd > 90*FRACUNIT))
{
dup_viewangle += ANGLE_90;
HWR_ClearClipSegs();
HWR_RenderBSPNode((INT32)numnodes-1); //left
dup_viewangle += ANGLE_90;
if (((INT32)aimingangle > ANGLE_45 || (INT32)aimingangle<-ANGLE_45))
{
HWR_ClearClipSegs();
HWR_RenderBSPNode((INT32)numnodes-1); //back
}
dup_viewangle += ANGLE_90;
HWR_ClearClipSegs();
HWR_RenderBSPNode((INT32)numnodes-1); //right
dup_viewangle += ANGLE_90;
}
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// Check for new console commands.
NetUpdate();
#ifdef ALAM_LIGHTING
//14/11/99: Hurdler: moved here because it doesn't work with
// subsector, see other comments;
HWR_ResetLights();
#endif
// Draw MD2 and sprites
#ifdef SORTING
HWR_SortVisSprites();
#endif
#ifdef SORTING
Transform sprites in world space rather than screen space Transformation based on screen space would make sense if we didn't want anything in the world to effect the sprites. This should allow sprite splitting and sorting of sprites with level geometry easier. stransform is no longer needed.
2018-03-21 19:45:37 +00:00
HWR_DrawSprites();
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
#endif
#ifdef NEWCORONAS
//Hurdler: they must be drawn before translucent planes, what about gl fog?
HWR_DrawCoronas();
#endif
#ifdef SORTING
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
if (numplanes || numpolyplanes || numwalls) //Hurdler: render 3D water and transparent walls after everything
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
HWR_CreateDrawNodes();
}
#else
OpenGL: Polyobject Planes
2014-10-27 20:57:45 +00:00
if (numfloors || numpolyplanes || numwalls)
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
HWD.pfnSetTransform(&atransform);
if (numfloors)
HWR_Render3DWater();
if (numwalls)
HWR_RenderTransparentWalls();
}
#endif
HWD.pfnSetTransform(NULL);
// put it off for menus etc
if (cv_grfog.value)
HWD.pfnSetSpecialState(HWD_SET_FOG_MODE, 0);
HWR_DoPostProcessor(player);
// Check for new console commands.
NetUpdate();
// added by Hurdler for correct splitscreen
// moved here by hurdler so it works with the new near clipping plane
HWD.pfnGClipRect(0, 0, vid.width, vid.height, NZCLIP_PLANE);
}
// ==========================================================================
// FOG
// ==========================================================================
/// \author faB
static UINT32 atohex(const char *s)
{
INT32 iCol;
const char *sCol;
char cCol;
INT32 i;
if (strlen(s)<6)
return 0;
iCol = 0;
sCol = s;
for (i = 0; i < 6; i++, sCol++)
{
iCol <<= 4;
cCol = *sCol;
if (cCol >= '0' && cCol <= '9')
iCol |= cCol - '0';
else
{
if (cCol >= 'F')
cCol -= 'a' - 'A';
if (cCol >= 'A' && cCol <= 'F')
iCol = iCol | (cCol - 'A' + 10);
}
}
//CONS_Debug(DBG_RENDER, "col %x\n", iCol);
return iCol;
}
static void HWR_FoggingOn(void)
{
HWD.pfnSetSpecialState(HWD_SET_FOG_COLOR, atohex(cv_grfogcolor.string));
HWD.pfnSetSpecialState(HWD_SET_FOG_DENSITY, cv_grfogdensity.value);
HWD.pfnSetSpecialState(HWD_SET_FOG_MODE, 1);
}
// ==========================================================================
// 3D ENGINE COMMANDS
// ==========================================================================
static void CV_grFov_OnChange(void)
{
if ((netgame || multiplayer) && !cv_debug && cv_grfov.value != 90*FRACUNIT)
CV_Set(&cv_grfov, cv_grfov.defaultvalue);
}
static void Command_GrStats_f(void)
{
Z_CheckHeap(9875); // debug
CONS_Printf(M_GetText("Patch info headers: %7s kb\n"), sizeu1(Z_TagUsage(PU_HWRPATCHINFO)>>10));
CONS_Printf(M_GetText("3D Texture cache : %7s kb\n"), sizeu1(Z_TagUsage(PU_HWRCACHE)>>10));
CONS_Printf(M_GetText("Plane polygon : %7s kb\n"), sizeu1(Z_TagUsage(PU_HWRPLANE)>>10));
}
// **************************************************************************
// 3D ENGINE SETUP
// **************************************************************************
// --------------------------------------------------------------------------
// Add hardware engine commands & consvars
// --------------------------------------------------------------------------
//added by Hurdler: console varibale that are saved
void HWR_AddCommands(void)
{
CV_RegisterVar(&cv_grrounddown);
CV_RegisterVar(&cv_grfov);
CV_RegisterVar(&cv_grfogdensity);
CV_RegisterVar(&cv_grfiltermode);
CV_RegisterVar(&cv_granisotropicmode);
CV_RegisterVar(&cv_grcorrecttricks);
CV_RegisterVar(&cv_grsolvetjoin);
}
static inline void HWR_AddEngineCommands(void)
{
// engine state variables
//CV_RegisterVar(&cv_grzbuffer);
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#ifndef NEWCLIP
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
CV_RegisterVar(&cv_grclipwalls);
More progress, NEWCLIP added to doomdef.h, sadly it actually all lags the game so I've disabled it for now Other notes: * on second thought I'll keep the hw_clip functions' gld prefixes rather than HWR, not like it matters either way * despite the extra lag it does fix the issues with translucent walls and such when displayed at different vertical angles, such as with the GFZ1 waterfall
2017-01-10 18:01:03 +00:00
#endif
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// engine development mode variables
// - usage may vary from version to version..
CV_RegisterVar(&cv_gralpha);
CV_RegisterVar(&cv_grbeta);
// engine commands
COM_AddCommand("gr_stats", Command_GrStats_f);
}
// --------------------------------------------------------------------------
// Setup the hardware renderer
// --------------------------------------------------------------------------
void HWR_Startup(void)
{
static boolean startupdone = false;
// setup GLPatch_t scaling
gr_patch_scalex = (float)(1.0f / vid.width);
gr_patch_scaley = (float)(1.0f / vid.height);
// initalze light lut translation
InitLumLut();
// do this once
if (!startupdone)
{
CONS_Printf("HWR_Startup()\n");
HWR_InitPolyPool();
// add console cmds & vars
HWR_AddEngineCommands();
HWR_InitTextureCache();
HWR_InitMD2();
#ifdef ALAM_LIGHTING
HWR_InitLight();
#endif
}
if (rendermode == render_opengl)
textureformat = patchformat =
#ifdef _NDS
GR_TEXFMT_P_8;
#else
GR_RGBA;
#endif
startupdone = true;
}
// --------------------------------------------------------------------------
// Free resources allocated by the hardware renderer
// --------------------------------------------------------------------------
void HWR_Shutdown(void)
{
CONS_Printf("HWR_Shutdown()\n");
HWR_FreeExtraSubsectors();
HWR_FreePolyPool();
HWR_FreeTextureCache();
Optimise the screen texture setup for SDL2, Post-processor, and wipes. Only use glCopyTexImage2D when first creating the screen texture, use glCopyTexSubImage2D anytime after that as it does not define a new texture each time. Flushing of the screen textures has been implemented for when the screen size changes (so that the screen textures don't stay at a wrong size) and the game is closed, I believe they would leave a memory leak before.
2018-03-07 05:19:06 +00:00
HWD.pfnFlushScreenTextures();
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
void transform(float *cx, float *cy, float *cz)
{
float tr_x,tr_y;
// translation
tr_x = *cx - gr_viewx;
tr_y = *cz - gr_viewy;
// *cy = *cy;
// rotation around vertical y axis
*cx = (tr_x * gr_viewsin) - (tr_y * gr_viewcos);
tr_x = (tr_x * gr_viewcos) + (tr_y * gr_viewsin);
//look up/down ----TOTAL SUCKS!!!--- do the 2 in one!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
tr_y = *cy - gr_viewz;
*cy = (tr_x * gr_viewludcos) + (tr_y * gr_viewludsin);
*cz = (tr_x * gr_viewludsin) - (tr_y * gr_viewludcos);
//scale y before frustum so that frustum can be scaled to screen height
*cy *= ORIGINAL_ASPECT * gr_fovlud;
*cx *= gr_fovlud;
}
//Hurdler: 3D Water stuff
#define MAX_3DWATER 512
#ifndef SORTING
static void HWR_Add3DWater(lumpnum_t lumpnum, extrasubsector_t *xsub,
fixed_t fixedheight, INT32 lightlevel, INT32 alpha, sector_t *FOFSector)
{
static size_t allocedplanes = 0;
// Force realloc if buffer has been freed
if (!planeinfo)
allocedplanes = 0;
if (allocedplanes < numfloors + 1)
{
allocedplanes += MAX_3DWATER;
Z_Realloc(planeinfo, allocedplanes * sizeof (*planeinfo), PU_LEVEL, &planeinfo);
}
planeinfo[numfloors].fixedheight = fixedheight;
planeinfo[numfloors].lightlevel = lightlevel;
planeinfo[numfloors].lumpnum = lumpnum;
planeinfo[numfloors].xsub = xsub;
planeinfo[numfloors].alpha = alpha;
planeinfo[numfloors].FOFSector = FOFSector;
numfloors++;
}
#endif
#define DIST_PLANE(i) ABS(planeinfo[(i)].fixedheight-dup_viewz)
#if 0
static void HWR_QuickSortPlane(INT32 start, INT32 finish)
{
INT32 left = start;
INT32 right = finish;
INT32 starterval = (INT32)((right+left)/2); //pick a starter
planeinfo_t temp;
//'sort of sort' the two halves of the data list about starterval
while (right > left);
{
while (DIST_PLANE(left) < DIST_PLANE(starterval)) left++; //attempt to find a bigger value on the left
while (DIST_PLANE(right) > DIST_PLANE(starterval)) right--; //attempt to find a smaller value on the right
if (left < right) //if we haven't gone too far
{
//switch them
M_Memcpy(&temp, &planeinfo[left], sizeof (planeinfo_t));
M_Memcpy(&planeinfo[left], &planeinfo[right], sizeof (planeinfo_t));
M_Memcpy(&planeinfo[right], &temp, sizeof (planeinfo_t));
//move the bounds
left++;
right--;
}
}
if (start < right) HWR_QuickSortPlane(start, right);
if (left < finish) HWR_QuickSortPlane(left, finish);
}
#endif
#ifndef SORTING
static void HWR_Render3DWater(void)
{
size_t i;
//bubble sort 3D Water for correct alpha blending
{
boolean permut = true;
while (permut)
{
size_t j;
for (j = 0, permut= false; j < numfloors-1; j++)
{
if (ABS(planeinfo[j].fixedheight-dup_viewz) < ABS(planeinfo[j+1].fixedheight-dup_viewz))
{
planeinfo_t temp;
M_Memcpy(&temp, &planeinfo[j+1], sizeof (planeinfo_t));
M_Memcpy(&planeinfo[j+1], &planeinfo[j], sizeof (planeinfo_t));
M_Memcpy(&planeinfo[j], &temp, sizeof (planeinfo_t));
permut = true;
}
}
}
}
#if 0 //thanks epat, but it's goes looping forever on CTF map Silver Cascade Zone
HWR_QuickSortPlane(0, numplanes-1);
#endif
gr_frontsector = NULL; //Hurdler: gr_fronsector is no longer valid
for (i = 0; i < numfloors; i++)
{
HWR_GetFlat(planeinfo[i].lumpnum);
Take out fixedheight-based hacks for checking if floor or ceiling! Use an "isceiling" boolean for that instead This apparently fixes most of the issues with ceiling slopes in Boinciel's SUGOI map
2016-07-27 18:56:21 +00:00
HWR_RenderPlane(NULL, planeinfo[i].xsub, planeinfo[i].isceiling, planeinfo[i].fixedheight, PF_Translucent, planeinfo[i].lightlevel, planeinfo[i].lumpnum,
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
planeinfo[i].FOFSector, planeinfo[i].alpha, planeinfo[i].fogplane, planeinfo[i].planecolormap);
}
numfloors = 0;
}
#endif
static void HWR_AddTransparentWall(wallVert3D *wallVerts, FSurfaceInfo *pSurf, INT32 texnum, FBITFIELD blend, boolean fogwall, INT32 lightlevel, extracolormap_t *wallcolormap)
{
static size_t allocedwalls = 0;
// Force realloc if buffer has been freed
if (!wallinfo)
allocedwalls = 0;
if (allocedwalls < numwalls + 1)
{
allocedwalls += MAX_TRANSPARENTWALL;
Z_Realloc(wallinfo, allocedwalls * sizeof (*wallinfo), PU_LEVEL, &wallinfo);
}
M_Memcpy(wallinfo[numwalls].wallVerts, wallVerts, sizeof (wallinfo[numwalls].wallVerts));
M_Memcpy(&wallinfo[numwalls].Surf, pSurf, sizeof (FSurfaceInfo));
wallinfo[numwalls].texnum = texnum;
wallinfo[numwalls].blend = blend;
#ifdef SORTING
wallinfo[numwalls].drawcount = drawcount++;
#endif
wallinfo[numwalls].fogwall = fogwall;
wallinfo[numwalls].lightlevel = lightlevel;
wallinfo[numwalls].wallcolormap = wallcolormap;
numwalls++;
}
#ifndef SORTING
static void HWR_RenderTransparentWalls(void)
{
size_t i;
/*{ // sorting is disbale for now, do it!
INT32 permut = 1;
while (permut)
{
INT32 j;
for (j = 0, permut = 0; j < numwalls-1; j++)
{
if (ABS(wallinfo[j].fixedheight-dup_viewz) < ABS(wallinfo[j+1].fixedheight-dup_viewz))
{
wallinfo_t temp;
M_Memcpy(&temp, &wallinfo[j+1], sizeof (wallinfo_t));
M_Memcpy(&wallinfo[j+1], &wallinfo[j], sizeof (wallinfo_t));
M_Memcpy(&wallinfo[j], &temp, sizeof (wallinfo_t));
permut = 1;
}
}
}
}*/
for (i = 0; i < numwalls; i++)
{
HWR_GetTexture(wallinfo[i].texnum);
HWR_RenderWall(wallinfo[i].wallVerts, &wallinfo[i].Surf, wallinfo[i].blend, wallinfo[i].wall->fogwall, wallinfo[i].wall->lightlevel, wallinfo[i].wall->wallcolormap);
}
numwalls = 0;
}
#endif
static void HWR_RenderWall(wallVert3D *wallVerts, FSurfaceInfo *pSurf, FBITFIELD blend, boolean fogwall, INT32 lightlevel, extracolormap_t *wallcolormap)
{
FOutVector trVerts[4];
UINT8 i;
FOutVector *wv;
UINT8 alpha;
// transform
wv = trVerts;
// it sounds really stupid to do this conversion with the new T&L code
// we should directly put the right information in the right structure
// wallVerts3D seems ok, doesn't need FOutVector
// also remove the light copy
for (i = 0; i < 4; i++, wv++, wallVerts++)
{
wv->sow = wallVerts->s;
wv->tow = wallVerts->t;
wv->x = wallVerts->x;
wv->y = wallVerts->y;
wv->z = wallVerts->z;
}
alpha = pSurf->FlatColor.s.alpha; // retain the alpha
// Lighting is done here instead so that fog isn't drawn incorrectly on transparent walls after sorting
if (wallcolormap)
{
if (fogwall)
LightLevelToLum gone LightLevelToLum unused by everything now, it seemed to be giving wrong values for lightlevels.
2014-03-24 16:11:25 +00:00
pSurf->FlatColor.rgba = HWR_Lighting(lightlevel, wallcolormap->rgba, wallcolormap->fadergba, true, false);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
else
LightLevelToLum gone LightLevelToLum unused by everything now, it seemed to be giving wrong values for lightlevels.
2014-03-24 16:11:25 +00:00
pSurf->FlatColor.rgba = HWR_Lighting(lightlevel, wallcolormap->rgba, wallcolormap->fadergba, false, false);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
else
{
if (fogwall)
LightLevelToLum gone LightLevelToLum unused by everything now, it seemed to be giving wrong values for lightlevels.
2014-03-24 16:11:25 +00:00
pSurf->FlatColor.rgba = HWR_Lighting(lightlevel, NORMALFOG, FADEFOG, true, false);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
else
LightLevelToLum gone LightLevelToLum unused by everything now, it seemed to be giving wrong values for lightlevels.
2014-03-24 16:11:25 +00:00
pSurf->FlatColor.rgba = HWR_Lighting(lightlevel, NORMALFOG, FADEFOG, false, false);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
pSurf->FlatColor.s.alpha = alpha; // put the alpha back after lighting
if (blend & PF_Environment)
HWD.pfnDrawPolygon(pSurf, trVerts, 4, blend|PF_Modulated|PF_Clip|PF_Occlude); // PF_Occlude must be used for solid objects
else
HWD.pfnDrawPolygon(pSurf, trVerts, 4, blend|PF_Modulated|PF_Clip); // No PF_Occlude means overlapping (incorrect) transparency
#ifdef WALLSPLATS
if (gr_curline->linedef->splats && cv_splats.value)
HWR_DrawSegsSplats(pSurf);
#ifdef ALAM_LIGHTING
//Hurdler: TODO: do static lighting using gr_curline->lm
HWR_WallLighting(trVerts);
#endif
#endif
}
void HWR_SetPaletteColor(INT32 palcolor)
{
HWD.pfnSetSpecialState(HWD_SET_PALETTECOLOR, palcolor);
}
INT32 HWR_GetTextureUsed(void)
{
return HWD.pfnGetTextureUsed();
}
void HWR_DoPostProcessor(player_t *player)
{
SRB2 2.1.11 release
2014-08-27 03:56:30 +00:00
postimg_t *type;
if (splitscreen && player == &players[secondarydisplayplayer])
type = &postimgtype2;
else
type = &postimgtype;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
// Armageddon Blast Flash!
// Could this even be considered postprocessor?
if (player->flashcount)
{
FOutVector v[4];
FSurfaceInfo Surf;
OpenGL Flashpal Flashes Fixed flashes for OpenGL Nukes/Teleports etc.
2014-03-23 00:04:35 +00:00
v[0].x = v[2].y = v[3].x = v[3].y = -4.0f;
v[0].y = v[1].x = v[1].y = v[2].x = 4.0f;
v[0].z = v[1].z = v[2].z = v[3].z = 4.0f; // 4.0 because of the same reason as with the sky, just after the screen is cleared so near clipping plane is 3.99
// This won't change if the flash palettes are changed unfortunately, but it works for its purpose
if (player->flashpal == PAL_NUKE)
{
Surf.FlatColor.s.red = 0xff;
Surf.FlatColor.s.green = Surf.FlatColor.s.blue = 0x7F; // The nuke palette is kind of pink-ish
}
else
Surf.FlatColor.s.red = Surf.FlatColor.s.green = Surf.FlatColor.s.blue = 0xff;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
Surf.FlatColor.s.alpha = 0xc0; // match software mode
HWD.pfnDrawPolygon(&Surf, v, 4, PF_Modulated|PF_Additive|PF_NoTexture|PF_NoDepthTest|PF_Clip|PF_NoZClip);
}
// Capture the screen for intermission and screen waving
if(gamestate != GS_INTERMISSION)
HWD.pfnMakeScreenTexture();
if (splitscreen) // Not supported in splitscreen - someone want to add support?
return;
#ifdef SHUFFLE
// Drunken vision! WooOOooo~
SRB2 2.1.11 release
2014-08-27 03:56:30 +00:00
if (*type == postimg_water || *type == postimg_heat)
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
// 10 by 10 grid. 2 coordinates (xy)
float v[SCREENVERTS][SCREENVERTS][2];
static double disStart = 0;
UINT8 x, y;
INT32 WAVELENGTH;
INT32 AMPLITUDE;
INT32 FREQUENCY;
// Modifies the wave.
SRB2 2.1.11 release
2014-08-27 03:56:30 +00:00
if (*type == postimg_water)
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
WAVELENGTH = 20; // Lower is longer
AMPLITUDE = 20; // Lower is bigger
FREQUENCY = 16; // Lower is faster
}
else
{
WAVELENGTH = 10; // Lower is longer
AMPLITUDE = 30; // Lower is bigger
FREQUENCY = 4; // Lower is faster
}
for (x = 0; x < SCREENVERTS; x++)
{
for (y = 0; y < SCREENVERTS; y++)
{
// Change X position based on its Y position.
v[x][y][0] = (x/((float)(SCREENVERTS-1.0f)/9.0f))-4.5f + (float)sin((disStart+(y*WAVELENGTH))/FREQUENCY)/AMPLITUDE;
v[x][y][1] = (y/((float)(SCREENVERTS-1.0f)/9.0f))-4.5f;
}
}
HWD.pfnPostImgRedraw(v);
disStart += 1;
OpenGL Screen Flip Screen flipping in OpenGL is no longer done by inverting the screen polygons.
2014-04-07 16:08:04 +00:00
// Capture the screen again for screen waving on the intermission
if(gamestate != GS_INTERMISSION)
HWD.pfnMakeScreenTexture();
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
OpenGL Screen Flip Screen flipping in OpenGL is no longer done by inverting the screen polygons.
2014-04-07 16:08:04 +00:00
// Flipping of the screen isn't done here anymore
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
#endif // SHUFFLE
}
void HWR_StartScreenWipe(void)
{
//CONS_Debug(DBG_RENDER, "In HWR_StartScreenWipe()\n");
HWD.pfnStartScreenWipe();
}
void HWR_EndScreenWipe(void)
{
OpenGL Fade masks I think I've done this all right, someone correct me if I haven't.
2014-09-04 00:35:29 +00:00
HWRWipeCounter = 0.0f;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
//CONS_Debug(DBG_RENDER, "In HWR_EndScreenWipe()\n");
HWD.pfnEndScreenWipe();
}
void HWR_DrawIntermissionBG(void)
{
HWD.pfnDrawIntermissionBG();
}
OpenGL Fade masks I think I've done this all right, someone correct me if I haven't.
2014-09-04 00:35:29 +00:00
void HWR_DoWipe(UINT8 wipenum, UINT8 scrnnum)
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
{
OpenGL Fade masks I think I've done this all right, someone correct me if I haven't.
2014-09-04 00:35:29 +00:00
static char lumpname[9] = "FADEmmss";
lumpnum_t lumpnum;
size_t lsize;
if (wipenum > 99 || scrnnum > 99) // not a valid wipe number
return; // shouldn't end up here really, the loop should've stopped running beforehand
// puts the numbers into the lumpname
sprintf(&lumpname[4], "%.2hu%.2hu", (UINT16)wipenum, (UINT16)scrnnum);
lumpnum = W_CheckNumForName(lumpname);
if (lumpnum == LUMPERROR) // again, shouldn't be here really
return;
lsize = W_LumpLength(lumpnum);
if (!(lsize == 256000 || lsize == 64000 || lsize == 16000 || lsize == 4000))
{
CONS_Alert(CONS_WARNING, "Fade mask lump %s of incorrect size, ignored\n", lumpname);
return; // again, shouldn't get here if it is a bad size
}
HWR_GetFadeMask(lumpnum);
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
OpenGL Fade masks I think I've done this all right, someone correct me if I haven't.
2014-09-04 00:35:29 +00:00
HWD.pfnDoScreenWipe(HWRWipeCounter); // Still send in wipecounter since old stuff might not support multitexturing
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
OpenGL Fade masks I think I've done this all right, someone correct me if I haven't.
2014-09-04 00:35:29 +00:00
HWRWipeCounter += 0.05f; // increase opacity of end screen
OpenGL Screen Fading Fix Screen Wipes now work as they did in 2.0.x.
2014-03-31 22:47:12 +00:00
OpenGL Fade masks I think I've done this all right, someone correct me if I haven't.
2014-09-04 00:35:29 +00:00
if (HWRWipeCounter > 1.0f)
HWRWipeCounter = 1.0f;
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
}
Virtual resolutions in OpenGL Also made fades use core functions if they are available.
2014-11-14 00:06:38 +00:00
void HWR_MakeScreenFinalTexture(void)
{
HWD.pfnMakeScreenFinalTexture();
}
void HWR_DrawScreenFinalTexture(int width, int height)
{
HWD.pfnDrawScreenFinalTexture(width, height);
}
SRB2 2.1 release
2014-03-15 16:59:03 +00:00
#endif // HWRENDER
Reference in a new issue Copy permalink