cvar_t r_subdivisions_collision_maxtess = {0, "r_subdivisions_collision_maxtess", "1024", "maximum number of subdivisions (prevents curves beyond a certain detail level, limits smoothing)"};
cvar_t r_subdivisions_collision_maxvertices = {0, "r_subdivisions_collision_maxvertices", "4225", "maximum vertices allowed per subdivided curve"};
cvar_t mod_q3bsp_curves_collisions = {0, "mod_q3bsp_curves_collisions", "1", "enables collisions with curves (SLOW)"};
-cvar_t mod_q3bsp_curves_collisions_stride = {0, "mod_q3bsp_curves_collisions_stride", "16", "collisions against curves: optimize performance by doing a combined collision check for this triangle amount first"};
-cvar_t mod_q3bsp_curves_stride = {0, "mod_q3bsp_curves_stride", "16", "particle effect collisions against curves: optimize performance by doing a combined collision check for this triangle amount first"};
+cvar_t mod_q3bsp_curves_collisions_stride = {0, "mod_q3bsp_curves_collisions_stride", "16", "collisions against curves: optimize performance by doing a combined collision check for this triangle amount first (-1 avoids any box tests)"};
+cvar_t mod_q3bsp_curves_stride = {0, "mod_q3bsp_curves_stride", "16", "particle effect collisions against curves: optimize performance by doing a combined collision check for this triangle amount first (-1 avoids any box tests)"};
cvar_t mod_q3bsp_optimizedtraceline = {0, "mod_q3bsp_optimizedtraceline", "1", "whether to use optimized traceline code for line traces (as opposed to tracebox code)"};
cvar_t mod_q3bsp_debugtracebrush = {0, "mod_q3bsp_debugtracebrush", "0", "selects different tracebrush bsp recursion algorithms (for debugging purposes only)"};
cvar_t mod_q3bsp_lightmapmergepower = {CVAR_SAVE, "mod_q3bsp_lightmapmergepower", "4", "merges the quake3 128x128 lightmap textures into larger lightmap group textures to speed up rendering, 1 = 256x256, 2 = 512x512, 3 = 1024x1024, 4 = 2048x2048, 5 = 4096x4096, ..."};
cvar_t mod_q3bsp_nolightmaps = {CVAR_SAVE, "mod_q3bsp_nolightmaps", "0", "do not load lightmaps in Q3BSP maps (to save video RAM, but be warned: it looks ugly)"};
+cvar_t mod_q3bsp_tracelineofsight_brushes = {0, "mod_q3bsp_tracelineofsight_brushes", "0", "enables culling of entities behind detail brushes, curves, etc"};
static texture_t mod_q1bsp_texture_solid;
static texture_t mod_q1bsp_texture_sky;
Cvar_RegisterVariable(&mod_q3bsp_debugtracebrush);
Cvar_RegisterVariable(&mod_q3bsp_lightmapmergepower);
Cvar_RegisterVariable(&mod_q3bsp_nolightmaps);
+ Cvar_RegisterVariable(&mod_q3bsp_tracelineofsight_brushes);
memset(&mod_q1bsp_texture_solid, 0, sizeof(mod_q1bsp_texture_solid));
strlcpy(mod_q1bsp_texture_solid.name, "solid" , sizeof(mod_q1bsp_texture_solid.name));
mnode_t *node, *nodestack[1024];
if (!model->brush.num_pvsclusters)
return -1;
- node = model->brush.data_nodes;
+ node = model->brush.data_nodes + model->brushq1.hulls[0].firstclipnode;
for (;;)
{
#if 1
mnode_t *node, *nodestack[1024];
if (!model->brush.num_pvsclusters)
return true;
- node = model->brush.data_nodes;
+ node = model->brush.data_nodes + model->brushq1.hulls[0].firstclipnode;
for (;;)
{
#if 1
mnode_t *node, *nodestack[1024];
if (!model->brush.num_leafs)
return true;
- node = model->brush.data_nodes;
+ node = model->brush.data_nodes + model->brushq1.hulls[0].firstclipnode;
for (;;)
{
#if 1
mnode_t *node, *nodestack[1024];
if (!model->brush.num_leafs)
return true;
- node = model->brush.data_nodes;
+ node = model->brush.data_nodes + model->brushq1.hulls[0].firstclipnode;
for (;;)
{
#if 1
surface = info->model->data_surfaces + *mark;
if (surface->texture->supercontents & SUPERCONTENTS_SOLID)
{
- if(surface->num_bboxstride)
+ if(surface->num_bboxstride > 0)
{
int i, cnt, tri;
cnt = (surface->num_triangles + surface->num_bboxstride - 1) / surface->num_bboxstride;
}
//#endif
-static void Mod_Q1BSP_TracePoint(struct model_s *model, int frame, trace_t *trace, const vec3_t start, int hitsupercontentsmask)
+static void Mod_Q1BSP_TracePoint(struct model_s *model, const frameblend_t *frameblend, const skeleton_t *skeleton, trace_t *trace, const vec3_t start, int hitsupercontentsmask)
{
RecursiveHullCheckTraceInfo_t rhc;
Mod_Q1BSP_RecursiveHullCheckPoint(&rhc, rhc.hull->firstclipnode);
}
-static void Mod_Q1BSP_TraceLine(struct model_s *model, int frame, trace_t *trace, const vec3_t start, const vec3_t end, int hitsupercontentsmask)
+static void Mod_Q1BSP_TraceLine(struct model_s *model, const frameblend_t *frameblend, const skeleton_t *skeleton, trace_t *trace, const vec3_t start, const vec3_t end, int hitsupercontentsmask)
{
RecursiveHullCheckTraceInfo_t rhc;
if (VectorCompare(start, end))
{
- Mod_Q1BSP_TracePoint(model, frame, trace, start, hitsupercontentsmask);
+ Mod_Q1BSP_TracePoint(model, frameblend, skeleton, trace, start, hitsupercontentsmask);
return;
}
#endif
}
-static void Mod_Q1BSP_TraceBox(struct model_s *model, int frame, trace_t *trace, const vec3_t start, const vec3_t boxmins, const vec3_t boxmaxs, const vec3_t end, int hitsupercontentsmask)
+static void Mod_Q1BSP_TraceBox(struct model_s *model, const frameblend_t *frameblend, const skeleton_t *skeleton, trace_t *trace, const vec3_t start, const vec3_t boxmins, const vec3_t boxmaxs, const vec3_t end, int hitsupercontentsmask)
{
// this function currently only supports same size start and end
double boxsize[3];
if (VectorCompare(boxmins, boxmaxs))
{
if (VectorCompare(start, end))
- Mod_Q1BSP_TracePoint(model, frame, trace, start, hitsupercontentsmask);
+ Mod_Q1BSP_TracePoint(model, frameblend, skeleton, trace, start, hitsupercontentsmask);
else
- Mod_Q1BSP_TraceLine(model, frame, trace, start, end, hitsupercontentsmask);
+ Mod_Q1BSP_TraceLine(model, frameblend, skeleton, trace, start, end, hitsupercontentsmask);
return;
}
#if 1
colbrushf_t cbox;
colplanef_t cbox_planes[6];
+ cbox.isaabb = true;
+ cbox.hasaabbplanes = true;
cbox.supercontents = boxsupercontents;
cbox.numplanes = 6;
cbox.numpoints = 0;
}
}
-static int Mod_Q1BSP_TraceLineOfSight_RecursiveNodeCheck(mnode_t *node, double p1[3], double p2[3])
-{
- double t1, t2;
- double midf, mid[3];
- int ret, side;
-
- // check for empty
- while (node->plane)
- {
- // find the point distances
- mplane_t *plane = node->plane;
- if (plane->type < 3)
- {
- t1 = p1[plane->type] - plane->dist;
- t2 = p2[plane->type] - plane->dist;
- }
- else
- {
- t1 = DotProduct (plane->normal, p1) - plane->dist;
- t2 = DotProduct (plane->normal, p2) - plane->dist;
- }
-
- if (t1 < 0)
- {
- if (t2 < 0)
- {
- node = node->children[1];
- continue;
- }
- side = 1;
- }
- else
- {
- if (t2 >= 0)
- {
- node = node->children[0];
- continue;
- }
- side = 0;
- }
-
- midf = t1 / (t1 - t2);
- VectorLerp(p1, midf, p2, mid);
-
- // recurse both sides, front side first
- // return 2 if empty is followed by solid (hit something)
- // do not return 2 if both are solid or both empty,
- // or if start is solid and end is empty
- // as these degenerate cases usually indicate the eye is in solid and
- // should see the target point anyway
- ret = Mod_Q1BSP_TraceLineOfSight_RecursiveNodeCheck(node->children[side ], p1, mid);
- if (ret != 0)
- return ret;
- ret = Mod_Q1BSP_TraceLineOfSight_RecursiveNodeCheck(node->children[side ^ 1], mid, p2);
- if (ret != 1)
- return ret;
- return 2;
- }
- return ((mleaf_t *)node)->clusterindex < 0;
-}
-
static qboolean Mod_Q1BSP_TraceLineOfSight(struct model_s *model, const vec3_t start, const vec3_t end)
{
- // this function currently only supports same size start and end
- double tracestart[3], traceend[3];
- VectorCopy(start, tracestart);
- VectorCopy(end, traceend);
- return Mod_Q1BSP_TraceLineOfSight_RecursiveNodeCheck(model->brush.data_nodes, tracestart, traceend) != 2;
+ trace_t trace;
+ model->TraceLine(model, NULL, NULL, &trace, start, end, SUPERCONTENTS_VISBLOCKERMASK);
+ return trace.fraction == 1;
}
static int Mod_Q1BSP_LightPoint_RecursiveBSPNode(dp_model_t *model, vec3_t ambientcolor, vec3_t diffusecolor, vec3_t diffusenormal, const mnode_t *node, float x, float y, float startz, float endz)
*/
void R_Q1BSP_LoadSplitSky (unsigned char *src, int width, int height, int bytesperpixel)
{
- int i, j;
- unsigned solidpixels[128*128], alphapixels[128*128];
+ int x, y;
+ int w = width/2;
+ int h = height;
+ unsigned *solidpixels = Mem_Alloc(tempmempool, w*h*sizeof(unsigned char[4]));
+ unsigned *alphapixels = Mem_Alloc(tempmempool, w*h*sizeof(unsigned char[4]));
// allocate a texture pool if we need it
if (loadmodel->texturepool == NULL && cls.state != ca_dedicated)
if (bytesperpixel == 4)
{
- for (i = 0;i < 128;i++)
+ for (y = 0;y < h;y++)
{
- for (j = 0;j < 128;j++)
+ for (x = 0;x < w;x++)
{
- solidpixels[(i*128) + j] = ((unsigned *)src)[i*256+j+128];
- alphapixels[(i*128) + j] = ((unsigned *)src)[i*256+j];
+ solidpixels[y*w+x] = ((unsigned *)src)[y*width+x+w];
+ alphapixels[y*w+x] = ((unsigned *)src)[y*width+x];
}
}
}
}
bgra;
r = g = b = 0;
- for (i = 0;i < 128;i++)
+ for (y = 0;y < h;y++)
{
- for (j = 0;j < 128;j++)
+ for (x = 0;x < w;x++)
{
- p = src[i*256 + j + 128];
+ p = src[x*width+y+w];
r += palette_rgb[p][0];
g += palette_rgb[p][1];
b += palette_rgb[p][2];
}
}
- bgra.b[2] = r/(128*128);
- bgra.b[1] = g/(128*128);
- bgra.b[0] = b/(128*128);
+ bgra.b[2] = r/(w*h);
+ bgra.b[1] = g/(w*h);
+ bgra.b[0] = b/(w*h);
bgra.b[3] = 0;
- for (i = 0;i < 128;i++)
+ for (y = 0;y < h;y++)
{
- for (j = 0;j < 128;j++)
+ for (x = 0;x < w;x++)
{
- solidpixels[(i*128) + j] = palette_bgra_complete[src[i*256 + j + 128]];
- p = src[i*256 + j];
- alphapixels[(i*128) + j] = p ? palette_bgra_complete[p] : bgra.i;
+ solidpixels[y*w+x] = palette_bgra_complete[src[y*width+x+w]];
+ p = src[y*width+x];
+ alphapixels[y*w+x] = p ? palette_bgra_complete[p] : bgra.i;
}
}
}
- loadmodel->brush.solidskytexture = R_LoadTexture2D(loadmodel->texturepool, "sky_solidtexture", 128, 128, (unsigned char *) solidpixels, TEXTYPE_BGRA, TEXF_PRECACHE, NULL);
- loadmodel->brush.alphaskytexture = R_LoadTexture2D(loadmodel->texturepool, "sky_alphatexture", 128, 128, (unsigned char *) alphapixels, TEXTYPE_BGRA, TEXF_ALPHA | TEXF_PRECACHE, NULL);
+ loadmodel->brush.solidskyskinframe = R_SkinFrame_LoadInternalBGRA("sky_solidtexture", TEXF_PRECACHE, (unsigned char *) solidpixels, w, h);
+ loadmodel->brush.alphaskyskinframe = R_SkinFrame_LoadInternalBGRA("sky_alphatexture", TEXF_ALPHA | TEXF_PRECACHE, (unsigned char *) alphapixels, w, h);
+ Mem_Free(solidpixels);
+ Mem_Free(alphapixels);
}
static void Mod_Q1BSP_LoadTextures(lump_t *l)
tx->reflectfactor = 1;
Vector4Set(tx->reflectcolor4f, 1, 1, 1, 1);
tx->r_water_wateralpha = 1;
+ tx->specularscalemod = 1;
+ tx->specularpowermod = 1;
}
if (!m)
if (cls.state != ca_dedicated)
{
// LordHavoc: HL sky textures are entirely different than quake
- if (!loadmodel->brush.ishlbsp && !strncmp(tx->name, "sky", 3) && mtwidth == 256 && mtheight == 128)
+ if (!loadmodel->brush.ishlbsp && !strncmp(tx->name, "sky", 3) && mtwidth == mtheight * 2)
{
data = loadimagepixelsbgra(tx->name, false, false);
- if (data && image_width == 256 && image_height == 128)
+ if (data && image_width == image_height * 2)
{
R_Q1BSP_LoadSplitSky(data, image_width, image_height, 4);
Mem_Free(data);
Mem_Free(freepixels);
}
else if (mtdata) // texture included
- skinframe = R_SkinFrame_LoadInternalQuake(tx->name, TEXF_ALPHA | TEXF_MIPMAP | TEXF_PRECACHE | (r_picmipworld.integer ? TEXF_PICMIP : 0), false, r_fullbrights.integer, mtdata, tx->width, tx->height);
+ skinframe = R_SkinFrame_LoadInternalQuake(tx->name, TEXF_PRECACHE | TEXF_MIPMAP | (r_picmipworld.integer ? TEXF_PICMIP : 0), false, r_fullbrights.integer, mtdata, tx->width, tx->height);
}
// if skinframe is still NULL the "missing" texture will be used
if (skinframe)
tx->basematerialflags |= MATERIALFLAG_WATERSCROLL | MATERIALFLAG_LIGHTBOTHSIDES | MATERIALFLAG_NOSHADOW;
else
tx->basematerialflags |= MATERIALFLAG_WATERSCROLL | MATERIALFLAG_LIGHTBOTHSIDES | MATERIALFLAG_NOSHADOW | MATERIALFLAG_WATERALPHA | MATERIALFLAG_WATERSHADER;
- if (tx->skinframes[0] && tx->skinframes[0]->fog)
+ if (tx->skinframes[0] && tx->skinframes[0]->hasalpha)
tx->basematerialflags |= MATERIALFLAG_ALPHA | MATERIALFLAG_BLENDED | MATERIALFLAG_NOSHADOW;
}
else if (!strncmp(tx->name, "mirror", 6)) // Tenebrae
tx->basematerialflags = MATERIALFLAG_SKY | MATERIALFLAG_NOSHADOW;
else if (!strcmp(tx->name, "caulk"))
tx->basematerialflags = MATERIALFLAG_NODRAW | MATERIALFLAG_NOSHADOW;
- else if (tx->skinframes[0] && tx->skinframes[0]->fog)
+ else if (tx->skinframes[0] && tx->skinframes[0]->hasalpha)
tx->basematerialflags |= MATERIALFLAG_ALPHA | MATERIALFLAG_BLENDED | MATERIALFLAG_NOSHADOW;
// start out with no animation
tx->currentframe = tx;
tx->currentskinframe = tx->skinframes[0];
+ tx->specularscalemod = 1; // not supported here
+ tx->specularpowermod = 1; // not supported here
}
}
}
#endif
-/* Maximum size of a single LM */
-#define MAX_SINGLE_LM_SIZE 256
-
-struct alloc_lm_row
-{
- int rowY;
- int currentX;
-};
-
-struct alloc_lm_state
-{
- int currentY;
- struct alloc_lm_row rows[MAX_SINGLE_LM_SIZE];
-};
-
-static void init_alloc_lm_state (struct alloc_lm_state* state)
-{
- int r;
-
- state->currentY = 0;
- for (r = 0; r < MAX_SINGLE_LM_SIZE; r++)
- {
- state->rows[r].currentX = 0;
- state->rows[r].rowY = -1;
- }
-}
-
-static qboolean Mod_Q1BSP_AllocLightmapBlock(struct alloc_lm_state* state, int totalwidth, int totalheight, int blockwidth, int blockheight, int *outx, int *outy)
-{
- struct alloc_lm_row* row;
- int r;
-
- row = &(state->rows[blockheight]);
- if ((row->rowY < 0) || (row->currentX + blockwidth > totalwidth))
- {
- if (state->currentY + blockheight <= totalheight)
- {
- row->rowY = state->currentY;
- row->currentX = 0;
- state->currentY += blockheight;
- }
- else
- {
- /* See if we can stuff the block into a higher row */
- row = NULL;
- for (r = blockheight; r < MAX_SINGLE_LM_SIZE; r++)
- {
- if ((state->rows[r].rowY >= 0)
- && (state->rows[r].currentX + blockwidth <= totalwidth))
- {
- row = &(state->rows[r]);
- break;
- }
- }
- if (row == NULL) return false;
- }
- }
- *outy = row->rowY;
- *outx = row->currentX;
- row->currentX += blockwidth;
-
- return true;
-}
-
-extern cvar_t gl_max_size;
+extern cvar_t gl_max_lightmapsize;
static void Mod_Q1BSP_LoadFaces(lump_t *l)
{
dface_t *in;
lightmaptexture = NULL;
deluxemaptexture = r_texture_blanknormalmap;
lightmapnumber = 1;
- lightmapsize = max(256, gl_max_size.integer);
+ lightmapsize = bound(256, gl_max_lightmapsize.integer, (int)vid.maxtexturesize_2d);
totallightmapsamples = 0;
totalverts = 0;
if (i == -1)
{
surface->lightmapinfo->samples = NULL;
-#if 0
+#if 1
// give non-lightmapped water a 1x white lightmap
if (surface->texture->name[0] == '*' && (surface->lightmapinfo->texinfo->flags & TEX_SPECIAL) && ssize <= 256 && tsize <= 256)
{
// small maps (such as ammo boxes especially) don't need big lightmap
// textures, so this code tries to guess a good size based on
// totallightmapsamples (size of the lightmaps lump basically), as well as
- // trying to max out the gl_max_size if there is a lot of lightmap data to
- // store
+ // trying to max out the size if there is a lot of lightmap data to store
// additionally, never choose a lightmapsize that is smaller than the
// largest surface encountered (as it would fail)
i = lightmapsize;
- for (lightmapsize = 64; (lightmapsize < i) && (lightmapsize < gl_max_size.integer) && (totallightmapsamples > lightmapsize*lightmapsize); lightmapsize*=2)
+ for (lightmapsize = 64; (lightmapsize < i) && (lightmapsize < bound(128, gl_max_lightmapsize.integer, (int)vid.maxtexturesize_2d)) && (totallightmapsamples > lightmapsize*lightmapsize); lightmapsize*=2)
;
// now that we've decided the lightmap texture size, we can do the rest
if (cls.state != ca_dedicated)
{
int stainmapsize = 0;
- struct alloc_lm_state allocState;
+ mod_alloclightmap_state_t allocState;
+ Mod_AllocLightmap_Init(&allocState, lightmapsize, lightmapsize);
for (surfacenum = 0, surface = loadmodel->data_surfaces;surfacenum < count;surfacenum++, surface++)
{
int i, iu, iv, lightmapx = 0, lightmapy = 0;
tsize = (surface->lightmapinfo->extents[1] >> 4) + 1;
stainmapsize += ssize * tsize * 3;
- if (!lightmaptexture || !Mod_Q1BSP_AllocLightmapBlock(&allocState, lightmapsize, lightmapsize, ssize, tsize, &lightmapx, &lightmapy))
+ if (!lightmaptexture || !Mod_AllocLightmap_Block(&allocState, ssize, tsize, &lightmapx, &lightmapy))
{
// allocate a texture pool if we need it
if (loadmodel->texturepool == NULL)
if (loadmodel->brushq1.nmaplightdata)
deluxemaptexture = R_LoadTexture2D(loadmodel->texturepool, va("deluxemap%i", lightmapnumber), lightmapsize, lightmapsize, NULL, TEXTYPE_BGRA, TEXF_FORCELINEAR | TEXF_PRECACHE, NULL);
lightmapnumber++;
- init_alloc_lm_state (&allocState);
- Mod_Q1BSP_AllocLightmapBlock(&allocState, lightmapsize, lightmapsize, ssize, tsize, &lightmapx, &lightmapy);
+ Mod_AllocLightmap_Reset(&allocState);
+ Mod_AllocLightmap_Block(&allocState, ssize, tsize, &lightmapx, &lightmapy);
}
surface->lightmaptexture = lightmaptexture;
surface->deluxemaptexture = deluxemaptexture;
}
portal_t;
-static portal_t *portalchain;
-
-/*
-===========
-AllocPortal
-===========
-*/
-static portal_t *AllocPortal(void)
-{
- portal_t *p;
- p = (portal_t *)Mem_Alloc(loadmodel->mempool, sizeof(portal_t));
- p->chain = portalchain;
- portalchain = p;
- return p;
-}
-
-static void FreePortal(portal_t *p)
-{
- Mem_Free(p);
-}
+static memexpandablearray_t portalarray;
static void Mod_Q1BSP_RecursiveRecalcNodeBBox(mnode_t *node)
{
static void Mod_Q1BSP_FinalizePortals(void)
{
- int i, j, numportals, numpoints;
- portal_t *p, *pnext;
+ int i, j, numportals, numpoints, portalindex, portalrange = Mem_ExpandableArray_IndexRange(&portalarray);
+ portal_t *p;
mportal_t *portal;
mvertex_t *point;
mleaf_t *leaf, *endleaf;
VectorSet(leaf->mins, 2000000000, 2000000000, 2000000000);
VectorSet(leaf->maxs, -2000000000, -2000000000, -2000000000);
}
- p = portalchain;
numportals = 0;
numpoints = 0;
- while (p)
+ for (portalindex = 0;portalindex < portalrange;portalindex++)
{
+ p = (portal_t*)Mem_ExpandableArray_RecordAtIndex(&portalarray, portalindex);
+ if (!p)
+ continue;
// note: this check must match the one below or it will usually corrupt memory
// the nodes[0] != nodes[1] check is because leaf 0 is the shared solid leaf, it can have many portals inside with leaf 0 on both sides
if (p->numpoints >= 3 && p->nodes[0] != p->nodes[1] && ((mleaf_t *)p->nodes[0])->clusterindex >= 0 && ((mleaf_t *)p->nodes[1])->clusterindex >= 0)
numportals += 2;
numpoints += p->numpoints * 2;
}
- p = p->chain;
}
loadmodel->brush.data_portals = (mportal_t *)Mem_Alloc(loadmodel->mempool, numportals * sizeof(mportal_t) + numpoints * sizeof(mvertex_t));
loadmodel->brush.num_portals = numportals;
// process all portals in the global portal chain, while freeing them
portal = loadmodel->brush.data_portals;
point = loadmodel->brush.data_portalpoints;
- p = portalchain;
- portalchain = NULL;
- while (p)
+ for (portalindex = 0;portalindex < portalrange;portalindex++)
{
- pnext = p->chain;
-
+ p = (portal_t*)Mem_ExpandableArray_RecordAtIndex(&portalarray, portalindex);
+ if (!p)
+ continue;
if (p->numpoints >= 3 && p->nodes[0] != p->nodes[1])
{
// note: this check must match the one above or it will usually corrupt memory
}
}
}
- FreePortal(p);
- p = pnext;
}
// now recalculate the node bounding boxes from the leafs
- Mod_Q1BSP_RecursiveRecalcNodeBBox(loadmodel->brush.data_nodes);
+ Mod_Q1BSP_RecursiveRecalcNodeBBox(loadmodel->brush.data_nodes + loadmodel->brushq1.hulls[0].firstclipnode);
}
/*
// create the new portal by generating a polygon for the node plane,
// and clipping it by all of the other portals(which came from nodes above this one)
- nodeportal = AllocPortal();
+ nodeportal = (portal_t *)Mem_ExpandableArray_AllocRecord(&portalarray);
nodeportal->plane = *plane;
// TODO: calculate node bounding boxes during recursion and calculate a maximum plane size accordingly to improve precision (as most maps do not need 1 billion unit plane polygons)
}
// the portal is split
- splitportal = AllocPortal();
+ splitportal = (portal_t *)Mem_ExpandableArray_AllocRecord(&portalarray);
temp = splitportal->chain;
*splitportal = *portal;
splitportal->chain = temp;
static void Mod_Q1BSP_MakePortals(void)
{
- portalchain = NULL;
- Mod_Q1BSP_RecursiveNodePortals(loadmodel->brush.data_nodes);
+ Mem_ExpandableArray_NewArray(&portalarray, loadmodel->mempool, sizeof(portal_t), 1020*1024/sizeof(portal_t));
+ Mod_Q1BSP_RecursiveNodePortals(loadmodel->brush.data_nodes + loadmodel->brushq1.hulls[0].firstclipnode);
Mod_Q1BSP_FinalizePortals();
+ Mem_ExpandableArray_FreeArray(&portalarray);
}
//Returns PVS data for a given point
static unsigned char *Mod_Q1BSP_GetPVS(dp_model_t *model, const vec3_t p)
{
mnode_t *node;
- node = model->brush.data_nodes;
+ node = model->brush.data_nodes + model->brushq1.hulls[0].firstclipnode;
while (node->plane)
node = node->children[(node->plane->type < 3 ? p[node->plane->type] : DotProduct(p,node->plane->normal)) < node->plane->dist];
if (((mleaf_t *)node)->clusterindex >= 0)
}
if (!merge)
memset(pvsbuffer, 0, bytes);
- Mod_Q1BSP_FatPVS_RecursiveBSPNode(model, org, radius, pvsbuffer, bytes, model->brush.data_nodes);
+ Mod_Q1BSP_FatPVS_RecursiveBSPNode(model, org, radius, pvsbuffer, bytes, model->brush.data_nodes + model->brushq1.hulls[0].firstclipnode);
return bytes;
}
mod->Draw = R_Q1BSP_Draw;
mod->DrawDepth = R_Q1BSP_DrawDepth;
mod->DrawDebug = R_Q1BSP_DrawDebug;
+ mod->DrawPrepass = R_Q1BSP_DrawPrepass;
mod->GetLightInfo = R_Q1BSP_GetLightInfo;
mod->CompileShadowMap = R_Q1BSP_CompileShadowMap;
mod->DrawShadowMap = R_Q1BSP_DrawShadowMap;
// textures and memory belong to the main model
mod->texturepool = NULL;
mod->mempool = NULL;
- mod->brush.TraceLineOfSight = NULL;
mod->brush.GetPVS = NULL;
mod->brush.FatPVS = NULL;
mod->brush.BoxTouchingPVS = NULL;
mod->firstmodelsurface = bm->firstface;
mod->nummodelsurfaces = bm->numfaces;
+ // set node/leaf parents for this submodel
+ Mod_Q1BSP_LoadNodes_RecursiveSetParent(mod->brush.data_nodes + mod->brushq1.hulls[0].firstclipnode, NULL);
+
// make the model surface list (used by shadowing/lighting)
mod->sortedmodelsurfaces = (int *)datapointer;datapointer += mod->nummodelsurfaces * sizeof(int);
Mod_MakeSortedSurfaces(mod);
if (!COM_ParseToken_Simple(&data, false, false))
break; // error
strlcpy(value, com_token, sizeof(value));
- if (!strcmp("gridsize", key))
+ if (!strcasecmp("gridsize", key)) // this one is case insensitive to 100% match q3map2
{
#if _MSC_VER >= 1400
#define sscanf sscanf_s
#endif
+#if 0
if (sscanf(value, "%f %f %f", &v[0], &v[1], &v[2]) == 3 && v[0] != 0 && v[1] != 0 && v[2] != 0)
VectorCopy(v, loadmodel->brushq3.num_lightgrid_cellsize);
+#else
+ VectorSet(v, 64, 64, 128);
+ if(sscanf(value, "%f %f %f", &v[0], &v[1], &v[2]) != 3)
+ Con_Printf("Mod_Q3BSP_LoadEntities: funny gridsize \"%s\" in %s, interpreting as \"%f %f %f\" to match q3map2's parsing\n", value, loadmodel->name, v[0], v[1], v[2]);
+ if (v[0] != 0 && v[1] != 0 && v[2] != 0)
+ VectorCopy(v, loadmodel->brushq3.num_lightgrid_cellsize);
+#endif
}
else if (!strcmp("deluxeMaps", key))
{
{
q3dbrush_t *in;
q3mbrush_t *out;
- int i, j, n, c, count, maxplanes;
+ int i, j, n, c, count, maxplanes, q3surfaceflags;
colplanef_t *planes;
in = (q3dbrush_t *)(mod_base + l->fileofs);
Mem_Free(planes);
planes = (colplanef_t *)Mem_Alloc(tempmempool, sizeof(colplanef_t) * maxplanes);
}
+ q3surfaceflags = 0;
for (j = 0;j < out->numbrushsides;j++)
{
VectorCopy(out->firstbrushside[j].plane->normal, planes[j].normal);
planes[j].dist = out->firstbrushside[j].plane->dist;
planes[j].q3surfaceflags = out->firstbrushside[j].texture->surfaceflags;
planes[j].texture = out->firstbrushside[j].texture;
+ q3surfaceflags |= planes[j].q3surfaceflags;
}
// make the colbrush from the planes
- out->colbrushf = Collision_NewBrushFromPlanes(loadmodel->mempool, out->numbrushsides, planes, out->texture->supercontents);
+ out->colbrushf = Collision_NewBrushFromPlanes(loadmodel->mempool, out->numbrushsides, planes, out->texture->supercontents, q3surfaceflags, out->texture, true);
// this whole loop can take a while (e.g. on redstarrepublic4)
CL_KeepaliveMessage(false);
;
// i is now 0 for 128, 1 for 256, etc
- for (power = 1;power + i <= mod_q3bsp_lightmapmergepower.integer && (size << power) <= gl_max_texture_size && (1 << (power * 2)) < 4 * (count >> (loadmodel->brushq3.deluxemapping ? 1 : 0)); power++)
+ for (power = 1;power + i <= mod_q3bsp_lightmapmergepower.integer && (size << power) <= (int)vid.maxtexturesize_2d && (1 << (power * 2)) < 4 * (count >> (loadmodel->brushq3.deluxemapping ? 1 : 0)); power++)
loadmodel->brushq3.num_lightmapmergepower = power;
loadmodel->brushq3.num_lightmapmerge = 1 << loadmodel->brushq3.num_lightmapmergepower;
int j, k, cnt, tri;
float *mins, *maxs;
const float *vert;
+ *collisionstride = stride;
if(stride > 0)
{
- *collisionstride = stride;
cnt = (num_triangles + stride - 1) / stride;
*collisionbbox6f = (float *) Mem_Alloc(loadmodel->mempool, sizeof(float[6]) * cnt);
for(j = 0; j < cnt; ++j)
}
}
else
- {
- *collisionstride = 0;
*collisionbbox6f = NULL;
- }
}
typedef struct patchtess_s
if (l->filelen)
{
if (l->filelen < count * (int)sizeof(*in))
- Host_Error("Mod_Q3BSP_LoadLightGrid: invalid lightgrid lump size %i bytes, should be %i bytes (%ix%ix%i)", l->filelen, (int)(count * sizeof(*in)), loadmodel->brushq3.num_lightgrid_isize[0], loadmodel->brushq3.num_lightgrid_isize[1], loadmodel->brushq3.num_lightgrid_isize[2]);
+ {
+ Con_Printf("Mod_Q3BSP_LoadLightGrid: invalid lightgrid lump size %i bytes, should be %i bytes (%ix%ix%i)", l->filelen, (int)(count * sizeof(*in)), loadmodel->brushq3.num_lightgrid_isize[0], loadmodel->brushq3.num_lightgrid_isize[1], loadmodel->brushq3.num_lightgrid_isize[2]);
+ return; // ignore the grid if we cannot understand it
+ }
if (l->filelen != count * (int)sizeof(*in))
Con_Printf("Mod_Q3BSP_LoadLightGrid: Warning: calculated lightgrid size %i bytes does not match lump size %i\n", (int)(count * sizeof(*in)), l->filelen);
out = (q3dlightgrid_t *)Mem_Alloc(loadmodel->mempool, count * sizeof(*out));
//Con_Printf("result: ambient %f %f %f diffuse %f %f %f diffusenormal %f %f %f\n", ambientcolor[0], ambientcolor[1], ambientcolor[2], diffusecolor[0], diffusecolor[1], diffusecolor[2], diffusenormal[0], diffusenormal[1], diffusenormal[2]);
}
+static int Mod_Q3BSP_TraceLineOfSight_RecursiveNodeCheck(mnode_t *node, double p1[3], double p2[3])
+{
+ double t1, t2;
+ double midf, mid[3];
+ int ret, side;
+
+ // check for empty
+ while (node->plane)
+ {
+ // find the point distances
+ mplane_t *plane = node->plane;
+ if (plane->type < 3)
+ {
+ t1 = p1[plane->type] - plane->dist;
+ t2 = p2[plane->type] - plane->dist;
+ }
+ else
+ {
+ t1 = DotProduct (plane->normal, p1) - plane->dist;
+ t2 = DotProduct (plane->normal, p2) - plane->dist;
+ }
+
+ if (t1 < 0)
+ {
+ if (t2 < 0)
+ {
+ node = node->children[1];
+ continue;
+ }
+ side = 1;
+ }
+ else
+ {
+ if (t2 >= 0)
+ {
+ node = node->children[0];
+ continue;
+ }
+ side = 0;
+ }
+
+ midf = t1 / (t1 - t2);
+ VectorLerp(p1, midf, p2, mid);
+
+ // recurse both sides, front side first
+ // return 2 if empty is followed by solid (hit something)
+ // do not return 2 if both are solid or both empty,
+ // or if start is solid and end is empty
+ // as these degenerate cases usually indicate the eye is in solid and
+ // should see the target point anyway
+ ret = Mod_Q3BSP_TraceLineOfSight_RecursiveNodeCheck(node->children[side ], p1, mid);
+ if (ret != 0)
+ return ret;
+ ret = Mod_Q3BSP_TraceLineOfSight_RecursiveNodeCheck(node->children[side ^ 1], mid, p2);
+ if (ret != 1)
+ return ret;
+ return 2;
+ }
+ return ((mleaf_t *)node)->clusterindex < 0;
+}
+
+static qboolean Mod_Q3BSP_TraceLineOfSight(struct model_s *model, const vec3_t start, const vec3_t end)
+{
+ if (model->brush.submodel || mod_q3bsp_tracelineofsight_brushes.integer)
+ {
+ trace_t trace;
+ model->TraceLine(model, NULL, NULL, &trace, start, end, SUPERCONTENTS_VISBLOCKERMASK);
+ return trace.fraction == 1;
+ }
+ else
+ {
+ double tracestart[3], traceend[3];
+ VectorCopy(start, tracestart);
+ VectorCopy(end, traceend);
+ return !Mod_Q3BSP_TraceLineOfSight_RecursiveNodeCheck(model->brush.data_nodes, tracestart, traceend);
+ }
+}
+
static void Mod_Q3BSP_TracePoint_RecursiveBSPNode(trace_t *trace, dp_model_t *model, mnode_t *node, const vec3_t point, int markframe)
{
int i;
static int markframe = 0;
-static void Mod_Q3BSP_TracePoint(dp_model_t *model, int frame, trace_t *trace, const vec3_t start, int hitsupercontentsmask)
+static void Mod_Q3BSP_TracePoint(dp_model_t *model, const frameblend_t *frameblend, const skeleton_t *skeleton, trace_t *trace, const vec3_t start, int hitsupercontentsmask)
{
int i;
q3mbrush_t *brush;
Mod_Q3BSP_TracePoint_RecursiveBSPNode(trace, model, model->brush.data_nodes, start, ++markframe);
}
-static void Mod_Q3BSP_TraceLine(dp_model_t *model, int frame, trace_t *trace, const vec3_t start, const vec3_t end, int hitsupercontentsmask)
+static void Mod_Q3BSP_TraceLine(dp_model_t *model, const frameblend_t *frameblend, const skeleton_t *skeleton, trace_t *trace, const vec3_t start, const vec3_t end, int hitsupercontentsmask)
{
int i;
float segmentmins[3], segmentmaxs[3];
if (VectorCompare(start, end))
{
- Mod_Q3BSP_TracePoint(model, frame, trace, start, hitsupercontentsmask);
+ Mod_Q3BSP_TracePoint(model, frameblend, skeleton, trace, start, hitsupercontentsmask);
return;
}
Mod_Q3BSP_TraceLine_RecursiveBSPNode(trace, model, model->brush.data_nodes, start, end, 0, 1, start, end, ++markframe, segmentmins, segmentmaxs);
}
-static void Mod_Q3BSP_TraceBox(dp_model_t *model, int frame, trace_t *trace, const vec3_t start, const vec3_t boxmins, const vec3_t boxmaxs, const vec3_t end, int hitsupercontentsmask)
+static void Mod_Q3BSP_TraceBox(dp_model_t *model, const frameblend_t *frameblend, const skeleton_t *skeleton, trace_t *trace, const vec3_t start, const vec3_t boxmins, const vec3_t boxmaxs, const vec3_t end, int hitsupercontentsmask)
{
int i;
float segmentmins[3], segmentmaxs[3];
msurface_t *surface;
q3mbrush_t *brush;
- colbrushf_t *thisbrush_start, *thisbrush_end;
+ colboxbrushf_t thisbrush_start, thisbrush_end;
vec3_t boxstartmins, boxstartmaxs, boxendmins, boxendmaxs;
if (mod_q3bsp_optimizedtraceline.integer && VectorCompare(boxmins, boxmaxs))
VectorAdd(start, boxmins, shiftstart);
VectorAdd(end, boxmins, shiftend);
if (VectorCompare(start, end))
- Mod_Q3BSP_TracePoint(model, frame, trace, shiftstart, hitsupercontentsmask);
+ Mod_Q3BSP_TracePoint(model, frameblend, skeleton, trace, shiftstart, hitsupercontentsmask);
else
{
- Mod_Q3BSP_TraceLine(model, frame, trace, shiftstart, shiftend, hitsupercontentsmask);
+ Mod_Q3BSP_TraceLine(model, frameblend, skeleton, trace, shiftstart, shiftend, hitsupercontentsmask);
VectorSubtract(trace->endpos, boxmins, trace->endpos);
}
return;
VectorAdd(start, boxmaxs, boxstartmaxs);
VectorAdd(end, boxmins, boxendmins);
VectorAdd(end, boxmaxs, boxendmaxs);
- thisbrush_start = Collision_BrushForBox(&identitymatrix, boxstartmins, boxstartmaxs, 0, 0, NULL);
- thisbrush_end = Collision_BrushForBox(&identitymatrix, boxendmins, boxendmaxs, 0, 0, NULL);
+ Collision_BrushForBox(&thisbrush_start, boxstartmins, boxstartmaxs, 0, 0, NULL);
+ Collision_BrushForBox(&thisbrush_end, boxendmins, boxendmaxs, 0, 0, NULL);
if (model->brush.submodel)
{
for (i = 0, brush = model->brush.data_brushes + model->firstmodelbrush;i < model->nummodelbrushes;i++, brush++)
if (brush->colbrushf)
- Collision_TraceBrushBrushFloat(trace, thisbrush_start, thisbrush_end, brush->colbrushf, brush->colbrushf);
+ Collision_TraceBrushBrushFloat(trace, &thisbrush_start.brush, &thisbrush_end.brush, brush->colbrushf, brush->colbrushf);
if (mod_q3bsp_curves_collisions.integer)
for (i = 0, surface = model->data_surfaces + model->firstmodelsurface;i < model->nummodelsurfaces;i++, surface++)
if (surface->num_collisiontriangles)
- Collision_TraceBrushTriangleMeshFloat(trace, thisbrush_start, thisbrush_end, surface->num_collisiontriangles, surface->data_collisionelement3i, surface->data_collisionvertex3f, surface->num_collisionbboxstride, surface->data_collisionbbox6f, surface->texture->supercontents, surface->texture->surfaceflags, surface->texture, segmentmins, segmentmaxs);
+ Collision_TraceBrushTriangleMeshFloat(trace, &thisbrush_start.brush, &thisbrush_end.brush, surface->num_collisiontriangles, surface->data_collisionelement3i, surface->data_collisionvertex3f, surface->num_collisionbboxstride, surface->data_collisionbbox6f, surface->texture->supercontents, surface->texture->surfaceflags, surface->texture, segmentmins, segmentmaxs);
}
else
- Mod_Q3BSP_TraceBrush_RecursiveBSPNode(trace, model, model->brush.data_nodes, thisbrush_start, thisbrush_end, ++markframe, segmentmins, segmentmaxs);
+ Mod_Q3BSP_TraceBrush_RecursiveBSPNode(trace, model, model->brush.data_nodes, &thisbrush_start.brush, &thisbrush_end.brush, ++markframe, segmentmins, segmentmaxs);
}
static int Mod_Q3BSP_PointSuperContents(struct model_s *model, int frame, const vec3_t point)
mod->TraceLine = Mod_Q3BSP_TraceLine;
mod->TracePoint = Mod_Q3BSP_TracePoint;
mod->PointSuperContents = Mod_Q3BSP_PointSuperContents;
- mod->brush.TraceLineOfSight = Mod_Q1BSP_TraceLineOfSight;
+ mod->brush.TraceLineOfSight = Mod_Q3BSP_TraceLineOfSight;
mod->brush.SuperContentsFromNativeContents = Mod_Q3BSP_SuperContentsFromNativeContents;
mod->brush.NativeContentsFromSuperContents = Mod_Q3BSP_NativeContentsFromSuperContents;
mod->brush.GetPVS = Mod_Q1BSP_GetPVS;
mod->Draw = R_Q1BSP_Draw;
mod->DrawDepth = R_Q1BSP_DrawDepth;
mod->DrawDebug = R_Q1BSP_DrawDebug;
+ mod->DrawPrepass = R_Q1BSP_DrawPrepass;
mod->GetLightInfo = R_Q1BSP_GetLightInfo;
mod->CompileShadowMap = R_Q1BSP_CompileShadowMap;
mod->DrawShadowMap = R_Q1BSP_DrawShadowMap;
loadmodel->brush.num_leafs = 0;
Mod_Q3BSP_RecursiveFindNumLeafs(loadmodel->brush.data_nodes);
+ if (loadmodel->brush.numsubmodels)
+ loadmodel->brush.submodels = (dp_model_t **)Mem_Alloc(loadmodel->mempool, loadmodel->brush.numsubmodels * sizeof(dp_model_t *));
+
mod = loadmodel;
for (i = 0;i < loadmodel->brush.numsubmodels;i++)
{
// textures and memory belong to the main model
mod->texturepool = NULL;
mod->mempool = NULL;
- mod->brush.TraceLineOfSight = NULL;
mod->brush.GetPVS = NULL;
mod->brush.FatPVS = NULL;
mod->brush.BoxTouchingPVS = NULL;
mod->brush.AmbientSoundLevelsForPoint = NULL;
}
mod->brush.submodel = i;
+ if (loadmodel->brush.submodels)
+ loadmodel->brush.submodels[i] = mod;
// make the model surface list (used by shadowing/lighting)
mod->firstmodelsurface = mod->brushq3.data_models[i].firstface;
Host_Error("Mod_MAP_Load: not yet implemented");
}
+#ifdef OBJWORKS
typedef struct objvertex_s
{
float v[3];
{
objvertex_t vertex[3];
int textureindex;
+ // these fields are used only in conversion to surfaces
+ int axis;
+ int surfaceindex;
+ int surfacevertexindex[3];
+ float edgeplane[3][4];
}
objtriangle_t;
+typedef objnode_s
+{
+ struct objnode_s *children[2];
+ struct objnode_s *parent;
+ objtriangle_t *triangles;
+ float normal[3];
+ float dist;
+ float mins[3];
+ float maxs[3];
+ int numtriangles;
+}
+objnode_t;
+
+objnode_t *Mod_OBJ_BSPNodeForTriangles(objnode_t *parent, objtriangle_t *triangles, int numtriangles, const float *mins, const float *maxs, mem_expandablearray_t *nodesarray, int maxclippedtriangles, objtriangle_t *clippedfronttriangles, objtriangle_t *clippedbacktriangles)
+{
+ int i, j;
+ float normal[3];
+ float dist;
+ int score;
+ float bestnormal[3];
+ float bestdist;
+ int bestscore;
+ float mins[3];
+ float maxs[3];
+ int numfronttriangles;
+ int numbacktriangles;
+ int count_front;
+ int count_back;
+ int count_both;
+ int count_on;
+ float outfrontpoints[5][3];
+ float outbackpoints[5][3];
+ int neededfrontpoints;
+ int neededbackpoints;
+ int countonpoints;
+ objnode_t *node;
+
+ node = (objnode_t *)Mem_ExpandableArray_AllocRecord(array);
+ node->parent = parent;
+ if (numtriangles)
+ {
+ VectorCopy(triangles[0].vertex[0].v, mins);
+ VectorCopy(triangles[0].vertex[0].v, maxs);
+ }
+ else if (parent && parent->children[0] == node)
+ {
+ VectorCopy(parent->mins, mins);
+ Vectorcopy(parent->maxs, maxs);
+ }
+ else if (parent && parent->children[1] == node)
+ {
+ VectorCopy(parent->mins, mins);
+ Vectorcopy(parent->maxs, maxs);
+ }
+ else
+ {
+ VectorClear(mins);
+ VectorClear(maxs);
+ }
+ for (i = 0;i < numtriangles;i++)
+ {
+ for (j = 0;j < 3;j++)
+ {
+ mins[0] = min(mins[0], triangles[i].vertex[j].v[0]);
+ mins[1] = min(mins[1], triangles[i].vertex[j].v[1]);
+ mins[2] = min(mins[2], triangles[i].vertex[j].v[2]);
+ maxs[0] = max(maxs[0], triangles[i].vertex[j].v[0]);
+ maxs[1] = max(maxs[1], triangles[i].vertex[j].v[1]);
+ maxs[2] = max(maxs[2], triangles[i].vertex[j].v[2]);
+ }
+ }
+ VectorCopy(mins, node->mins);
+ VectorCopy(maxs, node->maxs);
+ if (numtriangles <= mod_obj_leaftriangles.integer)
+ {
+ // create a leaf
+ loadmodel->brush.num_leafs++;
+ node->triangles = triangles;
+ node->numtriangles = numtriangles;
+ return node;
+ }
+
+ // create a node
+ loadmodel->brush.num_nodes++;
+ // pick a splitting plane from the various choices available to us...
+ // early splits simply halve the interval
+ bestscore = 0;
+ VectorClear(bestnormal);
+ bestdist = 0;
+ if (numtriangles <= mod_obj_splitterlimit.integer)
+ limit = numtriangles;
+ else
+ limit = 0;
+ for (i = -3;i < limit;i++)
+ {
+ if (i < 0)
+ {
+ // first we try 3 axial splits (kdtree-like)
+ j = i + 3;
+ VectorClear(normal);
+ normal[j] = 1;
+ dist = (mins[j] + maxs[j]) * 0.5f;
+ }
+ else
+ {
+ // then we try each triangle plane
+ TriangleNormal(triangles[i].vertex[0].v, triangles[i].vertex[1].v, triangles[i].vertex[2].v, normal);
+ VectorNormalize(normal);
+ dist = DotProduct(normal, triangles[i].vertex[0].v);
+ // use positive axial values whenever possible
+ if (normal[0] == -1)
+ normal[0] = 1;
+ if (normal[1] == -1)
+ normal[1] = 1;
+ if (normal[2] == -1)
+ normal[2] = 1;
+ // skip planes that match the current best
+ if (VectorCompare(normal, bestnormal) && dist == bestdist)
+ continue;
+ }
+ count_on = 0;
+ count_front = 0;
+ count_back = 0;
+ count_both = 0;
+ for (j = 0;j < numtriangles;j++)
+ {
+ dists[0] = DotProduct(normal, triangles[j].vertex[0].v) - dist;
+ dists[1] = DotProduct(normal, triangles[j].vertex[1].v) - dist;
+ dists[2] = DotProduct(normal, triangles[j].vertex[2].v) - dist;
+ if (dists[0] < -DIST_EPSILON || dists[1] < -DIST_EPSILON || dists[2] < -DIST_EPSILON)
+ {
+ if (dists[0] > DIST_EPSILON || dists[1] > DIST_EPSILON || dists[2] > DIST_EPSILON)
+ count_both++;
+ else
+ count_back++;
+ }
+ else if (dists[0] > DIST_EPSILON || dists[1] > DIST_EPSILON || dists[2] > DIST_EPSILON)
+ count_front++;
+ else
+ count_on++;
+ }
+ // score is supposed to:
+ // prefer axial splits
+ // prefer evenly dividing the input triangles
+ // prefer triangles on the plane
+ // avoid triangles crossing the plane
+ score = count_on*count_on - count_both*count_both + min(count_front, count_back)*(count_front+count_back);
+ if (normal[0] == 1 || normal[1] == 1 || normal[2] == 1)
+ score *= 2;
+ if (i == -3 || bestscore < score)
+ {
+ VectorCopy(normal, bestnormal);
+ bestdist = dist;
+ bestscore = score;
+ }
+ }
+
+ // now we have chosen an optimal split plane...
+
+ // divide triangles by the splitting plane
+ numfronttriangles = 0;
+ numbacktriangles = 0;
+ for (i = 0;i < numtriangles;i++)
+ {
+ neededfrontpoints = 0;
+ neededbackpoints = 0;
+ countonpoints = 0;
+ PolygonF_Divide(3, triangles[i].vertex[0].v, bestnormal[0], bestnormal[1], bestnormal[2], bestdist, DIST_EPSILON, 5, outfrontpoints[0], &neededfrontpoints, 5, outbackpoints[0], &neededbackpoints, &countonpoints);
+ if (countonpoints > 1)
+ {
+ // triangle lies on plane, assign it to one child only
+ TriangleNormal(triangles[i].vertex[0].v, triangles[i].vertex[1].v, triangles[i].vertex[2].v, normal);
+ if (DotProduct(bestnormal, normal) >= 0)
+ {
+ // assign to front side child
+ obj_fronttriangles[numfronttriangles++] = triangles[i];
+ }
+ else
+ {
+ // assign to back side child
+ obj_backtriangles[numbacktriangles++] = triangles[i];
+ }
+ }
+ else
+ {
+ // convert clipped polygons to triangles
+ for (j = 0;j < neededfrontpoints-2;j++)
+ {
+ obj_fronttriangles[numfronttriangles] = triangles[i];
+ VectorCopy(outfrontpoints[0], obj_fronttriangles[numfronttriangles].vertex[0].v);
+ VectorCopy(outfrontpoints[j+1], obj_fronttriangles[numfronttriangles].vertex[1].v);
+ VectorCopy(outfrontpoints[j+2], obj_fronttriangles[numfronttriangles].vertex[2].v);
+ numfronttriangles++;
+ }
+ for (j = 0;j < neededbackpoints-2;j++)
+ {
+ obj_backtriangles[numbacktriangles] = triangles[i];
+ VectorCopy(outbackpoints[0], obj_backtriangles[numbacktriangles].vertex[0].v);
+ VectorCopy(outbackpoints[j+1], obj_backtriangles[numbacktriangles].vertex[1].v);
+ VectorCopy(outbackpoints[j+2], obj_backtriangles[numbacktriangles].vertex[2].v);
+ numbacktriangles++;
+ }
+ }
+ }
+
+ // now copy the triangles out of the big buffer
+ if (numfronttriangles)
+ {
+ fronttriangles = Mem_Alloc(loadmodel->mempool, fronttriangles * sizeof(*fronttriangles));
+ memcpy(fronttriangles, obj_fronttriangles, numfronttriangles * sizeof(*fronttriangles));
+ }
+ else
+ fronttriangles = NULL;
+ if (numbacktriangles)
+ {
+ backtriangles = Mem_Alloc(loadmodel->mempool, backtriangles * sizeof(*backtriangles));
+ memcpy(backtriangles, obj_backtriangles, numbacktriangles * sizeof(*backtriangles));
+ }
+ else
+ backtriangles = NULL;
+
+ // free the original triangles we were given
+ if (triangles)
+ Mem_Free(triangles);
+ triangles = NULL;
+ numtriangles = 0;
+
+ // now create the children...
+ node->children[0] = Mod_OBJ_BSPNodeForTriangles(node, fronttriangles, numfronttriangles, frontmins, frontmaxs, nodesarray, maxclippedtriangles, clippedfronttriangles, clippedbacktriangles);
+ node->children[1] = Mod_OBJ_BSPNodeForTriangles(node, backtriangles, numbacktriangles, backmins, backmaxs, nodesarray, maxclippedtriangles, clippedfronttriangles, clippedbacktriangles);
+ return node;
+}
+
+void Mod_OBJ_SnapVertex(float *v)
+{
+ int i;
+ float a = mod_obj_vertexprecision.value;
+ float b = 1.0f / a;
+ v[0] -= floor(v[0] * a + 0.5f) * b;
+ v[1] -= floor(v[1] * a + 0.5f) * b;
+ v[2] -= floor(v[2] * a + 0.5f) * b;
+}
+
+void Mod_OBJ_ConvertBSPNode(objnode_t *objnode, mnode_t *mnodeparent)
+{
+ if (objnode->children[0])
+ {
+ // convert to mnode_t
+ mnode_t *mnode = loadmodel->brush.data_nodes + loadmodel->brush.num_nodes++;
+ mnode->parent = mnodeparent;
+ mnode->plane = loadmodel->brush.data_planes + loadmodel->brush.num_planes++;
+ VectorCopy(objnode->normal, mnode->plane->normal);
+ mnode->plane->dist = objnode->dist;
+ PlaneClassify(mnode->plane);
+ VectorCopy(objnode->mins, mnode->mins);
+ VectorCopy(objnode->maxs, mnode->maxs);
+ // push combinedsupercontents up to the parent
+ if (mnodeparent)
+ mnodeparent->combinedsupercontents |= mnode->combinedsupercontents;
+ mnode->children[0] = Mod_OBJ_ConvertBSPNode(objnode->children[0], mnode);
+ mnode->children[1] = Mod_OBJ_ConvertBSPNode(objnode->children[1], mnode);
+ }
+ else
+ {
+ // convert to mleaf_t
+ mleaf_t *mleaf = loadmodel->brush.data_leafs + loadmodel->brush.num_leafs++;
+ mleaf->parent = mnodeparent;
+ VectorCopy(objnode->mins, mleaf->mins);
+ VectorCopy(objnode->maxs, mleaf->maxs);
+ mleaf->clusterindex = loadmodel->brush.num_leafs - 1;
+ if (objnode->numtriangles)
+ {
+ objtriangle_t *triangles = objnode->triangles;
+ int numtriangles = objnode->numtriangles;
+ texture_t *texture;
+ float edge[3][3];
+ float normal[3];
+ objvertex_t vertex[3];
+ numsurfaces = 0;
+ maxsurfaces = numtriangles;
+ surfaces = NULL;
+ // calculate some more data on each triangle for surface gathering
+ for (i = 0;i < numtriangles;i++)
+ {
+ triangle = triangles + i;
+ texture = loadmodel->data_textures + triangle->textureindex;
+ Mod_OBJ_SnapVertex(triangle->vertex[0].v);
+ Mod_OBJ_SnapVertex(triangle->vertex[1].v);
+ Mod_OBJ_SnapVertex(triangle->vertex[2].v);
+ TriangleNormal(triangle->vertex[0].v, triangle->vertex[1].v, triangle->vertex[2].v, normal);
+ axis = 0;
+ if (fabs(normal[axis]) < fabs(normal[1]))
+ axis = 1;
+ if (fabs(normal[axis]) < fabs(normal[2]))
+ axis = 2;
+ VectorClear(normal);
+ normal[axis] = 1;
+ triangle->axis = axis;
+ VectorSubtract(triangle->vertex[1].v, triangle->vertex[0].v, edge[0]);
+ VectorSubtract(triangle->vertex[2].v, triangle->vertex[1].v, edge[1]);
+ VectorSubtract(triangle->vertex[0].v, triangle->vertex[2].v, edge[2]);
+ CrossProduct(edge[0], normal, triangle->edgeplane[0]);
+ CrossProduct(edge[1], normal, triangle->edgeplane[1]);
+ CrossProduct(edge[2], normal, triangle->edgeplane[2]);
+ VectorNormalize(triangle->edgeplane[0]);
+ VectorNormalize(triangle->edgeplane[1]);
+ VectorNormalize(triangle->edgeplane[2]);
+ triangle->edgeplane[0][3] = DotProduct(triangle->edgeplane[0], triangle->vertex[0].v);
+ triangle->edgeplane[1][3] = DotProduct(triangle->edgeplane[1], triangle->vertex[1].v);
+ triangle->edgeplane[2][3] = DotProduct(triangle->edgeplane[2], triangle->vertex[2].v);
+ triangle->surfaceindex = 0;
+ // add to the combined supercontents while we're here...
+ mleaf->combinedsupercontents |= texture->supercontents;
+ }
+ surfaceindex = 1;
+ for (i = 0;i < numtriangles;i++)
+ {
+ // skip already-assigned triangles
+ if (triangles[i].surfaceindex)
+ continue;
+ texture = loadmodel->data_textures + triangles[i].textureindex;
+ // assign a new surface to this triangle
+ triangles[i].surfaceindex = surfaceindex++;
+ axis = triangles[i].axis;
+ numvertices = 3;
+ // find the triangle's neighbors, this can take multiple passes
+ retry = true;
+ while (retry)
+ {
+ retry = false;
+ for (j = i+1;j < numtriangles;j++)
+ {
+ if (triangles[j].surfaceindex || triangles[j].axis != axis || triangles[j].texture != texture)
+ continue;
+ triangle = triangles + j;
+ for (k = i;k < j;k++)
+ {
+ if (triangles[k].surfaceindex != surfaceindex)
+ continue;
+ if (VectorCompare(triangles[k].vertex[0].v, triangles[j].vertex[0].v)
+ || VectorCompare(triangles[k].vertex[0].v, triangles[j].vertex[1].v)
+ || VectorCompare(triangles[k].vertex[0].v, triangles[j].vertex[2].v)
+ || VectorCompare(triangles[k].vertex[1].v, triangles[j].vertex[0].v)
+ || VectorCompare(triangles[k].vertex[1].v, triangles[j].vertex[1].v)
+ || VectorCompare(triangles[k].vertex[1].v, triangles[j].vertex[2].v)
+ || VectorCompare(triangles[k].vertex[2].v, triangles[j].vertex[0].v)
+ || VectorCompare(triangles[k].vertex[2].v, triangles[j].vertex[1].v)
+ || VectorCompare(triangles[k].vertex[2].v, triangles[j].vertex[2].v))
+ {
+ // shares a vertex position
+ --- FIXME ---
+ }
+ }
+ for (k = 0;k < numvertices;k++)
+ if (!VectorCompare(vertex[k].v, triangles[j].vertex[0].v) || !VectorCompare(vertex[k].v, triangles[j].vertex[1].v) || !VectorCompare(vertex[k].v, triangles[j].vertex[2].v))
+ break;
+ if (k == numvertices)
+ break; // not a neighbor
+ // this triangle is a neighbor and has the same axis and texture
+ // check now if it overlaps in lightmap projection space
+ triangles[j].surfaceindex;
+ if (triangles[j].
+ }
+ }
+ //triangles[i].surfaceindex = surfaceindex++;
+ for (surfaceindex = 0;surfaceindex < numsurfaces;surfaceindex++)
+ {
+ if (surfaces[surfaceindex].texture != texture)
+ continue;
+ // check if any triangles already in this surface overlap in lightmap projection space
+
+ {
+ }
+ break;
+ }
+ }
+ // let the collision code simply use the surfaces
+ mleaf->containscollisionsurfaces = mleaf->combinedsupercontents != 0;
+ mleaf->numleafsurfaces = ?;
+ mleaf->firstleafsurface = ?;
+ }
+ // push combinedsupercontents up to the parent
+ if (mnodeparent)
+ mnodeparent->combinedsupercontents |= mleaf->combinedsupercontents;
+ }
+}
+#endif
+
void Mod_OBJ_Load(dp_model_t *mod, void *buffer, void *bufferend)
{
-#if 0
+#ifdef OBJWORKS
const char *textbase = (char *)buffer, *text = textbase;
char *s;
char *argv[512];
float *vt = Mem_Alloc(tempmempool, maxvt * sizeof(float[2]));
float *vn = Mem_Alloc(tempmempool, maxvn * sizeof(float[3]));
objvertex_t vfirst, vprev, vcurrent;
+ float mins[3];
+ float maxs[3];
#if 0
int hashindex;
int maxverthash = 65536, numverthash = 0;
loadmodel->Draw = R_Q1BSP_Draw;
loadmodel->DrawDepth = R_Q1BSP_DrawDepth;
loadmodel->DrawDebug = R_Q1BSP_DrawDebug;
+ loadmodel->DrawPrepass = R_Q1BSP_DrawPrepass;
loadmodel->GetLightInfo = R_Q1BSP_GetLightInfo;
loadmodel->CompileShadowMap = R_Q1BSP_CompileShadowMap;
loadmodel->DrawShadowMap = R_Q1BSP_DrawShadowMap;
loadmodel->DrawShadowVolume = R_Q1BSP_DrawShadowVolume;
loadmodel->DrawLight = R_Q1BSP_DrawLight;
+ VectorClear(mins);
+ VectorClear(maxs);
+
// parse the OBJ text now
for(;;)
{
VectorCopy(v + 3*index1, vcurrent.v);
Vector2Copy(vt + 2*index2, vcurrent.vt);
VectorCopy(vn + 3*index3, vcurrent.vn);
+ if (numtriangles == 0)
+ {
+ VectorCopy(vcurrent.v, mins);
+ VectorCopy(vcurrent.v, maxs);
+ }
+ else
+ {
+ mins[0] = min(mins[0], vcurrent.v[0]);
+ mins[1] = min(mins[1], vcurrent.v[1]);
+ mins[2] = min(mins[2], vcurrent.v[2]);
+ maxs[0] = max(maxs[0], vcurrent.v[0]);
+ maxs[1] = max(maxs[1], vcurrent.v[1]);
+ maxs[2] = max(maxs[2], vcurrent.v[2]);
+ }
if (j == 1)
vfirst = vcurrent;
else if (j >= 3)
Mem_Free(texturenames);
texturenames = NULL;
+ // copy the model bounds, then enlarge the yaw and rotated bounds according to radius
+ VectorCopy(mins, loadmodel->normalmins);
+ VectorCopy(maxs, loadmodel->normalmaxs);
+ dist = max(fabs(loadmodel->normalmins[0]), fabs(loadmodel->normalmaxs[0]));
+ modelyawradius = max(fabs(loadmodel->normalmins[1]), fabs(loadmodel->normalmaxs[1]));
+ modelyawradius = dist*dist+modelyawradius*modelyawradius;
+ modelradius = max(fabs(loadmodel->normalmins[2]), fabs(loadmodel->normalmaxs[2]));
+ modelradius = modelyawradius + modelradius * modelradius;
+ modelyawradius = sqrt(modelyawradius);
+ modelradius = sqrt(modelradius);
+ loadmodel->yawmins[0] = loadmodel->yawmins[1] = -modelyawradius;
+ loadmodel->yawmins[2] = loadmodel->normalmins[2];
+ loadmodel->yawmaxs[0] = loadmodel->yawmaxs[1] = modelyawradius;
+ loadmodel->yawmaxs[2] = loadmodel->normalmaxs[2];
+ loadmodel->rotatedmins[0] = loadmodel->rotatedmins[1] = loadmodel->rotatedmins[2] = -modelradius;
+ loadmodel->rotatedmaxs[0] = loadmodel->rotatedmaxs[1] = loadmodel->rotatedmaxs[2] = modelradius;
+ loadmodel->radius = modelradius;
+ loadmodel->radius2 = modelradius * modelradius;
+
+ // make sure the temp triangle buffer is big enough for BSP building
+ maxclippedtriangles = numtriangles*4;
+ if (numtriangles > 0)
+ {
+ clippedfronttriangles = Mem_Alloc(loadmodel->mempool, maxclippedtriangles * 2 * sizeof(objtriangle_t));
+ clippedbacktriangles = clippedfronttriangles + maxclippedtriangles;
+ }
+
// generate a rough BSP tree from triangle data, we don't have to be too careful here, it only has to define the basic areas of the map
+ loadmodel->brush.num_leafs = 0;
+ loadmodel->brush.num_nodes = 0;
+ Mem_ExpandableArray_NewArray(&nodesarray, loadmodel->mempool, sizeof(objnode_t), 1024);
+ rootnode = Mod_OBJ_BSPNodeForTriangles(triangles, numtriangles, mins, maxs, &nodesarray, maxclippedtriangles, clippedfronttriangles, clippedbacktriangles);
+
+ // convert the BSP tree to mnode_t and mleaf_t structures and convert the triangles to msurface_t...
+ loadmodel->brush.data_leafs = Mem_Alloc(loadmodel->mempool, loadmodel->brush.num_leafs * sizeof(mleaf_t));
+ loadmodel->brush.data_nodes = Mem_Alloc(loadmodel->mempool, loadmodel->brush.num_nodes * sizeof(mnode_t));
+ loadmodel->brush.data_planes = Mem_Alloc(loadmodel->mempool, loadmodel->brush.num_nodes * sizeof(mplane_t));
+ loadmodel->brush.num_leafs = 0;
+ loadmodel->brush.num_nodes = 0;
+ loadmodel->brush.num_planes = 0;
+ Mod_OBJ_ConvertAndFreeBSPNode(rootnode);
+
+ if (clippedfronttriangles)
+ Mem_Free(clippedfronttriangles);
+ maxclippedtriangles = 0;
+ clippedfronttriangles = NULL;
+ clippedbacktriangles = NULL;
- // generate surfaces by recursing triangles into BSP tree and ensuring they do not overlap in the lightmap projection axis
+--- NOTHING DONE PAST THIS POINT ---
loadmodel->numskins = LittleLong(pinmodel->num_skins);
numxyz = LittleLong(pinmodel->num_xyz);
projects / xonotic / darkplaces.git / blobdiff