eessppeelllloo/processing4
1
0
Fork 0
mirror of https://github.com/processing/processing4.git synced 2026-02-14 02:45:36 +01:00
Code Issues Packages Projects Releases Wiki Activity

fix several triangulation issues, bugs #97, #390, and #111

Browse Source
This commit is contained in:
benfry
2006-11-05 15:49:20 +00:00
parent a441b32606
commit a5eedba1d8
3 changed files with 113 additions and 63 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
core/src/processing/core/PGraphics.java
View File

@@ -1078,7 +1078,7 @@ public abstract class PGraphics extends PImage implements PConstants {
}
// when the coords are Float.MAX_VALUE, then treat as a 2D curve
int dimensions = (x == Float.MAX_VALUE) ? 2 : 3;
int dimensions = (z == Float.MAX_VALUE) ? 2 : 3;
if (dimensions == 3) {
vertex[MZ] = z;

109
core/src/processing/core/PGraphics3D.java
View File

@@ -553,9 +553,19 @@ public class PGraphics3D extends PGraphics {
vertex_order = temp2;
//message(CHATTER, "allocating more vertices " + vertices.length);
}
float vertex[] = vertices[vertexCount++];
float vertex[] = vertices[vertexCount];
//if (polygon.redundantVertex(x, y, z)) return;
if (vertexCount > 0) {
float pvertex[] = vertices[vertexCount-1];
if ((abs(pvertex[MX] - x) < EPSILON) &&
(abs(pvertex[MY] - y) < EPSILON) &&
(abs(pvertex[MZ] - z) < EPSILON)) {
// this vertex is identical, don't add it,
// because it will anger the triangulator
return;
}
}
// user called vertex(), so that invalidates anything queued
// up for curve vertices. if this is internally called by
@@ -606,6 +616,8 @@ public class PGraphics3D extends PGraphics {
vertex[NZ] = normalZ;
vertex[BEEN_LIT] = 0;
vertexCount++;
}
@@ -832,16 +844,17 @@ public class PGraphics3D extends PGraphics {
//case CONVEX_POLYGON:
{
// store index of first vertex
int first = lineCount;
//int first = lineCount;
stop = vertex_end - 1;
add_path();
for (int i = vertex_start; i < stop; i++) {
add_line(i, i+1);
//System.out.println("adding line " + i);
}
if (mode == CLOSE) {
// draw the last line connecting back to the first point in poly
add_line(stop, lines[first][VERTEX1]);
add_line(stop, vertex_start); //lines[first][VERTEX1]);
}
}
break;
@@ -918,7 +931,7 @@ public class PGraphics3D extends PGraphics {
// TRANSFORM / LIGHT / CLIP
if (lightCount > 0 && fill) {
handle_lighting();
handle_lighting();
}
else {
handle_no_lighting();
@@ -1452,18 +1465,72 @@ public class PGraphics3D extends PGraphics {
* bit of code from the web.
*/
private void triangulate_polygon() {
// this clipping algorithm only works in 2D, so in cases where a
// polygon is drawn perpendicular to the z-axis, the area will be zero,
// and triangulation will fail. as such, when the area calculates to
// zero, figure out whether x or y is empty, and calculate based on the
// two dimensions that actually contain information.
// http://dev.processing.org/bugs/show_bug.cgi?id=111
int d1 = MX;
int d2 = MY;
// this brings up the nastier point that there may be cases where
// a polygon is irregular in space and will throw off the
// clockwise/counterclockwise calculation. for instance, if clockwise
// relative to x and z, but counter relative to y and z or something
// like that.. will wait to see if this is in fact a problem before
// hurting my head on the math.
// first we check if the polygon goes clockwise or counterclockwise
float area = 0.0f;
float area = 0;
for (int p = vertex_end - 1, q = vertex_start; q < vertex_end; p = q++) {
area += (vertices[q][MX] * vertices[p][MY] -
vertices[p][MX] * vertices[q][MY]);
//area += (vertices[q][X] * vertices[p][Y] -
// vertices[p][X] * vertices[q][Y]);
area += (vertices[q][d1] * vertices[p][d2] -
vertices[p][d1] * vertices[q][d2]);
}
// rather than checking for the perpendicular case first, only do it
// when the area calculates to zero. checking for perpendicular would be
// a needless waste of time for the 99% case.
if (area == 0) {
// figure out which dimension is the perpendicular axis
boolean foundValidX = false;
boolean foundValidY = false;
for (int i = vertex_start; i < vertex_end; i++) {
if (vertices[i][MX] != 0) foundValidX = true;
if (vertices[i][MY] != 0) foundValidY = true;
}
if (foundValidX && !foundValidY) {
//d1 = MX; // already the case
d2 = MZ;
} else if (!foundValidX && foundValidY) {
// ermm.. which is the proper order for cw/ccw here?
d1 = MY;
d2 = MZ;
} else {
// screw it, this polygon is just f-ed up
return;
}
// re-calculate the area, with what should be good values
for (int p = vertex_end - 1, q = vertex_start; q < vertex_end; p = q++) {
area += (vertices[q][d1] * vertices[p][d2] -
vertices[p][d1] * vertices[q][d2]);
}
}
// don't allow polygons to come back and meet themselves,
// otherwise it will anger the triangulator
// http://dev.processing.org/bugs/show_bug.cgi?id=97
float vfirst[] = vertices[vertex_start];
float vlast[] = vertices[vertex_end-1];
if ((abs(vfirst[MX] - vlast[MX]) < EPSILON) &&
(abs(vfirst[MY] - vlast[MY]) < EPSILON) &&
(abs(vfirst[MZ] - vlast[MZ]) < EPSILON)) {
vertex_end--;
}
// then sort the vertices so they are always in a counterclockwise order
int j = 0;
//if (0.0f < area) { // def <
if (area > 0) {
for (int i = vertex_start; i < vertex_end; i++) {
j = i - vertex_start;
@@ -1494,14 +1561,12 @@ public class PGraphics3D extends PGraphics {
int w = v + 1; if (vc <= w) w = 0; // next
// triangle A B C
//float Ax, Ay, Bx, By, Cx, Cy, Px, Py;
float Ax = -vertices[vertex_order[u]][MX];
float Ay = vertices[vertex_order[u]][MY];
float Bx = -vertices[vertex_order[v]][MX];
float By = vertices[vertex_order[v]][MY];
float Cx = -vertices[vertex_order[w]][MX];
float Cy = vertices[vertex_order[w]][MY];
float Ax = -vertices[vertex_order[u]][d1];
float Ay = vertices[vertex_order[u]][d2];
float Bx = -vertices[vertex_order[v]][d1];
float By = vertices[vertex_order[v]][d2];
float Cx = -vertices[vertex_order[w]][d1];
float Cy = vertices[vertex_order[w]][d2];
// first we check if <u,v,w> continues going ccw
if (EPSILON > (((Bx-Ax) * (Cy-Ay)) - ((By-Ay) * (Cx-Ax)))) {
@@ -1516,10 +1581,8 @@ public class PGraphics3D extends PGraphics {
continue;
}
//float Px = -vertices[vertex_order[p]][X];
//float Py = vertices[vertex_order[p]][Y];
float Px = -vertices[vertex_order[p]][MX];
float Py = vertices[vertex_order[p]][MY];
float Px = -vertices[vertex_order[p]][d1];
float Py = vertices[vertex_order[p]][d2];
float ax = Cx - Bx; float ay = Cy - By;
float bx = Ax - Cx; float by = Ay - Cy;
@@ -3672,11 +3735,11 @@ public class PGraphics3D extends PGraphics {
return (float)Math.pow(a, b);
}
/*
private final float abs(float a) {
return (a < 0) ? -a : a;
}
/*
private final float sin(float angle) {
return (float)Math.sin(angle);
}

65
core/todo.txt
View File

@@ -5,9 +5,24 @@ X null incompatible with Global antialiasing enable key
X fix issue where ambientLight(r, g, b) was instead ambientLight(r, g, r)
X http://dev.processing.org/bugs/show_bug.cgi?id=412
X createFont() should always use native fonts
_ need to warn that fonts may not be installed
_ recommend that people include the ttf if that's the thing
_ or rather, that this is only recommended for offline use
X need to warn that fonts may not be installed
X recommend that people include the ttf if that's the thing
X or rather, that this is only recommended for offline use
X fix 3D tessellation problems with curveVertex and bezierVertex
X actually was z = Float.MAX_VALUE regression
X http://dev.processing.org/bugs/show_bug.cgi?id=390
X two examples in sketchbook
X this has been reported several times
X concave polygons having trouble if points come back to meet
X tesselator/triangulator gets confused when points doubled up
X might need to avoid previous vertex hitting itself
X http://dev.processing.org/bugs/show_bug.cgi?id=97
X graphics gems 5 has more about tessellation
X polygons perpendicular to axis not drawing
X is this a clipping error?
X probably a triangulation error, because triangles work ok
X http://dev.processing.org/bugs/show_bug.cgi?id=111
X problem is that the area of the polygon isn't taking into account z
_ saveFrame() produces a black background because bg not set correctly:
_ http://dev.processing.org/bugs/show_bug.cgi?id=421
@@ -28,15 +43,6 @@ _ in which case it should be set to opaque
_ have createGraphics() create a completely transparent image
_ and also not require defaults() to be called
_ this produces a dark blue background:
colorMode(RGB, 100);
background(128);
_ fix 3D tessellation problems with curveVertex and bezierVertex
_ http://dev.processing.org/bugs/show_bug.cgi?id=390
_ two examples in sketchbook
_ this has been reported several times
_ image outofmemoryerror for casey's students
_ http://dev.processing.org/bugs/show_bug.cgi?id=355
_ this may be related to a ton of other memory bugs
@@ -47,26 +53,9 @@ _ maybe the image binding not getting unbound?
_ loading lots of images is a problem, describe how to unload
_ is it possible? necessary to call delay(5) or something?
_ make this work for JAVA2D and P3D
_ createGraphics() having problems with JAVA2D, and sometimes with P3D
_ http://dev.processing.org/bugs/show_bug.cgi?id=419
PGraphics pg;
void setup() {
size(100, 100);
pg = createGraphics(80, 80, P3D);
pg.beginDraw();
pg.background(102);
pg.stroke(255);
pg.line(20, 20, 80, 80);
pg.endDraw();
noLoop();
}
void draw() {
image(pg, 10, 10);
}
_ detect when using full screen
_ and if so, remove decoration and don't bother with bg present frame
_ frame.setUndecorated(true);
@@ -110,6 +99,7 @@ _ http://processing.org/discourse/yabb_beta/YaBB.cgi?board=SoftwareBugs;action
_ alloc() stuff not fixed because of thread halting
_ problem where alloc happens inside setup(), so, uh..
_ http://dev.processing.org/bugs/show_bug.cgi?id=369
_ should instead create new buffer, and swap it in next time through
opengl texture handling
_ maybe break out a separate timage object in textures?
@@ -775,15 +765,6 @@ _ rect() changes size as it changes position
_ http://dev.processing.org/bugs/show_bug.cgi?id=95
_ lines skip on 200x200 surface because of fixed point rounding error
_ http://dev.processing.org/bugs/show_bug.cgi?id=267
_ polygons perpendicular to axis not drawing
_ is this a clipping error?
_ probably a triangulation error, because triangles work ok
_ http://dev.processing.org/bugs/show_bug.cgi?id=111
_ concave polygons having trouble if points come back to meet
_ tesselator/triangulator gets confused when points doubled up
_ might need to avoid previous vertex hitting itself
_ http://dev.processing.org/bugs/show_bug.cgi?id=97
_ graphics gems 5 has more about tessellation
_ P3D not doing bilinear interpolation in text and images
_ because smooth() has to be set (and smooth throws an error in P3D)
_ how should this be handled? a hint? allowing smooth()?
@@ -942,3 +923,9 @@ _ pdf export ignoring transparency on linux
_ check to see if this is the case on other linux machines
_ seems to be working fine on windows
_ http://dev.processing.org/bugs/show_bug.cgi?id=345
LATER
_ setting a gray that's greater than the colorMode() can produce strange colors
_ http://dev.processing.org/bugs/show_bug.cgi?id=432