Generated by StatCvs v0.2-dev
Pages: << Previous 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 Next >>
Summary Period: 2002-12-11 to 2002-12-09 (Commits 10045-10094 of 12744)
Eliminated compilation warnings.
2 lines of code changed in:
Removed incorrect run-time dependency from netmtst.mak target. Only
compile-time dependencies should be specified.
5 lines of code changed in:
fixed g2dtest - however it seems the line clipping routines behave wrong (or the test... someone should investigate this)
2 lines of code changed in:
Eliminated compilation warning.
2 lines of code changed in:
Eliminated compilation warning.
1 lines of code changed in:
Eliminated compilation warnings.
4 lines of code changed in:
Correction for command-line invocation on MacOS/X.
1 lines of code changed in:
fixed a circular Ref: TxtMgr->texture->TxtMgr. Opengl plugin unloads correctly now and software should leak less...
8 lines of code changed in:
fixed a bug where CS crashs when a curve has no lightmap assigned
6 lines of code changed in:
fixed a bug in csBasicVector::Pop which wasn't testing if root==NULL. opengl should work again.
3 lines of code changed in:
Partially fixed a bug in Dynavis which causes lots of potential
loss in occlusion. The problem happend with occluders which
go beyond the screen at the right side. This fix should
improve performance of Dynavis levels a bit. Note that the
fix is not complete. I only fixed a few common trivial cases.
The hard general case hasn't been fixed yet.
45 lines of code changed in:
Made sure that the normal calculation in terrfunc first
clamps the x,y values given to the height function to avoid
crashing. Thanks to Joseph Miller for pointing me to the bug
and suggesting a fix (which I didn't use).
23 lines of code changed in:
Update.
2 lines of code changed in:
Another compile fix.
2 lines of code changed in:
Fixed compile error in recent addition to engine sequence manager.
3 lines of code changed in:
- Optimization in frustvis where the updates of objects in the
visibility culler are put on a queue and only performed when
actually needed. This means that movable->UpdateMove() no longer
causes a kdtree rebuild every time.
- Fixed a rather big bug in the frustvis kdtree->Flatten() function.
It would not properly unlink objects from the parent leafs in all
cases. Also the leafs were not destructed which is a memory
leak.
- Optimization in the frustvis kdtree so that MoveObject()
flattens() the affected part of the tree every 50 times it is
called. This greatly increases the quality of the tree while
keeping the overhead of calling Flatten() considerably.
- Optimized frustvis->Flatten() by flattening directly to the
top-node instead of doing this in a recursive manner.
- The result of all the optimizations in the kdtree means that
walktest flarge will no longer keep slowing down when looking
at the robot.
- Added all frustvis/kdtree changes above also to dynavis/kdtree.
267 lines of code changed in:
Added <lightcolor> trigger condition tag so now sequences can be
triggered by changes in light colors.
7 lines of code changed in:
Added <lightcolor> trigger condition tag so now sequences can be
triggered by changes in light colors.
13 lines of code changed in:
Added iLightCallback access list. These callbacks are called whenever
a light changes color.
86 lines of code changed in:
Added <lightcolor> trigger condition tag so now sequences can be
triggered by changes in light colors. Corridor on 2nd level of large room
now brightens its ambient light whenever the light by the stairs gets
very dim, and turns back down again as the stairs get brighter.
109 lines of code changed in:
Added <lightcolor> trigger condition tag so now sequences can be
triggered by changes in light colors. Corridor on 2nd level of large room
now brightens its ambient light whenever the light by the stairs gets
very dim, and turns back down again as the stairs get brighter.
39 lines of code changed in:
Added iLightCallback access list. These callbacks are called whenever
a light changes color.
11 lines of code changed in:
Added FindLight function to LoaderContext because standard FindLight
returns iStatLight instead of iLight.
16 lines of code changed in:
Added <lightcolor> trigger condition tag so now sequences can be
triggered by changes in light colors.
7 lines of code changed in:
Added iLightCallback access list. These callbacks are called whenever
a light changes color.
30 lines of code changed in:
Added <lightcolor> trigger condition tag so now sequences can be
triggered by changes in light colors.
54 lines of code changed in:
Removed SCF_DEBUG define.
0 lines of code changed in:
Eric Sunshine applied a patch from Andreas Busch <nobita@t-online.de> to
convert the cslua plugin over to csRef/csPtr.
2 lines of code changed in:
Eric Sunshine applied a large and thorough typographical-error correction
patch to the Texinfo documentation from Andreas Busch
<nobita@t-online.de>. Also applied a few minor corrections from Eric
Sunshine.
271 lines of code changed in:
Automated Visual-C++ SLN and VCPROJ project file repair.
2 lines of code changed in:
fixed a refcount bug in the opengl canvas
1 lines of code changed in:
cleaned up effects server code
231 lines of code changed in:
cleaned SCF usage of effects code a bit
26 lines of code changed in:
Made the opengl plugin abort if it wasn't able to find the effectserver plugin.
11 lines of code changed in:
- Javier Santos fixed another bug in 3ds2lev where the
<lighting> option wouldn't be correctly saved to XML.
- Fixed a memory leak with texture layers for a material. There
was no DecRef(). I fixed this by using csRef.
6 lines of code changed in:
fixed a big memory leak in the engine (MeshWrappers have been never released)
1 lines of code changed in:
- <goodoccluder> was not parsed correctly.
- Implemented -flaggood, -flagbad, and -flagclear options for
levtool. With these options you can control the placement
of <goodoccluder> and <badoccluder> in things. Check out
./levtool -help for more info on how to use these options.
- Javier Santos fixed an initialization bug in 3ds2lev.
229 lines of code changed in:
Fixed a crash bug in iso engine. An IncRef() was recently
removed.
4 lines of code changed in:
- When a good occluder is found, Dynavis will now use the write
queue but with the minimum depth instead of maximum depth.
That improves the usage of the coverage buffer but still maintains
the good potential for using the good occluder polygons.
- Fixed a problem that Dynavis would be using outline coverage
or polygon coverage culling based on wrong information. This
happens when inserting polygons from the write queue.
- Also fixed a similar problem with comparing a bool with a
non-bool value for coverage buffer insertion.
37 lines of code changed in:
- Removed the call to 'Fire()' from Trigger() in the engine sequence
manager. That's not needed. Additionally fixed FireTriggerByName()
so it works correctly now. To do that I added
iSequenceTrigger::ForceFire() which forces a trigger to be fired
no matter what conditions are true. This is the best way to
fire the sequence as the trigger holds the parameters required
by some sequences.
- Added 'now' parameter to FireTriggerByName() (engine sequence
manager).
70 lines of code changed in:
- Fixed a bug I introduced yesterday in parsing of HardMove.
The transform was inverted.
- Apparently the meaning of HARDMOVE was inverted for factories.
I changed HARDMOVE for factories so that the matrix is inverted.
This means that if you have a <meshfact> you now have to invert
your matrices. For example, a matrix that scales .5 will now
have to be a matrix that multiplies with 2. That way it is more
consistent with HARDMOVE for <meshobj> but also it is more
correct.
22 lines of code changed in:
Added direct triggering of sequences and direct running of sequences.
49 lines of code changed in:
Added lightning command, which directly runs a sequence in flarge.
37 lines of code changed in:
Forgot to commit.
1 lines of code changed in:
Fixed bug in animating frames when using mixed modes of displacements
and delays in the same animation.
6 lines of code changed in:
- Added iVisibilityObject::GetCullerFlags(). With this you can set
various hints for the visibility culler to help optimize
performance. The visibility culler can choose to ignore or honor
these hints. The following hint flags are defined at this moment
but no visibility culler honors them yet:
- CS_CULLER_HINT_CONVEX: object is convex.
- CS_CULLER_HINT_CLOSED: object is closed (i.e. if you are
outside the object you can only see the visible side
of the object.
- CS_CULLER_HINT_GOODOCCLUDER: this object is a good occluder
(i.e. it will probably be able to cull away a lot of the
geometry).
- CS_CULLER_HINT_BADOCCLUDER: this object is a bad occluder
(i.e. it is too small or not convex enough to be a useful
occluder).
- Generalized parameter parsing for mesh objects. A large part of
this was now duplicated in two routines (csloader plugin).
- The <convex/> flag now also sets CS_CULLER_HINT_CONVEX.
- Added <closed/>, <goodoccluder/>, and <badoccluder/> flags for
mesh objects.
- Dynavis now honors the <badoccluder/> and <closed/> flags. In
case of <closed/> it will use outline based culling in any case.
In case of <badoccluder/> it will simply not use the object
as an occluder.
- Dynavis now also honors the <goodoccluder/> flag. In that
case it will not use the write queue to delay writing of the
occluder and it will also use polygon mode instead of outline
mode to have more accurate filling of the coverage buffer.
The idea is to only use this for large objects that don't have
too many small polygons.
414 lines of code changed in:
- Made the following modifications to levtool:
- Added -splitpoly and -analyzep options. With this it is possible
to seperate large polygons in seperate objects.
- Fixed bug with -minsize. That option was simply ignored.
- Experimental optimation in Dynavis for clamp in polygon mode (not
yet outline mode). Not enabled yet since it isn't working properly
yet.
- Fixed a bug in Dynavis where it would consider polygons as possible
occluders even if Z very close to 0. This could cause bad overflow.
270 lines of code changed in:
Automated Visual-C++ SLN and VCPROJ project file repair.
2 lines of code changed in:
* Added test for ARB_vertex_program
* Added new objects for render3d-shaders
7 lines of code changed in:
* Added test for ARB_vertex_program
119 lines of code changed in:
Pages: << Previous 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 Next >>
Generated by StatCvs v0.2-dev