Subversion Repositories gelsvn

Rev

Go to most recent revision | Show changed files | Details | Compare with Previous | Blame | RSS feed

Filtering Options

Rev Age Author Path Log message Diff
325 6712 d 15 h jab /trunk/src/CGLA/ An alternative 4x4 determinant function is included and commented out.  
324 6713 d 13 h jab /trunk/src/ Some changes. Mostly a nasty precision issue has been fixed with regard
to 4x4 determinants. KDTrees have also been improved.		  
323 6719 d 7 h bj /trunk/src/CGLA/ added generic make_rot method with rotation matrix as input  
321 6721 d 3 h jab /trunk/ No log message.  
320 6728 d 13 h awk /trunk/src/CGLA/  
319 6729 d 3 h awk /trunk/src/Geometry/ Partitioned interface into public / private..
Changed all TriMesh* to const TriMesh*
Commented normal computation code in intersect in..
Fixed bug with normals (object space normals were returned, not world space)		  
318 6729 d 3 h awk /trunk/src/Geometry/ Changed hit_object to const pointer to TriMesh  
317 6731 d 5 h awk /trunk/src/CGLA/  
316 6734 d 7 h jab /trunk/src/CGLA/ Added a constructor to Vec2i that accepts a single int.  
315 6734 d 13 h jab /trunk/ Some slight changes to QEM based decimation  
314 6767 d 14 h jrf /trunk/src/Geometry/  
313 6782 d 8 h jrf /trunk/ BSPTree ray tracing now works well !
The problem was a way too large epsilon causing triangles to appear unnecessarily in several different space division boxes. Precision has now been increased and the code has been made more robust. Try it out on the bunny.		  
311 6782 d 10 h jrf /trunk/src/Geometry/  
310 6782 d 10 h jrf /trunk/ BSPTree corrections  
309 6786 d 12 h jab /trunk/ numerous changes making ray tracing work with bbox hierarchy  
306 6788 d 10 h jrf /trunk/ quaternion_cleanup branch has been merged with main branch  
304 6788 d 12 h jab /trunk/src/ I have made minor changes to ensure that GEL compiles on Linux.  
299 6800 d 12 h jrf /trunk/ The SPD branch has been merged with the main branch.
This means that Bounding Box Hierarchies, BSP Trees, and two new applications have been added.		  
295 6800 d 13 h jrf /trunk/src/Geometry/  
277 6880 d 14 h jrf /trunk/src/ The int_pow function in CGLA has been updated to run with O(log(n)) complexity.

The QuatTrackBall has been updated.		  
271 6894 d 11 h bj /trunk/ removed .vcproj and .sln files from SVN  
248 6940 d 13 h bj /trunk/src/CGLA/  
230 6978 d 10 h bj /trunk/src/Geometry/ resolving jab sloppyness ;-)  
229 6978 d 10 h bj /trunk/src/HMesh/ warnings removed  
223 6984 d 12 h bj /trunk/ resolved LIBC default lib issue with a #pragma
clapack.lib needs to have libc ignored		  
222 6996 d 8 h jab /trunk/src/ minor changes  
220 6996 d 11 h jab /trunk/ Fixed QEM problems  
219 6996 d 12 h bj /trunk/src/HMesh/ #include "Util/BinaryHeap.h" removed  
215 6997 d 10 h jab /trunk/src/HMesh/ Our version of QSlim has been greatly improved. It now performs
consistency checks to avoid bad edge collapses which will flip
triangles. Another fixed problem was a voracious memory hunger.

Fuzzy vector median smoothing has been added.

Other changes are minor.		  
214 6997 d 10 h jab /trunk/src/Geometry/ Load raw can now load a signed short volume
Elsewhere Vec3d has replaced Vec3f		  
211 7086 d 11 h jab /trunk/src/ Removed IMesh from documents  
206 7094 d 13 h jrf /trunk/src/CGLA/ Added a constructor to Vec3f taking a Vec4f as argument. The last element of the Vec4f is simply thrown away.

Added a Vec2ui. A 2D vector of unsigned integers is useful for specifying dimensions.		  
204 7130 d 13 h jrf /trunk/src/GLGraphics/ Minor bug corrected with respect to zero rotation  
203 7133 d 14 h jrf /trunk/src/Geometry/ pragmas corrected  
198 7137 d 12 h bj /trunk/ Added Visual Studio 2005 compatibility  
197 7138 d 11 h jrf /trunk/src/CGLA/ Redefinition of CGLA::isnan to have better cross platform support  
196 7138 d 11 h jrf /trunk/src/CGLA/ cgla_nan() has been corrected. A minor bug caused it to return 0.0 rather than nan ...  
195 7138 d 14 h jrf /trunk/ warning pragmas corrected.
Added messages stating that some preprocessor directives are needed for HMesh to work with MSVC8.		  
192 7144 d 13 h jab /trunk/ I made some changes to the polygonizer.  
191 7144 d 15 h jab /trunk/ So some changes have been made. Things are partly foobar in the volume
polygonizer and the trilinfilter. There is a dilemma - it is hard to
allow trilinear interpolation of vectors while disallowing gradient
computation of unsigned entities.