| Rev |
Age |
Author |
Path |
Log message |
Diff |
| 602 |
4623 d 13 h |
jab |
/trunk/ |
Removed the stale app rangescanviewer. Several changes to the XCode project which now compiles a proper framework for GEL and also compiles all test apps. |
|
| 601 |
4625 d 0 h |
jab |
/trunk/ |
The include statements for header files have been changed. Instead of including, like say
#include <CGLA/Vec3f.h>
we now use
#include "../CGLA/Vec3f.h"
for all files in the GEL library source tree (i.e. apps and test are not altered).
The point is that if GEL is used as an OSX framework and I include a GEL header, I have to do it like this:
#include <GEL/HMesh/myheader.h>
Now, inside myheader.h I may include CGLA/Vec3f.h, but how should this file be found? If I just add the GEL
framework as a framework, the path to the GEL headers is not added to the header search path because the
framework name (in this case GEL) is part of the header path. If all headers had been in one directory that
would not have mattered, since Vec3f.h would be in the same directory as myheader.h. But it is not. So, I
have made things more relative and it works well. |
|
| 600 |
4627 d 7 h |
jab |
/ |
Build system. I have both edited the bespoke Xcode project file, played with a premake4 build generator and our cmake system. Nothing works perfectly, but cmake might end up as the best solution. The cmake file now automatically finds the GEL files, but this entails that cmake is run when a new file is added.
Functions were added to Manifold: remove_face, remove_edge, remove_vertex. These do as advertised. I also added an add_face function (not tested but very simple ... hmmm famous last words)
the cpp/h file pairs close_holes and caps_and_needles were merged into "cleanup". To cleanup I added some functions that stitch meshes together, i.e. edges which share geometric end points are stitched. |
|
| 599 |
4628 d 1 h |
jab |
/trunk/src/ |
adding glew and a change for compatibility to ItemVector |
|
| 598 |
4628 d 12 h |
jrf |
/trunk/src/ |
Bug: CMake unable to generate project. Fix: Removing deleted source files from CMake lists. |
|
| 597 |
4628 d 13 h |
jab |
/ |
Some console stuff was not up to the right version |
|
| 596 |
4628 d 13 h |
jab |
/trunk/ |
merging |
|
| 595 |
4628 d 13 h |
jab |
/trunk/src/ |
Merged version |
|
| 594 |
4628 d 13 h |
jab |
/trunk/src/GLGraphics/ |
Resolved conflicts |
|
| 593 |
4628 d 14 h |
jab |
/branches/ctl/hmesh_vector/ |
Minor changes. An extra constructor for Vec3i and a function for creating a face in Manifold |
|
| 592 |
4630 d 2 h |
jab |
/branches/ctl/hmesh_vector/makefiles/ |
A Darwin 64 bit makefile. Does not work quite yet. |
|
| 591 |
4630 d 2 h |
jab |
/branches/ctl/hmesh_vector/ |
Various changes to clean up mostly |
|
| 590 |
4630 d 3 h |
jab |
/branches/ctl/hmesh_vector/makefiles/ |
Old build stuff |
|
| 589 |
4634 d 9 h |
jab |
/branches/ctl/hmesh_vector/src/HMesh/ |
After much deliberation, I changed it back so that ConnectivityKernel is now contained in Manifold - not an ancestor. |
|
| 588 |
4634 d 13 h |
jab |
/branches/ctl/hmesh_vector/ |
In a rather structural change, Manifold now inherits from ConnectivityKernel as opposed to containing it. What this means is that we now see the connectivityKernel as an abstract manifold. Methods for directly adding and removing vertices are protected and thus the user cannot see them. However, in a big change, the functions next, prev, opp, last, vert, which map mesh entities to other mesh entities (referenced by ID) are now exposed. This means that the Walker is no longer the sole means of getting to things in the mesh. I think this is an advantage, but I am not positive. Need to reflect before merging back into trunk. |
|
| 587 |
4635 d 13 h |
jab |
/branches/ctl/hmesh_vector/ |
A number of changes. Importantly
Vertex positions in Manifold are now stored as Vec3d
HalfEdgeWalker has been renamed Walker |
|
| 586 |
4636 d 0 h |
jab |
/branches/ctl/hmesh_vector/ |
Numerous changes. In particular, I changed the ItemID, ItemVector, and ConnectivityKernel in HMesh. This was done to make interfaces more clear.
Now ItemVector knows ItemIDs and conversion from ItemID to the actual integer ids happens in the ItemVector class and not connectivity kernel. |
|
| 585 |
4637 d 8 h |
jab |
/branches/ctl/hmesh_vector/ |
Committed many small changes. resolved conflicts. This commit merges in trunk. Now to merge back ... |
|
| 584 |
4640 d 15 h |
jab |
/branches/ctl/hmesh_vector/ |
Number of minor changes |
|
| 583 |
4648 d 5 h |
jab |
/branches/ctl/hmesh_vector/ |
now c++ 11ish |
|