Age | Commit message (Collapse) | Author | |
---|---|---|---|
2019-08-26 | Cloth: fix memory leaks from recent changesoc-2019-adaptive-cloth | ishbosamiya | |
2019-08-23 | Cloth: pinned vertices now works with adaptive remeshing | ishbosamiya | |
The customdata for newly created vertices is now modified to remove any weight for that vertex. Also the correct mesh is being passed to cloth_apply_vgroup(). | |||
2019-08-19 | Cloth: some code to better understand problems with pinning | ishbosamiya | |
2019-08-17 | Cloth: cache read check for adaptive cloth | ishbosamiya | |
2019-08-15 | Cloth: silence some print statements | ishbosamiya | |
2019-08-15 | Cloth: new method for collapsing an edge | ishbosamiya | |
2019-08-15 | Cloth: attempt at fixing extra face after collapsing edge | ishbosamiya | |
2019-08-13 | Cloth: debugging statements for finding bug in collapse edges | ishbosamiya | |
There is a bug in collapse edges which leaves 3 triangles for an edge in a special case. This might be an issue with the bmesh api. Now that the problem has been isolated, it can now be worked on. | |||
2019-08-12 | Cloth: revert back to using previous splitting method | ishbosamiya | |
The problem arises when a particular edge has more than 2 faces. This shouldn't ever happen. There must be problem with collapse edges or flip edges that introduces these extra number of faces for the edge. | |||
2019-08-10 | Cloth: attempt fix for rare error when not all faces remain as triangles | ishbosamiya | |
After spliting an edge, there are some instances when faces around the new_vert are not triangles. This attempts to fix this. | |||
2019-08-09 | Cloth: change intial values of the remeshing parameters | ishbosamiya | |
2019-08-09 | Cloth: UI to choose static or dynamic remeshing | ishbosamiya | |
2019-08-09 | Cloth: add support for turning adaptive remeshing on or off | ishbosamiya | |
2019-08-07 | Cloth: new attempt at sewing | ishbosamiya | |
Now it works for the most part but the mesh that is generated has some problems with the rendering. Need to look into this. The cloth doesn't seem to get the updated vertices. | |||
2019-08-07 | Cloth: attempt at fixing caching | ishbosamiya | |
clmd->mesh was being changed when it was returned as mesh_result, so now we are creating a local copy of the same. This is introduced another problem of the mesh not being correct when the playback is stopped. Need to look into this. | |||
2019-08-06 | Cloth: fix memory leak while removing vertex | ishbosamiya | |
ClothSizing was not being freed when the vertex was removed, now it is. | |||
2019-08-06 | Cloth: fix bug while removing vertex from cvm | ishbosamiya | |
2019-08-06 | Cloth: try storing localized mesh | ishbosamiya | |
2019-08-05 | Cloth: switch cloth flag to be part of clmd directly | ishbosamiya | |
This is important so that the flag is not reset every frame rather the information is actually stored. | |||
2019-08-05 | Cloth: Added flagging if previous frame was read from cache | ishbosamiya | |
It might be necessary to move this part to clmd itself instead of the clothObject. | |||
2019-08-04 | Cloth: Added comments for what needs to be added for the caching | ishbosamiya | |
2019-08-02 | Cloth: initial steps towards dynamic caching | ishbosamiya | |
2019-08-01 | Cloth: fix aspect ratio calculation | ishbosamiya | |
This does bring out the bugs in collapse edges, need to figure out where and how to fix them. | |||
2019-08-01 | Cloth: add back support for preserve verts and label edges | ishbosamiya | |
This is required for keeping the structure of the original mesh. | |||
2019-08-01 | Cloth: optimization: uv offset calculated only every frame | ishbosamiya | |
2019-08-01 | Cloth: better preverse vertex labeling and optimization: new way to update ↵ | ishbosamiya | |
active faces Massive speed up, because it no longer goes through the entire mesh every single time flip edges is called while updating the faces. Same optimization can be applied to updating active faces in collapse edges. | |||
2019-08-01 | Cloth: changes in matrices | ishbosamiya | |
Some matrices were transposed than what they should be leading to logical errors. There still might be more, but a lot of it has been fixed. | |||
2019-08-01 | Cloth: fix in eigen decomposition, b now refers to correct element of mat | ishbosamiya | |
2019-08-01 | Cloth: added refine obstacle metric parameter | ishbosamiya | |
2019-07-31 | Cloth: more debugging statements for face sizing | ishbosamiya | |
2019-07-31 | Cloth: fixed eigen decomposition floating point errors | ishbosamiya | |
This should generally work but needs further testing. Currently, the plan to check for floating point errors in compression metric. | |||
2019-07-31 | Cloth: multiply object matrix with bmesh vertices | ishbosamiya | |
2019-07-31 | Cloth: preserve count, labeled edges and improved debug statements | ishbosamiya | |
2019-07-31 | Cloth: find nearest planes only every frame for much faster excecution | ishbosamiya | |
2019-07-31 | Cloth: fix bug in eigen decomposition | ishbosamiya | |
2019-07-30 | Cloth: fix bug in invert_m2_m2() | ishbosamiya | |
2019-07-30 | Cloth: more face sizing debugging statements | ishbosamiya | |
2019-07-30 | Cloth: optimize obstacle metric by returning if planes is empty | ishbosamiya | |
2019-07-30 | Cloth: fix issue in compute_face_sizing, epsilon for invert_m2, face_sizing ↵ | ishbosamiya | |
debug statements Problem was that the det(facedm) was 0.0f or at least close to 0.0f which means inverse would lead to division by 0.0f error which resulted in -nan. | |||
2019-07-29 | Cloth: initial steps towards edge label support | ishbosamiya | |
2019-07-29 | Cloth: revert back to using bm_prev | ishbosamiya | |
This is important to ensure correct collisions and other stuff, maybe there is a more effictient way without bm_prev, need to check later. | |||
2019-07-29 | Cloth: prevent unnecessary copy of vertices | ishbosamiya | |
2019-07-26 | Cloth: intial steps towards preserving original mesh data | ishbosamiya | |
2019-07-26 | Cloth: different seam or boundary detection | ishbosamiya | |
2019-07-26 | Cloth: fix some bugs in a hacky way, must see again later | ishbosamiya | |
2019-07-26 | Cloth: switch to cpp style while finding bad_edges | ishbosamiya | |
2019-07-26 | Cloth: remove unnecessary code | ishbosamiya | |
2019-07-26 | Cloth: fix bug while reindexing vertices | ishbosamiya | |
Ensure it uses the ClothVertMap size while allocating memory for cloth->verts. | |||
2019-07-25 | Cloth: stop using bm_prev | ishbosamiya | |
2019-07-25 | Cloth: nearly completed transition to using ClothVertMap | ishbosamiya | |