January 2, 2015, 18:21 (GMT) |
Add request for HDR targets - all compositing is done in HDR targets now. |
January 2, 2015, 16:38 (GMT) |
Add 'dup_poin' callback param to new `BLI_duplicate_filelist()` func, to allow handling direntry's poin if needed. |
January 2, 2015, 16:24 (GMT) |
Merge branch 'master' into asset-experiments |
January 2, 2015, 16:12 (GMT) |
Disable bound textures just in case they come and bite us somewhere unpleasant. |
January 2, 2015, 15:57 (GMT) |
SSAO - use concentric mapping. Concentric mapping is a nice way to create a circular distribution from orthogonal 2D coordinates. Using this allows us to expose directly the number of samples to the user instead of quality options. Concentric mapping will likely be used for DOF sampling too later. References: [1] A Low Distortion Map Between Disk and Square - Peter Shirley, Kenneth Chiu [2] http://psgraphics.blogspot.nl/2011/01/improved-code-for-concentric- |
January 2, 2015, 14:44 (GMT) |
Merge remote-tracking branch 'refs/remotes/origin/master' into temp-text_editor_cursor_api |
Revision 5a51b93 by Martin Felke (fracture_modifier, fracture_modifier-master, temp-fracture-modifier-2.8) January 2, 2015, 14:15 (GMT) |
fix: restore kinematic state for regular triggered rigidbodies always now, they very likely were kinematic before, user must disable triggered checkbox if this is not desired |
January 2, 2015, 13:56 (GMT) |
Depsgraph: Fill in missing updateDepsgraph() callbacks for modifiers Doesn't mean it's 100% working, but preliminary checks for most of the modifiers seems to be fine. More intense testing is needed tho. |
January 2, 2015, 12:47 (GMT) |
Merge branch 'master' into viewport_experiments |
January 2, 2015, 12:38 (GMT) |
WIP incorrect inconvenient code that is supposed to draw PBVH colors. |
January 2, 2015, 12:37 (GMT) |
Depsgraph: Fix some stupid issues - Adding empty would crash. This is because of missing ob->data NULL check in id type tagging. - Fixed wrong assert statements in graph rebuild. - Silenced error about using char for array indices. Not sure why it's needed since chars are unsigned in blender any way. Could be some mismatch in C++/C flags, need to investigate further. |
January 2, 2015, 12:23 (GMT) |
Merge branch 'master' into depsgraph_refactor |
January 2, 2015, 08:05 (GMT) |
Depsgraph: Pole targets for IK constraints now work |
January 2, 2015, 04:13 (GMT) |
Depsgraph Fix: build_animdata() for relations was already called for obdata, so no need to repeat in build_rig() |
January 2, 2015, 04:09 (GMT) |
Depsgraph: Drivers on armature bones now have links to pose bone eval so that they cause updates |
January 2, 2015, 03:50 (GMT) |
Depsgraph: Fix for memory leak with unfreed links Now that relations are back in the base node, relations should be freed in the destructors for DepsNode instead of OperationDepsNode. This fixes memory leaks when using links between IDDepsNodes and also for TimeSourceDepsNodes |
January 2, 2015, 03:40 (GMT) |
Depsgraph ID Users: Order was inversed for the relationships added so far It's important to remember here that relationships here mean "B is USED by A", so doing the intuitive "A -> B" doesn't work out correctly |
January 2, 2015, 03:38 (GMT) |
Added "ID Users" depsgraph builder, for setting up links between ID datablocks This is useful for one-to-many situations when building relationships between operations, but is likely to come in handy for other parts of Blender later on. For example, when constructing drivers on armature bones, we don't have any easy way of referring back to the object (and thus the pose bones) to make those depend on these as well. Just a few skeleton types have been implemented now. |
January 2, 2015, 01:30 (GMT) |
Resolve TODO: We can check for specific components by using the type instead of using strings |
January 1, 2015, 16:00 (GMT) |
Merge branch 'master' into mesh-transfer-data |
|
|
|


Master Commits
MiikaHweb | 2003-2021